Pavement Engineering Notes 2012

U NIVERSITY NIVERSITY OF NOTTINGHAM Pavement Engineering - Module H23P01

DEPARTMENT

OF

CIVIL E NGINEERING NGINEERING Course Notes

PAVEMENT ENGINEERING

INTRODUCTION A pavement is a structure designed to allow trafficking, usually of wheeled vehicles. Most pavements are roads, but airfields, industrial hardstandings, cycle tracks etc. are all included. Key points: a) Pavements Pavements are high-vol high-volume ume constructi constructions; ons; the materi materials als used must must therefore therefore be cheap and environmentally acceptable . b) There is no exact definition of failure; they simply have to remain ‘serviceable ’. c) The definiti definition on of serviceabi serviceability lity will will vary from from applicati application on to application application.. d) Maintaining serviceability is an important part of pavement engineering.

The basic building blocks Soil: unpredictable; water susceptible; sometimes low strength Gra Granula ular Material: al: more ore predic dictabl able; less water susceptible; strong onger Hydraulically-Bound Material: bound with cement or something similar Asphalt: stones stuck together with bitumen; good quality material

A Typical Pavement Structure Surface course (or Wearing course) – Asphalt Asphalt Binder course (or Basecourse) – Asphalt Asphalt Base –

Asphalt , Hydraulically-bound Hydraulically-bound (e.g. (e.g. Pavement Quality Concrete), or Granular or Granular (often in more than one layer)

Sub-base – Hydraulically-bound Hydraulically-bound or or Granular Granular Capping (or Lower Sub-base) – Hydraulically Hydraulically-

bound or bound or Granular Granular (only (only used over poor subgrade; often in more than one layer) Subgrade (or Substrate) – Soil – Soil

1


DEPARTMENT


OF

CONSTRUCTION A competent pavement designer must understand the practicalities of material production and pavement construction if sensible decisions are to be taken. 1.

Unbound Material

Natural Soils • Check it to see that it is as strong as it was expected to be (CBR test – see later). • Protect it . It’s easy to turn a basically sound material into a muddy soup! So leave a thin layer of overlying material until the very last moment. Granular Materials • Make sure you have material that meets the specification. a) Part Particl iclee Size Size Distri Distribut butio ion n : achieved by crushing larger rocks and/or by blending materials from more than one source. Put a sample through a set of sieves to check it.

Sample

Shake 100 90 80

e.g. typical sub base limits:

g n i s s a p e g a t n e c r e P

Upper Limit Lower Limit

70 60 50 40 30 20 10 0 0. 01

0. 1

2

1 Sieve size (mm)

10

100


DEPARTMENT

OF


b) Particle So Soundness : Use a test such as the Los Angeles Abrasion test; sometimes also tests for frost damage and chemical weathering (MnSO4 soundness). c) Particle Shape : This isn’t always specified. Most common requirement = % crushed faces (trying to make sure that rounded gravel isn’t used); sometimes also by limits on flakiness (% of particles of a given size able to pass through a special thin sieve opening) and elongation (% of particles with one dimension over 1.8 times the nominal size) of particles.

The shape depends on the equipment used to carry out the crushing: Cone

Jaw

Roll

Impact

d) Wate Waterr Cont Conten entt: All unbound materials are sensitive to water. Dry Density

Desired in-service condition after drying out

0% air voids (i.e. full saturation) Maximum Dry Density (heavy compaction) Maximum Dry Density (light compaction)

Heavy Compaction (e.g. on site)

Light Compaction (e.g. in laboratory) Dry

Maximum suction

Optimum

Saturated

Water Content

Low water content: negative pore pressure (or suction) . Makes compaction difficult; BUT good once in the road.

3


DEPARTMENT

OF


Higher water content: positive pore pressure. Makes compaction easier; BUT bad once in the road. So: compact at Optimum Water Content (OMC); let it dry o ut to develop suction.

• Transport it to site in a suitable delivery truck • Place it and compact it properly

Delivery Truck

Delivery Truck

Dozer

Compactor

Motor grader

Paver

Compactor

Dozer + grader = good enough for lower layers Paver = for high quality base layers (consistent thickness and surface level) Compactors:

Vibratory

2.

Pneumatic

Static

Hydraulically-Bound Material (HBM)

HBM means any material that needs water to activate a binder – usually cement.

4


DEPARTMENT

OF


In-situ Stabilised Mixing unit

Stabiliser

Compactor

This is usually just soil improvement . Converts a soft soil into something you can build a road on. As well as cement, lime and/or fly ash (also called pulverised fuel ash – PFA) are used. Plant-mixed HBM base/sub-base • Get the right aggregate. You need a good durable rock, either river gravel or quarried and crushed. Particle size distribution = similar to granular. Water Content : Problem: you must have the right amount for compaction (OMC – similar to granular materials) BUT you must also have just the right amount for the hydraulic reaction to take place.

Too little water  not all the binder will be activated  reduced strength. Too much water  free water after the reaction has finished  air voids left after evaporation  reduced strength again. [This restricts the practical combinations of particle size distribution and strength]

•

Batch it and mix it (or mix it continuously) This is a twin-shaft batch mixer.

•

The alternative is a drum mixer, allowing HBM to pass through continuously. This gives higher productivity. Transport it to site – usually in a normal no rmal delivery truck.

5


DEPARTMENT

OF


• Place it and compact it. Basically the same as for granular material, except that you would usually put it through a paver to get good level and thickness control.

• Cure it – usually by spraying a bitumen seal to stop the water evaporating. Pavement Quality Concrete (PQC) • Get the mix design right. This is a real concrete, which means it will be too wet for roller compaction (usually); it will need vibrating.

Water Content

Range of mixture design options for wetforming

Range required to allow chemical reaction to take place

Range for wet-form workability

• •

Range for rollerBatch it and mix it.compaction workability

Range of mixture design options for roller-compaction

Transport it to site – usually in a purposepu rpose built concrete truck.

•

Pave it. Wet concrete needs to be enclosed by formwork of one sort or another.

Range required to achieve design strength

Options:

Cement Content

Fixed for form m; ch checke cker boa boarrd pattern – slow process. Fixed form; continuous side rails – much quicker. Slip form; with a purpose built slip-form paver – quicker still; most commonly used nowadays.

6


DEPARTMENT

OF


• Get the surface texture right A set of steel tines is dragged across the surface of the fresh concrete immediately after paving. [can also be achieved by sawing hardened concrete]

Grooving:

The fresh concrete surface is brushed using an appropriately heavy-duty brush to form a ridged finish.

Brushed finish:

A retarder is sprayed onto the finished concrete and loose mortar is brushed away from around the larger aggregate pieces about 12 hours afterwards.

Exposed aggregate:

A sheet of rough fabric is dragged over the surface of the wet concrete, leaving a rough finish.

Burlap drag:

• Form joints (usually) Joint Types

Dowel bar (smooth)

Joint Forming

Joint Seal Slip Coating

Filler Board Dowel bar (smooth)

Expansion Cap

Expansion Joint

Crack

Joint Seal

Slip Coating

• Cure it, usually by means of a colourless Contraction Jointwater spray usebarwet cloth or regular Tie (ribbed) Joint Seal

Bottom crack inducer Sealant groove cut once joint has been formed aluminium-based sealant but can

application.

also

Joint former strip

Crack Warping Joint

Saw Cut (at early age)

• Cure it – usually by spraying a colourless seal to stop the water evaporating.

