Building on fill - geotechnical aspects.pdf

BRE Garston, Watford, WD25 9XX

2001

Building on fill: geotechnical aspects Second edition

J A Charles and K SWatts

BRE Centre for Ground Engineering and Remediation

constructing the future

Prices for all available BRE publications can be obtained from: CRC Ltd 151 Rosebery Avenue London, EC1R 4GB Tel: 020 7505 6622 Fax: 020 7505 6606 email: [email protected]

BR 424 ISBN 1 86081 509 X

© Copyright BRE 1993, 2001 First published 1993 Second edition 2001

BRE is committed to providing impartial and authoritative information on all aspects of the built environment for clients, designers, contractors, engineers, manufacturers, occupants, etc. We make every effort to ensure the accuracy and quality of information and guidance when it is first published. However, we can take no responsibility for the subsequent use of this information, nor for any errors or omissions it may contain. Published by Construction Research Communications Ltd by permission of Building Research Establishment Ltd Requests to copy any part of this publication should be made to: CRC Ltd PO Box 202 Watford, WD25 9ZW

BRE material is also published quarterly on CD

Construction Research Communications

Each CD contains BRE material published in the current year, including reports, specialist reports, and the Professional Development publications: Digests, Good Building Guides, Good Repair Guides and Information Papers.

CRC supplies a wide range ofbuilding and construction related information products from BRE and other highly respected organisations.

The CD collection gives you the opportunity to build a comprehensive library of BRE material at a fraction of the cost of printed copies. As a subscriber you also benefit from a 25% discount on other BRE titles. For more information contact: CRC Customer Services on 020 7505 6622

Contact: post: CRC Ltd 151 Rosebery Avenue London, EC1R 4GB fax: phone: email: website:

020 7505 6606 020 7505 6622 crc@co nstruct.emap.co.uk www.constructionplus.co.uk

iii

Contents

Foreword to the first edition

vi

Foreword to the second edition

vii

Acknowledgements

viii

Notation

ix

Abbreviations

xii

Glossary

xiii

Part I: Fills in context Chapter 1

1.2 1.3 1.4 1.5

Chapter 2

Brownfield sites Definitions Scope Research BRE at

Fill formation and deposits 2.1 Opencast mining backfill 2.2 Colliery spoil 2.3 Pulverised fuel ash 2.4 Industrialandchemicalwastes 2.5 Urban fill 2.6D omestic refuse 2.7 Infilleddocks,pitsandquarries 2.8 Hydraulic fill

Part II: Engineering behaviour of fills Chapter 3

1 2

Introduction 1.1 Historical background

Properties of fills 3.1 Characteristicsoffilldeposits 3.2 Indexandclassificationproperties 3.3 Compactness 3.4 Stiffnessandcompressibility 3.5 Shear strength 3.6 Dynamic properties 3.7 Permeability

2 3 5 5 6 7 7 9 10 11 12 12 15 15 17 19 20 21 23 25 30 32 34 (continued)

iv Chapter 4

Chapter 5

4.1S elf-weight of fill

37

4.2 4.3 4.4 4.5 4.6 4.7D

39 42 43 43 45 46

Weight of buildings Changeinground-waterlevel Changeinmoisturecontent Decompositionofbiodegradablefill Chemical reactions ynamic loading

Collapse c ompression on we tting 5.1 Mechanismsofcollapsecompression 5.2 Laboratoryinvestigations 5.3 Field investigations 5.4 Magnitude ofcollapsepotential 5.5T ime dependency 5.6 Identificationofcollapsepotential 5.7 Buildingsdamagedbycollapsecompression

Chapter 6

36

Volume changes in fills

Boundaries and variable depth 6.1 Influenceoffillproperties 6.2 Influenceoffillgeometry 6.3 Vertical highwall 6.4B uried highwall 6.5 Long shallow slope 6.6E xclusion zones

48 48 49 49 50 54 54 55 57 57 58 59 60 60 61

Part III: Construction on fills

63 65 66 66 66 67 67 68 69 70

Chapter 7

Investigation and monitoring 7.1 Historical review 7.2 Site reconnaissance 7.3 Ground investigation 7.4 Laboratory tests 7.5 In-situ tests 7.6 Load tests 7.7 Geophysical tests 7.8 Monitoring

Chapter 8

Treatment of fills 8.1 Dynamic compaction 8.2 Vibro techniques 8.3 Preloading 8.4P re-inundation 8.5 Other methods

72 73 78 84 87 89

Chapter 9

Engineered fill 9.1 Types of specification

93

9.2 9.3F 9.4 9.5 9.6 9.7

94 94 95 95 97 98

Chapter 10

Site investigation categories ill End product criteria Sitepreparationandfillplacement Quality management Excavationandrecompaction

Foundations on fills 10.1 Differentialmovement 10.2 Classificationofpotentialmovement 10.3 Shallow foundations 10.4 Deep foundations 10.5 Implicationsofgroundchemistry

92

100 100 102 103 104 105

v

Part IV: Performance of fills Case histories 1 OpencastminingbackfillatCorby(A) 2 OpencastminingbackfillatCorby(B) 3 OpencastminingbackfillatCorby(C) 4 Opencast mining backfill at Horsley, Northumberland 5 OpencastminingbackfillatIlkeston 6 OpencastminingbackfillatTamworth 7 OpencastminingbackfillatWestAuckland 8 OpencastminingbackfillnearEdinburgh 9 CollieryspoilatCoalville(A) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

107 107 108 109 114 115 120 122 124 125 126

CollieryspoilatCoalville(B) 128 Colliery spoil at Methil 130 LagoonpfaatPeterborough 132 Slag bank at Hartlepool 134 Urban fill at Greenwich 136 UrbanfillatManchester(A) 138 OlddomesticrefuseatRedditch(A) 139 OldrefuseintheeastendofLondon 142 Old domestic refuse at trunk road widening in Hertfordshire 144 OlddomesticrefuseatRedditch(B) 146 OldrefuseatLiverpool(A) 148 Recent domestic refuse landfill at Brogborough 151 RecentdomesticrefuselandfillatCalvert 154 Infilled dock at Hull 158 InfilleddockatLiverpool(B) 159 ClayfillinformergravelpitatAbingdon 161 BuildingwastesatWaterbeach 162 AlluvialsanddepositatManchester(B) 164 Urban fill Bacup at 165

Appendices 167 Appendix A Stress distribution below building foundations 167 Appendix B Settlement of foundations calculated using elastic theory 168 Appendix C Delineation of an exclusion zone over a highwall 169 Appendix D Effectiveness of field compaction by impact loading 173 Appendix E Analysis of stone columns under widespread load 175 AppendixF Modelspecificationforengineeredfill 178 References

181

vi

Foreword to the first edition

One result of the scarcity and cost of good building land is that building development increasingly takes place on sites where there are deep deposits of waste fills. As these fills have considerable economic significance for land values, it may seem surprising that until recently they received relatively little attention from geotechnical engineers. This was not because of an absence of problems; many of the fills are poorly compacted and variable, and their behaviour as foundation materials may be unsatisfactory. However, the heterogeneous nature of many waste fills makes characterisation and analysis difficult, and it is easy to understand why the attention of geotechnical engineers has usually been focused on more promising and better behaved natural soils. Research at the Building Research Esta blishment (BRE) has attempted to redress the neglect by:  monitoring field performance at a large number of filled sites with emphasis on long-term observations of settlement; characterising fills on the basis of observed field performance, to assist in the selection of appropriate foundation solutions; and  assessing the effectiveness of various ground-treatment techniques, by field observations at selected sites. 

This report provides a detailed account of BRE research findings and their significance for appropriate and successful building developments on fill. Part A deals with the engineering behaviour of fills, and Part B examines construction on fills. Brief case histories of field performance are presented in Part C; these mostly describe sites where BRE has made measurements of fill behaviour, but some additional case histories, in which monitoring has be en carried out by other parties, are included where necessary to give a more complete picture. (Parts A and B make extensive use of these records.) While the report describes experience with fills in the United Kingdom, it has relevance to similar materials found in many other parts of the world. Field monitoring has shown that in most situations the fill settlement that damages buildings has causes other than the weight of the building. This means that the concept of bearing capacity is not adequate to define the loadcarrying characterist ics of many fills. Settlements caused by other physical factors, and in some cases by chemical or biological processes, need to be assessed. A particular hazard for poorly compacted partially saturated fills is a reduction in volume which can occur when the fill is first inundated with water.

vii

Foreword to the second edition

In the eight years since the first edition was published, the term ‘brownfield’ has come into everyday use and the importance of locating building developments on such sites has been widely accepted. A precise definition of brownfield has yet to find universal agreement, but the basic concept of land adversely affected by previous human activity is clear. The sustainab ility agenda requires the long-term productive re-use of brownfield land. The problem is that previous usage may have left a wide range of physical, chemical and biological hazards. Three systems whic h may be at risk in brownfield developments can be identified: the human population, the natural environment and the built environment. Physical problems may include buried foundations and settlement of filled ground. The range of problems associated with chemical contamination is vast and can present an immediate or long- term threat to human health, to plants, to amenity, to construction operations and to buildings and services. Biodegradation of organic matter may lead to the generation of gas. The objective is to build safe, durable and economic structures. The site and the building development form an interactive system and it is important to evaluate the risk of adverse interactions during the lifetime of the development. For many years the redevelopment of derelict land and brownfield sites has been dominated by the hazards associat ed with contamination and the risks posed to human health. The physical problems have received less attention and it is hoped that this book will help to redress the balance. Although brownfield land is a world-wide phenomenon, the issues are particularly acute for Great Britain, a heavily populated island with a long industrial history. The scale of the problem was illustrated by the size of the £1 billion plus package which was announced in 1996 for the regeneration of major coalfields. Some 91 0 ha of land were to be reclaimed for residential, commercial and retail uses. Many of these sites will involve building on colliery spoil. This example illustrates how the rede velopment of brownfield sites is closely linked to building on fill, the subject of this book. This second edition of Building on fill: geotechnical aspects updates and expands the first edition which was published in 1993. Three new chapters have been added covering, re spectively, collapse compression on wetting, problems associated with a variable depth of fill, and engineered fill. Records of BRE field monitoring have been brought up to date. The book has been reorganised into four parts and five appendices have been added.

viii

Acknowledgements

The preparation of this fully revised version of the book has been carried out under the DETR Sustainability Business Plan. The approach to the specification on enginee red fill, described in Chapter 9 and Appendix F, was developed as part of a study carried out by Halcrow for BRE, and the major contribution made by Mr Neil T renter is gratefully acknowledged. The assistance of a number of colleagues is also gratefully acknowledged: Miss Hilary Skinner made a substantial contribu tion to the work on boundary effects and variations in depth of fill, described in Chapter 6 and Appendix C. Dr Paul Tedd has commented on the sections involving ground chemistry, contamination and landfill gas. Mr A P Butcher commented on the section on dynamic properties. Mr R M C Driscoll commented on a complete draft of the book.

Acknowledgements to the first edition The research described in this report has been carried out at B RE over the last 20 years, and a number of colleagues have been involved. Most of the field measurements of settlement have been made by D Burford and K S Watts. Geophysical measurements of soil properties have been made by Dr C P Abbiss, who also commented on the sections of the report dealing with dynamic properties. J J M Powell commented on the chapter on investigation and monitoring of fills. Dr R J Collins gave some assistance on chemical aspects of fills. R M C Driscoll and Dr A D M Penman made helpful comments on a complete draft of the report. On a number of sites, observations have been made in collaboration with other organisations, and this is acknowledged in the relevant case histories.

ix

Notation

a a b b b cu cv d e fs fν fd g

h ki k mv mα n n p q q qu s sn sM smax si sα s r

su t v w wopt wP wL x xE

length of foundation coefficient in various empirical relationships width of foundation, footing or load test diameter of footing or weight coefficient in various empirical relationships undrained shear strength coefficient of consolidation depth of foundation voids ratio shape and rigidity factor Poisson’s ratio factor depth factor acceleration due to gravity g = 9.81 m/s height of sample hydraulic gradient coefficient of permeability or hydraulic conductivity coefficient in various empirical relationships coefficient of compressibility ratio mα = sα/si porosity (%) number of impacts at any point in dynamic compaction line load applied pressure rate of flow compressive strength settlement normalised settlement sn = s/sM maximum settlement maximum settlement immediate settlement logarithmic creep settlement rate settlement reduction factor for stone columns ε = s ε where εv is vertical compression in treated groundv andr voεvo is vertical compression in untreated ground ( sr = 1 indicates that no improvement in stiffness has resulted from ground treatment) undrained settlement time discharge velocity v = (q/A) moisture content (%) optimum moisture content (%) plastic limit (%) liquid limit (%) horizontal distance horizontal distance from top of highwall to beginning of exclusion zone

x

Notation xEn z ze

normalised horizontal distanc e from top of highwall to beginning of exclusion zone xEn = xE /H depth below ground level or foundation level depth of compacted fill in dynamic compaction, or depth significantly stressed by applied load

A A Ac Ar At C

area of sample area of impact in dynamic compaction area of stone columns area replacement factor for stone columns Ac = Ar At total area of treatment coefficient in Hazen’s permeability relationship k = C D102

CR

relative compaction; percentage ratio of in-situ dry density to maximum dry density in specified laboratory compaction test coefficient of uniformity CU = D60/D10 coefficient of secondary consolidation Cα = ∆h/(h log[ t2/t1]) constrained modulus secant constrained modulus normalised secant constrained modulus tangent constrained modulus constant equivalent constrained modulus particle size such that x % by mass is finer depth of burial of highwall normalised depth of burial of highwall Dn = D/H damping ratio Young’s modulus total applied energy in dynamic compaction fines content; percentage of silt and clay size particles shear modulus

CU Cα D Dsec Dsecn Dtan D* Dx D Dn Dr E E Fc G Gdyn H H H ID IL IP IS Ko L M N N N P Q R S r

V Va VP VR VS W W Wn WE WEn

dynamic modulus height of shear fill in embankment height of highwall height of fall of weight in dynamic compaction density index (relative density) liquidity index IL = wL – wP plasticity index IP = (w – wP)/IL index for strength of rock (also I1 and I2, see section 3.2) coefficient of earth pressure at rest length of wall mass of weight used in dynamic compaction SPT blow count total number of impacts in dynamic compaction number of blows in MCV test point load quality factor principal effective stress ratio R = σ'1/σ'3 degree of saturation of a soil; ratio of volume of water to volume of pores (%) volume air voids (%) compression wave velocity Rayleigh wave velocity shear wave velocity weight in dynamic compaction width of ground over highwall where tilt is not zero normalised width of ground over highwall where tilt is not zero Wn = W/H width of exclusion zone over highwall normalised width of exclusion zone over highwall WEn = WE/H

Notation

β δ ε εv εvo γ

tilt logarithmic creep compression rate parameter logarithmic compressi on rate parameter describing reduction in volume due to biodegradation in biodegradable fill logarithmic compressi on rate parameter describing reduction in volume due to physical creep in biodegradable fill angle of highwall to horizontal angle of friction at fill/pile interface or fill/highwall interface strain vertical strain vertical strain induced in untreated fill limit angle over highwall

γ γd γw ν φ' φ'cv ρ ρd ρdmax ρdmin ρs ρw σ σv σ' σ' 1 σ' 3

unit weight dry unit weight unit weight of water Poisson’s ratio effective angle of shearing resistance constant volume angle of shearing resistance bulk density dry density maximum dry density minimum dry density particle density density of water ρw = 1.0 Mg/m3 stress vertical stress effective stress major principal effective stress minor principal effective stress

∆

relative deflection — maximum vertical displacement relative to the straight line connecting two points

α α αb αc

xi

xii

Abbreviations

AOD ASCE ASTM BRE BS BSI CBR CIRIA CPT CSW DD DMT DP EN ENV

Above Ordnance Datum American Society of Civil Engineers American Society for Testing Materials Building Research Establishment BritishStandard British Standards Institution Californiabearingratio Construction Industry Research and Information Association cone penetration test continuous surface wave method Draft for Development flat dilatometer test dynamicprobing European Standard European Pre-standard

ICRCL

Interdepartmental Contaminated LandCommittee on the Redevelopment of moisture condition value Menard pressuremeter National Coal Board (subsequently British Coal) National House Building Council OrdnanceDatum pulverised fuel ash pressuremeter spectral analysis of surface waves self-boring pressuremeter standard penetration test Transport Research Laboratory UnitedKingdom United States of America

MCV MPM NCB NHBC OD pfa PMT SASW SBP SPT TRL UK USA

xiii

Glossary

allowable bearing pressure

angular distortion backfill band drain

bioconsolidation biodegradation brownfield land coarse soils collapse compression

colliery spoil compaction cone penetrometer cone pressuremeter consolidation

constrained modulus contaminated land

creep compression cross-hole tomography deflection ratio

The maximum allowable net loading intensity at the base of the foundation, taking into account the ultimate bearing capacity and required margin against failure, the amount and kind of settlement expected and the ability of the structure to accommodate this settlement. The ratio of the differential settlement ( δs) between two points and the distance ( L) between them, relative to the tilt. Material that has been used to fill an excavation or placed behind a retaining wall. A type of prefabricated vertical drain usually consisting of a plastic core surrounded by a geotextile sleeve; the core provides a flow path along the drain and supports the sleeve which in turn acts as a filter separating the core and its flow channels from the soil. Reduction in volume of domestic refuse fills due to biodegradation . Biological degradation; much of the organic constituents of domestic refusethat fills is biodegradable. Land has been previously developed, including derelict land. Gravels and sands are coarse soils; the term is applied to soils with more than about 65% of sand and gravel sizes. In this phenomenon a partially saturated soil undergoes a reduction in volume that is attributable to an increase in moisture content without there necessarily being any increase in applied stress. Waste from the deep mining of coal. The process of densifying soils by some mechanical means such as rolling, ramming or vibration to reduce the volume of voids. This in-situ testing device comprises a cone, a friction sleeve, any other sensors and measuring systems, as well as connections to push rods. This in-situ testing device consists of a pressuremeter module mounted behind a cone pene trometer. The process of densifying soils by increasing the effective stress using some form of static loading; consolidation of a saturated clay soil is a time-dependent process which results from the slow expulsion of water from the soil pores. The ratio of vertical stress to vertical strain in confined compression. Land that contains substances that, when present in sufficient quantities or concentrations, are likely to cause harm, directly or indirectly, to man, to the natural environment or to the built environment. Compression which occurs under constant effective stress. This common form of seismic velocity tomography is carried out from two boreholes. The maximum vertical displacement (∆) relative to the straight line connecting two points divided by the length be tween the two points (L).

xiv

Glossary density index

derelict land

dynamic compaction

earthfill engineered fill exclusion zone fill

fine soils foundation foundation fill geogrid

geosynthetic

ground treatment hardcore hazard heave highwall hogging kentledge

landfill limit angle

liquefaction low-rise buildings made ground magnet extensometer

The degree of packing in coarse soils, such as sand and gravel, can be described by the density index, ID, which relates the in-situ density of a granular fill to the limiting conditions of maximum density and minimum density; sometimes known as relative density . Land that has been so damaged by industry, mining and urban development that it can no longer be put to bene ficial use without treatment. A ground treatment method in which deep compaction is effected by repeatedly dropping a heavy weight onto the ground surface from a great height. Fill composed of natural soil materials such as sand, silt and clay. Fill which is selected, placed and compacted to an appropriate specification, so it will e xhibit the required engineering behaviour. An area of ground where building development is not permitted because of perceived hazard such as excessive differential settlement. Ground that has been formed by material deposited through human activity rather than geological processes; it is sometimes referred to as made ground. Clays and silts are fine soils; the term is applied to soils with more than about 35% of silt and clay sizes. That part of a structure designed and constructed to be in direct contact with, and transmitting loads to, the ground. Fill material that supports a building or other type of structure. A type of geosynthetic with a planar structure formed by a ne twork of tensile elements with apertures of sufficient size to allow interlocking with the surrounding ground. A generic term for civil e ngineering materials such as geotextiles, geogrids, geomembranes and geocomposites that are used to modify or improve ground behaviour. The controlled alteration state,an nature or mass behaviour of ground materials in order of tothe achieve intended satisfactory response to existing or projected environmental and engineering actions. A limited amount of selected fill put down as infill within the foundations of a building unit or beneath an oversite concrete slab. A situation which in certain circumstanc es could lead to harm to the human population, the built environment or the natural environment. Upward displacement of the ground. A steep excavation slope at the deep end of an opencast mining excavation. The mode of deformation of a foundation or beam undergoing upward bending — the opposite of sagging . A form of dead-weight loading providing a reaction over a jack or directly loading a plate in a large-scale load test; it may be concrete blocks, scrap metal, containers filled with sand or water, or any other convenient material. This expression is often used specifically to describe domestic refuse. The angle measured from the horizonta l defining the extent of the ground affected by differential settlement over a highwall. It is similar to the angle used to de fine the subsidence trough caused by tunnelling or mining. In this phenomenon a saturated sandy soil loses shear strength due to an increase in pore pressure. Buildings not more than three storeys in height. See fill. A device which can be installed in a borehole to monitor settlement at depth; it consists of a series of ring magnets anchored to the borehole wall and connected by plastics tubing; the position of the magnets can be located using a reed switch sensor.

Glossary maximum dry density

non-engineered fill

observational method

opencast mining optimum moisture content organic content pad foundation particle density penetration test

preloading primary consolidation

pulverised fuel ash

raft foundation rapid impact compaction

relative compaction

relative density risk rockfill sagging secondary compression seismic wave serviceability limit states settlement

xv

The maximum dry density obtained using some specified compaction procedure. With a fine fill, there will be an optimum moisture content at which this maximum density is achieved. The maximum dry density and optimum moisture content of a soil refer to a particular compaction procedure and can be misleading if taken out of the context of that procedure. Fill which has arisen as a by-product of human activity, usually involving the disposal of waste materials; it has not been placed with a subsequent engineering application in view. In ground treatment, the observational method is a managed and integrated process of design, treatment control, monitoring and review that enables previously defined modifications to be incorporated during or after treatment as appropriate; the aim is to achieve greater economy without compromising technical adequacy. Mining carried out by excavation from the ground surface. The moisture content at which a maximum dry density is achieved using some specified compactive procedure. The average organic mat ter content of a soil expressed as a percentage of the srcinal dry weight. A foundation usually provided to support a structural column, consisting of a simple circular, square or rectangular slab. The mass per unit volume of the solid particles of a soil; sometimes termed specific gravity. An in-situ test in which a device is pushed or driven into the ground while the resistance of the soil to penetration is recorded (eg standard penetration test, cone penetration test). A ground treatment method in which consolidation is achieved by temporarily loading the ground usually with a surcharge of fill. The reduction in volume of a fine soil caused by the expulsion of water from the to soilthe pores transfer of load from the excess pore water pressure soil and particles. A waste product from power stations which burn pulverised coal composed of nearly spherical particles predominantly in the coarse sand and fine silt range. A foundation continuous in two directions, usually covering an area equal to or greater than the base area of the structure. A ground treatment method in which compaction of the ground is effected by repeated impacts of a weight onto a plate which remains in contact with the ground surface. The ratio of the in-situ dry density to the maximum dry density achieved with a specified degree of compaction in a standard laboratory compaction test. See density index . The likelihood that a particular adverse event occurs during a specified period of time. A fill produced by blasting or ripping rock strata which may contain large rock fragments. The mode of deformation of a foundation or beam undergoing downward bending — the opposite of hogging . The reduction in volume of a fine soil caused by the adjustment of the soil structure after primary consolida tion has been completed . Seismic waves transmit energy by the vibration of soil particles; they can be generated by some form of impact or vibration source. These states correspond to conditions beyond which specified service criteria for a structure or structural element are no longer met . Downward movement of a structure and its foundation resulting from movement of the ground below it; total se ttlement of a structure may interfere with some aspect of its functions, such as connections to services, but it is differential settlement that causes structural damage.

xvi

Glossary specific gravity stone columns strip foundation tilt ultimate bearing capacit y

ultimate limit states uniformly graded soil vibro

well graded soil

See particle density . Method of ground treatment in which poor ground is stiffened by the installation of columns of stone. A foundation normally provided for a loadbearing wall. Rigid body rotation of whole structure. The value of the gross loading intensity for a particular foundation at which the resistance of the soil to displacement of the foundation is fully mobilised. These states are associated with collapse or other similar forms of structural failure. The soil has a majority of soil grains which are very nearly the same size. A generic term for deep vibratory ground treatment achieved by penetration into the ground of a large vibrating poker; usually the cylindrical hole formed by the vibrator is backfilled in stages with stone, forming stone columns . The soil has a wide and even distribution of particle sizes.

1

Part I: Fills in context The industrial revolution and the creation of parks around the country houses have taken us down to the later years of the nineteenth century. Since that time, and especially since the year 1914, every single change in the English landscape has either uglified it or destroyed its meaning, or both. Of all the changes in the last two generations, only the great reservoirs of water for the industrial cities of the North and Midlands have added anything to the scene that one can contemplate without pain. (Hoskins, 1981)

The long industrial history of Great Britain has le d to this small, heavily populated island having a large proportion of its land surface affected by human activity. Many would agree with Professor Hoskins’ view, expressed so trenchantly in his The making of the English landscape, that these human activities have generally had an adverse effect on the environment, and, it might be added, not only in England, but also in Scotland and Wales. Since the early 1970s there has been increasing concern over this situation. In the preface to his book, Derelict land, Kenneth Wallwork reports the Secretary of State for the Environment saying in 1971 that: ‘The scars left behind by industrial development of the past, the abandoned waste heaps, disused excavations and derelict installations and buildings no longer needed by industry, are an affront to our concept of an acceptable e nvironment in the 197 0s.’

The reclamation of derelict land became an important aspect of government policy. More recently there has beenthan an e greenfield mphasis onsites siting new building developments on brownfield rather and many of these brownfield sites are covered in substantial depths of fill. The subject of building on fill has thus acquired considerable prominence in recent years, although it should be remembered that the practice of building on fill can be traced back to antiquity. The geotechnical problems in achieving safe and e conomic developments on filled ground are substantial. Before examining these, the two chapters in Part I of this book present the background context of building on fills. The first chapter provides some historical background and defines the scope of the book. The engineering behaviour of a fill is strongly influenced by the method of deposition and subsequent stress history; it is pertinent, therefore, in Chapter 2 to examine the srcins of the principal types of fill encountered in the UK.

2

Chapter 1 Introduction All made ground should be treated as suspect because of the likelihood of extreme variability. (British Standard Code of Practice for Foundations, BS 8004:1986)

The term ‘fill’ is used to describe ground that has been formed by material deposited by man. Thus fill, or ‘made ground’ as it is sometimes called, results from human activity, in contrast to natural soil which has its srcin in geological processes. There is a great need for building land at reasonable cost within and adjacent to built-up areas. Shortage of land and a long industrial history ensure that much of the land now being used for building purposes in the UK has been previously affected by human activities which may cause serious problems during redevelopment. Despite the warning from BS 8004:1986 quoted above, many b uildings have been and are being founded on fill. The problems occasioned by building on fill have to be e valuated against a

placed to achieve a particular objective such as reclaiming low-lying marshy land or providing a suitable elevation for defence. A new Flavian city was erected at Chichester, over the remains of the first Roman city, on a 1 m deep platform of rammed gravel (Carver, 198 7). Although fills accumulated in urban locations during both Roman times and the Middle Ages, it was with the coming of the Industrial Revolution that man’s capacity to generate waste materials, and to cover significant portions of the e arth’s surface with them, greatly increased. Large areas of land have been used for the deposition of mining, industrial, chemical, building, dredging, commercial and domestic wastes. In a country with a long industrial history, much of the land used for

background ofissues, growing concern over sustainability which means thatenvironmental the beneficialand reuse of brownfield sites and environmental protection to prevent further land being damaged are of increasing importance.

building development will have a history of previous In recent years large-scale opencast mining has leftuses. great depths of fills. Many of these sites are close to centres of population where building developments may be proposed. This situation is not confined to the U K. Lange (1986) reported that 1200 residential buildings and farms have been established on the dee p uncompacted backfills in the Rhenish brown coal area of Ge rmany. In urban redevelopment programmes, old buildings are demolished and new buildings have to be built over infilled basements and on the rubble of the de molished buildings. Although this type of redevelopment has continued throughout recorded history, in modern urban redevelopment programmes it is carried out at a rate and on a scale not seen before. Low-lying wet areas have been reclaimed by filling. Again, this type of filled ground has been formed throughout history and is found in many parts of the world. Bordering the Baltic Sea, reclamation of the lowlying marshes on which St Petersburg is built, began in 1703. Rutledge (1970) commented that it was instructive to note how much of downtown Manhattan Island was constructed on filled land created be fore 1900. In contrast with waste and demolition fills, large quantities of fill materials are placed as part of carefully controlled civil engineering works. For many years these engineered fills have been placed to form embankment dams and highway embankments. With increasing frequency they are now placed specifically as foundation material for new buildings.

1.1 Historical background In urban locations where there has been continuous occupation of the land for centuries, there are likely to be large areas of filled ground. Fills may have arisen inadvertently from the rubble of demolished buildings and the slow accumulation of refuse. Old urban fills of these types may contain soil, rubble and refuse. They can be quite extensive in area but usually are relatively shallow. They may be very old. Striking examples of the unplanned accumulation of fills in inhabited areas are provided by many towns in the Middle East. The most common building material was mud brick, and walls of mud brick have to be thick. New construction took place on the ruins of the old, and in Syria and Iraq villages stand on mounds of their own making. The ruins of an ancient city may rise 30 m above the surrounding plain. This gradual rise of debris has bee n much less common in Great Britain, although in some situations deep fills have accumulated. By the third century AD the Wallbrook in the City of London was already half buried, and mosaic pavements of Roman London lie 8 m to 9 m below the streets of the modern city. Some fills were

1.2 Brownfield sites

3

Serious problems have occurred when building on fill. Occasionally prestigious buildings are involved and detailed investigations are carried out and described in the technical press. The settlement of the Royal Scottish Academy in Edinburgh provides such a case (Masters, 2000). The building was completed in 1826 on the Mound which had been formed in the late 1700s using clay spoil from the construction of the New Town. The building was founded on square timber piles which in the course of time rotted, leaving large voids under the stone footings. Carefully monitored remedial works involving compensation grouting were carried out in 2000. Many of the problems which have been reported in the technical press and daily newspapers refer to houses and other small buildings. In these cases, the reports usually have insufficient reliable technical information for a proper judgement to be formed on the cause of the particular problems at that site and just refer to a ‘filled- in tip’ or ‘rubbish dump’. However, they give an indication of how these situations are popularly perceived, as the following examples indicate:  ‘Based on a sound Victorian foundation’ ( Hardware Trade Journal, 27.2.75)  ‘Rubbish tips root of house problem’ ( Construction News, 28.10.76)  ‘Firm to buy back faulty homes’ (Daily Telegraph, 11.12.76)  ‘NHBC seeks re medy for “blighted” estate’ ( New Civil Engineer, 25.1.79)

Figure 1 Houses built on deep opencast mining backfill

the Department of the Environment had the complementary aims of reclaiming the dere liction inherited from the industrial past and securing, by planning control and other means, that we do not create fresh areas which are left without effective treatment. The first official survey of derelict land in England and Wales took place in 1954 and recorded 51 274 ha of dereliction (Ministry of Housing and Local Government, 1956). In commenting on this and subsequent surveys, Wallwork (1974) pointed out that at that time derelict land continued to grow more rapidly than reclamation could restore earlier dereliction to beneficial use. Although in the early 1970s an annual reclamation rate of



‘ “Egham experiment” sinks to failure’ ( Construction News, 28.6.79) ‘Rise and fall of a real Reggie Perrin’ ( Manchester Evening News, 22.12.79)  ‘Landslip homes forcing out the old’ ( Sheffield Morning Telegraph, 7.1.83)  ‘Sinking suburb’ (New Civil Engineer, 31.10.85)  ‘Tortuous tort’ ( New Civil Engineer, 2.8.90)  ‘£1m to shore up sinking homes’ ( Watford and Rickmansworth Review, 28.11.91) 

Although such reports appear from time to time and rightly give cause for concern, they should not be allowed to obscure the fact that many buildings have been successfully built on fill. Figure 1 shows low-rise housing built on a deep opencast mining backfill. Problems and failures on filled ground emphasise the importance of developing an adequate understanding of the behaviour of fills, and of identifying potential hazards so that appropriate types of building development can be successfully undertaken on suitable fill sites.

1.2 Brownfield sites While the history of building on fill can be traced back to antiquity, the relatively recent concentration on environmental and sustainability issues has given it an increased importance. In opening the Land Reclamation Conference at Grays, Essex, in October 1976, the Parliamentary Under Secretary of State emphasised that

2200 ha was reached, this has surveys not be en wholly (Barnett, 1976). In successive of de relictsustained land in England there were found to be 43 000 ha in 1974, 45 600 ha in 1982, 40 500 ha in 1988 and 39 600 ha in 1993. Thus the area of derelict land in England decreased by 8% in the 20 years between 1974 and 1993 (Parliamentary Office of Science and Technology, 1998). However, the difficulties in developing an appropriate definition of derelict land we re bound to reduce the value of such reclamation statistics (Wallwork, 1974). Of the 14 000 ha of derelict land reclaim ed between 1982 and 1988, 27% were for hard uses. In recent years the te rm ‘brownfield’ has come into use. Not surprisingly, it has proved just as difficult to obtain an acceptable and agreed definition of brownfield as it was to define dereliction (Alker et al , 2000). The Parliamentary Office of Science and Technology (1998) commented that because there is no agreed definition, many practitioners regard brownfield land as any land that has been previously developed. In a parliamentary answer, the Construction Minister was reported (Raynsford, 1998) as saying that: ‘There is, as yet, no specific definition of brownfield land, but it is usually taken to mean land previously used for urban uses…these include residential, transport and utilities, industry and commerce, community services, previously developed vacant land and derelict land.’

In November 1996 a target was proposed that 60% of ne w housing should be built on brownfield land. However,

4

Chapter 1 Introduction

previous usage may have left a wide range of physical, chemical and biological hazards. Physical problems may include buried foundations and settlement of filled ground. The range of problems associated with chemical contamination is vast, and chemical contamination can present an immediate or long- term threat to human health, to plants, to amenity, to construction operations and to buildings and services. Biodegradation of organic matter may lead to the generation of gas. The risks that are posed by these hazards need to be evaluated. One unfortunate side effect of evaluating risk is that there is, of necessity, a concentration on problems and potentially

various aspects of the natural environment, let alone the built environment. Where building development is proposed, risk assessment should consider a broad spectrum of potential targets which may be vulnerable to different hazards. Since brownfield sites are frequently contaminated, risk assessment should include health concerns, and from the early stages of investigation through to the final use of the site a range of people may be at risk. Where low-rise housing is built on the land, the occupiers are likely to be the people most at risk from many of the hazards. Society is increasingly risk averse and matters concerning health

negative effects. It needs to be emphasised that the redevelopment of brownfield sites can have major advantages and, while it is important to identify hazards and to evaluate risks, the benefits should not be overlooked. Three systems which may be at risk in brownfield developments can be identified:  the human population  the natural environment  the built environment.

and safety rightly receive much attention and publicity (Bevan and Hind, 1999). Threats to the natural environment are often relatively narrowly conceived to concern soil and ground-water contamination. However, a much wider range of issues may need to be examined against a background of growing concern over degradation of the natural environment and the increasing prominence of environmental pressure groups promoting concepts such as biodiversity. This whole area is diffuse and difficult to evaluate as, unlike matters of human health and building damage, there are no widely agreed objectives or ground rules. It is not easy to see how an appropriate balance can be obtained between economic well-being and the natural environment. Furthermore, proposals to reclaim derelict land may now run into serious difficulties with opposition from those who consider that ecosystems

These three systems are, of course, interdependent and the issues need to be examined against a background of growing concern over degradation of the natural environment and the increasing prominence of the concept of sustainable construction involving the rate of use of renewable and non-renewable resources, and the emission of pollutants. 



‘Sustainable development meets the needs of the present without compromisin g the ability of future generations to meet their own needs’ (World Commission on Environment and Development [Brundtland Commission], 198 7) ‘Sustainable development also involves a changing conception of the relationship between the natural environment and mankind’ (Institution of Civil Engineers, 1996a)

The sustainability agenda requires the long-term productive re-use of brownfield land. There is a great need for building land at re asonable cost within and adjacent to built-up areas. Furthermore, there is strong opposition to the urbanisation of the countryside. The growing concern over environmental issues means that reclamation of derelict land for beneficial use, and environmental protection to prevent further land being damaged, are of increasing importance. Some of the issues involved in sustainability and acceptability in infrastructure development have been reviewed by the Institution of Civil Engineers (1996b). For many years the redevelopment of derelict land and brownfield sites has been dominated by the hazards associated with contamination and the risks posed to human health. For example, in a paper on risk assessment and soil contamination it was affirmed that: ‘Risk assessment is appraising the possibility and severity of potential adverse health events’ (Aldrich et al, 1998). No mention was made of potentially adverse impacts on the

established amidstbe the dereliction, andwarning presenthas as abeen direct result of it, should preserved . The given that: ‘If you take an abandoned quarry, though, you have to be quick if you want to develop it. Otherwise it will turn into a Site of Special Scientific Interest’ (Guthrie, 1999).

The objective is to build safe, durable and economic structures. The site and the building development form an interactive system and it is important to evaluate the risk of adverse interactions during the lifetime of the development. Hazards to the built environment on a brownfield site can be physical, chemical or biological in character and concerns could include the following:  poor load-carrying properties of the ground  interaction between building materials and aggressive ground conditions  combustion  gas generation from biodegradation of organic matter and from other deleterious substances in the ground. The shortage of land and a long industrial history ensure that much of the land now being used for building purposes in the UK has been previously affected by human activities. Filled ground is one of the potential hazards which may be encountered on brownfield sites. Although brownfield land is a world-wide phenomenon, the issues are particularly acute for Great Britain, a heavily populated island with a long industrial history . An example which illustrates the scale of the problem is that

1.4 Scope

5

in 1996, a package of more than £1 billion was announced for the regeneration of major coalfields. Some 910 ha of land were to be reclaimed for residential, commercial and retail uses (Sleep, 1996). Many of these sites will involve building on colliery spoil (Skinner et al , 1997). Many brownfield sites are covered by fill and, in some cases, may also be chemically contaminat ed. However, not all filled sites will necessarily be regarded as brownfield; for example opencast mining sites restored to agriculture. Sites currently being used for building developments include the following ground conditions:  opencast mining backfills

definitions for these terms. In this book the following definitions are used:  shallow fill — less than 3 m deep  medium-depth fill — between 3 m and 10 m deep  deep fill — more than 10 m deep.

 

‘engineered fills’ and ‘non-engineered fills’. An engineered fill has been placed to an appropriate specification which has been enforced during placement. In this way a fill has be en produced which has engineering properties that are known and which are considered adequate for its purpose. Where an engineered fill is to be used as a foundation material, th e selection of suitable material and compaction plant, together with appropriate supervision of placement and compaction, should ensure adequate performance and reasonably uniform properties throughout the deposit. Engineered fill placed as a foundation material has sometimes been described as ‘structural fill’ or ‘controlled fill’, but in this book is termed ‘foundation fill’. In contrast, non-engineered fills arise as the by-product of human activities generally associated with the disposal of waste materials. Little control may have

colliery spoil infilled pits and quarries  urban fills  lagoon pulverised fuel ash (pfa)

1.3 Definitions It is difficult to find a suitable generic name for re -used or recycled ground, and brownfield is probably the most widely used term. There is a wide range of engineering and chemical problems associated with land which has been adversely affected by previous use. Four situations can be defined, although at a particular site the ground is likely to be in more than one of these categories:  Fill — ground that has been made by man placing either natural soils or rocks or waste materials over natural ground

In considering the ge otechnical problems which may arise when building on fill, it is important to distinguish between sites where filling has yet to take place and earthmoving can be controlled and supervised, and sites which have already been filled with little or no control being exercised. The basic distinction is between



Contaminated land — ground which has beenabove polluted by hazardous substances at concentrations natural levels; the contaminant may be solid, liquid or gas  Derelict land — ground that has been so damaged by industry, mining, etc, that it can no longer be put to beneficial use without treatment  Brownfield sites — land previously used for urban uses including residential, transport and utilities, industry and commerce, community services, previously developed vacant land and derelict land. The terms ‘fill’ and ‘made ground’ are used interchangeably in this book to denote a large quantity of material which has been deposited over a wide area, and which has raised the level of a site or a substantial part of a site. In contrast, a limited amount of ‘hardcore’ may be put down either as ‘infill’ within the foundations of a single building unit or beneath an oversite concrete slab. Problems involving hardcore are of a different nature and are not dealt with in this book; reference should be made to Digest 276 (BRE, 1992). Where material is used to fill an excavation, or is placed behind a retaining wall, it is termed ‘backfill’. The expression ‘landfill’ is sometimes used specifically to describe fill that is composed of domestic refuse, and hence the expression ‘landfill gas’ for the gas that is produced by the de composition of such refuse. The terms ‘shallow’, ‘medium’ and ‘deep’ describe the depths of fills, but different authorities adopt different

beeninexercised, and there is thevariability. possibility, and some cases theconsequently probability, of extreme This makes it difficult to characterise the engineering properties and predict behaviour. W here non-enginee red fills are used as foundation materials, problems may arise and caution is essential. Nevertheless, the scarcity of good building land and the desirability of redeveloping the inner cities and areas of industrial dereliction increasingly lead to building developments on heterogeneous, non-engineered fills. In such cases ground treatment may be required prior to building development. A comprehensive list of definitions of the more important terms used in this book is given in the Glossary.

1.4 Scope Fills can be composed of either natural soils and rocks or various types of waste material. Some fills composed of natural granular materials, for example sands, may have similar properties to the material in its natural undisturbed condition, although there can be important differences. While this book considers all types of fill, the emphasis is on those fills which behave least like natural soils, for example fills containing waste materials, rockfills and fills composed of lumps (or clods) of clay. The development of filled sites can present a wide variety of problems. There are two main categories, arising from:

6

Chapter 1 Introduction

the limited ability of the fill to support the building or other form of construction without excessive settlement  the presence in the fill of materials which could be hazardous to health or harmful to the environment or the building.

‘Suitable sites for new buildings and e states in industrial areas are becoming more difficult to find and it is more frequently necessary to build on made- up ground or filling.’



This book is confined to the former category of problem and is concerned with the geote chnical aspects of construction on filled ground. More general accounts of the problems of derelict land, which include all the various aspects, can be found in publications of the

Later in the Digest the following warnings were given: ‘Mixed fills, when composed of household and industrial wastes, may also contain materials liable to decay or decompose.’ ‘In general, sites of this nature are best avoided for building purposes.’

This Digest has been revised and updated on four

Institution of Civil Engineers and CI RIA such as Recycling derelict land (Fleming, 1991) and Building on derelict land (Leach and Goodger, 1991). The geoenvironmental problems of brownfield sites are also described in proceedings of conferences such as the British Geotechnical Society conference held in Cardiff in 1997 (Yong and Thomas, 1997). Fire and explosion hazards have been re viewed by Crowhurst and Beever (1987). The problems presented by the development of filled sites will depend on the type of construction, for example low-rise housing, light industrial buildings, high- rise structures, oil tanks, roads and embankments. The more flexible types of construction are likely to be le ss damaged by settlement. Buildings will generally be much more sensitive to differential settlement than embankments or storage tanks. It may be economical to take the foundations of large structures down through the

occasions since 1949:  Digest 142 (1972)  Digest 222 (1975)  Digests 274 and 275 (1983)  Digest 427 Parts 1, 2 and 3 (1997/1998)

full depth fill into underlying natural stratum using pilesofwhere nean cessary. Wherefirm small structur es are, built on deep fills this approach is unlikely to be economical, and many problems with construction on fill are associated with building small, low-rise structures on medium-depth and deep fills. The engineering behaviour of fills is described in Part II of this book. Properties of re levance to the load-carrying characteristics of fills are examined, with particular consideration given to the causes of volume changes. Important aspects of construction on fills are examined in Part III. The identification and classification of a fill through appropriate investigation and monitoring are described. Ground treatment techniques and foundation solutions are reviewed. Part IV comprises 28 case histories of sites where movements have been monitored. These are mostly filled sites where BRE has carried out investigations. Parts II and III make extensive reference to these case histories.

The present programme of began research carrying characteristic s of fill in into 1973the andloadthere has been an emphasis on field measurements of long-term settlement of fills. Surface settlement and settlement of buildings on fill have been monitored by precise levelling. Settlement at depth within fills has been monitored by installing magnet extensometers (Marsland and Quarterman, 1974) in boreholes drilled through the fill. Observations of ground-water level have been made using standpipe piezometers. The effectiveness of ground improvement techniques, including dynamic compaction, preloading and vibro techniques, in improving the load-carrying characteristics of various types of fill, has been examined. The data collected in these field investigations form the basis of this book. In addition to the general guidance in BRE Digests and Information Papers, numerous papers have been published in technical journals and conferences describing BRE investigations into the behaviour of filled ground. These are included in the list of refere nces. Other closely related research topics at B RE, such as the performance of building materials in chemically contaminated land and the hazard to buildings from landfill gas, are mentioned where appropriate, but the book does not attempt to deal fully with them.

1.5 Research a t BR E Investigations into the load-carrying properties of fill and the performance of buildings on fill have continued at BRE for over 50 years. In 1949, the Building Research Station (now BRE) published Digest 9 Building on made-up ground or filling which stated that:

These revisions reflect both the interest in the subject due to the growing pressure to build on fill re sulting from the shortage of good building land, and the greater experience in building on filled sites that has been acquired since 1949. An early investigation of the settlement of a factory built on fill was reported by Meyerhof in 1951. Some long-term monitoring of the settlement of experimental houses built on opencast ironstone mining backfill started in 1963 (Penman and Godwin, 1974); this is described in case history 1.

7

Chapter 2 Fill formation and deposits Is there anything of which it may be said, see this is new? It has already been in ancient times before us. (Ecclesiastes 110)

It has been noted that construction on fill is not a recent development, and that throughout history there have been examples of man building on fill with varying degrees of success. What has changed in the last few years is the scale and timing of such building development, and this has been particularly marked in Great Britain where the circumstances of a small, heavily populated island with a long industrial history, frequently necessitate building on waste materials. In November 1996, the Environment Secretary proposed a target of 60% of new housing to be built on previously used land and, while not all of this previously used land is covered by significant deposits of fill, much of it is. Some knowledge of the way in which various fill deposits have been of formed presentsinvolved a helpful to the examination the problems inbackground building on fill. No study of natural soil behaviour would be complete without some consideration of the srcin of the soil deposits, because the mode of formation affects subsequent engineering behaviour. However, the geological processes involved in the formation of a natural soil deposit have to be inferred from a study of the deposit in its present condition. In contrast, the formation of a fill deposit may have bee n witnessed and information about it may have been recorded . The study of the formation of fills is therefore different from the study of the formation of natural deposits in the following respects:  the evidence is of a different nature  the depositional processes are usually of a different form. Fills may be composed of natural soils and rocks, or they may be formed from industrial, chemical, mining, dredging, building, commercial and domestic wastes. Chapter 1 referred to the important distinction between engineered and non-engineered fill. Fills can be described and classified by:  degree of control during placement, eg engineered or non-engineered  method of deposition, eg whether the fill was placed in dry conditions or in submerged conditions, in thin layers or in high lifts, or transported and placed hydraulically

Engineered fill is selected, placed and compacted to an appropriate specification, so that it will exhibit the required engineering behaviour. Engineering design focuses on the specification and control of filling. Engineered fills are the subject of Chapter 9.

Non-engineered fill has arisen as a by-product of human activity, usually involving the disposal of waste materials. It has not been placed with a subsequent engineering application in view. Most problems are associated with developments on non-engineered fills. This chapter is primarily concerned with non-engineered fills.



type of fill material, eg colliery spoil, pfa, chemical

wasteof deposition, eg clay pit, old dock, opencast place mine  age. 

The following sections of this chapter give an indication of the range of situations that commonly may be encountered in the UK:  opencast mining backfill  colliery spoil  pulverised fuel ash (pfa)  industrial and chemical wastes  urban fill  domestic refuse  infilled docks, pits and quarries  hydraulic fill While these headings are useful as an indication of the wide range of circumstances in which fills occur, they are not mutually exclusive categories. For example, colliery spoil may be used to infill a dock. Guidance on the use of industrial wastes as fill materials in building and civil engineering is given in B S 6543:1985 (BS I, 1985).

2.1 Opencast mining backfill When the mineral has been extracted in an opencast mining operation, the overburden soils and rocks are replaced in the excavation. The backfill is therefore composed of these natural soils and rocks, although in

8

Chapter 2 Fill formation and deposits

some cases imported material has been placed within the backfill. Many opencast backfills have been placed without systematic compaction. BRE has carried out major investigations, some of them collaborative, at a number of opencast sites, and brief case histories are presented in Part IV (case histories 1 to 8). These include cases where damage to buildings is believed to have bee n caused by collapse compression on inundation of the backfill (case histories 5 and 7). A number of other studies are of importance. Knipe (1979) presented settlement observations for a number of opencast sites in the West Midlands. A major review of

largest dragline has a bucket capacity of 50 m 3. The excavations usually go below the water table and dewatering techniques have been required ( Norton, 1982, 1983). Sites have typically been restored to agricultural use, often without systematic compaction of the fill during the backfilling. Other minerals have also been e xtracted by opencast mining, and have produced large quantities of non-engineere d fill. In the vicinity of Corby over 10 00 ha of land affected by opencast ironstone mining have been restored to agricultural use. Figure 2 shows an aerial view of opencast mining at Corby, including the experimental

information relating to the controlled re storation of opencast coal sites which had been backfilled for a specific after-use, and for which settlement and compaction monitoring data are available, has been reported by Hodgetts et al (1993). Detailed accounts of the performance of engineered fills at particular sites have been given by Goodwin and Holden (1993), and Trenter (1993). There is published information concerning the settlement of similar backfills in other countries including Canada, the Czech Republic, Germany and the USA. Lange (1986) has described settlement measurements in the very deep backfills of the Rhenish brown coal area. The open-pit mining of brown coal has led to similar settlement problems in North Bohemia (Vanicek, 1991). An unusual situation in which the relocation of a church away from a site scheduled for open- pit mining involved

site described in case history 2.

moving the across a by previously backfill, hasbuilding been described Skopek surcharged (1979). The University of Alberta has carried out a research programme on open-pit mine backfill (Thomson and Schulz, 1984; Schulz et al , 1986) which has included deep settlement instrumentation, surface settlement monitoring, prototype foundations and inundation testing on both granular and cohesive backfills. Koutsoftas and Kiefer (1989) have described field tests to evaluate the feasibility of using preloading and dynamic compaction to improve a 30 m deep fill left by coal strip mining in southern Illinois.

Development

Extent

Opencast coal mining has been a major producer of deep non-engineered fills in Great Britain. It started in 19 42 as an emergency measure during the Second World War, and continued for many years to be a major producer of deep non-e ngineered fills. During the 1980s opencas t coal production was maintained at about 14 million tonnes per annum (British Coal, 1988). The ratio of overburden to extracted mineral is large; in 1942 the acceptable ratio of overburden to workable coal was 4:1, 30 years later it was 40:1. On average, it is necessary to remove 30 tonnes of overburden for every tonne of coal. Thus the extraction of 1 tonne of coal typically involves the excavation of 15 m3 of overburden. Typical working depths have been 30 m to 40 m, with a maximum depth of 250 m at Westfield in Scotland. Methods of excavation include scraper, truck and shovel, and dragline. The

s lm fi o r e A

Figure 2 Opencast mining at Corby

Where building development is foreseen at the time of opencast mining, an engineered fill should be specified, with the backfill placed in thin layers and adequately compacted under supervision. However, proposals for development of a site are often made after opencast mining and backfilling have been completed. It is then necessary to determine the suitability of the nonengineered ope ncast backfill for such developments. Details of the opencast mining operations, such as the location of an overburden heap or lagoon, can be significant for future development of the site (Charles et al, 1977), but detailed information will not always be readily accessible. Restored opencast coal mining sites have been widely used for housing, light industrial and road developments (Gilbert and Knipe, 1979; Smyth-Osbourne and Mizon, 1984; Buist and Dutch, 1984; Charles and Burford, 1987; Kilkenny, 1988). An opencast coal site was reinstated with engineered fill to a depth of ne arly 100 m for the Sheffield and Rotherham City airport (Morgan et al, 1993). The expansion of Corby has involved extensive housing and industrial development on opencast ironstone mining backfill (Penman and Godwin, 1974; Charles et al , 1978; Burford and Charles, 1991; Burford, 1991). Case histories 1, 2 and 3 refer to sites in Corby. Problems

Problems of a ge otechnical nature are usually associated

2.2 Colliery spoil

9

with the poorly compacted state of the backfill. Early work on the settlement of ope ncast backfills led to recommendations that a certain period of time should elapse after restoration before development took place (Kilkenny, 1968). F or sites over 30 m dee p a minimum period of 12 years was suggested. This approach was largely based on observations that showed that the rate of settlement of fills due to self-weight decreased rapidly with time. Many building developments have now taken place on restored opencast mining sites and in a few cases problems have occurred due to settlement of the fill, even when a long period of time has elapsed between

means of dealing with this hazard, but in many situations it is difficult to carry out in an effective and controlled manner (section 8.4).  Eliminate or greatly reduce susceptibility to collapse compression of backfill by some form of pre-treatment which increases fill density, for example preloading (section 8.3) or dynamic compaction (section 8.1 ). Alternatively, excavate the fill and replace with an engineered fill either by re -using the excavated fill or by replacement with better quality imported fill (Chapter 9).  Prevent inundation of the backfill during the life of the

restoration and development (Smyth-Osbourne and Mizon, 1984; Charles and Burford, 1987; Kilkenny, 1988; Charles and Watts, 1996). It is now recognised that collapse compression on wetting of an uncompacted, partially saturated backfill can be the cause of such damaging settlements and that many problems that have occurred when building on deep opencast mining backfills are associated with the movement of ground-water or the ingress of surface water (Charles et al , 1977, 1978; Charles and Burford, 1987; Reed and Singh, 1986; Hodgetts et al , 1993; Goodwin and Holden, 1993; Morgan et al , 1993; Reed and Hughes, 1990; Trenter, 1993). Settlement can occur at depth as the ground-water table rises (case histories 4 and 7) or close to ground level as surface water penetrates into the fill via shallow excavations through the surface crust (case histories 2

structure. For example, it may be possible to prevent water penetrating the surface crust by care in design and construction. However, sloping sites present particular problems as do leaking drains, and a rise in the level of the ground-water table could occur subsequent to construction due to factors outside the control of those responsible for the development of the site.  Design structures to survive the movements due to collapse compression without serious damage by either making them independent of backfill settlement by providing deep foundations (eg piles penetrating to natural ground) or making them able to tolerate large movements. The former solution is unlikely to be economically feasible where small structures are built on deep fills. When the latter solution has been attempted, houses have commonly been built on

and 5). The case histories shown that in both situations settlement can behave of sufficient magnitude seriously to damage structures built on backfills. Where water infiltrates into the backfill from surface trenches, substantial movements may continue for long periods after the trenches have been backfilled (case history 2). At a site where tilt increased to 300 mm over the 9 m width of a detached house, it was concluded that vibro stone columns, which had been installed in the opencast backfill, would have provided a means of introducing water into the backfill and encouraging collapse compression (Thompson, 1998). The magnitude of the collapse compression will depend on factors such as the type of backfill, its density and moisture content and the stress level and stress history; these factors are examined in Chapter 5. The implications of these results for building on opencast backfill are serious. If opencast backfill has been placed without systematic compaction and has not yet been inundated, it is probable that it will be susceptible to collapse compression on wetting. In this situation it is not necessarily sufficient to allow a certain period of time to elapse between restoration and development. Several approaches to the problem may be considered for building development on non-engineered opencast mining backfills, and in practice combinations of these approaches may be adopted:  Eliminate collapse compression by pre-inundation of the backfill. Flooding the site prior to building development might appear to be the most logical

lightly reinforced rafts with(Chapter edge beams, still be unacceptably large 10).but tilts could

2.2 Colliery s poil Waste from the deep mining of coal is derived from the rocks adjacent to the coal seams. These are mainly siltstones and mudstones with seat-earths and occasionally other rock types. During mining operations quantities of these rocks, unavoidably extracted with the coal or in driving the tunnels which give access to the coal faces, are brought to the surface. In the coal preparation plant most of this material is separated to yield coal of the required quality. Case histories 9, 10 and 11 illustrate situations in which coal mining wastes have been used in reclamation schemes. The different placement methods had a controlling effect on the subsequent performance of the colliery wastes. Where the waste has been placed in thin layers and adequately compacted its performance should be adequate for many building applications. Where this degree of enginee ring control is not practicable or was not exercised, a le ss satisfactory performa nce can be expected. Extent

In 1974 it was estimated that there were 3000 × 106 tonnes of colliery spoil in the UK (Gutt et al , 1974). In 1984 Taylor reported that the coal mining industry produced 55 × 106 tonnes of waste annually comprising

10

Chapter 2 Fill formation and deposits 6

50 × 10 tonnes of coarse discard placed in tips adjacent to the collieries, and 5 × 106 tonnes of fine discard which was carried in suspension from the washing process, much of it being discharged into lagoons. The rapid decline of the coal industry in the 1990s meant that the annual production of these wastes dramatically reduced, but large stocks of colliery spoil remain in the coalfields near the many closed collieries as well as the few working ones. In April 1988 there were 4700 ha of derelict land associated with colliery spoil heaps in England alone (Department of the Environment, 1991a). Colliery spoil

spoils has been measured in case histories 9, 10 and 11. Geotechnical problems are largely overcome by adequate compaction during placement. There has been concern that the material would break down in the long term, but observations have indicated that only the surface layers are affected by weathering. Pyritic shales can cause problems through expansion induced by slow oxidation. Where these materials underlie foundations, damage has often been caused by chemical attack of sulfate on concrete rather than by swelling. Chemical reactions are examined in section 4.6. Older spoil heaps often ignited and the materials

heaps accounted for 23% of the 14 000 ha of derelict land reclaimed between 1982 and 1988. In 1996 it was announced that British Coal non-operational properties would be transferred to English Partnerships to expedite regeneration and environmental improvements (Department of the Environment, 1996). It was planned that some 900 ha of land would be reclaimed for residential, commercial and retail uses and a further 1100 ha for forestry, leisure and agricult ure.

consequently may be a mixture of unburnt, partially burnt and well burnt spoil. Variability is then a problem if the fill is used in reclamation works which involve building developments on the fill. Turner (1979) suggested that the coal content of a spoil should be determined using float and sink tests rather than loss on ignition which could overestimate the amount of coal by 5%. Where the fill is well compacted and the air voids are small, the possibility of spontaneous combustion is remote (Isaac and Troughton, 1971; Isaac, 1972). More serious is the potential for combustion initiated by an external source, such as:  deposition of burning materials (eg hot ashes)  lighting a bonfire to dispose of demolition waste  indirect initiation by an external source of heat which may so accelerate the processes of natural oxidation that an uncontrolled and accelerating reaction occurs,

Development

Many reclamation schemes have removed old spoil heaps and regraded to specified contours. Colliery spoil used for landscaping may not have been placed to an engineering specification. Figure 3 shows colliery spoil being placed in 2 m high lifts in a land re clamation scheme (case history 1 0). Colliery spoil has been used extensively in1975a; engineered fills 1987a) for roadand embankments (Sherwood, Sleeman, has found some use on building sites including housing de velopments (Kirkpatrick and Webber, 1987; Hart et al, 1987) and industrial estates (Colman, 1977).

leading to ignition (egyard). a boiler house sited above coal left in a coal-stocking A survey of subterranean fires in the UK for the threeyear period to June 1987, found 64 incidents (Bee ver, 1989). Some 25% of these were related to coal or colliery spoil.

2.3 Pulverised fuel ash

Figure 3 Colliery spoil in reclamation scheme

Problems

Possible causes for concern are associated with geotechnical, combustio n and chemical properties of the spoil (Sleeman, 1987b). Following the Aberfan disaster in 1966 when the failure of an unstable colliery spoil heap caused great loss of life, colliery spoil was the subject of major geotechnical investigations (Thomson and Rodin, 1972; Wimpey Laboratories Ltd, 1972; Taylor, 1984). It has become the usual practice to place the spoil in thin layers with compaction. Lack of adequate compaction can render the spoil susceptible to significant settlement including collapse compression on wetting. Settlement of colliery

This waste product is produced by power stations which burn pulverised coal. Approximately 80% of the ash is pulverised fuel ash (pfa) and 20% is furnace bottom ash. Pulverised fuel ash is composed principally of nearly spherical particles predominantly in the coarse silt and fine sand range. Metal oxides constitute the major proportion of pfa; some trace constituents and pozzolanic chemicals are also present. It is a lightweight material. General information was presented in two reports by what was then the Central Electricity Generating Board (CE GB, 1967; Barber et al , 1972). Extent

There was an expansion of electricity generating capacity in Britain in the early 1960s, and the majority of new power stations were coal fired. Associated with this development there was an increase in the amount of pfa produced. In 1978, 10 × 106 tonnes were produced; 40% was utilised in the building and construction industries, and 60% was used to reinstate old gravel and clay pits so

2.4 Industrial and chemical wastes that the land could be returned to agricultural or building use (Brown and Snell, 1978). By 1986 the annual rate of production of pfa was 11 million tonnes. Where pfa is used as a bulk fill in civil engineering applications it is usually mixed with an optimum amount of water and described as ‘conditioned pfa’. Thanks to its low particle density it forms a useful lightweight fill with some pozzolanic properties. Its use as a structural fill has been described by Leonards and Bailey (1982). Where pfa is disposed of as a waste material it is usually mixed with sufficient water to enable it to be transported in suspension. ‘Lagoon pfa’ arises from the disposal of such hydraulical ly transported pfa. As lagoon pfa e xists at a higher moisture content, it can be expected to have properties that are different from, and possibly inferior to, conditioned pfa. Variability may also be a problem with lagoon pfa; particle size will be a function of distance from the outfall and so will vary across the site. Filling of worked-out clay pits started at Peterborough in 1965 (Wright and Brown, 1979; Humpheson et al, 1991). Dry ash was brought to Peterborough in pressurised rail tankers, which were emptied by compressed air and immediately mixed with water to form a slurry containing 25% ash. The slurry was pumped to the pit being filled and discharged at several points around the perimeter. When the ash had settled out of the slurry, the water was decanted. The resulting deposit had a moisture content of about 40%. The depth of pfa ranges up to a maximum of about 20 m. When a pit had been filled, topsoil was placed. Development

The use of conditioned pfa as a lightweight fill in civil engineering works is well established (Barber et al, 1972; Sherwood, 1975b). It has been used extensively for road embankments, its lightweight properties being particularly useful for bridge approach embankments on soft foundation soils. Building developments have taken place on conditioned pfa (Swain, 1979; Weatherley, 1979; Wilde and Crook, 1979). The development of the use of pfa as an engineered fill has been de scribed by Fox (1984) and Cabrera et al (1984). Developments on lagoon pfa are rare. The Metro Centre at Gateshead is one e xample and Ballisager and Sorensen (1981) have described the construction of an oil tank on a 5 m depth of lagoon pfa in Denmark. A new township at Peterborough is being built partly on lagoon pfa. Extensive investigations, including cone penetration and laboratory testing, were undertaken to demonstrate the feasibility of building development. Humpheson et al (1991) have described field loading tests, vibration trials and the monitoring of ground water levels. Case history 12 describes a preloading trial at this site.

11

particularly where there is a high ground-water level . The possibility of liquefaction should be examined, although in case history 12 laboratory and field testing indicated that there should not be a problem at this particular site. Settlement and ground-water level measurements in lagoon pfa are described in case history 12. Environmental problems associated with the physical and chemical properties of pfa may need to be addressed, including toxic trace elements, radon emission and dust control.

2.4 Industrial and chemical wastes Derelict sites of former industrial land may contain extensive areas on which considerable depths of process wastes have been deposited, either simply as a means of disposal or, in some situations, to form the base material for subsequent construction. Industrial wastes have been used as bulk fill in land reclamation schemes. Wastes may be of considerable age, the residue from processes which have long been obsolete. Industrial and chemical waste deposits could be found in any of the following forms:  hard coarse material such as metalliferous slag and foundry sand  fine unconsolidated sludges and silts in lagoons  biodegradable materials including refuse. The manufacture of iron and steel has been accompanied by the production of large quantities of slag. Blastfurnace slag is a by-product of iron steel slags are created at a later stage in themanufacture; process to facilitate refining. Recently produced slags may be of consistent composition, but at older sites slags may be very variable reflecting changes in steel making processes. Extent

Out of a total of 40 500 ha of derelict land reported in the 1988 survey of dere lict land in England (Department of the Environment, 199 1a), 8500 ha were attributed to general industrial dereliction. In 1971 nine million tonnes of blastfurnace slag and four million tonnes of steel slag were produced in the UK. Most of the blastfurnace slag was processed and used in the building and construction industry. Thomas (1983) stated that nearly all the blastfurnace slag and much of the steel slag was being utilised either within the steel industry or e xternally. Other examples of wastes from industrial and chemical industries include:  waste from the heavy chemical industry in Cheshire, which amounted to 21.5 × 106 tonnes during the year 1979/80, most of which went to lagoons  alkali wastes found at, for example, Glasgow and St Helens.

Problems

There may be both geotechnical and environmental problems. Sherwood (1975b) listed the compaction requirements for conditioned pfa. Construction on lagoon pfa may involve geotechnical problems,

Development

Many former industrial sites have been redeveloped, including gas works and stee l works. Thorburn and Buchanan (1983) have reported redevelopment of four

12

Chapter 2 Fill formation and deposits

sites in Glasgow. One of these involved the construction of a new church on old chemical waste from the Leblanc process for the manufacture of washing soda; reinforced concrete footings were constructed on compacted granular fill in trench. Industrial waste materials such as foundry sand and blastfurnace slags have been used in derelict land reclamation schemes.

the building industry, particularly demolition work (Nixon, 1976).

Problems

Waste fills tend to be poorly compacted, very heterogeneous, and they may be chemically unstab le, so that in addition to geotechnical problems a range of chemical problems can be encountered. Chemical reactions which can cause volume changes in the waste fills are described in section 4.6, and this includes problems with iron and steel slags. Reclamation of the old steelworks waste tip at Brenda Road, Hartlepool is described in case history 13. After excavation and recompaction, chemical reactions caused significant expansion of the old industrial wastes, which included iron and steel slags. In addition to the engineering problems associated with poor load -carrying characteristics, industrial sites can present a number of problems for development related to the past usage of the land. These include:  vulnerability of the materials of construction to aggressive ground conditions  hazards to health and safety of people working on or 

using thetosite hazards the environment.

Waste materials such as metalliferous slags often contain small amounts of toxic elements such as lead, cadmium and zinc, although the availability to the e nvironment may be low due to the crystalline or glassy nature of such materials (Emery, 1979; Emery and Machett, 1979). Account should be taken of possible risks to the health and safety of personnel during investigatory work (Chapter 7), and of possible hazards to the gene ral public in the long term. There is a wide range of potential shortand long-term health and safety hazards and environmental problems which may arise depending on previous land use (Smith, 1979a, b; Smith and Beckett, 1982). However, re cords of previous use of the site may be incomplete. Demolition may have spread contamination over the whole site and buried old foundations and service pipes. When combustible materials are present there may be cavities where areas have been burnt out, and indee d combustion may still be in progress. High moisture slurry tailings may remain at a very low strength (Green, 1977; Devenish and Green, 1980); hydraulic fill is described later in this chapter.

Extent

In Chapter 1 it was explained how continuous human occupation of land leads to the accumulation of fill. This has happened in historical times, as street leve ls have risen due to the disposal of household re fuse and houses have been rebuilt on the ruins of demolished buildings. These urban fills may contain soil, rubble and re fuse. Low-lying marshy areas close to rivers have bee n reclaimed by raising the ground level with a layer of fill; such fills may be extensive, but shallow. Although the redevelopment of urban sites has continued throughout history, it now occurs on a much greater scale and much more rapidly than in the past. It has been estimated that some 20 × 106 tonnes of demolition wastes are produced annually in the UK. As would be expected these wastes occur mostly in urban areas. Development

Programmes of urban renewal have accelerated and rapid and extensive redevelopment programmes have taken place. Redevelopment may occur on the site of demolished buildings and basements. Concrete and brick rubble from the demolition of buildings is commonly used as fill in urban situations and may be found in many non-engineere d fills. Old basements have frequently been infilled with demolition rubble which, in addition to concrete brick, wood,in glass Old dockand areas havecould beeninclude redeveloped citiesand suchplaster. as London, Cardiff and Liverpool. Case history 14 describes redevelopment on shallow urban fill at Gree nwich. Case history 15 describes a housing development on deep urban fill treated by vibro stone columns. Problems

Geotechnical problems are associated with the uncontrolled nature of many urban fills which are unlikely to have been adequately compacted. The fill may contain unsuitable material. Large obstructions may be present, causing difficulties for piling and certain types of ground treatment. In some cases the underlying natural soils may be soft and compressible. Where construction takes place on sites covered by a considerable depth of demolition wastes in old infilled basements, the heterogeneous material may be in a loose state. Demolition wastes may contain potentially deleterious materials such as wood, glass, steel reinforcing bars, gypsum plaster or asphalt. Sulfatebearing wastes include plasterboard and similar products generally based on gypsum. Expansion may not be a problem, but these materials may be sources of sulfate attack on concrete. Wood may rot.