7

U NIVERSITY NIVERSITY OF NOTTINGHAM Pavement Engineering - Module H23P01 3.

DEPARTMENT

OF


Hot-Mix Asphalt (HMA)

• Heat the bitumen to 140 to 180°C; keep it in a hot storage tank. • Dry the aggregate thoroughly in a drum dryer. •

Sieve the hot dried aggregate into size fractions; store in hot bins.

Can combine in a drum mix plant

•

Mix bitumen + aggregate (about 30 secs)  asphalt. Hot bins

Batch Mix Plant: Bitumen tank

Elevator Dryer

Mixer Aggregate feed

Drum Mix Plant:

Combined Dryer-Mixer Drum

Hot storage hopper

Cold bins

Drum mixers give higher productivity. They rely on accurate proportioning of moist aggregate since the bitumen is fed directly into the drying chamber, which takes the form of an inclined rotating drum. The drying chamber doubles as the mixing chamber and the hot h ot mixture is fed out continuously from the drum to a hot storage hopper before being dropped into the back of a waiting truck.

• Transport to site in a thermally insulated truck.

8


DEPARTMENT

OF


•

Pave while the mixture is still hot, e.g. 110- 130°

•

Compact before it cools down too much; either pneumatic or vibratory for main compaction, dead weight steel drum for the final finish

How quickly does the mat cool? For example, assuming a 110°C paving temperature: 120 110 ) C ° ( 100 e r u t 90 a r e p m 80 e T

80m layer 40m layer @ 2 minutes

70

20m layer

@ 4 minutes

60 0

20

40

60

80

Depth (mm)

The practicalities of compaction mean that layer thicknesses tend to be between 25mm and 120mm.

9


DEPARTMENT

OF


MATERIALS 1.

Sustainability and Cost

Pavements have to be cheap – that is an absolute requirement. However, we also have to try to limit environmental costs. The key concept is embedded (or embodied) energy , which is the total energy used to manufacture, transport, process etc. every component of the pavement. Approximate costs and embedded energies for pavement component materials: Material Embedded Cost Energy (£/Tonne) (MJ/Tonne) Embedded Direct Energy Ingr Ingred ediient ents Sand Sandss and and grav gravel elss 5-10 0.1-0.2 5-8 Crushed rock aggregate 20-25 0.4-0.5 12-15 Bitumen 3200-3800 Portland cement 4500-5000 Reinforcing steel 23000-27000 Mixtures Hot-mix as asphalt 600-800 12-16 25-40 Cold-mix asphalt 150-200 3-4 15-30 Lean concrete 450-500 9-10 15-25 Pavement quality concrete (PQC) 750-1000 15-20 35-50 Reinforced concrete 1100-1500 22-30 40-55 Transport All materials (per journey km) 12-20 The question is: what is a MJ worth? In terms of fuel cost it is only about £0.01-0.02! It also represents about 65g of CO2, and this might be valued anywhere from £0.002 to £0.02. To be environmentally conservative, the embedded energy costs in the table are based on an equivalence of £0.02 per MJ. So: although the table may be exaggerating, hot-mix asphalt and concrete carry a significant embedded environmental cost. BUT what about the issue of traffic , responsible for about 36% of all energy consumed in the UK? A Nottingham research project found that energy losses attributable to road stiffness were around 100MJ/m2 for a heavily trafficked concrete pavement over a 40year life. For asphalt this went up to around 250MJ/m2. These translate to about 300 and 1000MJ/Tonne assuming normal pavement thicknesses – which as you can see are quite significant numbers. And this does not include energy loss due to surface roughness, which is likely to be a much bigger factor and definitely needs researching! Conclusion: we really should take the environmental cost of road pavements seriously.

10


DEPARTMENT

OF


Unbound Material

What’s really going on?

A N

A

F

B

B

Initial State

After Strain

Individual stones have to translate and rotate. stones slide against one another. this is resisted by friction (typically around 30-35° for crushed rock, less for gravels); stone shape is also obviously important. 

Shear strength Can be measured in a shear box

After Strain

Initial State

Can also be measured in a triaxal apparatus sin φ = ½(σ1 – – σ2) / ½(σ1+σ2)

σ1 σ2

Shear stress τ

Angle of internal friction φ (typically 55° for a crushed rock

limiting τ /σ ratio (typically 10 for a crushed rock)

σ2 α

σ1

σ2

increasing σ1 Normal stress σ

Don’t confuse stone-stone friction angle with .

11


DEPARTMENT

OF


What properties affect shear strength?       

particle shape; angular is good stone-stone friction; not much effect stiffness modulus of the rock; not much effect particle size; big is good particle size distribution; broadly graded is good particle packing; dense is good water content; high suction is good (see p3) – increases effective stresses. Shear stress

τ

Apparent cohesion c (due to stone interlock)

Angle of internal friction φ Normal stress σ

Stiffness With an unbound material you can’t really talk about a Young’s Modulus because the behaviour is so non-linear and stressdependent . On the other hand it is convenient to pretend that it has a Young’s Modulus, so instead we call it Resilient Modulus or Stiffness Modulus and we have to remember that its value changes depending on the level of applied stress.

Approximate shear modulus

Shear Stress Ultimate stress (= shear strength) Applied Stress

Hysteresis loop – represents energy loss

Cycle no: 1 2 3

10

100

1000

10000

Typical values: Solid rock is approximately linear elastic with a stiffness modulus of 100 and 200GPa; unbound materials typically have a modulus in the range 20-250MPa . What properties affect stiffness?      

particle shape; not much effect stone-stone friction; high friction is good stiffness modulus of the rock; stiff is good particle size; big is good particle size distribution; not much effect particle packing; not much effect

12

Shear Strain


DEPARTMENT

OF




water content; high suction is good (see p3) – increases effective stresses. California Bearing Ratio (CBR)

Plunger Soil under test

50mm

Force (kN)

Force F Displacement d (50mm/min)

F2 F1

152mm diameter

125mm height

Steel mould

1.27m m

2.54m m Displacement

CBR = max {(F1/13.2) ; (F2/20.0)} × 100 where F1 and F2 are in kN

This is a convenient general measure of quality, but has no fundamental meaning. Very Very,, ver very y, ver very y app appro roxi xim mate ate rel relat atiionsh onship ipss wi with sti stiffne ffnesss: a) b)

E = 10 × CBR CBR E = 17.6 × CBR 0.64

Confined Compression A triaxial test is better (more fundamental meaning) but it is complicated; and the stress conditions are usually not right for a pavement. Confined compression is an alternative. Load applied via full-face platen Load applied via

Eight wall segments

full-face platen Springs

Calibrated steel and rubber band

Side plates free to move

Locking nuts fix side plates in place

PUMA (Pr ecision eSpringbox cision Unbound Material Analyzer)

Adjustment control for start conditions

13


DEPARTMENT

OF


These tests are designed to give about the right level of stress in the material and so hopefully about the right stiffness modulus for pavement d esign. You can also get a measure of resistance to deformation accumulation under repeated load. Dynamic Cone Penetrometer (DCP)

Number of blows

Drop weight

CBR (%) Granular

Scale Anvil

Penetration rate p (mm/blow)

Core hole

Soil

Depth

Depth Equation used in UK : log10[CBR] = 2.48 – 1.057 log10[p]

The DCP is especially convenient for doing down a core hole during evaluation of a failing pavement. Plate Tests Nowadays the usual way of doing these is by means of a portable dynamic plate test (DPT). It is a quick, practical method for getting the in-situ stiffness of a pavement foundation. You have to remember of course that it is affected by any layer within about 1m of the surface.

Peak load (P) Peak deflection (δ )

Drop weight Load

Deflection

Load cell and velocity transducer (geophone)

Rubber buffers

Time delay due to ground inertia Time

Loading plate (radius r)

Boussinesq’s equation for deflection under a rigid circular plate load:

Modulus E Poisson’s ratio

δ Therefore:

14

= P (1 – ν ν2) / 2rE

E = P (1 – ν 2) / 2r δ


DEPARTMENT

OF


Static plate tests are also possible. The standard test to evaluate an airfield pavement subgrade is a static 762mm diameter plate. This table illustrates the fact that different stress conditions give different d ifferent stiffnesses. Material

Very soft clay soil Firm clay soil Sandy soil Gravel capping Sub-base Granular Base

Stiffness Modulus (MPa) Triaxial DPT In the Pavement (confining stress (100kPa contact (K-Mould, PUMA and 20kPa; deviator pressure) Springbox generally stress 0-100kPa) give similar results) 10 5 15 50 30 80 75 30 50 125 50 80 250 75 150 500 100 250

Permeability Usually it is assumed that a sub-base or capping layer will be ‘free-draining’ – which is not entirely true so you might want to measure it.

Material Description

Well graded gravels Poorly graded gravels Silty gravels Clayey gravels Well graded sands Poorly graded sands Silty sands Clayey sands Low plasticity silts Low plasticity clays High plasticity silts High plasticity clays

Typical permeability range (m/sec) 10-5 to 10-3 5× 10-5 to 10-3 10-8 to 10-4 10-8 to 10-6 5× 10-6 to 5× 10-4 5× 10-7 to 5× 10-6 10-9 to 10-6 10-9 to 10-6 10-9 to 10-7 10-9 to 10-8 10-10 to 10-9 10-11 to 10-9

Measurement of volume flow Q

Water supply Head difference H

Overflow

Permeability k = QL/AH

Porous end restraints L

Lid

Specimen (cross-sectional area A)

15


DEPARTMENT

OF


Hydr Hydra aulic lically-B ly-Bo ound Mater terial (H (HB BM)

The word ‘Hydraulic ’ means that the binder needs water in order to be activated. The most common binder of this type is Ordinary Portland Cement (OPC), but fly ash (a.k.a. pulverised fuel ash – PFA), lime or ground granulated blast-furnace slag (GGBS) are often mixed in or even used without OPC to give a slow-setting (and cheaper) material. HBMs come in a range of different strengths / qualities: Stabilised soil – insitu mixing process; roller compacted; results in a partially bound material. [Compressive strength < 2MPa] HBM subbase – plant mixed; uses gravel or crushed rock aggregate; still roller compacted [Compressive strength 2-10MPa] HBM base – fully bound crushed rock; usually roller compacted; also known as ‘lean concrete’. [Compressive strength 5-20MPa] Pavement Quality Concrete (PQC) – strong, fully bound concrete; wet-formed; vibratory compaction. [Compressive strength 30-50MPa]

Strength Strength can be measured in several ways. Flexural is most realistic for a pavement. Compressive (cube or cylinder) is the most convenient.

Indirect Tensile

or

Compressive Tensile

Flexural

16


DEPARTMENT

OF


Tensile Strength In the end HBMs fail in Tension . The problem is that tensile strength is not easy to measure. You could do Indirect Tension . This is much easier and can be done on a core of material taken from a road – but the stress conditions in the test are a little complicated and it is hard to be certain what the result means. Flexural Strength This is much closer to what happens in a pavement. ½P

½P L/3

½P

L/3

L/3

h

Max Bending Moment (M) = P × L / 6

[beam width = b] ½P Compression

σ

Assumption:

M = 2×

∫ bσ(2x/h)x dx

[0 to h/2]

= 4(bσ/h).[⅓x3]

x

= σh2/6 = 6M/h2

σ

Tension

>σ The non-linear relationship between and ε in tension means that:

Reality:

M >

σ

σ

σ

σh2/6

< 6M/h2

It all adds up to the flexural strength (i.e. the tensile strength deduced from a flexural test) being 10-15% higher than the real tensile strength. But this doesn’t really matter since the pavement will behave more or less like a flexural test specimen. Compressive Strength This is the least meaningful strength test – but the most common, because beca use it is so convenient. For purposes of quality qu ality control it is therefore ideal.

Idealised shape

c t

Real shape

Low friction graphite applied to platens

Analysis: 1. Failure occurs when εt,failure is reached. 2. If zero friction: εt = ν × εc = ν σc / E But things aren’t quite linear, so: 3. εt > σt / E 4. So, combining: σc > σt / ν 5. And if friction ≠ zero: σc >> σt / ν

Adding all this uncertainty together, the compressive-tensile strength ratio can be anything from around 5 to 15.

17


DEPARTMENT

OF


Conclusion: Compressive strength is not a fundamental measure . It should never be relied upon in pavement design. Rule of thumb: 1. Limestone aggregate results in low compressive-tensile strength ratio (e.g. 7). 2. Granite aggregate results in moderate compressive-tensile strength ratio (e.g. 10). 3. Gravel aggregate results in high compressive: tensile strength ratio (e.g. 12). Strength Gain with Time All HBMs become stronger with time. We usually design on a 28day strength but often use 3-day or 7-day for quality control.

60

) a P 50 M ( h 40 t g n e r t S 30 e v i s s 20 e r p m o 10 C 0 1

10

100

1000

Age (days)

Fatigue All HBMs tend to follow a similar fatigue characteristic when plotted as numbers of load applications to failure against the ratio of applied stress to failure strength.

1

o i t a R h t g n e r t S : s s e r t S l a r u x e l F

0.9

/s f = 1.064-0.064log(N)

σ

0.8

0.7

9.2MPa Compressive strength 0.6

32.9MPa Compressive strength 25.9MPa Compressive strength; 0.5% steel fibres

0.5 1

10

100

1000

10000

100000

1000000

Number of Load Applications to Failure

Durability The key to durability is a low permeability – which means small non-interconnecting voids. The principal cause of damage is water; so if water can’t penetrate then damage won’t occur. The danger comes when voids become filled with water and there is no clear escape path. So be careful in design to avoid water becoming trapped. Frost is another potentially serious problem. Water expands when it turns to ice, which means that if the water is inside a nearly-closed void at the time it freezes then the expansive pressure is likely to fracture the surrounding material. In PQC this can be seen as areas of surface flaking, giving a very rough ride quality to vehicles. Measures to combat frost damage include:  Air entrainment ; introduce tiny air bubbles (e.g. 5% by volume) using a chemical additive in the mix.  Keep the strength high.