2.5 Urban fill This section reviews the various types of fill which have accumulated in urban areas and also the considerable quantities of waste materials which are associated with

2.6 Domestic refuse A major proportion of the domestic wastes generated in the UK has be en disposed to landfill. For example, in

2.6 Domestic refuse

13

1986, over 90% of controlled domestic, commercial and industrial solid wastes (excluding mining and quarrying wastes) were disposed of directly by landfill methods (Department of the Environment, 1986). In the year 1986–87 nearly 20 × 106 tonnes of household or domestic waste were disposed of by the Waste Disposal Authorities in England and Wales (Chartered Institute of Public Finance and Accountancy, 1988), and this figure remained substantially unchanged over the subsequent ten years (Chartered Institute of Public Fi nance and Accountancy, 1998). Despite e nvironmental initiatives, it is likely that landfilling will continue on a large scale. In

content is rapidly used up during ae robic decomposition which usually lasts for only a few days, after which anaerobic decomposition processes start. Decomposition leads to the production of carbon dioxide and methane which, along with other minor constituents, comprise landfill gas. During the decay process, the composition of the landfill gas changes. Initially carbon dioxide predominates, but later the gas consists typically of 65% methane and 35% carbon dioxide. Figure 5 illustrates the mixed nature of old refuse. Figure 6 shows recent domestic refuse being landfilled. The volume reduction which occurs with time is

1995, 70% of controlled wastes in the UK were be ing disposed to landfill (Department of the Environment, 1995). The composition by we ight of household refuse in Britain has changed markedly over the years as shown in Figure 4. The ash content reduced from over 50% in 1935 to only 10% in 1980; the major change occurred between 1960 and 1970 following implementation of the 1956 Clean Air Act which resulted in a major re duction in the use of solid fuel and hence the ash content of refuse. The loss of this inert, denser constituent and the increase in volume of paper and rag, and the use of plastic containers and bin liners, has had an adverse e ffect on the geotechnical properties of refuse fill. The biodegradable vegetable content was about 15% before 1960, but has been about 25% from 1980 onwards. Older waste fill also has had a greater period to decompose and is inherently a

described in section 4.5. BRE has investigated the effect of ground treatment techniques such as preloading and dynamic compaction at several old refuse sites and these case histories (16 to 20) are presented in Part IV. Also presented are case histories of the settlement of recently placed refuse landfill at two large disposal sites in the south of England, Brogborough (case history 21) and Calvert (case history 22). Settlements measured at these two sites since 1985 have shown that recently placed domestic refuse landfill is subject to large reductions in volume due to biodegradation and that it is highly compressible under applied loads. The measurements strongly suggest that large movements will continue for many years.

better engineering Recycling es and the landfill taxation arematerial. likely significantly toinitiativ alter further composition of refuse landfill. Blight et al (1999) have provided comparative figures for waste composition in the UK, the USA and a number of developing countries. Recent domestic refuse contains a large proportion of organic material, which is readily degradable. When waste is deposited and compacted at a landfill site, air is trapped in the voids within the waste. The oxygen

Figure 5 Old refuse

Figure 4 Composition by weight of domestic refuse in Britain (after Watts and Charles, 1999)

Figure 6 Compaction of recent domestic refuse

14

Chapter 2 Fill formation and deposits

Extent

measurements of Chang and Hannon (1976) on an old sanitary landfill in the USA corresponded to a modulus of 2.5 MPa. Case histories 16, 19 and 20 have given values varying between 2 MPa and 5 M Pa for untreated refuse materials dating from about 1960. Some load tests on domestic refuse have been reported by Harris (1979). Sargunan and Rajamani (1986) have described the investigation of waste disposal fills in Madras, including plate and pile load tests. Long-term settlement is likely to be the major problem (section 4.5). Meyerhof (1951) published an account of a BRE investigation of the settlement of a large factory

In the past, landfill sites tended to be relatively small and served a local community. They were usually situated on the outskirts of a town or city. The situation has changed as the older, smaller sites have reached their capacity and have been engulfed by urban development. Many large sites are currently being developed for the disposal of domestic refuse by landfill methods. Clay pits associated with the brickmaking industry are particularly suitable for waste disposal, as the excavation usually remains surrounded by clay. This limits the ingress of groundwater, the transfer of harmful leachate into natural water courses and aquifers, and the lateral migration of landfill gas. Examples of this use of old clay pits are found in the Stewartby area of Bedfordshire (case histories 21 and 22). Development

Small, old refuse sites are often proposed for building development as they may be the only areas in an urban location that have not already been de veloped. As such developments are small, they attract little publicity unless there are serious problems. Some of the press headlines quoted in section 1.1 refer to such cases, but these reports rarely contain enough information to draw any firm technical conclusions. Experience of construction on domestic refuse in the USA, where domestic refuse is termed ‘sanitary landfill’, has been summarised in a number of technical papers, for example by Eliassen (1942, 1947), Sowers (1968), and Stearns and Petoyan (1984).placed domestic refuse are Sites containing recently best avoided for building development. Gas protection measures are particularly difficult to maintain in private housing. Waste Management Paper 27 (Department of the Environment, 1991b) concluded that: ‘Domestic housing should not therefore be built on landfills which are gassing or have the potential to produce significant quantities of gas’. Warnings are also given for developments adjacent to such sites. Although there are severe hazards for building developments on recent domestic refuse, some developments are known to have taken place on this type of fill. As there is little publicity unless major problems develop, it is difficult to assess the extent to which these sites have been built on. Problems

Where sites containing degradable material have been built on, there frequently have been problems; these can be associated with settlement, landfill gas and leachate. The older the refuse the less severe the problems are likely to be. A survey of subterranean fires in the UK found that almost half the 64 incidents which were reported over a three-year pe riod were associated with domestic waste (Beever, 1989). Old refuse is likely to have a high compressibility. At Stockley Park (Gordon et al, 1986), the constrained modulus measured under a small surcharge was 2 MPa (constrained modulus is described in section 3.4). The

built on fill, where end tipping of industrial and domestic waste had occurred at the site from the middle of the previous century. When the factory was built in 1937, all the fill had been in position for 20 years, yet substantial settlement occurred and serious damage was caused to the building. The various case histories in Part IV which deal with old domestic refuse give information on longterm rates of settlement resulting from biodegradation of the refuse and from applied loads. With recent refuse, particularly serious problems for built development could be associated with the following:  generation and migration of landfill gas  major settlement due to biodegradation of the refuse  high compressibility of the refuse  contamination of ground-water by leachate. Recently placed domestic refuse isdevelopment likely to be an unsuitable foundation for building both because of the generation of landfill gas and because of large settlements. In 1986, a bungalow at Loscoe was destroyed by a methane gas e xplosion and a nearby domestic waste landfill placed from 1977 to 1982 was identified as the primary source of the gas (Williams and Aitkenhead, 1991). When gases accumulate in a confined space there are hazards posed by:  the danger of an e xplosion (eg when gases accumulate under the floor slab of a building)  the danger of asphyxiation or other harmful effects to occupants of buildings in which the gases accumulate. A BRE report (Crowhurst, 198 7) gives guidance on the measurement of gas emissions, and CI RIA have produced a series of reports on gas-related properties:  Protecting development from methane (Card, 1995)  Methane investigation strategies (Raybould et al , 1995)  Interpreting measur ements of gas in the ground (Harries et al, 1995) Waste Management Paper No 27 (Department of the Environment, 1991b) has provided guidance on the problems of gas generation in refuse landfills. While suggesting that after-use of landfill sites should normally be restricted to agriculture or similar uses, it recognises that other forms of development should be possible provided adequate precautions are taken. It recommends that no housing should be built within 50 m of any landfill

2.8 Hydraulic fill

15

site with gas concentrations in excess of certain specified limits. It also states that where development is proposed within 250 m of the site, specialist advice should be sought as to what measures are required to ensure the safety of such development. Measures for keeping landfill gas out of buildings basically compris e a gas-proof barrier with a highpermeability layer beneath it from which gas can be extracted in a controlled manner. Reports from BRE (1991) and CIRIA (Card, 1995) give examples of suitable construction techniques. Protective measures are also described by Johnson (2001).

development, it will not be feasible to compact the fill in thin layers where it is not practicable to de-water an old dock or gravel pit. In this case, selection of a good-quality granular fill can mitigate the problems which might otherwise arise. Differential settlement may be a major problem at the edge of the infilled hole (Chapter 6). A dock wall provides an extreme situation and it may be advisable to excavate some of the wall to reduce differential settlement (case histories 23 and 24). In old docks there may be silt which is difficult to remove (case history 23). In the London Docklands there was a large quantity of contaminated silt at Western Dock

2.7 Infilled docks, pits an d quarries

(Lord et al, 1986). In the redevelopment of Surrey Docks there were problems with methane and ground contamination (Thomson and Aldridge, 1983).

Many redundant excavations, including docks, pits and quarries, have been backfilled with little or no control over the type of fill and method of placement. Filling may have been in dry conditions or into a water-filled hole. Where the excavation had steep or vertical sides it is probable that the fill was end-tipped in high lifts. Several backfilled docks which hav e been used for various types of development have been investigated by BRE ( case histories 11, 23 and 24). Extent

Backfilling may have been carried out with subsequent building development in view. Alternatively and more commonly, backfil ling may have bee n simply to dispose of wasteonly materials, with theMany possibility of de excavations velopment arising at a later stage. backfilled are relatively small, although some very large clay pits have been infilled. The fill may be deep, and the sides of the hole may be steep or, in the case of an old dock, vertical. Development

Redundant dock areas ha ve been redeveloped in London, Liverpool and many other ports. Sometimes the water has been retained as a feature, but often the old docks have been infilled. Major works have bee n associated with London docklands: Thoms on and Aldridge (1983) have described the dynamic compaction of the infilled Surrey Docks; the backfilling of Western Dock has been described by Lord et al (1986). These reclamation schemes involved housing development. Many infilled pits have been used for building developments. Construction on an infilled gravel pit at Abingdon is described in case history 25. Clay pits at Peterborough have been infilled with lagoon pfa and a new township is planned on the site (case history 12). Problems

Geotechnical problems are associated with the looseness and variability of fill. Little information about the fill materials may be available. There may have been virtually no compaction, and in some cases the fill may have been end- tipped into water. Even where infilling takes place specifically for subsequent building

2.8 Hydraulic fill Natural soils or waste materials can be mixed with a sufficient quantity of water to enable them to be transported in suspension. Whitman (1970) suggested that the ratio of volume of solid particles to volume of sluicing water should be 1 to 6. For the lagoon pfa at Peterborough, the ratio was 1 to 4 ( Humpheson et al, 1991). Usually the suspension is pumped through a pipe and then discharged onto the surface being filled. The deposit which is formed in this way is described as hydraulic fill. As the suspension flows away from the discharge point, the larger soil particles settle out almost immediately and the water and f low Eventually the fine material alsofines settles out,away. but at a much greater distance from the discharge point. Where hydraulic fill is being used to re claim land for building, such segregation could be a problem and the use of a single fixed discharge point might be undesirable. Whitman (1970) concluded that for clean hydraulic sand fills, density index ( ID) was likely to be in the ra nge 0.45 to 0.6 (the significance of density index is considered in section 3.3). An ID of 0.7 has been measured for the lagoon pfa at Peterborough (Humpheson et al , 1991). The placement of hydraulic sand fills beneath water involves settling from a slurry. There are two different placement methods for subaqueous hydraulic fills:  bottom dumping  pipeline placement From observations on hydraulic fills in the Canadian Beaufort Sea, Sladen and Hewitt (1989) concluded that the range of de nsities achievable by subaqueous placement of sand straddled the boundary between acceptable and unacceptable potential performance. A dense sand could not be obtained by simple hydraulic placement; at best, a mean ID of 0.6 might be anticipated, but ID could be lower than 0.2. Considerable variations in ID were found. Some fills were susceptible to liquefaction flow slides. Material placed by bottom dumping is significantly denser than pipeline-placed material. Lee et al (1999) studied the e ffects of the placement method on geotechnical properties at five sites in Hong Kong. They

16

Chapter 2 Fill formation and deposits

concluded that the weakest zone of conventionally placed hydraulic fills extends a few metres below the water surface, because in this region the water is too shallow to permit bottom dumping from barges; above the water surface, compaction equipment can be used. When fine soils are dredged, much depends on the particular method of dredging and deposition that is used. Hartlen and Ingers (1981) reported that when clay was dredged using a cutter suction dredger, the material leaving the discharge pipe consisted of well rounded lumps, with diameters up to 300 mm, with a suspension of fine eroded material. The non-uniform structure of the

Development

filling was reflected in a wide scatter of measured values of undrained shear strength. Sully (1986) described a dredged fill which consisted of rounded clay lumps, up to 0.45 m in diameter, within a sand slurry. Cragg and Walker (1986) reported on a hydraulic clay fill that at the dredge outlet consisted of 10 mm to 25 mm diameter balls of intact clay with the interstices filled with sandy silt. Far from the dredge outlet the material had a varved appearance and ranged from clayey silt to silt. Engineering properties of hydraulic fills are often far from ideal, but in some situations hydraulic filling is the only feasible method of fill placement. In other situations economic considerations may strongly favour its use.

construction sites. As far as is known, little development has occurred yet on old tailings lagoons. However, a new township at Peterborough is being partly built on lagoon pfa filled clay pits (Humpheson et al , 1991). Case history 12 describes a trial of preloading at this site.

Extent

Hydraulic filling methods are used both to dispose of materials and to reclaim land. The former activities include thewastes. disposal oflatter dredgings and include various mining industrial The activities the and reclamation of low-lying wetlands and worked-out pits and quarries. There are also cases where, although hydraulic fill was deposited simply as a means of disposal of material, subsequently the area is designated for some form of development. Ball (1979) has described the investigation of a quarry waste silt lagoon on the M42 motorway. The hydraulic disposal into lagoons of waste materials from mining, quarrying and industrial processes in the UK has been summarised by Penman and Charles (1990). For many years, the coal mining industry annually produced 5 × 106 tonnes of fine discard. This was carried in suspension from the washing process and much of it was then discharged into lagoons (Taylor, 1984). The mining and processing of china clay leaves a fine tailings containing mica (Penman, 1986). The industry located in Cornwall and Devon at one time discharged these residues into rivers, but with increasing concern for the environment the micaceous residues are now discharged into lagoons. Large quantities of pfa have bee n pumped into lagoons (Brown and Snell, 1978). Maintenance dredging produces an organic silty clay of high plasticity which has poor engineering properties ( Bishop and Vaughan, 1972). Subaqueous hydraulic fills have been used to form artificial sand islands for drilling in the Beaufort Sea (Sladen and Hewitt, 1989).

Many building developments have taken place on lowlying land reclaimed with dredged hydraulic fill. The construction of storage tanks on soft hydraulic fill at Teesmouth has been described by Penman and Watson (1967). There are many large-scale examples of this type of development overseas. Pyke et al (1978) have described how the major expansion from the 1920s onwards of the port facilities of Los Angeles and Long Beach have involved hydraulic filling with dredged material. It seems likely that the shores of rivers and se as near large metropolitan areas will continue to be in de mand for

Problems

Hydraulic fills have a high placement moisture content; if granular they are likely to be in a relatively loose state and if cohesive they will be soft. The engineering properties of hydraulic fills will generally be relatively poor. Effective stresses are likely to be small, particularly if there is a high ground-water le vel. Where fills have underdrainage, properties may be much improved. A thin surface crust caused by desiccation can give a misleading impression of the strength of the underlying hydraulic fill. Some hydraulic fills may be susceptible to liquefaction. This vulnerabil ity is a function of particle size and relative density (section 4.7). Load-carrying properties can be improved by increasing the density of the fill, and many improvement techniques involve loading, drainage or compaction (Chapter 8). Pyke et al (1978) investigated the liquefaction potential of hydraulic fills in Los Angeles. Increasing the density of hydraulic fills is also of interest to those concerned with maximising the storage capacity of disposal areas. Most hydraulic fills placed to re claim land specifically for building purposes are coarse. It is unlikely that such fills will present major foundation problems unless there is a large fines content or silty layers are present. Construction on fine hydraulic fills can pose major problems and they can be expected to have poor loadcarrying characteristics due to their high compressibility . Salem and Krizek (1976) have described the stress–deformatio n–time behaviour of dredgings. It is unlikely that fine hydraulic fills will have been used in reclamation schemes where the intention was to build on the reclaimed land; they will gene rally have srcinated in the disposal of dredgings or industrial and mining wastes. Nevertheless, these fills may subsequently be considered for building development and some form of ground treatment may be essential. Preloading, probably combined with drainage measures to speed up the rate of consolidation, could be appropriate.

17

Part II: Engineering behaviour of fills However, as soon as we pass from steel and concrete to earth, the omnipotence of theory ceases to exist. (Terzaghi, 1936)

The objective of the four chapters which comprise Part II is to provide a framework within which the engineering behaviour of fills can be described and assessed. The basic properties that control fill movement are identified and, where possible, simple elastic parameters and relationships are used to describe fill behaviour. Although fills are not linear elastic materials, elastic theory provides a simple basis for settlement calculations which is adequate in many practical situations. Using this basis, typical values of the parameters can be determined from field observations. The practical approach adopted in this book may be criticised from two different perspectives. First, it might be cons idered that many non-engineere d fills are so heterogeneous that the task of providing a framework of engineering behaviour is neither profitable nor, indeed, feasible; this represents an overly pessimistic view as the following chapters of the book demonstrate. Secondly, it might be considered that the approach is too simplistic. The approach adopted is essentially experience- based; despite the great advances in theoretical soil mechanics since 1936truth. it is believed that Terzaghi’s dictum quoted above still contains much The experiencebased approach is particularly appropria te for non-e ngineered fills. The variability of many fills means that it is not worthwhile to adopt more complex soil models, such as those of critical state soil mechanics, and, in many situations, it is unnecessary. In many respects, the engineering behaviour of fills is similar to that of natural soils. For example, the principle of effective stress, which states that strength and deformation of soils are controlled by effective stress, is as relevant to fills as it is to natural soils. It would not serve any useful purpose to repeat the type of information given in an elementary textbook of soil mechanics, and the following chapters emphasise those aspects of fill behaviour that are distinctive to fills. The development of soil mechanics as a rational and coherent scientific discipline in the earlier part of the twentieth century was largely associated with deep foundations for heavy civil e ngineering structures. The subject therefore was concerned with the behaviour of soil at depth under high stresses and usually in a saturated state. This might seem sensible on the premise that deep foundations for heavy structures presented the major challenge. Harding and Glossop (1951) reflected this view: ‘Since the construction of shallow foundation s is gene rally a simple matter, this paper deals chiefly with deep foundations, such as may be required for heavy structures on sites where ground of adequate bearing capacity occurs at some depth below the surface.’

In reality, it is usually easier to predict behaviour at high stresses than at low stresses, and certainly far simpler to predict the behaviour of saturated soil at depth than the behaviour of partially saturated soil close to the ground surface. Cooling and Ward (1948) identified some of the problems of shallow foundations:

18

Part II: Engineering behaviour of fills ‘In the last five years a very large number of examples of building damage due to foundation movement has been encountered in which the effect of the weight of the building on the ground is negligible. Broadly speaking the damage arises from movement of the ground below the level of the foundations due to either climatic changes or the action of he at from industrial buildings.’

It might be argued that low-rise buildings are less important because the value of the structures is much less. While this is true of buildings viewed individually, it is not the case when low-rise buildings are considered as a generic entity. In some years the value of insurance claims for subsidence and heave damage has exceeded £500 million. This account of the enginee ring behaviour of fills begins in Chapter 3 with a consideration of the relevant physical properties. V olume changes in fills are often the major hazard when building on fill, and the causes and likely magnitudes of the resulting movements are considered in Chapter 4. Special emphasis is given to the phenomenon of collapse compression on wetting as this often represents the major hazard for buildings on fill. This is the subject of Chapter 5. In Chapter 6, the problems associated with the edges of filled areas and variations in depth of fill are e xplored.

19

Chapter 3 Properties of fills Working at Robert College in Constantinople from 1916, he [Terzaghi] made a study of those physical properties of soils which control their mechanical behaviour under stress and he succeeded in establishing the fundamenta l principles on which modern soil mechanics is based. (Cooling, 1945)

Those basic modes of soil behaviour which are associated with the principle of effective stress are as applicable to fills as they are to natural soils. However, it is helpful to compare and contrast the properties of fills with the properties of natural soils with the following factors in mind:  Nature of the material The fill may be composed of the same type of material as a natural soil (such as clay, sand or rock), but it could be composed of something quite different (such as wastes from chemical and industrial processes).  Method of deposition The fill may have been deposited in a manner that is quite similar to the deposition of some natural soils (for

stresses and ground movements are calculated to determine whether or not the allowable bearing capacity of the ground will be exceeded and excessive settlement will be caused. However, it has long been appreciated that this type of problem is not the dominant one for low-rise building foundations (Cooling and Ward, 1948). Most problems are due to ground movements caused by factors not related to the weight of the building. In these situations the building is subjected to imposed deformations which are modified by the relative stiffness of the foundation and the soil. The deformations may srcinate close to the ground surface (for example, seasonal movements of a shrinkable clay) or may be deep-seated ( for example, mining and t unnelling

example sedimentation water), but it(such may have beenslow deposited in a quiteunder different manner as compaction with a heavy vibrating roller).  Age Many fills have been placed quite recently and are much younger than natural soils.

movements). suchat as collapse compression on wetting canPhenomena produce effects both shallow depths and large depths. Shallow foundations for low-rise buildings are rarely on saturated ground and an understanding of the behaviour of partially saturated soils is fundamental to the design of such foundations. Major developments have taken place in the study of partially saturated soils and advanced numerical models have been developed. However, the subject appears to be largely the preserve of academics, and the utility of these developments to the design of low-rise building foundations has yet to be demonstrated. Low-rise buildings typically apply relatively small stress increments to soil close to the ground surface and, consequently, soil behaviour at low stresses is relevant to foundation design. The behaviour of partially saturated soils at low stresses is particularly complex and is not addressed in most standard soil tests. The properties of fills are examined under the following headings:  characteristics of the deposit  index and classification properties  compactness  stiffness and compressibilit y  shear strength  dynamic properties  permeability

Fills exhibit a range of engineering properties quite as wide as that of natural soils. For example, there is a great difference between the behaviour of an engineered heavily compacted sand-and-gravel fill and recently placed domestic refuse. The engineering properties of hydraulic fills can be expected to be very different from those of fills placed at lower moisture contents in dry conditions. The behaviour of natural soils and rocks may be strongly influenced by structure arising from cementation or bonding together of particles. Le roueil and Vaughan (1990) pointed out that, although structure arises from many different causes, its e ffect on behaviour is similar. Strength is increased and the stress domain over which the soil exhibits stiff behaviour is enlarged. In fills formed from natural soils, such structure will have been largely destroyed during excavation and placement of the fill. Often construction takes place soon after fill placement, and there may have been little time for structure to develop in the fill. Consequently, fill behaviour may be inferior to that of apparently similar natural soils. In a typical model for assessing foundation behaviour, loads are imposed on the ground surface and the resulting

The following sections are concerned with the physical properties of fills, be cause these physical characteristics

20

Chapter 3 Properties of fills

are generally the most significant in determining the load-carrying characteristic s of a fill. However, with some waste fills, chemical, and in certain cases biological, properties and processes can be of importance. Barry (1991) listed the following types of hazard: toxicity, carcinogenicity, corrosivity, combustibility, inflammability and explosiveness, and asphyxiation. The effects can be grouped under a number of headings:  hazards for buildings, from volume changes in fills, caused by chemical reactions  hazards for construction materials, because of their vulnerability to aggressive ground conditions

Age



There may be properties of the fill itself that change simply with the passage of time without any alteration in external conditions. Schmertmann (1991) reviewed the mechanical ageing of soils, and concluded that ageing effects occur in nearly all soil types and can be responsible for large improvements in pe rformance. Age strengthening is not ne cessarily associated with chemical bonding or cohesion, but could often be caused by mechanical processes such as grain slippage and increased interlocking. Age can have a major effect on further movements caused by the phenomena connected with fill placement. The rate of long-term settlement under self-weight is related to the time that has elapsed since fill placement; creep compression of granular fill and secondary compression of a saturated clay fill are described in section 3.4. The length of time that has elapsed since fill

hazards for buildings and occupants, arising from combustion and gas generation  health and safety hazards for people on site  environmental hazards. It is not the purpose of this book to review all the various chemical and biological hazards, but some matters are given consideration as follows:  volume change (reduction) caused by biodegradation (section 4.5)  volume change (principally expansion) caused by chemical reactions (section 4.6)  contaminated ground (section 10.5).

3.1 Characteristics of fi ll dep osits Some of the load-carrying characteristics of a fill are related thetonature and extent of the as such, and notto just the material of which it isdeposit composed. Surface extent and depth

The boundaries of the filled area need to be established reliably. With well defined infilled excavations such as old docks this may be simple, but with infilled excavations resulting from extractive industries it may be more difficult. The depth of the fill is important in assessing its behaviour, and in evaluating the applicability of various ground treatment techniques and foundation designs. A simple depth classification is proposed:  shallow <3 m  medium 3 m to 10 m  deep >10 m Abrupt changes in depth may lead to differential settlement. Case history 4 shows the effect of changes in depth of an opencast backfill on observed surface settlement. Some old excavations such as docks have vertical sides, and it may be nece ssary to excavate the sides to a flatter slope to reduce subsequent settlement. The differential settlement that can occur at a dock wall is illustrated by case history 24. When a road embankment was constructed across a dock in Hull, the dock wall was demolished to reduce differential settlement (case history 23). The problems assoc iated with the edges of filled areas and changes in depth are e xamined in Chapter 6.

The age of a fill deposit can have several different types of effect on its enginee ring behaviour. As time passes, a fill deposit may be subjected to various changes in its environment which affect its engineering behaviour. An old fill may have already been subjected to many changes in moisture content, water level and loading conditions, which may make it less vulnerable to large movements under future changes. Other time-related phenomena affecting a fill deposit include desiccation and weathering leading to the formation of a surface crust.

placement proportion is of major importance for fillsmaterials containing a significant of biodegradable where settlement is largely related to decomposition. Case histories 21 and 22 describe measurements of the large settlements that can occur with recently placed domestic refuse. Method of placement

The mode of formation will have a major influence on subsequent behaviour . Fills may have been placed in thin layers and heavily compacted or they may have be en end-tipped in high lifts in dry conditions or into standi ng water. The method of placement affects the density of the fill and the homogeneity of the de posit. Many of the problems with non-engineered fills are related to their heterogeneity. Depending on the method of placement and the degree of control exercised during placement, there may be variability in materials, density and age. Much of the usefulness of ground treatment techniques (Chapter 8) is related to reducing this variability. Hydraulic transport and deposition of materials generally produces fills with high moisture content in a relatively loose or soft condition. Segregation of different particle sizes is likely. Ground-water level

The ground-water level within the fill will be controlled by a variety of factors:  hydrogeology of surrounding natural ground

3.2 Index and classification properties formation of an impermeable crust on the surface of the fill  permeability of the fill  any drainage layers or pipes installed within or at the base of the fill. 

The current level of the water table, the amount by which it has fluctuated in the past and the possibility of future fluctuations may have important consequences. A rise in the water level which inundates for the first time a partially saturated, poorly compacted fill may cause collapse compression in the fill (Chapter 5). Lowering the water table can cause settlement due to the increase in effective stress within the fill (Chapter 4). Stress history

The stress history of a fill is usually simple and largely depends on the following factors:  placement conditions  subsequent changes in ground level  fluctuations in ground-water level  weight of buildings  any previous construction on, or occupation of, the site. Where a fill has been preloaded it is in an overconsolidated condition and should exhibit improved load-carrying properties.

21

particle size distributi on is finer than a grain size of 10 mm. The coefficient of uniformity ( CU) is defined by:

CU = D60/D10 There are basic differences in behaviour between coarse granular soils and fine cohesive soils. Coarse soils tend to have high shear strength and permeability whereas fine soils generally have lower strength and permeability. The percentage of silt and clay size particles ( ie finer than 0.06 mm), Fc, is important as when this percentage is high (usually taken as Fc > 35%), the soil will cease to behave as a coarse soil. In re ality, the distinction between granular and cohesive behaviour is a function of other properties in addition to particle size. With a well graded soil, it may not be immediately obvious whether the behaviour will be that of a coarse or a fine soil. The Code of Practice for site investigations gives helpful advice on this transition (BSI, 1999). If the soil deposit is heterogeneous it will be difficult to represent it by a grading curve. The behaviour of granular fills will be influenced by their particle size distribution. Th ey can be described as well graded where there is a wide range of particle sizes, or as uniformly graded where all the particles are within a narrow range. An uncompacted clay fill composed of lumps of stiff clay may behave more like a granular fill than a cohesive fill when loaded, but if inundated it undergoes collapse compression and subsequently behaves like a soft saturated clay fill.

3.2 Index and cl assification p roperties

Moisture content and degree of saturation

Simple index and classification properties enable fills to be classified in ways that are relevant to their loadcarrying characteristics. BS 1377-2:1990 describes classification tests for soils.

Water is invariably present in fill materials. The moisture content or water content , w, of the soil is the mass of water expressed as a fraction or percentage of the mass of the solid particles. Moisture contents for saturated fills typically are from 20% to 80%. BS 1377-2:1990 describes the definitive method of moisture content determination by oven drying at 105 to 110 °C. The placement moisture content affects the properties of fills, and the behaviour of clayey fills is very moisture dependent. The fill moisture content when building development takes place may be similar to that at placement or may have changed due to consolidation of the fill, penetration of surface water, surface drying or movement of ground-water. Moisture content changes can cause volume changes in fills (section 4.4 and Chapter 5). The degree of saturation, Sr, is the ratio of the volume of water to the volume of pores and is usually expressed as a percentage; Sr = 100% indicates a fully saturated soil. The degree of saturation can have a major influence on both collapse potential and liquefaction potential. For loose, partially saturated fills, the volume of air voids, Va , expressed as a perce ntage of the total volume of the soil is also a useful parameter:

Nature of material

Fills can be formed from excavated and redeposited natural soils and rocks, or from the waste products of human activities. Many fill materials are inert under normal conditions but, where this is not the case, the nature of the fill material may be of particular importance. Chemical contaminants may present hazards during investigation and subsequent use of the site. The presence of a significant amount of biodegradable material will have an adverse effect on load- carrying characteris tics as well as causing problems due to gas generation ( section 4.5). A large coal content in a colliery spoil could give concern in relation to combustion if there was a supply of air through large air voids in the fill (section 2.2). Particle size distribution

Particle size has a strong effect on soil behaviour and provides a useful method of classification. Sieving and sedimentation can be used to dete rmine the particle size distribution. Th e distribution can be described by a number of values of Dx where Dx is the grain size such that x % by mass of the particle size distribution is finer than Dx. Thus, D50 = 10 mm means that 50% of the

{

Va = 1 –

ρd ρw

[

ρw ρs

+

w 100

] } 100%

22

Chapter 3 Properties of fills 3

where ρw is the density of water (1.0 Mg/m ). Shallow foundations will usually rest on partially saturated ground and the behaviour of soils in this condition is a key factor in foundation performance. Soil above the water table is not subjected to a positive hydrostatic pressure; evaporation and evapotranspiration may lower the pore water pressures sufficiently to draw out water from the voids in the soil near the ground surface. Thus a zone is created, part of which is saturated and part of which is partially saturated, within which the pore pressures are negative. Suction in fine soils can result in an apparent cohesion with

A relationship of this form can only be very approximate for the following reasons:  different remoulded clays show somewhat different correlations between cu and IL  the relationships are not strictly linear  the measured values of cu depend on the test method and the rate of test.

enhanced resistance to particle movement. However, small perturbations in ground conditions can de stroy the suction and this can lead to ground movements. Collapse compression on wetting is described in Chapter 5.

strengths. For a normally consolidated soil (hydraulic fills as placed are normally consolidated) cu is related to the effective overburden pressure, σ'v, and typically the ratio cu/σ'v lies in the range from 0.2 to 0.4. With a high water table effective stresses in a hydraulic fill will be low. Some desiccation is likely to occur at the surface of the fill, forming a stiff crust.

Liquidity and plasticity indices

In this context, the term plasticity describes the response of fine fills to changes in moisture content. The plasticity of a soil is related to the amount of water required to be added to change the consistency of the soil from hard and rigid to soft and pliable. The wetter end of the plasticity range is described by the liquid limit ( wL), and the drier by the plastic limit (wP). The liquid limit is the moisture content beyond which the soil exhibits ‘liquid’ behaviour; conversely, the plastic limit is the moisture content below which the soil loses its intact behaviour and begins to break up into discrete pieces. The moisture content of

Plastic silt and clay hydraulic fills are cohesive and can be characterised by a profile of undrained shear strength versus depth. Some typical profiles are plotted in Figure 7 and it is seen that these fills generally have very low

fine soils may be related to (the limit by the liquidity index IL):liquid limit and the plastic

IL = (w – wP)/(wL – wP) This property of cohesive fills is analogous to the density index of coarse fills. The plasticity index ( IP) is defined as follows:

I P = wL – wP The methods for the determination of IP and IL are described in BS 1377-2:1990. The moisture content used to calculate IL should be determined on the same soil fraction that is used to determine wP and wL (that is the fraction passing the 425 µm sieve). Undrained shear strength

Undrained behaviour is of practical importance for saturated clay fills when applied loads change faster than induced pore pressures can dissipate. Undrained shear strength ( cu) is a function of moisture content. For many remoulded clays there is an approximately linear relationship between log cu and IL, and cu can therefore be considered as an index property. As a very crude guide: log cu = 2(1 – IL) where cu is measured in kPa.

Figure 7 Strength of cohesive hydraulic fills

Particle strength

The strength of individual fragments of rock has some influence on the behaviour of rockfills. Rocks are affected by wetting and drying; the strength of saturated rock fragments may be much smaller than the strength of the fragments when dry. Weak rocks such as shales will swell or disintegrate when exposed to atmospheric wetting and drying and are thus permanently weakened . Strength can be measured in several ways and helpful background information can be found in Eurocode 7 Geotechnical design — Part 2 Design assisted by laboratory testing (BSI, 2000a) and Engineering rock mechanics — an introduction to the principles by Hudson and Harrison (1997). Uniaxial compressive strength ( qu) of cylindrical specimens of rock The drawback to this standard test is the time and effort required to prepare the cylindrical sample of rock. Point load strength test Cylindrical specimens or irregular lumps are loaded to

3.3 Compactness failure between the conical platens of a portable tester (Broch and Franklin, 1972). A strength index, IS, is measured such that:

I S = ( P / d 2) where P is the maximum load and d is the length of the rock in the direction of load application. If d is significantly different from 50 mm a correction is applied and there are corrections for different rock shapes. Typically qu = 20IS, but wide variations occur. Strength test in which irregular lumps are crushed between flat hardened steel end plattens Two strength indices were proposed by Hobbs (1963):

I 1 = ( P / A) I 2 = ( P / d 2) where P is the maximum load, A is the average area in contact at failure, and d is the length of the rock fragment in the direction of load application. The relationship proposed by Hobbs (1963) between I1 for irregular specimens of about 30 g in mass and the compressive strength, qu, of 25 mm diameter cylindrical samples, was as follows:

qu = 0.91I1 — 22 where qbeand I arewhen in MPa. this itreislationship cannot correct I1 isClearly small, and preferable simply to consider that I1 is of similar magnitude to qu. The correlation between I1 and I2 is poor, but I1 is about 10 times as large as I2. u

1

Slake durability test The resistance of a rock to wetting and drying cycles can be assessed by immersing samples in water and noting the rate of disintegration (Franklin and Chandra, 1972). This test is useful for relatively weak rocks which are sensitive to the test.

3.3 Compactness The density or compactness of a fill has a major influence on its behaviour. Compactness is a function of the method of placement of the fill and its subsequent stress history. The term ‘compaction’ is used to describe processes in which equipment is used to compress fill into a smaller volume, thus increasing its density. Usually, engineering properties are improved because a dense fill generally has superior properties to the same fill in a loose condition. As solid soil particles and water are virtually incompressib le, compaction is concerned with reducing the percentage air voids. The density may be described by:  voids ratio ( e) — the ratio betwee n the volume of the voids and the volume of the solid particles  porosity ( n) — the volume of the voids expressed as a

23

percentage of the total volume dry density ( ρd) — the mass of solid particles contained in unit volume of soil  dry unit weight ( γd) — the weight of solid particles contained in unit volume of soil  bulk density ( ρ) — the mass of solid particles plus water contained in unit volume of soil  bulk unit weight ( γ) — the weight of solid particles plus water contained in unit volume of soil. 

These parameters are closely related as the following relationships demonstrate:

n = [e/(1 + e)] 100 (%) ρd = [ρs(100 — n)]/100 γ d = ρd g ρ = ρd[1 + ( w/100)] where ρs is the particle density, and w is the moisture content in %. For many natural soils and rocks ρs is in the range 2.6 Mg/m3 to 2.7 Mg/m 3. However, some fill materials have much lower values, for e xample organic wastes, coal and pfa. Some typical fill densities are presented in Table 1. For a fairly clean coarse hydraulic fill, typically the dry density is 1.6 Mg/m3. Coarse fill

In assessing field behaviour, it is helpful to re late the insitu compactness to thattests. determined in standard laboratory compaction The degree of packing in coarse fills can be described by the de nsity index, ID, sometimes known as relative de nsity. Density index provides a useful way of relating the in-situ dry density of a granular fill, ρd, to the limiting conditions of maximum dry density, ρdmax, and minimum dry density, ρdmin:

ID = ρ

ρd — ρdmin

dmax — ρdmin

ρdmax ρd

Methods of determining the maximum and minimum densities of sands and gravelly soils are given in BS 13774:1990. In-situ density tests are described in BS 13779:1990. As these three densities are difficult to determine accurately, the calculated value of ID will be liable to large errors. Nevertheless, it is an important parameter and strongly influences the compressibility, shear strength and vulnerability to liquefaction of granular fills. In Table 1 the engineered sandstone rockfill is heavily compacted and might be used in an embankment dam or a highway embankment. Such fills have a porosity typically in the range 20% to 25%, corresponding to ID in the region of 0.8 to 0.9. In contrast, an uncompacted opencast backfill of similar material has a much higher porosity of about 40%, corresponding to ID of about 0.5 to 0.6. Generally for fairly clean coarse hydraulic fills ID is of the order of 0.5 (Whitman, 1970). However, much lower values can occur in some situations (section 2.8) and the hydraulically placed pfa at Peterborough had a somewhat

24

Chapter 3 Properties of fills

Table 1 Typical field placement densities Placement method Mass per

F i l lt y p e

F i l lm a t e r i a l

Engineered

Clay fill

Layer Number thickness of

metre Typical compacted state width of rollρd w ρs n

( m)

(t /m)

0.2

p a s s es 6

( Mg/ 3m) ( % )

3

1.68

20

Va

( M g/ m3) ( % ) 2.65

37

( %) 3

Non-engineered

Clay fill

—

—

—

1.53

20

2.65

42

12

Engineered

Sand fill

0.25

6

3

1.85

6

2.65

30

19

Non-engineered

Sand fill

—

—

—

1.55

6

2.65

42

32

Engineered

Sandstone rockfill

1.0

4

6

2.03

7

2.65

23

9 28

Non-engineered

Sandstonerockfill

—

—

—

1.60

7

2.65

40

Engineered

Colliery spoil

0.3

6

4

1.86

11

2.55

27

7

Non-engineered

Collieryspoil

—

—

—

1.56

11

2.55

39

22

Engineered

Conditioned pfa

0.3

8

4

1.30

25

2.20

41

8

Non-engineered

Lagoon pfa

—

—

—

1.17

40

2.20

47

0

larger value of ID (case history 12). Because of the difficulties of re liably determining ID, penetration tests and other types of in-situ tests are frequently used to assess the properties of fills (section 7.5). Such tests are not practicable in fills such as rockfills, with large particles, but where they can be used they frequently give valuable information. The standard penetration test, the cone penetration test and the pressuremeter have been commonly used. Many different have design been proposed these tests and correlations ID. Alternatively, strengthbetween and deformation parameters may be directly correlated with in-situ test re sults. Nixon (1982) reviewed much of this work in his state-of-the- art report on the standard penetration test. Lunne et al (1997) have provided a comprehensive review for the cone penetration test. Fine fill

The amount of densification that can be achieved for a given compactive effort on a clayey soil is a function of moisture content. Laboratory compaction test results are plotted as the variation of dry density with moisture content. Such plots are a simple and useful way of representing the condition of a partially saturated fill and can be used to ide ntify the two parameters of prime interest: the maximum dry density and the corresponding optimum moisture content. The percentage ratio of the in-situ dry density to the maximum dry density achieved with a specified degree of compaction in a standard laboratory compaction test is termed the relative compaction ( CR). When quoting a value of relative compaction, the type of compaction must be specified, and this is usually, although not always, the standard Proctor compaction test. Modern procedures for the placement and compaction of fill using simple laboratory and field tests were introduced by Proctor (1933). Figure 8 shows laboratory test results obtained by Proctor (1948) for a clay fill using different compactive efforts. The basic laboratory testing

carried out by Proctor demonstrated some fundamental features of the compaction of clay soils.  The density to which a particular compactive effort will bring a clay fill depends on the moisture content of the fill.  For a given compactive effort there is a maximum dry density which is achieved at the optimum moisture content ( wopt). The expressions ‘maximum dry density’ and ‘optimum moistu re content’ refer to a specified compaction procedure can be misleading if taken out of the context of thatand procedure. At a moisture content dry of optimum, the specified compaction procedure will result in a fill with large air voids.  At a moisture content significantly wet of optimum moisture content, the specified compaction procedure will produce a fill with a minimum of air voids, typically between 2% and 4%. 

These fundamental factors demonstrate that the control of moisture content is crucial for field control of compaction. Holtz (194 8) proposed two moisture content limits for clay fills because it was recognised that:

Figure 8 Compaction of clay fill using different compactive efforts (after Proctor, 1948)

3.4 Stiffness and compressibility a moisture content that was too high could lead to excessive construction pore pressures  a moisture content that was too low could lead to excessive settlement of the fill when the fill became saturated. 

With respect to fills for building developments, the upper limit for placement moisture content corresponds to a fill where the bearing capacity is too low and the settlement produced by the weight of the building is too great. The lower placement moisture content for a clay fill was defined by Holtz as the moisture content at which

25

expression can be used:

Ko = 1 — sin φ' Under normally consolidated conditions, a clay fill might typically have a Ko of 0.6, and a sand fill a Ko of 0.4. For fills within this range of values of Ko:

D = 1.3E to 1.8E Constrained modulus

When a uniform load is applied over a wide area of the

saturation would have no effect on consolidation. However, collapse compression on saturation is re lated not only to the moisture content, but also to the dry density and percentage air voids. Proctor (1948) showed that the energy required to reduce the air voids in the fill to a minimum was related to the undrained shear strength of the soil and that the subsequent compression behaviour was also related to the strength after compaction. BS 1377-4:1990 includes several laboratory compaction test procedures. Clause 3.3 describes a test using a 2.5 kg rammer and clause 3.5 describes a test using a 4.5 kg rammer. The former test is a metricated version of the standard Proctor compaction test with a total energy input of 596 kJ/m 3. The latter test is a metricated version of the modified Proctor compaction test with an energy input of 2682 kJ/m 3.

surface of a fill, the resulting compression is largely one dimensional. One-dimensional consolidation properties have often been described by a compression index, Cc, or a coefficient of volume compressibil ity, mv. In this book the constrained modulus, D, is used. This parameter was proposed by Stamatopoulos and Kotzias (1973, 1978). An application of load will produce an immediate compression in a partially saturated fill. Effective stress is not a simple concept for partially saturated soil and, for most practical purposes, compressibility can best be described by a constrained modulus expressed in terms of total stress. The constrained modulus is the ratio of the increment of applied total vertical stress, ∆σ v, to the increment of vertical strain, ∆εv, which it produces under drained conditions:

3.4 Stiffness and compressibility

The constrained modulus has been de noted by a variety of symbols including M and Eoed. It should be noted that D = 1/mv. The constrained modulus can be measured in oedometer tests. Conventional small oedometers may be used for fine cohesive fills, but coarse fills such as rockfills require large-scale equipment. Side friction may significantly reduce the vertical stress in a large rockfill sample, and Penman and Charles (1976) derived a correction for this. The compressibility of heavily compacted coarse fills (125 mm maximum particle size) measured in a 1 metre diameter oedometer at B RE is shown in Figure 9. The fills were compacted in thin layers with an electric vibrating hammer, all the fills receiving a comparable degree of compaction. The slate and sandstone rockfills were more compressible than the sandy gravel. The sandy gravel was tested at two other initial densities and the major influence of initial density on compression is clearly indicated in Figure 10. In oedometer tests on coarse granular fill materials generally there are no excess pore pressures. In the field this will also be the case during embankment construction. However, if the fill is subsequently submerged, σ'v < σv, and the constrained modulus is best defined in terms of effective stress. Fills do not usually exhibit linear elastic behaviour and the constrained modulus is stress dependent. The value of the constrained modulus will depend on stress history, initial density and moisture content of the fill. The one-

The various causes of volume changes within fills are considered in detail in Chapter 4; some basic background on stiffness, compressibili ty and time-de pendence is presented in this section. When fills are not close to failure, it is often convenient to describe their behaviour using elastic parameters even though fills generally exhibit non-linear non-recoverable stress-dependent behaviour. Conventionally Young’s modulus, E, and Poisson’s ratio, ν, are used to describe linear elastic behaviour. However, these parameters have little direct relevance to the stress conditions that generally obtain for fills. Settlement in many fill situations is closely related to one-dimensional compression, and the constrained modulus, D, is of more direct application. The coefficient of earth pressure at rest, Ko, which is the ratio of minor to major principal effective stress in one-dimensional compression, is also useful in some situations. For an isotropic linear elastic material, Young’s modulus and Poisson’s ratio are related to D and Ko by the following expressions:

E/D = [(1 — Ko)(1 + 2 Ko)]/(1 + Ko) ν = Ko/(1 + Ko)

For normally consolidated fills the following approximate

D = ∆σ v/∆εv

26

Chapter 3 Properties of fills dimensional compression behaviour of many well graded granular fills can be approximated to a relationship of the form: σv = a εv2

where σv is the applied vertical stress and εv is the induced vertical strain. The secant constrained modulus for a stress increment from 0 to σv can be expressed as:

Dsec = σv/εv= a0.5σv0.5 Similarly the tangent constrained modulus, Dtan, at a stress σv is given by:

Dtan = ∆σ v/∆εv= 2a0.5σv0.5 Figure 9 Compressibility of heavily compacted granular fills measured in 1 metre diameter oedometer (H is the height of the embankment built of the fill; points marked ‘a’ for the different fills indicate the vertical stress, σv, such that σv = γH for the embankment built of that fill; DγH is the secant constrained modulus with σv = γH; D* is the constant equivalent constrained modulus for the embankment of height H)

The ratio Dtan/Dsec is 2 which illustrates the importance of specifying the type of modulus and the stress range when quoting moduli for non-linear fill behaviour. The above two equations can be expressed in terms of a normalised constrained modulus, Dsecn, as suggested by Schanz and Vermeer (1998). If the normalised secant modulus Dsecn at a specified vertical stress σvn is used as a reference, then:

Dsec = Dsecn (σv/σvn)0.5 If σvn is taken as 0.1 MPa and all the stresses and moduli are in MPa then the above equation can be represented as:

Dsec= 3.16 Dsecnσv0.5

Figure 10 Compressibility of sandy gravel fill measured in 1 metre diameter oedometer a γd = 21.5 kN/m3 ID = 0.82 b γd = 21.2 kN/m3 ID = 0.77 c γd = 17.4 kN/m3 ID = 0

The compression behaviour of some granular fills is shown in Figure 11 in the form of plots of Dsecn versus ID. The data for sands from Schultze and Moussa (1961) are plotted together with data from tests carried out in large oedometers on granular fills with large particle sizes (Penman and Charles, 1985a). These fills were compacted in thin layers using an ele ctric vibrating hammer. The constrained modulus of the sandy gravel was typically four times as large as the constrained modulus of the sandstone rockfill. This large difference can be attributed to greater crushing at points of contact of the angular fragments of sandstone than in the more rounded gravel fill. It should be noted that tests on uniformly graded rockfills have sometim es given a more linear type of relationship. It would seem that although the rock material may be quite strong, the absence of finer material means that inter-particle loads are high and rock fracture occurs as the applied stress is increased. This tends to prevent the increase in modulus that would otherwise occur. Compacted clay fills may show similar stress–strain behaviour in one-dimensional compression. Dsecn is plotted against cu for some clay fills in Figure 12. There is an approximately linear relationship between Dsecn and

27

3.4 Stiffness and compressibility Sandstone rockfill (Charles, 1973)

Constrained modulus for embankment construction

Gravel fill (Penman and Charles, 1985a)

Although the strains in coarse fills which occur as fill is placed are not of much significance for subsequent construction on the fill, some consideration is given to them in this section. The maximum construction settlement ( smax) of a wide embankment of height H built of fill of constant constrained modulus D will occur at mid-height and is given by:

Range for sands (Schultze and Moussa, 1961)

smax = 0.25 γH 2/D Generally, the stress–strain relationship is non-linear, but a constant equivalent constrained modulus, D *, can be chosen which gives the correct maximum settlement at mid-height, and which also predicts settlement at other heights to an acceptable accuracy (Penman et al, 1971). If the stress–strain relationship measured in the oedometer is of the form σv = aεv2:

D * = 1.28 Dsec = 1.28 (aγH )0.5 Figure 11 Relationship between reference constrained modulus and density index for coarse fills

where Dsec is the secant constrained modulus with σv = γH. Generally it is found that 1.0 < D */Dsec < 1.28. The constrained moduli of coarse fills from several dams in the UK have been measured in a 1 metre diameter oedometer. Values calculated for Dsec and D * for some of these dams are shown in Figure 9. Constructional settlements have been measured by field monitoring, and Charles and Penman (1988) have shown that the maximum constructional settlement, smax, is related to D * by an e mpirical relationship:

smax = 0.30 γH 2/D *

Figure 12 Relationship between reference constrained modulus and undrained shear strength for clay fills (after Proctor, 1948)

log c . Nwabuokei and L ovell (1985) have examined the u compressibilit y of such compacted fills. The following should be noted:  large compressions may occur  excess pore pressures induced in saturated clay by the loading decay slowly. Where a fill has significant collapse potential, there may be an almost linear stress–strain relationship prior to wetting. However, following collapse compression on wetting, the stress–strain relationship usually will be close to the parabolic form described earlier. Collapse compression is examined in Chapter 5.

This is a little larger than the 0.25( γH 2/D *) which would theoretically be expected in a wide embankment. The difference could be associated with the following factors:  the difference between the laterally confined conditions in the oedometer and the absence of lateral restraint at the slopes of the embankments  the difference between the compressibilit y of the 125 mm maximum size rockfill tested in the oedometer and the compressibility of the larger material in the field. Both these effects would tend to increase the settlement. Some typical values of D * are given in Table 2. Table 2 Typical values of constant equivalent constrained modulus (D*) for various fill types and embankment heights (H) D* (MPa) Fill type H = 10 m H = 30 m H = 100 m Sandy gravel (ID0.8) =

50

90

170

Sandy gravel (ID0.5) =

30

50

90

Sandstone rockfill (ID0.8) =

15

25

45

Sandstone rockfill (ID0.5) =

6

10

20

Clay fill (IP = 15%, IL0.1) =

6

10

18

28

Chapter 3 Properties of fills

Constrained modulus on unloading and reloading

Table 3 Typical values of secant constrained modulus (Dsec) and induced strain (εv) for stress increment∆σv = 100 kPa from σvo = 30 kPa Fill type Dsec (MPa) εv (%)

Although in many cases it is only the stiffness of the fill as it is first loaded that is of practical significance, in some situations behaviour during unloading and reloading is important. Unloading could occur with removal of fill or its submergence. In the latter case, fill behaviour may be affected by collapse compression (Chapter 5). Where a layer of fill is removed there will be some heave of the remaining fill. However, the stiffness of the fill on unloading may be much greater than the stiffness on first loading. Any subsequent reloading will be controlled by a stiffness similar to that on unloading, until the previous maximum stress is exceeded . In general, fills are much stiffer on unloading and reloading than on initial loading. Figure 13 shows the compression on first loading of three samples (a, b and c) of a sandstone rockfill compacted to different initial densities, and the unloading and reloading behaviour of the sample in the loosest condition (a). The behaviour of the loosest sample on reloading (ar) has been re plotted from the srcin, as though it were a new sample, to show that it is much stiffer than even the most heavily compacted sample (c) on first loading.

Sandy gravel fill (ID0.8) =

50

0.2

Sandy gravel fill (I0.5) D=

25

0.4

Sandygravelfill(preloaded)

200

0.05

Sandstone rockfill (ID0.8) =

12

0.83

Sandstone rockfill (ID0.5) =

6

1.67

Sandstonerockfill(preloaded)

40

Collieryspoil(compacted) Collieryspoil(uncompacted)

6 3

Clay fill (IP =15%,0.1) IL = Old urban fill Old domestic refuse Recentdomesticrefuse

0.25

5

1.67 3.3 20

4

2.5 3 1

3.3 10

have been derived from the laboratory oedometer tests described earlier in this chapter, and from field loading tests described in Chapter 4. Foundation loads usually significantly increase the stress within the fill only to a shallow depth. The constrained modulus near the surface of the fill can be affected by surface desiccation which can have an important effect in stiffening the surface layer of fill. An increase in D as a result of preloading is of considerable significance for foundation performance. Shear modulus

The shear modulus, G, is defined in terms of stress difference and is therefore independent of pore pressure. As a consequence, it is not a function of drainage conditions, and is the same for drained and undrained conditions. It is important in describing the behaviour of saturated clay fills. For linear elastic materials the relationship between E and G is:

E/G = 2(1+ν)

unloading and reloading

For saturated clay fills under undrained conditions: ν = 0.5 and G = E/3. In this situation D = ∞. Shear modulus, like constrained modulus , is dependent on initial density and moisture content, stress level and stress history. The shear modulus at low strain is examined later in this chapter, in the section on dynamic properties (section 3.6).

Constrained modulus for foundation loading

Time dependency for fine fill

The stresses induced by loading a small foundation do not strictly correspond to one-dimensional compression, but where foundation loads apply stresses to granular fills that are much lower than those required to cause bearing capacity failure, settlement can be calculated using a constrained modulus. The appropriate modulus is a secant modulus corresponding to the stress increment due to the applied foundation loading. T able 3 presents some typical values of D for this type of situation. These

A fundamental difference between the de formation behaviour of fine fills, such as clay fill, and coarse fills, such as sand fill and rockfill, is related to permeability. Although the permeability of fills is considered later in this chapter (section 3.7), some consideration of its effect on deformation behaviour is required here. The loading of a saturated clay fill initially induces excess pore pressures in the clay and there may be little immediate settlement. The fill only compresses in

Figure 13 Compressibility of sandstone rockfill during loading,

3.4 Stiffness and compressibility response to the increase in applied loading as water is squeezed out of the clay and this reduction in moisture content occurs slowly owing to the low permeability of the soil. Consequently, the initial effect of applying a load to a saturated clay fill is to increase the pore pressure, without affecting the effective stresses within the soil or causing immediate compression. Tim e-depe ndent movements of two types then occur: primary consolidation followed by secondary compression. Primary consolidation The flow of water out of the affected area is controlled by

the hydraulic gradient generated by the excess pore pressures and the drainage conditions at the soil boundaries. As the water flows away from the loaded area, the pore pressures dissipate and the soil compresses. This process is known as primary consolidation; it slowly continues until all the excess pore pressures have dissipated. The amount of primary consolidation can be described by the constrained modulus, D, and the rate of consolidation by the coefficient of consolidation, cv:

cv = kD/γw where k is the coefficient of permeability in m/s D is the constrained modulus in kPa γw is the unit weight of water in kN/m 3 cv is in m 2/s Secondary compression Primary consolidation is followed by secondary compression which continues under conditions of constant effective stress. The secondary compression of saturated clay fills exhibits a linear relationship between compression and the logarithm of the time that has elapsed since the load was applied. The continuing rate of compression of clay fills in one-dimensional compression can be described by the coe fficient of secondary compression, Cα:

Cα = ∆h/[h log ( t2/t1)] where compression ∆h/h occurs between times t2 and t1 after the load was applied. Time dependency for coarse fill

If a load were applied to a saturated coarse fill totally surrounded by an impermeable barrier, the immediate response would be similar to that of a clay fill; the applied load would be carried by the increase in pore pressure without affecting the effective stresses. While such conditions can be simulated in the laboratory, the drainage conditions and dissolved gases in the pore water that exist in the ground allow the water to flow and/or compress instantaneously . Consequently, coarse fills do not have a time-dependent consolidation of the type observed in clay fills. With coarse granular fills, most of the deformation occurs immediately the load is applied. However, nearly all coarse fills do have a component

29

of long-term compression that occurs under conditions of constant effective stress and this is usually termed ‘creep’. Although the movements caus ed by creep are relatively small, in many circumstances it is these longterm movements that are of most interest for foundation performance. Many coarse fills show a linear relationship between settlement and the logarithm of the time that has elapsed since the fill was placed. Sowers et al (1965) observed this behaviour in embankment dams built of dumped rockfill and described the behaviour in terms of a logarithmic creep compression parameter, α , such that: α = ∆s/[H log ( t2/t1)]

where ∆s is the settlement of an embankment of height H between times t2 and t1 since construction. Creep compression can occur because of the weight of the structure built on the fill as well as because of self -weight of the fill. The expression for the creep parameter, α, can also be represented by the following equation: α = 2.303 t (d εv /dt ) × 100%

where d εv/dt is the rate of vertical compression strain occurring at time t after the end of construction. This type of approach was termed the velocity method by Parkin (1978) who advocated its use in the interpretation of laboratory consolidation tests. Time relationships Somedependency: other types of other relationship between creep compression or settlement and time have been proposed. From visco-elastic theory the compression rate can be related to the damping ratio, Dr (Abbiss and Lewin, 1990):

(d ln ε)/(d ln t ) = 4Dr/π Damping ratio is described later in the chapter. This equation implies that: ε = A t (4Dr/π)

where A is a constant. Edil et al (1989) have successfully used a power law relationship of this type to describe the time-dependent settlement of municipal refuse. If the damping ratio is less than 5%, a plot of ε versus log time would show relatively little departure from linearity over three log cycles of time. In this case, the power law relationship predicts behaviour that can reasonably be described by an α parameter. The results of Edil et al (1989) for refuse indicated much larger values of damping and, if the power law correctly describes the be haviour of refuse fill, the use of an α parameter will give a poor approximation to actual behaviour. Gibson and Lo (1961) developed an expression for the settlement, s, of a clay soil of depth H exhibiting secondary compression under an applied stress, q,

30

Chapter 3 Properties of fills

such that: s = qH [a + b(1 — e—λt/b)]

applied loads will be much smaller than the failure load and bearing capacity will not be a problem; the allowable settlement is usually exceeded be fore soil rupture considerations become significant (Sutherland, 197 4). An example of the magnitude of the bearing capacity of a granular soil has been provided by Ortigosa et al (1981), who reported that the allowable contact pressure for foundation design on the Santiago gravel has been increased to 1180 kPa. Allowable bearing pressures should be specified that will provide an adequate factor of safety against soil rupture and keep total and differential settlement within tolerable limits.

When t = 0, si = qHa; when t = ∞, s = qH(a + b ). This relationship also has been used by Edil et al (1989) to describe the settlement of municipal refuse. Hagenaar and Wolsleger (1986) proposed a relationship of the following form for the settlement of hydraulic fills:

s = t/(a + b t ) When t = 0, d s/dt = 1/a; when t = ∞, s = 1/b. In general, long-term settlement or compression is plotted against the logarithm of time. The use of the logarithmic creep compression rate parameter α is based on the assumption that this plot will be linear. If the visco-elastic relationship proposed by Abbiss and Lewin (1990) is valid, then the plot will be non-linear with the gradient α increasing with log time. The relationships proposed by Gibson and Lo (1961) and Hagenaar and Wolsleger (1986) both predict a non-linear plot with the gradient α finally decreasing with log time so that settlement reaches an ultimate value.

3.5 Shear strength Most geotechnical problems that arise with construction on fill concern settlement, which is often associated with volume in theasfill. However, shear strength can also be ofchanges significance it controls the bearing capacity of foundations. It also has a major effect on slope stability, earth pressure and side friction of the fill on adjoining structures or piles. Shear strength may be measured in the laboratory on undisturbed or recompacted samples of fill under either drained or undrained conditions. Alternatively, in-situ testing techniques may be used. The shear strength of fills is a large subject and is only briefly reviewed with respect to the following types of fill:  engineered granular fills  engineered clay fills  non-engineered fills  hydraulic fills Engineered coarse fill

This category includes sand fill, gravel fill and rockfill when placed in layers and heavily compacted. These coarse fills usually have high permeability and the drained angle of shearing resistance, φ', is the significant shear strength parameter. It is assumed that there is no cohesion in granular fills ( c' = 0) and therefore: sinφ' = [( σ'1/σ'3)max — 1]/[(σ'1/σ'3)max + 1] where ( σ'1/σ'3)max is the maximum principal effective stress ratio. The angle of shearing resistance of heavily compacted granular fills is high. As a consequence, they usually have a large bearing capacity. In most practical situations the

The strength of engineered granular fills can be determined by laboratory tests on samples re compacted to the field density. The drained shearing resistance is strongly dependent on dilatancy and he nce is a function of ID, stress level, stress history and strain conditions in the test. In fill materials the stress history is usually simple and the fill is normally consolidated. In a dense state the failure envelope shows marked curvature, while in a loose state the failure envelope may be virtually linear (Charles and Watts, 1980). It is important that laboratory testing is carried out at stresses which correspond to the range of stresses which will be encountered in the field. Dense fills tested at low confining pressures exhibit a peak strength that is associated with strongly dilatant behaviour. This is followed by strain softening to a constant volume or critical state strength φ'cv. There is no evidence that this type of brittle behaviour in coarse fills leads to progressive failure unless they contain or clay-type materials. The effect is suppressed at shale very large confining pressures. The shear strength, therefore, has two components:  a basic angle of shearing resistance which is commonly identified with the constant volume or critical state or angle of shearing resistance, φ'cv, but occasionally with an angle of friction, φµ  a component due to dilatancy. Strength testing is most commonly carried out under axisymmetric conditions in triaxial testing equipment. Drained tests are often the easiest to perform and most closely resemble field conditions, but in some situations undrained testing may be required. The direct shear test is appropriate for testing free -draining fills. For many granular fills the difference between φ' measured on dry and saturated samples is relatively small. Strain rate should not greatly affect the measured strength. In many field situations associat ed with, for example, embankments, retaining walls, deep excavations and foundations, plane strain conditions apply. Plane strain laboratory tests have shown greater shear strengths than the corresponding triaxial tests. The use of triaxial test parameters is therefore conservative and may offset any small over-estimate of strength which might be attributed to scale or density effects. Some suggested relationships between the plane strain angle of shearing resistance (φ'ps) and the triaxial angle of shearing resistance, φ'ax, for sands are as follows.

3.5 Shear strength Danish Code of Practice (Stee nsen-Bach, 1989): φ'ps = 1.1φ'ax  Schmertmann (Leonards and Frost, 1988): φ'ps = φ'ax + 0.5( φ'ax —32°) (when φ'ax >32°) φ'ps = φ'ax (when φ'ax < 32°)  Bolton (1986): φ'ps = φ'ax +(2/3)( φ'ax — φ'cv) typically φ'cv =32° to 37° 

31

chapter, the value of cu depends on test method and rate of testing, and the value measured in the laboratory is not necessarily identical to that governing field be haviour. Non-engineered fill

Fine fills, such as clay fills, should be compacted in thin layers to achieve low air voids. If the initial Va is more than 5%, there is a possibility that, after the fill has bee n

Although these fills can be very variable and difficult to characterise, with many of them bearing capacity is not a problem. Settlement due to re duction in volume of the fill is generally the critical factor for foundations. Representative samples for testing are not easily obtained and in-situ geote chnical and geophysical tests may be more appropriate methods of investigation.

loaded, softening with associated collapse compression will occur on first wetting of the fill. The clay fill may be specified in terms of its placement undrained shear strength, typically 50 kPa < cu < 120 kPa. If the field placement strength is much below 50 kPa there may be problems in placing the clay; on the other hand, if the field placement strength is much above 120 kPa it may be difficult to compact the clay adequately and reduce the air voids to 5%. As an engineered clay fill should be nearly saturated, excess pore pressures are likely to be set up during placement due to the weight of overlying fill. Primary consolidation will occur as these pore pressures slowly dissipate with time. In the construction of embankment dams drainage layers are frequently incorporated into clay fills to speed up consolidation. Vaughan et al (1978) reviewed the stability of clay fills.

Standard penetration tests have often been used to investigate such fills but the results can be difficult to interpret. The cone penetration test is now often used in appropriate types of fill. Non-engineered fills are likely to be in a loose condition. Consequently, their strength may be close to the constant volume angle of shearing resistance, φ'cv. For a granular fill this angle of shearing resistance is usually similar to the angle of repose. Some typical ranges of values of φ'cv are given in Table 4. A number of investigations of the shear strength of domestic refuse have been reported. These fills may have been engineere d in terms of waste disposal factors, bu t are best considered as non-enginee red in terms of their engineering properties. Large direct shear tests on some old refuse in Canada provided quite high values of φ', of the order of 35° (Landva and Clark, 1989). In contra st, the

Two types of fill, clayIfill 27%, residual strength much below peak strength.

measured shear strength shredded φrefuse much smaller. Duplancic (1989)ofmeasured ' = 34°was in triaxial compression tests on a landfill material containing waste from various industrial and municipal sources at a site in California. Siegel et al (1989) measured some very high strengths on refuse in Southern California, but expressed caution about the validity of direct shear test results where relatively strong pieces of re fuse may have become wedged betwee n the shear boxes, producing artificially large values. Van Impe (1998) has tabulated the results of a large number of published investigations which show a wide range of values of φ', mainly between 20° and 40°. The low shear strengths that may be operative on interfaces within multi-layer liner systems have been described by Mitchell et al (1989). If instability should occur, loose fill materials may be vulnerable to liquefaction and flow slides. In 1966 a slip occurred in a 67 m high colliery waste tip at Aberfan in Wales. The large magnitude of the slip caused a flow slide to develop owing to the presence of a substantial volume of loose saturated material at the base of the tip ( Bishop etal , 1969). The flow slide travelled 500 m, a school was engulfed and 144 lives were lost.

Engineered fine fill



However, there is evidence that some clays with IP < 27% may have residual strengths considerably below pe ak strength (Rowe, 1991). Plastic clay fills may have a residual strength which is significantly smaller than φ'cv. Perry (1989) has surveyed the condition of motorway embankment slopes in England and Wales. From an analysis of all the recorded cases of instability, maximum slope angles are quoted for different geologies and fill heights based on restricting failures to below 1% within 25 years of construction. This can give assistance in the preliminary design of slopes for engineered fills. Where building will take place adjacent to a slope, a factor of safety greater than that typically required for road embankments will be required. Rapid construction of buildings on saturated clay fill could cause excess pore pressures to develop. These will subsequently dissipate during primary consolidation. Consequently, the undrained as well as the drained strength of the clay fill may need to be considered. During rapid construction the undrained strength may have a controlling influence on soil behaviour and bearing capacity, so with clay fills cu may be of importance as well as the effective stress shear strength parameters. However, as mentioned earlier in this

Table 4 Typical values of φ'cv for fills Fi ll ty pe Typic al ra nge foφr'cv Clayfill

20°— 30°

Sand fill

32°— 37°

Rockfill

35°— 42°

32 Hydraulic fill

These fills are deposited at a high moisture content; if granular they may be in a loose condition, and if cohesive they may be in a soft condition. Where the ground-water level remains at the surface of the fill, effective stresses remain low and strengths are small. Bearing capacity can be small, but this could be hidden by a desiccated crust on the deposit. Much depends on the drainage conditions; where there is effective under-drainage the situation may be greatly improved. Profiles of cu (as measured by the vane) with depth are shown in Figure 7. Some hydraulic fills may be vulnerable to liquefaction and flow slides.

3.6 Dynamic properties Knowledge of dynamic properties such as deformation moduli, damping and strength can be useful in several ways:  they are directly applicable to situations involving dynamic loading (for e xample earthquakes, traffic and machine vibrations)  they can be used to characterise heterogeneous fills  in some cases, it may be possible to relate field deformations under static loads to certain dynamic properties. The wave velocity in a fill is dependent on the mechanical properties of that fill, and in particular its stiffness; the stiffer the material the faster the wave propagation. However, dynamic moduli, such as theshear dynamic compression modulus and the dynamic modulus, do not necessarily have the same magnitude as the corresponding static moduli. The dynamic moduli are low-strain moduli. The dynamic shear modulus and the damping ratio are of particular importance in de scribing the behaviour of fills. Dynamic moduli

The dynamic moduli of fills can be calculated from the insitu measurement of wave velocities, and are readily determined as a function of depth . Measurements may be made in boreholes or at the ground surface. Dynamic moduli are measured at low strain, but it may be possible to relate properties controlling the behaviour of fills under typical monotonic or static loading conditions to dynamic parameters derived from the insitu measurement of wave velocities. This type of measurement has the advantage that the fill is not disturbed by the measurements. Several different types of wave velocity can be measured. Body waves are propagated in the interior of a body and are of two types:  compression or longitudinal (P) waves, in which the direction of particle motion is the same as the direction of wave propagation (the velocity of propagation is denoted by VP)  shear or transverse (S) waves in which the direction of particle motion is normal to the direction of wave

Chapter 3 Properties of fills propagation (the velocity of propagation is denoted by VS). Surface waves travel along or near the surface and two types should be noted:  Rayleigh waves travel along the ground surface, with a particle motion which is elliptical and retrograde in the vertical plane containing the direction of propagation (the velocity of propagation is denoted by VR)  Love waves travel along the ground surface, with a particle motion which is transverse to the dire ction of propagation with no vertical motion. Waves may be continuous or pulsed:  Continuous waves can be generated by a vibrator on a plate standing on the ground surface. A shear wave and a compression wave will be radiated into the ground. In addition there will be a surface or Rayleigh wave travelling in a direction parallel to the surface and in a layer about one wavelength deep. It is relatively easy to measure VR with geophones on the surface of the ground.  Pulsed waves can be produced by striking a plate resting on the ground surface and, as with continuous waves, three types of wave are produced simultaneously. To measure the velocity of the pulsed shear wave it is easiest to strike a plate sideways with a hammer; the plate must be attached firmly to the ground surface. Two dynamic moduli, the constrained modulus and the shear modulus, can be derived directly from a wave velocity measurement using an elastic solution to the wave equations:

Ddyn = ρVP2 Gdyn = ρVS2 where ρ is the mass density in Mg/m 3, wave velocities VP and VS are in m/s, and moduli Ddyn and Gdyn are in kPa. Sometimes Gdyn is termed Gmax or Go. Unlike the constrained modulus, the shear modulus is not a function of drainage conditions and the most useful dynamic measure of ground deformability seems to come from VS. This is of similar magnitude to VR and can be derived from V which is more easily measured. R Typically VS = 1.05 VR. Figure 14 shows BRE measurements of VR at a number of fill sites . Most fill types showed an increase in VR with depth. Values of VR in refuse shown in Figure 14 were generally smaller than 100 m/s and there was little increase over the 4 m de pth. Sharma et al (1989) quoted VS = 200 m/s for a refuse landfill in California but this was over a much greater depth. Cuellar et al (1998) measured VS = 160 m/s in a near-surface layer of refuse at a waste disposal site near Madrid; this corresponded to a shear modulus of 15 MPa, ten times greater than that measured by a 0.45 m diameter plate loading test.

3.6 Dynamic properties

33

Figure 14 Rayleigh wave velocity measured in fills

The dynamic shear modulu s can be expressed as a function of effective stress:

Gdyn = k1(σ'm)n Where k 1 is a function of voids ratio and overconsolidation and σ'm is the mean effective stress. In field situations the vertical effective stress is a more useful parameter and it has been suggested in many sandthe deposits a relationship of the followingthat form, in which exponent n = 0.5, is a reasonable approximation:

Gdyn = k2(σ'v)0.5 Dobry et al (1982) summarised field and laboratory data for sands from which a correlation between k 2 and ID can be deduced. It can be concluded, very approximately, that with ID = 0.5, k 2 = 8000 and that with ID = 0.8, k2 = 25 000 (where Gdyn and σ'v are measured in kPa). According to this approximate relationship:

VS = (k 2/ρ)0.5(σ'v)0.25 The relationship between Gdyn and σ'v is similar in form to the relationship betwee n static constrained modulus, D, and σ' for many well graded coarse fills. v ratio can be determined from the following Poisson’s expression: ν = (0.5 VP2 — VS2)/(VP2 — VS2)

Sharma et al (1989) have reported ν = 0.46 for some refuse landfill in California. Damping

The resistance which reduces vibration by energy absorption is termed damping. It gives a measure of the extent to which the fill differs from an ideal elastic solid.

The critical damping is the least amount of damping which will prevent free oscillatory vibration in a one degree of freedom system. Damping ratio ( Dr) is the ratio of the actual damping to the critical damping and can be evaluated by in-situ measurements. It can be defined as:

Dr = (1/4π)(∆E/E) where ∆E is the energy maximum energy stored.lost per cycle and E is the Damping ratio can be calculated from the ratio of two successive amplitudes, A n/An+1 in the free motion of a damped system: 2πDr /(1 — Dr2)0.5 = ln(A n/An+1) Where Dr is small:

Dr = (1/2π) ln(A n/An+1) Damping is sometimes expressed as a quality factor, Q, which is the ratio of the elastic to the non -elastic part of the modulus:

Q = 1/(2Dr) Damping ratio and quality factor are, like the shear modulus, strain dependent. Whereas the shear modulus decreases with increasing shear strain, the damping ratio increases with increasing shear strain. Liquefaction

The phenomenon in which effective stress in sandy soils is reduced by an increase in pore water pressure and the soil loses its shear strength is called liquefaction. Causes of the rise in pore pressure include fluctuations in groundwater level and wave action as well as the repeated actions of shear stresses during earthquakes. Historically,

34 the term ‘liquefaction’ has been applied to a variety of phenomena involving soil deformations caused by monotonic, transient or repeated disturbance of saturated cohesionless soils under undrained conditions. Cyclic loading of saturated sands under conditions of no drainage results in a marked reduction of strength associated with an increase in pore pressure. If the effective stress drops to zero, the sand is said to have liquefied. The term ‘liquefaction’ is sometimes restricted to this complete loss of shear strength. However, it is also used to denote a partial loss of strength due to build up of pore water pressure. Two phenomena can be

Chapter 3 Properties of fills q = kiA where q is the volume of water flowing through an area A under the influence of hydraulic gradient i in the direction of flow. Permeability is controlled by pore size within the fill which in turn is a function of both particle size and porosity. For fairly uniform sands Hazen (1892) suggested that:

k = C(D10)2

distinguished: cyclic mobility and flow liquefaction. Cyclic mobility occurs when the static shear stress is less than the shear strength of the liquefied soil and the deformations develop incrementally during earthquake shaking. Flow liquefaction can occur when the shear stress required for static equilibrium of a soil mass is greater than the shear strength of the soil in its liquefied state. Once triggered, the large deformations produced by flow liquefaction are driven by static shear stresses. The vulnerabilit y of a fill to liquefaction is a function of the particle size distribut ion and density index of the fill as well as the magnitude, frequency and duration of the dynamic forces to which it is subjected. There appears to be a threshold strain below which cyclic straining is non-destructive and there is no re arrangement of soil grains.