18


DEPARTMENT

OF


Thermal Properties HBMs are rigid solids; they are therefore susceptible to thermal expa nsion and contraction. This is the reason for the use of joints in concrete pavements. The key property is the coefficient of thermal expansion (α). Aggregate River gravel Igneous rock Limestone

Coefficient of thermal expansion 13 × 10-6 10 × 10-6 7 × 10-6

(per degree C)

Stiffness HBMs are more or less linear elastic . Stiffness is not usually measured in the laboratory because it is more difficult to do + a bit less important important than strength. Here are some typical values: Mixture Type [compressive strength]

Typical stiffness

Pavement quality concrete [40MPa] Str Strong cement-bound base [10-20MPa] Weak cement-bound base [5-10MPa] Slag etc. bound base [5-10MPa] Hydraulically-bound sub-base [2-5MPa] Stabilised soil [< 2MPa]

30000-40000 MPa 15000-25000 MPa 5000-15000 MPa 3000-10000 MPa 2000-5000 MPa 100-300 MPa

The Stiffness of a Discontinuous Layer A HBM is often designed to end up in a cracked state; sometimes it is deliberately cracked when joints are formed. The effective stiffness in-situ will therefore be less than that of the intact material. A HBM base with an initial stiffness of 10000-20000MPa can easily end up with an apparent in-situ stiffness of no more than 5000MPa. M

1. Curvature = M/EI = 12M/Eh 3

3

I = h /12 per m width

1.

h

M

Stress distribution

2.

L

L

h

Uneven stress zone

2. Strain at top = ( θh/4)/d ≈ θ /3 Stress at top = θE/3 ( 0 at mid depth) Moment = (3/256)×σ h2 = θh2E/256 Curvature = θ /L = 256M/(h 2EL)

E

L

3. Moment = τhL/2 = gδ hL/2 Radius of curvature = (L2/4)/2δ = L2/8δ Curvature = 8δ /L2 = 16M/ghL3

L

Combine: Curvature = M[12/Eh3+256/(h2EL)+16/ghL 3] Curvature also = M/E eff I = M[12/Eeff h3]

d ≈ 0.75h L

3.

h

Eeff = 1/[1/E+64h/3EL+4h2/(3gL3)]

Slip stiffness g = τ/δ

19


DEPARTMENT

OF


Asphalt

Bitumen (otherwise known as binder) Bitumen is a liquid , even in service. This means that we need to worry about its viscosity , which is a function of temperature . The problem is that viscosity is a little complicated to measure directly, so the following two substitute tests are commonly used. Penetration : the penetration of a needle into a container of bitumen.

Ring and ball softening point : the temperature at which a steel ball drops through a prepared disk of bitumen.

Bitumen at 25°C

15g

Water bath (stirred continuously in most specifications)

Allowed to fall for 5 seconds

Disk of bitumen

Metal ring

Steel balls

p

Softening point is the temperature at which the bitumen disks sag by 25mm (touching the plates positioned beneath them)

Penetration p measured in tenths of a millimetre

Heat at 5 °C per minute

Penetration (pen)

Ring & Ball Softening Point (SP) Penetration Index (PI):

(20-PI)/(10-PI) = 50 (log 10[100/pen(at temperature T)])/(SP-T)

100000000

50 pen

10000000

) s . a P ( y t i s o c s i V

100 pen

1000000

200 pen

100000 10000 1000 100

Compaction

10 1

Mixing

0.1 0.01 0

Real bitumen behaviour

25

50

75

100

125

Temperature (Celsius) 20

150

175


DEPARTMENT

Load

OF


There are two main components to bitumen behaviour, elastic and viscous , with a visco-elastic bit between them just to make things awkward. Time

Displacement Elastic Viscous Visco-elastic

Visco-elastic Viscous

Elastic

Time

Bitumen can also fracture . Tensile stress at failure is generally around 2-3MPa. Ageing Bitumen is that it changes with time due to oxidation and absorption by aggregate. Short-term ageing occurs during mixing, transporting, placing and compacting while the bitumen is at high temperature; pen goes down by about 25%. Long-term ageing occurs gradually such that bitumen becomes ever harder during its lifetime.

100 90

) m m d ( n o i t a r t e n e P

80 70 60 50 40 30 20

Texas Texas (Bens on, 1976) Texas Texas (Bens on, 1976) Michigan (Corbett and Schweyer, 1981)

10 0 0 .0 1

0.1

1

10

10 0

Time (years)

Binder Modification Bitumen chemistry is a complicated subject and there are many different products on the market that are claimed to improve the properties of bitumen. You should be aware of: 

Polymers : Styrene-Butadiene-Styrene (SBS), Styrene-Butadiene Rubber (SBR ) and Ethyl Vinyl Acetate (EVA) will all increase the viscosity at high temperatures but not at low temperatures. They may also increase fatigue life, particularly SBS. to recycle vehicle tyres.  Natural rubbers : This has been driven by the need to  Sulphur: enhances workability at high temperature (>115°C); becomes solid at lower temperature.  Manganese: increases the cross-linkage between molecules and thereby increases viscosity and stiffness. The problem is that the bitumen beco mes brittle. Bitumen-Filler Mortar

21


DEPARTMENT

OF


Filler is the term used for silt-size particles (2-75μm) which are added to enhance the performance of the bitumen. It stiffens and strengthens the bitumen + gives extra resistance to fatigue cracking. 10

Pure bitumen +5% filler +15% filler +35% filler +50% filler +65% filler

9

) a P M ( e r u l i a F t a s s e r t S

8 7 6 5 4 3 2

Warm and/or slow loading

Cold and/or rapid loading

1 0 0.1

1

10

Strain a t Failure (%) (%)

Typically filler is added at slightly more than the mass of bitumen – which means about a 2:1 bitumen to filler ratio by volume. Bitumen-Aggregate Adhesion This is a vital property if an asphalt is going to work properly. A lack of adhesion will mean that fracture can take place along the interfaces between aggregate particles and the bitumen-filler mortar – and this will shorten shorten the fatigue life of the asphalt; i.e. it will crack. But what causes poor adhesion? The problem is that all aggregates would really prefer water to bitumen – hence the need to ensure that particles are absolutely dry before mixing with bitumen. Aggregate chemistry determines how easily the bond with bitumen is broken down, but it all happens much more rapidly if there is water about. What can be done about poor adhesion? Answer: include an additive, usually hydrated lime . Asphalt Stiffness Modulus Asphalt does not have a single modulus value; it is both temperature and loading rate dependent because bitumen is temperature and loading rate dependent. Predicting Asphalt Stiffness Modulus

22


DEPARTMENT

OF


Step 1 – Binder stiffness : The most widely adopted approach is to use van der Pohl’s nomograph, nomograph, a chart which relates binder stiffness ( E binder (T ), ), softening binder ) to temperature (T point (SP (SP ), ), penetration index ( PI PI ), ), and load pulse duration (t (t ). ). The following formula matches the nomograph over a restricted range of input parameters.

E binder [in MPa] = 1.157 10 –7 × t-0.368 × 2.718 –PI × (SP – T)5 Step 2 – Mixture Stiffness : The other controlling parameter here is Voids in Mixed Aggregate (VMA), VMA), = the percentage of the mixture which is not aggregate = binder % + air %. Many different equations have been proposed. Here is one.