If k is measured in m/s and D10 in mm, then the empirical constant C is approximately 0.01. Hazen’s tests were limited to sands of grain size betwee n 0.1 and 3.0 mm, with a uniformity coefficient ( CU) less than 5. In practice, considerable variation in C is found. When D10 is much larger than 1 mm, flow is likely to become non-laminar and Darcy’s law is no longer applicable. In permeability testing, laminar conditions can be confirmed by obtaining a linear relationship between discharge velocity ν (= q/A) and i. Coarse and fine fills have very different permeabilities, as is illustrated by the typical values presented in Table 5. The range of values of permeability exhibited by fills is much greater than the range of values for most other properties. Permeability depends not only on the properties of the porous soil, but also on the fluid. Usually, only the flow of

3.7 Permeability

watertoiswater, of interest andcontaminated the quoted values of permeability refer but in land the flow of liquid contaminants may be important. A poorly compacted fill composed of lumps of clay with large interconnecting air voids between these lumps will initially have a high permeability. However, when water flows through the fill, the lumps will soften, collapse compression will occur and pe rmeability will greatly reduce. The permeability of cohesive hydraulic fills is low and reduces as the effective stress (and therefore the density) increases. Hydraulic deposition of fine waste materials is likely to lead to the formation of a horizontally laminated structure and horizontal permeability may be larger than vertical permeability. Consequently, small oedometer tests with vertical drainage could significantly overestimate the time, t, required for excess pore pressures to dissipate and primary consolidation to be completed. However, t varies as the square of the length of the drainage path and a high horizontal permeability will

A number of important effects and processes are influenced by water movement and pe rmeability, including the following:  rate of dissipation of excess pore pressures and associated primary consolidation of saturated clay fills  collapse compression of loose unsaturated fills  liquefaction of saturated fills  loss of ground due to erosion of fills. The significance of the first three phenomena in causing settlement is described in Chapters 4 and 5. Uncontrolled seepage can cause the migration of fine particles out of the fill or into the pore s of a coarser soil. Where water flows out of a fill, local instability may occur at the exit point. Where water is flowing through a fill which is susceptible to erosion, erosion can be prevented or controlled by protection of the fill with adequate filters which will halt the loss of material. Hutchinson (1981) described the damage to slopes produced by see page erosion in sands. Filter design can be related to particle size (Sherard et al , 1984a, 1984b; Sherard and Dunnigan, 1985) or permeability (Vaughan and Soares, 1982). The flow of water through fills is controlled by the permeability of the fill and the drainage conditions. Darcy’s law governs the laminar flow of water through fills. The coefficient of permeability or hydraulic conductivity, k, is defined by:

Table 5 Typical ranges of permeability for fills Fill type k (m/s) Clayfill(compacted) Sand fill

10 10

–8

– 10–10

–2

– 10–4

–2

– 10–5

Rockfill(wellgraded)

10

Colliery spoil (coarse)

10 –3 – 10–8

Pulverisedfuelash

10

–6

– 10–8

Refuselandfill

10

–3

– 10–6

3.7 Permeability have little effect where there is a long horizontal drainage path. Permeability may be measured in the laboratory or in situ. In the laboratory it can be measured in a permeameter, in a hydraulic oedometer or a triaxial cell. The constant-head method for granular soils is described in BS 1377-5:1990. It may be difficult to obtain a representative sample of the fill for laboratory testing. In the field, local permeability can be measured using hydraulic or standpipe piezometers. Sometimes permeability is estimated from infiltration tests carried out in pits excavated in the fill. Landva and Clark (19 89) have carried out such tests in refuse landfill. The interpretation of such tests is difficult, but in many practical situations only the order of magnitude of the permeability is of concern. Pumping tests can be used to measure the permeability over a larger volume of fill.

35

36

Chapter 4 Volume changes in fills In the last five years a very large number of examples of building damage due to foundation movement has been encountered in which the effect of the weight of the building on the ground is negligible. (Cooling and Ward, 1948)

Many of the examples referred to by Cooling and Ward (1948) were associated with the shrinkage of natural clay soils. Nevertheless, a similar observation would be equally valid for examples of damage to buildings on non-engineered fills. The major hazard to buildings on fill is typically associated with long-term movement within the fill due to volume change rather than inadequate bearing capacity. For example, in an investigation of 60 damaged houses built on fill in Hungary, Rethati (1961) found that the cause of damage was more frequently water seepage than deadload. Volume changes within the fill may be attributable to a variety of physical, chemical and biological conditions

fine fills and refuse fills. Coarse granular fills are relatively free draining. Excess pore pressures are unlikely under normal loading conditions. Much of the settlement due to self-weight is immediate and occurs as the fill is placed or re moved. Some time-dependent creep can be expected. Saturated clay fills come about in two distinct ways: first, clay excavated and placed by earthmoving machinery with compaction adequate to reduce air voids to a minimum, and, secondly, clay dredged and placed hydraulically. Saturated clay fills are relatively impermeable. Time-dependent movement will initially be associated with primary consolidation as e xcess pore pressures induced during loading slowly dissipate. When

and processes, most of which to be the weight of the structure built onare thenot fill.related They can classified by the cause of the change in volume as follows:  change in effective stress caused by weight of overlying fill  change in effective stress caused by weight of structures built on the fill  change in effective stress caused by rise or fall in ground-water level  change in moisture content  decomposition of biodegradable fill  chemical reactions  dynamic loading.

excess porecompression pressures have dissipated, some secondary willfully continue. Refuse fills with a high organic content will experience a large reduction in volume as a result of biodegradation. Some fills may not readily fit into these broad categories, for example:  silty fills may have properties, particularly permeability, which are intermediate in character between coarse fill and clay fill  fills composed of poorly compacted lumps or clods of clay may have some characteristics of coarse fills until saturated  chalk fill presents particular problems; a classification has been proposed by Ingoldby (1978), and a case history where problems developed has been described by Clayton (1980).

These volume changes may cause various types of movement which could adversely affect foundations:  settlement, caused by a reduction in volume of the fill, is the most common problem  heave, caused by expansion of the fill, could be associated with chemical reactions within the fill or the swelling of plastic clay  horizontal movement could result from volume changes; this is likely to be much smaller than the vertical movement where the fill is laterally confined; where fill is not laterally confined, horizontal movement could be important. In describing the various types of volume change, it is helpful to distinguish between types of fill with different modes of behaviour; for example, coarse fills, saturated

Where volume change is brought about by a change in effective stress, simple elastic relationships are used to describe these movements. This approach is adopted for simplicity, not because fills generally e xhibit isotropic linear elastic behaviour. In many cases the fill immediately under the foundation will be partially saturated and the actual behaviour may be complex (Toll, 1990; Skinner et al, 1998). In reviewing the causes and magnitude of settlement within a fill, it should be remembered that it is differential rather than total settlement that damages structures. Differential movement is examined in Chapter 10. Other types of movement could be caused by bearing

4.1 Self-weight of fill capacity failure or slope instability. Damaging ground movement may also occur due to deformation of the natural ground below the fill. Mining subsidence can cause compression within the fill as well as settlement of the base of the fill (case history 10).

4.1 Self-weight of fill Stress changes within the fill will cause movements, the magnitude of which will depend on the compressibility and stiffness of the fill. Important causes of a change in stress are the placement of further layers of fill and the removal of some overlying fill. The changes in stress which result from the placement of fill over a wide area differ from the stress changes due to buildings, which usually apply loads to relatively small areas. The wide extent of fill placement has the following consequences:  loading and compression behaviour of the fill is largely one-dimensional  the full depth of existing fill may be affected by the extra fill  the underlying natural strata may be significantly loaded by the fill. Similar conclusions hold for removal of layers of fill although the stiffness properties may be very different on unloading the fill, with the fill exhibiting a much greater stiffness (section 3.4). Movement caused by compression or expansion of the fillimmediate occurs in two stages: as the applied stress increases or movement 

decreases  time-dependent movement after the change in applied stress. Movement during fill placement

The immediate compression, which occurs during the earthmoving operation before construction of buildings on the fill, has no practical effect on the structures. As only long-term movement is of significance for building on the fill, the immediate movements during fill placement are only briefly reviewed. Coarse fill As extra fill is placed, most of the resulting compression of the underlying coarse fill occurs immediately. T he onedimensional compression behaviour of coarse granular fills has been described in section 3.4 in terms of the constrained modulus, D, which depends on fill type, placement conditions including initial density and moisture content, stress level and stress history. Saturated clay fill The behaviour of saturated or almost saturated clay fill when extra fill is placed, differs considerably from that of coarse fill. The load initially induces excess pore pressures in the clay fill, and there is little immediate settlement. However, there may be some immediate settlement due to compression of any remaining air voids

37

(but this is likely to be very small), and some primary consolidation which occurs as excess pore pressures dissipate during the period of fill placement. Refuse fill The immediate compression behaviour is similar to that of granular soils but the constrained modulus is small (Table 3). Case histories 19 and 20 provide information about old domestic refuse and case histories 21 and 22 about recent domestic refuse. Long-term movement

When movements due to the self-weight of the fill occur subsequent to building on the fill, damage to structures may result. The magnitude and rate of movement subsequent to fill placement are therefore of major interest. The nature of time-dependent movements has been discussed in section 3.4, and in this section examples of the magnitude of such movements are given. Coarse fill For most coarse fills, there is an approximately linear relationship between creep compression and the logarithm of time that has elapsed since the load was applied. Sowers et al (1965) reported that when long -term settlement of many rockfill dams in the U SA was plotted against the logarithm of time that had elapsed since the middle of the construction period, approximately linear relationships were obtained with values of α ranging from

0.2% to 1.0%. of settlement as to a percentage ofThe damamount height did not appearexpressed to be related rock type, form of construction (for example upstream membrane, central core) or height. The significant factor was the method of placement of the rockfill. The dam material with the greatest settlement had bee n dumped with limited sluicing, the dam material with least settlement had been compacted by rolling while sluiced. Kilkenny (1968) quoted α = 0.74% for an opencast mining backfill at Chibburn in Northumberland. Measurements by BRE on an uncompacted mudstone, siltstone and sandstone opencast backfill (case history 8) indicated that α = 1%. Data presented by Lange (1986) suggested that for backfills in the Rhenish brown coal area, typically α = 0.5% to 1.0% for backfills less than 100 m high, but greater v alues for deeper fills. The settlement of the crest of a 73 m high embankment of heavily compacted sandstone and mudstone rockfill has been monitored for a period of almost 20 years since the end of construction (Charles, 1990). The settlement, s, is shown plotted against log time in Figure 15. There is an approximately linear relationship between s and log t corresponding to α = 0.17%. In Figure 16 the creep rates of three very different types of fill are compared: (a) heavily compac ted sandsto ne/mudstone rockfil l (b) opencast sandst one/mudstone fill witho ut systematic compaction (case history 8) (c) recent domestic refuse (case history 22)

38

Chapter 4 Volume changes in fills

Figure 16 Long-term settlement of fills (a) heavily compacted sandstone/mudstone rockfill (b) opencast sandstone/mudstone backfill without system atic compaction (c) recent domestic refuse

Figure 15 Long-term settlement of 73 m high heavily compacted sandstone and mudstone rockfill embankment

All three fills show a reasonably linear increase of compression with log time, although there is an indication for the refuse of some steepening of the gradient with time. α differs by orders of magnitude between the different fill types. BRE has monitored settlement within the downstream shoulders of rockfill dams, and creep deformations have been examined at different depths in regions of the embankments not seriously affected by reservoir

compression, have been described in se ction 3.4. Secondary compression continues under conditions of constant effective stress after all excess pore pressures have dissipated. In a similar manner to creep compression of granular fills, there is a linear relationship between settlement and the logarithm of the time that has elapsed since the load was applied. The continuing

fluctuations. Compression linearly withasthe logarithm of time since the increased end of construction, expected. The logarithmic creep rate parameter α is stress dependent for these heavily compacted fills as shown in Figure 17 where, for two rockfill embankments, α is plotted against the mean vertical effective stress, σ'v. It is estimated that ID = 0.8 for these heavily compacted rockfills. BRE measurements are summarised in Table 6. More case histories are needed to confirm the gene ral applicability of this type of relationship.

rate of compression clay fillsby in one -dimensional compression can be of described the coefficient of secondary compression, Cα (section 3.4), which typically for clay fill might be in the range 0.001 to 0.005 (0.1% to 0.5%). Refuse fill Long-term movements are likely to be very large, and can be attributed to two different processes:  physical creep compression occurs as the particles become more closely packed, as in a granular fill  biodegradation of the organic content of the refuse causes a large reduction in volume of the material.

Saturated clay fill The two types of time-dependent movements for saturated clay fills, primary consolidation and secondary

Table 6 Creep compression rate param eter α for engineered and non-engineered fills F i ltl y p e Co m pa c t i o n α ( %) * Ca s e h i s t o r y † Sandygravelfill

Heavyvibratingroller

0.04

σ' v

Megget

Mudstonefill

Heavyvibratingroller

0.12

σ' v

Brianne

Sandstone/mudstone rockfill

Heavy vibrating roller

0.13 σ'v

Scammonden

Sandstone/mudstone rockfill

No systematic compaction

0.9–1.5

8

Stiffclayfill

Heavydynamiccompaction

0.5

2

* σ'v in MPa † The first three examples refer to case histories of embankment dam construction: Megget — Penman and Charles (1985a) Brianne — Penman and Charles (1973) Scammonden — Penman et al (1971)

4.2 Weight of buildings

39

significantly stressed only to a depth of 2 m to 3 m. This confirms that with a deep fill, settlement due to the weight of low-rise building is unlikely to be the critical issue. Apart from immediate settlement of clay fills under undrained conditions, most movement due to the weight of structures is attributable to a volume reduction resulting from an increase in effective stress. Although the stress and strain conditions under a small foundation are not strictly one-dimensional, movements can usually be related to one- dimensional parameters suc h as the constrained modulus. Movement occurs in two stages:  

as construction proceeds after completion of construction.

The distinction has important practical consequences. Although structur al damage can result from both the immediate and long-term ground movements, settlement occurring during construction can be built out and may have little harmful effect, whereas settlement occurring subsequent to construction can seriously damage a structure. Damage to finishes (for e xample partitions or plasterwork) is caused mainly by long-te rm movements. Movement during construction

The settlement caused by the weight of a structure will be a function of bearing pressure, foundation size and depth and stiffness of the fill and foundation. Immediate settlement of a granular fill or consolidation settlement of Figure 17 Relationship between logarithmic creep compression rate parameter, α, and stress level, σ'v, for heavily compacted rockfill

The latter process is usually the major cause of long-term settlement, and is examined later in this chapter (section 4.5).

4.2 Weight of b uildings Whereas fill placement is usually over a wide area, the weight of buildings will generally be concentrated on relatively small foundation areas. Where the foundation size is small, stresses will be significantly increased over only a relatively shallow depth of fill immediately under the foundation; if the fill is deep, the underlying natural strata may not be affected by the structural loads. In any evaluation of ground deformations caused by the weight of a building, it is important to have some knowledge of the stress distribution below the foundation, because in predicting settlement it is necessary to know to what de pth the foundation will effectively stress the fill. The stress distributio n below foundations is discuss ed in Appendix A . It is shown that, for typical low-rise foundations, 1 m below foundation level there is little difference in vertical stress despite large differences in footing widths. Furthermore, for a foundation load of 50 kN/m run, the ground is

a clay fill due the theory weightusing of theastructure maymodulus. be estimated by to elastic constrained Settlement, s, can be related to applied pressure, q, foundation length, a, foundation width or diameter, b, depth of foundation, d, shape and rigidity of foundation, constrained modulus of the fill, D, and Poisson’s ratio of the fill, ν, as follows:

s = fs fν fd (qb/D) where fs is a shape and rigidity factor, fν is the Poisson’s ratio factor and fd is a foundation depth factor. Details of the values of these factors and this method of calculating settlement using elastic theory are given in Appendix B. The type of movement that occurs as load is applied depends not only on the stiffness of the fill but also on the permeability of the fill, and consequently this type of movement is examined under the headings of coarse fill, saturated fine fill and refuse fill. B RE has carried out load tests on a number of different types of fill and the immediate settlements measured during the tests are summarised in Table 7. The immediate settlement, si, was measured 0.1 day (2.4 hours) after the load was applied. Coarse fill The shear strength of coarse granular fill is generally high, particularly when the fill has been heavily compacted. As a consequence, such fills have a high bearing capacity. Applied loads will generally be much smaller than the failure load and, consequently, bearing capacity is

40

Chapter 4 Volume changes in fills

Table 7 Immediate settlement measured in BRE load tests on non-engineered fills Fill type q (kPa) b (m) a /b si (mm) si /b (%) qb/si ( MP a ) Building wastes

24

1.8

1

Colliery spoil

35

2.0

1

Old urban fill Clay fill Alluvial sand*

30

1.8

50 50

1

2.0 0.75

5.0 23 13

1 3

0.3

9

26

1.1

3

10

0.7 3.2

Cas e hi s t or y

4

0.2

1.5

0.2

14

31 25

25 27

* The load test on alluvial sand has been included although this was not a fill; a sand fill would be likely to perform in a similar way

unlikely to be a problem when structures are founded on these fills. Allowable bearing pressures should be specified that will provide an adequate factor of safety against soil rupture and keep total and differential settlement within tolerable limits. In granular soils the allowable settlement is usually exceeded before soil rupture considerations become significant (Sutherland, 1974). Foundation movements may be largely due to reduction in volume of the fill and can be related to the constrained modulus D as described in Appendix B. Foundation design on coarse -grained soils is principally concerned with the prediction of settlement. Much of the settlement is immediate and occurs as the structure is built. Although this immediate settlement may affect the structure, it will not be so significant as long-term settlement that occurs after the structure is complete. The difficulty of obtaining undisturbed samples of the soil means that the required properties cannot usuallyisbe determined from laboratory tests and settlement calculated from the results of in-situ tests. However, Leonards and Frost (1988) suggested that penetration resistance tests are inherently incapable of sensing the effects of over-consolidation and prestressing on the compressibilit y of granular soils. It has been generally found that settlement is linearly related to applied stress. This would seem to be true over a wide range of working stresses. There must, however, be some limiting stress where the linearity ceases as bearing failure is approached. From an analysis of case histories, Burland and Burbidge (1985) have proposed an empirical relationship for the calculation of the settlement of normally consolidated sand (in the situation where there is a large factor of safety against bearing capacity failure) using the following equation: ∆s/∆q = 1.71b0.7/N 1.4

where ∆s is the settlement in mm produced by an increase in bearing pressure ∆q kPa over a foundation of width b m. N is the average standard penetration test (SPT) count over the depth of influence of the foundation (Appendix A). N is not corrected for effective overburden pressure, but corrections are made for fine and silty sand below the water table and for gravels. A method of relating settlements to cone penetration test (CPT) results is also given by Burland and Burbidge (1985). In Figure 18, ∆s/∆q is plotted against b for three values

Figure 18 Settlement of foundation of width b on sand (derived from Burland and Burbidge, 1985)

of N typically representing a loose, medium and de nse condition. The expected settlement is quite small except for loose sand with a high be aring pressure or wide foundation. The relationship empirically derived by Burland and Burbidge (1985) indicates that settlement is proportional to b0.7 whereas elastic theory predicts that settlement and b are linearly related. Holzlohner (1984) suggested that the increase of the ratio of settlement to applied stress, ∆s/∆q, with footing width, b, is best explained by a modulus which increases with depth so that it is proportional to the square root of depth. This is compatible with the case where the one-dimensional compression properties of the granular fill are represented by a relationship of the form σv = aεv2. Saturated fine fill Not only do saturated clay fills have a smaller angle of shearing resistance than coarse fills, but also, if construction is sufficiently rapid, excess pore pressures do not have time to dissipate during the construction period. Consequently, the bearing capacity may be controlled by the undrained shear strength with immediate settlement also largely a function of undrained behaviour. Bearing capacity may be low and undrained movements may be large. These undrained movements will be followed by primary consolidation and then secondary compression. Although most of the immediate settlement of a soft clay fill under an applied load is likely to be due to

4.2 Weight of buildings undrained deformation which is controlled by the shear modulus G, there could also be some immediate movement due to compression of small air voids, particularly as it is unlikely that a clay fill will be fully saturated close to ground leve l. The shear deformation may be estimated from calculations based on simple elastic theory and undrained elastic parameters such as the shear modulus G. The undrained settlement, sun, for a footing at ground surface can be calculated from:

sun = 0.25fs(qb)/G

41

foundation loads remain sensibly constant, but conditions in the foundation fills change, and further movements result; possible causes include change in moisture content, erosion, mining subsidence and seismic event  foundation loads and conditions within the foundation fills remain sensibly constant, but settlement still occurs. 

Long-term settlement attributable to foundations loading is of a similar type to that which occurs during the initial construction of the building. Settlement caused by

Values of fs for some common shapes of footing are given in Appendix B. With a soft clay fill, deformation could be large as the factor of safety against undrained failure may be relatively low. A crude estimate of G can be obtained from an undrained triaxial test (BS 1377-7:1990). A secant value of the modulus at an appropriate axial strain, ε, corresponding to a mobilised shear stress of, say, cu/3 could be calculated from G = (2/9)( cu/ε). A typical value of axial strain corresponding to cu/3 might be 1%, in which case G = 22cu, and the immediate undrained settlement at the centre of a 1 m diameter circular flexible footing loaded with q = 100 kPa and with cu = 50 kPa is 23 mm. In practice, a clay fill is unlikely to be saturated at the surface unless there is a high ground-water level. Otherwise desiccation is likely to have formed a stiff,

changes in depth to water table or inundation from ground surface are addressed in later sections of this chapter (sections 4.3 and 4.4) and in Chapter 5. The effect of seismic events is also examined later in this chapter (section 4.7). Case history 10 gives an example of the effect of deep mining activity on fill settlement. In many instances the loads applied to the foundations will not change significantly after the completion of construction. However, some time-dependent strains will occur under constant applied load:  with coarse fills, settlement will be caused by creep movements occurring at constant effective stress  with fine fills, settlement will be caused by primary consolidation as pore pressures induced by construction loading slowly dissipate, followed by secondary compression at constant effective stress.

partially crust. Immediate settlement may thenwill be causedsaturated by compression of air voids in the crust and be analogous to immediate settlement of a granular fill.

Coarse fill Creep compression will occur as the particles of fill become more closely packed. Settlement due to imposed foundation loads often shows a linear relationship with the logarithm of time since construction, analogous to creep settlement arising from self-weight of the fill. In general, long-term settlement can be described by a relationship of the form:

Refuse fill The type of behaviour is generally similar to a granular fill, but movements may be large. Harris (1979) reported some 0.3 m diameter plate loading tests on 16-year-old domestic refuse with a dry density of only 0.24 Mg/m 3 and moisture content of 28%. The measured constrained modulus of 10 MPa was surprisingly large. Harris commented that the refuse had been grassed over and used as a recreation ground and, as a result, the surface layer may have been compacted although the measured fill density did not indicate this. Landva and Clark (1989) have reported numerous load tests on refuse landfills in Canada in which a pressure of 150 kPa was applied through a 1.13 m diameter plate. Values of constrained modulus varied between 1 M Pa and 17 MPa, with a typical result of 3 MPa. Knochenmus et al (1998) reviewed a large amount of data and concluded that the modulus for municipal solid waste ranged from 0.5 MPa to 3 MPa. The results of BRE investigations, which are presented later in this chapter in the section on biodegradation of refuse fills (section 4.5), generally confirm this. Movement subsequent to construction

The following situati ons have to be considered:  foundation loads increase and cause further settlement

s = si + sα log( t/ti) where s is the total settlement (immediate plus longterm) at time t after completion of loading, and si is the immediate settlement occurring during loading and measured a short but finite time, ti, after the completion of loading. There is a difficulty with this type of logarithmic relationship in deciding the time ti. In a load test where the load is applied rapidly and the time by which loading has been completed can be accurately defined, ti = 0.1 days is appropriate. For a slowly constructed building, the time at which loading is completed is less well defined and ti = 10 days may be appropriate. The parameter sα can be defined as:

sα = (s — si)/log( t/ti) It can also be related to the settlement rate by the following relationship:

sα = 2.303 t(d s/dt )

42

Chapter 4 Volume changes in fills

Table 8 Long-term settlement measured in BRE load tests on non-engineered fills Fill type q (kPa) b (m) a /b si (mm) sα (mm) sα /b (%) mα

Case history

Building wastes

24

1.8

1

5.0

2.0

0.1

0.4

26

Colliery spoil

35

2.0

1

23

6.0

0.3

0.3

10

Old urban fill Clay fill

30

1.8

50

Alluvial sand†

50

2.0 0.75

1

13 1 3

1.5

3.2 1.5

2.8 0.3

0.1 0.1 0.04

0.1 0.9* 0.2

14 25 27

* A particularly high value of mα was found for the clay fill. The long-term settlement of the clay fill could be partly due to dissipation of excess pore pressure † The load test on alluvial sand has been included although this was not a fill

where d s/dt is the settlement rate at time t after the completion of construction. The ratio sα/b is analogous to α the self-weight logarithmic creep rate parameter, α. Some values of sα/b measured in BRE load tests on non-engineered fills are quoted in Table 8. Ortigosa et al (1981) derived values of a creep coefficient mα from plate loading tests on gravel:

m α = sα / si mα relates logarithmic creep rate, sα (settlement during a log cycle of time), to immediate elastic settlement measured after 2 hours, si. For the Santiago gravel it was found to be almost linearly related to the stress level such that: mα = 0.027σv where σv is the bearing pressure in MPa. BRE load tests on non- engineered fills have shown values of mα orders of magnitude larger than those suggested by the above relationship, and these are summarised in Table 8. It should be noted that if the BRE results were applied to buildings where si was taken at ten days rather than the 0.1 days in the load tests, mα would be significantly reduced; for the granular materials it would range from 0.1 to 0.2 and even for the clay fill would be only 0.3. These are still much greater than the values quoted by Ortigosa et al (1981) for a natural gravel deposit. Field measurements of creep compression caused by self-weight suggest that α is proportional to σv in well compacted fills, but less dependent on σ in loose fills. This finding may also be applicable to the v ratio sα/b. Saturated fine fill As the excess pore pressures induced by construction loads dissipate, primary consolidation will occur. Some primary consolidation will occur during the construction period. The magnitude and rate of consolidation can be analysed using the drained constrained modulus, D, and the coefficient of consolidation, cv, respectively. The use of the constrained modulus in calculating the settlement under a footing is described in Appendix B. After the dissipation of excess pore pressures induced

by construction of the building, secondary compression of clay fills will continue in the long te rm. This type of settlement can be of particular importance because it occurs subsequent to the completion of construction. The coefficient of secondary compression Cα has been described in section 3.4. In practice, a clay fill is unlikely to be fully saturated near ground level, and its behaviour will not necessarily closely correspond to the simple consolidation model of a saturated clay outlined above and described in section 3.4. Case history 25 describes load tests on a clay fill; initial settlement was small but the rate of long-te rm settlement was quite large. Refuse fill Although physical creep compression caused by self-

weight of the fill and the of buildings long-term settlement willwe beight controlled by will occur, biodegradation. This is examined later in this chapter (section 4.5).

4.3 Change in ground-water level The effective stress within a fill deposit depends strongly on ground-water level. A rise in ground-water level reduces effective stress, and a fall in ground-water level increases effective stress. The smaller the particle de nsity of the fill, the greater the pe rcentage change in effective stress caused by a given change in ground-water level. Refuse fills (case histories 21 and 22) and pfa (case history 12) will therefore be particularly affected. The magnitude of the movements caused by these changes in effective stress will depend on the stiffness of the fill. The changes will occur over large areas and so deformations will be largely one- dimensional. The movements can be described by a constrained modulus appropriate to the particular stress change. As the change in water level will usually result in unloading or reloading of the fill, the stiffness may be much larger than that operative during first loading of the fill (section 3.4). Granular fill has a high permeability and the groundwater table will adjust quickly to any changes in external constraints. The water level may thus rise or fall rapidly. In case history 4, the water table rose 34 m through an opencast backfill in three years. Changes in water level will change the moisture

4.5 Decomposition of biodegradable fill

43

content of the fill. This too can cause ground movements and is examined in the next section and in Chapter 5. With saturated clay fills of low permeability, any changes in the ground-water regime within the fill will only occur slowly. A clay fill which initially has large air voids between lumps of clay, due to inadequate compaction, will have a high permeability when first inundated. However, on inundation the clay lumps will soften, and the fill will then become a low permeability, saturated clay fill. Changes in effective stress can cause large volume changes in clay fills.

Table 9 Clay shrinkage potential (BRE, 1993a) Plasticity index,IP (% ) Cl ay fr ac ti on (% ) Sh ri nk age pot en ti al

4.4 Change in moisture content

by rainfall and by upward migration from the water table  by transpiration of soil moisture from greater depths through the action of the roots of vegetation; the larger the vegetation, the greater the demand for moisture in dry weather  heating of the ground below, for e xample by furnaces.

Ground movements caused by moisture changes are encountered in many situations. The complex phenomena controlling ground movements in partially saturated soils have been reviewed by Gens et al (1991). Changes in moisture content within a fill can have various causes including:  rise of ground-water level  downward infiltration of rain-water and percolation of water from leaking drains, etc  drying as a result of e vaporation and transpiration. Movements associated with changes in moisture content of the fill could be attributed to the following causes.  Partially saturated fills may undergo a re duction in volume known as collapse compression when inundated. Most types compression of partially saturated are susceptible to collapse under a fillwide range of applied stress when inundated or wetted for the first time if they have been placed in a sufficiently loose and/or dry condition. This phenomenon can occur without any change in applied total stress. When construction is about to take place on such a fill, susceptibility to collapse compression may be the most significant hazard because, where collapse compression occurs after construction has taken place, buildings may be seriously damaged. This most important facet of fill behaviour is examined in some detail in Chapter 5.  A rise or fall of ground-water level within the fill will change the effective stresses within the fill as well as changing the moisture content of the fill. This type of movement was considered in the previous section.  The volume of saturated clay fill, particularly highplasticity clay, is moisture-content dependent. Drying out may cause considerable shrinkage. This phenomenon is considered in this section. Clay volume change can occur as a result of change either in the applied load or in its moisture content. As the water content of a clay fill reduces, the clay will shrink. As the fill dries and shrinks, the soil moisture exerts more suction. If water is then brought into contact with the fill, it is absorbed and the clay swells. These types of movement have caused damage to houses with shallow foundations on natural clay soils, and clay fills are

>35

>95

Very high

22–48

60–95

High

12–32

30–60

Medium

<18

<30

Low

vulnerable to the same processes. Shrinkage can be brought about in several ways:  by moisture movement near the ground surface, as moisture evaporates in dry weather and is reple nished

BRE re search has indicated that significant shrinka ge is confined to the top 1 m to 1.5 m unless there are large trees, which can create deep zones of dried and shrunken clay around them. The amount of shrinkage or swelling depends on clay content and mineralogy. Table 9 gives an indication of clay shrinkage potential. The strength of clay fill subgrades has been investigated by the Transport Research Laboratory: Black and Lister (1978) described the influence of the strength of clay fill subgrades on road performance.

4.5 Decomposition of bi odegradable fill The age of refuse landfill is important in evaluating the potential for further settlement. Not only has older refuse landfill had longer for the decomposition of organic matter to occur, it is often an inherently better material with a much higher ash content than more recent refuse (section 2.6 and Figure 4). Settlement behaviour is also affected by a number of other factors:  composition of the waste  initial bulk density  initial moisture content  level of leachate within the fill  time over which the fill was placed. Movements can be attributed to the following causes:  immediate physical compression occurs due to the weight of successive layers of fill causing crushing, distortion, bending and reorientation of the materials within the fill (can be de scribed by a constrained modulus D)  long-term movements occur as biodegradation causes a reduction in volume (can be described by a logarithmic compression rate parameter αb)  long-term movements occur, caused by physical creep compression as the particles of the refuse become more closely packed (can be described by a

44 logarithmic compression rate parameter αc). Generally, the most important cause of settlement is volume reduction resulting from the decomposition of organic matter, which may continue over a very long period. Although α-type logarithmic compression rate parameters are used to describe the long-term settlement in this book, there is evidence that a power law such as that described earlier, in section 3.4, might more precisely represent the observed movements caused by biodegradation and physical creep compression (Edil et al, 1989; case histories 21 and 22). There may be chemical processes resulting in oxidation and corrosion. Fires on landfill sites are now not common. A number of laboratory and field studies have bee n carried out in the U SA into the rate and magnitude of settlement of refuse landfill, usually described as ‘sanitary landfill’. Some of the more important findings are summarised below.  Merz and Stone (1 962) studied six 6 m dee p test cells. They concluded that the greatest settlement of a 6 m deep landfill would occur in the first month after completion and that after the third month settlement would be relatively small.  Sowers (1973) suggested that the initial settlement could be described by a compression index which was related to voids ratio and that continuing settlement was analogous to secondary compression in soils.

Chapter 4 Volume changes in fills concluded that the power law was the more appropriate.  Ling et al (1998) have examined the validity of the logarithmic and power relationships, which have commonly been used to e stimate the settlement of municipal solid waste, for settlement monitored at three landfill sites. They proposed a hyperbolic function as an improved method of simulating the settlement–time relationship. Relatively little information is available on the settlement of refuse landfill in the U K. Cheyney (1983) described the settlement of domestic waste up to 18 m deep measured between about 0.5 years and 2.5 years after completion of the fill. Data from up to 70 settlement points over a fill area of 7.5 ha were plotted and gave an approximately linear relationship with the logarithm of time since completion of the landfill. There was a great variation in the rate of settlement between different locations. Table 10 summarises BRE measurements of αb and αc at a number of sites (Watts and Charles, 1990, 1999). All the sites, with the exception of Heathfield, form case histories in Part IV. At Heathfield, BRE and NHBC collaborated to construct an experimental building on an 8 m deep refuse landfill to develop practical and costeffective solutions for the protection of houses threatened by landfill gas migration. The results shown in Table 10 confirm that the more recent refuse has properties which are inferior to older



Yen and Scanlon (1975) records taken over a nine-year pestudied riod at settlement three sites and concluded that the rate of settlement decreased linearly in proportion to the logarithm of the median age of the fill.  Chang and Hannon (1976) measured the settlement of landfill when loaded by a 3 m high highway trial embankment. The six-year-old refuse was 6 m deep and immediate settlement was of the order of 0.15 m.  Rao et al (1977) measured the settlement of small experimental cells and larger zones of compacted refuse under imposed loads as well as carrying out laboratory tests. The laboratory studies indicated that settlement characteristics were dependent on such factors as stress history, initial density, load increment ratio and magnitude of pressure. Field studies indicated the importance of environmental factors such as temperature and rainfall which affected decomposition.  Dodt et al (1987) monitored the settlement of the existing surface of a landfill when a further 11 m of landfill were placed. The settlement rate of the old landfill was of a similar form to that expected in the primary and secondary phases of consolidation of a cohesive soil. Settlement of 0.9 m was measured corresponding to a compression of 6% of the depth of landfill.  Edil et al (1989) used both a powe r creep law and the Gibson and Lo (1961) expression to represent longterm settlement of refuse (section 3.4). They

αb is refuse. It should bealmost noted that muchrefuse. greater than αc and that αb can be 20% for recent However, the two parameters have different time zeros. For αb time is reckoned from the placement of the re fuse. For αc zero time is reckoned from the application of the load. Case histories 16 and 20 illustrate the significance of this. These two case histories confirm that with refuse the relationship between settlement and log time is not as nearly linear as it is with many other types of fill. High leachate levels in a refuse landfill can greatly reduce effective overburden stresses in the fill. The compression of saturated fill results in upward movement of leachate within the refuse fill, and compression caused by a widespread load may be significantly smal ler than Table 10 Properties of refuse fills (after Watts and Charles, 1999) Max depth Bulk Cas e A ge o f o f r ef u s e D density αc αb h istory refu s e ( m) ( MP a) ( % ) ( % ) ( Mg/3)m 16

1960

6.0

6

0.2

2

17

1935

6.5

—

—

—1

18

1960

7.8

—

—

2

—

19

1960

5.8

2

—

2

—

20

1960

8.0

3

0.2–0.3

6

—

21

1983–71 1.0

0.7

—

18

0.6

22

1982–6 20.0

0.9

2.0

18

0.8

—

0.7

7

Heathfield1 989

8.0

1.8 .8

—

4.6 Chemical reactions expected because the increase in effective stress caused by the applied load is offset by the reduction in effective stress caused by the rise in leachate level.

4.6 Chemical re actions Volume changes, usually expansion, may occur due to chemical reactions within the fill. Loss of material due to solution of carbonates could occur in some fills. While this book does not deal with these chemical processes in detail, an outline is given of some of the more common hazards. In addition to volume change of the waste fill, there may be interaction between construction materials and chemically aggressive ground and this is examined in section 10.5. Chemical reactions may have been progressing slowly ever since the waste fill was deposited. However, building development on the site could lead to an acceleration of these reactions as a result of effects connected with construction, such as the following:  earthmoving may lead to mixing of wastes, which brings different substances into contact and may introduce oxygen into the waste fill  leaking drains may cause water to infiltrate the waste fill. Fills which should be treated with caution include the following:  sulfate-bearing waste materials

45

for decades may undergo expansion when newly disturbed and exposed to the atmosphere. Hydration of free lime and periclase is almost impossible to inhibit. There are reported incidents of the failure of floor slabs where steel slag has been used as hardcore. When used as bulk fill to support buildings, such old slags should be carefully tested for stability and, if used, should be placed carefully and monitored for any subsequent heave. A laboratory procedure for the identification of potentially expansive steel slag has been proposed by Emery (1979).  The most serious problems with older blastfurnace slags derive from their sulfur content (Thomas, 1983). Weathering can produce a reaction between the slag minerals, sulfate and water. The formation of ettringite or thaumasite can cause expansion. Free sulfate in contact with concrete can provoke a reaction with Portland cement, causing disintegration of the concrete. Further information on the problem of expansive slags is given in the Environment Agency Report P331 (Garvin et al, 1999). Pyritic shales

Many shales contain iron pyrites (FeS 2) which may be oxidised to sulfuric acid by microbiological activi ty (B RE, 1992; Collins, 1990). This is the srcin of most cases of high-sulfate soil which can cause a severe attack on

  

iron and steel slags pyritic shales alkali wastes

Sulfate-bearing waste materials

Building wastes, for example, may include plasterboard or similar materials generally based on gypsum. If the waste also contains broken concrete, expansion may occur when the concrete is exposed to sulfate solution which srcinated in the other components of the waste. Even if expansion in the fill material is not a problem, sulfate attack on concrete may occur. BRE Special Digest 1 describes sulfate a nd acid resistance of c oncrete in the ground ( BRE, 2001). Iron and steel slags

Two main types of slag are produced during the manufacture of iron and steel. Blastfurnace slag arises when iron ore is smelted to produce pig iron, and steel slag arises when pig iron is converted into steel. Modern blastfurnace slag is generally a good fill material, but is usually too valuable to use as bulk fill. In contrast, there are problems with steel slags and older blastfurnace slags. Large deposits of slags and other wastes are found around iron- and steel -making plants. Case history 13 describes the reclamation of an old slag bank.  Steel slag is potentially e xpansive. Volume changes of 10% or more can occur on exposure to moisture, caused by hydration of free calcium and magnesium oxides. Old banked slag that has re mained undisturbed

ordinary Portland cement there isthe a shale, sourcethe of acid-soluble calcium (suchconcrete. as calcite)Ifwithin formation of gypsum can cause heave. The process may be initiated by the access of air when the shale is used as a fill or in bedrock by trenches for foundations or services. Nixon (1978) described floor heave in buildings caused by pyritic shales used as fill material. Hawkins and Pinches (1987a, 1997) have drawn attention to how naturally occurring pyritic shale has damaged buildings founded on this rock, and have described floor he ave at Llandough Hospital, Cardiff due to gypsum growth. They (1987b) considered that the effects of the growth of selenite (crystal form of gypsum) in discontinuities in mudstones was not fully appreciated. Expansion can be as large as 10%. Pye and Miller (1990) listed reported instances of heave of buildings on pyritic shale or compacted shale fill in Canada, France, Norway, Sweden and the USA. Following the failure during construction of the shale embankment of Carsington Dam, a de tailed investigation of the chemical weathering of the mudstone fill was carried out (Pye and Miller, 1990). It was concluded that significant post-emplacement changes occurred in the fill associated with the following:  oxidation of pyrite  leaching of carbonates by sulfuric acid  formation and growth of gypsum and jarosite. The comment was made that engineering problems associated with volume loss due to carbonate dissolution,

46

Chapter 4 Volume changes in fills

or volume increase due to gypsum formation, are only likely to occur in mudrocks which contain both finely divided pyrite and calcite (or other re adily soluble carbonate or calcium). However, acid generation will still occur in pyritic rocks which do not contain carbonate. The mudstone rockfill at Roadford Dam contained significant quantiti es of pyrite and it was recognised that some of this might oxidise during and after embankment construction. Monitoring of the chemistry and mineralogy of the fill during construction and the subsequent monitoring of the chemistry of the see page water confirmed that pyrite oxidation was occurring but

Institution of Civil Enginee rs, 1998). Loose sands are particularly vulnerable to vibrations which may produce different effects as accelerations increase:  resilient (elastic) behaviour  compaction (reduction in volume)  liquefaction (loss of strength)

the quantities were very small and should not significantly affect the geotechnical properties of the fill (Macdonald and Reid, 1990; Davies and Reid, 1997). Unburnt colliery spoils have been used extensively in the UK as fill materials without causing problems due to heave, as far as is known. However, the possibility that some colliery spoils could suffer pyritic e xpansion should not be overlooked.

soils exist under most pavements, and pavement performance can be sensitive to soil suctions. Repeated load triaxial testing has given some understanding of non-linear resilient or recoverable elastic stiffness properties and resistance to permanent deformation. Gorle and Thijs (1989) have examined the use of new high-performance geosynthetics to improve the behaviour of granular fills in low deformation road structures.

Resilient behaviour

This may be of interest for coarse fills in the foundations beneath highway pavements which are subjected to repeated stresses from traffic loading. Partially saturated

Alkali wastes

Volumetric expansion has been attributed to the formation of ettringite by a reaction between calcium sulfate and calcium aluminate.

4.7 Dynamic loading Dynamic loads may change sufficiently rapidly with time to comparison induce inertia forces the fill that are significant in with thewithin static forces. Dynamic cyclic loading may produce stress–strain relationships in fills that are different from those under static loading and which are described by dynamic moduli and damping (section 3.6). Non-engineered loose sandy fills may be subject to loss of strength and associated reduction in volume when subjected to dynamic loading. They may be vulnerable to dynamic loading from a number of causes such as:  earthquakes  traffic vibrations  machine foundation vibrations  pile driving  blasting The UK is an area of low seismicity where the probability of earthquakes causing significant damage is relatively low. Irving (1985) commented that until 1970 the only structures built in Britain which allowed for significant earthquake forces were a number of dams in an area of Scotland that had suffered lengthy and repetitive swarms of moderately sized earthquakes. However, following on from the work of the nuclear industry, increasingly seismic risk is being addressed in the design of major structures. The Department of the Environment commissioned a review of seismic hazard and the vulnerability of the built environment (Ove Arup and Partners, 1991). Guidance has been provided on seismic risk for the assessment of dam safety (Charles et al, 1991;

Compaction

A major effect of dynamic loading on many coarse fills is that the vibrations compact the soil into a smaller volume. Vibration is used both in the field and in the laboratory to achieve dense conditions. Laboratory compaction tests for coarse fills may be performed by vibration using either a vibrating table or a vibrating hammer. The maximum of coarse fills is usually determined by these types ofdensity te st. Field compaction of coarse fills is best achieved by heavy vibrating rollers. Compaction can be monitored by measuring the vibration of the roller (Forssblad, 1981). Insitu compaction of natural coarse soils and loose fills can be carried out using vibro-compaction. It has bee n generally agreed that e ffective compaction can only be achieved when the amount of silt-sized particles is less than about 15% to 20%, but, with modern higherpowered vibrators, soils with a higher fines content can be effectively compacted (Slocombe et al , 2000) (section 8.2). The usefulness of vibration as a means of densifying loose coarse fills serves as a reminder that such fills in their loose state are susceptible to a reduction in volume if accidentally subjected to vibration through human activity (such as pile driving) or from a natural cause (such as an earthquake). Earthquakes have caused significant ground subsidence. In Valdivia, Chile, the settlement due to densification during the 1960 earthquake was 1 m with volumetric strains of the order of 1% to 5% (Nagase and Ishihara, 1988). Although these effects during earthquakes are often associated with the liquefaction of saturated sands, this is not always the case. Seed and Silver (19 72) report cases of settlement of dry and moist sand during the 1971 San Fernando earthquake and the 1963 Skopje earthquake. Tokimatsu and Seed (1987) have proposed simplified methods of estimating settlement of both saturated and partially saturated sand

4.7 Dynamic loading

47

deposits subjected to e arthquake shaking. Fills that are already dense will be much less affected than fills in an initially loose condition. Uniformly graded fills are more susceptible to this form of compaction than are well graded fills. Some laboratory investigations (Brand, 1972; Dobry and Whitman, 1972; Morgan and Markland, 1981; Chang and Chae, 1987) identified some of the main features of this type of compaction:  acceleration was the principal factor controlling compaction; amplitude, frequency and surcharge were not so important  there was an optimum acceleration between 1.5 g and

(1986) have described a site at Barrow-in- Furness where hydraulic sand fill was de nsified to improve liquefaction resistance to earthquakes. In the 1989 7.1 magnitude Loma Prieta earthquake, a large area of saturated hydraulic fill placed in 1915 liquefied and caused extensive damage to buildings. However, Pyke et al (1978) concluded that the risk of damage due to liquefaction of hydraulic fills in seismic areas was not as great as had sometimes been stated. Humpheson et al (1991) have demonstrated that a hydraulically placed pfa fill in Peterborough was not vulnerable to liquefaction under the maximum strength earthquake likely in that area .

3g which produced maximum compaction final density did not depend on initial density  dry and saturated samples showed similar behaviour  vibration had a greater effect on uniform sand than it had on well graded sand.

Several investigations of the effect of earthquakes on refuse landfills in the USA have been reported (Orr and Finch, 1989; Siegel et al, 1989; Duplancic, 1989; Rathje and Bray, 1999). The 1989 Loma Prieta earthquake produced only minor damage to refuse landfills close to the epicentre. It was concluded that the wastes tended to attenuate the effects of the earthquake. Practical methods of evaluating liquefaction potential of sand deposits during earthquakes are of major concern in seismically active regions (Ishihara, 1977). It is difficult to determine the in-situ density of sands in the field and therefore density index and liquefaction potential are often inferred from penetration testing. Liquefaction potential can be related to SP T, CPT and dynamic probing results (Schneider et al, 1999). In-situ test methods are described in Chapter 7. In addition to determining the critical conditions under



Liquefaction

This phenomenon has been described in se ction 3.6. Earthquake, blasting or some other form of vibratory or shock loading could cause liquefaction, and major damage has occurred when fills have liquefied during large earthquakes. Seed (1987) considered that there were two main problems confronting the geotechnical engineer dealing with a situation where soil liquefaction may occur:  determining the stress conditions required to trigger liquefaction 

determining the consequences liquefaction in terms of potential sliding and potentialofdeformations. Density index ( ID) is an important parameter in determining the resistance to liquefaction of coarse fill. The build-up in pore pressure is considered to be due to a potential reduction in volume produced by the cyclic loading. Thus, one requirement for liquefaction is that densification is possible. Greater ID requires greater energy for liquefaction and it is not likely to occur in dense fills owing to their tendency to dilate. The particle size distributi on for the most easily liquefiable soils is between 0.07 mm and 0.6 mm. Sandy materials are therefore most susceptible, while an increased fine content reduces the tendency to liquefaction. The boundary for potentially liquefiable soils is sometimes quoted as between 0.02 mm and 2.0 mm. The permeability of the material limits the time for which the excess pore pressures remain. A high permeability will mean that they dissipate quickly and may not even develop. Liquefaction is therefore less likely in gravelly soils due to their greater permeability. However, Andrus and Youd (1989) have described the liquefaction of gravelly soils during the Borah Peak, Idaho, magnitude 7.3 earthquake in 1983. Generally, sandy and silty hydraulic fills have been reckoned to be susceptible to liquefaction under dynamic loading. Although this has been of concern mainly in the more seismically active regions of the world, Bell et al

whichan liquefaction it may also desirable to make estimate ofwill theoccur, displacement thatbewould occur after failure. Liquefaction failure can lead to instability in large masses of soil. These flow slides have been observed in both fills and natural deposits. Flow slides have also been reported where no significant source of cyclic loading has ever been identified ( Kramer and Seed, 1988) and may occur in underwater slopes of loose sands owing to dredging, scour, filling or change in water level . If it is decided that some form of ground improvement is required prior to construction, it would seem sensible to select a tre atment technique that is closely related to the perceived hazard. If settlement caused by the weight of the buildings is considered to be the major problem, then preloading might be the preferred solution. If liquefaction under dynamic loading is the major hazard, then a treatment technique involving dynamic loading such as vibro-compaction, dynamic compaction or blasting could be appropriate. Ground treatment is described in Chapter 8.

48

Chapter 5 Collapse compression on wetting It may now be seen that it is possible to compact a soil so firm and hard as to appear entirely suitabl e for a dam and for this same soil to become very soft and unstable when percolating water saturates it. (Proctor, 1933)

Most poorly compacted, partially saturated fills undergo a reduction in volume when inundated or submerged for the first time and if this occurs subsequent to construction on the fill, buildings can suffer serious damage. This often represents the most serious hazard for buildings on fill. The phenomenon is usually termed ‘collapse settlement’ or ‘collapse compression’ . It was srcinally described as collapse settlement because it was considered to be associated with a collapse of the soil structure. Collapse compression is a widespread phenomenon affecting both natural soils and fills and can occur without any change in applied total stress. Most types of partially saturated fill are susceptible to collapse compression under a wide range of applied stress when first inundated if they have

successful. The most serious problem was water finding a route out of the embankment and causing slope instability or erosion. The volume of water required was 15% of the fill volume. In contrast to these highway embankment fills where collapse compression was used as a method of ground improvement, collapse compression was identified by Proctor (1933) as a potential hazard in embankment dam fills. Proctor developed a procedure involving field and laboratory controls to ensure that cohesive fills were placed at the correct moisture content to reduce settlement or softening when the e mbankment dam fill was subsequently saturated. These procedures for the placement and compaction of fill were introduced by

been placedcompression in a sufficiently loose and/or dry condition. Collapse received little attention in the early years of modern soil mechanics. This may have been because the phenomenon might see m to conflict with the principle of effective stress; for example, the submergence of a fill by a rising ground-water le vel increases pore pressures thus reducing effective stresses and consequently heave of the ground might be expected, whereas in practice saturation often causes settlement due to collapse compression. Despite this lack of attention from those leading the development of soil mechanics as a scientific discipline, the phenomenon was recognised in practice. Collapse compression was utilised to compact fills more than 70 years ago. Water jetting was used by the Illinois State Division of Highways to induce settlement of clay fills in highway embankments prior to pavement construction (Hathaway, 1926). A 1 inch pipe fitted with a nozzle was forced downwards by means of the water jet until it reached the bottom of the fill. Treatment was at 3 m to 6 m spacings. Two or th ree applications of wat er might be required at each point be fore saturation was reached. Most settlement occurred within two days of treatment with little occurring after the first week. Similar methods were used to induce settlement in embankment fills on the Houston–Galveston road (Wise, 1929). Ponding was used on fills up to 1.2 m high and je tting for deeper fills. Holes were at 1.5 m centres and pe netrated to not closer than 0.6 m from the bottom of the fill. The method, described as water tamping, appeared to be very

Proctor (1933) in aRecord series. of four articles published in Engineering News Problems caused by collapse compression in foundation fills were reported by Jaky (1948). In 1940, buildings along the bank of the Danube in Budapest suffered damage when the river level rose 2 m above the flood mark; banks of loose gravel were subject to large settlements on first inundation (about 6%), but only one tenth to one-hundredth of this initial settlement on subsequent inundations. The solution was compaction prior to building construction. Much early work on collapse compression was concerned with natural sands and silts. For example, Clevenger (1956 ) studied the characteristics of loess as a foundation material in the USA; Jennings and Knight (1957) described large sudden settlements of sandy soils in Southern Africa where movements were associated with the presence of water in the ground (such as from broken drains); Pilyugin (196 7) carried out a major field infiltration trial in a loess soil in the Northern Caucasus. The assessment of the collapse potential of fills and its significance for building on fill has been described by Charles and Watts (1996).

5.1 Mechanisms of collap se com pression Various mechanisms may cause collapse compression on wetting. The phenomenon was well known in loosely tipped rockfills and Terzaghi (1960) suggested that saturation reduced the compressive strength of the rock

5.3 Field investigations thus causing collapse compr ession. In a re view of collapsing soils, Dudley (1970) concluded that for collapse to occur the soil must start with a structure that is open (ie large voids ratio) and must have a temporary source of strength to hold the soil grains in position against shearing forces. These temporary sources of strength are reduced by the addition of water. The source of strength could be capillary tensions or cementing agents. Collapse compression in fills on wetting can be attributed to three basic processes. Weakening of interparticle bonds

There may be intergranular bonds within the fill which are weakened or e liminated by an increase in moisture content. These bonds may be due to the following effects:  capillary forces  silt and clay particles  cementing agents This may be the predominant cause of collapse compression in some sand fills. Recent developments in understanding the behaviour of partially saturated soils have particular relevance to collapse compression caused by weakening of the interparticle bonds attributable to suction. Weakening of particles in coarse fill

The parent material from which the fill is formed may lose some strength as its moisture content increases and

49

examined in oedometer tests. Sometimes doubleoedometer tests are carried out on pairs of ‘identical’ samples (Jennings and Knight, 1957). One sample is loaded in the as-compacted condition and then inundated with the maximum load applied to the sample, the other is soaked prior to loading. The difference in vertical strain at each loading increment is a measure of the vulnerability to collapse compression under that applied stress. The double oedometer test procedure has been one of the most common approaches to evaluating collapse potential and the test procedure can give a measure of collapse potential over a wide range of stress. There have been many laboratory investigations of collapse compression of fills. Walker and Holtz (1951) presented data for a silty clay showing the way in which collapse compression reduced as the air voids subsequent to compaction reduced. Cox (1978) carried out tests on Keuper Marl. Clayton and Simons (1980) reported re sults from tests on a compacted hoggin fill. Lawton et al (1989) carried out a laboratory test programme on a moderately plastic clayey sand (wL = 34%, IP = 15%) and provided contours of collapse compression; the following conclusions were reached:  depending on the stress level at which it was wetted, the fill exhibited either collapse compression or swelling at constant vertical stress  collapse compression could be eliminated for all usual stress levels by compacting at a moisture content greater than or equal to the optimum moisture content 

approaches saturation. This may be the predominant cause of collapse compression in relatively uniformly graded rockfills. Weakening or softening of aggregations of particles in fine fill

Where a fill is formed of aggregations of fine particles, such as lumps or clods of clay, these aggregations may lose strength as the moisture content increases. This may be the principal cause of collapse compression in stiff clay fills. Inundation may cause more than one effect:  in some coarse fills it may reduce the stiffness of the fill by weakening both interparticle bonds and the particles themselves  in some fine fills it may cause both a loss of strength and a swelling of the aggregations of particles; whether this causes a net reduction or increase in volume will depend on, among other factors, the stress level. Collapse of sand fills and silt fills associated with the destruction of interparticle bonding may be quite similar to collapse of natural sands and silts. Collapse in rockfills and clay fills associated with weakening of rock fragments and aggregations of particles is dissimilar to collapse in natural soils.

5.2 Laboratory investigations The effect of inundation on a partially saturated fill can be

the double oedometer test was adequate to pre dict collapse compression. The effect of stress ratio on collapse compression was examined by Lawton et al (1991) using a double -triaxial test procedure analogous to the double-oedometer test. Tests on a compacted clayey sand e stablished that the magnitude of the volumetric strain depended on mean normal total stress and was independent of principal stress ratio. However, the individual components of volumetric strain, the axial and radial strain, depended significantly on principal stress ratio. A laboratory test programme has been carried out at BRE in a 150 mm diameter oedometer on recompacted samples of three fill materials: a granular colliery spoil, an opencast mining mudstone and a boulder clay. The test results are described in section 5.4.

5.3 Field inv estigations Building developments and road construction in the UK have been increasingly taking place on opencast mining backfills and, in 1973 , BR E started an investigation to determine whether ope ncast backfills, which at that time were usually placed without systematic compaction, were susceptible to collapse compression. The backfills could be composed of mudstone and sandstone fragments, shale or clay. This work identified collapse compression as a major hazard and determined its magnitude in a variety of field situations. Table 11

50

Chapter 5 Collapse compression on wetting

Table 11 Collapse compression measured in non-engineered fill Collapse Cause of F i l lt y p e c o m pr e s s i o n( % ) w et t i n g *

moisture content as measured in the heavy compaction test (4.5 kg rammer). Fills up to 30 m deep were placed for a housing development in an area where irrigation produces a change from a natural annual precipitation of 200 mm to an equivalent annual rainfall of 1.8 m. The specification required compaction to 90% of modified Proctor at moisture contents close to optimum. At one site the average dry density achieved was 92% of modified Proctor maximum at an average field moisture content 0.4% dry of the modified Proctor optimum; at another site relative compaction was 93% and average moisture content 0.2% dry of optimum. Some heave was

Mudstone/sandstone Clay/shale fragments

2

Mudstone/sandstone 6

clay Stiff

3

Colliery spoil

4

1 1

7 1

2

8 2

2 7

h istory

1

5

clay Stiff

Case

5 2

10

*1 = rise in ground- water; 2 = downward infil tration of surface water

summarises the magnitude of collapse compressions that have been measured in these investigations. T here are two different causes of wetting: 1 a rise in gr ound-water lev el can trig ger settl ement at depth within the fill 2 downward infiltration of surf ace water can trigg er settlement close to ground level . The infiltration of ponded water beneath a reinforced concrete footing on collapsible soil was investigated by El-Ehwany and Houston (1990). It was found that partial wetting produces only partial collapse. Houston et al (1988) have also described a field ponding test. Although collapse compression can affect most poorly compacted partially saturated fills, there are some significant differences between coarse fills and fine fills. Coarse fill

The effect of rising ground-water levels on the settlement of opencast coal mining backfills compos ed of mudstone and sandstone fragments at Horsley (case history 4) and composed of clay/shale fragments at West Auckland (case history 7) has been monitored (Table 11). Clayton and Simons (19 80) have described how a 5 m to 6 m depth of fill was placed below a concrete apron slab abutting a major structure during a dry autumn period, and underwent a small collapse upon becoming wetted up in the winter rains. The hoggin fill was a well graded sand and gravel with 6% to 12% clay content. Vibrating rollers could not be used adjacent to structures and compaction was effected by three passes of a 55 kgf vibrotamper on 100 mm layers. Air voids were believed to be greater than 12%. Heavy rainfall caused settlements of 200 mm. Laboratory tests at 75 kPa vertical stress showed no significant collapse compression with initial air voids less than 10%; with 25% air voids, collapse could be 5%.

expected as moisture contents increased after construction. This was observed in shallower parts of the fill. However, five years after construction some structures built on deeper fill began to suffer distress due to settlement. Ten years after construction some areas had settled as much as 0.45 m. Laboratory testing showed heave when specimens were wetted under vertical stresses below 50 kPa and collapse above 50 kPa. At a stress of 600 kPa there was 3% collapse compression which is comparable with the field measurements. Chalk fill

Clayton (1980) investigated the collapse of compacted chalk fill. Collapse compression of 4.5% occurred in chalk fill behind a motorway bridge abutment. The main danger occurred when soft chalk was excavated, allowed to dry out in a stockpile and then compacted as fill. Laboratory tests on blocky chalkinitial showed 1% collapse compression with 20% air only voids, increasing to 5% with 35% air voids. In contrast, dried-out soft chalk showed collapse compressions of 40% of the initial air voids.

5.4 Magnitude of collapse potential The variety of mechanisms causing collapse compression makes it difficult to de rive widely applicable relationships linking collapse potential to physical properties such as density and moisture content, air voids and degree of saturation. The magnitude of the collapse potential of a particular fill is dependent on both its present state and its history; it is a function of both the past and the present density and moisture content of the fill and the stresses to which the fill has been subjected since it was placed. The magnitude of collapse compression is examined under the following headings:  placement conditions  moisture content history  stress history

Fine fill

Case histories 2 and 5 describe the effect of infiltration of surface water into stiff clay opencast mining backfills. Measured collapse compressions are summarised in Table 11. Brandon et al (1990) have described collapse settlement of engineered sandy clay fills in southern California which were placed close to the optimum

Placement conditions

The conditions in which the fill is placed have a major influence on subsequent collapse potential. To take an extreme example, a fill placed hydraulically would be most unlikely to be vulnerable to collapse compression. The collapse potential of fills placed by normal earthmoving plant will decrease as placement density

5.4 Magnitude of collapse potential

51

and moisture content increase; heavy compaction and placement at a high moisture content should eliminate or at least greatly re duce collapse potential. Figure 19 shows collapse compression of some opencast mining backfills plotted as a function of the percentage of air voids immediately before inundation. A laboratory test programme has been carried out on recompacted samples 80 mm high at BRE using a 150 mm diameter oedom eter (Charles and Skinner, 2001a). Three fill materials were tested: a coarse colliery spoil, an opencast mining mudstone and a boulder clay. The samples were wetted to the appropriate moisture content, compacted in the oedometer ring to the required density, loaded to 30 kPa and then submerged. Generally there was a rapid re sponse to submergence. The effects of the initial moisture content and density on the collapse compression of a coarse fill (Skinner et al , 1997) are illustrated in Figures 20 and 21. Three samples of colliery spoil at 8% moisture content were compacted to different densities. Figure 20 shows that, on inundation at an applied vertical stress of 60 kPa, the densest sample suffered no collapse compression, whereas the loosest sample suffered a collapse compression of 6%. Two samples of a colliery spoil which were formed at the same low dry density, one sample at a moisture content close to the Proctor optimum and the other completely dry, both showed major collapse compression (Figure 21). Contours of percentage volume change on submergence are plotted for the three fills in Figures 22, 23 and 24. of collapse compression are in plotted on aThe drycontours density versus moisture content plot, which dry density is expressed as relative compaction, CR, and moisture content is plotted in re lation to Proctor optimum, wopt. An initial dry density corresponding to 95% of the maximum dry density achieved in the standard Proctor compaction test was sufficient to eliminate collapse potential in the colliery spoil and the mudstone fill. This was not the case for the clay fill when compacted dry of optimum. The results for a clayey sand fill at a vertical stress of 400 kPa (L awton et al , 1989) are plotted in Figure 25, and 95% relative compaction is not adequate in this case at a moisture content dry of optimum. The results of the BRE test programme suggest that 95% relative compaction based on the standard Proctor compaction test might largely eliminate collapse potential with some coarse fills, but with fine fills there could be a collapse potential of as much as 2% where the fill is compacted dry of optimum moisture content. Thus, on its own, a 95% relative compaction criterion is not an adequate compaction requirement. It might be satisfactory if it was linked to a criterion of air voids of 5% or smaller or, possibly for some fills, to a criterion of a moisture content greater than standard Proctor optimum. An important consideration with engineered fills (Chapter 9) is to eliminate entirely, or at least largely, collapse potential.  Heavy compaction of granular fills during placement in layers at moisture contents typically found in the U K will eliminate or greatly reduce subsequent

Figure 19 Collapse compression of opencast backfills in oedometer tests as a function of percentage air voids

Figure 20 Effect of initial dry density on collapse compression of a coarse colliery spoil

Figure 21 Effect of initial moisture content on collapse compression of a coarse colliery spoil

52

Chapter 5 Collapse compression on wetting

Figure 22 Collapse potential of coarse colliery spoil at applied vertical stress of 60 kPa in relation to standard Proctor compaction (contours show collapse potential in per cent) (after Charles and Skinner, 2001a) Figure 24 Collapse potential of boulder clay fill at applied vertical stress of 30 kPa in relation to standard Proctor compaction (contours show collapse potential in per cent) (after Charles and Skinner, 2001a)

3

) /m g M ( d ρ

Figure 23 Collapse potential of mudstone fill at applied vertical stress of 30 kPa in relation to standard Proctor compaction (contours show collapse potential in per cent) (after Charles and Skinner, 2001a)

Figure 25 Collapse potential of clayey sand at applied vertical stress of 400 kPa (contours show collapse potential in per cent) (after Lawton et al, 1989)

5.4 Magnitude of collapse potential vulnerability to collapse compression. Adding a small quantity of water to the fill during placement may also be beneficial.  With a clay fill it is necessary to reduce air voids to about 5% to remove vulnerability to collapse compression. It is the large pores and voids be tween lumps of clay that have to be largely e liminated. The vulnerability to collapse compression is linked to the bulk permeability of the clay fill.  Where air voids are reduced to 5% or less there will be low permeability (typically 10 –9 m/s to 10–10 m/s) and 

no vulnerability to collapse compression. Where air voids are high there will be high permeability (typically 10 –6 m/s) and vulnerability to collapse compression; after collapse compression permeability will be reduced.

Research on compacted clay liners which act as barriers to the movement of contaminants has given an improved understanding of the compaction required to achieve low permeability and, therefore, to eliminate collapse compression. Daniel and Benson (1990) examined the influence of moisture content and compaction method on permeability. Elsbury et al (1990) investigated an unsatisfactory section of liner and attributed the poor performance to the use of a roller too light to remould the lumps of clay at the placement moisture content. Two sections of the clay liner had permeabilities of –7

53

front. Laboratory tests indicated that a degree of saturation of 50% produced about 85% of the full collapse potential. There was full collapse with a degree of saturation above 70%. Stress history

The BRE results plotted in Figures 22, 23 and 24 refer to submergence at a low stress. Clearly collapse compression is stress dependent and both Cox (1978) and Lawton et al (1989) have provided data showing the extent of stress depende ncy on collapse compression contours for the soils that they tested. The results obtained with collapse at 30 kPa vertical stress will tend to under-estimate the collapse at higher stresses. Contours of collapse potential for a clayey sand at a much higher stress are shown in Figure 25. Many fills have a simple stress history as the fill has been loaded under conditions of confined compression to its present stress level, but there are important exceptions. The temporary preloading of a fill, usually with a surcharge of fill, produces an over-consolidated state and reduces air voids. Charles et al (1977) have shown that preloading a mudstone and sandstone opencast mining backfill greatly reduces the collapse potential. The general effects of preloading as a method of improving the load-carrying characteristics of fills have been reviewed by Charles et al (1986). Most fills are normally consolidated, and many coarse fills with this simple stress history exhibit the following

–7

3 × 10 m/s andthrough 10 × 10them/s respectively. Seepage predominantly macro-voids between thewas soil clods and along the interlift boundary. It was concluded that density and initial degree of saturation had little meaning if the basic compaction objectives of remoulding soil clods and bonding between lifts were not achieved.

collapse characteristics: a minimum stress, often very small and typically about 10 kPa, below which collapse does not occur  a wide range of stress which is of practical interest, typically from 20 kPa to 400 kPa, within which the stress level has a relatively minor e ffect on the collapse potential 

Moisture content history

The moisture content of the fill usually remains unchanged during fill placement while the density increases due to consolidation under the weight of overlying fill. Although collapse potential will be largely controlled by the moisture content and density immediately prior to inundation, there are some important qualificati ons to this.  If a partially saturated fill has previously been fully saturated, very little collapse should be expected on subsequent wetting. Jaky (1948) found that settlement of a gravel on a subsequent inundation was between one-tenth and one- hundredth the settlem ent on first inundation. Strohm (19 78) reported that repeated cycles of soaking and draining a shale fill did not produce any significant increase in compression.  If a partially saturated fill has previously had a higher degree of saturation, but has never be en fully saturated with consequential compression, there could still be a substantial potential for compression on wetting. ElEhwany and Houston (1990) found that, in an infiltration test on a natural collapsible soil, the degree of saturation averaged about 50% behind the wetting

Figure 26 Collapse compression of a coarse colliery spoil at an initial moisture content of 0% and dry density of 1.60 Mg/m3 submerged at different applied stresses

54 

suppression of collapse potential at very high stresses.

Chapter 5 Collapse compression on wetting 5.6 Identification o f coll apse po tential

Blanchfield and Anderson (2000) have provided laboratory data for a particular mudstone fill showing a linear relationship between collapse potential and applied stress. Charles and Skinner (2001b) have discussed why this is not typical of field behaviour observed in BRE monitoring of opencast backfills. Figure 26 shows that collapse com pression of a coarse colliery spoil was not too greatly affected by applied stress between 20 kPa and 120 kPa.

In view of the significance of collapse settlement for building development, a key element in an investigation of a filled site prior to building should be to identify and quantify collapse potential. The general application of site investigation methods to fills is described in Chapter 7. In this section some matters directly associated with the identification of collapse potential are highlighted.  All available information should be used in the assessment, including historical information, site investigation data and, where appropriate, special in-

5.5 Time dependency

situ infiltration tests. In many cases conventional site investigation methods alone will not provide a satisfactory answer.  The underlying basis of the assessment should be the assumption that any partially saturated fill placed without systematic compaction is vulnerable to collapse compression unless there is adequate evidence to the contrary, such as knowledge that the fill has been submerged at some stage in its history.

The practical consequences of collapse compression are related not only to the magnitude of the compression but also to the rate at which collapse movements occur. Time dependency is a common feature of collapse compression. There are two factors which have a major influence on the time dependency of the movements associated with collapse compression: the supply of water and the nature of the fill material.

History of fill Rate of supply of water

The rate at which collapse compression takes place will be largely controlled by the rate at which the fill is wetted. Where there is inundation from the ground surface the rate of infiltration will be dependent on the supply of water at the ground surface and the permeability water of the partially saturated rate at which to a ground-water level rises withinfill. a fillThe may be largely controlled by the hydrogeology of the natural strata surrounding the fill. Rate of response to wetting

Collapse compression may continue for a significant period after the source of the wetting has been removed. At Horsley, the ground-water level rose 34 m through a mudstone and sandstone opencast coal mining backfill during a three-year period (case history 4) but the rate of surface settlement did not reduce to the rate that it was prior to the ground-water rising for several ye ars. At Corby, where an inundation test was carried out from 1 m deep trenches in the stiff clay opencast ironstone mining backfill, the surfa ce settlement did not re duce to the preinundation rate for another six years and about 40% of the total settlement occurred after the addition of water had ceased (case history 2). Where wetting is due to infiltration from the ground surface, this type of timedependent behaviour may be at least partly due to the continuing percolation through the fill of water initially added to the fill. However, it can be concluded that the mechanisms caus ing collapse have time-de pendent effects, particularly for clay fills. The practical consequences of this time dependency should be recognised. Where collapse compression is causing a problem, removing the source of the wetting is unlikely to halt the damaging movements immediately.

Historical information should be collected and evaluated. Where there is adequate knowledge of the srcin and subsequent history of a fill, it should be possible to make some preliminary judgement as to whether or not collapse compression is a hazard.  Origin The information about placement conditions includes the type required and extent of fill, the date and method of placement and compaction, and data concerning the initial moisture content and density of the fill.  Subsequent history Information is needed about the subsequent stress and moisture content history of the fill, including the placement of any additional fill or the removal of fill and any fluctuations in ground-water level or substantial surface inundations. Present state of fill

A site investigation involving trial pits and boreholes should be undertaken to provide information on the extent and type of fill and its variability . Large bulk samples enable particle size and moisture content to be measured. Where high-quality small undisturbed samples can be obtained, laboratory oedometer tests can be undertaken to measure collapse compression. The double-oedometer test procedure makes it possible to measure collapse potential over a wide stress range. When undisturbed sampling is not feasible, an assessment of collapse potential may be based on a correlation between collapse potential and parameters such as dry density, moisture content and percentage of air voids derived from tests on re compacted samples of the fill under specified stress conditions. A knowledge of the in-situ density and moisture content of the fill is required, and a fill with a substantial proportion of large

5.7 Damage by collapse compression

55

particles presents particular problems. The approach, while useful, may not accurately model the actual moisture content history and stress history of the fill. In-situ tests such as dynamic probing and standard penetration test may give some indication of the density of the fill and hence provide additional information. Appropriate geophysical tests may also be of assistance in some situations. Infiltration tests may be used to assess collapse potential. For near-surface ground conditions, the test may involve filling a trench with water and observing ground settlements (see, for e xample, case history 5). The

vertical compression of the newly saturated fill was measured. It was found that, following closure of local mines, pumping had ceased in November 1972 and the water table had risen 40 m and inundated the bottom of the opencast backfill. Piles were ineffective as they did not penetrate through the full depth of the backfill.

feasibility of using borehole infiltration tests to assess collapse potential at depth has been examined by Charles and Watts (1996). Although an initial assessment of a borehole infiltration test for a granular fill with some fines was promising, further work is needed to examine the applicability of the test for other fill types. Air permeability is a useful measurement for clay fills. Where air pressure can be maintained in a pie zometer, there are no interconnecting macro-voids and collapse compression is unlikely to occur.

air voids, and vibrated stone columns were installed. Within a few months of construction, settlement of 110 mm had been observed. Th e settlement has been attributed to compression of the chalk fill due to the combined effect of new loading from some sand and gravel fill and subsequent wetting from soakaways, just after the buildings had been roofed, following periods of heavy rain.