Emixture [in MPa] = E binder × [1 + (257.5 – 2.5VMA)/(n × (VMA–3))]n where:

n = 0.83 × log10[4 × 104/E binder ] VMA = V binder + Vair in %; E binder is in MPa

These are easy enough to use, but the problem is that we don’t normally know the volume of binder, only the mass percentage; so there is more calculating to be done. Example : predict the stiffness modulus of an asphalt with 5% binder by mass and 7% air voids at 10°C under fast highway traffic if the softening point of the binder is 49 °C and the penetration index is -0.5.

a) What loading loading time? time? Fast Fast highway highway traffic traffic:: say 100kph; 100kph; i.e. i.e. 27.8ms 27.8ms-1. Tyre contact is typically 300mm long, therefore loading time at the surface = 0.3/27.8 = 0.0108s. Rule of thumb: loading time (secs) = 1/speed (kph); i.e. 0.01s. So:

E binder = 1.157 10 –7 × 0.01-0.368 × 2.7180.5 × (49 – 10)5 = 93.7MPa

b) What VMA? We need to estimate the density of the rock in the asphalt, generally in the range 2500-2900kg/m3, say 2700 here. We also need the density of bitumen, typically 1030kg/m3. In 1000kg of asphalt: asphalt: 5% binder = 50kg; binder binder volume volume = 50/1030 = 0.0485m 0.0485m3. 95% rock = 950kg; rock volume = 950/2700 = 0.3519m3. Combine: 0.4004m3. But 7% air voids  total volume of (100/93) × 0.4004 = 0.4305m3.  binder volume percentage of 0.0485/0.4305 × 100 = 11.3% Therefore VMA = 11.3% + 7% = 18.3% So:

n = 0.83 × log10[4 × 104/93.7] = 2.183 Emixture = 93.7 × [1 + (257.5 – 2.5 × 18.3)/(2.183 × (18.3–3))]2.183 = 7270MPa

23


DEPARTMENT

OF


Measuring Asphalt Stiffness Here are three possible tests: TensionCompression F

Four-point Bending (Flexure)

Indirect Tensile F

F/2

F/2

Diameter = d d

h

L

F/2

L

L

L

F/2

Thickness = b F

E = 4FL/πd2δ

Thickness = b E = F( ν+0.27)/bδ

E = FL[(23L 2/4h2)+1+ ν]/hbδ

The tension-compression test is not very practical to carry out. The indirect tensile test is very quick and easy, eas y, but has complex stress conditions; nevertheless it is widely used. The four-point bending test is usually only carried out when testing for fatigue, but it also gives a good stiffness measurement. Note: stiffness depends on loading rate. Therefore you usually have to correct the result before using it in pavement design. Typical Stiffness Values Material

Dense asphalt base (50 pen binder) Dense asphalt base (100 pen binder) Surfacing

Stiffness Modulus (MPa) at 20 C In the Laboratory In the Pavement (e.g. 125 milliseconds (e.g. 10 milliseconds to peak load) to peak load) 5000 7000 3500 5000 2000 3000

These values are for new asphalt , immediately after laying. But bitumen ‘ages’, which means that the stiffness of a mixture will increase throughout its life as the viscosity of the binder increases. For example, if a new dense asphalt base with 50 pen binder has an initial stiffness (in the road) of 7000MPa at 20 °C (by which time the penetration of the bitumen has already decreased to around 35 purely as a result of mixing and laying), then this will probably have increased to about 9000MPa after 10 years in a climate such as the UK, with much more rapid stiffness increase in hotter climates. Note approximate temperature correction suggested by the Transport Research Laboratory: log10(ET) = log10(E20C) – 0.0003 × (20-T)2 + 0.022 × (20-T) Fracture and Fatigue of Asphalt

24


DEPARTMENT

OF


Low-temperature fracture can occur in continental climates. Asphalt expands and contracts with temperature changes, with a typical thermal expansion coefficient α of around 1.8 × 10-5 per degree C. If it is too brittle at the low point of the temperature cycle then it may simply break.

More usually, an asphalt cracks due to fatigue under millions of load applications. Localised fractures begin to form at particle-to-particle contacts and then slowly grow, eventually joining to form proper cracks. Growth rate is primarily controlled by the magnitude of strain in the mixture under load. You can see the effect in a loss of stiffness even before any cracks can actually be seen.

1 0.9

s s e n 0.8 f f i t 0.7 S l a 0.6 i t i n 0.5 I f o 0.4 n o i t 0.3 r o p 0.2 o r P

10°C 20°C

0.1

30°C

0 0

0.2

0.4

0.6

0.8

1

Proportion of Life to Failure

Measuring Fatigue Resistance The following three tests are commonly used.

Indirect tensile

4-point bending

Trapezoidal (or 2-point bending)

1000

Usually what you do is to carry out a series of tests at different stress/strain levels and to plot the lives to failure against strain (under load at early stages of the test). Failure is generally taken to be a 50% loss in stiffness. The results tend to be only slightly affected by temperature. Permanent Deformation

Fatigue characteristic: Fatigue characteristic Slope typically about -0.25 in log-log space

)

6 -

0 1 x ( n i 100 a r t S l a i t i n I

10 100

1000

10000

100000

Number of Load Applications to Failure

25

1000000


DEPARTMENT

OF


Rutting is always a danger. To avoid it you need:     

Good angular aggregate A sensible gradation (i.e. particle size distribution of the aggregate) Good compaction A hard binder (so rutting only happens in hot weather) At least 2% air voids, otherwise you begin to lose particle-particle contacts.

Measuring Permanent Deformation The most usual forms of test are: Repeated load axial (RLA)

Vacuum repeated load axial

Wheel tracking

2

Typical RLA data: (100kPa, 40°C) Note the effect of moisture conditioning – i.e. soaking in water Rule-of-thumb: 1% strain is a safe limit; 2% spells danger; 3% means trouble!

1.8 1.6

) 1.4 % ( n 1.2 i a r 1 t S l 0.8 a i x A 0.6

Mix A - Dry Mix A - Moisture conditioned Mix B - Dry Mix B - Moisture conditioned Mix C - Dry Mix C - Moisture conditioned

0.4 0.2 0 0

500

1000

1500

2000

2500

3000

3500

4000

Number of Load Cycles

Durability As mentioned already, bitumen ageing means asphalt becomes stiffer – a good thing – but also more brittle – not so good!

Water bath

But we also have to worry about water damage, leading to loss of adhesion between aggregate and bitumen. So:

test test one one bat batch ch – usu usual ally ly indi indire rect ct tens tensil ilee str stren engt gth; h; Simple Soaking soak a second batch for a while (many different specifications); Procedure test the second batch; express the result as retained strength , i.e. ratio of soaked to unsoaked (in %); if > 75% (specifications vary) then OK Mixture Design 26


DEPARTMENT

OF


Aggregate Particle Size Distribution The choice depends on where the material is going in the pavement.

Bas Base: Surface:

usua usuall lly y lar large gerr si sized zed par partticles cles – bec becau ause se layer ayerss are are thick hick;; als also o les lesss bi bitumen umen.. usually smaller particles – better ride quality; better durability; thin layers; high binder content, therefore expensive.

Gradation – usually go for a broadly graded mixture to give optimum aggregate packing, deformation resistance, stiffness + minimise minimise binder content. For example the US Superpave specification uses a Fuller curve with n = 0.45 for what is termed asphalt concrete (sometimes called dense bitumen macadam in the UK). Fuller curves:

% passing size d = (d/D)n

Another option is gapgrading. This means that there are large stones and smaller particles but not much in between. Hot rolled asphalt and stone mastic asphalt are both gap-graded mixtures. A very different surface material is porous asphalt , with a near single-sized gradation designed to allow easy water drainage.