5.7 Buildings damaged by collapse compression When buildings on fill are damaged owing to settlement it is not always easy to determine the cause of the ground movements with certainty. However, in many cases there has beencompression good re asonof tothe attribute tomost collapse fill andthe it ismovements probably the frequent cause of damaging settlements in poorly compacted fill. A number of brief case studies are used to illustrate the range of situations in which collapse compression has damaged buildings. More detailed accounts of the case histories can be found in Part IV for the first two case histories and in the listed refere nces for the other four. (A) Houses at Ilkeston (case history 5)

A block of eight two-storey houses was built on a stiff clay opencast backfill with a maximum depth of 12 m. The opencast site had been restored in 1959. Soon after the completion of the brickwork early in 1973 excavations for drains began and damage was observed close to one gable end. Following heavy rain, movement took place in the centre of the block. An inundation test carried out by BRE in 1975 using 3 m dee p trenches indicated that the fill was susceptible to collapse compression. The block was never occupied and was de molished in 1982 at which time a total settlement of 0.3 m was estimated. (B) Factory at West Aukland (case history 7)

In 1971, many years after restoration, a single-storey factory was built on an 18 m deep opencast mining backfill composed of clay with shale fragments. Se rious differential settlement occurred and measurements in July 1977 showed a maximum of 0.21 m. A 2 m to 3 m rise in ground-water table occurred early in 1979 and 4.7%

(C) Industrial estate in Brighton

Collapse compression in a 100-year-old chalk fill caused damage and delay to the construction of an industrial estate in Brighton (Stroud and Mitchell, 1989). The 6 m deep chalk fill was in the form of a loose rubble with high

(D) Houses in southern California

Although most problems with building on fills have been associated with non-engineered fills, collapse compression has occurred on occasions where fills have been placed to a specification under controlled conditions. Sandy clay fills up to 30 m deep were placed to a specification which required 90% of modified Proctor dry density at moisture contents close to optimum in an area where irrigationof produces from a natural annual precipitation 200 mmatochange an e quivalent annual rainfall of 1800 mm (Brandon et al , 1990). In areas where the fill was shallow, heave was observed. Five years after construction, some stru ctures built on deeper fill began to suffer distress due to settlement. Ten years after construction some areas had settled as much as 0.45 m. (E) North Tyneside District General Hospital

A low-rise hospital development has taken place since 1982 on an opencast coal mining site some 30 years after working and restoration. The 20 m deep clay and shale fill had been placed without systematic compaction. The buildings have suffered significant settlements of up to 160 mm requiring some remedial work. It has been concluded that the acceleration of settlements at constant stress which have been observed at the site could only be attributed to wetting up of the fill (Kilkenny, 1998). (F) Clifford’s Tower in York

The 15 m high mound was built in York in 1069 by William I during his campaign to subdue the north of England. It was built on low-lying ground adjacent to the river. The mound was built in horizontal layers of a fill comprising stones, gravel and clay. The fortifica tion was initially a timber structure. The stone tower known as Clifford’s Tower was not built until the middle of the thirteenth century. Some years after its construction, in 1315–16, severe floods softened the fill in the mound

56

Chapter 5 Collapse compression on wetting

Figure 27 Clifford’s Tower, York

(English Heritage, 1987) and in 1358 the tower was described as ‘cracked from top to bottom in two places’. These cracks are still visible. Although little was done to remedy the damage to its foundations until it was underpinned in 1903, the tower remained standing as seen in Figure 27. Conclusions from case studies

Ground movements due to collapse compressions may be large, and vertical compressions of the order of 3% to 5% are not unusual. Serious damage is therefore possible (houses in case A were demolished).  Many types of buildings have been damaged by 

collapse compression houses A and D), industrial buildingsincluding (cases B and C) (cases and historic buildings (case F).  Fills of different composition and arising in different circumstances can be vulnerable to collapse compression, for example opencast mining backfills (cases A, B and E), old chalk fill (case C) and engineered clay fill (case D).

Inundation can be due to either submergence from a rising ground-water level or river level (cases B and F), or water infiltrating downwards from the ground surface (cases A, C, D and E).  Collapse compression can occur many years after the fill is placed: HousesinsouthernCalifornia 5years HousesatIlkeston 14years FactoryatWestAuckland 24years North Tyneside District General Hospital 30 years IndustrialestateatBrighton 100years Clifford’sTower,York 246years Thus mere passage of time does not eliminate collapse 



potential. The degree of time dependency of the ground movements is a function of both the type of fill and the source of the wetting. Movements are likely to occur over a particularly long period where a clay fill is affected by the downward infiltration of water (case A).

57

Chapter 6 Boundaries and variable depth Special care should be taken with buildings sited near the edges of filled ground; in particular placing a building partly on the natural ground and partly on fill should be avoided. Instead of this, the foundations should be carried down to the natural ground by piers or piles. (Digest 9, Building Research Station, 1949)

Distortion and damage of buildings are related to the magnitude of the differential settlement across the building rather than the total settlement . Variation in the depth of fill under a building can be a cause of differential settlement and may also result in damaging horizontal movements. There are particularly acute problems associated with differential settlement at the edges of a filled area. This topic is of major practical and economic importance. In reviewing the performance of an engineered ope ncast backfill at Hurst Industrial Park, Dudley, Waite and Knipe (1991) concluded that: ‘In the vicinity of the high and side walls of the opencast excavation, the potential for long term unacceptable differential settlement exists due to the rate of change in fill thickness although the levelling surveys have shown a re latively smooth transition across these areas. Thus, a “restriction zone” straddling the margins of the opencast pit was designated with no building development within this area, although it could be used for parking, storage or other non-sensitive use.’

Relatively little has been written on the problems occasioned by changes in depth of fill. Digest 9 Building on made-up ground or filling , published by BRE (then the Building Research Station) in 1949, gave the warning quoted at the beginning of this chapter. A BRE investigation of a factory built on fill described by Meyerhof (1951) led to a similar recommendation that: ‘In general, buildings should not be constructed partly on fill and partly on natural ground unless the former is compacted to the same density as the latter, nor should they be constructed near the edge of made-up ground, e specially where this ends in a free



settlement and extension were calculated. Problems due to varying depths of fill have occurred in California and Rogers (1992) recommended that the variation in fill depth under a building should not exceed 15% of the depth.

In this chapter the implications of soil behaviour adjacent to the boundaries of the filled area, and at other locations where there are significant changes in depth of fill, are explored. There is a nee d to define the areas from which buildings should be excluded because the rate of change in fill thickness gives the potential for unacceptable longterm differential settlement. If unsafe decisions are reached and exclusion zones are too small, building damage andan consequent legal action can result. other hand, over-conservative approach willOn the needlessly sterilise large areas of land and could make many sites uneconomic to develop. In determining the area of land where ground surface movements may be unacceptably large owing to variations in the depth of the fill, consideration needs to be given to both the properties of the fill and the geometry of the fill.

6.1 Influence of fill properties Two of the most important fill properties which control the differential settlement caused by a variable depth of fill are:  the volume reduction potential of the fill, expressed as vertical strain, εv  the stiffness of the fill, expressed in the form of the shear modulus, G.

slope, which is generally in a loose state.’

Such advice deals only with one aspect of the more general problem occasioned by variations in depth of fill and it presupposes that the edge of the fill has been identified and accurately located. The problems are not, of course, confined to the U K, as the following examples demonstrate.  Lange (1986) described work associated with the construction of pipelines across brown coal opencast mining areas in the Rhineland. Ground deformations over steep opencast mine slopes were analysed and the results presented in terms of influence values; both

Differential settlement associated with variations in the depth of fill will only be a problem if there is some volume reduction in the fill, and, from the standpoint of this particular hazard, the volume reduction potential is the most critical property of the fill. Most forms of volume reduction will be mitigated, if not eliminated, by heavy compaction of the fill during placement. Compression of the fill could occur due to a variety of processes, as described in Chapter 4, with the following being of particular significance:  compression may be caused by the application of

58

Chapter 6 Boundaries and variable depth

additional load to the surface of the fill, in which case εv will be a function of the constrained modulus, D  with engineered we t clay fills, primary consolidation under self-weight of the fill may be the major problem; this was a major issue in the redevelopment of the old brickpits at Peterborough where decisions about exclusion zones had significant economic consequences  in many cases, collapse compression will be the most likely cause of movement, and wetting will trigger the settlement; information on the likely magnitude of collapse compression on wetting is given in Chapter 5.

The highwall is likely to be the location of severe differential settlement and forms a particular hazard for buildings on fill. Figure 29 shows a simplified cross-section through a typical backfilled excavation. Significant changes in depth of fill are not confined to locations at the edges of the filled area and changes in the depth of the fill at any location will constitute a potential hazard. The precise extent of the areas within which the variability in depth of fill is likely to cause serious differential settlement problems is not immediately obvious. The geometry of the fill in the vicinity of a highwall can

If volume reduction potential is uniform throughout the fill deposit, and the fill has no stiffness, then the settlement profile will be purely a function of the fill geometry and the surface settlement at every point will be in proportion to the depth of fill at that location. In practice, the stiffness of the fill modifies this simple situation. Fill does not behave as a linear elastic material and it is difficult to characterise the stiffness in a simple way. The problem of characterising the stiffness is made more difficult by the complex patterns of stress changes which occur in the fill near to a highwall. On the basis of a BRE programme of research involving field studies, an instrumented pilot scale test in a 4 m deep test pit and finite element analyses, it has been concluded that in many cases the pattern of movement can be pre dicted

be defined by the following three parameters as shown in Figure 30:  height of the highwall, H  depth of burial of top of highwall, D  angle of highwall to horizontal, β.

with reasonable confidence using linear tilt and model, which is independent of filla simple stiffness (Charles Skinner, 2001c). T his is described in Appendix C .

6.2 Influence of fill geometry In a backfilled excavation, the severity of the differential settlement will depend on the steepness of the face of the excavation, and many different situations are commonly encountered:  an old dock with vertical sides  a quarry in hard rock with a ne ar vertical face.  gravel and clay pits with slopes that are not so steep. A wide variety of situations may be encountered in opencast mining in dipping strata. A steep slope at the deep end of the excavation is termed a ‘highwall’. Figure 28 shows fill being placed against a highwall in an opencast operation. The highwall may emerge at ground surface or may be buried under a substantial depth of fill.

Figure 28 Fill placed against highwall in opencast operation

Figure 29 Simplified cross-section of a backfilled excavation

6.3 Vertical highwall

59

Figure 30 Basic geometry of highwall Figure 32 Typical settlement profile for case A (after Skinner and Charles, 1999) Case A

Case B

Case C

Case D

Figure 31 Variable depth of fill: different geometries

The other geometrical factor relates to the location of the building in relation to the highwall. A wide variety of fill geometries may be met in practice, and it is helpful to examine the four cases which are illustrated in Figure 31.

6.3 Vertical highwall With a ve rtical, or near vertical, highwall the pattern of ground movements will be strongly influenced by the conditions at the interface between the highwall and the fill. The angle of friction, δ, at the interface of the fill and the highwall will be of primary importance. In general

0 < δ < φ' where φ' is the angle of shearing resistance. In many field situations, the interface will be relatively rough, but an old dock wall could be quite smooth. Two extreme cases are illustrated in Figure 31; case A represents a smooth interface with δ = 0 and case B represents a rough interface with δ = φ'. In case A there is a perfectly smooth interface between a vertical highwall and the fill and no fill has been placed over the top of the highwall. An infilled dockof may this condition at the location the approximate dockwall. Thetosituation is characterised by the following: β = 90° D=0 δ= 0 When there is volume reduction in the fill, slippage along the interface between the fill and the undisturbed ground will result in a step at ground level at the top of the wall. This is illustrated in Figure 32. Thus the building exclusion zone is, in theory, simply a line. This situation is very serious if a building has been located over the highwall. Otherwise it may seem beneficial be cause movements are concentrated at an identifiable location and the area from which buildings are excluded is reduced to a minimum. Case B in Figure 31 is characterised by the following: β = 90° D=0 δ = φ' The friction between the fill and the undisturbed ground means that volume reduction in the fill is unlikely to cause a vertical step in the ground surface at the top of the highwall. However, defining the area of land above the highwall which must not be built on is not simple, although differential settlement still will be relatively localised in the vicinity of the wall. Although friction at the fill/highwall interface is suffi cient to prevent a step in

60

Figure 33 Typical settlement profiles and exclusion zones for vertical highwalls (after Skinner and Charles, 1999)

Chapter 6 Boundaries and variable depth

Figure 34 Typical settlement profile and exclusion zone over buried highwall (after Skinner and Charles, 1999)

the ground surface, differential movement is still likely to

6.5 Long shallow slope

be severe a limitedthe area. With a large angle will of be a friction atover the interface, deformation pattern function of the shear modulus of the fill, G. From the point of view of building development, the key issue is the location of the highwall and the ne ed to avoid building over it. Figure 33(a) prese nts a comparison of the simple situation for case A with a typical settlement profile and exclusion zone derived from a finite e lement analysis of case B with εv = 1%. The exclusion zone in Figure 33(b) is based on an acceptable tilt criterion of 1/500 and is plotted as a function of εv.

In case D in Figure 31, fill has been placed on an undisturbed stratum with a relatively shallow slope (β < 15°). If the fill compresses uniformly and the bu ilding is small compared with the length of the slope and far from either end of it, differential movement of the fill will be entirely due to the variation in depth of fill under the building and will cause tilt of the building but not distortion. The question whic h has to be addressed is what variation in depth of fill under the building is acceptable. If the tilt of the building is α, then:

6.4 Buried highwall Deep burial of a highwall ensures that the severity of differential movements is greatly reduced, but the area of ground affected by differential settlement is increased. The deformed profile of the ground is principally a function of the ratio D/H and does not depend too strongly on the angle of the highwall or the fill properties, other than, of course, the magnitude of the volume reduction of the fill. Case C in Figure 31 illustrates a typical case of a buried highwall with the following characteristics: β > 30° D /H > 1 Settlement profile and exclusion zone derived from finite element analysis are shown in Figure 34 for the case where D/H = 4.

α = ε v δ z /L

where εv is the constant vertical compression strain in the fill δz is the variation in depth of fill under a building of length L δz = L tanβ

Combining the above two equations gives: α = εv tanβ

Figure 35 shows limiting values of β corresponding to different values of εv for an acceptable design tilt, α = 1/500. These results indicate that even relatively shallow slopes could cause problems if εv > 2% but that if εv < 0.5% there should be few problems. This finding

6.6 Exclusion zones

61

Figure 35 Critical slope angles for long shallow slope (after Skinner and Charles, 1999)

confirms the classificat ion of fills based on compression potential proposed by Charles and Burland (1982) in which 0.5% was identified as the value below which fill would generally form a good foundation material with few problems and 2% was identified as the value above which problems would be severe.

6.6 Exclusion zones In delineating the area of land above a highwall which should not be built on, two primary factors control ground movements:  the geometry of the highwall, which can be described by D, H and β  the volume reduction potential of the fill, which can be expressed in terms of a vertical compressive strai n, εv. Other properties of the fill such as stiffness and friction at the fill/bedrock interface, will in most cases affect the pattern of deformations to a smaller extent. The criteria for tolerable deformations of buildings and infrastructure, which are discussed in Chapter 10, have a

critical effect on the size of the e xclusion zones . When building on fill, many acute problems arise with small buildings for which deep foundations are not an economically viable solution. Fo r this type of building it is usually feasible to provide a stiff raft foundation which will prevent distortion of the building due to differential settlementinterest and will resist tensile forces. principal then lieshorizontal in the delineation of the The zone from which buildings should be excluded because tilt is unacceptably large. Ground deformations can be analysed using finite element techniques, but uncertainties concerning fill properties make this approach of limited value, and a simple linear tilt model for ground surface settlements has been developed and is de scribed in Appendix C . Using the linear tilt model, charts have been pre pared showing the dimensions of the exclusion zones in terms of normalised parameters. The basic variable is the normalised tilt, α/εv, where α is the tilt and εv is the vertical compression in the fill below the top of the highwall at some distance beyond the base of the highwall.

63

Part III: Construction on fills The value of an engineering science is determined by what it can accomplish as a tool in the hands of the practicising engineer. (Terzaghi, 1939)

The study of the engineering properties of fill is not an end in itself. It should always be closely linked to application. Some relevant aspects of engineering behaviour have been described in Part II and these are now applied to practical considerations in construc tion on fill. However, the warning in the Code of Practice for Foundations (BS 8004:1986) should always be kept in mind when building on non-engineered fills: ‘All made ground should be tre ated as suspect because of the likelihood of e xtreme variability.’

Successful buildi ng development on land with a history of previous uses requires input from several technical disciplines as well as legal and planning skills. A practical handbook for site selection and investigation has been produced by Lampert and Woodley (1991) which includes contributions from civil engineers and from specialists in environmental assessment, project management, planning, insurance, construction and contract law, surveying and property The handbook should help to provide some knowledge ofdevelopment. the different specialisms. Although this book focuses on experience within the U K, construction on fill is a world-wide phenomenon as the following examples indicate:  Rutledge (1970) commented that one only had to superimpose present-day maps on old maps of cities such as Boston and New York to realise that extensive areas were built on made land.  The New York World’s Fair was located at Flushing Meadow Park on a large depth of ash fill. Spread footings were widely used ( Foster and Glick, 1938). A few years later, housing construction took place in the city on refuse landfill (Eliassen, 1947).  Large settlements of industrial reclaimed land in Kawasaki City, Japan, were caused by the placement of a thick layer of fill ( Lambe, 1969).  Although Dmitrenko (1976) described four damaged buildings in Kiev which had been founded on fill, it was asserted that many buildings in the city had been successfully built on fill.  Buildings founded on uncompacted rubbish in Singapore settled more than 150 mm (Tan, 1977).  Since the late 1950s, there have been many large earthworks projects in the coastal hills of California’s largest urban areas to create level areas for residential developments; long-term saturation of compacted clay fills has led to both e xpansion and collapse compression (Noorany et al, 1992; Rogers, 1992).  Massive reclamati on projects have been carried out in Hong Kong, Japan and Singapore for the construction of new airports. The construction of the 1248 ha platform for Hong Kong’ s new airport at Chek Lap Kok involved excavating and placing 118 × 106 m3 of rockfill and the dredging and placement of 76 × 106 m3 of sand fill from marine borrow areas.

64

Part III: Construction on fills The four chapters in Part III deal with the following subjects:  the use of investigatory and monitoring techniques to identify, characterise and classify fills  the different types of ground treatment that are available, and the improvement in properties that can be produced  the specification and quality management of engineered fills  various aspects of foundations on fills.

65

Chapter 7 Investigation and monitoring Any proposal to found a structure on made ground should be investi gated with extreme care.

In any building or construction project, it should be recognised that the ground is a major area of risk. Adequate site investigation and, where appropriate, ground treatment (Chapter 8) should reduce that risk. A general account of site investigation practice can be found in Clayton et al (1995) and authoritative information on particular matters is contained in reports and codes prepared by BSI, CE N, ICE and AGS:  BS 5930:1999 Code of practice for site investigations (BSI, 1999)  BS 10175:2001 Investigation of potentially contaminated sites — Code of practice (BS I, 2001)  DD E NV 1997-3:2000 Eurocode 7: Geotechnical design — Part 3: Design assisted by field testing (BSI, 2000b)







Site investigation in construction (Institution of Civil Engineers Site Investigation Steering Group, 1993)  Code of conduct for site investigation, Guidelines for good practice in site investigation and Guidelines for combined geoenvironmental and geotechnical investigations (Association of Geotechnical and Ge oenvironmental Specialists, 1998a, 1998b, 2000) Construction on non-engineered fills should be preceded by careful investigation of the site which should be both appropriate and adequate. Recommendations for the procurement of ground investigati on are given in CI RIA Special Publication 45 (Uff and Clayton, 1986). A wide background to the subject including professional liabilities and insurance issues is provided in a practical handbook on site selection and investigation by Lampert and Woodley (1991) and in a book on risk management for construction professionals by Hatem (1998). It is important to determine the type, condition, variability and depth of the fill and the nature of the underlying soils. In carrying out the investigation, reference should be made to the Code of practice for site investigations (BS 5930:1999). Methods of determining the fill properties which control or influence fill deformations are of particular interest. Standard procedures for many field and laboratory tests are provided in BS 1377:1990. The investigation to assess the suitability of the site for de velopment may include:  historical review  site reconnaissance

(BS 8004:1986)

ground investigation laboratory tests  in-situ tests  load tests  geophysical tests  monitoring 

Site investigation for low-rise buildings presents particular problems due to the small scale of many developments and inadequate geotechnical input. BRE has issued Digests giving guidance on best practice for site investigation for low-rise building:  318 — desk studies (B RE, 1987a)  322 — procurement (B RE, 1987b) 348 — — trial the walk-over 381 pits (BRE,survey 1993c)(BRE, 1989b) 383 — soil description (BRE, 1993d)  411 — direct investigations (BRE, 1995c)  

This book is concerned with geotechnical considerations, but it is emphasised that a multi-disciplinary approach to the assessment of sites should be adopted ( Smith, 1980). The investigation of geotechnical aspects of fill behaviour, as described in the following sections, should be integrated with the investigation of other rele vant factors which could, in some cases, include chemical attack, toxicity, gas generation and combustibility. The presence of old foundations and other buried objects is also important. These potential hazards should not be seen in isolation; they can compound and add to one another. A comprehensive assessment of a site should avoid the situation in which action designed to remedy one problem either adds to another or undermines remedial action aimed at some other hazard. The presence of gaseous pollutants can present particular difficulties. According to the code of practice Investigation of potentially contaminated sites, BS 10175:2001 (B SI, 2001), the objective of the investigation is to gather the information required to form a conceptual model which will facilitate an assessment of the presence and significance of contamination. McEntee (1991) has reviewed site investigation for recycled land and Smith (1991) has described the analysis and interpretation of

66

Chapter 7 Investigation and monitoring

data. The significance of chemical contamination should be investigated with respect to the following:  health or environmental hazards  chemical attack on foundations, pipework, etc  volume changes in the fill due to chemical reactions (section 4.6)  presence of combustible materials (Palmer, 1979).

7.2 Site rec onnaissance

Published guidance on methods of chemical analysis is limited; BS 1377-3:1990 gives test specifications for a number of chemical tests, but it is not specifically related to problems of contaminated ground. At all the stages of the site investigation, the safety of personnel carrying out the investigation should be of paramount importance. This will involve not only precautions against physical hazards such as instability of excavations into which personnel enter, but also precautions against chemical hazards posed by possible contamination of the site.

7.1 Historical review Whereas with natural soils the geological processes which created them have to be inferred from observations of the soils themselves, with man- made fills there may be direct evidence of the manner in which they were formed and the date of their srcin. All available historical evidence should be collected together in a preliminary desk study to determine the history of the use of the site. Inreviewed. particular,Inthe placement ofbe fill should thoroughly practice, it may difficultbeto ensure that the historical record of the fill which has been discovered is, in fact, complete. Digest 318 (BRE, 1987a) gives guidance on preliminary desk studies in site investigations for low-rise buildings. TRRL report LR 403 by Dumbleton and West (1976) will also be found to be helpful. Information may be obtained from the following sources:  previous owners and occupants of the site  local, regional and national authorities, including local planning and environmental health de partments, waste disposal authorities, Environment Agency  libraries. The historical evidence could include the following:  oral testimony  business records, works plans  old plans and maps, including current and historical Ordnance Survey, Geological Survey  air photographs. Sources of maps and aerial photographs have been listed by Thomas (1991). From this review of historical evidence it should be possible to make an initial assessment of the nature, extent and variability of a fill deposit. This will be of considerable assistanc e in planning the subsequent stages of the investigation.

Inspection of the site by a preliminary visit can provide information which supplements the historical review. Digest 348 (BRE, 1989b) gives guidance for a walk-over survey. Attention should be paid to surface topography, site layout, and poor vegetation growth which may indicate the presence of fill and contaminants. With the information that previously has been obtained from the historical review, the site reconnaissance should have the following objectives:  to identify physical and chemical hazards so that appropriate safety precautions can be taken to improve the definition of the extent of the filled area  to find any evidence of chemical contaminat ion  to obtain samples of surface soils  to make a photographic record  to identify locations for trial pits and boreholes. 

The site reconnaissance should be an important preliminary stage in the investigation. Although it is usually limited to surface features, it may be possible to do some investigation at shallow depths using simple portable hand-operated boring and sampling equipment (West, 1991). For example, the penetration resistance of the Mackintosh probe may give a crude indication of the density index of a coarse fill or the undrained shear strength of a clay fill. Safety aspects of the reconnaissance should not be overlooked, chemical contamination, presented byincluding surface features and, where some hazards exploratory sub-surface investigation is carried out at this stage, underground hazards.

7.3 Ground inve stigation There is a considerable choice of methods of ground investigation, and it is necessary to select a combination of techniques which are appropriate to the particular filled site under investigation. A number of possible facets of the geotechnical investigation are dealt with in subsequent sections covering:  laboratory tests  in-situ tests  load tests  geophysical tests It is again emphasised that adequate precautions must be taken to ensure the safety of personnel carrying out the ground investigation. Possible hazards include chemical contamination, gas emission and unstable excavations. The possibility that a site may be contaminated should be addressed at the outset. On contaminated sites special precautions may be necessary. Ground investigation should generally include trial pits and boreholes:  trial pits are particularly useful, as they permit the examination of large quantities of fill and can also be used to establish the boundaries of the filled areas

7.5 In-situ tests 

boreholes make it possible to examine the fill at greater depths, to determine the depth of deep fills and to investigate the underlying natural strata.

The historical review and the site reconnaissance usually form the basis for the design of the ground investigation including the location and number of boreholes, trial pits and any in-situ tests. The proposed layout of structures will also influence the location of boreholes and trial pits. Shallow trial pits are usually dug using a hydraulic back-hoe excavator . Where personnel are to enter pits it is essential that the sides of the pit are safe against sudden

67

methods (section 7 .5) may be used or, in some cases, it may be satisfactory to measure compressibility and shear strength on recompacted samples. Parts 1 to 8 of BS 1377:1990 provide detailed specifications for many laboratory soil tests and Head (1998) has given guidance on most commonly used test methods. Requirements for the execution, interpretation and use of geotechnical laboratory tests are provided in Eurocode 7: Geotechnical design — Part 2: Design assisted by laboratory testing (BS I, 2000a). Guidance on t he selection of laboratory tests and practical considerations involving the procurement, administration and scheduling of tests

collapse. Safety is discussed in BS 5930:1999 (BS I, 1999), BS 6031:1981 (BSI, 1981) and BS 10175:2001 (BSI, 2001). The inspection of trial pits should identify materials which could have an adverse impact on future development. Buried objects could affect the applicability of some ground treatment techniques. It is important that trial pits are backfilled in layers with adequate compaction to ensure that the investigation itself does not cause subsequent settlement problems. Light cable percussion boring, commonly termed shell and auger, can be used on many filled sites. Rotary drilling may be required in rockfill-type materials. It is important that boreholes are properly grouted up; failure to do so could cause pollution of an aquifer. Samples for laboratory testing can be obtained from either trial pits or boreholes. Appropriate sampling techniques are required and care is neede d in transport

has been provided by the Association of Geotechnical and Geoe nvironmental Specialists (19 98c). In some fills it may be important to determine the organic matter content. A procedure based on Wa lkley and Black’s method, using dichromate oxidation, is described in BS 1377-3:1990 which also includes a specification for the loss-on-ignition test. It warns that although for some soils (for example sandy soils with little or no clay, chalky material, peats and organic clays with less than 10% organic matter) the mass loss on ignition is related to the organic content, in other soils factors unrelated to organic matter could be responsible for the major proportion of the mass loss on ignition.

and storage of samples. Some field vane tests such as the standard penetration test and field can be carried out in boreholes.

undisturbed samples for laboratory testing, field be of assistance in determining the strength and tests may compressibility of the fill. They can also be a soil profiling tool. In-situ testing te chniques include:  standard penetration test (SPT)  cone penetration test (CPT)  dynamic probing (DP)  flat dilatometer (DMT)  pressuremeter (PMT) including the Menard pressuremeter (MPM) and the self-boring pressuremeter (SBP)  field vane

7.4 Laboratory tests Samples of fill obtained during the ground investigation should be brought back to the laboratory for testing. It will be difficult to ensure that the samples adequately represent a heterogeneous fill. Disturbed samples obtained from trial pits or boreholes can be used to determine index and classification properties including density, moisture content, liquid and plastic limits, particle size distribution, and particle density. The quantiti es of fill required for the different tests are de scribed in B S 1377-1:1990 and BS 5930:1999. All samples will undergo some disturbance during sampling, but the term ‘undisturbed’ is used for samples which have been obtained with a minimum of disturbance. The quality of such samples will vary depending on the sampling techniques adopted and the fill being sampled. Samples are often obtained by pushing or driving a tube into the bottom of a borehole. Block samples may be obtained in a trial pit. Undisturbed samples may be tested to de termine compressibilit y (section 3.4) and strength (section 3.5) of the fill. In many fills it may be difficult to obtain an adequate quality of sample, and the fill may be so variable that such samples are not representative. In-situ testing

7.5 In-situ tests Where it is difficult or impractical to obtain small

Tests should be performed in accordance with national or international standards whenever possible. BS 13779:1990 specifies test methods for SPT, CPT, DP, field vane and plate loadi ng test. BS 5930:1999 also contains important information on field tests and, in particular, covers permeability testing. For a number of commonly used field tests, Eurocode 7: Geotechnical design — Part 3: Design assisted by field testing (DD E NV 1997-3:2000) provides requirements for equipment and test procedures, and reporting, presentation and interpretation of test results. Generally, these in-situ tests do not directly measure parameters that can be used in settlement or bearing capacity calculations, and many correlations between the various tests and geotechnical parameters have been proposed. Nixon (1982) and Stroud (1988) have reviewed the use of the SP T. Methods and interpretation for the

68

Chapter 7 Investigation and monitoring

CPT are described by Meigh (1987) and Lunne et al (1997), and for pressuremeter testing by Mair and Wood (1987). While many of these in-situ tests have a major application in those fills which are similar to natural soils, they may not be particularly helpful or even practicable in many miscellaneous waste fills:  the fill may be so heterogene ous that the test results have little meaning  the testing equipment may not be sufficiently robust and may be damaged.



In-situ testing has been used to assess liquefaction potential in sandy soils (sections 3.6 and 4.7). The SPT N value, suitably corrected to standard conditions, has been widely used (Seed et al , 1983). Efforts have been made to establish a reliable relationsh ip between CPT and S PT data (Robertson and Campanella, 1985 ; Seed and de Alba, 1986; Lunne et al , 1997). A fundamental approach relating liquefaction potential to the angle of dilation measured in self-boring pressuremeter tests has been suggested by Vaid et al (1981). Schneider et al (1999) have summarised the observed boundaries for liquefaction in terms of SPT, CPT, VS and DMT results. Where collapse compression (Chapter 5) is a possible problem, a field inundation test at shallow depth using a trench or borehole may be of assistance in assessing the collapse potential. Some examples are given in case histories 2 and 5. A shallow trench may be dug through

if any, information about the long-term settlement characteristics of the fill under working load . The longterm settlement is usually of much greater significance to the satisfactory functioning of the structures built on the site than the movements that occur during construction (section 4.2). Generally, the principal aim is to estimate the long-term settlement under working load of a foundation on a shallow fill and any poor underlying ground. The type of test that is required is one in which the load can be kept constant over a comparatively long period, and this can be done most simply by the direct application of dead weight. An appropriate specification for this is given in BS 1377-9:1990 Determination of the settlement characteristics of soil for lightly loaded foundations by the shallow pad maintained load test . In specifying such a test the following parameters have to be decided:

the test result will be controlled by the stiffness of only the surface layer of fill  settlement may occur at depth within the fill owing to causes other than structural load, and the load tests will give no indication of this. Where plate loading tests are carried out to de termine bearing capacity, the loading may be applied to a plate by jacking against the reaction provided by a heavy tracked vehicle such as a crane or bulldozer. The immediate settlement can be measured as the load is increased. Such a test is completed in a matter of hours and provides little,



the surface crust of themovement fill, adequately supported and then filled with water. The of adjacent se ttlement stations is monitored by precise levelling. The rate of fall of the water level in the trench is also recorded. Collapse compression may continue for some time after completion of the test as the water flows slowly away through the fill. In a clay fill, both the pe rmeability and susceptibility to collapse compression are closely related to percentage air voids and, as a consequence, the susceptibility to collapse compression can be investigated by a field permeability test. A high permeability in a clay fill points to high air voids and the potential for collapse compression on inundation. It should be recognised that there are practical difficulties in determining collapse potential (Charles and Watts, 1996).

7.6 Load tests Direct measurements of settlement characteristics in field loading tests can form an important part of ground investigations. Lightweight structures on strip footings typically stress the ground only to depths of 1.5 m to 2.5 m (section 4.2 and Appendix A). Consequ ently, it is relatively easy to reproduce the actual stress level and stress distributi on with depth in a full-scale load test which is simple, cheap and provides direct evidence about the settlement of the foundations. The limitations of such tests should be recognised:  a few load tests may not be representative of a heterogeneous fill

the pressureload to apply (this will depend on the foundation and widths) the depth to which pressure should be extended (this will depend on the nature of the ground and the proposed foundation widths)  the length of time the test should be maintained ( the results will have to be extrapolated to predict long term foundation settlement)  the number of tests to conduct on a particular site (this will depend on the variability, composition, density and thickness of the fill and on the underlying material). 

The use of a sand-filled rubbish skip as kentledge provides a simple form of test (Charles and Driscoll, 1981; McEntee, 1991). Settlement measurements can be made by precise levelling from benchmarks established away from the test area on ground which is not expected to move significantly during the period of the test. The actual period over which the test is carried out is inevitably a compromise between the theoretically desirable requirement of a period comparable with the life of the structure and the practical requirement of early development of the site. A month would seem to be a minimum for the test and it would be highly desirable for tests to be carried out over periods of 3 to 6 months, whenever possible. Settlement can be plotted against the logarithm of time since the load was applied and extrapolated to predict settlement during the lifetime of the structure. The test results can be used to calculate the

7.7 Geophysical tests following parameters:  constrained modulus D = [(qb)i] fs fn fd (Appendix B); examples are given in Table 7 (page 40)  logarithmic creep settlement rate parameter sα; examples are given in Table 8 (page 42). Case histories 9, 10, 14, 22, 25, 26 and 27 include this type of test. Tests have been carried out by placing the skip on a thin bedding layer of sand . This can induce some small bedding error, and it is preferable to place the skip on a concrete pad cast onto the fill. The concrete pad can be

69

Much larger-scale loading tests have been described by McEntee (1991). Large tests will be expensive.

7.7 Geophysical tests Geophysical tests may find applications in the investigation of filled sites. Although methods such as ground-probing radar and electrical resistivity are sometimes used, in many situations seismic methods are the most useful geophysical techniques. Shear wave velocity and damping are of particular interest, and their significance has been examined in section 3.6. BRE has

cast to the desired dimensions whereas if the fill is loaded directly by the skip the loaded area has the dimensions of the base of the skip (typically about 1.7 m × 1.7 m). Where there are definite areas with different types and depths of fill, or soft underlying material, tests need to be carried out on each area. Even where this is not the case, a shallow fill on an inner city site is likely to be quite variable and three tests would be a minimum requirement to give some indication of the likely variability in settlement behaviour across the site. The greater the depth of ground to be tested the larger must be the loaded area. This may put a practical limit on the pressure that can be applied conveniently. The maximum loading from a typical modern, two- storey, semi-detached house of 85 m 2 area will be at the party wall, which will apply a load of about 50 kN/m run to the foundations. A typical foundation width of 0.5 m would,

made measurements of surface waves and carried out refraction surveys at selected fill sites. In refraction surveys the time is measured for waves to pass from the source to a number of geophones placed on the surface of the fill at different distances from the source. The analysis of refraction measurements requires the assumption of an increasing velocity with depth and a relatively consistent soil profile along the survey line. The seismic refraction technique is commonly used to delineate the depth to bedrock. The importance of the measurement of Rayleigh wave velocity in fills has been explained in section 3.6. There are two forms of surface wave source in use:  impact sources such as a hammer or drop weight produce a transient pulse  vibrators produce continuous waves.

therefore, apply stress 100 kPa the ground. However, on theatype ofof filled site to under consideration, likely solutions to the foundation problems could involve reducing bearing pressures either by using wider, reinforced concrete strip footings or some semi-raft design incorporating reinforced edge beams with an integral reinforced slab. The effective width of the foundation transmitt ing the wall load to the fill might be between 0.5 m and 1 m in the former case and somewhat greater in the latter situation. A practical form of load test can be achieved by applying to a square test area the pressure applied by the foundation. As the vertical stress decreases more rapidly with depth under a square footing than under a strip footing of equal width, the sides of the square test load area should be of a greater dimension than the proposed strip footing width. Table 12 suggests dimensions for square test areas appropriate for footings of various widths. The table is based on elastic theory with the square test area dimensions calculated to give reasonable agreement between vertical stress under the centre of the test load and under the strip footing, down to a depth of 1.5 m.

Where anusing impact is used, theofdata are usually analysed thesource spectral analysis surface waves (SASW) method. Vibrator sour ces have been widely used with the continuous surface wave (CSW) system. The SASW method is a non- intrusive seismic test for determining wave velocity profiles ( Matthews et al , 1996). The method uses a hammer or other type of impact as an energy source. The test is based on the principle that the depth of the soil profile sampled by surface waves varies with frequency and hence wavelength. In most soils the Rayleigh waves travel at a depth of between one- half and one-third wavelength be low ground surface. Thus Rayleigh waves of different wavelengths propagate at different depths and if the stiffness of the soil varies with depth, surface waves of different wavelengths will propagate at different velocities. The method uses the spectral analysis of the propagating Rayleigh wave to determine the frequency wavelength dispersion. The SASW method has become an important method for evaluating the variation with depth of stiffness moduli at small strains. It has also been used:  to detect underground obstacles and cavities (Ganji et al, 1997)  to evaluate liquefaction potential (Andrus et al , 1998)  to evaluate the effectiveness of dynamic compaction (Kim and Kim, 1997)  to evaluate the effectiveness of vacuum consolidation (Haegeman and Van Impe, 1998)  to evaluate the effectiveness of vibro treatment methods.

Table 12 Appropriate test load areas for strip footings St ri p f oo ti ng widt h ( m) Squa re te st area (m ) 0.5

1.0 × 1.0

1.0

1.5 × 1.5

1.5

2.0 × 2.0

70

Chapter 7 Investigation and monitoring

The continuous surface wave system is also used on the ground surface and makes use of Rayleigh waves. The wave source is a frequency controlled vibrator which makes it possible to derive the relationship between Rayleigh wave frequency and wavelength and hence calculate velocity as a function of depth (Abbiss, 1981). The method has been used at an old chalk quarry at Swanscombe, which had been infilled with Thanet sand, to monitor the improvement in stiffness resulting from the installation of vibro stone columns. Butcher and McElmeel (1993) described the use of C SW to determine the depth and degree of effectiveness of rapid impact

protection such as concrete manhole rings. Knipe (1979) found that 1.2 m long × 20 mm diameter steel rods driven into the ground so that less than 100 mm protruded, recorded almost identical settlement to that measured by more sophisticated stations formed from steel rods concreted into the base of shallow holes. Much more important are the establishment of stable reference stations and the use of sufficiently accurate surv eying equipment. Although filled sites can be subject to large settlements, accurate measurements are required to establish reliably the rate of settlement over a re latively

compaction in the treatment of a loose building waste fill. The CSW method requires a relatively costly energy source but has the advantage of good frequency resolution. Matthews et al (1996) have discussed the relative merits of the two surface wave methods. Geophysical testing has some major advantages:  it is non-de structive  some tests can be carried out from ground surface  it can measure representative values of parameters in a heterogeneous fill.

short period. Measurements should be made to an accuracy of at least 1 mm. It can be helpful to locate levelling stations in a number of traverses with the stations at quite close intervals. In this way an indication of the likely differential settlement over the area of a building can be obtained. Settlement of the ground surface should be compared with rainfall in the area to see whether there is any correlation. Settlement at different depths within a fill can be measured by installing a magnet extensometer (Marsland and Quarterman, 1974). A borehole 0.15 m or 0.2 m in diameter is drilled through the full depth of fill into the underlying natural stratum. At selected depths ring magnets are anchored to the sides of the bore hole by strong springs, to act as markers. The position of each magnet is detected by a reed switch, which is lowered down a central access tube. The probe is attached to a

It also has limitations:  it is necessary to carry out other forms of ground exploration to enable geophysical measurements to be interpreted with confidence  the measured dynamic properties have to be related to the static properties of interest for deformation prediction.

In assessing the condition of a fill it is often helpful to carry out some monitoring which can include:  surface levelling stations to measure the settlement of the fill surface  magnet extensometers to measure the settlement of incremental depths of fill  standpipe piezometers to measure ground-water levels within the fill  load tests for direct estimation of settlement produced by the load.

steel measuring tape,The which incorporates electrical wiring for the probe. rigid access tubethe should be isolated from the fill, which may undergo substantial compression, by a helically reinforced outer tube. After installation the space between the outer tubing and the borehole wall is usually filled with dry sand. This type of instrumentation (shown in Figure 36) has been widely used by BRE and many of the case histories described in Part IV are based on measurements made with this system. It is also important to measure water levels within a fill. Simple standpipe piezometers can be sealed into boreholes. The water level can be located and measured using an electric dipmeter. This will prove satisfactory in

Such measurements can form one of the most important parts of an investigation of a filled site. Surface settlement can be monitored by precise optical levelling. The settlement of the levelling stations comprises the vertical compression of the fill, in which the station is installed, together with any movement of the underlying natural ground. Surface levelling stations are easy to install and very e ffective. The station s need not be elaborate, simple steel rods set in concrete may be quite adequate. The levelling stations should be sufficiently robust to resist damage due to construction traffic and a round-headed bolt cast into a 1 metre concrete cube set 300 mm into the fill has be en found effective. Where the stations protrude above the ground surface, it is usually necessary to provide substantial

Figure 36 Monitoring settlement using a magnet extensometer

7.8 Monitoring

7.8 Monitoring many types of fill, but in low-pe rmeability fills the response time of the standpipe piezometer to a change in piezometric pressure may be long. Pneumatic piezometers can be installed in boreholes and will give a much more rapid response. It has to be dete rmined how long monitoring should be continued. It is clear that to establish reliably trends in settlement and ground-water levels a minimum period of at least a year will be required, and several years would be preferable. Where a site is being investigated for building development, it will rarely be possible to monitor over such long periods. However, even where the period is quite short, it may be very advantageous to perform some monitoring. Post-construction settlement of buildings on fill should be monitored where possible as shown in Figure 37. Digests 343 and 344 (B RE, 1989a, 1995b) provide guidance on measuring movements of low-rise buildings. In biodegradable fill, gas monitoring may be required. Guidance on the measurement of gas emissions has been given in BRE and CI RIA reports (Crowhurst, 198 7; Crowhurst and Manchester, 1993), Waste Management Paper No 27 (Department of the Environment, 1991b) and in a report of the Institute of Wastes Management (1990). CIRIA Report 151 (Harries et al , 1995) provides guidance on interpreting measurements of gas in the ground. Figure 37 Monitoring settlement of house

71

72

Chapter 8 Treatment of fills In most civil engineering works worthy of the name, the unexpected happens. To be prepared for such eventualities, and to forestall their effects, is the test of good constructional practice. (Harding, 1946)

With non-engineere d fills the unexpected is only too likely to occur. Ground treatment before construction can mitigate or forestall the effects of the unexpected. Together with adequate site investigation, appropriate ground treatment can greatly reduce the risks to a construction project from difficult or variable ground conditions. In this chapter some common types of ground treatment are described. The subject is more comprehensively covered in textbooks such as Moseley (1993), Van Impe (1989) and Xanthakos et al (1994). Where a site investigation has indicated that significant differential movements may occur over the area of a proposed building, improving the load-carrying



excavation and recompaction in thin layers as an engineered fill (Chapter 9)  pre-inundation (section 8.4)

characteristics of the fill bybefore the usedevelopment of an appropriate ground treatment method, of the site, may be an attractive solution. Ground treatment may reduce subsequent movement of the fill, particularly differential movement. It is unlikely to e liminate movement and, therefore, may not be an adequate solution for buildings that are particularly sensitive to small settlements. A thorough site investigation is essential for successful ground treatment. There has been growing interest in ground tre atment techniques in the UK as scarcity of good building land has increasingly led to construction on fill hitherto considered unsuitable for development. In selecting an appropriate ground treatment technique, the problem with the fill must first be correctly diagnosed. Generally, this is associated with the loose, uncompacted nature of the fill, and most improvement te chniques are methods of increasing the density. Densification should make the fill less compressible, and consequently movements in the fill subsequent to construction will be reduced. Treatment should reduce the he terogeneity of the fill and make its subsequent performance more predictable, in effect converting a non-engineered fill into something closer to an engineered fill. Ground treatment techniques described subsequently include the following:  dynamic compaction (section 8.1)  vibro techniques (section 8.2)  preloading (section 8.3)

effects. Theon treatment mayinterparticle destroy strength which was dependent ageing and bonds in the fill. The stiffening effect of a desiccated surface layer may be lost. Appropriate methods of ground treatment and foundation construction should be employed which fully take into account the conditions on the site, for example:  dynamic compaction might not be appropriate at a site where liquid or gaseous pollution is a major problem, because it would be squeezed out like water out of a sponge; similar constraints could apply to surcharging  vibro techniques are not usually applicable when putrescible matter is present, be cause these techniques rely upon lateral constraint from the ground, which will reduce as the organic matter decays.

The first two ground treatment methods listed above, dynamic compaction and vibro techniques, are carried out by specialist contractors; the next two methods involve major earth-moving operations, either temporarily preloading with a surcharge of fill or excavating the fill and then refilling in thin layers with adequate compaction. The fifth method, pre-inundation, has limited applicability. While these treatment methods may be neficially increase the density of the fill, there may also be adverse

Other types of problem e ncountered with some fills may require different forms of ground treatment, for example:  low effective stresses within hydraulic fill with associated looseness or softness of the deposit  chemical contamination  presence of old mine workings, where shafts must be located and properly capped, and underground workings must be traced and filled as nece ssary. All site investigation data and foundation levels should be related to a fixed datum. Full account shou ld be taken of any changes in ground level and foundation level. The addition or removal of ground after treatment has be en completed is of particular concern; the weight of e xtra fill

8.1 Dynamic compaction may cause significant further settlement. The depth and location of excavations for sewers, drains and other services should be available when the design is prepared.

8.1 Dynamic compaction The repeated dropping of a weight onto the ground surface is one of the simplest and most basic methods of compacting loose, partially saturated fill. Although the method has been used from earliest times, only with the development of modern civil engineering plant has it become practical on a large scale. BR E research has addressed two basic factors:  the depth to which fill can be effectively compacted  the magnitude and rate of movements which will occur in the long term subsequent to ground treatment. Principle

Deep compaction of the fill is effected by repeated impacts of a heavy weight onto the ground surface. Th e Menard technique of ‘dynamic consolidation’ was introduced into the U K in 1973 and, initially, involved the use of a 15 tonne weight dropped from a height of 20 m. More recently, weights of from 8 to 15 tonnes have been dropped from heights of up to 15 m to achieve treatment depths of about 6 m. Findlay and Sherwood (1986) considered that problems have arisen due to inexact terminology and distinguished three different forms the method can take:

73

civil engineering for compacting shallow fills. The basic principle of high-energy impact compaction is identical to that used in dynamic compaction, but the following differences should be noted:  the energy imparted by each impact is much smaller  the impact rate is much greater  the steel tamping foot, through which energy is transmitted to the ground, remains in contact with the ground. Application

Types of fill that have been treated by dynamic compaction have included:  domestic refuse (case histories 16, 17, 18; Perelberg et al, 1986; Swain and Holt, 1986)  industrial wastes (Downie and Treharne, 1979)  demolition wastes  granular fill (Slocombe and Moseley, 1986)  dredged sand and gravel ( Pearce, 1974)  clay fill (case history 2; Thomson and Herbert, 1978)  colliery spoil (West and Slocombe, 1973; Slocombe and Moseley, 1986)  pfa (Pearce, 1974)  steelworks slag (Slocombe and Moseley, 1986) Certain limitations of the method should be recognised.  High mobilisation costs associated with the large crane required to drop the weight usually mean that areas smaller than 5000 m 2 cannot be treated economically.

  

dynamic consolidation compaction ofofgranular soils dynamic saturated fine soils dynamic replacement in which large pillars of imported granular fill are formed within the body of the soil by heavy tamping.

In the second and third methods excess pore pressures are established and require a period of hours or days to dissipate. The major use of the method in the UK has been to compact loose, partially saturated fills; hence the term ‘dynamic compaction’ has been used in this book, but the terms ‘heavy tamping’ and ‘pounding’ are also found in the literature on the subject. The mechanism of compaction is simple: repeated impacts of a heavy weight compact the fill by reducing the air voids. This is analogous to the mechanism of compaction in a standard laboratory Proctor compaction test. Despite the enormous difference in scale between field and laboratory, the energy input per unit volume of soil, E/V, is comparable; typically the compactive effort applied in the field is similar to that applied in the standard Proctor compaction test (600 kNm/m 3). Further information is given in Appendix D. Impact compaction of fills can also be carried out using the rapid impact compactor, which was srcinally developed in collaboration with the Ministry of Defence for fast compaction of backfill materials in the repair of craters in airfield runways. Following field research by BRE, it is finding increasing application in building and

The smaller equipment for rapid impact compaction means that required mobilisation costs are relatively low. As a consequence, rapid impact compaction is finding increasing application in compacting shallow fills 3 m to 4 m deep (Watts and Charles, 1993).  It would be inadvisable to use the method close to existing structures because of the potentially damaging vibrations caused by the impact of the weight onto the surface of the fill. Minimum distances of 30 m have been quoted, but much depends on individual circumstances, such as whether the building is also on the filled ground or is built on natural ground. Lukas (1980) presented a method of estimating particle velocity at a given distance from the point of impact.  Flying debris constitutes a hazard for personnel, vehicles or structures close to the impact point. Some shielding of vulnerable targets may be required. Although flying debris is not a problem for rapid impact compaction, another environmental issue, noise levels, may be.  With many fills it will be necessary to provide a granular blanket to form a working platform for the crane. The cost of this granular material may be substantial in relation to the total cost of the treatment. A granular blanket is not usually required for rapid impact compaction.  Energy may be inefficiently transferred to the fill by a large weight falling onto an irregular fill surface, particularly where it falls into a deep crater produced

74

Chapter 8 Treatment of fills

by previous impacts. The rapid impact compactor does not have this drawback because the steel tamping foot remains in contact with the ground.  In some fills compaction could cause gas generation. After a small 1940s landfill was dynamically compacted, gas concentrations increased to such an extent that preventive measures were required for the proposed buildings (Institute of Wastes Management, 1990); compaction had caused a change in water le vel, saturating previously dry waste. At Surrey Docks it was found that the rate of emission of methane into perforated plastics borehole tubes generally increased after dynamic compaction (Thomson and Aldridge, 1983). Method and equipment

Where the energy input per blow has been achieved by dropping a 15 tonne weight from heights of up to 20 m, a large crane has been required. A greater energy input per blow requires special lifting equipment and, unless a very large site required treatment, it would not be economic to use greater impact loads. Figure 38 shows the use of dynamic compaction at a site in the UK. Primary tamping has usually consisted of repeated impacts at a number of grid points 5 m to 10 m apart. The objective of this stage of the treatment process is to achieve compactio n at depth. The creation of a dense surface layer at this early stage would absorb a large amount of the subsequent compaction energy and inhibit densification at depth. A suitable small earthmoving machine is needed to infill the craters formed by tamping. A second stage involves repetition of the first stage on a grid offset from the srcinal one. The craters formed by this compaction are then filled. These initial stages involve high-energy impacts. In later stages a more uniform tamping is carried out over the whole area, with a reduced drop height. This should compact the nearsurface fill in as uniform a manner as possible. Total energy input has typically been in the region of 2000 kNm/m2. A granular blanket is usually required to form a working platform for the crane. The equipment for rapid impact compaction drops a 7 tonne weight 1.2 m onto a 1 .5 m diameter steel tamping foot as shown in Figure 39. Although the energy per blow is not large, the equipment permits a large number of impacts to be applied rapidly at a rate of about 40 blows per minute. The operator monitors the number of impacts, the total energy input and the foot penetration per blow. When any operating parameter reaches the specified value, the equipment is moved to the next tamping location. As the foot remains in contact with the fill, the energy should be much more efficiently used in compacting the fill than in dynamic compaction where the weight may fall on an irregular fill surface in such a way that much of the energy is dissipated in deforming the irregularities of the fill. Case history 26 describes a trial of this equipment on a fill composed of building wastes.

Figure 38 Dynamic compaction

Figure 39 Rapid impact compactor

8.1 Dynamic compaction Analysis of treatment

The energy required to compact a fill is a complex function of the method of compaction (mass and dimensions of weight, height of fall, spacing of grid points, number of blows per grid point, number of passes) and the soil properties (fill type, particle size distribut ion, existing density and moisture content, depth to be compacted, degree of improvement required). The assessment of the required energy is based largely on past experience and empirical correlations which do not take account of all the factors. Consequently, only a relatively crude assessment can be made of the required energy. There is some correlation between the depth of compaction and the energy per blow. This has usually been expressed as:

ze = k(MH )0.5 where ze is the depth of compacted soil in metres, M is the mass in tonnes, H is the height of fall in metres and k is a coefficient which depends on fill type. Generally it has been found that 0.3 < k < 0.8. These matters are examined in more detail in Appendix D. Testing

Some form of testing may be required to mee t some or all of the following objectives:  to investigate the properties of the fill prior to treatment — testing may be carried out as part of the siteassess investigation to the degree of improvement effected by ground treatment — testing may be carried out to give a comparison of properties before and after treatment  to assess the effectiveness of the treatment in connection with an observational approach to ground treatment or with quality management — some testing may be carried out during dynamic compaction  to confirm that a specified improvement has been achieved — testing may be carried out after treatment  to determine the load-carrying characteristics of the treated fill — field load tests may be carried out subsequent to treatment  to determine long-term movement subsequent to treatment — field monitoring may be undertaken. 

Close monitoring of the effect of impact loading during treatment can be helpful in identifying soft areas where further treatment is required. Briaud et al (1989) have proposed a test in which, after it has been dropped by the crane, the dynamic compaction weight is hit with an instrumented sledge hammer while it rests on the fill. Geophones on the weight record the response and give a measure of the stiffness of the fill. Rapid impact compaction can be closely monitored during treatment using in-cab monitors. At each impact point, the behaviour of the ground can be assessed and soft or hard spots can be quickly identified. Watts and Charles (1993) have described a case history where rapid impact compaction identified a linear feature in filled

75

ground which was later shown to be soft cohesive fill within ash deposits. In-situ testing will usually be preferable to laboratory testing. In the selection of an appropriate type of testing the following should be considered:  the nature of the fill that is being treated  the type of development which will take place on the treated ground. In fills composed of natural sands or clays in- situ testing techniques such as cone penetration test (CPT), standard penetration test (SP T), dynamic probing (DP), pressuremeter (PMT) and flat dilatometer (DMT) may be appropriate. However, in fills composed of variable waste products these techniques may not be feasible. Geophysical methods have been used to obtain a measurement of the improvement in soil properties resulting from ground treatment (Butcher and McElmeel, 1993). Rayleigh wave velocity measurements have been used to assess the effect of rapid impact compaction (Watts and Charles, 1993) . Holeyman and Vanneste (1987) have described the use of vibration monitoring to assess the effect of treatment on adjacent buildings. An accelerometer on the tamper has been used to monitor the compaction process. With rapid impact compaction, measurement of the depth of imprint and surface levels are often considered to be sufficient, but loading tests are sometimes undertaken. Several in-situ tests are also used. Performance

Five case histories in Part IV refer to dynamic compaction and some basic data are summarised in Table 13. Case history 26 describes a trial of rapid impact compaction on a fill composed of building wastes. Three factors are of particular significance in assessing field performance:  depth of compaction induced by ground treatment  stiffness of the treated fill  post-treatment settlement of the fill. Induced compaction It is helpful to compare the energy which has been used, and the compaction which has been induced, on dynamic compaction sites. The energy input at the four case histories of heavy dynamic compaction described in Part IV (2, 16, 17 and 18) varied from 2200 kNm/m 2 to 2800 kNm/m 2. A stiff clay fill was compressed by 4% and old refuse by about 10%. There are a number of published reports on the use of dynamic compaction on refuse sites (for e xample, Downie and Treharne, 1979; Perelberg et al , 1986; Swain and Holt, 1986). Welsh (1983) described the use of dynamic compaction on a recently placed sanitary landfill in the USA. The treatment of this recent refuse produced compressions of up to 2.5 m or 25% of the srcinal depth of fill. Dynamic compactio n has been widely used on coarse fill, sometimes with a high water level in the fill. Some instructive records, mostly from outside the UK are

76

Chapter 8 Treatment of fills

Table 13 Dynamic compaction case histories Induced Case h i s t o r y F i l lt y p e

M ( t on ne)

H ( m)

A 2 (m )

E/At ( k N m / m2)

n

surface

Fill

settlement ( m)

depth ze (m) (m)

2

Stiff clay

15

20

4

2800

3.7

0.24

16

Old refuse

15

20

4

2600

3.5

0.5

17

Old refuse

14

14

4

2600

5.3

0.58

6.5

5.5

18

Old refuse

15

20

4

2200

2.9

0.5

8

—

26

Building waste

7

1

1.8

1500

0.3

6.5

4

39

24

6

6

—

Notes M H

Mass of weight Height of fall

A

Area in contact with fill on impact

n

Number of impacts at any point; if treatment comprises a total ofN impacts uniformly distributed over a total area At, then n = (NA)/At

E/At Average energy input per unit area; E/At = MgHn/A ze

Depth of compacted fill

briefly summarised here:  At Changi airport, Singapore, an 8 m depth of hydraulically placed coarse sand fill was compacted using an average applied energy of 1230 kNm/m 2 (Choa et al, 1979). Weights of up to 40 tonne were dropped from heights of up to 30 m. The improvement resulting from dynamic compaction has been described by Ramaswamy and Yong (1983) in terms of SPT, CPTincreases and PMT different tests indicated inresults. densityThe index, ID, varying between 0.15 and 0.4 or more. Typically, the tests suggested that ID increased from 0.55 to 0.85, with the SPT indicating the greatest increase.  At Nice airport, France, weights of up to 170 tonne dropped from heights of up to 22 m were used to compact down to 40 m (Gambin, 1983). The upper 9 m consisted of sand and gravel fill which had be en dumped into water. An energy input of 400 kNm/m 2 produced 2% compression; 1200 kNm/m 2 produced 6% compression and 4120 kNm/m2 produced 9% compression. On this site the energy per drop was extremely large (36 700 kNm).  At Vancouver Island, Canada, dumped rockfill up to 15 m deep was dynamically compacted using a 20 tonne weight dropped from 30 m (Wightman and Beaton, 1984). Th e treated rockfill was to be the foundation for a building. Tota l energy per unit area was 5600 kNm/m2 for 15 m depth of fill, 3800 kNm/m2 for 9 m depth of fill. The compactive effort was about 400 kNm/m3. The induced settlements were 0.92 m (6%) and 0.58 m (6.5%) re spectively.  At Uddevala shipyard, Sweden, compaction of 30 m deep dumped rockfill was effected with a 40 tonne weight dropped from a maximum height of 40 m (Hansbo, 1977). The total impact energy per unit volume of fill was 200 kNm/m 3. Compression of the rockfill varied between 4% and 12% of the initial depth of rockfill.



At Rubislaw quarry, Aberdeen, the quarry waste fill was composed of sand to boulder-size fragments of granite (Slocombe and Moseley, 1986). The fill was between 3 m and 22 m dee p with an average depth of 8.5 m. The applied energy of 5500 kNm/m 2 induced 650 mm settlement. A high level of treatment was required as four-storey offices were to be built on the site. The quarry waste was probably closer to a rockfill than to a sand.

While the information from case histories helps to set broad limits for the likely required energy, there are too many differences between the sites in terms of soil types and depths, and energies per blow, to draw firm conclusions. Stiffness of the treated fill Case history 16 describes the dynamic compaction of old domestic refuse with an average energy input of 2600 kNm/m2. When a 3 m high embankment was built over the treated fill, the immediate compression of the fill suggested that the constrained modulus D = 20 MPa. Similar loading of some adjacent untreated refuse indicated that this fill was more than three times as compressible as the treated fill; treatment had increased D by a factor of 3. Some published results of load tests on treated ground are summarised in Table 14. A load test on untreated fill is given for comparison. Post-treatment settlement A major uncertainty in developing fill sites is associated with settlement subsequent to ground treatment. There is no simple way in which long-term settlement behaviour can be reliably related to the amount of energy used during dynamic compaction. At best, relatively crude estimates can be made. Dynamic compaction should increase the constrained

8.1 Dynamic compaction Table 14 Load tests on ground treated by dynamic compaction E /At q b Sit e F i ltly pe ( k N m / m2) ( k Pa ) ( m) Aberdeen*

Quarry waste

5400

215

a /b

3.0

1

Gartsherrie*

Sandyclay/ash

2000

100

2.5

1

Briton Ferry*

Silty sand

2100

100

2.0

1

1600

( M Pa )

7.7

52

1.7

118

Glasgow†

Demolitionwastes

Glasgow†

Demolition wastes§

0

125

2.4

1

7.3

41

Allington‡

Clay fill

1200

300

2.0

1

28–34

18–21

Allington‡

Granularfill

1200

300

2.0

1

9–20

30–67

2.4

3

84

4.8

Sand/ash/clay

125

1.0

qb/si

Glasgow*

1200

150

si (mm)

1

77

11.0 3.2

14 94

* Slocombe and Moseley (1986) † Reid and Buchanan (1986) ‡ Slocombe (1993) § Untreated fill

modulus of the fill and so reduce settlement under any subsequent loading condition. However, creep settlement under self-weight may initially be increased as the compaction process in effect creates a ne w fill. The logarithmic creep compression rate parameter α should be smaller after treatment of the fill than it was before, but there is effectively a new zero time, corresponding to the date of dynamic compaction rather than the date the fill was placed. On the clay fill at Corby ( case history 2) the settlement of houses built on fill treated by dynamic compaction was similar to that of dynamically compacted ground that was not loaded.by This thatofmovement has been controlled theindicates self-weight the fill rather than the weight of the buildings. Over a period of 25 years, settlements typically increased to about 50 mm. If it is assumed that all the settlement occurs in the upper 6 m of fill which were effectively compacted, the settlement corresponds to α = 0.5%, with zero time reckoned as the date of treatment. It is e stimated that prior to treatment α = 1.5% in this upper 6 m of fill with zero time reckoned as the date of fill placement. Although the movements have not caused significant damage to the buildings, they are considerably greater than those experienced on an adjacent area temporarily preloaded with a 9 m high surcharge of fill. On old refuse fills dynamic compaction would seem more appropriate than preloading with a surcharge, because the high impact load should crush any buried containers. At Redditch the immediate compressibilit y of old domestic refuse was reduced to one-third of its srcinal value by treatment (case history 16). However, treatment did not eliminate subsequent creep settlement. The effectiveness of dynamic compaction will largely depend on the age of the refuse (older refuse having had longer for decomposition to take place and also having been a better material with a high ash content when placed). In considering the use of dynamic compaction on a particular site, much depends on the type of development and the capacity of structures to survive some differential settlement.

Design and specification

A major concern with the use of dynamic compaction has been the difficulty of preparing an appropriate specification. The specification may define:  a method of compaction, which could be in terms of the mass of the weight, height of fall, grid spacing, number of blows at each grid point, number of passes  an end product or end result to be achieved by the compaction, which could be in terms of an improvement in some fill property measured in situ (for example SPT, CPT, DP, DMT, PMT)  an aspect of the performance of the treated ground such as the maximum allowable post-treatment settlement. Each approach has certain advantages and certain disadvantages. With a method specification it is necessary to assess the total energy input in kNm/m 2 needed to achieve the required improvement. There are problems with an end product specification:  variability of the fill may make it difficult to characterise the fill reliably by an in-situ test value  if mean values of test results are to be used, then it has to be decided whether any values should be rejected as unrepresentative; for example a very high value might be caused by a boulder, while a low value might be caused by a pocket of silt and clay in a granular fill  test results may not correlate well with density inde x (ID) and hence the degree of compaction. While a performance specification may seem the best way of ensuring that the client obtains the required ground behaviour, post-treatment settlement may occur slowly over many years, and there may be no effective redress if settlement goes outside the specified maximum at a late stage. With rapid impact compaction, a limiting energy input is determined on site by observing the blow count above which impacts produce negligible further penetration of the foot. For fills, an input of 1500 kNm/m 2 is typical. The layout of abutting tamping points is usually based on a square or rectangular grid.

78

Chapter 8 Treatment of fills

8.2 Vibro techniques

large areas with a uniform pattern of treatment points. It is particularly suited to the redevelopment of sites. The method can be used relatively close to existing structures, much closer than dynamic compaction could be safely used. Well over half the work in the UK is in fills, with some work in soft clays. The work is carried out by specialist contractors. Typically, vibro techniqu es are used to treat areas of shallow fill overlying less compressible soil. The fills may be very variable, and a major purpose of the treatment is to create a more uniform foundation. Most, though not all, types of fill can be treated but caution should be

These deep vibratory ground treatment techniques srcinated in Germany in the early 1930s and were introduced into the UK around 1960. Initially, many of the applications were associated with civil engineering works, but subsequently these techniques have been used extensively for low- rise buildings, particularly housing and industrial units, and this is now the main use of these methods in Britain (St John et al, 1989). Vibro stone columns have been used on urban rede velopment sites and the wider acceptance of the method grew from its successful use in northern cities. The use of vibro concrete columns was developed in Germany and introduced into the U K in 1991. Although it might be considered to be a piling method rather than a form of ground treatment, it has been mentioned because of its similarity to vibro stone columns. Principle

Rotation of an eccentric weight in the depth vibrator results in vibrations in a horizontal plane being transmitted to the soil as the vibrator penetrates into the ground. The method was srcinally developed to compact loose sands. It is generally reckoned that soils can be effectively compacted only if the pe rcentage of silt size particles is relatively small. In cohesive soils the long cylindrical hole formed by the depth vibrator, which is sometimes called a vibrating poker, is backfilled with granular material to form stone columns the soil. While sand can be compacted, clay iswhich only stiffen stiffened by stone columns. The method can be used both in fills and in natural soils. Terminology can be a problem with the terms vibroflotation, vibro-compaction, vibro-replacement, vibrodisplacement and vibro stone columns all being used. In this book, the following processes are identified:  vibro-compaction describes the densification of a sand in which extra material may or may not be added  vibro stone columns is a term which describes the stiffening of a soil by stone columns  vibro concrete columns are designed to transmit structural loading to a suitable underlying bearing stratum and thus are effectively acting as piles; satisfactory performa nce is far less depe ndent on the support of the surrounding soil than is the case for vibro stone columns. Application

Vibro techniques can provide a cost-effective way of reducing settlement and creating more uniform foundation conditions. The installation process is relatively simple, quick and flexible. Treatment can be localised under footings and slabs and, therefore, is well adapted to traditional housing and light industrial units. It is not necessary to treat large areas in a uniform way, as with dynamic compaction or preloading. Projects range in size from a few treatment points be neath strip footings for a pair of semi-detached houses, to the treatment of

exercised at the investigation and design stages in certain conditions. Vibro-compaction can be used to densify granular fills. Previous experience showed that the method was effective without the addition of stone provided that the fines content was not greater than about 15%. However, Slocombe et al (2000) affirm that the development of new vibrators and modified construction techniques has enabled sands with significantly higher fines content to be treated, and a case history is described where treatment of a sand with a fines content of up to 30% increased ID from 0.45 to 0.6. Vibro stone columns can be used to treat soft clay fill. A lower limit of undrained shear strength of 15–20 kPa is often quoted, but much depends on the installation process as described in the ne xt sub-section. The amount of improvement that can be achieved in a clay fill may be limited.beTypically, settlement under a foundation might reduced to half the settlement under theload same load on untreated ground. This means that on this type of ground the method is often more appropriate to low-rise industrial buildings than to settlement-sensitive housing. Some limitations of the technique, and re strictions on its applicability, should be noted.  Access is required for crawler-mounted cranes.  If the fill contains discrete zones of organic matter, or any other substance that may decrease in volume with time due to chemical reaction or solution, these may have to be removed and re placed with granular fill. Significant volumes of such unacceptable materials render a fill unsuitable for vibro treatment. The reduction in volume of the fill would lead to a loss of lateral support for the stone columns, and consequent settlement.  If it is not considered to be economic to provide fulldepth treatment, the appropriateness of partial-depth treatment needs to be critically assessed before proceeding with the treatment. Varying depths of fill, particularly at the boundary of the filled area, can give cause for concern.  Stone columns could form paths for water to penetrate into the untreated fill at depth, which could have serious consequences in a loose, partially saturated fill susceptible to collapse compression on wetting. They could also permit the contamination of an aquifer or the upward migration of landfill gas.  Extensive buried obstructions can seriously impair a

8.2 Vibro techniques treatment programme if not identified and removed ahead of the works.  A firm or desiccated crust may need to be pre-bored to avoid impeding the pene tration of the vibrating poker.  Although the noise level is relatively low, there may still be environmental problems.  The possibility that vibrations will cause settlement of adjacent existing structures and buried services needs to be carefully considered. Limiting distances of 2 m to 5 m have been quoted.  Although treatment should reduce ground movements, it is unlikely to eliminate them. Vibro is suitable only for structures that are not unduly sensitive to very small ground movements, and treatment cannot guarantee that movements will be virtually eliminated. Heavy concentrated loads may be difficult to accommodate. Adding extensions to structures founded on vibro can present problems unless provision is made at the time of the initial ground treatment (the same problems may also occur with other alternatives such as piling). Vibro concrete columns are be ing used in soft soil ground treatment. The technique is being used not only to treat very soft and organic soils which are unsuitable for stone column application, but also some denser soils. They can be used in soils with very low values of cu. Method and equipment

The basic item of e quipment is a depth vibrator which is sometimes referred to as a vibrating poker, a vibrating probe or a vibroflot. Figure 40 shows the depth vibrator in use at a fill site. The vibrator is usually suspended from extension tubes with air or water jetting systems. The crane or base machine supports the vibrator and extension tubes. The depth vibrator has an eccentric weight assembly rotating rapidly within a heavy tubular steel casing. Vibratory motion is therefore horizontal with the vibrator cycling around the vertical axis. The nose of the depth vibrator is tapered to aid penetration in the ground, whilst vertical fins prevent the vibrator rotating during penetration. Diameters of the depth vibrator range from 300 mm to 450 mm and they are about 2 m to 3.5 m long. Their weight varies according to size and purpose, but is usually in the range of 2 tonnes to 4 tonnes. Greenwood (1991) reported that machines, srcinally 35 kW, were now powered up to about 200 kW with frequencies of usually 30 Hz or 50 Hz . More recently, Slocombe et al (2000) have reported that centrifugal forces of over 300 kN are currently in use at frequencies of 20 Hz to 30 Hz, in some cases with variable frequencies. There are flushing jets in the nose cone and sides. The depth vibrator is suspended by means of a flexible coupling to which extension pieces may be added . Treatment depth is commonly not greater than 6 m, but where soil conditions make it necessary, extension pieces can be added to the vibrator to treat greater depths. Depths of 56 m have been achieved (Degen, 1997). T he

79

depth vibrator responds differently to various ground conditions and much can be learnt about the site during treatment. It is important that full advantage is taken of the information so gained in order to modify treatment patterns. The introduction of automatic recording of instrumented depth vibrators should be beneficial as detailed records of each penetration are obtained. A loose sand can be compacted with no necessity to add any material, although in this situation the level of the ground surface will be lowered. This process, vibrocompaction, is rarely applicable in the UK. In most fills stone columns are used. The cylindrical hole which is formed by the depth vibrator is backfilled in stages with stone, each stage be ing compacted by the vibrator. There are three principal methods of installing vibro stone columns and all the processes use a similar type of depth vibrator. Th e presence of ground-water within the depth of treatment will influence the choice of process. Although a high ground-water table should not be a major problem, it can increase the possibility of problems occurring during treatment. Vibro stone columns: dry top-feed process The whole assembly is suspended from a crawlermounted crane and the vibrator is lowered onto the ground. Penetration of the fill is effected by a combination of the weight of the vibrator, the highfrequency vibration and compressed air, which emerges as high-pressure jets from nozzles in the main steel

housing just above the vibrator tip. After reaching the

Figure 40 Installation of vibrated stone columns

80

Chapter 8 Treatment of fills

required depth, the vibrator is held in the ground for a short time and then withdrawn. A small charge of clean, inert stone is tipped into the hole and the vibrator is lowered again to compact the stone and interlock it with the surrounding soils. By adding successive small charges of stone and compacting each one to chosen levels of power consumption, a dense column of stone is built up to ground level. If the dry process is employed in a clay fill, an undrained shear strength in excess of 30 kPa is required to ensure that the hole created by the vibrating poker stays open during backfilling with stone. Typically gradings for the stone are within the range from 40 mm to

vibrations and its weight, using an additional pull-down force if necessary, the vibrator penetrates the ground to the required depth. The stone column is then formed and compacted by lifting the vibrator, holding the lift for a short time to allow the stone to run, and then forcing the vibrator down on the charge of stone to compact and tightly interlock it with the surrounding soil. This is repeated, charging the system with stone as necessary, until a compact stone column is formed up to ground level. Typically gradings for the stone are within the range from 25 mm to 50 mm.

75 mm.