100 90 80 g n i s s a p e g a t n e c r e P

where D = maximum particle size

Asphalt Concrete Hot Rolled Asphalt Porous Asphalt Stone Mastic Asphalt Fuller Curves (n = 0.45)

70

Critical particle size (Hot (Hot Rolled As phalt)

60 50 40 30 20 10 0 0. 01

0.1

1

10

100

Sieve size (mm)

The concept of a critical particle size can be useful. This is the point where the actual gradation just touches a Fuller curve. The gradation to the right of the critical size is less steep than the Fuller curve, which means that coarser particles are always separated by plenty of smaller-sized particles, right down to the critical size. However, the gradation to the left of the critical size is steeper than the Fuller curve, which means there are never enough small-size particles to fill the gaps between larger ones. This means that particles above the critical size form the aggregate skeleton; particles smaller than the critical size are really just floating in the binder. So:

Asphalt concrete – similar to the Fuller curve, therefore no clear critical size. Hot rolled asphalt – very clear critical size at around 1mm; needs good sand-sized aggregate. Stone mastic asphalt – should be a large critical size if well designed, but with almost enough particles to fill the gaps; therefore very sensitive to errors in gradation. Porous asphalt – definitely a large critical size. Binder Content

27


DEPARTMENT

OF


Don’t forget: binder content as a percentage by mass is quite different from its percentage by volume – because the specific gravity of bitumen is so much less than that of rock. The volume percentage (typically 8-12%) determines mixture properties; the mass percentage (typically 4-6%) is most easily measured and therefore specified. The Marshall Mix Design Method A practical technique developed in the 1950s:      

Select a gradation. Make up a series of mixes at different binder contents. Prepare specimens using a Marshall hammer for compaction. Measure achieved densities. Carry out Marshall tests to derive stability and flow values. Determine optimum binder content based on stability, flow and density.

60°C

(50mm /min)

Stability

F

Density

• • • Marshall Test

•

•

•

F 10kg

•

Stability

Binder content

Flow maximum

50mm 100mm

• • optimum

optimum Binder content

0.45m

•

•

•

• • •minimum

Flow

Marshall Hammer

Binder content

The Superpave Mix Design Method This design approach grew out of research in the US in the 1990s and is principally concerned with optimising mixture volumetrics:       

Select a gradation (only broadly graded mixtures covered; filler-binder ratio by mass between 0.6 and 1.2). Make up a series of mixes at different binder contents. Prepare specimens using a gyratory compactor . Measure achieved densities. The optimum binder content is the one that gives a void content of 4%. Prepare further specimens at the optimum binder content. Check voids at light and heavy compaction.

28


DEPARTMENT

OF


Void Content

Void content check on specimens at ‘design’ number of gyrations (50125 depending on traffic)

• • 4%

optimum

•

•

•

Binder content content

Check void content after ‘maximum’ level of compaction (75-205 gyrations depending on traffic)

Gyratory Compactor

>2%

Check void content after ‘initial’ compaction (6-9 gyrations depending on traffic) > 11% (unless traffic < 3msa)

The additional checks at the end ensure that compaction doesn’t occur too easily, an indicator of poor aggregate interlock, and that void content will never fall below 2%, even under heavy trafficking. Both checks are intended to avoid the danger of rutting. Binder Grade (i.e. Penetration or similar measure) This depends largely on climate. The key point is that the binder should be able to perform satisfactorily over the full range of temperatures temperatures experienced in the pavement.

Low temperature  danger of fracture and fatigue High temperature  danger of rutting Desirable working range range of binder viscosity viscosity ≈ 5 × 103 to 107 Pa.s. e.g. 50 pen binder gives a working temperature range of around -10 to +45°C. In some climates it is just not possible to find a conventional binder which covers the expected temperature range satisfactorily. In these cases there are two options:  

Accept that damage will occur and plan accordingly. Pay extra and use a modified binder , extending the working temperature range.

Filler Filler is an extremely important part of the mixture. It is a very effective binder additive , multiplying stiffness, fracture and fatigue strength by a factor of up to about 3. The successful use of a good quality filler will:     

enhance mixture stiffness and fatigue strength; assist chemically in promoting aggregate-bitumen adhesion; inhibit drainage of hot binder off the aggregate during transportation; not prevent proper mixing; not prevent proper bitumen-aggregate contact.

29


5.

DEPARTMENT

OF


For best results, filler % by mass ≈ bitumen % by mass, maybe a little more. Cold-Mix Asphalt Conventional ‘Hot-Mix’ materials have excellent mechanical properties BUT: -

aggregate has to be heated and dried thoroughly, this means that certain potential aggregates are excluded, the material has to be placed and compacted before it cools.

It would therefore be extremely useful to find an alternative which could be mixed with cold, wet aggregate. There might be significant environmental benefits as well as a reduced energy demand. Options:

Bitumen Emulsion Foamed Bitumen

Bitumen Emulsion Bitumen emulsion is a ‘suspension’ of bitumen droplets in water, created as follows: Break the bitumen into very small droplets, typically 1-20 microns in size. This requires a Colloid Mill (which unfortunately is expensive and uses up significant energy!) Water + emulsifier

Hot bitumen (100-140°C) 10006000rpm The Emulsifier Emulsifiers are hydrocarbon chains with positively or negatively charged ions at the end of the chain. Cationic:

positively ch charged

Anionic:

negatively charged

+

Emulsion (<90°C)

_

The emulsion works because the polymer parts of the emulsifier molecules attach themselves to the bitumen droplets. This leaves each droplet surrounded with charge and means that droplets repel each other. These forces are enough to prevent droplets

30


DEPARTMENT

OF


coalescing (combining) since bitumen and water have very similar specific gravities (1.00 and 1.03 respectively).

Bitumen droplet Charged emulsifier molecules Water + free emulsifier

Typical proportions: 40-70% bitumen. Big advantage of emulsions: they can last for months (with the occasional stir). stir). Foamed Bitumen Bitumen foaming is an alternative technique to produce a binder which is workable at normal ambient temperatures. This is done as follows: Expansion Ratio Maximum expansion ratio (typically 10-20)

Hot Bitumen (140-200°C)

Water (2-6% of bitumen) Air

Foam Time Half lives (typically 15-30 seconds)

Advantage of foamed bitumen over emulsion: it doesn’t need so much water. Disadvantage : you have to use it within about a minute! (i.e. straight into the mixer)

How Cold Mix Works For both emulsion and foamed bitumen to achieve reasonable mixing in of the binder the aggregate must be wet . The water content required is typically 2-3%. in emulsion or flakes of bitumen foam) heads Problem : the binder (bitumen droplets in straight for the water, which is mostly found amongst the fine aggregate particles.

31


DEPARTMENT

OF


Result: coarse particles often don’t get coated properly with bitumen, leaving a partially bound material.

In detail:

1. Particles coated with a film of water before mixing

2. Bitumen droplets or flakes of foam coalesce onto the aggregate

3. Most of the water evaporates; some is trapped

Look in even more detail: Compaction squeezes particles together, forcing the water away from contact areas and creating bitumen bonds.