Top vibrator Deep vibratory compaction of sands can be achieved by methods which use either a depth vibrator or a top vibrator. Methods using a depth vibrator are similar in principle to the methods for installing vibro stone columns previously described, although a stone column may not always be formed. Where a top vibrator is used, it is connected to the top of a compaction probe which is designed to transfer the vibrations to the soil as e fficiently as possible. Several different types of compaction probe are available including the vibro-wing (Massarsch and Broms, 1983). Although the top vibrator usually vibrates vertically, the probe will cause horizontal accelerations which may locally be greater than the vertical ones. The compaction increases when resonance is created between the vibrating system and surrounding soil. Compaction is achieved by inserting the probe at

Vibro stone columns: wet process The depth vibrator, which is similar to that used for the dry process but is equipped with water jetting, is suspended from a suitable crane, lowered onto the ground and the water jets are ope ned. The vibrator penetrates quickly through weak fills under its own weight aided by the water je tting and vibrations. After reaching the required depth, the vibrator is partially withdrawn and is sometimes surged to flush out the weak soils accumulati ng in and adjacent to the hole. With loose particles washed out, the hole becomes larger than the vibrator. Following formation of an open hole the vibrator is kept in the ground and the water flow reduced whilst clean inert stone is heaped around the top of the vibrator bore at ground level. The stone then passes

down between the vibrator thecolumn surrounding soils to permit the construction of aand stone in short lifts. The vibrator compacts the stone infill and interlocks it tightly with the surrounding soil. The cycle is re peated until a compact stone column is built up to ground level. The diameter of a wet stone column is usually greater than that of a dry stone column. Typically gradings for the stone are within the range from 25 mm to 75 mm. The wet process has considerable attendant problems of water supply, drainage ditches, settlement lagoons and final disposal of the effluent in a manner acceptable to the Statutory Authorities, which may be particularly difficult on a contaminated site. Vibro stone columns: dry bottom-feed process The bottom-feed process obviates the need to support the sides of the hole with water. The stone is introduced directly to the tip of the vibrating poker through a he avyduty stone supply tube permanently attached to the depth vibrator. The stone can thus be placed and compacted at the same time that the means of side support is withdrawn. The method is a dry process which allows stone to be introduced into soft and loose soils in a coherent column, where previously, with cranesuspended vibrators, the wet process would have been required. The vibrator is positioned on the ground at the treatment location and the whole system is charged with stone. With the stone in the supply tube acting as a plug at the tip of the vibrator, assisted as necessary by compressed air and under the combined action of the

treatment points onfrom a triangular grid. Spacings areusually typically 1 m to 4 ormrectangular depending on the type and size of the compaction probe and vibrator capacity. Vibro concrete columns Vibro concrete columns are installed using a modified guided bottom-feed vibro rig. Construction is generally carried out from a compacted working platform. Concrete is pumped into the bottom of the cylindrical void formed by the vibrating poker and an enlarged base is formed. As the vibrator operates, a continuous record is kept of the depth, the energy output, the concrete pressure and the installed concrete volume. The columns are not normally reinforced. Analysis of treatment

In the design of the treatment some analysis is needed of the likely performance of the structure built on the treated ground. The following factors are important:  the type of fill and whether or not the properties of fill have been modified by treatment  whether or not the columns fully penetrate the poor fill and are themselves founded on a firm stratum  the type of foundation, and in particular the size of foundation in relation to the column length and spacing. For vibro stone columns, a useful indicator of effectiveness is the settlement reduction factor, sr (the

8.2 Vibro techniques factor by which the settlement that would have occurred under the same loading on untreated ground must be multiplied to give the settlement on the treated ground). This is often expressed as a function of the area replacement factor, Ar (the factor by which the total area, At, should be multiplied to give the area now occupied by columns, Ac). Methods of analysis for three cases of practical interest are now examined:  vibro-compaction of sand fill  fully penetrating stone columns under widespread load  stone columns under strip or pad footing. The situation with vibro concrete columns is quite different. The key design feature is the pile capacity and this is controlled by the properties of the bearing layer. Vibro-compaction of sand fill Where vibro-compaction is adopted as the treatment process for a loose sand, the sand should be compacted into a denser state. Typically a density index of up to 0.8 can be obtained. Extra material may or may not be added, but the important factor is the increase in density of the sand fill induced by treatment. Soils e xhibiting cohesion due to cementation, suction or some other cause may not be suitable for this type of ground treatment. Silt- and clay-size particles dampen vibrations and reduce the effectiveness of the treatment. Deep vibratory compaction is usually restricted to granular soils be cause

a fines content aboveSlocombe 15% will reduce the compaction efficiency. However, et al (2000) have shown that, with newly developed e quipment, sands with significantly higher fines content can be treated. In saturated soils, pore pressures will increase during treatment reducing the shear strength and facilitating compaction. The spacing of compaction points is a principal factor in predicting the e ffectiveness of compaction. Vibration energy is attenuated with distance from the source and there is a limiting radius beyond which the energy is insufficient to achieve denser particle packing. Attenuation depends on the power of the vibrator, the nature of the fill including its particle size distribution, and the position of the water table. Mitchell (1981) indicated that the zone of improved soil extends from 1.5 m to 4 m from the vibrator depending on soil type and the powe r of the depth vibrator. Several relationships betwee n spacing of compaction points and ID have been observed or proposed.  Thorburn (1975) presented an empirical relationship for clean sand between the density index at points midway between centres of vibration and probe spacing. This indicated that a spacing of 1.5 m was required for ID = 0.85 and a spacing at 2 m centres for ID = 0.7.  Johnson et al (1983) using a 75 kW hydraulically operated vibrating poker in a sand fill, reported ID = 0.7 with compaction points at 3.5 m centres.



81

BRE measurements on a natural sand site, described in case history 27, showed little improvement at a distance greater than 2 m from the centre of a compaction point.

The above variat ions could be due to differences in equipment and treatment technique, and differences in the soil conditions and initial densities, together with differences in the way ID is measured. Relationships between ID and compaction point spacing can only be a crude initial guide to the treatment required at a particular site. Field trials can be used to establish an appropriate spacing for the compaction points. Some 70 × 106 m3 of dredged sand fill was used in the reclamation works for Chek Lap Kok airport in Hong Kong (Covil et al , 1997). Below the water table, ID was in the range 0.2–0.4. Vibro-compaction was carried out to densify the sand and produce more uniform conditions. The grid spacing was related to the type of rig. Using a rig with a 120 kW, 300 mm diameter depth vibrator, light compaction was carried out on a 4.0 m triangular grid and heavy compaction on a 3.5 m grid. A minimum value for qc of 8 MPa was specified for light compaction and 15 MPa for heavy compaction. The behaviour of densified sand under load can be analysed by the methods commonly adopted to assess the settlement of foundations on sand. The effectiveness of the method is assessed on the basis of the expected improvement in properties of the sand fill. For a normally consolidated sand, increasing the densitybyindex to 0.8 might typically reduce settlement 50%.from If, in 0.5its undisturbed state, the sand is over-consolidated, treatment might lead to little if any reduction in settlement. Also, any age-related bonding between sand particles will have been destroyed by the treatment. However, the increase in density could still be of value, because the treated fill will now be in a more stable state of packing in which the vulnerability to the following hazards will be eliminated or reduced:  liquefaction of a saturated fill  collapse compression on wetting of a partially saturated fill. Fully penetratin g stone columns under widespread load The situation is relatively simple to examine from a theoretical standpoint and is considered in Appendix E . Without stone columns the compression of the fill would be one dimensional such that: εvo = q/Do

where εvo is the vertical strain induced in the untreated fill q is the applied pressure Do is the constrained modulus of the untreated fill The stone columns effectively reduce the compressibility of the fill into which they are installed. The vertical strai n in the stiffened fill is reduced to εv and the improvement

82

Chapter 8 Treatment of fills A B C D

Elastic analysis (Balaam & Booker, 1981) Plastic columns (Priebe, 1995) Large oedometer tests (Charles & Watts, 1983) Centrifuge tests (Craig & Al-Khafaji, 1997) Field data:  Munfakh et al (1983)  Raju (1997)  Watts & Serridge (2000)  Watts et al (1992)  Watts & Charles (1991)

r

s

t, n e m e tlt e S

Area ratio,Ar

Figure 41 Comparison of field results with theoretical predictions for fully penetrating granular columns under widespread load

can be expressed as a settlement reduction factor, sr, such that: εv = sr εvo

A value of sr = 1 indicates no improvement due to treatment, whereas sr =The 0.5analyses means that has been reduced by 50%. andsettlement experimental work described in Appendix E lead to the conclusion that with Ar < 0.1, improvement is unlikely to be significant. From practical considerations in the field, generally Ar < 0.3. With Ar = 0.3, sr = 0.5 might typically be achieved. Figure 41 includes some field data which confirm these conclusions. Stone columns under strip or pad footing A single column or a small group of columns may be installed under a small footing. For a column or group of columns it may be necessary to calculate the following:  load-carrying capacity  settlement at working load

The mechanism of behaviour and distribution of stress can be quite complex. Under a small strip or pad footing, the depth to which there is a significant increase in vertical stress will usually be very limited. The installation of stone columns may increase that depth considerably. Hughes and Withers (1974) examined the case where columns are installed in a soft saturated clay. Cavity expansion theory was used to de rive an expression for the load-carrying capacity of a column. Also, an e xpression was derived for critical column length at which failure in bulging and end- bearing occur simultaneously . Balaam and Poulos (1983) analysed the behaviour of a single stone column in a clay soil. Some assumptions had to be made about the behaviour at the interface be tween

the column and the surrounding soil. Priebe (1990, 1995) has derived a chart which presents the settlement of a rigid footing on a limited number of stone columns as a proportion of the settlement of an infinite raft on an infinite grid of columns. The reduction factor is a function of the ratio of column depth, z, to column diameter, b, and the number of columns under the raft, N. For example with four stone columns under the raft and z/b = 5, the reduction factor is 50%. It should be noted that this reduction factor is not the same as the settlement reduction factor sr which relates the situation with columns to the situation without columns. Priebe’s method presupposes that one can calculate the settlement of an infinite raft. This type of approach assumes uniform soil conditions, but in practice an important effect of column installation may be the transmitting of load to greater depths where soil conditions may be significantly different. Greenwood and Kirsch (1983) reported that sr = 0.5 for two 0.91 m diameter plate tests on a soft clay; one test was on clay stiffened with a 0.58 m diameter column and the other test was on clay without a column. Case history 28 describes the treatment of between 3 m and 5 m of miscellaneous fill at a site in Bacup. Ash and stone fill overlies a mixed cohesive fill and stone columns were installed at 1.8 m centres ( Ar = 0.21) through the fill to the underlying glacial till. A 9 m × 0.75 m test foundation s trip was subsequently loaded in three increments to a be aring pressure of 123 kPa. T he performance of the strip was compared with strip constructed and loaded on untreated fill.a similar The effectiveness of the stone columns in reducing maximum settlement increased with applied load with a value of sr = 0.59. Testing

The different objectives that testing may be required to fulfil are broadly similar to those specified for dynamic compaction and discussed earlier in the chapter (section 8.1). Monitoring the treatment process itself is an important element in the control and assessment of the treatment. Plant with continuous in-cab recording of vibrator depth and power output is now available and offers a significant advantage in control of the vibro works on site. The introduction of rig -guided depth vibrators with automatic stone-delivery systems has led to the increasing use of in-cab instrumentation systems and onboard computers. This enables data on power demand during initial penetration and column construction, stone consumption, and vibrator mov ement to be related to time as well as column location. Records should include depth of penetration and amount of stone used for each column. Instrumentation to measure vibrations of the depth vibrator may be helpful in assessing the treatment. Increasingly vibro treatment is being applied at geotechnically complex sites and a commensurate degree of process control and data feedback are e ssential. Automatic in-cab recording has clear advantages for

8.2 Vibro techniques process control and contract purposes. In-situ penetration tests such as SPT, CPT and DP (section 7.5) have been used to provide a comparison of treated and untreated ground. The comparison is, however, largely qualitative, and of value mainly in cohesionless fills. Greenwood (1970, 1990) has reviewed load tests on stone columns. The most common form of testing for vibro is the plate loading test. A 600 mm diameter plate is placed on top of a column and the load deformation behaviour is determined during a quick loading and unloading cycle. The load is applied by a hydraulic jack, using the weight of a vehicle or crane as reaction. This type of test is carried out as a routine control procedure. It may give some indication of workmanship and uniformity, but the results of the test cannot be used for design or to predict the long-term movements of structures which stress a large number of columns and the intervening ground. To predict movements of structures, it is necessary to load a representative area that includes a number of columns and the surrounding ground in the same way that the structure will apply load to the treated ground. It is also necessary to maintain the load for a reasonable period, to obtain an indication of the rate of settlement in the long term, after the immediate response to the application of the load. This involves the use of kentledge to apply the load. Such tests are sometimes called zone or area tests. A concrete slab is cast over a number of columns andthe loaded to about times test the working Whereas standard plate1.5 loading may be load. relatively inexpensive, a large-scale zone test may be costly. On small jobs the cost of testing can be a significant proportion of the total cost. Simpler and cheaper forms of area test, appropriate for typical housing loads, are the model footing test and the skip te st. A model footing can be cast over two columns and loaded with kentledge. A test has also been developed using a portable footing. In the skip test a small area is loaded by a rubbish skip filled with sand (section 7.6). Larger stresses can be applied by placing a second skip on top of the first, but are limited typically to loads from walls of two-storey housing. The ultimate evaluation of the effectiveness of vibro is the long-term performance of the structure built on the treated ground. There are not many reported instances of unsatisfactory performance. However, there are few documented case histories where settlements have been measured. There is a need to carry out more monitoring of structures built on stone columns. The testing of vibro concrete columns is carried out in a similar way to pile testing, employing both static and dynamic tests. In the static test a high-strength re inforced concrete loading cap is cast onto the head of the column and load is applied by jacking against a reaction frame loaded with heavy kentledge. Where columns form the support for widespread loading, often incorporating a granular load transfer platform reinforced with ge ogrids, a zone test may be carried out in which several columns

83

and the intervening ground are loaded to simulate the working situation. Dynamic testing of concrete columns is increasingly used, often to supplement the more expensive static tests which can be used as reference tests on larger sites. Performance

Some case histories are described in Part IV.  In case history 15 a deep urban fill was treated to a maximum depth of 5 m. Where the full depth of fill was treated the maximum settlement of houses built on the treated ground was 15 mm. Elsewhere a settlement of 78 mm was recorded. In case history 25 a soft clay fill was treated . Load tests did not show any improvement due to treatment.  In case history 27 an alluvial sand was treated. Settlement of houses built on the treated ground was small, but would not have been very large if the ground had not been treated.  In case history 28 a miscellaneous fill comprising ash and stone overlying a mixed cohesive fill was treated. The performance of a strip foundation on the treated fill was compared with a similar strip constructed on untreated fill. The effectiveness of the stone columns in reducing maximum settlement increased with applied load. 

A loose chalk fill placed more than 100 years ago to form a railway goods yard was treated by installing stone columns when it was for light industrial buildings (Stroud anddeveloped Mitchell, 1989). Excessive settlement occurred as the buildings were being completed, when rain-water from the ne w roofs was discharged into soakaways. Wilde and Crook (1991) have described the settlement of a factory unit of area 90 m × 20 m in Warrington. Soft alluvial soils varied in thickness from 5 m to 10 m. The site was brought to the required level by adding up to 1.5 m depth of fill. The site was then treated by vibro stone columns, but not to the full depth of the alluvial soil. The unit was built on simple pad foundations. Monitoring of the structure showed that over a period of 6 years, 120 mm of settlement occurred. There was proba bly another 50 mm of settlement during the construction period which was not recorded. Levelling of the floor slab revealed similar movements to those suffered by the foundation. It is clear that the cause of the movement was the weight of the fill, not the re latively small weight of the industrial unit. Greenwood and Kirsch (1983) plotted a graph showing reduction in settlement from published results of field measurements and laboratory testing programmes. Although field data can be very variable in quality, the following broad conclusions can be reached:  where Ar < 0.25, usually there has been re latively little improvement  where Ar > 0.25, usually there has been major improvement.

84

Chapter 8 Treatment of fills

Figure 41 indicates that with Ar = 0.3, sr might typically be in the region of 0.5. Vibro concrete columns have more in common with piling than with ground treatment and this is confirmed by monitored field pe rformance. In applications involving embankments a load transfer platform is often required. Ground treatment for the toll plaza for the Second Severn Crossing (Maddison et al , 1996) included both vibro concrete columns and a load transfer platform incorporating geogrids.

2 m centres, with additional treatment points under ground-bearing slabs if ground-bearing slabs are deemed to be suitable. However, in some cases suspended ground-floor slabs may be required. The soil profile should control the required depth of treatment. In fills this is normally the full depth of the fill and any underlying soft natural soils. Where the fill is deep and full depth treatment is not economic, the adequacy of partial depth treatment should be carefully assessed before it is adopted. Excavation adjacent to treated areas can result in settlement due to loss of support to the stone columns.

Design and specification

It is not easy to predict performance on filled ground accurately. Reliance is placed on comparison with past projects on similar sites. Some basic analytical approaches have been outlined for different situations. With loose granular fills, treatment should densify the fill significantly and the design should be based on the behaviour of the densified fill. The most important parameter is the spacing of treatment points. Generally, treatment points should be not more than 2.0 m apart. With cohesive fills, it can be assumed that the fill is not significantly affected by the installation of stone columns, and the design should be based on the composite action of fill stiffened by columns. The area replacement ratio Ar is an important parameter. Sometimes there is little engineering input to the design of vibro, apart from that of the specialist contractor. When a suitably qualified ge otechnical engineer is involved in the of initial investigation, site supervision, continuity design philosophydesign can beand maintained throughout the project. Appropriate supervisory personnel, and not only the operatives, should be on site during ground treatment. The responsibilities of the different parties should be clearly defined. A technical specification for vibro stone columns with notes for guidance has been published by BRE (2000). It has been prepared for use by specialist contractors, consulting engineers and other building professionals concerned with the design, procurement and supervision of vibro treatment. It provides a technically prescriptive specification, including design procedures, which is based on accepted best practice. A specification for ground treatment with companion notes for guidance published by the Institution of Civil Engineers (1987a, 1987b) includes vibro. It deals with materials and workmanship but does not attempt to specify design criteria. The notes for guidance describe suitable materials and gradings for the stone and state that stone columns shall be located within 150 mm of the plan positions shown on the layout drawings. Treatment typically comprises the placement of lines of stone columns beneath loadbearing walls or to a grid pattern beneath rafts. The geometry of the proposed structure constra ins the spacing of the columns in plan. For small structures such as low-rise housing there is little scope for variation in treatment pattern. Treatm ent points along each length of footing are gene rally at about

8.3 Preloading Temporary preloading over- consolidates the fill prior to construction. In the simplest form of this treatment, the fill is loaded with a surcharge of fill and the process does not require specialist equipment or skills. While this is advantageous in many respects, usually there is no commercial interest in promoting the method and for this reason it may be overlooked. Principle

The superior load-carrying characteristics of many natural soils can be attributed to preloading during their geological history. This over-consolidation has made them stiffer under applied loads than a comparable normally consolidated soil. Similarly, the load-carrying characteristics of a fill canthen be improved byon temporary preloading. Construction takes place overconsolidated fill. The temporary increase in effective stress can be achieved by applying a surcharge of fill but, in some situations, it could be induced by lowering the ground-water level or applying a vacuum. Application

The application of a surcharge of fill can be used to treat two different types of fill:  Uncompacted fills with large air voids Compression is largely immediate, and it should not be necessary to leave the surcharge in position for an extended period.  Saturated fine fills Consolidation may take time, and it may be necessary to install vertical drains to speed it up (Hansbo, 1993). In the latter situation, it may be necessary to control the rate of placement of the surcharge so that the preloading does not cause instability due to excess pore pressures in the fill. This book is primarily concerned with the former situation. The method has a wide field of application but there are some restricting factors. A relatively large area is needed for preloading with a surcharge of fill to be practical and, because the cost depends on the haul distance, a local supply of fill is usually required. In some coarse fills, it may be practicable to increase the effective stress by lowering the ground-water table. On very soft fine fills, vacuum preloading may be

8.3 Preloading advantageous because it may not be feasible to place a surcharge of fill which is several metres high. Water is drained from the ground by the application of a vacuum. A hydraulic gradient towards the drains is created by reducing the pressure in the drains rather than by creating excess pore water pressure in the surrounding ground. Method and equipment

Preloading with a surcharge of fill requires only the normal earthmoving machines. The appropriate type of plant for a particular job will depend on:  

the quantity of surcharge fill to be moved the haul distance  the trafficability of the surcharge fill. Figures 42 and 43 illustrate the use of surcharges to preload fills. Where uncompacted fills are being compressed and the surcharge does not need to be left in position, the surcharge fill can be moved around the site in a continuous earthmoving operation. This may greatly reduce the amount of fill which is required and make efficient use of the earthmoving plant. If the surcharge fill needs to be left in position for a short period, two surcharges located at opposite ends of the site can be moved alternately, thus keeping the earthmoving plant working. This procedure was used in case history 20. Where there is a high water table, it may be desirable to leave part of the surcharge in position to raise the ground

85

level permanently. The part left in position should be placed as an engineered fill ( Chapter 9). The practicability of preloading by lowering the ground-water table depends on the permeability of the fill. Some means of pumping water out of the ground is required. Methods of ground-water lowering, including sump pumping, well-points, deep we lls and horizontal drainage, have been reviewed by Bell and Cashman (1985) and Preene et al (1997). Continuous pumping may be required over a considerable period and the lowering of the water table could have deleterious e ffects on adjacent buildings. In vacuum preloading, an impermeable membrane is placed over a granular filter layer and sealed into the clay fill at its edges. Vertical drains are usually installed in the clay fill. A vacuum pump is used to pump air out of the granular filter layer causing the clay fill to consolidate. Analysis of behaviour

The magnitude of surcharge fill needed to effect the required improvement of the load-carrying characteristic s has to be determined. This necessitates an examination of the causes of settlement, because longterm settlement rather than bearing capacity is likely to be the main geotechnical problem. The variation of vertical effective stress with depth at a number of stages in the history of the fill should be e xamined:  the maximum effective stresses which existed in the past, σ'va 

σ'vb the present situation with the surcharge inprior place,to σsurcharging, 'vc with the surcharge removed, σ'vd  with the structure built, σ 've.  

The stress ratios σ'vc/σ'va and σ'vc/σ've are useful in assessing the effectiveness of preloading. Figures 44 and 45 present typical stress histories for two cases:  preloading with a surcharge of fill  preloading by reducing the ground-water level. Testing Figure 42 Preloading opencast backfill

The settlement of the ground surface induced by the treatment is a measure of the effectiveness of preloading. In some instances the installation of magnet extensometers to measure the distribution of vertical strain with depth will be justified. A programme of in-situ testing may have several objectives:  to determine the properties of the fill prior to treatment  to assess the improvement brought about by preloading  to determine the load-carrying characteristics of the treated fill, by load tests  to assess long-term performance, by monitoring settlement after treatment. Performance

Figure 43 Preloading old domestic refuse

BRE has monitored the effectiveness of preloading on a number of different types of fill, including restored

86

Chapter 8 Treatment of fills

Figure 44 Stress history due to preloading with a surcharge of fill

Figure 45 Stress history due to preloading by lowering groundwater level compared with stress history due to surcharge

opencast mining sites, old domestic refuse sites, infilled docks and a pfa lagoon. Case history 2 describes preloading an area of opencast mining backfill with a 9 m high surcharge of fill. The surcharge was placed by towed scrapers over a threeweek period, left in position for a month and then removed. Most of the settlement occurred as the surcharge was being placed. The movements measured while the surcharge was left in position were small. A small amount of heave occurred as the surcharge was removed. The stresses produced by preloading were much greater than those subsequently applied by foundation loads. The settlement versus depth profile suggests that the surcha rge was effective down to a depth of 10 m where σ' /σ' = 1.7. Case history 3 is a vc va of preloading following the commercial application experimental work described in case history 2. Case history 4 describes the effect of pre loading by a 30 m high overburden heap on the collapse compression behaviour of an opencast backfill. The total settlement measured at ground level between 1973 and 1992 was much greater where the backfill had not been pre loaded than where it had been . In the lower part of the backfill, which was saturated by the rising ground-water table, the average collapse compression was five times greater in non-preloaded ground. It can be concluded that although the preloading did not eliminate collapse compression, it

very substantially reduced it. Virtually all the settlement in the preloaded ground was located at depths more than 34 m below ground le vel. The depth to which surcharging was effective was therefore a little greater than the height of the surcharge and corresponded to σ'vc/σ'va > 1.8. Case history 19 describes the construction of a road interchange over old domestic refuse. It was specified that suitable fill was to be placed to form an embankment with a minimum height of 3 m. This was to be left in position for at least three months. Subsequently, fill would be removed to the final earthworks profile. Most of the settlement occurred as the embankment fill was placed, and the settlement monitoring demonstrated that it was unnecessary to leave the surcharge fill in position for a long period. The specified three-month period was more than adequate. Long-term movement has been very small. However, the stress ratio σ'vc/σ'va was greater than 1.8 throughout the full depth of fill at both locations where settlement was measured. Case history 11 describes how a 2 m high surcharge of colliery spoil was placed over a 50 m wide finger of a dock at Methil which had been infilled with colliery spoil. It appears that the surcharge effectively compressed the colliery spoil infill to a de pth of about 5 m, that is where σ'vc/σ'va > 1.6. Case history 12 describes the preloading of lagoon pfa

8.4 Pre-inundation

87

with an 8 m high surcharge of brickbats. The surface of the pfa settled 0.1 2 m. Two- thirds of this settlement took place as the surcharge was being placed and 90% of the settlement had occurred within a month of the completion of the surcharge. The bulk unit weight of the brickbats was 10 kN/m3 and thus a preload pressure of 80 kPa was applied to the pfa. Th e saturated bulk unit weight of the pfa was only 16 kN/m 3, and consequently the distribution of vertical effective stress was very dependent on the level of the water table. This varied seasonally from ground level to 4 m below ground level. The water table was falling quite rapidly during surcharge placement and this may have affected the settlement response of the pfa. The stress ratio σ'vc/σ'va = 1.9 at the bottom of the pfa and there appears to be compression throughout its full depth. Tomlinson and Wilson (1973) have described the preloading of some colliery spoil which ha d been end tipped into a flooded clay pit. The SPT gave a mean N = 10. The existing spoil formed a 1 5 m deep fill and this was surcharged with 5.5 m of additional spoil providing an applied pressure of 100 kPa. Additionally, the water level was lowered by 2.5 m. The surcharge was moved progressively across the site. Vertical compression of up to 3.3% was measured, corresponding to a constrained modulus of about 3 MPa. There are a number of examples of the use of vacuum preloading in China. The method has been used on 480 000 m 2 of reclaimed land at Xingang Port (Shang et al ,

Figure 46 Depth of influence of surcharge loading of fills (after Charles and Watts, 2000) ze = depth of effectiveness H = height of surcharge B = width of surcharge γs = bulk unit weight of surcharge γ = effective unit weight of the loaded ground

ratio of the increment of vertical stress produced by the surcharge to the existing overburden stress, H has been

2

1998). Chu et ala 5(2000) have described 50 000 site in Tianjin where m thick very soft clayalayer hadm been formed from dredged slurry. Burland and Burbidge (1985) suggested that for sands the settlement due to foundation loading on preloaded sand will only be one-third of the settlement that would have occurred without preloading. Design and specification

The height of the surcharge should be calculated in relation to the properties of the uncompacted fill that it is necessary to improve. This is simple where settlement caused by the weight of buildings will be the major problem, but more difficult where other causes of settlement are perceived to be major hazards. It is useful to examine the variation of effective vertical stress with depth within the fill at various stages. The surcharge should ensure that:  the fill is everywhere preloaded to stresses greater than those which will subsequently be imposed by building development, that is σ'vc/σ've > 1.2  the vertical effective stress ratio σ 'vc/σ'va > 1.8 to the depth to which it is decided that the fill should be improved. Figure 46 is based on the BRE field studies of preloaded fills and shows the relationship between the ratio of the depth of effectiveness to the height of the surcharge, ze/H, and the ratio of the height of the surcharge to the width of the surcharge, H/B. Since ze is a function of the

s is the bulk unit γ , where multiplied by asurcharge factor γs/and weight of the γ is theγ effective unit weight of the loaded ground. The effective depth of influence of the surcharges is based on the depth at which 90% of the settlement had occurred. Attempts to estimate ze from simple linear elastic theory tend to over-estimate it when compared with field data, and Figure 46 shows a relationship derived by Charles (1996) which gives better agreement with the field data. In the range of most practical interest with 0.1 < H/B < 0.5, and with γs/γ = 1, ze/H is likely to be of the order of 1.0 to 2.0. Where fills are in a loose unsaturated state, compression will occur mainly as the surcharge is placed and consequently there is no ne ed to leave the surcharge in position for an extended period. Where saturated clay fills are preloaded, it may be necessary to install vertical drains to speed up the rate of consolidation. The following should be specified:  type of fill to be used for surcharge  extent in plan, height and slope angle of surcharge  any restriction on rate of placing of surcharge to maintain stability (usuall y only a problem where a saturated low-permeability fill is preloaded)  period surcharge to be left in position.

8.4 Pre-inundation Wetting of a loose, partially saturated fill material can cause collapse compression and, if this happens after construction has taken place on the fill, serious problems

88

Chapter 8 Treatment of fills

can result (Chapter 5). It might be assumed, therefore, that inundation prior to construction could form a ground treatment technique for increasing the density of loose fills. In practice it is of limited applicability. Principle

Most loose, partially saturated fills undergo a reduction in volume when first wetted. The mechanisms involved have been discussed in section 5.1. Pre- inundation should remove the hazard of collapse compression occurring subsequent to construction on the fill. For many fills, collapse compression is the primary hazard and preinundation could be an attractive solution where it is practicable. Application

While pre-inundation is a ground improvement technique that should be considered in some situations, its range of applicability is limited. In some instances it may be usefully combined with another treatment method such as preloading. Pre-inundation is of limited application because:  it is only relevant to certain types of fill which, being in a loose condition and partially saturated state, are vulnerable to collapse compression  it is only bene ficial where a uniform treatment can be achieved (for example by a rising ground-water table); inundation from the surface may produce a variable and unsatisfactory treatment in some types of fill

Figure 47 Pre-inundation trial



it is difficultmay to control as movements caused inundation continue for some time after by inundation; in such situations pre-inundation is not suitable if construction has to take place immediately after treatment. An attempt to inundate a clay fill via surface trenches at the restored opencast ironstone mining site at Corby was not very successful (case history 2). It seems likely that the method will be successful only either when the ground-water table rises and saturates the fill in a uniform manner (case history 4) or when uniform inundation can be achieved from the ground surface by surface ponding. Method and equipment

Fills may be above an artificially lowered ground-water level:  during opencast mining the ground-water level may have been lowered to permit extraction of the mineral  during deep coal mining the ground-water level may have been lowered very substantially over a wide area. In these situations, inundation may be achieved when the pumps are turned off and the ground-water rises to a new equilibrium level. Inundation by infiltration downwards from the surface will be less effective in many fills, particularly heterogeneous fills. The water will tend to run away down the largest voids and fissures and will not produce a uniform treatment.

Wheremay surface ponding is used, the volume required be large. In case history 2, whereofa water square area, 50 m × 50 m, was inundated via 1 m deep trenches as shown in Figure 47, some 90 m 3 of water were absorbed during the first ten days of the e xperiment. Analysis of behaviour

The amount of collapse compression that a fill will undergo on saturation is difficult to predict. The improvement in behaviour that it produces is largely confined to rendering the fill no longer vulnerable to collapse compression. Other properties such as compressibility may not have been improved, indeed they may be poorer after inundation; for example, a clay fill will have been softened . Nevertheless, vulnerability to collapse compression may be the major hazard for building on many fill sites and the elimination of this risk is of great importance (Chapter 5). Testing

Vulnerability to collapse compression is difficult to determine by laboratory testing or the usual in-situ te sts. Inferences can be drawn from the placement conditions and from the e xisting moisture content and density. It should be assumed that an uncompacted or poorly compacted partially saturated fill that has never been inundated will be vulnerable to collapse compression. A field inundation test may be helpful in assessing the potential problem. The rate at which water infiltrates into the fill should be measured and the settlement of the

8.5 Other methods surrounding fill should be monitored (section 5.6). BRE has evaluated possible approaches for testing to assess susceptibility to collapse compression including the use of geophysical testing (Charles and Watts, 1996). Performance

BRE has monitored several sites where inundation has occurred:  as a method of treatment prior to house building, using inundation from the ground surface via trenches ( case history 2)  to investigate the mechanism of settlement which had damaged houses, using inundation from the ground surface via trenches (case history 5)  to investigate fill performance as the ground-water level rose (case history 4) Case histories of fills where pre-inundation has been used as a ground treatment are not common. Section 5.7 summarises some published case histories of buildings damaged by collapse compression. Design and specification

It is initially necessary to determine if pre-inundation is practical and beneficial and the following questions should be addressed:  is the fill susceptible to collapse compression?  can water be effectively applied to the fill?

89

into the fill slope drainage, to preserve and enhance slope stability  natural processes of evaporation and transpiration. 

The following are some of the techniques that might be appropriate in particular circumstances:  vertical drains to accelerate the consolidation of a preloaded saturated clay fill; sand drains, sand wicks and band drains  well-points to de -water fill temporarily; Steger (19 86) reported a successful de-watering of a fill with permeability 1.4 × 10-4 m/s 

vacuum well-points to accelerate the consolidation of fine fills; the cost of electrical power can make this an expensive method  horizontal drainage layers incorporated in an engineered clay fill during construction to accelerate consolidation  electro-osmosis to reduce the moisture content of a fill; when an anode and a cathode are placed in a fine soil and a voltage is applied, water flows towards the cathode at a rate that depends on the voltage; if water is removed at the cathode and not replaced at the anode, consolidation of the soil occurs; the method can be effective in de -watering silts and silty sands, but power requirements are high and the method is expensive  fine fills usually acquire a crust of stronger material through the natural process of desiccation; the crust

Where pre-inundation is feasible, extent of it itswith effectiveness and the possibility ofthe combining some other form of treatment should be examined. The source of water, the method of infiltration, the period over which treatment will be carried out and the required monitoring should be specified.

may quite thinofand a misleading impression of thebecondition thecan bulkgive of the fill (Hartlen and Ingers, 1981)  underdrainage installed prior to hydraulic filling may mean that the water table can be lowere d substantially and pe rmanently.

8.5 Other m ethods

Chemical stabilisation

Having considered in detail some of the principal methods of ground treatment, four other ways of modifying ground conditions are briefly reviewed: drainage, chemical stabilisation, reinforcement, and explosive compaction. Finally, some remedial strategies for contaminated fill are considered. Drainage

Fill behaviour is controlled by the principle of effective stress, and the use of drainage methods to control pore water pressure is therefore of major importance in many practical situations. Successful building development on some types of filled ground will be largely influenced by, and contingent on, appropriate drainage measures. Reduction of pore pressure can increase slope stability, reduce compressibility and reduce the possibility of fills eroding. Drainage has a wide variety of applications and a number of aspects need to be considered:  deep drainage within the body of the fill, to control ground-water level  surface drainage, to control infiltration of surface water

The behaviour of fills may be improved by additives which may either modify fill behaviour by physico-chemical processes or cement the fill material together. This type of stabilisation may have one or more of the following objectives:  to improve load-carrying properties (for e xample improve bearing capacity, reduce compressibility , increase resistance to water softening)  to reduce permeability  to stabilise chemical contaminants. Cement, lime, bitumen and various chemicals may be mixed with the fill in the following situations. During placement of the fill in layers Strength can be improved by the addition of, for example, lime to clay fills or cement to granular fills. Lime stabilisation of soft clay fills involves two stages: a rapid modification into a material that is granular and friable in nature, accompanied by a reduction in moisture content and plasticity index and an increase in strength, followed by stabilisation with long-term increase in strength due to

90

Chapter 8 Treatment of fills

formation of cemented structure in the pores of the compacted clay–lime fill.

During deposition of waste slimes, slurries, sludges, muds and fine dredged material Behaviour during sedimentation may be modified by chemical treatment, because flocculation can increase the rate of sedimentation of clay-size particles but the total amount of consolidation may be reduced (Bishop and Vaughan, 1972). BS 1924:1990 (BS I, 1990b) describes methods of testing cement- and lime-stabilised soils.

Subsequent to the completion of filling by some in-situ mixing process Deep stabilisation by the formation of stabilised soil columns has developed on two separate lines:  deep foundation stabilisation of soft clay was first used in Sweden in 1967 and the method has bee n widely used in Scandinavia since 1975  from the early 1970s , deep soil mixing has been developed in Japan to improve the properties of

cohesive soils to considerable depths; cement or lime have been used and depths of as much as 50 m have been treated. Dry mixing methods are commonly used in Scandinavia to form lime, lime/cement and cement columns (Broms, 1999). Deep in -situ mixing of lime with the soft clay is achieved by feeding a measured quantity of lime through a rotary drill equipped with special auger bit to both advance to the required depth and mix the soil and lime thoroughly during withdrawal. The diameter of the columns is 0.5 m to 0.6 m and columns of 15 m length have been installed. Lime columns have been used to reinforce soft clays (Bredenberg and Broms, 1983) and might be used in fine hydraulic fills (Broms and Anttikoski, 1983). Soyez et al (1983) have described loading tests on a clayey hydraulic fill stabilised with lime columns: with m diameter columns at 1of mthe centres, settlement was0.5 reduced to nearly a quarter settlement of untreated ground but columns at 2 m centres appeared to have little effect on the magnitude of the settlement. In the deep soil mixing method developed in Japan, the two principal stabilising agents have been lime and cement, but cement is now the primary agent (Toth, 1993). The method has bee n used primarily to improve bearing capacity and settlement performance. Stabilising agents such as cement slurry or quick lime are forced into the ground under pressure and mechanically mixed with soft clay soils down to depths of 50 m. The mixing equipment penetrates to the specified depth and is afterwards withdrawn with rotation. By injection into the fill Grouting techniques can find applications in the treatment of contaminated land where they could both mitigate pollution problems and improve ground engineering conditions. Technical problems include the difficulty of assessing the effectiveness of the treatment, possible incompatibili ty of contaminants and grout materials, and inability to carry out ground injection at depths of less than about 2 m . High costs are also likely to be a problem. Jet grouting has been used to improve loose fills (Oteo and Sopena, 1991).

Reinforcement

The properties of a fill may be improved by physical inclusions introduced into the fill at appropriate locations. The inclus ions may be linear, planar or in the form of a grid. They may be extensible or non-extensible. The main application of these reinforcing techniques is associated with slope stability. Two cases should be distinguished:  reinforcement of engineered fills where inclusions are placed in the fill as the fill is placed and compacted in layers; in the reinforced earth technique, flexible strips of galvanised steel are laid on the layers of the fill and bolted to retaining panels at the vertical or inclined boundary of the fill  reinforcement of non-engineered fills where inclusions are installed subsequent to the completion of fill placement; soil nailing can be used by driving rods into the fill from the surface. Various types ofbe geosynthetics such asreinforcing geogrids and geotextiles may used as horizontal elements within engineered fill. They can be combined with vertical reinforcement systems to support widespread loads such as large floor slabs or embankments. The toll plaza for the Second Severn Crossing involved the construction of an e mbankment over highly compressible soils and the foundation system comprised vibro concrete columns and a load transfer platform of granular fill incorporating low-strength geogrids (Maddison et al , 1996). The proceedings of the international reinforced soil conference held in Glasgow in September 1990 (McGown et al , 1990) and the third international conference on ground improvement geosystems held in London in June 1997 (Davies and Schlosser, 1997) provide a useful review of the state of this rapidly developing subject. Explosive compaction

Explosive compacti on has been used in projects throughout the world, usually to compact sandy soils below the water table. Reviews of the method have been presented by Mitchell (1981), Van Impe (1989) and Gohl et al (2000). Blast patterns are generally based on a rectangular grid of boreholes at spacings of 4 m to 9 m. Boreholes are drilled to the depth of fill to be densified using a bentonite suspension or by introducing plastic casing (Van Impe, 1989). An explosive charge is placed and the borehole filled up. Explosives are detonated at

8.5 Other methods

91

predetermined intervals in a carefully selected sequence. Detonation of an explosive produces two forms of energy release: the shock wave from the detonation front and the work done by the high- pressure gas formed in the explosion as it expands. About one- third of the charge energy is available for work in expanding the cavity containing the charge, and it is this component which is of interest for compacting the soil (Gohl et al , 2000). The shock wave is predominantly a compression wave, but shear waves are also formed. High pore water pressures may be induced by the shock wave in a saturated fill. This can cause a sand fill to liquefy, with subsequent



readjustment, as pore pressures quickly dissipate, to a denser and more stable condition under the influence of the weight of the overlying material. The method has been used in the U SSR on sands and silty sand ( Ivanov, 1980). The densification of sand tailings has been described by Klohn et al (1981). Gohl et al (2000) have summarised nine case histories including three field trials: typically ID was increased from 0.45 to 0.75 in sandy soils with a volume change of 4% to 8%. The possibility of damage to existing structures is clearly an important consideration. Blasting within 30 m to 40 m of e xisting structu res requires a reduction in the charge weights with a consequent reduction in blast hole spacings, and in the number of holes detonated at any one time (Gohl et al, 2000). Explosives are an inexpensive source of readily transported energy and the method of e xplosive

and cost of fill removal need to be considered, and this option may be come increasingly financially unattractive as measures are taken to limit environmental impact in transit and at the other site. Where fills contain undesirable materials, particularly biodegradable or contaminated wastes, it may be necessary to transport the fill a considerable distance to find a waste disposal site which will accept it. The cost of fill removal is related to the distance it has to be taken for disposal, and consequently it may be very e xpensive to remove all the fill. Excavation of colliery spoil can be hazardous where this has caught fire and there are high temperatures. The suitability of the site for building de velopment subsequent to fill removal has to be assessed. The topography may not be suitable for building development and removal of the contaminated fill may have to be followed by the placement of a new imported engineered

compaction densificat ionconcerns at great depths. mainly been permits used where there are over It has liquefaction. Successful use of the technique requires considerable experience and Faraco (1981) suggested that lack of experience with the method was the re ason why poor results were obtained with a granular hydraulic fill at a site in Spain. As far as is known the method has not found application in the UK.

fill Containment (Chapter 9). of contamination has a vital role in the

Remedial strategies for contaminated fill

Where fill has become chemically contamina ted, it has to be decided whether any action is required and, if it is, the most appropriate type of remedial action, such as:

remove material contain material in the ground  treat the ground to remove the contamination. 

The vast majorit y of contaminated land in the U K is remediated directly by on-site containment or indirectly by excavating and transporting to contained landfill. A complete solution to physical and chemical problems occasioned by the presence of contaminated fill is to remove all the fill. There may be a particular preference for removal of contaminated fill off-site due to concerns over future liabilities. However, the practicality

utilisation of many brownfield sites. The objective is to block or control the pathway between a hazard and potential receptors. Engineered cover layers and inground vertical barriers such as slurry trench cut- off walls have long been accepted as reliable means of containing contamination. Slurry trench cut-off walls using selfhardening cement–bentonite are the most common type of in-ground vertical barrier used in the UK to control the lateral migration of pollution and gas from contaminated land and landfill sites. A specification for these walls has been published by the Institution of Civil Engineers (1999).

92

Chapter 9 Engineered fill Since the reasonable but otherwise arbitrary choice of 95% Proctor densities as the target value for fill control work of the 1920s was made, there has developed a fixation on the part of specification writers to require this percentage compaction, irrespective of loadings, fillthickness, or other factors that should logically influence compaction requirements. (Monahan, 1994)

Where suitable fill material is placed to an appropriate specification under controlled conditions, most of the problems associated with construction on fills should be greatly reduced if not e liminated. The two major considerations are first the preparation of an appropriate specification, and secondly the provision of adequate site supervision to ensure compliance with it. Through such a specification and its enforcement during placement, a fill can be produced which has engineering properties which are known, have an acceptable degree of uniformity and which are adequate for the purpose for which the fill has been designed. The required behaviour of the fill should be well defined and it should be possible to make reliable estimates of settlement under working loads including

compression to very small amounts except for great depths of fill (Chapter 4). At moisture contents norma l in the UK for coarse fills, heavy compaction should virtually eliminate vulnerability to collapse compression (Chapter 5).  Heavy compaction of clay fills at moisture contents close to or above Proctor optimum should reduce Va to below 5% and largely eliminate the risk of collapse compression (Chapter 5). The required compactive effort is roughly proportional to the undrained shear strength and with very stiff clays it may not be possible to reduce air voids to below 5% with normal compaction plant. In such cases the addition of water may be required during the fill placement process.

long-termsometimes settlement have and be aring capacity. However, problems been experienced be cause inappropriate specification or inadequate control of the placement of engineered fills have led to excessive settlement or heave in service. The specification of engineered fill should define acceptable materials for the fill and the method of placement and compaction. Many of the materials described in the preceding sections may be used in engineered fills, including both natural soils and rocks and some waste materials. The fills resulting from the use of natural materials may commonly be described as follows.  Rockfill is produced by blasting or ripping rock strata and contains rock fragments which may be up to 1 m in size; it is usually compacted in thick layers with a heavy vibratory roller.  Sand fill may have properties similar to those of a natural sand although there could be important differences; compaction can also be by vibratory roller but thinner layers are used.  Clay fill is produced by compacting lumps or clods of clay; pneumatic tyred and sheepsfoot rollers have often been used.

While the specified fill may be adequate to support the building, other possible causes of settlement which could affect the building include:  settlement of underlying natural ground due to weight of fill  settlement of engineered fill due to self-weight  movement in engineered fill due to changes in groundwater level, pore pressures or seasonal changes in moisture content.

Selection of fill with appropriate properties and placement in thin layers with adequate compaction should ensure that the fill has the required engineering performance characteristics.  Heavy compaction of coarse fills should reduce creep

From a technical standpoint, a period of monitoring prior to building development is prudent. However, there may be strong economic reasons for building to take place almost immediately after the completion of filling. Case histories of the performance of engineered fills can be helpful. While there are re latively few case histories that relate specifically to engineered fills placed for building purposes in the UK, there are many which relate to embankment dams (Penman et al , 1971; Penman and Charles, 1973; Penman and Charles, 1985a, 1985b) and highway embankments (Farrar, 1978) which provide important informatio n on the settlement of fill and thus have relevance for engineered foundation fills. Brandon et al (1990) have presented a foundation fill case history from the USA of particular interest which describes the behaviour of deep clay fill placed for a housing development. Where subsequent building and road developments are foreseen, ope ncast mining backfills (section 2.1 ) are

9.1 Types of specification

93

placed as engineered fills. Roads, housing and other lowrise buildings have been built on such engineered fills (Thompson and Holden, 1990; Goodwin and Holden, 1993; Hodgetts et al, 1993; Morgan et al, 1993; Trenter, 1993). Colliery spoil (section 2.2) has been used as an engineered fill for road embankments and building developments. An 11 m depth of sand fill was placed at the site of the UK terminal for the Channel Tu nnel to bring the site to the required level; major structures have been founded on shallow spread foundations on the fill. A new township has been built at Peterborough on former brickpits and there have been a number of

In a method specification, the procedure for the work, including all the steps in the compaction process, is specified. This will include the type and mass of the compaction plant, the number of passes, the moisture content of the fill and its layer thickness. The process will be related to fill type and should be designed so that the technical requirements listed earlier in this chapter are met. Site control will involve inspection of the works to ensure compliance with the specification. In an end- product specification, compa ction is specified in terms of a required value for some property or properties of the fill when placed. The usual

geotechnical problems on this brownfield site. Some of the pits previously had been filled with pfa (case history 12). Other pits had been partially backfilled with weathered Oxford Clay, and in a major earthmoving operation undertaken in 1996, some 2 × 106 m3 of earth were moved to transform 160 ha of derelict clay workings into an attractive new landscape with hills and lakes. The earthmoving involved the placement of wet clay fill as an engineered fill using scrapers (Patel, 1995). The plastic limit of the clay was typically 25%, the liquid limit 60%, and the clay fill as-placed was some 5% to 10% wet of Proctor optimum moisture content. From 1997, low-rise housing has been built on the e ngineered clay fill, which has a maximum depth of 12 m. Settlement has been small.

measurements of the in-situ state of compaction are density, moisture content and air voids. Density will normally be related to some measure of the de gree of compaction such as 95% of a specified laboratory compaction test or 5% air voids. Alternatively, for a clay fill, the specification could be written in terms of undrained shear strength. Control by on- site testing as the fill is placed should be rigorous to e nsure that the specified value is being achieved. In a performance specification, some aspect of the behaviour of the completed fill is specified. This might be in terms of a maximum permissible settlement over a specified period following the end of fill placement or a required result from a loading test. With this approach, the specification is directly related to one or more aspects of the structure’s performance requirements. The performance specification may appear attractive

9.1 Types of specification Specifications have often been for highway embankments, butbased theseon arethose not develope necessarilyd adequate for fills on which buildings will be founded. Foundation fills re quire particularly strict quality control. A specification for a fill which is to form a foundation for buildings must address the following technical requirements:  a well constructed excavation, with all soft or hard spots removed, reasonably dry and well drained  sound fill without undesirable material and capable of compaction as specified, with the provision of starter and capping layers as necessary  placement and compaction to ensure that fill performance meets serviceability limit state criteria  monitoring. The specification should adequately describe the design requirements, be easily understood by the parties to the contract, be practicable and capable of both enforcement and measurement, and not be unnecessarily costly or time consuming in its application. It should be capable of being monitored by an e ffective form of quality assuran ce procedure with due regard to safety as re quired by the Construction (Design and Management) Regulations (Health and Safety Commission, 1994). Three approaches to a specification for engineered fills may be identified:  method specification  end-product specification  performance specification

to a client as itand provides link withfor pesuch rformance requirements placesaredirect sponsibility performance on the contractor, which may seem contractually clear cut. However, although compliance with a post- construction performance requirement may be easy to check, it may be difficult at such a late stage to obtain any adequate redress where non- compliance with the specification is established. Compliance with even a large load test may only demonstrate that the upper layers of fill are adequately compacted and may say nothing about the performance characteristics of the underlying fill. Such an approach also makes demands on the contractor’s quality procedures. For these reasons the performance specification is not often used. Successive editions of the earthworks specification included within the Specification for Highway Works have been widely used as a basis for foundation fills. However, this specification was prepared for highway purposes and therefore may not always be appropriate for foundation fills with respect either to the technical requirements or to the compaction procedures adopted. Tolerable settlement is likely to be significantly smaller for a building than for a road and hence a more stringent specification may be necessary for foundation fills than for highway purposes. Furthermore, highway schemes are often major civil engineering projects, whereas schemes involving low-rise buildings founded upon engineered fill are often relatively small in scale. Typically, a highway scheme is let to a major contractor with earthworks experience and the contract is

94

Chapter 9 Engineered fill

supervised by a consulting engineer with earthworks experience. The contract is normally of such a size that obtaining the compaction equipment necessary to fulfil the requirements of an elaborate method specification is not difficult. In contrast, engineered fills for buildings are often of shallow depth and restricted lateral extent. The work is unlikely to be e xecuted by a major contractor and the consulting engineer, if one is e mployed, may not be experienced in compaction work. Moreover, specific types of compaction equipment for small projects may not be readily available. In these circumstances, it is considered most

indicate the likely moisture content at which the fill will be placed. The compaction tests provide the optimum moisture content and maximum dry density values which are necessary for quality control. Liquid and plastic limit tests indicate the range of moisture contents over which cohesive materials are plastic (plasticity index IP = wL – wP) and draw attention to high-plasticity fills which could present heave problems or, in extreme cases, could render them unsuitable. The number of tests required at the investigation stage depends on the volume of fill to be placed and the homogeneity of the fill. For example, on a typical

appropriate to use an end-product specification. A model specification which has been prepared by Trenter and Charles (1998) is given in Appendix F.

relatively small site3with a volume of fill to be placed of less than 50 000 m it would be appropriate to carry out one set of compaction tests for every 5000 m 3 of fill with no fewer than two sets of tests for each distinct type of fill. The requirement that the tests are carried out on material which is representative of the fill to be placed is as important as the number of tests performed.

9.2 Site investigation A site investigation is an essential precursor to the design and construction of works containing engineered fill. General aspects of the site investigation of existing fills are presented in Chapter 7, but a site investigation carried out prior to the placement of an engineered fill has some special features. An investigation is necessary not only to determine the characteristics of the site under consideration, including the strata succession, groundwater conditions and parameters governing allowable bearing pressures and settlement, but is also required to identify and characterise the fills to be placed during construction Suchbut fillscould may srcinate not neceatssarily bepits located at thework. site itself, borrow some distance away. Certain tests have to be performed on the fills in order to create a basis for the design of the engineered fill. Great care should be placed on se lecting representative fill samples and carrying out the necessary tests since all subsequent control testing, to be undertaken during fill placement, will relate to the tests made on the fill during the site investigation stage. To provide the information required for design of the engineered fill, the following tests should be undertaken on the fill at this stage using the test procedures de scribed in B S 1377:  natural moisture content, w (BS 1377-2:1990 section 3)  liquid limit, wL (BS 1377-2:1990 section 4), and plastic limit, wP (BS 1377-2:1990 section 5), for cohesive soils only; to derive liquidity index, IL, such that I = (w – w )/(w – w ), where w is the moisture L a of the P fineLfraction P a content of the material required for the plastic and liquid limit te sts  compaction tests to determine maximum dry density, ρdmax, and optimum moisture content, wopt, at the appropriate compactive effort, 2.5 kg rammer (BS 1377-4:1990 section 3.3 or 3.4) or 4.5 kg rammer (BS 1377-4:1990 section 3.5 or 3.6)  particle density, ρ s, test (BS 1377-2:1990 section 8), needed to assist in evaluating the compaction test. The natural moisture contents of the fill at the borrow pit, with due allowance for seasonal wetting or drying, should

9.3 Fill ca tegories Fill should be clearly categorised into material which may or may not be used. In the model specification presented in Appendix F, the following four categories are adopted:  unsuitable fill  general fill  restricted fill  special fill Fill which is designated unsuitable shall not be used at any location on the site in any circumstance. General fill is all material e xcept that which is unsuitable, restricted or special. It may include both natural soils and rocks as well as some waste products. Restricted fill is material which would be classified as general fill except that it contains minerals hostile to the built environment. These restricted fills include such natural materials as pyritic shales and gypsiferous clays and waste materials such as burnt colliery discard and steel slag. Their use will be precluded in conditions where ground-water could rise to the underside of the de epest foundation or where authorities could reject them on pollution grounds. Special fill is high-quality material such as well graded natural sands and gravels, crushed rock or clean demolition rubble. Its use will normally be reserved for specifically defined purposes such as a capping layer or backfill to retaining walls. In selecting suitable material, there are both economic and technical factors to consider. While granular materials such as sand, gravel and rockfill are usually preferable to clay soils, the latter may provide acceptable fills in many circumstances. The use of clay fills will, however, involve the same types of problem that are encountered on many natural clay soils as clay fills, like natural clay soils, undergo changes in volume due to seasonal changes in moisture content.

9.5 Site preparation and fill placement

95

9.4 End product cri teria As already mentioned, high bearing pressures are not normally a problem with low-rise structures and the greatest threats to successful in-service performance are:  collapse settlement due to inundation of dry or inadequately compacted fills  excessive consolidation settlement of wet compressible fill  heave or settlement of clay fills due to climatic changes or the effects of vegetation. These features are dependent upon moisture movement and by restricting the voids in the fill the opportunities for excessive in-service movements should be reduced. A maximum allowab le air voids of 5% is a suitable criterion for most clay fills. However, specifying an air voids content is insufficient, since a 5% value may easily be achieved by adding water to the fill, without increasing compactive effort. An alternative control method would be to specify a minimum acceptable relative compaction, CR, that is a minimum acceptable dry density expressed as a proportion of the maximum dry density measured in a standard laboratory compaction test. Limits on moisture content would also be required as, if the fill is too wet, there may be excessive consolidation settlement, and if the fill is too dry it may be vulnerable to collapse compression. Reliance on a certain relative compaction is inadequate because itFor does not take the nature of itthe structure into account. a particular application, is necessary to consider what level of compaction should be adopted and in particular whether to use the 2.5 kg or 4.5 kg rammer compaction method as the basis for the specification (BS 1377-4:1990). Most of the work done by the Transport Research Laboratory which formed the basis of the successive highways specifications was based upon the 2.5 kg rammer method. The choice of air voids content was about 10% although reduced to 5% within 600 mm of the formation for certain soils. Road pavements are generally significantly more flexible than most buildings and for some low-rise structures the heavier 4.5 kg rammer method should be specified. Such cases could include:  structures with large open or fenestrated areas  structures of mixed (single- and two-storey) height  structures containing long loadbearing walls, without regular construction joints  structures containing heavily loaded floor areas  heavily loaded column bases associated with, for example, wide spans or machine gantries. Figure 48 illustrates diagrammatically compaction curves for the 2.5 kg and 4.5 kg rammer methods. Th e shaded area OPQR is bounded by the zero air voids line (OP), the 5% air voids lines (RQ), the maximum dry density for the 4.5 kg rammer method (RO), and the maximum dry density for the 2.5 kg rammer method ( QP). The shaded area OPQR, therefore, represents the range of moisture

Figure 48 Basis for design of engineered fill; moisture contents and dry densities to fall within stippled areas (after Trenter and Charles, 1998) OMC (2.5 kg) = Optimum moisture content obtained in the standard Proctor test OMC (4.5 kg) = Optimum moisture content obtained in the modified Proctor test MDD (2.5 kg) = Maximum dry density obtained in the standard Proctor test MDD (4.5 kg) = Maximum dry density obtained in the modified Proctor test

contents and densities expected to be acceptable for most fills designed to support low-rise structures. The particular part of OPQR in which the moisture content and density values should fall will depend the structure and, for relatively heavily loaded upon foundations or settlement-sensitive structures, moisture content and density values should be sought in the upper part of the diagram, near line RO or above. Lightly loaded buildings which are not sensitive to settlement could remain in the lower part of the diagram, near line QP and below. It should be kept in mind that heavily compacted plastic clays are prone to heave. While the approach based on end-product criteria determined by the 2.5 kg and 4.5 kg rammer laboratory compaction tests will be appropriate for many types of commonly used fills, there are situations where other methods may be preferable or necessary. Different compaction criteria are likely to be required for a uniform-size coarse fill or for a rockfill. Control on a large clay fill site might be in terms of a range of acceptable undrained shear strength. Chalk fill can present peculiar difficulties due to the degree of breakdown during earthmoving and this needs to be reflected in the specified compaction criteria.

9.5 Site p reparation an d fill pla cement The site should be cleared of all topsoil and other unsuitable material. Soft spots and hard spots, such as derelict foundations, should also be excavated and taken away. Surface water should be removed from ponds and depressions. Drainage grips or trenches should be dug on the uphill side of filled areas which are constructed in

96

Chapter 9 Engineered fill

excavations on side-long ground. Sump pumping may be necessary when filling some excavations below the ground-water level. Construction over the face of a quarry or an opencast mining highwall must be avoided. This is considered in Chapter 6. Special measures may have to be taken at the location of highwalls by providing flexible connections for services and by thickening construction for service and estate roads. As far as is practicable, fill thicknesses should be reasonably constant beneath a given structure to avoid differential settlement. Feather-edges which might result

200 mm or less may be necessary. W ith rockfills, thi cker layers may be acceptable provided that adequately heavy compaction plant is available. Where landscape or other non-loadbearing areas form part of a development, they need not necessarily receive the same level of compaction as the loadbearing areas. Where a lower level of compaction is deemed to be acceptable, there should be a transition zone around the loadbearing area as shown in Figure 50. Sensitive structures may merit a capping layer formed from special fill. Moisture content at fill placement is fundamental to subsequent performance. If the fill is too dry there is the

in foundations set partly on fill and partly on natural ground should be avoided and the site should be worked in such a way that structures are located either directly on natural ground or directly over a roughly equal thickness of fill. In order to do this, some ‘stepping’ of the natural ground may be necessary as shown in Figure 49. The effect of a variable depth of fill on the differential settlement of a building is examined in Chapter 6. For a given item of plant, fill layer thickness, fill moisture content and the number of passes of the compaction equipment determine compaction performance, although there are other factors such as the need to avoid e xcessive handling with materials suc h as chalk. Layer thicknesses should be such as to allow the compactive energy to spread throughout the layer in order to promote the specified fill density and low air voids content. For earthfills compacted to support low-

possibility of heave or collapse settlement; if it is too wet there is the possibility of insufficient strength and high compressibility. It will be difficult to achieve an air voids content of 5% or smaller when the moisture content is low. The importance of moisture content in achieving adequately low air voids is demonstrated by Parsons (1992), who described a case where ( wopt – 6.5%) required 32 passes to achieve 5% air voids, against just five passes for ( wopt + 5%). However, at ( wopt – 6.5%), the dry density was equivalent to 107% of the maximum dry density in the 2.5 kg rammer method, whilst ( wopt + 5%) was equivalent to only 94% of the same standard. Dry density decreases as moisture content increases wet of wopt and this should be kept in mind when attempting to restrict air voids content. Compaction of clay fill should endeavour to reduce air voids to below 5%, which should eliminate the possibility

rise structures, loose layer than of 250 mm are unlikely to bethicknesses satisfactory,greater and layers

of collapse compression air voids are a good indicator of fillon bewetting. haviour,Although it is difficult to determine the percentage air voids accurately. A clay fill can be specified in terms of undrained shear strength, cu, and there is a range of values of cu, typically 50kPa< cu < 120 kPa, which will generally be acceptable.  where cu < 50 kPa, it will be difficult for normal construction plant to place the fill; large excess pore pressures may be generated with significant consolidation settlements occurring after the completion of filling. Several investigations have been reported by TRL and others (Farrar and Darley, 1975; Forde and Davis, 1978; Parsons and Darley, 1982)  where cu > 120 kPa, it will be difficult to compact the fill adequately and reduce air voids to 5%.

Figure 49 Benching of sloping natural ground for engineered fill (after Trenter and Charles, 1998)

Figure 50 Fill beneath buildings, hard standing and landscaped areas (after Trenter and Charles, 1998)

9.6 Quality management

97

In the same way that the addition of too much water can detract from the performance of engineered fill, soil can be over-compacted. Granular soils and cohesive soils dry of optimum when rolled excessively become overstressed and a firm compacted surface is replaced by a loose tilth. This is an undesirable engineering characteristic, and, furthermore, it represents a waste of compacting resources. Whenever possible, site trials should be undertaken to determine the correct layer thickness, moisture content and number of passes, for the particular material in hand. These trials may be most conveniently carried out at the

As already noted, for cohesive soils, undrained shear strength, cu, forms an alternative basis for specification and control testing. However, different methods of measuring cu, such as the unconfined compression test and the vane test, can give significantly different values. For a particular type of test, the measured value of cu can be sensitive to the detailed test procedure, such as the rate of shearing, and it is important that the method of testing should be closely specified. The unconfined compressive strength ( qu = 2cu) also may be used as a basis for control. Unfortunately, the portable version of the unconfined compression test

beginning of the project when the work may be done using billed rates with some minor amendments.

machine which is commonly available for site use in the UK is re stricted to 38 mm sample size which is not suitable for cohesive soils containing significant gravel quantities. Larger capacity unconfined compression test equipment capable of testing 100 mm diameter samples is available but may not always be sufficiently rugged and portable for site operations. Moisture content and wet and dry density may be measured in addition to qu. All methods of control have their limitations and suitability mus t also be judged on cost, experience required in interpretation, and range of soils to which the particular method of control is applicable. Compaction tests are time consuming to perform and their interpretation requires considerable experience. However, the corresponding field control tests are relatively quick and easy to interpret and both the compaction test and the field control tests are applicable

9.6 Quality ma nagement Quality control procedures should be implemented to ensure compliance with the specification. The nature of the control procedure will depend on the type of specification adopted, and the end-product specification requires an appropriate type and quantity of testing of the fill during placement to ensure that the desired end product is being achieved. Depending upon the contract, quality control may be the responsibility of the contractor working under the supervision of the engineer or the work may be done by the e ngineer. In the model specification given in Appendix F, it is assumed that the former method is adopted but the appropriate changes could bethis readily made to permit control by the engineer, should be desired. Control parameters should be of technical significance in compaction operations and will necessarily be the same as those determined during the site investigation stage. The equipment necessary to obtain them must be sufficiently light and portable for transport in a LandRover or similar vehicle. For the model specification in Appendix F, the following are the most significant control parameters:  moisture content in respect of an optimum moisture content established at the site investigation stage  dry density in respect of the already established maximum dry density  air voids content which depends on moisture content and dry density  undrained shear strength which is an alternative to monitoring moisture content and dry density for clay fills. In addition to being of technical significanc e, both design and control parameters must be reproducible, a term which denotes the range within which measurements made on the same fill, by different operators using different equipment, should agree. Sherwood (1970) undertook a major study of the reproducibility of earthworks tests, involving the study of three soils and some forty different test laboratories. The results of this study indicated that the compaction tests were amongst the most reproducible of the common e arthworks tests.

to a wideusing range of soil or fill types. In these circumstances, control a laboratory compaction test together with field dry density and moisture content determinations is the most generally applicable approach. Nevertheless, for a limited range of cohesive fill types, control by undrained shear strength offers a viable alternative. Sampling and testing frequency may be based either on the number of samples ne cessary to achieve a certain level of statistical confidence or upon prescriptive codes and guidelines, usually written in terms of number of samples or tests per unit volume of fill. The statistical approach is probably best suited for large sites and is not dealt with further here. Small sites are more difficult to work than large sites; finished work may be damaged more easily in confined working areas and deficiencies in site preparation usually reflect more readily in poorerquality compaction than on larger sites. Consequently it would seem appropriate to test to a higher frequency for a small site than for a large one. A guide on sampling and testing frequency is given in the Australian Standard AS3798-1990 (Standards Association of Australia, 1990) which specifies a minimum frequency of testing of one test per 200 m 3 for small-scale operations, such as small residential developments, and one test per 500 m 3 for large-scale operations, such as industrial developments and road embankments. The testing has to be carried out reasonably evenly throughout the full depth and area. Minimum frequencies of testing in terms of tests per layer per unit area are also defined.

98

Chapter 9 Engineered fill end product in terms of the selected geotechnical parameters for the various fills (based on site investigation information)  a list of the required suitability tests, one form to be completed for each borrow pit under investigation  the suitability test results for each borrow pit  a list of the required control tests  the results of the control tests on each fill type or layer or area as appropriate  a list of post-compaction monitoring requirements  the results of post- compaction monitoring.

Figure 51 Guidelines for minimum quality control test frequency; test types — moisture content and in-situ density (after Trenter and Charles, 1998)

A suggested minimum test frequency is presented in Figure 51 which is broadly in agreement with the Australian Standard. However, each site should be judged on its own merits with careful note being taken of any problems revealed during site investigation. In very variable or difficult conditions more frequent testing may be required. Tests in visually doubtful areas, and re-tests of failed areas, should be carried out and are additional to the testing recommended in Figure 51. In the bulk of the fill the pattern should be stratified, but itofistesting also necessary tosystematic ensure thatand testing is conducted at locations where compaction is known to be difficult to achieve, for example adjacent to manholes, kerbs, service trenches or retaining walls. The objectives of quality assurance are the provision of a product with the required quality and the provision of evidence that the product has attained it. Verifiable records, including adequate documentation and photographs, are required. The regimented nature of compaction and its control lends itself well to the quality assurance cultur e which requires a defined work plan, monitoring to ensure that the plan is being adhered to, feedback to ensure that the plan is appropriate and to allow its modification if necessary, and a course of action to be followed if procedures have departed from the plan. For compaction control, the plan must contain sufficient flexibility for the assessment of new or changed fill materials; any such changes should be formally assessed and documented and not appear to be the result of arbitrary decisions. Modern compaction and control requires laboratory and field testing before ( ie during the site investigation), during and, possibly, after the e arthworks. The results of this work should be recorded, collated and prese nted so that the quality of the operation can be demonstrated. Without such a demonstration, doubts may remain over the fitness of the work. The required documentation includes the following:  a summary of the specification requirements and the

All forms when completed should be signed and dated by the person responsible and a list prepared of any required action or remedial work to be carried out. Once these actions have been completed, the forms may be filed for future reference. Whereas quality management checks certain geotechnical properties relevant to the enforcement of the specification during compaction, monitoring is a check on performance of the fill after compaction. Settlement and differential settlement of a structure can be measured and compared with values deemed acceptable for the structure in question. Monitoring, which is described in section 7.8, is not widely used for small sites and shallow depth fills, and only appears to be necessary for sites where:  particularly vulnerable struc tures are to be built  a substantial ground-water rise within the fill is expected.

9.7 Excavation and recompaction Where the problems with a non-engineered fill are associated with its loose condition, a simple remedy is to excavate the fill and replace it as an engineered fill in thin layers with adequate compaction. This form of ground treatment has the advantage that during e xcavation unsuitable material can be identified and removed from the site. Unacceptable material could include:  biodegradable wastes  chemically contaminated wastes with hazardous properties  material which is burning or susceptible to spontaneous combustion  materials in a frozen condition  clay fill outside an acceptable range of c . u

The recompaction of fill in thin layers is the same process and has the same requirements as placing an enginee red fill in the first instance. The replacement of the fill should be carried out to an appropriate specification under controlled conditions. In addition to those factors that need to be specified for all enginee red fills, the following should be specified:  depth of excavation (preferably full depth of fill)  criteria of acceptability of fill for re- use. Normal earthmoving plant is required. The type and

9.7 Excavation and recompaction quantity of plant selected for a particular job involving excavation and recompaction will depend on:  nature of fill material to be excavated  area and depth of fill to be excavated  space available for temporary storage of excavated material  specification for recompaction  time available for earthmoving. Where the fill is deep it may not be practical to excavate the full depth of fill and it is necessary to dete rmine the required depth of excavation and recompaction. The following factors should be considered in determining this depth:  depth to which foundations will significantly stress the ground  other possible causes of settlement at depth in untreated fill, for example collapse settlement due to rising ground-water level  practicability and cost of deep excavation.

99

The treatment method has wide application but the following situations may present difficulties:  on a small congested site, fill handling and storage may be difficult  where there is a high ground-water level, excavation may necessitate de-watering the site  very wet fill will be difficult to handle  in contaminated ground, special safety techniques and equipment may be required  where combustion is occurring, fill may be too hot to handle  where some of the fill is biodegradable, and hence unsuitable, it may be difficult to find a suitable site to which it can be taken.

100

Chapter 10 Foundations on fills On account of the fact that there is no glory attached to the foundations, and that the sources of success or failure are hidden deep in the ground, building foundations have always been treated as stepchildren; and their acts of revenge for the lack of attention can be very embarrassing. (Terzaghi, 1951) …the most fundamental principle of humility in civil engineering design is as follows. As a first step, use every possible means to avoid having to deal with conditions determined by the statistics and probabilities of extreme values. (De Mello, 1977)

The warning given by Terzaghi (1951) is particularly relevant to foundations on non-engineered fills. There is always the possibility that a non-engineered fill will contain unpleasant surprises. If the advice given by De Mello (1977) is taken seriously, not too much time will be wasted trying to calculate the risk associated with many possible but unlikely hazards associated with building on a non-engineered fill. Rather, the objective will be to develop foundation solutions whic h reduce or eliminate those risks. One approach is to improve the fill by using an appropriate ground treatment technique prior to building on the site. This has been considered in Chapter 8. Some other aspects of foundations on fills are examined in this chapter.

(1978) and Atkinson (1993). Only those aspects of these subjects which are peculiar to fills are considered in this book. Digest 251 (BRE, 1995a) deals with the assessment of damage and Digest 352 (B RE, 1993b) describes underpinning. These matters are not covered in this book. This chapter begins with an examination of the significance of differential movement. A preliminary classification of the severity of the problems involved in building on a particular fill is based on the magnitude of potential compression (or expansion) in the fill during and subsequent to construction. Shallow foundations on non-engineere d fills are discussed. The most difficult foundation problems are usually associated with small

study of descriptions the geotechnical aspectsproperties of building on fillThis has included of relevant of fills, methods of measuring those properties and techniques of improving the properties. However, successful construction depends not only on the load- carrying characteristics of the fill, but also on the sensitivity of the structure to ground movement. Foundati on design has to be related to predicted fill–structure interaction. Embankments, roads, storage tanks and many different types of building are located on fills. The more flexible forms of construction are less likely to suffer damage due to fill settlement. Buildings are generally much more sensitive to settlement than, for example, embankments or storage tanks. Moreover, there are significant differences in ability to deform without incurring serious damage between different types of building and forms of construction such as:  low-rise residential  high-rise residential  warehouse and factory units  loadbearing brickwork  portal frame

structures on deep fills, where through fill to an underlying firm stratum is not piling likely to be an the economic solution and the structure has to be founded directly on the fill. In later sections alternatives, such as deep foundations bearing on a firm stratum underlying the fill or complete removal of the fill, are examined. Finally, the special features of contaminated ground, which can pose further problems for constru ction on non-engineere d fills, are described. In addition to geotechnical and chemical problems, former industrial land and derelict sites can present difficulties for redevelopment related to the past usage of the land which include:  existing buildings, underground structures and plant with underground pipework  massive foundations of old buildings  lack of infrastructure  poor environmental location.

Foundation design and construction have bee n dealt with in a comprehensive manner by Tomlinson (1995). Foundation design is covered in BS 8004:1986, the code of practice for foundat ions (BS I, 1986) and in DD ENV 1997-1:1995, Eurocode 7: Geotechnical design — Part 1: General rules (BS I, 1995). Other helpful references on foundations for low-rise buildings include Tomlinson et al

10.1 Differential movement Distortion and damage to buildings are caused by differential movements rather than the magnitude of total settlement. Figure 52 shows a house badly damaged by settlement of an opencast mining backfill. Situations in which large differential movements should be expected include:  heterogeneous non-engineered fill  edges of filled areas and other locations where the depth of fill changes rapidly.

101

10.1 Differential movement

other end (C) (Figure 53).  differential settlement of B relative to A, δsBA = sB – sA  rotation of line AB is ( sB – sA)/( L/2)  tilt of wall ABC is ( sC – sA)/ L  relative deflection of wall ABC, ∆ = sB – 0.5( sA + sC)  deflection ratio of wall ABC is ∆/L For relative deflection and deflection ratio, sag is positive (+) and hog is negative (–). Burland and Wroth (1974) reviewed the settlement of buildings and associated damage. Followin g the work of Polshin and Tokar (1957), the onset of visible cracking was related to a limiting critical tensile strain in masonry and finishes. A value of 0.075% was adopted for brickwork, and critical values of deflection ratio at which damage could be expected were calculated for different ratios of wall length to height in frame structures and structures with loadbearing walls. From the theoretical relationship and field data which they brought together, it is evident that loadbearing walls subject to hogging are much more vulnerable to damage than when subjected to sagging. Damage due to hogging of loadbearing walls could occur with values of ∆/L as small as 0.2 × 10–3. The critical value of ∆/L for sagging can be twice as great. Frame structure can tolerate more relative movement than loadbearing walls. Deflection ratios measured on some buildings on fill are given in Table 15. Also included in the table are the maximum settlement, smax, and the maximum differential

Figure 52 House damaged owing to differential settlement of fill

Figure 53 Notation for settlement of wall

Some important terms used in describing ground and

δs)max, measured settlement, at the building an adjacent (building, and alsoeither the maximum verticalor on strain ( εv)max measured in the fill close to the building. No major damage has been reported at any of these buildings except in case history 5. There would appear to be some correlation between (∆/L)max and ( εv)max. At a number of sites where no structures have been built, traverses of fill settlements have been measured by pre cise levelling. These have been interprete d as def lection ratios in Ta ble 16. If structures had been built at these locations, undoubtedly deflection ratios would have been much smaller due to the stiffness of the structure. This illustrates the major effect of fill–structure interaction on foundation performance.

foundation have been summarised by Burland and Wroth movement (1974) as follows:  differential settlement, δs: settlement of one point relative to another  rotation: change in gradient of straight line connecting two points  tilt: rigid body rotation of whole structure  relative deflection, ∆ : maximum displacement relative to straight line connecting two reference points a distance L apart  deflection ratio, ∆/L: ratio of relative deflection to distance between reference points. The above can be illustrated by a wall of length L which settles sA at one end (A), sB in the middle (B) and sC at the

Table 15 Deflection ratios measured on low-rise housing built on fill ∆ L s (δs) max max Ca s eh i s t o r y ( mm ) (m) ( /L)max ( m m) (mm) ∆

(ε ) v max ( %)

1

1

180

130

2(preloading)

–3

19

–0.16 × 10–3

22

13

0.2

2(dynamiccomp)

–2

19

–0.10 × 10–3

59

18

1

–11

19

–0.6 × 10–3

143

90

2

+0.16 × 10–3

40

28

0.5 1

2(inundation) 2(notreatment)

—

+2

—

12.8

—

15

+7

25

+0.3 × 10–3

78

43

27*

–2

16

–0.13 × 10–3

6

3

–34

23

–1.5 × 10–3

300

150

5

* Housing built on alluvial sand, not fill

0.1 3

102

Chapter 10 Foundations on fills

Table 16 Deflection ratios measured on fill which has not been built on Case smax (δs)max (εv)max history ∆ ( /L)max ( mm ) (mm) ( %)

the hazard posed by potential movements in the fill under particular loading conditions. A threefold preliminary hazard classificatio n was suggested by Charles and Burland (1982):  category A, small movements  category B, significant movements  category C, very large movements.

4*

+2.4 × 10–3

735

146

2

6

–3.4 × 10–3

544

80

4

24

+2.4 × 10–3

160

64

6

22

–1.5 × 10–3

1086

132

5

The system is based on the potential maximum vertical compression in the fill underlying the structure during and after construction. Damage is a function of differential movement which

* Settlement traverse E

The use of parameters such as deflection ratio to provide criteria for acceptable deformations is relatively complex as soil–structure interaction has to be taken into account. Furthermore, horizontal movements which cause extension can also cause serious damage. Where small structures are built on poor ground, the buildings are likely to have stiff foundations. Indeed, such foundations can be provided relatively cheaply. In this situation, it can be assumed that the foundation is sufficiently stiff to ensure that the structure does not deform, but simply tilts, and that the superstructure is not subjected to horizontal tensile forces. As a consequence, a suitable deformation criterion can be defined in te rms of tolerable tilt, αT. The problems caus ed by tilt will depend on the type of the building and its purpose, but it is suggested that the following represent reasonably typical situations for lowrise housing:

will usually be difficult to estimate accurately. Maximum compression of the fill is used for preliminary hazard classification purposes rather than maximum settlement of the structure because differential settlement is more likely to be re lated to maximum compression than to maximum settlement. It is an implicit assumpti on in this classification system that movement of the structure is caused by volume change of the fill. Circumstances in which this will not be the case, and in which the system will be inappropriate, include the following:  where natural ground underlying the fill undergoes significant movements because of the weight of the structure  where natural ground underlying the fill undergoes significant movements due to causes other than the weight of the structure, for example mining subsidence  where soft saturated clay fill deforms under the

  

structural concern/distress collapse at 1/20 at 1/50 structural noticeability at 1/200.



In the light of this, it is suggested that, for de sign purposes in a typical case, a suitable limit for an acceptable tilt is 1/500. A tilt of this magnitude would not normally be noticed and would have a substantial margin against the tilt where remedial measures might be considered desirable.

10.2 Classification of potential movement In applying knowledge about the causes and mechanisms of fill behaviour to practical problems, it is important to have some system of classification to assist in the recognition of similar types of behaviour and the grouping together of situations presenting comparable hazards. Fills can be classified in a variety of ways including the following physical properties and factors relating to the history of the fill:  nature of fill material (eg natural soil or rock, waste material, biodegradable)  size of fill material (eg coarse or fine)  srcin and method of deposition (eg compacted in layers, end tipped in high lifts, hydraulic placement)  stress history (eg normally consolidated or overconsolidated). Perhaps the most useful type of classification is related to

foundation loading in ancondition essentially undrained constant volume plastic where movement occurs due to slope instability .

The hazard classification summarised in Table 17 considers the maximum potential vertical compression under given loading conditions, and is therefore a classification of a fill associated with a particular form of construction. It is related mainly to movement occurring after construction, because this is a more likely cause of distress, but construction movement has some effect and so is also taken into account. Category A: small movements This category refers to situations in which the maximum potential vertical compression of the fill subsequent to construction at any location is not greater than 0.5% and also not greater than 1% during construction. This type of situation will principally occur with engineered fills (for example a granular fill that has been placed under controlled conditions and received Table 17 Preliminary hazard class ification Vertical compression Vertical compression subsequent to construction durin g c on st ruc t io n < 0.5% 0.5% – 2.0% > 2.0% <1.0%

4.0% – 1% >4.0%

A B

B B

C

C C

C

C

10.3 Shallow foundations adequate compaction). Such materials usually form good foundations, and movements are generally likely to be too small to present major problems unless the building is particularly sensitive to foundation movement. Category B: significant movements This category refers to situations in which the maximum potential vertical compression of the fill is larger than for Category A situations; the potential vertical compression at any location subsequent to construction is not greater than 2% and also not greater than 4% during construction. Many non-engineered fills will give rise to this

103

In addition large differential movements are likely where there is extreme heterogeneity in the fill and at the edge of filled areas.

10.3 Shallow foun dations The code of practice for foundations, BS 8004 (BS I, 1986), defines shallow foundati ons as being less than 3 m deep. Thus in a shallow fill (<3 m deep) a shallow foundation is all that is required to found the structure on the underlying natural stratum. With medium-depth (3 m to 10 m) and deep (>10 m) fills, either deep foundations

situation. The fill is a relatively poor foundation material and foundation movements are likely to be of major concern and have serious consequences for most types of building on the fill. A granular fill that has been placed without systematic compaction, but has little organic matter within it, whose vulnerability to collapse compression is quite low and which has already been in place for some years, could come into this category. In this situation special attention needs to be given to foundation design. If piling is considered to be uneconomic, the basic alternatives are either to use some ground treatment technique to improve the load- carrying properties of the fill and effectively produce a Category A situation, or to design the foundations to withstand the differential movements caused by settlement of the fill. Reinforced concrete rafts with edge beams have commonly been used for two-storey dwellings. It should

are required if the structure is to be founded on the underlying natural stratum, or a shallow foundation is adopted and the structure is founded on the fill. Where small structures are to be built on medium-depth or deep fills, the only economic solution may involve founding the structure on the fill. It should be ensured that:  bearing capacity of the fill is adequate  settlement under working load will not damage the structure  settlement due to causes other than building loads will not damage the structure.

be realised, however, large differential settlements can occur,that verywhere substantial foundations may be needed and building units should be kept small and simple in plan. Even if stiff foundations prevent structural damage, tilt could be unacceptably large.

construction on the fill should be considered. treatment methods, when appropriately used,Ground should limit and control settlement; they will not normally eliminate settlement (Chapter 8). Many types of ground treatment will not improve the full depth of a deep (>10 m) fill, but they may stiffen the upper part of the fill sufficiently to prevent excessive differential settlement. The following basic approaches to shallow foundation design may be adopted:  construct shallow foundations on fill which has been sufficiently improved by ground treatment so that deformations are acceptably small  design shallow foundations which are sufficiently stiff to withstand large deformations in the fill; this is not difficult with small buildings, but tilt may still be a problem  design shallow foundations and structures which are sufficiently flexible to tolerate large deformations in the fill — methods of articulated construction, suc h as the CLASP system, have been used in mining subsidence areas (Heathcote, 1965).

Category C: very large movements This category refers to situations in which the maximum potential vertical compression of the fill is so large that it cannot be classified as Category A or Category B. The potential vertical compression at any location is greater than 2% subsequent to construction or greater than 4% during construction, or both. In this category movement is likely to present problems for construction on the fill which are so severe that some types of construction may not be technically or economically feasible. Ground improvement techniques may be quite limited in what they can achieve. The following potential causes of large post-construction movements should be noted:  collapse compression of poorly compacted partially saturated fills on wetting (Chapter 5)  reduction in volume due to decay and decomposition of biodegradable wastes (section 4.5)  expansion caused by chemical reactions (section 4.6)  compression following liquefaction of loose saturated sandy fills (sections 3.6 and 4.7)  high compressibility of fine-grained hydraulic fills (sections 3.4, 4.1 and 4.2)

The preliminary hazard classific ation presented in the previous section defined three different levels of hazard on the basis of the potential vertical compression that could occur in the fill. Where fill is vulnerable to significant movements, ground improvement prior to

In practice, solutions may comprise both ground treatment and special foundation design. Fill which is susceptible to collapse compression on wetting is of particular concern. Several ways of dealing with the problem are listed below and a combination of these approaches may often be the most suitable:  eliminate susceptibility to collapse by pre- inundation (not always feasible)

104

Chapter 10 Foundations on fills

eliminate or greatly reduce susceptibility to collapse by some other form of ground treatment, for example preloading or dynamic compaction (depth of effectiveness may be limited)  prevent inundation of the fill during life of structure (this cannot be guaranteed)  design foundation and structure to survive collapse movements (difficult if potential movements are large).

with flexible connections may be sufficient. In more severe cases it may be ne cessary to use flexible pipes, or to carry services on piles be aring on a firm stratum beneath the fill. Falls given to drains should be adequate to reduce the risk that settlement will cause backfalls.



Shallow foundations are sometimes termed ‘spread’ foundations to distinguish them from deep foundations. Spread foundations comprise:  

strip foundations pad foundations  raft foundations Reinforced concrete rafts with edge be ams, sometimes termed semi-rafts or stiffened e dge rafts, have commonly been used on non- engineered fills. They have usually been designed so that a certain length will act as a cantilever over a potential void. However, where large differential movements may occur, very substantial foundations may be required and tilt may still be unacceptably large. Atkinson (1993) and Tomlinson (1995) have de scribed typical designs. In foundation design it is important to distinguish between settlement attributable to the weight of the building and settlement attributable to other causes such as self-weight of the fill or collapse compression. Th e concept of bearing capacity based on the premise that the settlement resulting fromisthe weight of the building is the critical factor. With small structures on deep fills, almost invariably settlement attributable to causes such as self-weight of the fill or collapse compression will predominate and, consequently, bearing capacity can be a misleading concept. (Allowable bearing pressure is defined as the maximum allowable net loading intensity at the base of the foundation, taking into account the ultimate bearing capacity and required margin against failure, the amount and kind of settlement expected and the ability of the structure to accommodate this settlement.) Founda tion design should be based on an assessment of the magnitude of differential movements of the fill subsequent to construction on it. It is nece ssary to identify first the cause of se ttlement. The small scale of many developments makes it difficult to ensure that there is an adequate investigation of the fill. The problems associated with shallow foundations for low-rise buildings on filled ground can be minimised by:  avoiding building across the edges of filled areas where the structure would be partly founded on fill and partly on undisturbed natural ground  restricting construction to small units and not building long terraces of houses. The relative movements between the building and services entering it, and betwee n various sections of the services, merit careful consideration. If the movements are not likely to be large, the use of short lengths of pipes

10.4 Deep foundations Where a large structure is to be built on shallow or medium-depth fill, poor load- carrying characteris tics of the fill can be circumvented by using piled foundations and a suspended floor. The piles will be end- bearing, their function being to transmit the loads applied by the weight of the building down through the fill to a competent stratum underlying the fill. Digest 315 (B RE, 1986) gives general guidance on different piling systems and on the choice of an appropriate type of pile. Vibrator-formed compacted concrete columns have been introduced into the UK. Similar to vibro stone columns, vibro concrete columns are formed by pumping concrete down a tremie tube to backfill a cylindrical hole as the poker vibrator which has formed the hole is withdrawn (section 8.2). There may be problems where driven piles hit obstructions. If the depth of fill is not accurately known at the pile location, it may be incorrectly assumed that the pile has reached natural ground. Case history 7 describes the settlement of a factory built on piles which did not penetrate through the full depth of the backfill and where damaging settlements occurred.for down-drag (negative The piles should be designed skin friction) caused by settlement of the fill. The downdrag can be calculated in terms of e ffective stress (Burland, 1973): τ = σ'vK tanδ

where τ = shaft friction σ'v = vertical effective stress K = earth pressure coefficient δ = angle of friction of the fill on the pile surface The appropriate value of K tanδ depends on the pile type as well as the type of fill and may be difficult to estimate. Atkinson (1993) gives some information. For a loose granular fill a value of K tanδ = 0.2 might be appropriate for an initial estimate. For a group of piles, an upperbound to the down-drag force may be calculated from the effective weight of the fill causing settlement, making due allowance for possible changes in ground-water pressure caused by ground-water lowering. In some situations it may be considered that down-drag is taking too large a proportion of the allowable bearing capacity of the pile. The use of a slip coating can greatly reduce its effect. Special attention is needed in the design of services which span from the filled ground into buildings founded on piles. In a fill in which methane gas is being generated by the decay and decomposition of organic matter, piles could form paths for the escape of the gas.

10.5 Implications of ground chemistry

105

The effect on load-carrying characteristics of volume change due to chemical reactions in fills has been examined in section 4.6. Some other types of problem related to ground chemistry are now mentioned. In general, these problems are not confined to filled ground. The interaction between foundation building materials and chemically aggressive ground is a major subject. Hawkins (1998) has examined the engineering significance of sulfates in dark pyritic mudrocks. Materials at risk include concrete, mortars, metals,

effects if coating materials used to protect piles from one contaminant were themselves attacked by another. Damage has generally been found on the tops of exposed piles or foundation walls. The form of concrete is important in determining the risks of attack. Slender sections are more at risk than massive foundations. Modern housing, which tends to be of lighter construction, shoul d be carefully designed and specified. Slabs and floors on the ground are at risk especially where they can dry from the top, which encourages the movement of contaminants into the concrete from the ground.

plastics and masonry . General guidance on the performance of building materials used in contaminated land has been given by Paul (1994), Garvin et al (1999) and Environment Agency and NH BC (2000). Particular hazards are associated with sulfate attack on concrete as described in BRE Special Digest 1 ( BRE, 2001) and corrosion of buried metals (King, 1977, 1989). The durability of building materials used in contaminated land has been reviewed by Nixon et al (1979). Aggressive ground conditions are likely to be encountered on former industrial sites including:  extensive, highly acidic conditions associated with spillages of acid such as may be found around the pickling tanks of a steel works  mixtures of organic chemicals, and a variety of salts and other chemicals, that may attack any of the wide range of materials typically used in construction.

The attack on a material buried in the ground depends not only upon the absolute aggressiveness of the substances or organisms present, and on the amount and concentration present, but also on its availability. Where the aggressive agent is consumed by reaction with the material under attack, the rate of attack will be limited if the immediate area is depleted of the aggressive agent and the supply is not replenished. It may be expected that the potential for attack of a water-soluble aggressive agent such as sulfate ions would be greater in a highly permeable sandy soil than in a relatively impermeable clay soil. This is the reverse of the potential for corrosion of buried steel, for which sandy soils provide the preferable environment (King, 1977). The movement of ions through soils is not, however, related only to permeability but also to the ability of the soil to bind the ions by various physico-chemical processes, the

The effects of sulfates, acids and chlorides give rise to concern for both unreinforced and reinforced concrete. Magnesium, ammonia and phenol are also known to cause deterioration of concrete. High-quality dense concrete is the primary prerequisite for durability and there is no substitute for quality of the materials used. The consequences of chemical attack on concrete are not easy to assess. Ultimately, chemical attack may cause the collapse of a structure, but such extreme examples are not often found. More commonly, localised deterioration and cracking will occur resulting in loss of strength. Chemical attack below ground is difficult to assess due to poor accessibility. W ith concrete piles there is the possibility that sulfate attack , and other forms of attack, could adversely affect the soil/pile interface so as to reduce bearing capacity in service. There could be similar

existence of concentration and otherand factors. The permeability of soils togradients, organic solvents alkaline solutions will not be the same as their pe rmeability to water, and may be considerably higher. A mixture of substances may be pre sent, concentrations may be uneven and the ground poorly consolidated. All these factors are likely to enhance rates of corrosion and make the selection of appropriate materials, protective systems (for example cathodic protection and applied coatings) or other means of overcoming the problem more difficult. For example, an organic coating that might ordinarily be used to protect against sulfate and acid attack on concrete, may itself be attacked by organic substances that are also present. Information on the performance of a material under particular conditions may be lacking. Gasworks sites are typical of these complex situations.

10.5 Implications of gr ound chem istry

107

Part IV: Performance of fills Case histories There is a tendency among the young and inexperienced to put blind faith in formulae, forgetting that most of them are based on premises which are not accurately reproduced in practice, and which inany case are frequently unable to take into account collateral disturbances which only observation and experience can foresee and common sense provide against. (Terzaghi, 1939)

This warning, given to geotechnical engineers by Karl Terzaghi in the 45th James Forrest Lecture, repeated the warning given to structural engineers nearly 50 years earlier by Dr William Anderson in the 1st James Forrest Lecture. It is particularly relevant to the behaviour of non-engineered fills. Their heterogeneous nature means that sophisticated analysis will rarely be worthwhile and the major emphasis should be on a realistic characterisation of properties and identification of hazards. Observation and experience are essential. A major part of the BRE research programme has been concerned with monitoring settlement at filled sites, and in Part IV of this book brief case histories are presented. At most of these sites observations were made by BRE but a few other cases have been included to give a more comprehensive view of fill behaviour. To aid rapid identification of relevant sites, the index tables (Tables 18 and 19) give fill type and type of ground treatment involved in the case histories.

Table 18 Index of cas e histories by fill typ e F i ltly p e Ca s eh i s t o r i e s Opencastminingbackfill

1 2 3 4 5 6 7 8

Colliery spoil

9 10 11

Pulverisedfuelash

12

Industrial and chemical wastes Urban fill

13 14 15 26 28

Domesticrefuse

161 71 81 92 02 12 2

Infilled docks,pits andquarries

11 23 24

Hydraulic fill

12

Engineered fill

8

Table 19 Index of case histories by ground treatment method Trea t m en tm e t h o d Ca s eh i s t o r i e s Dynamiccompaction

2 16 17 18 26

Vibrotechniques

152 72 8

Preloading

2 11 12 19 20 23 24

Pre-inundation Engineered fill

2( 45

7)*

8

* Although inundation was not adopted as a treatment technique at these sites, it did occur and provided useful information

108

Case history 1: Corby (A) Case history 1

Opencast mining backfill at Corby (A) Fill material

Boulder clay overlying oolitic limestone

Method of placement Age of fill

Large walking dragline

Depth of fill

18 m

Water table Ground treatment

None in fill so far as is known

Development

Low-rise housing

Date of development

1963

Period of monitoring Reference

1963 to 1972

1951

None

Penman and Godwin (1974)

Description of project

The expansion of Corby involved housing and industrial developments on land previously worked for ironstone by opencast methods. In 1963, 24 experimental semi-detached houses were built using four types of foundation:  type A — concrete strip footings  type B — concrete raft with fabric reinforcement on mass concrete edge beam  type C — concrete raft with fabric reinforcement on reinforced edge beam  type D — concrete raft with fabric reinforcement on deep reinforced edge beam. Fill properties

Opencast backfill similar to case history 2. Field measurements

The settlement themeasurements houses has been monitored. L evelling stations were installed and theoffirst made in July 1963. The houses were virtually complete by November 1963. Large differential movements were measured although damage was limited. Houses settled up to 180 mm in ten years. Figure 54 shows the maximum differential settlement measured during the period from 1963 to 1972 at a pair of semi-de tached houses which had traditional strip foundation s. The levelling stations 45 and 46, 47 and 48 were installed to measure the differential settlement across the end walls. It is seen that most of the settlement occurred within two years of the end of construction. A wide range of settlement was measured at the site but no correlation was found between settlement and type of foundation.

Figure 54 Maximum differential settlement of houses at Corby (A)

Case history 2: Corby (B) Case history 2

109

Opencast mining backfill at Corby (B) Fill material

Boulder clay overlying oolitic limestone

Method of placement Age of fill

Large walking dragline

Depth of fill

24 m

Water table Ground treatment

None within fill

Site restored 1970

Dynamic compaction Preloading

Inundation Date of ground treatment 1975 Development Low-rise housing

Date of development

1975

Period of monitoring

From 1974

References

Charles et al (1978), Burford and Charles (1991)

Description of project

The experimental site (see Figure 2, page 8) was divided into four 50 m × 50 m square areas. Three different types of ground improvement were used:  dynamic compaction using a 1 5 tonne weight dropped from 20 m  pre-inundation via 1 m deep trenches at 10 m centres  preloading with a 9 m high surcharge of fill. Houses were built on each of the three areas of treated ground and also on an untreated area. Trench fill foundations were 375 mm × 900 mm deep with the top of the concrete 75 mm below ground level. Fill properties

The upper shear part ofstrength the opencast backfill was Moisture a very stiffcontents clay fill with anbetween undrained of about 100 kPa. varied 7% and 28% with a mean value of 18%. The corresponding range of dry densities was from 1.5 Mg/m3 to 1.8 Mg/m3 with a mean value of 1.70 Mg/m3. A particle size analysis indicated that 54% was finer than 0.075 mm and the clay fraction was 19%. Typical values for plastic and liquid limit were 17% and 28% respectively. The lower part of the backfill consisted mainly of oolitic limestone. Field measurements

Field settlement was monitored using magnet extensometers. The average surface settlement produced by the surcharge was 410 mm, by dynamic compaction 240 mm and by inundation 100 mm. Dynamic compaction A 50 m × 50 m square area was treated using a 15 tonne weight with a base area of 4 m 2. At each point on a 10 m grid the we ight was dropped repeatedly from a height of 20 m using a heavy crawler crane. More than te n blows were usually required, and a hole some 2 m de ep was produced. Each hole was backfilled with the surrounding material. The second stage of treatment was to repeat the process, using the same compaction at the same grid points. In further stages, compaction was applied on a grid offset by 5 m from the srcinal grid. There was a final general tamping of the whole area with a reduced fall of the weight. The average energy input was 2800 kNm/m 2. The average settlement with depth profile induced by ground treatment is shown in Figure 55, which indicates that an average vertical compression of 4% was produced in the top 5 m to 6 m of fill with negligible effect on the fill below 6 m. The settlement of the fill sub sequent to dynamic compaction is sh own in Figure 56. The settlement of the houses built on the treated ground has continued at the same rate as the settlement of ground which has no structural

110

Case history 2: Corby (B) load on it, and corresponds to α = 0.5% in the upper 6 m of fill affected by dynamic compaction. Inundation Five 50 m long and 1 m de ep trenches (see Figure 47, page 88) were dug at 10 m centres. They were first filled with water in Februar y 1975 and were backfilled in June 1975. T he average surface settlement induced by this inundation was 0. 1 m. During the first ten days of this inundation experiment, about 90 m 3 of water was absorbed by the backfill, and comparatively little was absorbed subsequently. About half of this volume was lost from one trench. The largest settlement was measured at a magnet extensometer which was only 2 m from this trench. Figure 57 shows settlement induced by

inundation at this location. Settlement continued after the trenches had be en backfilled. Figure 58 shows settlement versus depth measured in August 1975 shortly after the trenches had been backfilled. It is helpful to consider four phases of movement behaviour which can be illustrated by the maximum settlement which was measured at the gauge shown in Figure 57: 1 prior to the inundation test (23.5.74 to 18.2.75) — the fill surface settled at about 1 mm per month 2 during the inundation test (18.2.75 to 26.8.75) — the fill surface settled 165 mm in six months, princip ally as a result of vertical compression of 5.6% between 2.3 m and 4.5 m depth below ground level 3 during the six years subsequent to the end of the inundation test (26. 8.75 to 19.11.81) — the fill surface settled 118 mm, mainly as a result of compression of 1.8% between 4.5 m and 12.1 m below ground le vel 4 during the following nine years (1 9.11.81 to 15.11.90) — the fill surface settled 25 mm, principally because of small compression between 4.5 m and 12.1 m below ground level. Preloading The 9 m high surcharge was placed over a three-week period, left in position for a month and then re moved. Most of the settlement occurred as the surcharge was being placed (Figure 59). A small amount of heave occurred as the surcharge was removed. The settlements measured at different depths within the backfill are plotted in Figure 60. The stresses produced by preloading were much greater than those subsequently applied by foundation loads. The surch arge was effective down to a de pth of 10 m. In the upper part of the clay fill the compression produced by the surcharge indicated D = 5 MPa for the untreated fill. Settlement of houses The settlement of the houses built on the four areas was monitored by precise levelling of settlement stations installed at damp-proof course level . Figure 61 shows settlement versus time for the maximum, minimum and typical settlement measured at levelling stations on houses built on e ach of the four areas. Table 20 summarises the movements during and subsequent to construction. Maximu m and minimum settlements of levelling stations on each area are listed as is the maximum differential settlement of a house measured between any two levelling stations. Settlement has been smallest in the preloaded area and the performance of the houses built on the preloaded opencast backfill has been very satisfactory . The settlement of the houses built in the untreated backfill was surprisingly small. Houses built on pre-inundated ground have settled most. Thus, inundation from the surface via trenches proved unsuccessful as a means of adequately pretreating the ground. However, the experiment did clearly de monstrate that water penetrating into the backfill from the ground surface could be a major hazard to construction on the site. The maximum tilt has occurred on the inundation area where one of the buildings has a tilt of 1 in 110.

111

Case history 2: Corby (B) Table 20 Settlement of experimental houses at Corby (B)

Total settlement during and after Sett le me nt du ri ng c on st ru ct io n ( mm ) G ro u n dt re a t m e n t t ech ni que Preloading Dynamic compaction Inundation No treatment

Me a n 1.4 7.0 6.1 2.7

Ma x 3.0 9.2 14.3 6.8

Mi n –0.4 3.2 2.8 1.4

co ns tr uc ti on t o 19 99 ( mm )

Ma x d i f f e re n t i a l 2.3 5.9 6.8 2.8

Ma x Me an 11

52 54 33

Ma x

Mi n

25 74 149 53

5 23 30 14

d i f f e re n t i a l 14 27 91 32

Figure 55 Settlement at depth induced by dynamic compaction at Corby (B)

Figure 56 Settlement subsequent to dynamic compaction at Corby (B)

112

Case history 2: Corby (B)

Figure 57 Maximum settlement induced by inundation at Corby (B)

Figure 58 Settlement at depth induced by inundation at Corby (B)

Figure 59 Maximum settlement induced by preloading at Corby (B)

Case history 2: Corby (B)

113

Figure 60 Settlement at depth induced by preloading at Corby (B)

Figure 61 Settlement of houses at Corby (B)

114

Case history 3: Corby (C) Case history 3

Opencast mining backfill at Corby (C) Fill material

Boulder clay overlying oolitic limestone

Method of placement Age of fill

Large walking dragline

Maximum depth of fill

24 m

Water table Ground treatment

None within fill

Site restored 1970

Temporarily preloaded with 7 m high surcharge

Date of ground treatment 1989 Development

Low-rise housing

Date of development Period of monitoring

1989

Reference

Burford (1991)

1988 to 1990

Description of project

A large housing development has been built on opencast backfill close to the area described in case history 2. Fill properties

The upper 12 m of fill consisted of boulder clay with fragments of oolitic limestone. Moisture content varied between 9% and 18% with a mean value of 15%. Dry density varied from 1.68 Mg/m3 to 1.87 Mg/m3, with an average of 1.81 Mg/m3. Typical values of plastic and liquid limit were 16% and 34% respectively. T he lower part of the backfill consisted mainly of oolitic limestone with fragments up to boulder size. Field measurements

Settlement during and subsequent to preloading has been monitored by twocaused magnetsurface extensometers. Preloading settlements of 300 mm. Figure 62 shows induced settlement versus depth and Figure 63 presents the same data as vertical compression versus depth. Behaviour was generally similar to that observed with preloading at case history 2.

Figure 62 Settlement at depth induced by preloading at Corby (C)

Figure 63 Vertical compression induced by preloading at Corby (C)

Case history 4: Horsley Case history 4

115

Opencast mining backfill at Horsley, Northumberland Fill material

Mudstone and sandstone fragments

Method of placement Age of fill

Truck and shovel, dragline

Depth of fill

70 m

Water table Period of monitoring

34 m rise between 1974 and 1976

References

Charles et al (1977, 1984, 1993)

1961–1970; restoration completed in 1973

1973 to 1992

Description of project

The effect of a rising water table on the settlement of an opencast coal mining backfill which had not been systematically compacted has been observed at this experimental site. The land was returned to agricultural use after the completion of backfilling and, so far as is known, no building development is planned. Fill properties

The average dry density of 100 mm diameter open-tube samples was 1.70 Mg/m3, the average moisture content was 7%, and the density index was about 0.6. The mean SPT N value was 29, confirming that the fill was of medium density. Typically about 10% of particles were finer than 0.075 mm and about 50% coarser than 2.36 mm. The fill was variable. Field measurements

The following types of measurement have been made:  settlement at different depths within the fill was monitored using five magnet extensometers installed in boreholes drilled through the full depth of fill settlement of the fill was monitored by precise levelling of traverses surface of surface settlement stations  ground-water level was monitored in standpipe piezometers. 

Settlement measured between 1973 and 1991 at the five magnet extensometers is given in Table 21. At these locations settlement has been measured by both magnet extensometers and precise le velling; the agreement is good at most locations, but not everywhere. The settlement at E12 is also listed; there was no magnet extensometer at this location, but it is here the maximum settlement has been recorded. Figure 64 shows the surface settlement measured by precise levelling along the various traverses during the period 1973–1983 when most of the settlement occurred. The maximum movement measured at any location within the site is at surface settlement station E12 where the fill has settled 0.8 m. The smallest movement measured at a magnet extensometer has occurred at gauge D1 and this can be attributed to the effect of preloading by a 30 m high overburden heap. The settlement at gauge C1 1 has also been small but it should be noted that, prior to the rise in ground-water level, the settlement rate at this gauge was greater than that in the other parts of the site. It may be that in the ye ars before monitoring started, the settlement of the lagoon area, which is composed of a wet and more cohesive fill, was large. The settlement versus depth profiles for four of the magnet extensometers are plotted in Figure 65. The largest vertical compression measured at any of the magnet extensometers is 3%. The settlement of the backfill during four successive periods shown is summarised in Table 22.

116

Case history 4: Horsley Period A: December 1973 to April 1974 In the few months prior to April 1974, settlement was monitored while pumping kept the water level down in the bedrock below the backfill. Considerable settlement may have occurred before the start of monitoring in December 1973. The rate of settlement was greatest at gauge C11, which had been the site of a lagoon during opencast mining. In the four months of monitoring during this period, the settlement at ground level measured at this gauge by precise levelling was 17 mm. The magnet extensometer indicated that compression was occurring over the full depth of the fill. The backfill in the locality of gauge D1 previously had bee n loaded by a large overburden heap with a maximum height above restored ground level of

30 m. This had been removed two years before the measurements began. Four months of precise levelling prior to April 1974 showed a heave of 10 mm at ground level. The magnet extensometer indicated that this movement was caused by expansion at the base of the fill. At gauge B2, close to the pump, the ground-water level varied between 5 m and 10 m above rockhead during this initial period in which pumping continued. The settlement measured by precise levelling was 2 mm during the four-month period. Period B: April 1974 to April 1977 In the three years from April 1974 to April 1977, ground movements were monitored while the ground-water level rose 34 m through the backfill. Th e ground-water level at gauge B2 rose by 20 m between April 1974 and April 1975, 9 m in the following 12 months and 5 m in the 12 months after that. From June 1975 onwards the water level measured in the five borehole gauges has been virtually the same height above OD. Having reached a new equilibrium water level in April 1977 at about 83 m AOD, subsequent

fluctuations in water level have been 1978. small.The A maximum ground-water 84 m AOD was recorded in August final equilibrium level oflevel the of ground-water level in the ope ncast backfill appears to have been controlled largely by the topography of the site. At gauge B2, where the fill was deepest, 0.33 m settlement occurred at the ground surface during this three-year period as the water level rose 34 m. The effect of the rising ground-water level at gauge B2 is shown in Figures 66 and 67. In Figure 66, the settlement at different depths within the fill and the ground-water level are plotted against time. The same information is presented as vertical compression at different depths within the fill in Figure 67. As the ground-water level rose, vertical compressions locally were as large as 2% but the average settlement measured over the full depth of inundated backfill was smaller than 1%. Although most of the measured settlement can be attributed to the rising ground-water level, maximum compress ion has occurred between magnets 10 and 11, well above the equilibrium water level. Collapse compressions were much smaller in the area of the backfill which, during opencast mining, had been temporarily preloaded by a 30 m high overburden heap. The rise in water level saturated 31 m of the backfill at the location of gauge D1 and a surface settlement of 0.1 m was observed. In the lower part of the preloaded backfill, which was satur ated by the rising groundwater table, the average collapse compression was less than one-fifth of the collapse compression in the ground which had not been preloaded . Thus although the preloading did not eliminate collapse compression, it substantially reduced it . At a location at which there had been a lagoon during opencast mining (gauge C11), the fill had been effectively pre-inundated. Consequently the rising ground-water table had very little effect on the settlement of the backfill, although the wet and more cohesive fill at this location may have suffered large settlements in the years immediately following backfilling. Precise levelling recorded 61 mm settlement during this period. Settlement

117

Case history 4: Horsley

occurred fairly uniformly through the full depth of the backfill. The rising ground-water level had little effect on settlement at gauge D15 because the gauge is situated on high ground and only the bottom 11 m of the backfill have been inundated (Table 21). Between April 1975 and April 1977, the ground-water inundated the bottom 11 m of the backfill and there was some increase in the rate of settlement. Period C: April 1977 to April 1981 During the four-year period immediately following the rise in ground-water level, ground movements were still affected by the rise in water level. At gauges A9, B2, D15 and E1 2, the rate of settlement was be tween three times and ten times as large as the rate prior to the rise in ground-water level. Large

movements continued at B2, but Figure 66 shows that the continuing settlement was largely caused by compression of the fill above the groundwater level. Precise levelling has shown that the greatest movement during this period occurred at E12 where 0.29 m settlement was observed. The cause of such large movements at this location is not known. Period D: April 1981 to December 1992 Ground movement monitoring continued to December 1992. This 1 1-year period started several years after the rise in ground-water level had ceased. At most locations, the rate of settlement during this period has been smaller than the rate prior to the rise in ground-water level. The only magnet extensometer that was an exception to this was gauge D15 in the most recent fill where the settlement rate has bee n greater than at the other gauges. At the other borehole gauges the rate of movement was 3 mm per year or smaller. The rate of movement at E12 is greater than at any of the five borehole gauges.

Table 21 Settlement at Horsley 1973–1991 Ma g n e t

D a t eo f

e x t e n s o m e t er

bac k fi l l i ng

Fill F i l lc o n di t i o n

Surface settlement M a gn e t Pr ec i s e

S u b m e rg e d

d e p t h( m )

d e p t h( m )

ex t en s o m et e r

A9

1961

Oldest

61

46

0.39

0.39

B2

1964

Deepest

63

45

0.50

0.50

C11

1965

Lagoon

46

35

0.12

0.12

D1

1966

Preloaded

56

31

0.06

D15

1970

Most recent

47

11

0.29

0.31

E12

1966

Intermediate age

48

17

—

0.78

l e v el l i n g

0.09

Table 22 Surface settlement measured by precise levelling at Horsley 1973–1992 P er i o dA Pe r i o dB Pe r i o dC P er i o dD Dec 73 – Dec 73 – Apr 74 Apr 74 – Apr 77 Apr 77 – Apr 81 Apr 81 – Dec 92 De c 9 2 To t a l Ra t e To t a l Ra t e To t a l Ra t e To t a l Ra t e G a u ge ( m m ) (mm) (mm/ y) ( m m) (mm/y) (mm) ( m m/ y ) (mm) ( mm / y ) A9

397

002

005

310

103

057

014

028

002

B2

499

002

006

331

110

133

034

033

003

C11

109

017

043

061

020

015

004

016

001

D1

095

–010

–021

098

033

006

002

001

000

D15

307

001

002

152

051

081

021

073

006

E12

788

008

017

354

118

288

074

138

012

118

Case history 4: Horsley

Figure 64 Surface settlement of opencast backfill at Horsley, 1973–1983

Figure 65 Variation of settlement with depth at Horsley, 1973–1992

Case history 4: Horsley

Figure 66 Settlement of deepest opencast backfill at Horsley

Figure 67 Vertical compression of deepest opencast backfill at Horsley

119

120

Case history 5: Ilkeston Case history 5

Opencast mining backfill at Ilkeston Fill material

Clay

Method of placement Age of fill

Scraper

Depth of fill

12 m

Water table Ground treatment

None within fill

Development

Two-storey blocks of eight houses

Date of development

1972

Period of monitoring References

1974 to 1981

1959

None

Guest (1974), Charles and Burford (1987)

Description of project

The brickwork for a block of eight two -storey houses was completed in February 1973. The block had a movement joint in the middle, and was built on a concrete raft with edge beams. Early in May 1973, when excavation for drains began close to the north gable end, a cracked floor and beam we re observed. Following heavy rain in June, movement took place in the ce ntre of the row of houses. In July it was reported that all the houses in the block were affected. Underpinning and pressure grouting were carried out but movements continued. At the end of January 197 4 floor levels showed a maximum differential settlement of 0.14 m across the 9 m wide block and the east wall was 0.065 m out of plumb. BRE was asked to investigate the reasons for the settlement of the houses, which had been built on a site with a slope of about 1 in 15. It was suspected that water penetrating into the fill through drain trenches had caused collapse compression within the backfill. The block was never occupied and was demolished in 1982. A total settlement of 0.3 m was e stimated. Other blocks on the site have also been demolished. Fill properties

The opencast backfill was predominantly a clay fill with, typically, a plastic limit of 23% and a liquid limit of 41%. The moisture content of samples varied between 12% and 25% with an average value 4% b elow the plastic limit; cu = 150 kPa approximately. The mean dry density of 100 mm diameter open drive samples was 1.78 Mg/m3. Field measurements

An inundation test was carried out in 1975. Trenches 2 m × 1 m × 3 m deep were excavated and supported by vertical steel sheets and struts. Into each trench 3.5 m 3 of water was poured. The trenches were filled to a depth of 1.8 m with water. The rates at which water levels fell in the different trenches varied from as little as 0.04 m/hour (0.08 m3/hour) to as much as 1 m/hour (2 m3/hour). Within 24 hours of filling the trenches with water, additional settlements of up to 50 mm had been recorded confirming that water penetrating into the ope ncast backfill via surface trenches could cause significant collapse compression. Figur e 68 shows the settlement measured at various depths within the fill at a magnet extensometer installed about 1.5 m from two of the trenches. Compression occurred immediately water was put into the trenches and was located between 0.5 m and 7 m below ground level at this gauge. Six days after the start of the test 40 mm of settlement had occurred and the trenches were backfilled. Settlement continued, however, at a significant rate. There was a temporary increase in the rate of settlement in late 1976 following a period of heavy rainfall.

Case history 5: Ilkeston

Figure 68 Inundation test on opencast backfill at Ilkeston

121

122

Case history 6: Tamworth Case history 6

Opencast mining backfill at Tamworth Fill material

Clay with shale fragments

Method of placement Age of fill

Scraper

Depth of fill

32 m

Water table Ground treatment

Variable within backfill

1972

Temporarily preloaded with 7 m high surcharge

Date of ground treatment 1995 Development

Low-rise housing

Date of development Period of monitoring

1995 to 1999

Reference

Charles and Burford (1987)

1977 to 1995

Description of project

This opencast mining bac kfill site has been investigated in order to assess its suitability for building development. Fill properties

Moisture contents ranged from 4% to 18% with a mean value of 9%. Dry densities ranged from 1.3 Mg/m 3 to 2.1 Mg/m3 with a mean value of 1.78 Mg/m3 corresponding to a porosity of 32% and air voids of 16%. Approximately 45% of the fill was in the silt and clay fractions. Field measurements

Settlement was monitored at this sloping site from 1977 to 1995. Figure 69 shows the settlement measured close to the deepest part of the backfill. It is unlikely that the settlement has been solely attributable to creep under the self-weight theincrease backfill:in settlement rate in the period 1979–1981, whereas there wasofan settlement due to creep would show a continuing decrease in settlement rate with time  settlement in the period 1981–1995 interpreted as creep settlement would correspond to α = 5% and this is too large to be credible for a fill of this type. 

It seems likely that the settlement was largely attributable to collapse compression effects although the small rise in ground-water level that occurred during the first three years of monitoring can only account for a small part of the settlement. An inundation test carried out in 1979 using a 30 m long and 2.4 m deep trench produ ced a negative result with n o measurable quant ity of water pene trating into the fill and no settlement occurring. There remains the possibility of water pene trating into the backfill from old unsealed workings and seeping through the backfill. The rise in ground-water level in 1994–199 5 caused an increased rate of settlement. Ground treatment by preloading with a 7 m high surcharge of fill was carried out in 1995 and induced some 0.2 m of surface settlement. This is not shown in Figure 69 as the extensometer did not survive the surcharging. Figure 70 shows vertical com pression plotted against depth below ground level from 1977 to 1995 and then induced by the surcharge loading.

Case history 6: Tamworth

123

Figure 69 Settlement of opencast backfill at Tamworth

Figure 70 Vertical compression of opencast backfill at Tamworth

124

Case history 7: West Auckland Case history 7

Opencast mining backfill at West Auckland Fill material

Clay with shale fragments

Method of placement Age of fill

Not known

Depth of fill

18 m

Water table Ground treatment

Rise in ground-water level through base of backfill

Development

Single-storey factory

Date of development

1971

Period of monitoring References

1977 to 1982

Site restored 1953, benched 1969

None

Leigh and Rainbow (1979), Smyth-Osbourne and Mizon (1984)

Description of project

This opencast coal mining site was restored in 1953 to contours similar to the srcinal land form. The backfill was mainly grey shale with coal fragments. In 1969 the site was levelled in a cut-and -fill earthmoving operation. A singlestorey light industrial factory, which covers an area 100 m × 60 m, was built in 1971. The sheeted roof is supported by steel trusses and stanchions with the stanchions found ed on pile s. Fill properties

Moisture contents varied between 10% and 13% with a mean value of 11%. The range of dry densities was from 1.9 Mg/m3 to 2.1 Mg/m3 with a mean value of 1.98 Mg/m3, corresponding to a porosity of 28%. Particle size distributions indicat ed that about 40% of the fill was finer than 0.075 mm and 30% coarser than 2.36 mm. Tests carried out by BRE on 100 mm diameter open drive samples indicated small collapse compressions with a maximum compression of It only 0.4% compared with 4.5% indicated field sufficient measurements. is considered that the sampling processby caused compaction of the fill to reduce the susceptibility to collapse compression greatly in the laboratory test. Field measurements

Settlement of the south-west corner of the factory was noted early in 1977. Levelling in July 197 7 showed that the settlement extended to over half the factory, with a maximum of 0.21 m at the south-west corner. There was little evidence of differential settlement betwee n stanchion bases, floor slab and outside ground surface. It was ascertained that underground mining activity had begun in 1949 and that the wat er table had been lowered we ll below the base of the opencast mine by pumping. Deep mining had ceased by 1967, and all associated pumping had stopped by 1972. As a consequence, the water table had risen and inundated the bottom of the ope ncast backfill. Th e piles were ineffective in preventing settlement because they did not penetrate through the full depth of backfill. On the assumption that all the settlement was located in the lower inundated part of the backfill, it was calculated that a collapse compression of 4.5% had occurred. A further 2 m to 3 m rise in ground-water level occurred early in 1979, and vertical compressions of 4.7% and 2.4% respectively were measured at two locations in the newly saturated fill.

Case history 8: near Edinburgh Case history 8

125

Opencast mining backfill near Edinburgh Fill material

Mudstone, siltstone and sandstone

Method of placement Age of fill

Dragline and face shovel

Depth of fill

60 m

Water table Ground treatment

15 m rise between 1997 and 1999

1982

Top 16 m of fill systematically compacted

Date of ground treatment During fill placement Development

Dual carriageway trunk road

Date of development Period of monitoring

1985

References

Ferguson (1984), Charles and Burford (1987)

From 1984

Description of project

To reduce settlement of a 1.4 km section of trunk road, the Tranent by-pass, which was to be built across the Blindwells opencast coal mine, the top 16 m of the 60 m dee p backfill was systematicall y compacted. During opencast mining the site has been de-watered and the ground-water level was kept below the bottom of the excavation. Fill properties

Fill properties were determined on 100 mm diameter open drive samples and cannot be representative of the fill as a whole, which included boulders. Moisture contents ranged from 4% to 11% with a mean value of 7%. Dry densities ranged from 1.4 Mg/m 3 to 1.6 Mg/m3 with a mean value of 1.56 Mg/m3 corresponding to a porosity of 38%. The silt and clay fraction was 20%. Field measurements

Magnet extensometers were installed in late 1984. During the first ten years of monitoring, settlements of 0.44 m in fill which had not be en systematically compacted and 0.2 m where the upper zone had been compacted were measured. Typically α values in the uncompacted fill were about 1%. In 1995 construction of a new carriageway led to the destruction of the two gauges on the compacted fill. In 1997 the ground-water level began to rise. Figure 71 shows that there has been an increase in surface settlement of 0.30 m during the period in which the ground-water level has risen by 15 m. Most of this settlement can be related to settlement of the submerged fill and corresponds to a vertical compression of 1.4% in the submerged fill.

Figure 71 Settlement of opencast backfill at Blindwells, 1985–1999

126

Case history 9: Coalville (A) Case history 9

Colliery spoil at Coalville (A) Fill material

Old colliery spoil

Method of placement Age of fill

Scrapers

Depth of fill

13 m

Water table Ground treatment

4 m below ground level

Development

Road

Date of development

1984

Period of monitoring References

1978 to 1995

1978

None

Charles (1984), Skinner et al (1997)

Description of project

A major land re clamation scheme was undertaken, which involved the backfilling of three clay pits described as north, middle and south pits or holes. Water was pumped out of the holes before backfilling began. The north hole was backfilled with colliery spoil from an adjacent waste heap. It was moved by scrapers and placed in thin layers but received no other systematic compaction. The middle and south holes are de scribed in case history 10. Fill properties

The mean value of moisture content was 15% and the mean dry density was 1.68 Mg/m3. The SPT N values ranged from 12 to 56 with a mean value of 30. About 40% of the particles were finer than 0.075 mm and 30% were coarser than 2.36 mm. Field measurements

Generally, movements very small. At ground level measured settlementathas been, typically, about 10have mmbeen but larger movements have been two locations at depth within the fill. The surface settlement measured by one of the magnet extensometers is plotted in Figure 72. Load tests

A loading test was carried out, in which a net bearing pressure of 35 kPa was applied through a 0.9 m × 0.9 m pad. Figure 73 shows settlement plotted against the logarithm of time. Between 1 day and 90 days after starting the test 4 mm of additional settlement occurred. Results are given in Table 23.

Table 23 Load tests on colliery spoil at Coalville (case histories 9 and 10) q b si sα si /b Tes tl o c a t i o n (k Pa) (m) (mm) (mm) mα ( %)

sα /b (%)

qb/si ( M Pa )

North hole*

35

0.9

5.6

2.0

0.35

0.62

0.46

5.6

Middle hole†

35

0.9

8.1

3.8

0.46

0.90

0.42

3.9

South hole† South hole†

35 35

0.9 2.0

10.9 22.8

2.3 6.3

0.21 0.28

1.2 1.1

0.26 0.32

2.9 3.1

* Case history 9 † Case history 10

Case history 9: Coalville (A)

Figure 72 Settlement of colliery spoil at Coalville

Figure 73 Load tests on colliery spoil at Coalville

127

128

Case history 10: Coalville (B) Case history 10

Colliery spoil at Coalville (B) Fill material

Fresh colliery spoil

Method of placement Age of fill

Uncompacted lifts 1.5 m to 2 m high

Depth of fill

15 m

Water table Ground treatment

Towards base of fill

1979

Dynamic compaction

Date of ground treatment 1984 Development

Light industry

Date of development Period of monitoring

1984

References

Charles (1984), Skinner et al (1997)

1978 to 1995

Description of project

Land reclamation in the middle and south holes at Coalville involved freshly mined waste material brought in lorries from a local colliery and tipped in lifts 1.5 m to 2 m high. There was no compaction other than that provided by the lorries running over the surface of each layer. It is understood that at a later stage part of the site where light industrial buildings were to be built was treated with dynamic compaction; this was subsequent to the BRE settlement monitoring described here. Fill properties

The mean moistur e content of the colliery spoil was 16%, and the mean dry density was 1.56 Mg/m3. The range of SPT N values was from 2 to 26 with a mean value of 10. About 40% of the particles were finer than 0.075 mm and 30% were coarser than 2.36 mm. Field measurements

Magnet extensometers were installed before filling had be en completed. At one location in the south hole where there was 6.5 m of fill, 0.47 m of settlement occurred be tween November 1978 and January 1979 corresponding to an average vertical compression of 7%. This appeared to be associated with very heavy rainfall in December 1978 (180 mm). In July 1979 a further 1.7 m of fill were placed at this location and 39 mm of settlement immediately occurred. Settlements of ground surface monitored by magnet extensometers in the middle and south holes, since backfilling was completed in these holes, are plotted in Figure 72. The ground-water table is 7 m and 11 m respectively below the surface of the fill at these two gauges. The sudden increase in settlement in the middle hole in the spring of 1982 was associated with deep mining. The settlement monitored by precise levelling at this location was 0.6 m, whereas the magnet extensometer only measured the settlement relative to the base of the backfilled hole. It can be deduced that most of the mining subsidence occurred below the base of the hole, but this deep- seated movement caused some compression of the backfill. Constrained moduli at different depths within the fill can be calculated for each occasion when further fill was added. This is shown in Figure 74 in which constrained modulus, D, is plotted against the average vertical effective stress in the layer of fill. Figure 74(a) shows results where the fill has been inundated prior to the application of load, and Figure 74(b) shows results where the fill has not been inundated. The following conclusions can be drawn:  there is a good de al of scatter in the results, indicating variability in the fill  there is some increase in stiffness with increasing stress level  fill that has previously been inundated is more than twice as stiff as fill that has not been inundated

Case history 10: Coalville (B) 

129

at a vertical stress of 100 kPa, typically D = 5 MPa in the pre-inundated fill and only 2 MPa in the fill that has not been pre- inundated.

Load tests

The results of loading tests with 0.9 m × 0.9 m pads carried out on the colliery waste backfill in the middle and south holes are plotted in Figure 73. The settlements were larger than those measured in the similar test carried out on the spoil in the north hole. Significantly more settlement occurred when a 2 m × 2 m square pad was loaded to apply the same net bearing pre ssure of 35 kPa. Results are tabulat ed in Table 23.

Figure 74 Constrained modulus of colliery spoil at Coalville

130

Case history 11: Methil Case history 11

Colliery spoil at Methil Fill material

Old colliery spoil

Method of placement Age of fill

Pushed into standing water

Depth of fill

13 m

Water table Ground treatment

2 m below ground level

1978

Temporary preloading with 2 m high surcharge

Date of ground treatment 1978 to 1979 Development

Industrial but none on fill

Period of monitoring Reference

1978 to 1994 Charles (1984)

Description of project

As part of an industrial development a 50 m wide finger of Number 3 dock was infilled. Scrapers were used to transport the fill from heaps adjacent to the dock onto the area being filled. The fill was then dozed into the standing water in the dock. There was little silt in the dock, and the dock bottom was sandstone. It was not practical to de-water the dock and place the fill in layers under dry conditions. A 2 m high surcharge of colliery waste was placed over the infilled section of the dock to pre load the backfill prior to any construction on it. The surcharge was placed in August 1978 and removed in February 1979. Fill properties

The mean moisture content was 12%. Bulk density of the colliery spoil above the water table was 1.73 Mg/m3 and below the water table the submerged density was 0.87 Mg/m3. The SPT N values ranged from 4 to 17, with a mean valuethan of 9.2Typically, about 15% the permeability fill was finer than 0.075 mm and 45% finer mm. A constant headoffield test gave a value of the coefficient of permeability k = 7 × 10–8 m/s for the colliery spoil backfill. Field measurements

Figure 75 shows surface settlement plotted against time. Large settlements occurred before the surcharge was placed and it is difficult to distinguish between the initial settlement caused by the self-weight of the colliery waste immediately it was placed, and the additional settlement caused by surcharging. After the surcharge was removed settlements continued for another year, but little further movement has occurred since the beginning of 1980. Over half the settlement was located in the upper 3 m of the spoil.

Case history 11: Methil

Figure 75 Settlement of colliery spoil in dock at Methil

131

132

Case history 12: Peterborough Case history 12

Lagoon pfa at Peterborough Fill material

Lagoon pfa

Method of placement Age of fill

Hydraulic fill

Depth of fill

11 m

Water table Ground treatment

Seasonal variation 0 m to 4 m below ground level

1971

Preloading with 8 m high surcharge of brickbats

Date of ground treatment 1983 to 1984 Date of development

None on treated fill

Period of monitoring References

1983 to 1985 Charles et al (1986), Humpheson et al (1991)

Description of project

A preload experiment, in which an 8 m high surcharge of brickbat s was placed during June and July 1983, was part of a BRE investigation into the feasibility of building development on the extensive lagoon pfa- filled brick pits (Charles et al, 1986). Subsequently, Ove Arup has carried out a comprehensive investigation over a wider area, including CPT and laboratory testing, field loading and vibration trials, and monitoring of ground-water levels (Humpheson et al, 1991). Following the Ove Arup investigations, a major retail development has been sited on the lagoon pfa without any ground treatment being required. Fill properties

In the BRE investigation it was found that, typically, the lagoon pfa had a moisture content of about 40%, a saturated bulk unit weight of 1.6 Mg/m 3 and a field permeability of 1 × 10–6 m/s. The extensive testing carried outplaced by Ove has indicated that the density index of the hydraulically pfaArup is surprisingly high, probably in the range 0.5 to 0.7. Field measurements

Settlement was monitored while the surcharge was in position and subsequent to its removal in August 1984 (Figure 76). The surface of the pfa settled 0.12 m. Two-thirds of this settlement took place as the surcharge was being placed and 90% of the settlement had occurred within one month of the completion of the surcharge. The bulk density of the brickbats was 10 kN/m3, and so a preload pre ssure of 80 kPa was applied to the pfa. As the submerged density of the pfa was only 0.6 Mg/m 3, the distribution of e ffective vertical stress was very dependent on the ground-water level. This was found to vary seasonally, from ground level to 4 m be low ground level. The water table was falling quite rapidly during surcharge placement and this may have affected the settlement response of the pfa. There appears to have be en compression throughout the full depth of the pfa. Ove Arup has carried out field loading trials (Humpheson et al, 1991). These included three 5.5 m × 11 m house raft foundations, two 1.2 m × 4.2 m strip foundations, and nine 2.2 m × 2.2 m pad foundations. Settlements have been small, usually less than 10 mm. The liquefaction potential of the pfa has been investigated by field and laboratory testing and it has be en concluded that there is no risk of liquefaction due to the maximum earthqu ake likely in the Peterborough area.

Case history 12: Peterborough

Figure 76 Preloading lagoon pfa at Peterborough

133

134

Case history 13: Hartlepool Case history 13

Slag bank at Hartlepool Fill material

Iron and steel slag

Method of placement Age of fill

Tipping

Depth of fill

15 m

Water table Ground treatment

Not known

Tip in use for 100 years

Excavation and refilling

Date of ground treatment 1980 Period of monitoring

1980 to 1983

References

Eakin (1983), Eakin and Crowther (1985)

Description of project

Reclamation was considered to be of major environmental importance and included preparation of an area for eventual building redevelopment. Ground treatment involved some excavation and refilling with compaction of the excavated fill under controlled conditions to form a 4 m deep compacted layer. Fill properties

The 40 ha steelworks waste tip had been in use for over 100 years and there was a very variable mixture of iron and steel slags, flue dust and general industrial waste. Several slurry ponds were also present. A chemical and mineralogical investigation identified several reactions which might result from mixing materials during earthmoving and which could cause volumetric instability:  hydration of magnesia (MgO) or pe riclase to form brucite (Mg(OH) 2)  hydration of free lime (CaO) 

formation glassy slag.of sulfo-aluminates (ettringite) as a result of sulfate attack on Tests were carried out by compacting samples into CBR-type moulds. The average expansion was 2.6% and more than 75% of the samples e xpanded by over 1%. Field measurements

Immediately after reclamation of the site, levelling stations were established and settlement was measured at ground surface and at the bottom of the 4 m deep compacted fill at six locations. D1 and D2 were loaded with kentledge to a bearing pressure of 100 kPa to simulate anticipated building loadings. The monitored movements are shown in Figure 77. Expansion of the 4 m layer varied between 1.55% and 3.25% between July 1980 and September 1983, with expansion still continuing at that time. This case history illustrates how in dealing with one problem (potential for reduction in volume due to looseness of the fill), another hazard (expansion due to chemical reactions) may be encountered. However, if recompaction had not triggered the expansion it would have been latent in the fill, and might have occurred subsequent to any building development on the site.

Case history 13: Hartlepool

135

Figure 77 Heave of compacted slags at Hartlepool (after Eakin and Crowther, 1985)

136

Case history14: Greenwich Case history 14

Urban fill at Greenwich Fill material

Medium dense sandy clay with brick fragments

Method of placement Age of fill

Unknown

Depth of fill

2m

Water table Ground treatment

Base of fill

Type of development

Housing

Period of monitoring

1980

Reference

Charles and Driscoll (1981)

Old

None

Description of project

At a site close to the Thames, old property had been de molished and the area was ready for redevelopment with two-storey housing. Fill properties

A site investigation had been carried out by drilling a limited number of boreholes. It appeared that the site was covered with about 2 m of fill, consisting mainly of medium-dense, sandy, gravelly clay with brick fragments. Beneath the fill was a layer of cohesive soil varying in thickness and composition across the site. Over part of the site the layer consisted of a firm-to-stiff brown, silty clay ( cu = 35 kPa to 80 kPa), whereas elsewhere there was soft, clayey peat ( cu = 18 kPa). Underlying this layer was dense, sandy gravel at a depth not exceeding 4 m. The water table was located near the base of the fill. Load tests

In-situ load testsexcavation were carried locations Atdeep eachlayer of location a small wasout dugattofour about 0.5 m, on andthe a site. 0.1 m dry sand was placed in it. A rubbish skip with a base area of 1.8 m × 1.85 m was then lowered onto the sand bed. The skips were filled with damp sand. The bearing pressure applied to the fill by e ach full skip was about 30 kPa . The skips were left in position for one month only. This was the maximum period that could be allowed for the tests owing to the timetable for development of the site. The measured settlements of the skips are plotted against log time in Figure 78. About 80% of the settlement which had been recorded by the end of the first month had occurred almost immediately, as the skips were being loaded. Subsequent settlements were small. The measurement of similar values of settlement for four skips suggested that the fill might not be very variable. This would need to be confirmed by visual inspection of numerous trial pits. Basic results from the four tests are summarised in Table 24.

Table 24 Load tests on old urban fill at Greenwich q b s s α i Tes t (k Pa) (m) (mm) (mm)

mα

s /b i ( %)

s /b α (%)

qb/s i ( M Pa )

A

30

1.8

11.9

1.6

0.13

0.66

0.09

4.6

B

30

1.8

13.3

1.5

0.11

0.74

0.08

4.1

C

30

1.8

10.7

1.4

0.13

0.59

0.08

5.1

D

30

1.8

9.7

0.9

0.09

0.54

0.05

5.6

Case history14: Greenwich

Figure 78 Load tests on urban fill at Greenwich

137

138

Case history15: Manchester(A) Case history 15

Urban fill at Manchester (A) Fill material

Ash, brick, rubble

Method of placement Age of fill

Tipped

Depth of fill

15 m

About 1900

Low Water table Type of ground treatment Vibro stone columns

Date of ground treatment 1974 Type of development

Low-rise housing

Period of monitoring Reference

1974 to 1977 Gray and Thomson, 1979

Description of project

A total of 358 dwellings was built on an infilled site. Vibro stone columns were installed to a maximum depth of 5 m. On the deep fill, lightly reinforced concrete raft foundations were used. Fill properties

The fill consisted of ash, brick and demolition rubble with occasional pockets of clayey material. The SPT N values varied between 1 and over 20 with an average of 6. No cavities were found. Load tests

On and between compaction centres, 0.6 m diameter plate load tests were carried out. At the design working pressure of 75 kPa, settlements of the order of 1 or 2 mm were measured. At the maximum test loading of 375 kPa the maximum settlement was less than 5 mm. Field measurements

Where there was full depth treatment, settlement was small; a maximum of 15 mm after three years. Where the fill was up to 10 m deep, a maximu m settlement of 78 mm was recorded after three years. The maximum deflection ratio was 0.3 × 10–3. Settlement was rapid in the first 12 months , but after that it was just under 1 mm per month.

Case history 16: Redditch (A) Case history 16

139

Old domestic refuse at Redditch (A) Fill material

Old domestic refuse

Method of placement Age of fill

Tipped

Depth of fill

6m

Water table Ground treatment

None in the fill

1960

Dynamic compaction

Date of ground treatment 1975 Development

Dual carriageway road with interchange

Date of development Period of monitoring

1975

References

Charles et al (1981), Charles (1991), Watts and Charles (1999)

1975 to 1995

Description of project

A dual-carriageway road with interchanges and slip roads was to be built across an old domestic refuse tip. It was planned initially to excavate the refuse and replace it with suitable fill. As an e conomic alternative, an area of 22 000 m 2 was treated by dynamic compaction. A granular fill was imported to form a working platform for the crane and compaction was achieved by dropping a 15 tonne weight with a base area of 4 m 2 from heights of up to 20 m. Primary tamping was carr ied out at points on a 5 m square grid and the weight was dropped up to ten times at each location. Craters 1 m deep were formed. After this primary tamping was complete the holes were backfilled and the ground was levelled off. The second stage of tamping was carried out at grid points offset from the srcinal grid. In some places there were three more stages of tamping and the maximum input of energy was 2600 kNm/m 2. Grids of levels before and after treatment showed an average enforced settlement about 10%.of 0.5 m, corresponding to a volume reduction in the refuse fill of Fill properties

The refuse fill was variable, and included black ash and clay with fragments of brick, wood, rags, plastics, bottles and metals. It had been in place about 15 years at the time of treatment and typical bulk densities and moisture contents were 1.8 Mg/m3 and 30% respectively. Standard penetration tests suggested that the fill was of loose to medium de nsity. The fill was underlain by stiff Keuper Marl. Field measurements

Immediately following the completion of ground treatment, magnet extensometers were installed. At gauge 1 there was 5.4 m of refuse fill and at gauge 2 there was 5.0 m. A 3 m high embankment for a slip road was constructed, and the surface of the refuse settled immediately by up to 20 mm owing to the weight of the e mbankment, corresponding to an average D of 20 MPa. The immediate settl ement of some untreated refuse under a similar weight of fill was more than three times greater than the immediate settlement of the refuse treated by dynamic consolidation. It is clear that ground treatment has substantially reduced the compressibility of the refuse fill. The continuing movements are shown plotted against time since embankment construction (logarithmic scale) in Figure 79. In Figure 80 the same settlement measurements are plotted against time since the refuse was deposited, again using a logarithmic time scale. An examination and comparison of these two plots suggests that three phases of movement behaviour occurred. (a) The first 12 months following embankment construction Settlement occurred primarily as a result of physical compression

140

Case history 16: Redditch (A) associated with dynamic compaction and embankment loading, corresponding to αc = 0.2% with zero time taken at completion of embankment construction (Figure 79). (b) An intermediate phase between one year and four years after embankment construction A significant contribution to the compression was made by both physical creep compression described in (a) and by biodegradation of the refuse as described in (c). (c) From four years after embankment construction Settlement is primarily the result of volume reduction of the refuse associated with biodegradation. When the settlement is plotted against the logarithm of time that has e lapsed since the refuse was srcinally deposited (Figure 80), a linear relationship is obtained corresponding to αb = 2%. Physical creep compression corresponding to αc = 0.2% is negligible in this period. Although treatment did not eliminate long-term settlement, the road was opened two years after dynamic compaction, and it is understood that it has performed satisfactorily.

Case history 16: Redditch (A)

141

Figure 79 Settlement of surface of old refuse at Redditch (A) plotted against logarithm of time elapsed since embankment was completed

Figure 80 Settlement of surface of old refuse at Redditch (A) plotted against logarithm of time elapsed since refuse was deposited

142

Case history 17: east end of London Case history 17

Old refuse in the east end of London Fill material

Old domestic refuse

Method of placement Age of fill

Tipped

Depth of fill

6.5 m

Water table Ground treatment

None within fill

About 1935

Dynamic compaction trial

Date of ground treatment 1976 Period of monitoring

1976

References

Charles (1979b), Charles et al (1981), Watts and Charles (1999)

Description of project

It was proposed to build a hospital of two -storey construction on a site which had been used as a gravel pit and subsequently as a refuse tip. A trial of dynamic compaction was arranged to test the possibility of treating the refuse to improve its load-carrying characteristics as an alternative to piling. A 12 m × 12 m square test area was chosen, and a 14 tonne weight with a base area of 4 m 2 was dropped from heights of up to 14 m. Primary tamping points were spaced at 5 m centres on a square grid and each point received about 12 blows from the falling weight. Typically, the first three blows produced a hole 1 m deep, and after 12 blows the imprint was 2 m deep . Secondary tamping points were interspersed among the primary points and received fewer blows. The average total energy applied to the fill was 2600 kNm/m 2, and the tamping was completed during one day. Fill properties

Standard penetration teststhat andmuch cone penetration tests carried outof loose during site investigation showed of the very variable fill was to the medium density. Th e fill was described as black, clayey, sand and ash, with gravel, brick fragments and traces of glass, wood, metal and organic matter. Gradings indicated that about 20% by weight of the fill was finer than 0.075 mm. The organic content was low. Typical valu es for bulk density and moisture content of 100 mm diameter open drive samples were 1.8 Mg/m 3 and 30% respectively. Field measurements

The average enforced settlement produced by dynamic compaction was 0.58 m. This represented an average vertical compressi on of 9% over the 6.5 m depth of fill. Some heave occurred around the te st area, so the average volume reduction was about 8%. Two magnet extensometers showed that the amount of vertical compression was not constant with the depth but was 13% near ground level and very small indeed in the bottom metre of fill. Figure 81 shows the settlement and compression versus depth profiles obtained by averaging the results from the two gauges. After dynamic compaction high water levels were recorded in both gauges, although previously there had been no water table within the fill. Figure 82 shows both the settlement of the ground surface and the decay of excess pore water pressure plotted against the logarithm of the time that had elapsed since ground treatment. About 60% of the settlement occurred above the water level. Almost half the excess pore pressure dissipated during the three months that observations continued, and it is estimated that one and a half ye ars would have been neede d for 90% dissipation. The large pore pressures might have been avoided if a larger area of fill had been tre ated and the energy had been applied in a number of stages separated by several wee ks.

Case history 17: east end of London

143

Figure 81 Settlement induced by dynamic compaction of old refuse in east end of London

Figure 82 Settlement and pore pressure subsequent to dynamic compaction of old refuse in east end of London

144

Case history 18: Hertfordshire Case history 18

Old domestic refuse at trunk road widening in Hertfordshire Fill material

Old domestic refuse

Method of placement Age of fill

Tipped

Depth of fill

8m

Water table Ground treatment

Base of fill

About 1960

Dynamic compaction

Date of ground treatment 1977 Development

Road

Date of development Period of monitoring

1977

References

Charles et al (1981), Watts and Charles (1999)

1977 to 1994

Description of project

A section of the road between Hatfield and Hertford has been improved by the construction of a second carriageway. About 600 m of this new carriageway had to pass over land which had been a gravel pit and had since been refilled with domestic refuse, sealed with clay and topsoil placed. Excavation and disposal of the refuse under the proposed new carriageway and replacement with suitable fill was considered, but dynamic compaction of the refuse was adopted as a cheaper alternative. Compaction was carried out using a 15 tonne weight with a base area of 4 m 2. Some granular material was imported to form a working platform. Treatment was carried out in three stages. Primary tamping took place at points on a 7 m triangular grid. Each point received about eight blows with the weight dropped from 20 m, producing craters over 2 m deep in places. In the second stage, tamping was carried out at closer ce ntres and this was followed by a continuous pass with the weight dropped from a reduced height. Much of the final stage was carried out using a 13.5 tonne weight with base area of 6.5 m 2. Maximum energy input was 2200 kNm/m 2 and average enforced settlement was about 0.5 m (equivalent to an average vertical compression of just under 10%). Fill properties

The soft and compressible refuse was very variable and contained black ash and brick rubble with bottles, paper, wood, plastic and metal. The average moisture content was around 20%, and the percentage of material finer than 0.075 mm was of the order of 15%. It appeared that a major proportion of the decomposition of the refuse was complete. Field measurements

Settlement subsequent to ground treatment has been measured at three magnet extensometers. Settlement with depth profiles measured ten years after treatment are shown in Figure 83. At that time the maximum settlement was 70 mm where the fill is deepe st. Monitoring continued for another seven years and the settlement increased to 100 mm. If it is assumed that the longterm settlement results from biodegradation, the average rate of settlement corresponds to αb = 2%.

Case history 18: Hertfordshire

145

Figure 83 Settlement of old refuse in Hertfordshire during ten-year period following dynamic compaction

146

Case history 19: Redditch (B) Case history 19

Old domestic refuse at Redditch (B) Fill material

Old refuse

Method of placement Age of fill

Tipped

Depth of fill

6m

Water table Ground treatment

Close to ground level

1960

Preloading

Date of ground treatment 1980 Development

Road interchange

Date of development Period of monitoring

1980

References

Charles et al (1986), Watts and Charles (1999)

1980 to 1995

Description of project

In 1980 a road interchange had to be constructed close to that described in case history 16. Dynamic compaction was not practicable because of the close proximity of adjacent buildings, and it was decided to preload the old refuse. It was specified that suitable fill should be placed to form an e mbankment with a minimum height of 3 m, and left in position for at least three months. Subsequently, fill would be removed to the final earthworks profile. Placement of embankment fill be gan in March 1980. By September 1980 all excess surcharge had been removed down to road formation level. Fill properties

Similar to case history 16. Field measurements

In October 1979 two magnet extensometers were installed with the following ground conditions:  5.8 m of refuse with ground-water level 3 m below surface of refuse; 3 m of clay fill was placed over the re fuse and when excess surcharge had been removed 0.5 m of clay fill was left permanently in position  3.8 m of refuse with ground-water level 1 m below surface of refuse; 5.5 m of clay fill was placed over the refuse and when excess surcharge had been removed 2.8 m of clay fill was left permanently in position. The settlement monitored at the two gauges is shown in Figure 84. Most of the settlement occurred as the embankment fill was placed, and the settlement monitoring demonstrated that it was unnecessary to leave the surcharge fill in position for a long period. The specified three- month period was more than adequate. The immediate compression corresponded to D = 2 MPa. The refuse appears t o have been effectively c ompressed through its full depth. However, the stress ratio σ'vc/σ'va is greater than 1.8 throughout the full depth of fill at both locations (section 8.3). Long-term movement has been relatively small. Only 20 mm settlement has been measured over a ten-year pe riod. If it is assumed that the long-term settlement is due to biodegradation, this rate of settlement corresponds to αb = 2%.