The final trapped water content is usually 0.5-1.0% (by mass). The rest of the wa ter can, in theory, evaporate; but this is highly weather dependent. Cold-mix therefore needs good weather. Additives Because the UK is not always warm and dry, it is common to add a small percentage (12%) of cement, lime or fly ash to the mixture in order to take up some of the water, helping the bitumen to attach itself to the aggregate. Air voids

Volumetrics Dry Density We need good compaction, and this depends on the ‘fluid’ content (‘fluid’ means water or emulsion or – more debatably – foam residue). Optimum compaction is achieved at

0% 5% 10%

Optimum Fluid content

32 Fluid content


DEPARTMENT

OF


optimum fluid content. This is typically about 6% (by mass). So, say 2% water content is present in the aggregate already. This means that we can add 4% extra fluids. In an emulsion , this can be up to 70% bitumen, which means that the upper limit on bitumen content is just under 2.8% (compared to a typical 4-5% for a hotmix asphalt). If you want more bitumen then you just have to accept a lower density. Foamed bitumen needs a slightly higher initial aggregate water content to get good mixing. The result is that you end up with more or less the same limit as for emulsion.

Illustration of volumetrics: By Volume:

78%

7%

10%

5%

FillerBinder BinderAir Air Hot mix; dense grading;

Aggregate By Mass:

87.5%

8%

By Volume:

78%

4% 6 % 7% 2%

By Mass:

90%

By Volume:

84.5%

By Volume:

69%

By Mass:

84.5%

0%

4.5% 2.5% 3% 0%

73%

By Mass:

4.5%

6%

7% 6%

7%

9%

4% 9%

4%

standard binder content

Cold mix; dense grading; low binder content

10% 2%

4.5% 0% 10%

4.5%

Cold mix; dense grading; standard binder content

6%

0%

Cold mix; open grading; standard binder content

We can use the same equations for predicting the stiffness modulus of a cold-mix as for hot-mix. A trapped water pocket has h as just the same effect as an air void. v oid. Equation:

Emixture [in MPa] = E binder × [1 + (257.5 – 2.5VMA)/(n × (VMA–3))]n

where:

n = 0.83 × log10[4 × 104/E binder ] VMA is in %; E binder is in MPa

Say, for example, S binder is 25MPa for hot-mix at 20°C but reduces to an average of 20MPa for cold-mix because of non-coated areas and the effect of the trapped water. Therefore n = 2.66 for hot-mix and 2.74 for cold-mix. Takin king the the volume umetric ex examples ab above: ve:

SDBM = 6090MPa Scold-mix low binder = 5350MPa Scold-mix std binder = 1790MPa

33


DEPARTMENT

OF


Fatigue and deformation resistance are both also likely to be poorer than for hot mix. Curing An important difference between cold-mix and hot-mix is the time taken to gain strength. HotHot-m mix: ix:

- fair fairly ly high high stre streng ngth th as soon soon as it has has cool cooled ed down down (e.g (e.g.. 2 hour hours) s);; - quite rapid rapid stiffe stiffening ning for a few weeks weeks as traff traffic ic slightly slightly reorie reorientates ntates particles for maximum effectiveness; - slow slow stiffe stiffenin ning g thereaf thereafter ter due due to binder binder agein ageing. g. Cold-m Cold-mix: ix: - initia initially lly lit little tle more more than than a gran granula ularr materi material; al; - binder binder effect effect clear clear wit within hin 24 hour hours; s; - continu continuing ing slow slow stre strengt ngth h gain gain over over a period of up to 6 months; - simu simula late te in in lab lab with with 5 days days at 40 40°C

6.0 Basalt

5.0

FoamMaster

Blast f urnace slag

Asphalt planings

RAP

Limestone

Average

4.0

3.0

2.0

a s e r n1.0 g 0.0 0

50

100

150

200

250

300

350

400

Days since Construction

Unfortunately, it is not really practical to keep traffic off the road for 6 months while the material stiffens up! Therefore there is a danger of early life damage taking place – but this is very hard to predict. So long as the trafic is light, no permanent damage will occur. The question is: just how much is too much?

Practical Use of Cold Mix Summary of key points: a) The binder content will probably be lower than for a hot mix. b) The void content is likely to be higher than an equivalent hot-mix. c) This means means that cold-mix cold-mix will will usually usually be less stiff, stiff, less resist resistant ant to deformatio deformation n and have lower fatigue life. It is a poorer material ! d) Water needs needs to evaporat evaporatee before sealin sealing g the surface. surface. This means means that that the aggregate grading used should be reasonably open. e) It is impo import rtan antt to to limit early trafficking . f) All of thes thesee points points mean mean that that there there is a significant risk associated with cold-mix. So why would anyone use cold-mix? a) The range range of possible possible aggregat aggregates es is extended. extended. For example, example, constructi construction on and demolition waste, incinerator ash, crushed concrete and recycled asphalt planings (RAP) can all be used. b) Cold-mix technology is ideal for in-situ recycling. c) Cold-m Cold-mix ix can can have a stor storage age life life of of severa severall months months..

34


DEPARTMENT

d) Lower Lower energy energy usage usage gives gives environ environmen mental tal benefi benefits. ts. For these reasons, cold-mix is a popular choice for minor roads .

35

OF



DEPARTMENT

OF


PAVEMENT DESIGN 1.

Traffic

Load Magnitude a) asphalt asphalt fatigue fatigue cracki cracking ng depends depends approxim approximately ately on the the 4th power of tensile strain – see graph on p25; b) concrete fatigue cracking depends approximately on the 10th to 12th power of tensile stress – see graph on p18; c) subgra subgrade de ruttin rutting g depends depends approxi approximat mately ely on the 4th to 7th power of subgrade compressive stress (or strain ?). So it’s all a bit complicated. The American Association of State Highway Officials (AASHO) undertook a series of pavement trials during the late 1950s using controlled trafficking with known loads. Conclusion: use a 4th power law. No. of equivalent design axles (Neq) = [axle load (P) / design axle load (Pdes)]4 i.e. if P is P is twice the design axle load P load P des da mage of a design axle. des, it will do 16 times the damage How much difference does the choice of power law really make? Typical highway traffic – 1 hour: Wt band 0-1T 1-2T 2-3T 3-4T 4-5T 5-6T 6-7T 7-8T 8-9T 9-10T 10-11T 11-12T

Average 0.5T 1.5T 2.5T 3.5T 4.5T 5.5T 6.5T 7.5T 8.5T 9.5T 10.5T 11.5T

Number 3875 1201 564 320 254 189 265 412 365 176 17 4

Convert to equivalent 8T axles: Neq = N × (Wav/8)n 5000

s e 4500 l x a 4000 T 8 3500 f o 3000 o 2500 n t 2000 n e 1500 l a v 1000 i u q 500 E 0

Wit hout 14-15T axle With 14-15T axle

0

2

4

6

8

10

12

Exponent n

So, for highways we convert traffic to an equivalent number of 8T (80kN) axles (called standard axles ) which means 40kN wheel loads . For airfields we have to choose a design aircraft and convert to numbers of equivalent design aircraft ; similarly for port pavements etc. But, for highways at least, we usually don’t know the detailed numbers in each weight band, so it is common to use a wear factor (or damage factor ) for each vehicle type.

36

14


OF


Wear Factors (= conversion factor to standard axles)

Vehicle Type

Hakim (1998b)

2 axle rigid 3 axle rigid 3 axle articulated 4 axle rigid 4 axle articulated 5 axle articulated 6 axle articulated

DEPARTMENT

1.16 0.39 1.75 0.84 2.02 1.78

Frith et al (1997)

0.40 1.26 0.65 2.80 1.00 2.50 1.69

UK Highways Agency (HD24)

Maintenance 0.40 2.30 1.70 3.00 1.70 2.90 3.70

New road 0.60 3.40 2.50 4.60 2.50 4.40 5.60

Collop (1999) Flexible Rigid Rutting Fatigue

1.16 2.32 1.79 2.85 2.71 3.70 3.94

1.46 2.39 1.63 3.12 2.26 3.94 3.03

0.68 1.29 0.68 2.12 1.10 2.65 1.48

Note: these keep changing as vehicle and tyre design changes.