Case history 19: Redditch (B)

Figure 84 Preloading old refuse at Redditch (B)

147

148

Case history 20: Liverpool(A) Case history 20

Old refuse at Liverpool (A) Fill material

Old refuse

Method of placement Age of fill

Tipped

Depth of fill

12 m

Water table Ground treatment

Near base of fill

Early 1960s

Preloading

Date of ground treatment 1982 Development

Road

Date of development Period of monitoring

1983

References

Eakin and Crowther (1985), Charles et al (1986),

1982 to 1996 Watts and Charles (1999)

Description of project

Much of the area of the International Garden Festival, close to Liverpool city centre, had been previously used as a refuse tip. The Riverside Spine Road crossed the Festival site and at its south-east end had to be built over old refuse. Preloading was the selected ground improvement technique. The sequence of earthmoving operations is illustrated in Figure 85. First, the refuse was regraded to the profile indicated by the line b–b–b–b. An embankment of sand fill was then constructed to the profile c–c–c. This was left permanently in position. Finally, a temporary surcharge of sand fill 4 m high was placed (line d–d–d–d). The bulk unit weight of the sand was estimated to be 20 kN/m3. The surcharge was 20 m in length and was moved along the line of the road. It was kept in each position for a week or so. Fill properties The old refuse contained ash, soil and rubble. Field measurements

The settlements observed at two locations during the preload period are plotted in Figure 86. The immediate settlement corresponded to D = 3 MPa in the untreated refuse. Where the refuse was 8 m deep, 0.27 m settlement occurred and three-quarters of the total settlement occurred in the upper 4.5 m of the refuse. W here the refuse was 4 m deep, settlement was 0.22 m. Small amounts of heave were monitored as the surcharge was removed. Movements in the sand fill permanently placed over the refuse have been negligible. Long-term movements after surcharge removal are plotted in Figure 87 and have averaged about 5 mm per year. It may be questioned whether this ongoing settlement is mainly due to physical effects associated with the temporary preloading or is principally related to biodegradation in old refuse. The latter seems most probable. The refuse was deposited in the early 1960s and, reckoning zero time as the date of placement, the rate of settlement observed between 1983 and 1996 corresponds to αb = 6%. Some distortion of the heavily trafficked road surface has occurred. When the vertical effective stresses in the old re fuse at gauge 1 are examined, it is found that at the level of magnet d the stress ratio σ'vc/σ'va is about 1.7. Above this level the surcharge has been very effective in compressing the refuse. At gauge 2, the refuse has been thoroughly compressed throughout its full depth, but the layer is thinner at this location and σ'vc/σ'va = 1.9 at its base. The limited information from this site therefore suggests that preloading is effective in compressing the refuse at least to depths at which σ'vc/σ'va > 1.7 (section 8.3).

Case history 20: Liverpool(A)

Figure 85 Preloading old refuse at Liverpool (A)

Figure 86 Settlement of old refuse induced by preloading at Liverpool (A)

149

150

Case history 20: Liverpool(A)

Figure 87 Settlement of old refuse after removal of temporary preloading at Liverpool (A)

Case history 21: Brogborough Case history 21

151

Recent domestic refuse landfil l at Brogborough Fill material

Domestic refuse landfill

Method of placement Age of fill

2 m lifts; steel wheel compactor

Depth of fill

16 m

Water table Ground treatment

Leachate level within refuse

Development

None

Period of monitoring

From 1985

References

Watts and Charles (1990, 1999)

1983 to 1987

None

Description of project

Mainly domestic refuse in plastic bin liners is being placed as landfill in a clay pit with a total area of 120 ha. When filling began in 1983, the fill was raised in 2 m high terraced lifts, and was compacted with a steel wheel compactor. This waste disposal site has been used for e xperimental purposes to monitor the settlement of recently placed domestic refuse. Fill properties

The as-placed in-situ bulk density of the domestic re fuse was 0.55 Mg/m 3 (0.64 Mg/m 3 including daily clay cover). Field measurements

A section through the landfill is shown in Figure 88. In May 1985 a layer comprising 1.1 m of compacted refuse and 0.7 m of clay was placed on an 11 m depth of refuse, which had been placed over a 15 month period. Until then the surface of the refuse was settling at approximately 0.5 mm a day. The extra load caused immediate surface settlement of After 200 mm with compression measuredadditional through the full depth of the fill. about ten days the rate of settlement reduced significantly. A further 0.65 m of clay was placed at the beginning of November 1985 and caused further immediate settlement. In September 1986, 0.8 m of re fuse and a 0.8 m thick clay sealing cap were placed with the final agricultural capping being added at the end of October 1987. The additional loading resulted in immediate settlement through the full depth of the refuse on both occasions. Figure 89 shows the depth and movement of the magnet markers within the re fuse as well as leachate level for the first three years of monitoring. By mid-1989, the settlement of the srcinal ground surface at Brogborough (magnet I) was 1.7 m as shown in Figure 89. Unf ortunately, the magnet extensometer became non-operational in 1989, and subsequently only surface settlement was monitored by precise levelling. In March 1985 the leachat e level was about 1 m above the base of the fill. This level rose by 2.5 m almost immediately on application of an additional layer of refuse plus clay during May 1985 (Figure 89). There was another immediate, but much smaller, rise when an extra layer of clay was placed on the surface in October 1985, and generally the leachate level continued to rise slowly until September 1986. At this time another substantial rise in leachate level was recorded when 0.8 m of refuse and 0.8 m of clay were placed, resulting in large additional settlement of the refuse. Again the leachate level within the fill continued to rise slowly, until the final layer of clay was placed at the end of October 1987, when a further rise in the leachate level in the fill was recorded. By spring 1989 the leachate had saturated the lower 8 m of refuse. The immediate compression of the refuse at Brogborough and Calvert (case history 22), which occurred as further fill was placed, has been summarised in Figure 90 in the form of a plot of constrained modulus, D, versus average vertical effective stress. Some small increase in D with

152

Case history 21: Brogborough increasing σv can be seen. Figure 91 shows the long-term compression of the fill at the two sites plotted against log time. The compression has been calculated from the settlement of the surface of the refuse fills measured by precise levelling. The initial settlement profile largely depends on when zero time is reckoned, but after about three years it is much less important. If ze ro time is taken when the fill was three-quarters complete, a more nearly linear relationship between compression and log time is found throughout the full period of monitoring. However, the relationship between compression and time could probably be better expressed by a power law (sections 3.4 and 4.5). By late 1991, the rate of settlement at Brogborough corresponded to αb = 10%. This rate later increased substantially and over the period to the end of monitoring in 1995 the rate corresponded to αb = 23%.

Figure 88 Section through refuse landfill at Brogborough

Figure 89 Settlement of refuse landfill at Brogborough

Case history 21: Brogborough

Figure 90 Constrained modulus of recent domestic refuse

Figure 91 Long-term settlement of recent domestic refuse

153

154

Case history 22: Calvert Case history 22

Recent domestic refuse landfill at Calvert Fill material

Recent domestic refuse

Method of placement Age of fill

2 m lifts; steel wheel compactor

Depth of fill

20 m

Water table Ground treatment

Leachate level within refuse

Development

None

Period of monitoring

From 1985

References

Watts and Charles (1990, 1999)

1982–1986

None

Description of project

In 1979 a landfill operation began to fill a clay pit with domestic re fuse, and from 1985 this waste disposal site has been used to monitor the settlement of recently placed domestic refuse. Fill properties

As-placed bulk density was up to 0.8 Mg/m 3 (including cover material). Field measurements

Figure 92 shows a section through the landfill. Settlement is plotted in Figure 93. From December 198 5 until June 1986, surface settlement continued at an almost constant rate of about 2 mm per day, with compression occurring through the full depth of the fill. Towards the end of June 1986, an additional 1.3 m of refuse and 0.1 m thick intermediate clay cover were placed, resulting in immediate surface settlement of 0.2 m. After about a week, the rate of settlement decreased significantly to a fairly constant rate of about 3 mm a day. Inresulting September 1986 the refuse was sealed withMonitoring a 1.1 m thick clay capping layer, in immediate settlement of 0.4 m. of settlement under this constant load continued until June 1988 when a final 0.75 m thick agricultural clay cap was placed. The rate of settlement remained almost constant during this period at just under 3 mm pe r day. The magnet extensometer became non- operational in mid-1987, and from then on only surface settlement was measured. The large movements that have occurred are shown in Figure 91. Figure 94 shows the settlement monitored at settlement stations on the surface of the landfill over a period of 18 months. The stations were located at appropriate distances so that the likely differential settlement over the area of a small building could be estimated. Load tests

Two load tests were carried out on the surface of the 1 m thick clay sealing cap layer. The settlement of the two skips and the magnet markers located at the top of the refuse beneath the clay cap is plotted against the logarithm of time in Figure 95. A continuing rate of settlement of 50 mm pe r log cycle of time can be attributed to the applied load. Most of this settlement is located in the refuse immediately below the clay cap as shown in Figure 96. If it is assumed that a depth of fill of 1.5B is affected by the applied load, the rate of settlement is equivalent to αc = 2% for physical creep compression. However, the linearity with log time is poor and the relationship between settlement and time might be better expressed in terms of a power law (sections 3.4 and 4.5).

Case history 22: Calvert

Figure 92 Section through refuse landfill at Calvert

Figure 93 Settlement of refuse landfill at Calvert

155

156

Case history 22: Calvert

Figure 94 Differential movement at surface of refuse landfill over a period of 18 months

Figure 95 Load tests on refuse landfill at Calvert

Case history 22: Calvert

157

Figure 96 Load test on refuse landfill at Calvert; settlement versus depth

158

Case history 23: Hull Case history 23

Infilled dock at Hull Fill material

Chalk fill

Method of placement Age of fill

End tipping into water

Depth of fill

9m

Water table Ground treatment

3 m below ground level

1977

Temporary preloading with 2 m high surcharge

Date of ground treatment 1978 to 1980 Development

Road

Date of development Period of monitoring

1980

Reference

Charles et al (1986)

1977 to 1984

Description of project

The construction of the Hull South Orbital Road involved crossing the west end of the disused Victoria Dock, which at this location was 100 m wide. It was decided to build the road on an embankment across the dock. A method of infilling the dock was required which would minimise long-term settlement and ensure that little or no damage would be caused to the road built over the dock. There was, typically, some 8 m of loose organic silt in the dock overlying boulder clay, but conditions we re variable. The surface of the silt was generally about 2 m below mean water level. End tipping of chalk fill displaced much of the loose silt, but eventually silt had to be e xcavated to make it possible to complete the filling, as shown in Figure 97. The average depth of the chalk fill was 9 m and there was an average depth of 4 m of silt left in place below it, but there were considerable variations in these depths across the site. Dock walls were demolished to reduce differential settlement at these locations.forEarly 1978, a 2years. m high surcharge of fill was placed and left in position moreinthan two Fill properties

The bulk density of the chalk fill was 1.7 Mg/m3 and the moisture content was 14%. Field measurements

Figure 97 Excavating dock silt

Of 12 pneumatic piezometers installed below the chalk fill, seven showed little change in pore water pressure as the surcharge was placed; the other five recorded an average rise of 3 m head of water with total excess pore pressures of between 5 m and 12 m at the completion of su rcharge placement. Dissipation of excess pore pressures was largely completed within one year. During the surcharge period settlements of up to 0.4 m were measured. The settlement showed an approximately linear re lationship when plotted against the logarithm of time elapsed since the placing of the surcharge. Generally, less than half the total settlement could be attributed to compression of the silt and hence associated with dissipation of pore pressures. Most of the settlement occurred because of compression of the chalk fill. Following road construction late in 1980, kerbs have been levelled to monitor any longterm settlement. After four years movements were still negligible and no problems had arisen.

Case history 24: Liverpool(B) Case history 24

159

Infilled dock at Liverpool (B) Fill material

Sand

Method of placement Age of fill

Hydraulic

Depth of fill

13 m

Water table Ground treatment

7 m below ground level

1982

Temporarily preloaded with 4 m high surcharge

Date of ground treatment 1982 Development

Road

Date of development Period of monitoring

1983

References

Eakin and Crowther (1985), Charles et al (1986)

1982 to 1996

Description of project

At its northern end, the Riverside Spine Road had to cross Herculaneum Dock. First, 8 m of loose silt was removed from the dock by dre dging. The dock was then filled with hydraulically placed sand to a height of 4 m above the quay level, to stockpile sand for use elsewhere at the International Garden Festival site. This acted as a surcharge on the sand used to infill the dock. Where the road crossed the dock walls, large wedges were cut out to minimise the effect of differential settlement. Fill properties

The SPT N values were very low in the loose dark grey silty sand. Field measurements

Two magnet extensometers were installed in boreholes. At one location a layer layer of loose siltwas wasnot found fromat11 to 13 m below the surface of thethe fill . This of silt present themother location. Figure 98 shows settlement measured where the silt layer was pre sent. The water table was 7 m below ground level at this location. Up to Febru ary 1985, the surface settlement was 146 mm and 75% of this movement was attributable to compression in the silt layer at the bottom of the fill. At the other monitored location, the settlement over the same period was 37 mm. During 1985 further settlement occurred, which was fairly evenly distributed through the depth of fill. This coincided with the opening of the road to heavy through-traffic, and vibrations and high axle loads may have been the cause of the additional movements. Little further movement was measured up to the loss of the gauges after 1989 when the carriageway was realigned and resurfaced. Beginning in April 1983, the levels of kerbs have been monitored across the infilled dock. Figure 99 shows settlements measured on the two sides of the road. The maximum settlement at chainage 180 m is at the location where the silt layer was present in the fill. There was damage to a boundary wall and deflection of the road surface where the road passes over the old dock wall.

160

Case history 24: Liverpool(B)

Figure 98 Settlement subsequent to removal of surcharge at infilled dock, Liverpool (B)

Figure 99 Settlement of kerbs on infilled dock at Liverpool (B), 1983–1989

161

Case history 25: Abingdon Case history 25

Clay fill in former gravel pit a t Abingdon Fill material

Miscellaneous clay fill with some organic material

Method of placement Age of fill

Uncontrolled 1955

Depth of fill

4m Vibro stone columns Ground treatment Date of ground treatment 1988

Development

Two-storey steel framed structure

Date of development

1988

Period of monitoring Reference

From 1988 Watts and Charles (1991)

Description of project

The southern half of the site was used for storage of heavy plant; the northern half had remained undisturbed since backfilling. A two-storey framed structure was built on the site with one wing on the southern half and one wing on the northern half. All structural columns are supported on reinforced concrete foundation pads. Before construction, stone columns were constructed through the full depth of the fill, using the dry method. Fill properties

The soft clay fill contained organic matter at some locations. Field measurements

Levelling stations were installed on the foundation pads and monitoring of settlement started before the steel frame was erected. Typically, settlement of the north wing has been twice as large as settlement of the south wing. After a year a maximum settlement of 21 mm had been measured. Load tests

Two load tests were carried out on 2 m square load pads cast on fill close to the north wing. The results are summarised in Table 25. An average bearing pressure of 50 kPa was applied in each test. One of the tests was on untreated fill, the other was situated over four stone columns. The load test on untreated ground settled 13 mm; that on treated ground settled 18 mm. Laboratory tests predict settlements of this order for untreated ground at the two locations. It would appear that the stone columns had little effect on the settlement. Figure 100 shows the results plotted on a log time scale. The clay fill shows poor linearity on this log time plot, with relatively large long-term settlement.

Figure 100 Load tests on clay fill at Abingdon

Table 25 Load tests on clay fill at Abi ngdon Test q b si lo c a tio n

( kPa )

(m)

(mm)

sα (mm)

si /b

sα /b

mα

( %)

(%)

qb/si ( MPa )

Untreated

50

2.0

3.2

2.8

0.88

0.16

0.14

31

Treated

50

2.0

5.5

3.8

0.69

0.28

0.19

18

162

Case history 26: Waterbeach Case history 26

Building wastes at Waterbeach Fill material

Inert building wastes

Method of placement Age of fill

End tipped in 1 m high lifts

Depth of fill

6.5 m

Water table Ground treatment

4.5 m below fill surface

From 1987

Rapid impact compaction

Date of ground treatment 1990 Development

None

Period of monitoring Reference

1990 Watts and Charles (1993)

Description of project

In 1990, a new ground improvement technique using rapid impact compaction was assessed by carrying out a field trial on loose fill, which had not received systematic compaction other than from a bulldozer spreading the wastes. An area of ground was treated with a 7 tonne weight falling 1 m onto the 1.5 m diameter compacting foot. Abutting treatment points were used at 1.5 m centres, and each treatment point rece ived 50 blows. The average energy input was 1500 kNm/m 2. Fill properties

Typical constituents of the fill comprised brick, concrete, wood, glass and rag with some soil. Dynamic probing at a number of locations showed large variations in blow count over short depths, indicating extreme variability of these building wastes. Field measurements Surface settlement was generally greater than 0.3 m. Vertical compression in the upper 2 m of the fill was of the order of 10% and significa nt compression was measured to a depth of 4 m. Settlement versus depth is plotted in Figure 101. Two techniques were used to assess the properties of the fill before and after treatment: dynamic probing and the measurement of Rayleigh wave velocity. Dynamic shear modulus has be en calculated from the latter and is shown on Figure 101. Load tests

Tests were performed on both treated and untreated fill. Results are presented in Table 26 and Figure 102. The behaviour monitored in the two tests was very similar and there was little difference in performance between the treated and untreated fill. However, the results of surface load tests are likely to be controlled by near-surface conditions and will give little indication of ground properties at depth or the long-term behaviour of the fill. The initial settlement may principally reflect looseness of a thin surface layer of fill.

Table 26 Load tests on fill at Wa terbeach Fill q b si c o nditio n (k Pa) (m) (mm)

sα (mm)

mα

si /b ( %)

sα /b (%)

qb/si ( M Pa )

Untreated

24

1.8

5.0

2.0

0.4

0.28

0.11

9

Treated

24

1.8

4.0

2.8

0.7

0.22

0.16

11

Case history 26: Waterbeach

Figure 101 Rapid impact compaction of building wastes at Waterbeach

Figure 102 Load tests on building wastes at Waterbeach

163

164

Case history27: Manchester(B) Case history 27

Alluvial sand deposit at Manchester (B) Fill material

None; 4 m of alluvial sand with thin peat layer

Method of placement Age of fill

Not applicable

Depth of fill

Not applicable

Not applicable

1 m below ground level Water table Type of ground treatment Vibro stone columns

Date of ground treatment 1986 Type of development

Low-rise housing

Date of development Period of monitoring

1986/1987

References

Watts et al (1989), Watts and Charles (1991)

1986 to 1989

Description of project

The housing development, consisting of 15 semi-detached and terraced twostorey blocks, has been built on ground treated by vibro stone columns, installed using the wet method. The treatment comprised a single row of stone columns along the line of all loadbearing walls, with a small number under ground-floor slabs. The full depth of alluvial sand was treated to depths varying between 2.8 m and 4.2 m. Fill properties

There was no fill on this site. Alluvial deposits of sand and sand and gravel are underlain by firm clay at about 4 m be low ground level. The case history has been included in this book because the properties of some sand fills should not be very different from those of this alluvial sand. Field measurements Settlement of six blocks of houses has been monitored. The houses have exhibited only small differential and total movements, well below the level at which structural damage would be expected. House blocks founded on the peat layer have, in ge neral, settled more than those located where there was no peat. Load tests

A load test was carried out on untreated ground so that a comparison could be made with the performance of the houses on the treated ground. The loaded pad was 0.75 m wide × 2.25 m long. In Table 27 the results are interpreted in two ways. The ‘uncorrected’ values show the test results from the measured settlement. The ‘sand only’ values show the results interpreted by removing those parts of the settlement which the magnet extensometer indicated occurred in the peat layer and the underlying clay.

Table 27 Load q test on untreated sandswith peat layer (a/b = 3) Fill b sα i

si /b

sα /b

qb/si

c o nditio n

( %)

(%)

( M Pa )

Uncorrected Sand only

(k Pa) 50 50

(m) 0.75 0.75

(mm) 2.8 1.5

(mm) 0.5 0.3

mα 0.2 0.2

0.37 0.20

0.07 0.04

13 25

Case history 28: Bacup Case history 28

165

Urban fill at Bacup Fill material

Ash fill overlying cohesive fill

Method of placement Age of fill

Unknown

Depth of fill

5m

Old

Base of fill Water table Type of ground treatment Vibro stone columns

Date of ground treatment 1990 Type of development

None

Period of monitoring References

1990 to 1992 Watts et al (2000), Wood et al (1996)

Description of project

Full-scale instrumented load tests have been carried out at an experimental site in Bacup to study the installation and performance of vibro stone columns supporting a strip foundation on a variable fill. The performance of a similar strip foundation on untreated ground has also been studied. The considerable variability of the fill provided the potential for unacceptable differential settlements of a foundation on this fill and the objective of ground treatment was to reduce total and differential settlement. Treatment was carried out using the dry top-feed vibro technique. Three isolated columns were installed to establish an appropriate treatment procedure for the trial footings. One column was installed to the maximum capability of the poker at each stage of the column construction, and two in accordance with generally adopted practice for these ground conditions. Five columns were then installed in accordance with the latter procedure at 1.8 m centres along the centreline of the treated foundation strip. Several additional testThe columns were installed using thein wet top-feed vibro method.building foundations were cast in situ accordance with normal practice. Load was applied using 4 tonne concrete cubes, with the bottom row supported on spreader beams so as not to increase foundation stiffness. Each of three load increments comprising a row of blocks applied a total bearing pressure of 41 kPa to the 0.75 m wide foundation strips. Fill properties

Coarse fill overlay cohesive fill with stiff glacial till below the fill. The coarse fill contained a high proportion of black ash, some small pieces of sandstone and limestone, slate, burnt slag, clinker, brick and concrete fragments, sand and gravel and broken sandstone and sandstone cobbles and boulders. Particle size analys es showed the upper fill to be slightly silty, sandy gravel. There were pockets of soft silty clay within the coarse fill. The lower cohesive fill comprised silty clay with granular fragments. Classification tests on the cohesive fill indicated an inorganic clay of medium plasticity. The organic content of the fill ranged from 1.4% to 4%. Field measurements

Pressure cells at 0.9 m and 1.5 m from the column axis and at depths of 0.4 m, 0.9 m and 1.8 m below foundation level all measured changes in radial earth pressure during poker pene tration and compaction of the stone column. The effect diminished with increasing distance from the column centre. Dynamic probing was carried out at increasing distances from a normal test column to investigate soil densification around the stone columns. There was a significant increase in blow counts close to the edge of the column, with a more modest increase at 0.6 m from the column centre. Greater increases in blow count were measured in granular fill, but significant increases above pre-treatment values were also apparent in the more cohesive fill. Much larger increases, averaging ten times pre-treatment values, were measured in

166

Case history 28: Bacup the granular fill immediately surrounding the heavily compacted test column. By contrast, probing around a column installed using the wet top-fee d process, which is designed to minimise disturbance of the surrounding soil, showed no change in resistance in the surrounding soil. The immediate settlement resulting from the initial load increment which applied a pressure of 41 kPa was similar for both strips, as shown in Figure 103. Increasing the pressure to 82 kPa produced slightly more settlement of the untreated strip than the treated strip and the rate of creep for the untreated strip was also greater after one month. The third load increment which increased the pressure to 123 kPa induced much larger settlement of the untreated strip. At the end of a further seven months of monitoring, the maximum settlements of the treated and untreated strips were 16 mm and 26 mm respectively. B oth foundations exhibit ed sagging along their length. The maximum settlement of the foundations was reduced by 40% by the installation of stone columns. Magnet extensometers installed beneath both treated and untreated strips showed that the presence of the columns gave rise to more dee p-seated settlements than beneath the untreated strip. These settlements occurred within the cohesive fill and the underlying stiff clay below the base of the columns. It appears that the presence of the columns encouraged load transfer to a greater depth resulting in the deeper seated settlements.

Figure 103 Settlement along strip foundations at Bacup

167

Appendices

Appendix A: Stress distribution below building foundations When calculating ground deformations which are likely to be caused by the weight of a building, it is necessary to make some assessment of the stress distribution in the ground beneath the load. Figure A1 shows the distribution of vertical stress, σv, with depth, on the centre-line of strip footings that apply a uniform stress to ground which has uniform linear elastic properties . The footing widths, b, vary from 0.5 m to 1.5 m for a net loading of 50 kN/m run and it can be seen that 1 m below foundation level there is little difference in vertical stress despite large differences in footing widths. The approximation of elastic real fillmedium behaviour to that of a homogeneous, isotropic, may appear crude, but the vertical stress distribution is not sensitive to the assumption of linear elasticity (Burland et al, 1977). With variable fill material any more sophisticated analysis would not be justified. However, where a relatively stiff fill overlies soft natural soils the stress distribution will be significantly modified with the fill acting to spread the load. The overburden pressure is also plotted in Figure A1; this is the vertical stress in the soil prior to construction and is designated by γ(z + d ), where γ is the bulk unit weight of the fill, z is the depth below foundation level and d is the depth of the foundation. Appreciable compression of the fill is only likely to occur at depths where the increase in vertical stress due to foundation load, ∆σ v, is a significant proportion of the existing in- situ effective stress, γ(z + d ). This assumes that ground-water level is below the depth of influence of the foundation loading. A reasonable limiting value of the ratio ∆σv/[γ(z + d )], below which ∆σv has little significance, is in the range 0.5 to 0.2. Lange (1986) found that the effective depth to which additional stresses caused movement in a silty sand backfill corresponded to a ratio of 0.25. Adopting a ratio of 0.25, it can be seen from Figure A1 that, for a foundation load of 50 kN/m run, the ground is significantly stressed by the foundation load to a depth below foundation level of only about 2.3 m. Using the stress increment ratio of 0.25, an initial estimate of the significant depth ze (m) affected by

Figure A1 Vertical stress (σv) on centre-line of strip footings of different widths with net load of 50 kN/m run

various loading conditions can be obtained from the following relationships:  an applied pressure q kPa on a very large raft foundation; z = 4(q/γ) e  a line load p kN/m acting on a strip foundation; ze = 1.5 (p/γ)0.5  a point load P kN acting on a small pad foundation; ze = 1.2 (P/γ)0.33 These relationships assume that the foundation is on a uniform fill of bulk unit weight γ (kN/m 3). If there is a high ground-water level, the submerged unit weight, γ', should be used instead of γ.

168

Appendix B

Appendix B: Settlement of foundations calculated using elastic theory

Fox (1948) presented depth factors as a function of the geometry of the foundation. Gazetas et al (1985) suggested that Fox’ s factors exaggerated the effects of embedment and presented an approximate expression for fd as follows:

The immediate settlement of a coarse fill, or the consolidation settlement of a fine fill, caused by the weight of a structure may be estimated from elastic theory using a constrained modulus. Immediate settlement, si, can be related to applied pressure, q, foundation length, a, foundation width or diameter, b, depth of foundation, d, shape and rigidity of foundation, constrained modulus of the fill, D, and Poisson’s ratio of the fill, ν, as follows:

fd = 1 – 0.08( d/b)[1 + (4/3)( b/a)] The expression is said to be valid for aspect ratios, a/b, up to 6. Although the expression for depth factor gives some indication that at greater foundation depths the effect of

si = fs fν fd (qb/D) Shape and rigidity factor

The shape and rigidity factor, fs, for some common shapes of footing is given in Table B1. Values are quoted for flexible and rigid footings; in general the settlement of a rigid foundation is 0.8 times the settlement at the centre of a flexible foundation. Poisson’s ratio fac tor

The Poisson’s ratio factor, fν, is such that:

fν = (1 – ν)2/(1 – 2ν) The value of fν for a range of values of ν is given in Table B2. In section 3.4 the relationship for normally ν and Ko and φ' is 0.25 consolidated between hence explained. Forfills coarse fills, typically, ν is in the range to 0.3. These values of fs and fν are based on the assumption that the fill behaves as a homogeneous isotropic elastic material. Oner (1990) has examined the situation where the modulus increases with depth according to a power law. Depth factor

The foundation depth factor, fd, is defined as:

fd = settlement of foundation at depth d settlement for surface foundation Table B1 Shape and rigidity factor fs for centre of footing Footing Footing F o o t i n g s h a pe d i m en s i o n b rigidity f

embedment is more pronounced, in re ality the effects of embedment are more complex than this simple factor based on footing geometry suggests. Gazetas et al (1985) described three possible effects of embedment on foundation settlement:  the trench effect due to weight and stiffness of the overlying soil may reduce settlement (the de pth factor fd is based on this effect)  the fill stiffness usually increases with depth and hence smaller settlement may be expected with greater depth of embedment (a representative value of the constrained modulus should be established)  foundation sidewall contact with the surrounding soil may reduce settlement (Gazetas et al (1985) have presented a method for making an appropriate allowance for this). The second the effect may emphasises need tobe de particularly termine theimportant, constrainedand it modulus at a relevant depth over an appropriate stress range. Burland and Wroth (1974) showed that under a circular footing of diameter b, 80% of the settlement occurs above a depth of 2 b. This assumed that D was constant with depth. In a situation where D increases with depth, 80% of the settlement may occur above a depth of b. However, there are situations where a stiff surface crust overlies less stiff material. Appendix A has demonstrated the major effect of the applied load in determining the depth to which the fill is significantly stressed.

Table B2 Poisson’s ratio fac tor fν fν

ν

Reference Scott(1965)

0.00 0.20

1.00 1.07

0.79

PoulosandDavis(1974)

0.25

1.13

1.12

Giroud(1970)

0.30

1.23

Gazetas et al (1985)

0.35

1.41

s

Circular

Diameter

Flexible

1.0

Circular

Diameter

Rigid

Square

Side

Flexible

Square

Side

Rigid

0.90

Rectangular (a/b =3)

Width

Flexible

1.72

Giroud(1970)

Rectangular (a/b =3)

Width

Rigid

1.37

Gazetas et al (1985)

Rectangular (a/b =5)

Width

Flexible

2.10

Giroud(1970)

Rectangular (a/b =5)

Width

Rigid

1.66

Gazetas et al (1985)

Rectangular

Width

Flexible

1.12(a/b)0.39

Giroud (1970)

Rectangular

Width

Rigid

0.90(a/b)0.38

Gazetas et al (1985)

Appendix C Constrained modulus

Load tests may be used to determine the constrained modulus (section 4.2). The ratio qb/si is determined in load tests (see for example Table 7) and the constrained modulus then can be derived from the following expression:

D = (qb/si) fs fν fd The tests were all carried out at shallow depths and all but one on footings for which a/b = 1; in these tests the product fs fν fd is about 1. In the test where a/b is 3, the product fs fν fd is about 1.4. The settlement ratios si/b measured in the load tests are also quoted in Table 7. The ratio of applied pressure to constrained modulus is closely related to the settlement ratio:

q/D = (si/b)/(fs fν fd) Settlement under a footing is often calculated by integrating compression with depth under the footing using an elastic distribution of vertical stress, which is independent of ν, on the centre-line of the footing. This calculation makes the implicit assumption that ν = 0. For a fill with ν = 0.25, Table B2 shows that the error is only 13%. To give an indication of the magnitude of settlement which is predicted by this method, a simple example is taken of a rectangular rigid footing with length a = 5 m, width b = 3.4, 1 m,some q = 100 kPavalues on fill of withDνhave = 0.3.been In Table 3 in section typical provided for σvo = 30 kPa and ∆σ v = 100 kPa which are appropriate to this example. The following settlements are calculated: sandy gravel fill, ID = 0.8 s = 4 mm sandy gravel fill, ID = 0.5 s = 8 mm colliery spoil (heavily compacted) s = 34 mm colliery spoil (poorly compacted) s = 68 mm The values listed above are applicable to a foundation at ground surface. The equation for the depth factor derived by Gazetas et al (1985) can be used to predict settlement of footings at depth. For the cases given above with d = 1 m, they ind icate fd = 0.9, and thus this relatively shallow depth of embedment has only a minor effect on settlement.

169

Appendix C: Delineation of an exclusion zone over a highwall The safe and economic redevelopment of many brownfield sites is dependent on an appropriate decision concerning the area of land which cannot be built on due to the potential for differential settlement associated with changes in the depth of fill. Where building is excluded from an excessively large area, redevelopment may not be economically feasible. Where the exclusion area sterilised is too small, building damage can result. The significance of a variable depth of fill has been described in Chapter 6. The geometry of the fill over a highwall can be described by the de pth of burial of the highwall, D, the height of the highwall, H, and the slope angle of the highwall, β, as shown in Figure 30 (page 59). In this appendix, a simple approach to delineating building exclusion zones is presented. Small structures can be stiffened so that tilt is the only problem. Consequently charts have been produced which relate the size and location of exclusion zones to tolerable tilt. A linear tilt model has been developed for surface settlement from which charts have been produced relating the size of exclusion zones to tolerable tilt. A large-scale experiment, field measurements and published data have been used to validate the linear tilt model (Charles and Skinner, 2001c). The charts help to identify where there is likely to be a problem, thus facilitating decisions concerning the need for ground treatment and the of buildings relation to hazards. While thelocation charts provide an aidinto engineering judgement, their limitations should be recognised. Heterogeneous fills are unlikely to settle in a uniform manner so that the charts can only be very approximate guides. Uncertainties about the true geometry and location of the highwall, and about soil properties together with the heterogeneity of many loose fills make any more complex approach unprofitabl e. The charts are not applicable where a substantial depth of compacted fill overlies much looser fill. For small buildings it is usually feasible to provide a stiff raft foundation which will prevent distortion of the building due to differential settlement and will re sist horizontal tensile forces. The charts are only applicable for building developments comprising small structures with stiff raft foundations, so that tolerable tilt can be adopted as the criterion defining the zone from which buildings should be excluded. A simple linear tilt model of surface settlement is proposed in which it is assumed that the slope of the ground surface, d s/dx, either varies linearly with horizontal distance, x, or is constant. For buildings, the slope d s/dx is generally referred to as tilt, α. It is assumed that differential settlement is caused by compression of the fill below the level of the top of the highwall. There is a uniform vertical compression potential, εv, and the maximum settlement, sM, is located away from the base of the highwall such that: sM = ε v H

170

Appendix C

Three-zone case

relation to the height of the highwall. The normalised depth of burial Dn = D/H and the normalised total length of ground where tilt is not zero Wn = W/H.

In the case shown in Figure C1, Hcotβ > (H + 2D)cot γ and, consequently, there are three zones:  zone T: over the top of the highwall there is a linear increase in tilt from zero to a maximum tilt of sM/(Hcotβ)  zone M: over the middle of the highwall there is a constant tilt of sM/(Hcotβ)  zone B: over the base of the highwall there is a linear decrease of tilt from the maximum, sM/(Hcotβ), to zero. The width of the three zones can be e xpressed in terms of a limit angle, γ, which is shown in Figure C1. If x = 0 is taken as the point on the ground surface vertically above the top of the wall, then the extent of the three zones are as follows:  zoneT: – Dcotγ < x < +Dcotγ  zone M: Dcotγ < x < Hcotβ – (H + D)cot γ  zone B: Hcotβ – (H + D)cot γ < x < Hcotβ + (H + D)cot γ The total width of ground where tilt is not zero, W, can be related to the highwall geometry and the limit angle:

W = (H + 2D)cotγ + Hcotβ The results obtain ed from the linear tilt model can be generalised by normalising the linear dimensions in

Wn = (1 + 2Dn)cot γ + cot β The settlement of the ground surface, s, is normalised in relation to the maximum settlement, sM, such that sn = s/sM. For this three-zone case, the linear tilt model predicts a surface settlement profile composed of two parabolae connected by a line of constant gradient as shown in Figure C1. In evaluating the magnitude and extent of the ground deformations, the value of the limit angle, γ, is important. For a settlement trough over a cavity, based on surface observations, Arato (1992) suggested γ = (45° + φ'/2) for dry granular soils where φ' is the angle of shearing resistance. This limi t angle has been used for fills by Lange (1986). For loose fills φ' is likely to be in the range of 30° to 40° and therefore γ = 60° to 65°. This is compatible with limit angles used in mining and tunnelling. For arithmetical simplicity γ has been taken as 63.4° so that c ot γ = 0.5 and for this case:

Wn = 0.5 + Dn + cotβ A small value of γ implies a large area of relatively small tilt whereas a large value of γ implies larger tilt affecting a smaller area. The value of γ that has been adopted is towards the upper bound of probableconservative values and value for therefore should give a moderately maximum tilt. However, it could slightly under-estimate the area affected by tilt. Expressed in normalised form and with cot γ = 0.5, the three-zone situation shown in Figure C1 is defined by cotβ > (0.5 + Dn). In this case the maximum tilt αM is as follows: αM = (ds/dx)M = sM/(Hcotβ)

This equation can be reformulated in normalised form: (d sn/dxn)M = (ds/dx)M (H/sM) = αM/εv = tanβ Two-zone case

The two-zone case of the linear tilt model is illustrated in Figure C2; there is no middle zone of constant tilt:  zone T: over the top of the highwall there is a linear increase in tilt from zero to a maximum  zone B: over the base of the highwall there is a linear decrease of tilt from the maximum to zero.

Figure C1 Linear tilt model and surface settlement: three zones (after Charles and Skinner, 2001c)

This two-zone case applies where Hcotβ < (H + 2D)cot γ; with cot γ = 0.5 this can be expressed as cot β < (0.5 + Dn). The equations which describe the total length of ground where the tilt is not zero are applicable to the two-zone case as well as to the three-zone case. It is assumed that the maximum tilt occurs at x = Hcotβ[D/(H + 2D)] or in normalised form,

Appendix C

171

(d s/dx)M = αM = 2sM/[(H + 2D)cot γ + Hcotβ] Normalising and putting cot γ = 0.5, this equation becomes: (d sn/dxn)M = αM/εv = 2/(0.5 + Dn + cot β) Charts for exclusion zones

Using the linear tilt model, charts have been pre pared showing the dimensions of the exclusion zones in terms of normalised parameters. The basic variable is the normalised tilt, α/εv, where α is the tilt and εv is the

Figure C2 Linear tilt model and surface settlement: two zones (after Charles and Skinner, 2001c)

xn = Dncotβ/(1 + 2Dn). The extents of the two zones are as follows:  zoneT: – Dcotγ < x < + Hcotβ[D/(H + 2D)]  zone B: Hcotβ[D/(H + 2D)] < x < Hcotβ + (H + D)cot γ Equating the maximum settlement, sM , to the integration of tilt between x = –Dcotγ and x = Hcotβ + (H + D)cot γ, it can be shown for the two zone case that:

vertical compression in the fill below the top of the highwall at some distance beyond the base of the highwall. Figure C3 provides a simple guide to help to identify situations where tilt will be a problem. Contours of αM/εv are plotted on a graph of Dn versus cot β, where αM is the maximum tilt for a particular slope geometry computed using the linear tilt model. The value of αM/εv which is read off from the chart should be multiplied by εv, the estimated potential vertical strain in the backfill, in order to obtain the predicted value of the maximum tilt, αM. If the tolerable tilt, αT , is smaller than αM there is likely to be a problem. This can be illustrated by the following example concerning a highwall with cot β = 1.0 and Dn = 1.0, fill with εv = 0.005 (0.5%) and tolerable tilt, αT = 0.002 (1/500). The predicted value of αM/εv from Figure C3 is αM = 0.8 ×there 0.8 and 0.005is = 0.004. This than αTtherefore and, consequently, a problem andisanlarger exclusion zone should be delineated. The range of values of εv of most practical interest is from 0.001 to 0.02 (0.1 to 2%). This is because:  εv is likely to excee d 0.1% even with a good granular fill and, therefore, smaller values of strain are of little practical interest  where εv exceeds 2%, differential movements are likely to be so large owing to heterogeneity that there will be

Figure C3 Identification of problem: predicted values of maximum tilt (after Charles and Skinner, 2001c)

172 a major problem even if there is no change in depth of the fill. A typical design criterion for tolerable tilt for small structures is 0.002 (1/500) (Skinner and Charles, 1999). Consequently the range of α/εv of most interest is from 0.1 to 2. With this in mind, some general findings from Figure C3 are as follows:  with αT = 0.002 (1/500) and εv = 0.001 (0.1%), there will be a problem if αM/εv > 2; from Figure C3 it can be seen that this is only likely to occur with a near-vertical highwall and a very small burial depth

Appendix C with αT = 0.002 (1/500) and εv = 0.005 (0.5%), there will be a problem where αM/εv > 0.4; from Figure C3 it can be seen that this will occur in most cases where there is a highwall (ie in most cases where cot β < 2.5 and Dn < 3)  with αT = 0.002 (1/500) and ε v = 0.02 (2.0%), there will be a problem where αM/εv > 0.1; from Figure C3 it can be seen that this will occur in all cases where there is any change in the depth of fill; however, with a fill of this type there will almost certainly be a problem e ven where there is no change in de pth of fill! 

Figure C6 Width of exclusion zone, WEn, for cotβ = 1.0 (after Charles and Skinner, 2001c)

Figure C4 Width of exclusion zone, WEn, for cotβ = 0 (after Charles and Skinner, 2001c)

Figure C5 Width of exclusion zone, WEn, for cotβ = 0.5 (after Charles and Skinner, 2001c)

Figure C7 Width of exclusion zone, WEn, for cotβ = 2.0 (after Charles and Skinner, 2001c)

Appendix D

173

For those situations where the tolerable tilt is exceede d, Figures C4, C5, C6 and C7 present contours of the e xtent of the exclusion zones on graphs of αT/εv versus Dn for cotβ = 0, 0.5, 1.0 and 2.0, respectively. The width of the exclusion zone, WE, is normalised as WEn = WE/H. With a near-vertical highwall and Dn = 0, the exclusion zone is very localised. With small values of β and larger values of Dn, tilt is smaller but in some circumstances the exclusion zone could be larger. For the range of αT/εv from 0.1 to 2, some findings from the charts are as follows:  cot β = 0 (Figure C4); for a situation where there is very little potential for vertical compression, εv = 0.1%, and

Appendix D: Effectiveness of field compaction by impact loading

with αT =1/500 and hence αT/εv = 2, there is likely to be a problem only with a small burial depth and the width of the exclusion zone, WE, should be less than 0.25 H  cot β = 0.5 (Figure C5); for a situation where there is potential for significant vertical compression, εv = 0.5%, and with αT =1/500 and hence αT/εv = 0.4, there is likely to be a problem in most cases and the normalised width of the zone where the tilt is unacceptable, WEn, is close to 1 (ie WE = H) for 0.5 < Dn < 3  cot β = 2 (Figure C7); for a situation where there is potential for large vertical compression, εv = 2%, and with αT =1/500 and hence αT/εv = 0.1, there is likely to be a problem and the length of the zone where the tilt is not tolerable is depende nt on the magnitude of Dn; with Dn = 0, WE is nearly 2.5 H; with Dn = 5, WE is about

compaction of building wastes (case history 26). Dynamic compaction has been described in Chapter 8. Despite the enormous difference in scale between field and laboratory, the energy input per unit volume of soil, E/V, is comparable. There is no generally accepted universal theory which relates energy input during dynamic compaction to improved soil properties and effective depth of compaction. The energy required to compact a partially saturated fill is a complex function of the method of compaction (mass and dimensions of the weight, height of fall, spacing of grid points, number of blows per grid point, number of passes), and the soil properties (fill type, particle size distribution, existing density and moisture content, depth to be compacted and degree of improvement required). Assessment of required energy is based largely on past experience and empirical

5 H. To delineate the exclusion zone it is necessary to determine the distance from the top of the highwall to the beginning of the zone, xE. The normalised distance from the top of the highwall to the beginning of the exclusion zone, xEn = xE/H , can be determined as follows:

xEn = (cot β – WEn) [Dn/(1 + 2Dn)] If xEn is positive, the exclusion zone starts over the highwall, whereas if xEn is negative the exclusion zone starts over the shallow depth of fill behind the highwall.

Repeated impacts of a heavy weight, sometimes refe rred to as a tamper or pounder, onto the surface of a partially saturated fill compact the fill by reducing the air voids. This is analogous to the mechanism of compaction in a standard laboratory Proctor compaction test. Table D1 gives a comparison between, first, the energy input in two standard laboratory compaction tests (BS 1377-4:1990) and, secondly, that used in dynamic compaction of a clay fill at Corby (case history 2) and in rapid impact

correlations. thea correlations do not take account of all these factors,As only relatively crude assessment can be made of the required energy. There is some correlation between the depth of compaction and the energy per blow. This has usually been expressed as:

ze = k(MH )0.5 where ze is the depth of compacted soil in m, M is the mass in tonnes, H is the height of fall in m, and k is a coefficient which depends on fill type. Mitchell (1981) and Mayne et al (1984) have presented summaries of data from many sites which showed that generally it has bee n found that 0.3 < k < 0.8. Field measurements by BRE indicated k = 0.35 on a stiff clay fill (case history 17) and k = 0.4 on old domestic refuse (case history 2). The values in Table D2 give an indication of likely magnitudes of k. Clearly this approach is an over-simplification of the problem and in practice the depth of compacted soil will be influenced by other factors such as the total energy applied and the base area of the tamper. The quoted values of k are likely to underestimate the depth of compaction where n > 10. Leonards et al (1980) suggested that de gree of compaction for a sand measured by cone resistance, qc, correlated best with the product of energy per blow and total energy applied per unit area. Such a relationship suggests that if a lower energy per impact is used, then a

174

Appendix D

Table D1 Impact compaction in the laboratory and in the field W H A z e 2 Ty pe ( k N) (m) (m ) n (m)

WH

W H/A

(k Nm)

2 ( kNm /m )

E/At

E /V

( k Nm / m2)

(kNm/m3)

ze /A0.5

Lab*

0.0245

0.3

0.001 96

6.1

0.038

0.0074

3.8

23

600

0.9

Lab*

0.0441

0.45

0.00196

6.1

0.023

0.0199

10.1

62

2680

0.5

Field†

150

20

Field‡

70

1

4 1.8

3.7 39

6

3000

750

2800

470

3.0

4

70

39

1500

375

3.0

* BS 1377-4:1990 clause 3 † Case history 2 ‡ Case history 26 Notation used in Table D1 W is the weight of the tamper H

is height of fall of the tamper

A

is area of tamper in contact with fill on impact

n

is the average number of impacts at any point; thus if treatment comprises a total ofN impacts distributed over a total areaAt, n = (NA)/At

ze

is depth of compacted fill

WH

is energy per impact

WH/A

is energy per impact per area of impact

E/At

is average energy input per unit area [E/At = (WHn)/A]

E/V

is average applied energy per unit volume of fill [E/V = (WHn)/Aze]

ze/A0.5 is ratio of depth of compacted fill to impact dimension circular weight with diameter b: A0.5 = 0.89b square weight with side b: A0.5 = 1.0b

Table D2 Typical values of k for dynamic compaction of fills (n < 10) Fill type k (m/tonne)0.5

soil is given by:

Stiff clay fill

0.35

Oldrefuse

0.4

Parson’s tests were in the range 10 kPa < cu < 100 kPa. Threadgold (1978) presented data confirming the relationship for values of cu up to 700 kPa. Edwards (1978) stated that with an MCV of 16 and above (cu > 300 kPa), samples may not be compacted to air voids of 5%. In the range 50 kPa < cu < 300 kPa in which it is likely that dynamic compaction might be applied to a cohesive fill, it is reasonable to approximate the relationship to:

Rockfill Sandfill

0.5 0.5

higher total average energy input per unit area will be required. The effectiveness of the treatment and the economics of using it on a partially saturated clay fill de pend on the amount of compactive effort required to reduce the air voids in the fill to a minimum. The minimum air voids that can be obtained in a clay fill is usually 2% to 3% and a generally acceptable value after treatment would be 5%. Parsons (1978) described a laboratory test programme in which soils were compacted into a cylindrical container with a rammer of almost the same diameter as the container. For several different types of cohesive soil a relationship between undrained shear strength (measured by a vane) and moisture condition value (MCV) of the following form was found: logcu = a + b(MCV) Parsons defined MCV = 10logN where N was the number of blows required to compact the soil. In Parson’s test the energy per blow was about 22 kN/m 3 and typical values for a and b were a = 0.8 and b = 0.11. Charles (1978) pointed out that these findings indicated that in this test the energy/unit volume, E/V, required to compact the

E/V = 4cu0.9

E/V = 2.5 cu This type of laboratory dynamic compaction is very efficient as the fill is restrained from lateral movement. E/V = 2.5 cu should therefore be considered a lower bound to the energy required in dynamic compaction in the field. The expression is important because it indicates that the required energy is a linear function of c . In order to discover the influence of various u parameters on the depth of e ffectiveness of impact compaction, some small-scale laboratory tests were carried out in a 100 mm diameter compaction mould constructed in such a way that it was possible to slice the soil sample horizontally into eight slices after compaction. By weighing these a profile of density versus depth could be obtained for the treated soil. Charles (1979a) presented results for tests on a sandy clay fill as follows:

ze/b = 0.4[ E/(Atbcu)]0.5

Appendix E where ze was the depth to which the soil was compacted to 5% air voids, and b was the diameter of the rammer. This expression takes no account of the proportion of the energy that was applied in a single blow. It can be expected that for a given value of E/At, the larger the value of n, the number of impacts at any point, the smaller will be ze. However, in the range 3 < n < 10, the tests showed little effect. Above n = 10 some reduction did occur. The expression can be rearranged to show the relationship between energy per unit volume required to compact the fill and ze/b and cu.

E / V = 6 cu ( z e / b ) The relationship holds for the sandy clay fill for 0.5< ze/b < 3, although there is a good deal of scatter in the experimental results. T ests carried out on a plastic clay have given a slightly different relationship. This relationsh ip for a sandy clay can be compared with the relationship derived from the MCV test results on many types of cohesive soil, E/V = 2.5cu. In that test the compaction is very efficient as the rammer is almost the same size as the compaction mould. Field data suggest that a rather smaller amount of energy than that indicated by the above relationship will suffice. Table D3 gives some typical values of energy per unit volume that have been used to compact different types of fill. Such values can only form a crude guide to the required energy at a particular site. Other factors such as the initial density

Appendix E: Analysis of stone columns under widespread load The case of fully penetrating stone columns subjected to a widespread load is illustrated in Figure E1. Without stone columns the compression of the fill would be one dimensional such that: εvo = q/Do

where εvo is the vertical strain induced in the untreated fill q is the applied pressure

Do is the constrained modulus of the untreated fill. The stone columns effectively reduce the compressibility of the fill into which they are installed. In a clay fill the length of the drainage path will be re duced and consolidation accelerated. The vertical strain in the stiffened fill is reduced to εv and the improvement can be expressed as a settlement reduction factor sr such that: εv = sr εvo

of the fill are important. Table D3 Typical energy input for dynamic compaction of fills (n < 10) Fill type E /V (kNm/m3) Stiff clay fill

600

Oldrefuse

400

Rockfill Sandfill

175

300 200

Figure E1 Fully penetrating stone columns under wide load

176

Appendix E large compared with the in-situ stresses prior to loading. On the basis of these assumptions it can be shown that the settlement reduction factor sr is as follows:

s r = ( D o/ D )

1 [1 + {( RA r)/(1 – Ar)} – Ar{1 + (R/2)}{1 – Ko}]

where D is the constrained modulus of the fill after columns installed Do is the constrained modulus of the fill without

Figure E2 Stone column unit cell

sr = 1 indicates no improvement due to treatment, sr = 0.5 means that settlement has been reduced by 50%. The situation can be simplified to a unit cell composed of a single column surrounded by an annulus of fill laterally confined at its outer boundary as shown in Figure E2. The area replacement factor Ar = (dg/de)2 = 0.91(dg/s)2. This problem is amenable to analysis and has been addressed by a number of workers including the following:  Goughnour and Bayuk (1979) based their analysis on elasto-plastic behaviour of the stonethat column Balaam and Booker (1981) assumed both column and surrounding soil behaved elastically  Goughnour (1983) presented simplified charts  Charles and Soares (1983) demonstrated that in most practical situations the stone column will be undergoing plastic yield  Van Impe and de Beer (1983) assumed the granular material was yielding  Balaam and Booker (1 985) extended their e arlier work by taking account of the yielding of the columns  Priebe (1995) presented the results of an analysis based on plastic yielding of columns at constant volume. 

Van Impe and de Beer (1983) and Balaam and Booker (1985) have compared their methods with work by Priebe. Similar results have been obtained from the different methods of analysis; typically with A = 0.1, sr = 0.7 to 0.8 ; with Ar = 0.3, sr = 0.4 to 0.5. Therresults depend, of course, on the parameters which are used for the shear strength of the granular columns, etc. A simple type of analysis is now outlined in which the composite action of the fill reinforced by stone columns is analysed on the following basis:  the columns yield at constant volume and constant principal effective stress ratio  the surrounding fill compresses elastically and the stress–strain relationships in the fill surrounding each column can be described by thick cylinder theory  the stress increments imposed by the applied load are

columns R is the principal stress ratio in the column Ar is the area replacement factor Ko is the coefficient of earth pressure at rest in the fill. To obtain a typical re lationship between sr and Ar the following assumptions have been made:  the constrained modulus is unaffected by column installation, ie Do/D = 1  for the column, φ'cv = 37° and therefore R = 4.0  for the fill, φ' = 30° and therefore Ko = 1 – sinφ' = 0.5. Using these values the equation becomes:

sr = (1 – Ar)/ [1 + 1.5Ar(1 + Ar)] The assumption Do/Dcould = 1 can be questioned installation of thethat columns significantly affectasthe constrained modulus of the fill in the following ways:  installation of the columns may densify the fill, and so D will be increased  the beneficial effects of any small preloading or ageing effect in the untreated fill is likely to be destroyed by installation of the columns thus reducing D  remoulding a clayey fill could lead to a decrease in D  the increase in vertical stress in the fill, caused by the applied pressure in the situation where columns have been installed, will be smaller than it would have been without columns. In many fills D is stress dependent and will be smaller at these lower stress levels. The initial stresses present in the fill due to self weight before treatment have been ignored in the analysis. This is reasonable in attempting to correlate the analysis with laboratory oedometer tests, but is less justifiable in using the analysis to predict field behaviour. In the field such an assumption will be realistic only close to the ground surface. Juran and Guermazi (1988) carried out laboratory triaxial tests on soft silty soil with a central sand column. They concluded that consolidation and partial drainage during loading had an important effect on the load transfer mechanism. Large-scale laboratory tests were carried out at BRE in a 1 m diameter oedometer as shown in Figure E3 (Charles and Watts, 1983). The effect of the granular

Appendix E

177

Figure E3 Tests on granular columns in soft clay in 1 metre diameter oedometer

Figure E4 Compression of soft clay with granular columns of different diameters

column diameter on the compression of the soft clay is shown in Figure E4. These results are compared with predictions from elastic theory in Figure E5 which also shows the relationship derived above. At small values of Ar, the plastic columns predict significantly less improvement than the elastic columns. Nevertheless the laboratory tests showed even less improvement than the plastic analysis indicated. It can be concluded that with Ar < 0.1, improvement is unlikely to be significant. From practical considerations in the field generally Ar < 0.3. With Ar = 0.3, sr = 0.5 might typically be achieved.

Figure E5 Comparison of laboratory results with theoretical predictions for fully penetrating granular columns under widespread load

178 Appendix F: Model specification for engineered fill This model specification (Trenter and Charles, 1998) is intended to be used in connection with the construction of low-rise buildings on engineered fill, typically no greater than 50 000 m 3 in volume and no more than 5 m in average depth. The specificati on should be suitable for most purposes but there may be special conditions at some sites which should be taken into account in arriving at a properly engineered fill. It is e mphasised that clay fill s can be at least as susceptible to settlement or heave due to climatic, vegetation or other e ffects, as naturally occurring cohesive soils. The specification is for contracts which are designed, let and supervised by a consulting engineer. However, the appropriate wording may readily be adapted for other forms of contract, such as design and build. It is assumed that there will have been an adequate site investigation prior to the works and that geotechnical properties of relevance to enforcing the specification will have been measured.

Appendix F (e) steel slag (f ) spent oil shale (g) incinerator waste (h) some demolition and construction indust ry waste. Such fill shall be precluded from use in designated zones including locations where ground-water may rise to the level of the underside of the deepest foundation and where its use will be condemned by the appropriate authorities on pollution grounds. Such fill shall not be placed to a depth less than 1 metre from the underside of the deepest foundation. [1.6] Special fill shall comprise material which would otherwise be classified as general fill but which contains durable well graded natural sand and natural gravel or crushed rock, other than argillaceous rock, or durable clean crushed demolition rubble of similar particle size and free from any contaminants. Such fill may be employed as capping layers be neath structure foundations, beneath roads or as backfill to retaining walls.

1. Engineered fill

[1.1] Engineered fi ll is defined as fill which is selected, placed and compacted to an appropriate specification so that it will exhibit the required engineering behaviour. [1.2] Fill shall be classified as follows: (a) unsuitable fill

2. Selection of end product requirements

(b) restricted general fillfill (c) (d) special fill [1.3] Unsuitable fill shall comprise any material so designated by the Engineer and shall include: (a) cohesive soils having a liquid limit in excess of 90% or plasticity index in excess of 65% (b) chalk having a fine fraction (<400 µm) in excess of 10% at the borrow pit (c) any material containing topsoil, wood, peat or lignite (d) any material containing biodegradables (e) any material containing scrap metal (f ) frozen or water-logged substances, (g) material from contaminated sites (h) material which by virtue of its particle size or shape cannot be properly and effectively compacted (eg some slate wastes). Unsuitable fill shall not be used at any location or part of the site, including landscaped areas. [1.4] General f ill shall comprise all fill except unsuitable fill, restricted fill and special fill. [1.5] Restricted fill s hall comprise material which would otherwise be classified as general fill but which contains minerals hostile to the built e nvironment and shall include: (a) pyritic shales (b) gypsiferous clays (c) burnt colliery discard (d) pulverised fuel ash

(i) natural section 3] moisture content [BS 1377-2:1990 (ii) liquid and plastic limits for cohesive soils [BS 1377-2:1990 sections 4 and 5] (iii) compaction tests to dete rmine maximum dry density and optimum moisture content at the appropriate compactive effort (2.5 and 4.5 kg rammer) [B S 1377-4:1990 section 3] (iv) particle density (specific gravity) to assist in evaluating the compaction test [BS 1377-2:1990 section 8]. (b) A graph such as that illustrated in Figure 48. [2.2] The graph will show the dry density plotted against moisture content for the 2.5 kg and 4.5 kg rammer method compaction tests; the corresponding optimum moisture contents and maximum dry densities; and the 0% and 5% air voids lines. [2.3] By reference to lines OPQR illustrated on Figure 48, the required level of compaction will be indicated by selecting appropriate moisture content and dry density values. This level of compaction shall form the basis of the compaction specification.

[2.1] On the basis of the results of a site investigation which the Engineer shall have carried out, the Engineer shall, at the time of tende r, provide the Contractor with the following: (a) The results of the tests shown below:

3. Preparation of site

[3.1] On side-long ground, drainage grips or trenches shall be excavated uphill of the area to be filled and compacted. Drainage shall be effected without causing siltation or erosion and water shall be

Appendix F

179

disposed of in a manner to be agreed by the Engineer. [3.2] The area to be filled, whether an existing excavation or otherwise undisturbed ground, shall be graded to falls, and sump pumping or other suitable de-watering facilities shall be provided by the Contractor to keep the base of the excavation dry at all times. [3.3] Where the area to be filled comprises an existing excavation, the excavation shall be inspected and subsequently monitored by the Contractor, to ensure that there is no danger of its collapse during

achieve the specified end product and shall, as far as practicable, be brought up at a uniform rate so that all parts of the site reach finished (formation) level at the same time. [5.2] The compaction plant selected, the number of passes made and the fill layer thickness and moisture content used shall have regard to the specified end product and the means and manner of control testing. [5.3] Where several different types of fill material (all meeting the requirements of Clause 1 of this specification) are to be employed, they shall be

the works with consequences for safety, for existing buildings, or for other construction adjoining. [3.4] All topsoil shall be stripped and, where required for further use, stockpiled in an area provided by the Contractor and agreed by the Engineer. [3.5] All soft and compressible soils or exi sting fill shall be removed and run to spoil in dumps provided by the Contractor and agreed by the Engineer (including licensed tips in the case of certain contaminated materials). The work shall be accomplished in such a way that there is no undercutting of the sides of existing excavations. [3.6] Existing foundations or ledges of hard rock, roots of trees or former pipelines or services at the base of the area to be filled shall be excavated and replaced with compacted general fill which shall

deposited in such a way that all parts of the site receive roughly equal amounts of a given material, in roughly the same sequence, thus ensuring a uniform distribut ion of fill types over the whole fill thickness. [5.4] The Contractor shall take all necessary steps to ensure that the fill is placed at the moisture content necessary to achieve the specified level of compaction and shall, where necessary, add water to or dry the fill, in order to obtain this value. Where it is necessary to add water, this shall be done as a fine spray and in such a way that there is time for the water to be absorbed into the fill before being rolled by the plant. [5.5] Cobbles, boulders, rock or waste fragments whose largest dimension is greater than two- thirds of the loose layer thickness shall not be incorporated into

be compacted to the same specification as that adopted for subsequent compaction works. [3.7] Where unsuitable material has been exca vated, the underlying natural ground shall be compacted to the same specification as that adopted for subsequent compaction works.

the fill fill.shall be placed and left uncompacted at the [5.6] No end of a working day. Compac ted fill shall be graded to falls to ensure free run-off of rain-water without ponding. [5.7] Compaction plant and compaction method shall be selected having regard to the proximity of existing trenches, excavations, retaining walls or other structures and all work shall be performed in such a way as to ensure that their e xisting stability is not impaired. [5.8] If weather conditions are such that the specified moisture content and density values cannot be achieved, the Contractor shall cease work until such time that the fill can be placed and compacted to meet specification requirements. [5.9] If the results of control tests (Clause 6.3) indicate that the fill is being placed and compacted in such a way that the desired level of compaction is not being achieved, the Contractor shall further compact or, if necessary, shall excavate the affected work and replace with new fill, compacted to meet the specification requirements. [5.10] If the results of control tests (Clause 6. 3) indicate that antecedent weather conditions (such as frost or heavy rain) have caused deterioration of finished work such that the work no longer mee ts specification, the Contractor shall take such steps as are necessary to bring the fill to specification requirements at the Contractor’s own cost.

4. Disposition of fill

[4.1] Where construction is required upon fill placed over sloping natural ground, and where fill thickness is less than 5 m, the natural ground shall be benched, with the maximum vertical height of each bench not exceeding 300 mm (see Figure 49). [4.2] Where a development contains landscaped areas on which no structures will be built, the underlying fill shall be selected, placed and compacted in the same way as the engineered fill, unless otherwise directed by the Engineer. Where some relaxation of the specification for fill compaction underlying landscaped areas is permitted, there shall be a transition zone between the fill underlying the landscaped area and the fill underlying the structure. The dimensions of the transition zone will depend on the degree to which fill compaction was relaxed for the fill in the landscaped area (see Figure 50). 5. Placing and compacting fill

[5.1] Fill shall be placed and compacted in near horizontal layers of the thicknesses required to

180

Appendix F

6. Control testing

[6.5] The Engineer will, from time to time and with reasonable notice, request the Contractor to make available equipment to enable the Engineer to perform control tests. The results of these tests shall be used by the Engineer in assessing the Contractor’s performance.

[6.1] The end product requirements selected in Clause 2.1 shall be controlled by in-situ and laboratory testing as follows: (a) in -situ dry density (B S 1377-9:1990, section 2) and moisture content determinations (BS 13772:1990, section 3) (b) where required these tests shall be augmented by moisture content–dry density relationships (BS 1377-4:1990 section 3) and particle density (BS 1377-2:1990, section 8). [6.2] At least 14 working days before the start of site work, the Contractor shall provide the Enginee r, for approval, with a list of the equipment that the Contractor proposes to use to undertake these tests. [6.3] Control tests shall be performed throughout the fill at such frequency as shall be directed by the Engineer. Test locations shall be agreed with the Engineer. [6.4] When requested by the Engineer, the Contractor shall make available a plot of in-situ dry de nsities against in-situ moisture content results on a graph such as that illustrated in Figure 48, showing that the results lie within or above the area O PQR or such area as has been selected by the Engineer. Should any results lie outside the selected area, the Contractor shall provide the Engineer with proposals for rectifying the e xisting situation and for improving future performance.

7. Monitoring of fill performance

[7.1] If instructed by the Engineer , the Contractor shall make arrangements for the performance of the fill, once placed, to be monitored. Monitoring may take one or more of the following forms: (a) optical levelling of surface monuments (b) standpipes or piezometers (c) load tests (d) other methods as directed by the Engineer. [7.2] The Contractor shall, within 21 working days of receiving notification of the Engineer’s intention to monitor fill performance, arrange for the procurement and supply of the equipment to the Engineer’s written specification and shall inform the Engineer of the date on which the equipment installation shall begin. The specification shall include: (a) a full description of the nature and type of instrument and the purpose it fulfils (b) the number required and the locations or depths, or both, at which it is to be installed (c) the frequency, accuracy and duration for which any readings are to be taken.

181

References

Abbiss C P (1981). Shear wave measurements of the elasticity of the ground. Géotechnique, vol 31, no 1, March, pp 91–104. A test of the generalised theory of Abbiss C P and Lewin P I (1990). visco-elasticity in London Clay.Géotechnique, vol 40, no 4, pp 641–646. Aldrich T E, Torres C and Lilquist D (1998). Risk assessment and soil contamination.Land Contamination and Reclamation, vol 6, no 4, pp 207–213. Alker S, Joy V, Roberts P and Smith N (2000). The definition of brownfield. Journal of Environmental Planning and Management, vol 43, no 1, pp 49–69. Andrus R D and Youd T L (1989). Penetration tests in liquefiable gravels. Proceedings of 12th International Conference on Soil Mechanics and Foundation Engineering, Rio de Janeiro, vol 1, pp 679–682. Andrus R D, Chung R M, Stokoe K H and Bay J A (1998).

Ballisager C C and Sorensen J L (1981). Flyash as fill material. Proceedings of 10th International Conference on Soil Mechanics and Foundation Engineering, Stockholm, vol 2, pp 297–301 Barber E G, Jones G T, Knight P G K and Miles M H (1972). PFA utilization. Central Electricity Generating Board. Barnett G (1976).Opening address by Parliamentary Under Secretary of State.Proceedings of Land Reclamation Conference, Grays, Essex, pp 2-4. Thurrock Borough Council. Barry D (1991).Hazards in land reycling. Chapter 3 of Recycling derelict land (ed G Fleming), pp 28–63. Thomas Telford, London. Beever P F (1989).Subterranean fires in the UK – the problem.BRE Information Paper IP3/89. CRC, Garston. Bell F G and Cashman P M (1985). Groundwater control by groundwater lowering. Groundwater in engineering geology (eds J C Cripps, F G Bell and M G Culshaw). Proceedings of 21st annual conference of Engineering Group of Geological Society, Sheffield,

Delineation of densified sand at Treasure Island by SASW testing. September 1985, pp 471–486. Geological Society, London, 1986. Geotechnical site characterisation.Proceedings of 1st international Bell A L, Slocombe B C, Nesbitt A M and Finey J T (1986). conference, Atlanta, vol 1, pp 459–464. Balkema, Rotterdam. Vibro-compaction densification of a deep hydraulic fill.Building on Arato A (1992). Deformation due to cavities.Soil settlement effects marginal and derelict land. Proceedings of conference, Glasgow, May on buildings (ed E Dulacska), pp 54–97. Elsevier, Amsterdam. 1986, pp 791–797. Thomas Telford, London, 1987. Birth defects cluster found near toxic Association of Geotechnical and Geoenvironmental SpecialistsBevan S and Hind G (1999). (1998a).Code of conduct for site investigation. AGS, Beckenham, dump. Sunday Times, 11 April. Kent. Bishop A W and Vaughan P R (1972). Consolidation of fine grained Association of Geotechnical and Geoenvironmental Specialistsdredged material after hydraulic deposition (2 vols). National Ports (1998b).Guidelines for good practice in site investigation. AGS, Council, London. Beckenham, Kent. Bishop A W, Hutchinson J N, Penman A D M and Evans H E Association of Geotechnical and Geoenvironmental Specialists(1969). Geotechnical investigations into the causes and circumstances of the disaster of 21st October 1966. A selection of (1998c).AGS Guide: the selection of geotechnical soil laboratory testing. AGS, Beckenham, Kent. technical reports submitted to the Aberfan Tribunal, Welsh Office. Association of Geotechnical and Geoenvironmental SpecialistsHMSO, London. (2000).Guidelines for combined geoenvironmental and geotechnical The strength of clay fill subgrades: Black W and Lister N W (1978). investigations. AGS, Beckenham, Kent. its prediction and relation to road per formance.Clay fills. Proceedings Atkinson M F (1993). Structural foundations manual for low-rise of conference, London, November 1978, pp 37–48. Institution of Civil buildings. Spon, London. Engineers, London, 1979. Analysis of rigid rafts Wetting collapse in Balaam N P and Booker J R (1981). Blanchfield R and Anderson W F (2000). supported by granular piles. International Journal of Numerical and opencast coalmine backfill. Proceedings of Institution of Civil Analytical Methods in Geomechanics, vol 5, pp 379–403. Engineers, Geotechnical Engineering, vol 143, no 3, July, pp 139–149. Balaam N P and Booker J R (1985). Effect of stone column yield on Blight G E, Fourie A B, Shamrock J, Mbande C and Morris J W F settlement of rigid foundations in stabilized clay.International Journal of (1999).The effect of waste composition on leachate and gas quality: a Numerical and Analytical Methods in Geomechanics, vol 9, no 4, pp study in South Africa. Waste Management and Research, vol 17, pp 331–351. 124–140. The behaviour of foundations The strength and dilatancy of sands. Balaam N P and Poulos H G (1983). Bolton M D (1986). Géotechnique, vol 36, no 1, March, pp 65–78. supported by clay stabilised by stone columns.Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Brand E W (1972).Some observations on the control of density by Helsinki, vol 1, pp 199–204. Balkema, Rotterdam. vibration. Evaluation of relative density and its role in geotechnical projects involving cohesionless soils, pp 121–132. STP 523, ASTM, Ball, M J (1979).The investigation of a quarry waste silt lagoon on the M42 motorway. Engineering behaviour of industrial and urban fill. Philadelphia. Proceedings of symposium, Birmingham, pp C11–C24. Midland Geotechnical Society.

182

References

). Brandon T L, Duncan J M and Gardner W S (1990 Hydrocompression settlement of deep fills.ASCE Journal of Geotechnical Engineering, vol 116, no 10, October, pp 1536–1548. Bredenberg H and Broms B B (1983). Lime columns as foundations for buildings. Piling and ground treatment. Proceedings of international conference, London, March 1983, pp 133–138. Thomas Telford, London, 1984. The Wak test to check the Briaud J L, Liu M L and Lepert P (1989). increase in soil stiffness due to dynamic compaction.Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 107-122. STP 1070, ASTM, Philadelphia, 1990. Opencast coal mining in Great Britain (2nd British Coal (1988). edition). British Coal Opencast Executive, Mansfield. BSI (1981).Code of practice for earthworks. BS 6031:1981. BSI, London. BSI (1985).Guide to use of industrial by-products and waste materials in building and civil engineering. BS 6543:1985. BSI, London. BSI (1986).Code of practice for foundations. BS 8004:1986. BSI, London. BSI (1990a).Methods of test for soils for civil engineering purposes. Parts 1-9. BS 1377:1990. BSI, London. BSI (1990b).Stabilized materials for civil engineering purposes. Parts 1 and 2. BS 1924:1990. BSI, London. BSI (1995).Eurocode 7: Geotechnical design – Part 1: General rules. DD ENV 1997-1:1995. BSI, London. BSI (1999).Code of practice for site investigations. BS 5930:1999. BSI, London. BSI (2000a). Eurocode 7: Geotechnical design – Part 2: Design assisted by laboratory testing. DD ENV 1997-2:2000. BSI, London. BSI (2000b).Eurocode 7: Geotechnical design – Part 3: Design assisted by field testing. DD ENV 1997-3:2000. BSI, London. BSI (2001).Investigation of potentially contaminated sites – Code of practice. BS 10175:2001. BSI, London. Broch E and Franklin J A (1972). The point load strength test. International Journal of Rock Mechanics and Mining Science, vol 9, pp 669–697. Broms B B (1999). Keynote lecture. Design of lime, lime/cement and cement columns. Dry mix methods for deep soil stabilisation (eds H Bredenberg, G Holm and B B Broms). Proceedings of international conference, Stockholm, pp 125–153. Balkema, Rotterdam. Broms B B and Anttikoski U (1983). Soil stabilization – general report. Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 3, pp 1289–1301. Brown J and Snell P A (1978). The use of PFA in land reclamation. Ash technology and marketing. Proceedings of conference, paper 2.3. Central Electricity Generating Board, London. BRE (1975).Fill and hardcore. Digest 222. (Revised 1979, Reprinted 1981.) CRC, Garston. BRE (1983).Fill. Part 1: classification and load carrying characteristics. Part 2: site investigation, ground improvement and foundation design. Digests 274 and 275. (Revised 1991 and 1992, respectively.) CRC, Garston.

BRE (1993a).Low-rise buildings on shrinkable clay soils: Part 1. Digest 240. CRC, Garston. BRE (1993b).Underpinning. Digest 352. CRC, Garston. BRE (1993c).Site investigation for low-rise building: trial pits.Digest 381. CRC, Garston. BRE (1993d).Site investigation for low-rise building: soil description. Digest 383. CRC, Garston. BRE (1995a).Assessment of damage in low-rise buildings with particular reference to progressive foundation movement.Digest 251. CRC, Garston. BRE (1995b).Simple measuring and monitoring of movement in low-rise buildings. Part 2: settlement, heave and out of plumb.Digest 344. CRC, Garston. BRE (1995c).Site investigation for low-rise building: direct investigations. Digest 411. CRC, Garston. BRE (1997).Low-rise buildings on fill: classification and load-carrying characteristics. Digest 427: Part 1. CRC, Garston. BRE (1998a).Low-rise buildings on fill: site investigation, ground movement and foundation design.Digest 427: Part 2. CRC, Garston. BRE (1998b). Low-rise buildings on fill: engineered fill. Digest 427: Part 3. CRC, Garston. BRE (2000).Specifying vibro stone columns. BRE Report BR391. CRC, Garston. BRE (2001). Concrete in aggressive ground. Special Digest 1. CRC, Garston. Building Research Station (1949). Building on made-up ground or filling. First Series Digest 9. CRC, Garston. Building Research Station (1972). Fill and hardcore. Digest 142. CRC, Garston. Deflection in M1 motorway Buist D S and Dutch W G (1984). carriageway due to settlement of backfilled opencast coal site, Strelley, Nottinghamshire, England.Ground movements and structures (ed J D Geddes). Proceedings of 3rd international conference, Cardiff, July 1984, pp 496–509. Pentech Press, Plymouth, 1985. Burford D (1991). Surcharging a deep opencast backfill for housing development. Ground Engineering, September, pp 36-–39. Burford D and Charles J A (1991). Long term performance of houses built on opencast ironstone mining backfill at Corby, 1975–1990. Ground movements and structures (ed J D Geddes). Proceedings of 4th international conference, Cardiff, July 1991, pp 54–67. Pentech Press, London, 1992. Shaft friction of piles in clay – a simple Burland J B (1973). fundamental approach.Ground Engineering, vol 6, no 3, pp 30, 32, 37, 38, 41, 42. ). Settlement of foundations on Burland J B and Burbidge M C (1985 sand and gravel. Proceedings of Institution of Civil Engineers, Part 1, vol 78, December, pp 1325–1381. Burland J B and Wroth C P (1974). Settlement of buildings and associated damage.Settlement of structures. Proceedings of conference, Cambridge, April 1974, pp 611–654. Pentech Press, London, 1975. Behaviour of Burland J B, Broms B B and De Mello V F B (1977).

BRE (1986).Choosing piles for new construction. Digest 315. CRC, Garston. BRE (1987a).Site investigation for low-rise building: desk studies. Digest 318. CRC, Garston. BRE (1987b).Site investigation for low-rise building: procurement. Digest 322. CRC, Garston. BRE (1989a).Simple measuring and monitoring of movement in low-rise buildings. Part 1: cracks. Digest 343. CRC, Garston. BRE (1989b).Site investigation for low-rise building: the walk-over survey. Digest 348. CRC, Garston. BRE (1991).Construction of new buildings on gas contaminated land. Report BR212. CRC, Garston. BRE (1992). Hardcore. Digest 276. CRC, Garston.

foundations and structures.Proceedings of 9th International Conference on Soil Mechanics and Foundation Engineering, Toyko, vol 2, pp 495–546. The ability of in-situ testing to Butcher A P and McElmeel K (1993). assess ground treatment. Engineered fills (eds B G Clarke, C J F P Jones and A I B Moffat).Proceedings of conference, Newcastle-uponTyne, pp 529–540. Thomas Telford, London. The use of Cabrera J G, Braim M and Rawcliffe J (1984). pulverised fuel ash for the construction of structural fills.Ash technology and marketing. Proceedings of 2nd international conference, pp 529–533. Central Electricity Generating Board, London. Card G B (1995) . Protecting development from methane. Report 149. CIRIA, London.

References

183

Carver M (1987).Underneath English towns. Fitzhouse Books, London. Central Electricity Generating Board (1967) . PFA data book: load bearing fill. CEGB. London. A parametric study of effect of Chang P W and Chae Y S (1987). vibration on granular soils.Soil dynamics and liquefaction (ed A S Cakmak). Developments in Geotechnical Engineering 42, pp 137–151. Elsevier, Amsterdam. Settlement performance of two Chang J C and Hannon J B (1976). test highway embankments on sanitary landfill.New horizons in construction materials, pp 139–157. Lehigh Valley PA; Envo Publishing Co. Charles J A (1973). Correlation between laboratory behaviour of rockfill and field performance with particular reference to Scammonden dam. PhD thesis, University of London. General report: Methods of treatment of clay fills. Charles J A (1978). Clay fills. Proceedings of conference, London, November 1978, pp 315–321. Institution of Civil Engineers, London, 1979. Charles J A (1979a). Contribution to discussion, session 8, Proceedings of 7th European Conference on Soil Mechanics and Foundation Engineering, Brighton, vol 4, pp 307–310. Charles J A (1979b ). Field observations of a trial of dynamic consolidation on an old refuse tip in the east end of London. Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, pp E1–E13. Midland Geotechnical Society. Charles J A (1984). Some geotechnical problems associated with the use of coal mining wastes as fill materials.Reclamation, treatment and utilisation of coal mining wastes. Proceedings of symposium, Durham, pp 50.1–50.16. National Coal Board Minestone Executive. Charles J A (1990). Laboratory compression tests and the deformation of rockfill structures.Advances in rockfill structures (ed E Maranha das Neves). Proceedings of NATO Advanced Science Institute, Lisbon, June 1990, pp 73–95. Kluwer, Dordrecht, 1991. Charles J A (1991). The causes, magnitudes and control of ground movements in fills. Ground movements and structures (ed J D Geddes). Proceedings of 4th international conference, Cardiff, July 1991, pp 3–28. Pentech Press, London, 1992. The depth of influence of loaded areas. Charles J A (1996). Géotechnique, vol 46, no 1, pp 51–61. Charles J A and Burford D (1987). Settlement and groundwater in opencast mining backfills. Proceedings of 9th European Conference on Soil Mechanics and Foundation Engineering, Dublin, vol 1, pp 289–292. Charles J A and Burland J B (1982). Geotechnical considerations in the design of foundations for buildings on deep deposits of waste materials. Structural Engineer, vol 60A, no 1, pp 8–14. Charles J A and Driscoll R (1981). A simple in-situ load test for shallow fill. Ground Engineering, vol 14, no 1, pp 31–36. The behaviour of Charles J A and Penman A D M (1988). embankment dams with bituminous watertight elements.Transactions of 16th International Congress on Large Dams, San Francisco, vol 2, pp 693–705.

The influence of confining Charles J A and Watts K S (1980). pressure on the shear strength of compacted rockfill.Géotechnique, vol 30, no 4, pp 353–367. Charles J A and Watts K S (1983). Compressibility of soft clay reinforced with granular columns.Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 1, pp 347–352. Charles J A and Watts K S (1996). The assessment of the collapse potential of fills and its significance for building on fill.Proceedings of Institution of Civil Engineers, Geotechnical Engineering, vol 119, January, pp 15–28. Treated ground: objectives, Charles J A and Watts K S (2000). properties and performance.Core Programme Funders Report. CIRIA, London. Charles J A, Naismith W A and Burford D (1977). Settlement of backfill at Horsley restored opencast coal mining site.Large ground movements and structures (ed J D Geddes). Proceedings of conference, Cardiff, July 1977, pp 229–251. Pentech Press, Plymouth, 1978. Treatment and Charles J A, Earle E W and Bur ford D (1978). subsequent performance of cohesive fill left by opencast ironstone mining at Snatchill experimental housing site, Corby.Clay fills. Proceedings of conference, London, November 1978, pp 63–72. Institution of Civil Engineers, London, 1979. Charles J A, Burford D and Watts K S (1981). Field studies of the effectiveness of ‘dynamic consolidation’.Proceedings of 10th International Conference on Soil Mechanics and Foundation Engineering, Stockholm, vol 3, pp 617–622. Charles J A, Hughes D B and Burford D (1984). The effect of a rise of water table on the settlement of backfill at Horsley restored opencast coal mining site, 1973–1983. Ground movements and structures (ed J D Geddes). Proceedings of 3rd international conference, Cardiff, July 1984, pp 423–442. Pentech Press, Plymouth, 1985. Charles J A, Burford D and Watts K S (1986). Improving the load carrying characteristics of uncompacted fills by preloading.Municipal Engineer, vol 3, no 1, pp 1–19. Charles J A, Abbiss C P, Gosschalk E M and Hinks J L (199 An 1). engineering guide to seismic risk to dams in the United Kingdom. BRE Report 210. CRC, Garston. Charles J A, Burford D and Hughes D B (1993). Settlement of opencast coal mining backfill at Horsley 1973–1992.Engineered fills. Proceedings of conference held in Newcastle-upon-Tyne, September 1993, pp 429–440. Thomas Telford, London. Chartered Institute of Public Finance and Accountancy (1988). Waste disposal statistics 1986–87 actuals. CIPFA, London. Chartered Institute of Public Finance and Accountancy (1998). Waste disposal statistics 1996–97 actuals. CIPFA, London. Cheyney A C (1983). Settlement of landfill.Proceedings of Landfill Completion Symposium, Harwell, pp 13–29. Choa V, Karunaratne G P,Ramaswamy S D, Vijiarat nam A and Lee S L (1979).Compaction of sand fill at Changi airport. Proceedings of 6th Asian Regional Conference on Soil Mechanics and

Charles J A and Skinner H D (2001a). Compressibility of foundation fills. Proceedings of Institution of Civil Engineers, Geotechnical Engineering, vol 149, no 3, pp 145–157. Discussion on Wetting Charles J A and Skinner H D (2001b). collapse in opencast coalmine backfill (Blanchfield and Anderson, 2000). Proceedings of Institution of Civil Engineers, Geotechnical Engineering, vol 149, no 2, pp 129–130. The delineation of building Charles J A and Skinner H D (2001c). exclusion zones over highwalls. Ground Engineering, vol 34, no 2, pp 28–33. Charles J A and Soares M M (1983). Estimation of settlement of ground treated by stone columns.Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 3, pp 1118–1120.

Foundation Engineering, Singapore, vol 1, pp 137–140. Chu J, Yan S W and Yang H (2000). Soil improvement by the vacuum preloading method for an oil storage station.Géotechnique, vol 50, no 6, December, pp 625–632. Clayton C R I (1980). The collapse of compacted chalk fill. Proceedings of International Conference on Compaction, Paris, vol 1, pp 119–124. The collapse of slabs on a Clayton C R I and Simons N E (1980). compacted hoggin fill. Large ground movements and structures (ed J D Geddes). Proceedings of 2nd international conference, Cardiff, 1980, pp 655–670. Pentech Press. Site Clayton C R I, Matthews M C and Simons N E (1995). investigation. Blackwell Science, Oxford.

184

References

Experiences with loess as a foundation Clevenger W A (1956). material. ASCE Journal of Soil Mechanics and Foundations Division, vol 82, no SM3, paper 1025. Collins R J (1990). Case studies of floor heave due to microbiological activity in pyritic shales. Microbiology in civil engineering. Proceedings of Federation of European Microbiological Societies Symposium No 59, Cranfield Institute of Technology, pp 288–295. Spon, London. Colman R B (1977).Colliery shale – useful in formation. Civil Engineering (UK), July/August, pp 66–67. Construction News (1976).Rubbish tips root of house problem. 28 October, p 2. Construction News (1979).‘Egham experiment’ sinks to failure. 28 June, pp 1–2. Cooling L F (1945). Development and scope of soil mechanics. Lecture given at Institution of Civil Engineers, June 1945 .The principles and appplication of soil mechanics, pp 1–30. Institution of Civil Engineers, London, 1946. Cooling L F and Ward W H (1948). Some examples of foundation movements due to causes other than structural loads.Proceedings of 2nd International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol 2, pp 162–167. Covil C S, Luk M C W and Pickles A R (1997). Case history: ground treatment of the sand fill at the new airport at Chek Lap Kok, Hong Kong. Ground improvement geosystems: densification and reinforcement (eds M C R Davies and F Schlosser). Proceedings of 3rd international conference, London, pp 148–156. Thomas Telford, London. Cox D W (1978).Volume change of compacted clay fill.Clay fills. Proceedings of conference, London, November 1978, pp 79–86. Institution of Civil Engineers, London. Cragg C B H and Walker B P (1986). To accept or avo id settlement when developing marginal land. Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 59–67. Thomas Telford, London, 1987. Craig W H and Al-Khafaji Z A (1997). Reduction of soft clay settlement by compacted sand columns.Ground improvement geosystems: densification and reinforcement (eds M C R Davies and F Schlosser). Proceedings of 3rd international conference, London, pp 218–231. Thomas Telford, London. Crowhurst D (1987). Measurement of gas emissions from contaminated land. BRE Report. CRC, Garston. Fire and explosion hazards Crowhurst D and Beever P F (1987). associated with the redevelopment of contaminated land.BRE Information Paper IP2/87. The measurement of Crowhurst D and Manchester S J (1993). methane and associated gases from the ground. Report 131, CIRIA, London. Cuellar V, Monte J L and Valerio J (1998) Static . and dynamic elastic moduli for waste landfills.Environmental geotechnics (ed P S Seco e Pinto). Proceedings of 3rd international congress, Lisbon, vol 1, pp 325–329. Daily Telegraph(1976).Firm to buy back faulty homes.

Degen W S (1997).56 m deep vibro-compaction at German lignite mining area. Ground improvement geosystems: densification and reinforcement (eds M C R Davies and F Schlosser). Proceedings of 3rd international conference, London, pp 127–133. Thomas Telford, London. De Mello V F B (1977). Reflections on design decisions of practical significance to embankment dams: 17th Rankine Lecture. Géotechnique, vol 27, no 3, pp 279–355. Landfilling wastes. Waste Department of the Environment (1986). Management Paper No 26, HMSO, London. Department of the Environment (1991a). Survey of derelict land in England 1988 (2 vols). HMSO, London. Department of the Environment (1991b). Landfill gas. Waste Management Paper No 27. Second edition. (First edition published in 1989) HMSO, London. . Landfill design, Department of the Environment (1995) construction and operation practice design, construction and operational practice.Waste Management Paper 26B. HMSO, London. Department of the Environment (1996). Major coalfields regeneration plans unveiled. News release no 467, 6th November. Devenish M and Green P V (1980). High moisture slurry tailings – final report LR 350. Warren Spring Laboratory, Stevenage. Errors in construction on fills. Soil Mechanics Dmitrenko N M (1976). and Foundation Engineering (translated from Russian), vol 13, no 1, pp 7–11. Dobry R and Whitman R V (1972). Compaction of sand on a vertically vibrating table. Evaluation of relative density and its role in geotechnical projects involving cohesionless soils, pp 156–170. STP 523, ASTM, Philadel phia. Dobry R, Ladd R S, Yokel F Y, Chung R M and Powell D (1982). Prediction of pore water pressure buildup and liquefaction of sands during earthquakes by the cyclic strain method.NBS Building Science Series 138. US Department of Commerce, Washington. Dodt M E, Sweatman M B and Bergstrom W R (1987). Field measurements of landfill surface settlements.Geotechnical practice for waste disposal ‘87 (ed R D Woods). ASCE Geotechnical Special Publication No 13, pp 406–417. Dynamic consolidation of Downie A R and Treharne G (1979). refuse at Cwmbran. Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, pp E15–E24. Midland Geotechnical Society. Dudley J H (1970). Review of collapsing soils. ASCE Journal of Soil Mechanics and Foundations Division, vol 96, no SM3, pp 925–947. Dumbleton M J and West G (1976). Preliminary sources of information for site investigation in Britain. Report LR 403 (revised edition). TRRL, Crowthorne. Duplancic N (1989). Landfill deformation monitoring and stability analysis. Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 303–314. STP 1070, ASTM, Philadelphia, 1990. Eakin W R G (1983). Hard development potential and design techniques; part 2 – some technical problems.Reclamation of former

11 December. Daniel D E and Benson C H (1990). Water content – density criteria for compacted soil liners.ASCE Journal of Geotechnical Engineering, vol 116, no 12, December, pp 1811–1830. Davies S E and Reid J M (1997). Roadford Dam: geochemical aspects of construction of a low grade rockfill embankment.Ground chemistry implications for construction (ed A B Hawkins), pp 111–131. Balkema, Rotterdam. Davies M C R and Schlosser(eds) F (1997).Ground improvement geosystems – densification and reinforcement. Proceedings of 3rd international conference, London. Thomas Telford, London.

iron and steelworks sites. Proceedings of conference, Windermere, pp E7–E17. Durham County Council. Eakin W R G and Crowther J (1985). Geotechnical problems on land reclamation sites.Municipal Engineer, vol 2, October, pp 233–245. Ecclesiastes.In The new King James version of the Bible. New York, Thomas Nelson, 1985. Chapter 1 verse 10. Edil T B, Ranguette V J and Wuellner W W (1989). Settlement of municipal refuse. Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 225–239. STP 1070, ASTM, Philadelphia, 1990. Contribution to discussion. Clay fills. Edwards R J G (1978). Proceedings of conference, London, November 1978, pp 253–258. Institution of Civil Engineers, London, 1979.

References

185

Settlement and moisture El-Ehwany M and Houston S L (1990). Garvin S, Hartless R, Smith M A, Manchester S and Tedd P movement in collapsible soils. ASCE Journal of Geotechnical (1999).Risks of contaminated land to buildings, building materials and Engineering, vol 116, no 10, October, pp 1521–1535. services: a literature review. Environment Agency R&D Technical Eliassen R (1942). Load-bearing characteristics of landfills. Report P331. WRc, Swindon. Engineering News Record, 10 September, pp 103–105. Gazetas G, Tassoulas J L, Dobry R and O’Rourke M J (1985). Eliassen R (1947). Why you should avoid housing construction on Elastic settlement of arbitrarily shaped foundations embedded in refuse landfills. Engineering News Record, 1 May, pp 90–94. half-space. Géotechnique, vol 35, no 3, pp 339–346. Elsbury B R, Daniel D E, Sraders G A and Anderson D C (1990).Gens A, Alonso E E, Batlle F and Lloret A (1991). Prediction of Lessons learned from compacted clay liner.ASCE Journal of ground movements in partially saturated soils.Ground movements and Geotechnical Engineering, vol 116, no 11, November, pp 1641–1660. structures (ed J D Geddes). Proceedings of 4th international Emery J J (1979).Assessment of ferrous slags for fill applications. conference, Cardiff, July 1991, pp 461–484. Pentech Press, London, Reclamation of contaminated land. Proceedings of conference, 1992. Eastbourne, October 1979, pp F1/1–10. Society o f Chemical Industry, Gibson R E and Lo K Y (1961). A theory of consolidation for soils London, 1980. exhibiting secondary compression. Norwegian Geotechnical Institute Emery J J and Matchett D (1979). Use of stabilized industrial sludge Publication No 41. for land reclamation.Reclamation of contaminated land. Proceedings Residential development on Gilbert J A and Knipe C (1979). of conference, Eastbourne, October 1979, pp F2/1–8. Society of reclaimed opencast mining sites.Engineering behaviour of industrial Chemical Industry, London, 1980. and urban fill. Proceedings of symposium, Birmingham, pp E67–E79. Clifford’s Tower. English Heritage, London. Midland Geotechnical Society. English Heritage (1987). Guidance for safe Stresses under linearly loaded rectangular area. Environment Agency and NHBC (2000). Giroud J P (1970). development of housing on land affected by contamination. The ASCE Journal of Soil Mechanics and Foundations Division, vol 96, no Stationery Office, London. SM1, pp 263–268. Faraco C (1981).Deep compaction field tests in Puerto de la Luz. Gohl W B, Jefferies M G, Howie J A and Diggle D (2000). Proceedings of 10th International Conference on Soil Mechanics and Explosive compaction: design, implementation and effectiveness. Foundation Engineering, Stockholm, vol 3, pp 659–662. Géotechnique, vol 50, no 6, pp 657–665. Farrar D M (1978). Settlement and pore water pressure dissipation Goodwin A K and Holden J M W (1993). Performance of an within an embankment built of London Clay.Clay fills. Proceedings of engineered fill at Lounge opencast coal site. Engineered fills. conference, London, November 1978, pp 101–106. Institution of Civil Proceedings of conference, Newcastle-upon-Tyne, pp 413–428. Engineers, London, 1979. Thomas Telford, London. The operation of earthmoving plant The Stockley Park Farrar D M and Darley P (1975). Gordon D L, Lord J A and Twine D (1986). on wet fill. Laboratory Report 688, Transport and Road Research project. Building on marginal and derelict land. Proceedings of Laboratory, Crowthorne. conference, Glasgow, May 1986, pp 359–379. Thomas Telford, Ferguson D (1984). Reclaimed and backfilled areas. Ground London, 1987. movements and structures (ed J D Geddes). Proceedings of 3rd Gorle D and Thijs M (1989). Geosynthetic reinforced granular international conference, Cardiff, July 1984, pp 443–462. Pentech materials. Proceedings of 12th International Conference on Soil Press, Plymouth, 1985. Mechanics and Foundation Engineering, Rio de Janeiro, vol 1, pp Findlay J D and Sherwood D E (1986). Improvement of a hydraulic 715–718. fill site in Bahrain using modified heavy tamping methods.Building on Settlement of vertically loaded stone Goughnour R R (1983). marginal and derelict land. Proceedings of conference, Glasgow, May columns in soft ground. Proceedings 8th European Conference on Soil 1986, pp 13–25. Thomas Telford, London, 1987. Mechanics and Foundation Engineering, Helsinki, vol 1, pp 235–240. Fleming G(ed) (1991).Recycling derelict land. Thomas Telford, Goughnour R R and Bayuk A A (1979). Analysis of stone columns London. and soil matrix interaction under vertical load.Proceedings of Forde M C and Davis A G (1978). Trafficability studies on wet clay. International Conference on Soil Reinforcement, Paris, vol 1, pp Clay fills. Proceedings of conference, London, November 1978, 271–277. pp 107–112. Institution of Civil Engineers, London, 1979. Some observations on Gray J and Thomson G H (1979). Forssblad L (1981). Vibratory soil and rockfill compaction. Dynapac settlements of houses on treated urban fill.Engineering behaviour of Maskin, Solna, Sweden. industrial and urban fill. Proceedings of symposium, Birmingham, Foster H A and Glick G W (1938). Foundations on deep fill. pp E51–E66. Midland Geotechnical Society. Engineering News Record, vol 121, 22 September, pp 361–366. Green P V (1977).Ashington project – main report. Report LR 268, Fox E N (1948).The mean elastic settlement of a uniformly loaded Department of Environment Contract DGR 482/18. Warren Spring area at a depth below the ground surface.Proceedings of 2nd Laboratory, Stevenage. International Conference on Soil Mechanics and Foundation Mechanical improvement of soils below Greenwood D A (1970). Engineering, Rotterdam, vol 1, pp 129–132. Fox N H (1984).Pulverised fuel ash as structural fill. Ash technology and marketing. Proceedings of 2nd international conference, pp 495–499. Central Electricity Generating Board, London. Franklin J A and Chandra R (1972). The slake durability test. International Journal of Rock Mechanics and Mining Science, vol 9, pp 325–341. Gambin M P (1983).The Menard dynamic consolidation method at Nice airport. Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 1, pp 231–234. Ganji V, Gucunski N and Maher A (1997). Detection of underground obstacles by SASW methods – numerical aspects.ASCE Journal of Geotechnical and Geoenvironmental Engineering, vol 123, no 3, March, pp 212–219.

ground surface. Ground engineering. Proceedings of conference, London, pp 11–22. Institution of Civil Engineers, London. Greenwood D A (1990). Load tests on stone columns.Deep foundation improvements: design, construction and testing. STP 1089, ASTM, Philadelphia, 1991. Greenwood D A (1991). Vibrational loading used in the construction process. Cyclic loading of soils: from theory to design (eds M P O’Reilly and S F Brown), pp 434–475. Blackie, Glasgow. Greenwood D A and Kirsch K (1983). Specialist ground treatment by vibratory and dynamic methods; State-of-the-Art Report.Piling and ground treatment. Proceedings of international conference, London, March 1983, pp 17–45. Thomas Telford, London, 1984.

186

References

Contribution to discussion. Settlement of Guest J E (1974). structures. Proceedings of conference, Cambridge, April 1974, p 662. Pentech Press, London, 1975. Guthrie P (1999).Quoted by A Mylius in ‘Brown paradise’. New Civil Engineer, 28 January, pp 20–22. Gutt W, Nixon P J, Smith M A, Harrison W H and Russell A D (1974).A survey of the locations, disposal and prospective uses of the major industrial by-products and waste material.BRE Current Paper 19/74, CRC, Garston. Haegeman W and Van Impe W F (1998). SASW control of a vacuum consolidation on a sludge disposal.Geotechnical site characterisation. Proceedings of 1st international conference, Atlanta, vol 1, pp 473–477. Balkema, Rotterdam. Hagenaar J and Wolsleger A (1986). Prediction of long-term settlements in hydraulic fill. Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 27–36. Thomas Telford, London, 1987. Hansbo S (1977).Dynamic consolidation of rockfill at Uddevalla shipyard. Proceedings of 9th International Conference on Soil Mechanics and Foundation Engineering, vol 2, pp 241–246. Hansbo S (1993).Band drains. Ground improvement (ed M P Moseley), pp 40–64. Blackie, London. The choice of expedients in civil engineering Harding H J B (1946). construction.Proceedings of Institution of Civil Engineers, vol 5, pp 435–467. Harding H J B and Glossop R (1951). The influence of modern soil studies on the construction of foundations.Proceedings of Building Research Congress, London, Division 1, Part III, pp 146–156. Hardware Trade Journal (1975).Based on a sound Victorian foundation. 27 February, p 16. Harries C R, Witherington P J and McEntee J M (1995). Interpreting measurements of gas in the ground. Report 151, CIRIA, London. Harris M R R (1979). Geotechnical characteristics of landfilled domestic refuse. Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, pp B1–B10. Midland Geotechnical Society. House building Hart I M, Couper A S and Stephenson D C (1987). on bog land. Reclamation, treatment and utilization of coal mining wastes (ed A K M Rainbow). Proceedings of conference, pp 145–162. Elsevier, Amsterdam. Hartlen J and Ingers C (1981). Land reclamation using fine grained dredged material. Proceedings of 10th International Conference on Soil Mechanics and Foundation Engineering, Stockholm, vol 1, pp 145–148. Hatem D J(ed) (1998).Subsurface conditions – risk management for design and construction management professionals. Wiley, New York. Hathaway C M (1926). Water jets consolidate road fills.Engineering News Record, 11 March, pp 417–418. Hawkins A B (1998). Engineering significance of ground sulphates. Geotechnical site characterisation (eds P K Robertson and P W Mayne). Proceedings of 1st international conference, Atlanta, vol 1,

Head K H (1998).Manual of soil laboratory testing (3 vols), 2nd edition. Pentech Press, Plymouth. Health and Safety Commission (1994). Managing construction for health and safety, Construction (Design and Management) Regulations (1994), Approved Code of Practice L54. HSE Books. Heathcote F W L (1965). Movement of articulated buildings on subsidence sites. Proceedings of Institution of Civil Engineers, vol 30, pp 347–368. Hobbs D W (1963).A simple method for assessing the uniaxial compressive strength of rock. International Journal of Rock Mechanics and Mining Science, vol 1, pp 5–15. Hodgetts S J, Holden J M W, Morgan C S and Adams J N (1993). Specifications for and performance of compacted opencast backfills. Engineered fills. Proceedings of conference, Newcastle-upon-Tyne, pp 262–280. Thomas Telford, London. A new approach to the Holeyman A and Vanneste G (1987). execution and control of dynamic compaction.Compaction technology. Proceedings of conference, London, October 1987, pp 111–124. Thomas, Telford, London, 1988. Holtz W G (1948).The determination of limits for the control of placement moisture in high rolled earth dams.Proceedings of American Society for Testing and Materials, vol 48, pp 1240–1248. Settlements of shallow foundations on sand. Holzlohner U (1984). Soils and Foundations, vol 24, no 4, pp 58–70. Hoskins W G (1981). The making of the English landscape. Book Club Associates, London. (First edition, 1955) Prediction of Houston S L, Houston W N and Spadola D J (1988). field collapse of soils due to wetting.ASCE Journal of Geotechnical Engineering, vol 114, no 1, pp 40–58. Hudson J P and Harrison J A (1997). Engineering rock mechanics – an introduction to the principles. Pergamon, Oxford. Hughes J M O and Withers N J (1974). Reinforcing of soft cohesive soils with stone columns. Ground Engineering, vol 7, no 3, pp 42–49. Humpheson C, Simpson B and Charles J A (1991). Investigation of hydraulically placed PFA as a foundation for buildings.Ground movements and structures (ed J D Geddes). Proceedings of 4th international conference, Cardiff, July 1991, pp 68–88. Pentech Press, London, 1992. Hutchinson J N (1981). Damage to slopes produced by seepage erosion in sands. Proceedings of International Seminar on Landslides and Mudflows, pp 250–268. UNESCO. The classification of chalk for embankment Ingoldby H C (1978). construction. Clay fills. Proceedings of conference, London, November 1978, pp 137–142. Institution of Civil Engineers, London, 1979. Monitoring of landfill gas. Institute of Wastes Management (1990). IWM, Northampton. Institution of Civil Engineers (1987a). Specification for ground treatment. Thomas Telford, London. Specification for ground Institution of Civil Engineers (1987b). treatment: notes for guidance. Thomas Telford, London. Institution of Civil Engineers (1996a). Whither civil engineering? ICE, London.

pp 685–692. Balkema, Rotterdam. Hawkins A B and Pinches G M (1987a). Cause and significance of heave at Llandough Hospital, Cardiff, – a case history of ground floor heave due to gypsum growth.Quarterly Journal of Engineering Geology, vol 20, pp 41–57. Hawkins A B and Pinches G M (1987b). Sulphate analysis on black mudstones. Géotechnique, vol 37, no 2, June, pp 191–196. Understanding sulphate Hawkins A B and Pinches G M (1997). generated heave resulting from pyrite degradation.Ground chemistry implications for construction (ed A B Hawkins), pp 51–75. Balkema, Rotterdam. Hazen A (1892).Some physical properties of sands and gravels with special reference to their use in filtration.24th Annual Report Massachusetts State Board of Health.

Institution of Civil Engineers (1996b). Sustainability and acceptability in infrastructure development. ICE, London. Institution of Civil Engineers (1998). An application note to ‘An engineering guide to seismic risk to dams in the United Kingdom’. ICE, London. Institution of Civil Engineers (1999). Specification for cementbentonite cut-off walls. Thomas Telford, London. Institution of Civil Engineers Site Investigation Steering Group (1993).Site investigation in construction. Thomas Telford, London. Irving J (1985).Earthquake hazard in Britain.Earthquake engineering in Britain, pp 261–277. Thomas Telford, London. An instrumented demonstration embankment at Isaac A S (1972). Cortonwood colliery – final report.Mineral Products Section Report MP 13. Stoke Orchard, National Coal Board Coa l Research Establishment.

References

187

An instrumented Isaac A S and Troughton S J (1971). demonstration embankment at Cortonwood colliery – interim report. Mineral Products Section Report MP 6. Stoke Orchard, National Coal Board Coal Research Establishment. Pore water pressure response and liquefaction of Ishihara K (1977). sand deposits during earthquakes. Proceedings of Conference on Dynamic Methods in Soil and Rock Mechanics, Karlsruhe, vol 2, pp 161–168. Balkema, Rotterdam. Consolidation of saturated soils by explosions. Ivanov P L (1980). Proceedings of International Conference on Compaction, Paris, vol 1, pp 331–337. . Influence of groundwater level oscillation on Jaky J (1948) subsidence of structures. Proceedings of 2nd International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol 2, pp 148–150. The additional settlement of Jennings J E and Knight K (1957). foundations due to a collapse of structure of sandy subsoils on wetting. Proceedings of 4th International Conference on Soil Mechanics and Foundation Engineering, London, vol 1, pp 316–319. Protective measures for housing on gasJohnson R (2001). contaminated land. BRE Report BR414. CRC, Garston. Johnson D, Nicholls R and Thomson G H (1983). An evaluation of ground improvement at Belawan Port, North Sumatra.Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 1, pp 45–52. Juran I and Guermazi A (1988). Settlement response of soils reinforced by compacted sand columns.ASCE Journal of Geotechnical Engineering, vol 114, no 8, August, pp 930–943. Kilkenny W M (1968). A study of the settlement of restored opencast coal sites and their suitability for building development.Bulletin No 38, Department of Civil Engineering, University of Newcastle upon Tyne. Kilkenny W M (1988). A low-rise hospital development on restored opencast fill. Case histories in geotechnical engineering. Proceedings of 2nd international conference, University of Missouri-Rolla, vol 1, pp 19–24. Kim S-I and Kim D-S (1997). SASW method for the evaluation of ground densification by dynamic compaction.Ground improvement geosystems: densification and reinforcement (eds M C R Davies and F Schlosser). Proceedings of 3rd international conference, London, pp 141–147. Thomas Telford, London. King R A (1977).A review of soil corrosiveness with particular reference to reinforced earth. Transport and Road Research Laboratory Supplementary Report No 316, Crowthorne. King R A (1989).Corrosion in reinforced earth.Proceedings of Symposium on Reinforced Earth and Other Composite Soil Techniques, Heriot Watt University. TRRL Supplementary Report 457. Kirkpatrick W M and Webber I P (1987). Compaction control and testing of a colliery spoil for landfill sites.Reclamation, treatment and utilization of coal mining wastes (ed A K M Rainbow). Proceedings of conference, pp 133–144. Elsevier, Amsterdam. Klohn E J, Garga V K and Shukin W (1981). Densification of sand tailings by blasting. Proceedings of 10th International Conference on

Initiation of soil liquefaction under Kramer S L and Seed H B (1988). static loading conditions.ASCE Journal of Geotechnical Engineering, vol 114, no 4, pp 412–430. Lambe T W (1969).Reclaimed land in Kawasaki City, Japan.ASCE Journal of Soil Mechanics and Foundations Division, vol 95, no SM5, pp 1181–1198. Lampert D and Woodley D (eds) R (1991).Site selection and investigation: a practical handbook. Gower, Aldershot. Geotechnics of waste fill. Landva A O and Clark J I (1989). Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 86–103. STP 1070, ASTM, Philadelphia, 1990. Lange S (1986).Building on uncompacted dumps in the Rhenish brown coal area of the F ederal Republic of Germany.Building on marginal and derelict land. Proceedings of conference, May 1986, pp 137–153. Thomas Telford, London, 1987. Lawton E C, Fragaszy R J and Hardcastle J H (1989). Collapse of a compacted clayey sand.ASCE Journal of Geotechnical Engineering, vol 115, no 9, pp 1252–1267. Stress ratio Lawton E C, Fragaszy R J and Hardcastle J H (1991). effects on collapse of compacted clayey sand.ASCE Journal of Geotechnical Engineering, vol 117, no 5, pp 714–730. Building on derelict land. CIRIA, Leach B A and Goodger H K (1991). London. Lee K M, Shen C K, Leung D H K and Mitchell J K (1999). Effects of placement method on geotechnical behaviour of hydraulic fill sands. ASCE Journal of Geotechnical and Geoenvironmental Engineering, vol 125, no 10, October, pp 832–846. Leigh, W J P and Rainbow, K R (1979). Observations of the settlement of restored backfill of opencast mine sites.Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, pp E99–E128. Midland Geotechnical So ciety. Leonards G A, Cutter W A and Holtz R D (1980). Dynamic compaction of granular soils.ASCE Journal of Geotechnical Engineering Division, vol 106, no GT 1, pp 35–44. Leonards G A and Bailey B (1982). Pulverised coal ash as structural fill. ASCE Journal of the Geotechnical Engineering Division, vol 108, no GT4, pp 517–531. Leonards G A and Frost J D (1988). Settlement of shallow foundations on granular soils.ASCE Journal of Geotechnical Engineering, vol 114, no 7, pp 791–809. The general and congruent Leroueil S and Vaughan P R (1990). effects of structure in natural soils and weak rocks.Géotechnique, vol 40, no 3, September, pp 467–488. Ling H I, Leshchinsky D, Mohri Y and Kawabata T (1998). Estimation of municipal solid waste landfill settlement.ASCE Journal of Geotechnical and Geoenvironmental Engineering, vol 124, no 1, January, pp 21–28. Western Dock and Lord J A, Mudd I G and Williams G J (1986). Hermitage Basin reclamation.Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 521–537. Thomas Telford, London, 1987.

Soil Mechanics and Foundation Engineering, Stockholm, 1981, vol 3, pp 725–730. Knipe C (1979).Comparison of settlement rates on backfilled opencast mining sites. Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, pp E81–E98. Midland Geotechnical Society. Knochenmus G, Wojnarowicz M and Van Impe W F (1998). Stability of municipal solid wastes.Environmental geotechnics (ed P S Seco e Pinto). Proceedings of 3rd international congress, Lisbon, vol 3, pp 977–1000. Balkema, Rotterdam. Improvement of mine spoils Koutsoftas D C and Kiefer M L (1989). in Southern Illinois. Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 153–167. STP 1070, ASTM, Philadelphia, 1990.

Lukas R G (1980). Densification of loose deposits by pounding.ASCE Journal of Geotechnical Engineering, vol 106, no GT4, pp 435–446 Lunne T, Robertson P K and Powell J J M (1997). Cone penetration testing in geotechnical practice. Blackie, London. Macdonald A and Reid J M (1990). Embankment dam behaviour: the contribution of geochemistry.The embankment dam. Proceedings of 6th Conference of British Dam Society, Nottingham, September 1990, pp 185–191. Thomas Telford, London, 1991. Maddison J B, Jones D B, Bell A L and Jenner C G (1996). Design and performance of an embankment supported using low strength geogrids and vibro concrete columns.Geosynthetics: applications, design and construction (eds M B De Groot, G Den Hoedt and R J Termaat).Proceedings of 1st European Geosnthetics Conference, Maastricht, pp 325–332. Balkema, Rotterdam.

188

References

Pressuremeter testing: methods Mair R J and Wood D M (1987). Nixon I K (1982).Standard penetration test – state of the art report. and interpretation. CIRIA/Butterworths, London. Proceedings of 2nd European Symposium on Penetration Testing, Manchester Evening News (1979).Rise and fall of a real Reggie Amsterdam, vol 1, pp 3–24. Perrin. 22 December 1979. Nixon P J (1976).The use of materials from demolition in Further development of construction. Resources Policy, December, pp 276–283. Marsland A and Quarterman R (1974). multipoint magnetic extensometers for use in highly compressible Nixon P J (1978). Floor heave in buildings due to the use of pyritic ground. Géotechnique, vol 24, no 3, pp 429–433. shales as fill material. Chemistry and Industry, 4 March 1978, Massarsch K R and Broms B B (1983). Soil compaction by pp 160–164. vibrowing method. Proceedings of 8th European Conference on Soil Reclamation of Nixon P J, Treadaway K and Harrison W H (1979). Mechanics and Foundation Engineering, Helsinki, vol 2, pp 275–278. contaminated land.Proceedings of conference, Eastbourne, October Masters J (2000).Keller plays to the gallery. Construction News, no 1979, pp E4/1–10. Society of Chemical Industry, London, 1980. 6684, 21 September, pp 28–29. Deformation of fill Noorany I, Sweet J A and Smith I M (1992). Matthews M C, Hope V S and Clayton C R I (1996). The use of slopes caused by wetting. ASCE Geotechnical Special Publication no surface waves in the determination of ground stiffness profiles. 31. Stability and performance of slopes and embankments, vol 2, Geotechnical Engineering, Proceedings of Institution of Civil Engineers, pp 1244–1257. vol 119, no 2, April, pp 84–95. Groundwater problems in surface coal mining in Norton P J (1982). Mayne P W, Jones J S and Dumas J C (1984). Ground response to Scotland. International Journal of Mine Water, vol 1, pp 17–24. dynamic compaction.ASCE Journal of Geotechnical Engineering, vol Norton, P J (1983). Groundwater rebound effects on surface coal 110, no 6, June, pp 757–774. mining in the United Kingdom. Proceedings of 2nd International Symposium on Surface Mining and Quarrying, Bristol, pp 255–263. McEntee J (1991).Site investigation. Recycling derelict land (ed G Fleming), pp 64–87. Thomas Telford, London. Institution of Mining a nd Metallurgy. McGown A, Yeo K C and Andrawes (eds) K Z (1990).Proceedings Nwabuokei S O and Lovell C W (1985). Compressibility and of International Conference on Performance of Reinforced Soil settlement of compacted fills.Consolidation of soils: testing and Structures, Glasgow, September 1990. Thomas Telford, London, evaluation. pp 184–202. STP 892, ASTM, Philadelphia. 1991. Oner M (1990).Vertical and horizontal deformation of an Meigh A C (1987).Cone penetration testing: methods and inhomogeneous elastic half-space.International Journal for Numerical interpretation. CIRIA/Butterworths, London. and Analytical Methods in Geomechanics, vol 14, no 9, pp 613–629. Merz R C and Stone R (1962). Landfill settlement rates.Public Orr W R and Finch M O (1989). Solid waste landfill performance Works, vol 93, no 9, pp 103–106, 210, 212. during the Loma Prieta earthquake.Geotechnics of waste fills – theory Meyerhof G G (1951). Building on fill with special reference to the and practice. Proceedings of symposium, Pittsburg, September 1989, pp 22–30. STP 1070, ASTM, Philadelphia, 1990. settlement of a large factory.Structural Engineer, vol 29, no 2, February, pp 46–57. Discussion, vol 29, no 11, November, Ortigosa P, Musante H and Kor t I (1981). Mechanical properties of pp 297–305. the gravel of Santiago.Proceedings of 10th International Conference Ministry of Housing and Local Government (1956). Derelict land on Soil Mechanics and Foundation Engineering, Stockholm, vol 2, and its reclamation.Technical Memorandum No 7. pp 545–548. Mitchell J K (1981). Soil improvement: state of the art report. Oteo C and Sopena L (1991). The use of jet grouting to improve Proceedings of 10th International Conference on Soil Mechanics and loose fills. Proceedings of 10th European Conference on Soil Foundation Engineering, Stockholm, vol 4, pp 509–565. Mechanics and Foundation Engineering, Florence, vol 2, pp 903–906. Mitchell J K, Seed R B and Seed H B (1989). Stability Ove Arup and Partners (1991). Preliminary study of UK seismic considerations in the design and construction of lined waste hazard and risk. Report prepared for the Department of the repositories. Geotechnics of waste fills – theory and practice. Environment. Proceedings of symposium, Pittsburg, September 1989, pp 207–224. Palmer K N (1979). Fire and explosion hazards with contaminated STP 1070, ASTM, Philadelphia, 1990. land. Reclamation of contaminated land. Proceedings of conference, Monahan E J (1994). Construction of fills. 2nd edition. Wiley, New Eastbourne, October 1979, pp D3/1–7. Society of Chemical Industry, York. London, 1980. Morgan C S, Holden J M W, O’Brien P S and Gardner J W (1993). Parkin A K (1978). Coefficient of consolidation by the velocity Engineered fill for Sheffield and Rotherham City Airpor t.Engineered method. Géotechnique, vol 28, no 4, pp 472–474. fills. Proceedings of conference, Newcastle-upon- Tyne, pp 441–454. Parliamentary Office of Science and Technology (1998). A Thomas Telford, London. brown and pleasant land – household growth and brownfield sites. Morgan J G D and Markland E (1981). The effect of vibration on Report 117. ballast beds. Géotechnique, vol 31, no 3, pp 367–386. Parsons A W (1978). Moisture condition test for assessing the Moseley M P(ed) (1993).Ground improvement. Blackie, Glasgow. engineering behaviour of earthwork material.Clay fills. Proceedings of

Moseley M P and Priebe H J (1993). Vibro techniques. Ground improvement (ed M P Moseley), pp 1–19. Blackie, London. Munfakh G A, Sarkar S K and Castelli R J (1983). Performance of a test embankment founded on stone columns.Piling and ground treatment. Proceedings of international conference, London, March 1983, pp 259–265. Thomas Telford, London, 1984. Nagase H and Ishihara K (1988). Liquefaction induced compaction and settlement of sand during earthquakes.Soils and Foundations, vol 28, no 1, pp 65–76. New Civil Engineer (1979).NHBC seeks remedy for ‘blighted’ estate. 25 January, p 13. New Civil Engineer (1985).Sinking suburb. 31 October, p 9. New Civil Engineer (1990).Tortuous tort. 2 August, p12.

conference, London, November 1978, pp 169–175. Institution of Civil Engineers, London, 1979. Parsons A W (1992). Compaction of soils and granular materials: a review of research performed at the Transport Research Laboratory. HMSO, London. Parsons A W and Darley P (1982). The effect of soil conditions on the operation of earthworks plant. Laboratory Report 1034, Transport and Road Research Laboratory, Crowthorne. Patel D (1995).As reported by K S Watts in report on joint meeting of British Geotechnical Society and ICE Ground Board on Building on fill, London, April 1995. Ground Engineering, vol 28, no 6, July/August, pp 32–35. Paul V (1994).The performance of building materials in contaminated land. BRE Report BR255. CRC, Garston.

References

189

Pearce R W (1974).Dynamic consolidation.Consulting Engineer, vol 38, April, 43–47. Penman A D M (1986). Tailings dams and lagoons.Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 95–109. Thomas Telford, London, 1987. Penman A D M and Charles J A (1973). Constructional deformations in rockfill dam.ASCE Journal of Soil Mechanics and Foundations Division, vol 99, SM2, pp 139–163. The quality and suitability of Penman A D M and Charles J A (1976). rockfill used in dam conmstruction.Transactions of 12th International Congress on Large Dams, Mexico, vol 1, pp 533–556. A comparison between Penman A D M and Charles J A (1985a). observed and predicted deformations of an embankment dam with a central asphaltic core. Transactions of 15th International Congress on Large Dams, Lausanne, vol 1, pp 1373–1389. Behaviour of rockfill dam Penman A D M and Charles J A (1985b). with asphaltic membrane.Proceedings of 11th International Conference on Soil Mechanics and Foundation Engineering, San Francisco, vol 4, pp 2011–2014. The safety of tailings dams Penman A D M and Charles J A (1990). and lagoons in Britain.The embankment dam. Proceedings of 6th conference of British Dam Society, Nottingham, September 1990, pp 53–58. Thomas Telford, London, 1991. Penman A D M and Godwin E W (1974). Settlement of experimental houses on land left by opencast mining at Corby.Settlement of structures. Proceedings of conference, Cambridge, April 1974, pp 53–61. Pentech Press, London, 1975. Penman A D M and Watson G H (1967). Foundations for storage tanks on reclaimed land at Teesmouth.Proceedings of Institution on Civil Engineers, vol 37, May, 19–42. Observed Penman A D M, Burland J B and Charles J A (1971). and predicted deformations in a large embankment dam during construction. Proceedings of Institution of Civil Engineers, vol 49, May, pp 1–21. Perelberg S, Boyd P J H, Montague K N and Greenwood J R (1986).M25 Bell Lane Pit: ground improvement by dynamic compaction.Building on marginal and derelict land. Proceedings of conference, Glasgow, pp 267–280. Thomas Telford, London. Perry J (1989).A survey of slope condition on motorway earthworks in England and Wales. Research Report 199, Transport and Road Research Laboratory, Crowthorne. Pilyugin A I (1967). Settlement of loess foundations under canal embankments. Proceedings of 3rd Asian Conference on Soil Mechanics and Foundation Engineering, pp 29–32. Maximum allowable non-uniform Polshin D E and Tokar R A (1957). settlement of structures.Proceedings of 4th International Conference on Soil Mechanics and Foundation Engineering, vol 1, pp 402–405. Poulos H G and Davis E H (1974). Elastic solutions for soil and rock mechanics. Wiley, New York. Preene M, Roberts T O L, Powrie W and Dyer M R (1997). Groundwater control: design and practices.Core Programme Funders Report FR/CP/50, CIRIA, London.

Chemical and biochemical weathering Pye K and Miller J A (1990). of pyritic mudrocks in a shale embankment.Quarterly Journal of Engineering Geology, vol 25, pp 365–381. Pyke R M, Knuppel L A, and Lee K L (1978). Liquefaction potential of hydraulic fills. ASCE Journal of Geotechnical Engineering Division, vol 104, GT 11, November, pp 1335–1354 . Raju V R (1997).The behaviour of very soft cohesive soils improved by vibro-replacement. Ground improvement geosystems: densification and reinforcement (eds M C R Davies and F Schlosser). Proceedings of 3rd international conference, London, pp 253–259. Thomas Telford, London. Evaluation of densification Ramaswamy S D and Yong K Y (1983). of sandfill. Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 1, pp 69–72. Rao S K, Moulton L K and Seals R K (1977). Settlement of refuse landfills. Geotechnical practice for disposal of solid waste materials. Proceedings of ASCE conference, University of Michigan, pp 574–598. Rathje E M and Bray J D (1999). Two dimensional seismic response of solid-waste landfills.Earthquake geotechnical engineering (ed P S Seco e Pinto). Proceedings of 2nd international conference, Lisbon, vol 2, pp 655–660. Balkema, Rotterdam. Raybould J G, Rowan S P and Barry D L (1995). Methane investigation strategies. Report 150, CIRIA, London. Raynsford N (1998). Quoted by S Aker in Talking point, Ground Engineering, vol 31, no 11, p3. Reed S M and Singh R N (1986). Groundwater recovery problems associated with opencast mine backfills in the United Kingdom. International Journal of Mine Water, vol 5, no 3, pp 47–74. Reed S M and Hughes D B (1990). Long term settlement of opencast mine backfills – case studies from the north east of England. Reclamation, treatment and utilization of coal mining wastes (edA K M Rainbow). Proceedings of conference, pp 313–322. Balkema, Rotterdam. Reid W M and Buchanan N W (1986). The Scottish Exhibition Centre. Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 435–448. Thomas Telford London, 1987. Rethati L (1961).Behaviour of buildings founded on embankments. Proceedings of 5th International Conference on Soil Mechanics and Foundation Engineering, Paris, vol 1, pp 781–784. Robertson P K and Campanella R G (1985) . Liquefaction potential of sands using the CPT. ASCE Journal of Geotechnical Engineering, vol 111, no 3, pp 384 –403. Rogers J D (1992). Long term behavior of urban fill embankments. Stability and performance of slopes and embankments II, vol 2, pp 1258–1273. ASCE Geotechnical Special Publication no 31. Rowe P W (1991).A reassessment of the causes of the Carsington embankment failure. Géotechnique, vol 41, no 3, September, pp 395–421. Rutledge P C (1970). Utilization of marginal lands for urban development. 5th Terzaghi Lecture.ASCE Journal of Soil Mechanics and Foundations Division, vol 96, no SM1, January, pp 1–48.

Priebe H J (1990). Vibro replacement: design criteria and quality control. Deep foundation improvements: design, construction and testing. Proceedings of symposium, Las Vegas, January 1990, pp 62–72. STP 1089, ASTM, Philadelphia, 1991. Priebe H J (1995).The design of vibro-replacement.Ground Engineering, vol 28, December, pp 31–37. Proctor R R (1933). Fundamental principles of soil compaction. Engineering News Record, vol 111, no 9, 31 August, pp 245–248. Proctor R R (1948). The relationship between the foot pounds per cubic foot of compactive effort and the shear strength of compacted soils. Proceedings of 2nd International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol 5, pp 219–223.

Salem A M and Krizek R J (1976). Stress-deformation-time behaviour of dredgings. ASCE Journal of Geotechnical Engineering Division, vol 102, GT2, February, pp 139–157. . Foundation problems of waste Sargunan A and Rajamani K (1986) disposal fills in Madras. Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 315–322. Thomas Telford, London, 1987. On the stiffness of sands. PreSchanz T and Vermeer P A (1998). failure deformation behaviour of geomaterials (eds R J Jardine, M C R Davies, D W Hight, A K C Smith and S E Stallebrass). Géotechnique Symposium in Print, pp 383–387. Thomas Telford, London.

190

References

The mechanical ageing of soils. 25th Schmertmann J H (1991). Terzaghi Lecture.ASCE Journal of Geotechnical Engineering, vol 117, no 9, September, pp 1288–1330. Schneider J A, Mayne P W, Hendren T L and Wise CM (1999). Initial development of an impulse piezovibrocone for liquefaction evaluation. Physics and mechanics of soil liquefaction (eds P V Lade and J A Yamamuro), pp 341–354. Balkema, Rotterdam. Factors affecting the Schultze E and Moussa A (1961). compressibility of sand. Proceedings of 5th International Conference on Soil Mechanics and Foundation Engineering, Paris, vol 1, pp 335–340. A study of Schulz T M, Sonnenberg R G and Thomson S (1986). urban restoration of surface mined land in Western Canada.Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 125–136. Thomas Telford, London, 1987. Principles of soil mechanics. Addison-Wesley, Scott R F (1965). Reading, Massachusetts. Seed H B (1987).Design problems in soil liquefaction.ASCE Journal of Geotechnical Engineering, vol 113, no 8, pp 827–845. Use of SPT and CPT tests for Seed H B and de Alba P (1986). evaluating the liquefaction resistance of sands.Use of in-situ tests in geotechnical engineering, pp 281–302. ASCE Geotechnical Special Publication no 6. Seed H B and Silver M L (1972) . Settlement of dry sands during earthquakes. ASCE Journal of Soil Mechanics and Foundations Division, vol 98, no SM4, pp 381–397. Evaluation of liquefaction Seed H B, Idriss I M and Arango I (1983). potential using field performance data.ASCE Journal of Geotechnical Engineering, vol 109, no 3, pp 458–482. Shang J Q, Tang M and Miao Z (1998). Vacuum preloading consolidation of reclaimed land: a case study.Canadian Geotechnical Journal, vol 35, no 5, October, pp 740–749. Sharma H D, Dukes M T and Olsen D M (1989). Field measurements of dynamic moduli and Poisson’s ratio o f refuse and underlying soils at a landfill site. Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 57–70. STP 1070, ASTM, Philadelphia, 1990. Sheffield Morning Telegraph (1983).Landslip homes forcing out the old. 7 Ja nuary. Sherard J L and Dunnigan L P (1985). Filters and leakage control in embankment dams.Seepage and leakage from dams and impoundments. Proceedings of ASCE symposium, Denver, pp 1–30. Sherard J L, Dunnigan J P and Talbot J R (1984a). Basic properties of sand and gravel filters.ASCE Journal of Geotechnical Engineering, vol 110, no 6, June, pp 684–700. Sherard J L, Dunnigan L P and Talbot J R (1984b). Filters for silts and clays. ASCE Journal of Geotechnical Engineering, vol 110, no 6, June, pp 701–718. Reproducibility of the results of soil Sherwood P T (1970). classification and compaction tests. Report LR 339. Road Research Laboratory, Crowthorne. The use of waste and low-grade materials in Sherwood P T (1975a).

Building on colliery Skinner H D, Watts K S and Charles J A (1997). spoil: some geotechnical considerations.Ground Engineering, vol 30, no 5, June, pp 35–40. Skinner H D, Crilly M S and Charles J A (1998). Numerical models for the design of shallow foundations for low-rise buildings.Application of numerical methods to geotechnical problems (ed A Cividini). Proceedings of 4th European Conference, Udine, Italy, October, pp 655–664. Springer, Wien, New York. Skopek J (1979).Geotechnical problems encountered in moving the church at Most. Canadian Geotechnical Journal, vol 16, no 3, pp 473–480. Influence of placement method on Sladen J A and Hewitt K J (1989). the in-situ density of hydraulic sandfills. Canadian Geotechnical Journal, vol 26, no 3, August, pp 453–466. Sleeman W (1987a). The economics of minestone utilization. Reclamation, treatment and utilization of coal mining waste. (ed A K M Rainbow). Proceedings of conference, pp 1–19. Elsevier, Amsterdam. Sleeman W (1987b). Colliery spoil in urban development. Reclamation, treatment and utilization of coal mining wastes (ed A K M Rainbow). Proceedings of conference, pp 163–177. Elsevier, Amsterdam. Sleep K (1996).Coal deal will yield £1 billion in work. Construction News, 14 November, p 6. Slocombe B C (1993). Dynamic compaction.Ground improvement (ed M P Moseley), pp 20–39. Blackie, London. Slocombe B C and Moseley M P (1986). Experience with dynamic compaction on derelict sites.Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 799–806. Thomas Telford, London, 1987. The densification of Slocombe B C, Bell A L and Baez J I (2000). granular soils using vibro methods. Géotechnique, vol 50, no 6, pp 715–725. Smith L C (1979a). Hazard to immediate workers on site. Reclamation of contaminated land. Proceedings of conference, Eastbourne, October 1979, paper D1, 1–11. Society of Chemical Industry, London, 1980. Smith M A (1979b). Redevelopment of contaminated land: landfill sites. Engineering behaviour of industrial and urban fills. Proceedings of symposium, Birmingham, pp B49–B69. Midland Geotechnical Society, Birmingham. Appraisal of contaminated land.Proceedings of Smith M A (1980). Annual Congress of Environmental Health Officers’ Association, Bournemouth, pp 53–56. EHOA, London. Smith M A (1991). Data analysis and interpretation.Recycling derelict land (ed G Fleming), pp 88–144. Thomas Telford, London. Smith M A and Beckett M J (1982). An international study of contaminated land.Proceedings of conference on Management of uncontrolled hazardous waste sites, Washington 1982. Hazardous Materials Control and Research Institute, Silver Spring, Maryland 1982. Smyth-Osbourne K R and Mizon D H (1984). Settlement of a factory on opencast backfill.Ground movements and structures (ed J D Geddes). Proceedings of 3rd international conference, Cardiff, July

road construction: 2 colliery shale. Laboratory Report 649, Transport and Road Research Laboratory, Crowthorne. Sherwood P T (1975b). The use of waste and low-grade materials in road construction: 3 pulverised fuel ash. Laboratory Report 686, Transport and Road Research Labo ratory, Crowthorne. Siegel R A, Robertson R J and Anderson D G (1989). Slope stability investigations at a landfill in southern California.Geotechnics of waste fills – theory and practice. Proceedings of symposium, Pittsburg, September 1989, pp 259–284. STP 1070, ASTM, Philadelphia, 1990. Skinner H D and Charles J A (1999). Problems associated with building on a variable depth of fill. Ground Engineering, vol 32, no 7, July, pp 32–35.

1984, pp 463–479. Pentech Press, Plymouth, 1985. Sowers G F (1968). Foundation problems in sanitary landfills.ASCE Journal of Sanitary Engineering Division, vol 94, no SA1, pp 103–116. Settlement of waste disposal fills.Proceedings of Sowers G F (1973). 8th International Conference on Soil Mechanics and Foundation Engineering, Moscow, vol 2.2, pp 207–210. Sowers G F, Williams R C and Wallace T S (1965). Compressibility of broken rock and the settlement o f rockfills.Proceedings of 6th International Conference on Soil Mechanics and Foundation Engineering, vol 2, pp 561–565. Soyez B, Magnan J P and Delfaut A (1983). Loading tests on a clayey hydraulic fill stabilised by lime treated soil columns.Proceedings of 8th European Conference on Soil Mechanics and Foundation Engineering, Helsinki, vol 2, pp 951–954.

References

191

The specific Stamatopoulos A C and Kotzias P C (1973). constrained modulus. Proceedings of 8th International Conference on Soil Mechanics and Foundation Engineering, Moscow, vol 1.2, paper 1/61, pp 397–402. Soil compressibility as Stamatopoulos A C and Kotzias P C (1978). measured in the oedometer.Géotechnique, vol 28, no 4, pp 363–375. (Discussion 1980, vol 30 ,no 3, pp 325–328) Guidelines on Standards Association of Australia (1990). earthworks for commercial and residential developments. AS37981990. Sydney. Stearns R P and Petoyan G S (1984). Utilization of landfills as building sites. Waste Management and Research, vol 2, 75–83. Steensen-Bach J O (1989). Strength evaluation of a natural sand. Proceedings of 12th International Conference on Soil Mechanics and Foundation Engineering, Rio de Janeiro, vol 1, pp757–762. Water pollution control centre, Tubli, Bahrain. Steger E H (1986). Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 69–81. Thomas Telford, London, 1987. St John H D, Hunt R J and Charles J A (1989). The use of ‘vibro’ ground improvement techniques in the United Kingdom.BRE Information Paper IP5/89. CRC, Garston. Strohm W E (1978). Design and construction of compacted shale embankments: volume 4 field and laboratory investigations, phase 3. Report no FHWA-RD-78-140. US Army Engineers Experimental Station, Vicksburg. Stroud M A (1988). The standard penetration test – its application and interpretation.Proceedings of Conference on Penetration Testing in the UK, Birmingham, July 1988, pp 29–49. Thomas Telford, London, 1989. Collapse settlement of old Stroud M A and Mitchell J M (1989). chalk fill at Brighton. Chalk. Proceedings of international symposium, Brighton, September 1989, pp 343–350. Thomas Telford, London, 1990. Sully J P (1986).Stability of tank foundations on a cohesive hydraulic fill. Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 841–852. Thomas Telford, London, 1987. Sutherland H B (1974). Review paper on granular materials. Settlement of structures. Proceedings of conference, Cambridge, April 1974, pp 473–499. Pentech Press, London, 1975. Swain A (1979).Field studies of PFA in par tially submerged conditions. Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, April, pp D49–D61. Midland Geotechnical Society. Swain A and Holt D N (1986). Dynamic compaction of a refuse site for a road interchange.Building on marginal and derelict land. Proceedings of conference, Glasgow, May 1986, pp 339–357. Thomas Telford, London, 1987. Tan S B (1977). Settlement of structures on uncompacted rubbish. Proceedings of 9th International Conference on Soil Mechanics and Foundation Engineering, Tokyo, vol 1, pp 395–397. Taylor R K (1984). Composition and engineering properties of British colliery discards. London, National Coal Board Mining Department.

Geological maps and sections for civil Thomas P R (1991). engineers. Blackie, Glasgow. Thompson R P (1998). The value of timely hazard identification.The value of geotechnics in construction. Proceedings of seminar, London, pp 3–11. CRC, Garston. Thompson J and Holden J M W (1990). Compaction of opencast backfill beneath highways and associated developments.Reclamation, treatment and utilisation of coal mining wastes. Proceedings of conference, Rotterdam, pp 293–312. Thomson G H and Aldridge J A (1983). London docklands – problems associated with their development with special reference to the Surrey Docks scheme. Reclamation 83. Proceedings of International Land Reclamation Conference, Grays, Essex, pp 124–132. Industrial Seminars, Tunbridge Wells. Thomson G H and Herbert A (1978). Compaction of clay fills in situ by dynamic compaction.Clay fills. Proceedings of conference, London, November 1978, pp 197–204. Institution of Civil Engineers, London, 1979. Colliery spoil tips – after Thomson G McK and Rodin S (1972). Aberfan. Proceedings of Institution of Civil Engineers, Paper 7522, pp 1–59. Thomson S and Schulz T M (1984). Settlement studies of an open pit mine backfill in Western Canada. Ground movements and structures (ed J D Geddes). Proceedings of 3rd international conference, Cardiff, pp 480–495. Pentech Press, Plymouth, 1985. Thorburn S (1975). Building structures supported by stabilised ground. Géotechnique, vol 25, no 1, pp 83–94 Thorburn S and Buchanan N W (1983). Redevelopment of urban areas in Glasgow and Edinburgh. Reclamation 83. Proceedings of International Land Reclamation Conference, Grays, Essex, pp 133–138. Industrial Seminars, Tunbridge Wells. Threadgold L (1978). Contribution to discussion. Clay fills. Proceedings of conference, London, November 1978, pp 251–253. Institution of Civil Engineers, London, 1979. Tokimatsu K and Seed H B (1987). Evaluation of settlements in sand due to earthquake. ASCE Journal of Geotechnical Engineering, vol 113, no 8, pp 861–878. A framework for unsaturated soil behaviour. Toll D G (1990). Géotechnique, vol 40, no 1, pp 31–44. Tomlinson M J (1995). Foundation design and construction. 6th edition (first published 1963). Longman Scientific a nd Technical, Harlow, Essex. Tomlinson M J and Wilson D M (1973). Pre-loading of foundations by surcharge on filled ground. Géotechnique, vol 23, no 1, pp 117–120. Tomlinson, M J, Driscoll, R and Burland, J B (1978). Foundations for low-rise buildings. Structural Engineer, Part A, vol 56A , no 6, pp 161–173. In-situ soil mixing. Ground improvement (ed M P Toth P S (1993). Moseley), pp 193–204. Blackie, London. Trenter N A (1993). Compaction, monitoring and performance of engineered fill at the Dixon opencast site, Chesterfield UK.Engineered

Terzaghi K (1936). Relation between soil mechanics and foundation engineering. Proceedings of 1st International Conference on Soil Mechanics and Foundation Engineering, Harvard, vol 3, pp 13–18. Terzaghi K (1939)Soil . mechanics – a new chapter in engineering science (45th James Forrest Lecture). Journal of Institution of Civil Engineers, vol 12, pp 106–141. Terzaghi K (1951)The . influence of modern soil studies on the design and construction of foundations.Proceedings of Building Research Congress held in London, 1951, division 1, part 3, pp 139–145. Terzaghi K (1960). Discussion on Salt Springs and Lower Bear River Dams. ASCE Transactions, vol 125, part 2, pp 139–159. Properties of iron and steel slags. Reclamation Thomas G H (1983). of former iron and steelworks sites.Proceedings of conference, Windermere, pp C1–C8. Durham County Council.

fills. Proceedings of conference held in Newcastle-upon-Tyne, pp 455–469. Thomas Telford, London. Trenter N A and Charles J A (1998). A model specification for engineered fills for building purposes. Proceedings of Institution of Civil Engineers, Geotechnical Engineering, vol 119, October, pp 219–230. Turner M D (1979). Lithology, behaviour and properties of coarse unburnt colliery spoils in relation to their use as fills.Engineering behaviour of industrial and urban fill. Proceedings of symposium, Birmingham, pp B35–B48. Midland Geotechnical Society Uff J F and Clayton C R I (1986). Recommendations for the procurement of ground investigation. Special Publication 45, CIRIA, London.

192

References

Dilation angle and An Vaid Y P, Byrne P M and Hughes J M O (1981). Watts K S, Johnson D, Wood L A and Saadi A (2000). liquefaction potential. Recent advances in geotechnical earthquake instrumented trial of vibro ground treatment supporting strip engineering and soil dynamics. Proceedings of international foundations in a variable fill.Géotechnique, vol 50, no 6, December, pp conference, University of Missouri-Rolla, vol 1, pp 161–165. 699–708. Foundation on high clayey spoil heaps.Proceedings Trench filled PFA in colliery waste supporting Vanicek I (1991). Weatherley N (1979). of 10th European Conference on Soil Mechanics and Foundation old peoples’ bungalows. Engineering behaviour of industrial and urban Engineering, Florence, vol 2, pp 629–632. fill. Proceedings of symposium, Birmingham, pp D71–D76. Midland Van Impe W E (1989). Soil improvement techniques and their Geotechnical Society. evolution. Balkema, Rotterdam. Dynamic deep compaction of sanitary landfill to Welsh J P (1983). Van Impe W E (1998). Environmental geotechnics: ITC 5 activities – support superhighway. Proceedings of 8th European Conference on State of the art.Environmental geotechnics (ed P S Seco e Pinto). Soil Mechanics and Foundation Engineering, Helsinki, vol 1, Proceedings of 3rd international congress, Lisbon, vol 4, pp pp 319–321. 1163–1187. West G (1991).The field description of engineering soils and rocks. Van Impe W and de Beer E (1983). Improvement of settlement Geological Society of London Professional Handbook. Open University behaviour of soft layers by means of stone columns.Proceedings of Press, Milton Keynes. 8th European Conference on Soil Mechanics and Foundation Dynamic consolidation as an West J M and Slocombe B C (1973). Engineering, Helsinki, vol 1, pp 309–312. alternative foundation.Ground Engineering, no 6, pp 52–54. Vaughan P R and Soares H F (1982). Design of filters for clay cores Whitman R V (1970). Hydraulic fills to support structural loads.ASCE of dams. ASCE Journal of Geotechnical Engineering, vol 108, no GT1, Journal of Soil Mechanics and Foundations Division, SM1, vol. 96, January, pp 17–31. January, 23–47. Vaughan P R, Hight D W, Sodha V G and Walbancke H J (1978).Wightman A and Beaton N F (1984). Improvement of a dumped Factors affecting the stability of clay fills in Britain.Clay fills. rockfill foundation by dynamic consolidation.Case histories in Proceedings of conference, London, November 1978, pp 205–217. geotechnical engineering. Proceedings of international conference, Institution of Civil Engineers, London, 1979. vol 3, pp 1365–1372. Waite N J and Knipe C V (1991). Land reclamation by opencasting – Wilde P M and Crook J M (1979). Problems and solutions in case study of ground movement within development area.Ground developing large areas of filled ground at Warrington New Town. movements and structures (ed J D Geddes). Proceedings of 4th Engineering behaviour of industrial and urban fill. Proceedings of international conference, Cardiff, July 1991, pp 42–53. Pentech Press, symposium, Birmingham, April, pp D39–D47. Midland Geotechnical Plymouth, 1992. Society. Derelict land. David and Charles, Newton The monitoring of ground Wallwork K L (1974). Wilde P M and Crook J M (1991). Abbot. movements and their effects on surface structures: a series of case Walker F C and Holtz W G (1951). Control of embankment material histories. Ground movements and structures (ed J D Geddes). by laboratory testing.ASCE Journal of Soil Mechanics and Foundations Proceedings of 4th international conference, Cardiff, July 1991, Division, vol 77, separate no 108. pp 182–192. Pentech Press, London, 1992. Watford and Rickmansworth Review (1991).£1m to shore up Williams G M and Aitkenhead N (1991). Lessons from Loscoe: the sinking homes. 28 November. uncontrolled migration of landfill gas.Quarterly Journal of Engineering Watts K S and Charles J A (1990). Settlement of recently placed Geology, vol 24, pp 191–207. domestic refuse landfill. Proceedings of Institution of Civil Engineers, Review of research on properties Wimpey Laboratories Ltd (1972). Part 1, vol 88, December, pp 971–993. of spoil tip materials. National Coal Board: Headquarters Research Watts K S and Charles J A (1991). The use, testing and Project. performance of vibrated stone columns in the United Kingdom.Deep Water tamping of fills in Texas expedites highway Wise A J (1929). foundation improvements: design, construction and testing, work. Engineering News Record, vol 103, 26 December, pp pp 212–223. STP 1089, ASTM, Philadelphia. 1016–1017. Watts K S and Charles J A (1993). Initial assessment of new rapid Wood L A, Johnson D, Watts K S and Saadi A (1996). Performance ground compactor.Engineered fills (eds B G Clarke, C J F P Jo nes and of strip footings on fill materials reinforced by stone columns. A I B Moff at. Proceedings of international conference, NewcastleStructural Engineer, vol 74, no 16, August, pp 265–271. upon-Tyne, pp 399–412. Thomas Telford, London. World Commission on Environment and Development Watts K S and Charles J A (1999). Settlement characteristics of (Bruntland Commission) (1987). Quoted in Institution of Civil landfill wastes. Proceedings of Institution of Civil Engineers, Engineers (1996b). Geotechnical Engineering, vol 137, October, pp 225–233. Wright D and Brown J (1979). Hydraulic disposal of pfa from CEGB Watts K S and Serridge C J (2000). A trial of vibro bottom-feed Midland region power stations. Proceedings of 6th International stone column treatment in soft clay soil.Grouting, soil improvement, Conference on Hydraulic Transport of Solids in Pipes, pp 379–387. geosystems including reinforcement (ed H Rathmayer). Proceedings of 4th international conference, Helsinki, pp 549–556. Building Information Ltd, Helsinki. Ground Watts K S, Charles J A and Butcher A P (1989). improvement for low-rise housing using vibro at a site in Manchester. Municipal Engineer, vol 6, no 3, pp 145–157. Watts K S, Saadi A, Woods L A and Johnson D (1992). Preliminary report on a field trial to assess the design and perfo rmance of vibro ground treatment with reinforced strip foundations.Polluted and marginal land. Proceedings of 2nd international conference, Brunel University, London, pp 215–221.

BHRA, Cranfield. Xanthakos P P, Abramson L W and Bruce D A (1994). Ground control and improvement. Wiley, New York. Sanitary landfill settlement rates. Yen B C and Scanlon B (1975). ASCE Journal of Geotechnical Engineering Division, vol 101, no GT5, pp 475–487. Yong R N and Thomas H (eds) R (1997).Geoenvironmental engineering – contaminated ground: fate of pollutants and remediation. Proceedings of British Geotechnical Society conference, Cardiff. Thomas Telford, London.