Contact Pressure The contact zone between a pneumatic rubber tyre and the road surface is obviously complicated. But should we worry? By the time we go down a few centimetres the pressure will have evened out so the exact details certainly won’t affect the lower layers – probably not the base either. Result: don’t worry about it except for high pressure, usually aircraft loads, when we need to choose a suitably strong surface course. Typical values: 250kPa for a car, 700kPa for a large truck, 1000-1500kPa for commercial aircraft and up to 3000kPa for some military planes.

2.

Standard Pavement Designs

As an example, here is the main chart from UK Highways Agency standard HD26.

Exam le

Composite design – HBM lower base

Composite design – asphalt upper base, binder course and surface course

37

Fully flexible design (asphalt/ foundation)

U NIVERSITY NIVERSITY OF NOTTINGHAM Pavement Engineering - Module H23P01 Note:

DEPARTMENT

OF


Traffic: (msa = millions of standard axles) HBM (hydraulically bound material) category: A. 9-12MPa (gravel) B. 9-12MPa (c (crushed ro rock); 12 12-16MPa (g (gravel) C. 12-16MPa (crush ushed rock ock); 1616-20M 20MPa (gravel) D. 16-20MPa (crushed rock) Foundation class: 1. Capping only 2. Granular subbase 3. Weak HB HBM subbase 4. Strong HBM subbase

50MPa at top of foundation 100MPa at top of foundation 200MPa at at top of of foundation 400MPa at top of foundation

Asphalt materials (= base layer): • DBM125 = Asphalt concrete with 125pen binder • HRA50 = Hot rolled asphalt with 50pen binder • DBM50/HDM50 = Asphalt concrete with 50pen binder • EME2 = Asphalt concrete with 15pen binder Here is the equivalent chart for concrete (rigid) pavements

Note:

Traffic: (msa = millions of standard axles) CRCP = continuously reinforced concrete pavement CRCR = continuously reinforced concrete roadbase (i.e. needs an asphalt surface) f r r = concrete flexural strength R = reliability

38


DEPARTMENT

OF


The AASHTO (1993) method is also still widely used Key Equation:

Structural Number (SN) = a1h1 + a2h2 + a3h3

where:

a1, a2, a3 = layer coefficients [Asphalt a ≈ 0.238 × modulus(GPa)0.55] [Granular a ≈ modulus(MPa)0.85 / 470] h1, h2, h3 = layer thicknesses (inches)

Structural Number (SN) is a measure of the strength of the pavement structure, which is related to pavement life (in msa) through a complicated equation. Advantage : pretty simple conceptually

Nowadays the AASHTO Mechanistic Empirical Pavement Design Method (MEPDM) is also available. This takes several minutes to run on a modern PC but accounts for detailed traffic distribution + changes in temperature day and night throughout the year. The problem is that all these methods + many others across the world are basically b asically ‘black boxes’. You input some parameters and you get a design out – and you just have to trust that the method applies to your particular case. You calculate nothing!

3.

Analytical Pavement Design – Flexible Pavements

Design Principles The pavement has to fulfil the following roles: a) Protect the subgrade : Natural ground will will not be usually usually be strong enough to bear traffic load directly; it would deform and rut. b) Guard against deformation in the pavement layers : All pavement materials materials must themselves be stable enough not to deform too much. c) Guard against break-up of the pavement layers : The strength strength of the pavement layers must be sufficient to prevent excessive cracking from developing. d) Protect from environmental attack : The materials used used must not lose their properties (too much) under environmental attack. e) Provide a suitable surface : The design has to be suitable to provide an appropriate pavement surface. f) Ensure ‘maintainability’ : The design must ensure that that it is possible possible to carry out necessary maintenance. Analytical design usually only looks at (a) and (c). The other points are covered by sensible material specifications and sensible combinations of layers. Protect the subgrade Recognising the difficulties involved in soil parameter estimation, most of the current analytical design methods use the so-called the subgrade strain criterion

39


DEPARTMENT

OF


This is a massive simplification – and relies on a massive assumption , namely: THE STRENGTH OF A SOIL IS DIRECTLY RELATED TO ITS STIFFNESS This is not really true!! Nevertheless, over a limited range of soils, for example UK heavy clays, it may be close enough to being true to be usable. The argument goes like this:   

life (to a limiting rut depth) = fn [ stress ÷ strength ] if: strength = fn [stiffness modulus ] …. …. then: life = fn [stress /stiffness modulus ] = fn [elastic strain ].

The great advantage of this assumption is that it is only necessary to calculate the elastic strain value in the subgrade under load, i.e. the vertical elastic strain at the top of the subgrade under a design wheel load. This can be done using multi-layer linear elastic analysis – programs like BISAR, ELSYM, JULEA, CIRCLY etc.

Hot-Mix Asphalt Subbase Subgrade

All we need now is a relationship between life and strain. Here are three:

Nf [in millions] = 3.09 1010 εz –3.95 [εz in microstr microstrain] ain] – UK Transport Transport Research Research Laborator Laboratory y -3 –7.14 Nf [in millions] = 8.511 10 εz [εz in millist millistrain] rain] – NAASRA NAASRA (Austr (Australia alia)) 8 –3.7 Nf [in millions] = 7.6 10 εz [in mi microst ostrain ain] – Br British Ai Airports Au Author hority Conclusion: think before you believe any of them! They have all been developed based on experience. For example, sandy soils will tend to have different relationship between strength and stiffness modulus from that of clay soils, which means they can carry many times the number of load applications app lications for a given level of elastic strain. This is about as good as it gets unless you really know something about the soil in your subgrade. However if you know enough to estimate the shear stress at failure at the top of your subgrade you might be able to be a little cleverer in your design, based on data like that shown here.

100 8

Clay Sand

) 7 % ( e l c 6 y c t s r i f 5 e h t g 4 n i d u l c 3 x e n i a 2 r t S r a e 1 h S

90%:

90

88%

ratio of applied stress to to failure stress

99%

83%

71%

90%

52%

73%

0 10

100 Number of Load Cycles

40

) % ( s 80 s e r t s e 70 r u l i a f / s 60 s e r t s d 50 e i l p p A 40

2% strain

1% strain

possible design line 30

54% 1

1% strain 2% strain

1000

10000

1

100

10000

1000000

Numbe r of Load Cycles Cycles


DEPARTMENT

OF


This data is from laboratory tests carried out in the triaxial equipment on a soft clay and an angular sand. Note that when the data is in the form of a stress ratio, the behaviour of the two soils is fairly similar, allowing a possible design line to be proposed. Guard against break-up of the pavement layers It is a fact that a relationship is generally found between tensile strain in asphalt under load and the number of load applications until failure occurs; it is therefore quite logical that the maximum tensile strain in the asphalt layers of a pavement under a wheel load should be related to cracking. Cracking can occur: A – at the bottom of the asphalt immediately under the load; B – near the surface just outside the loaded area; C – at the surface in the tyre tread contact zone.

Hot-Mix Asphalt Subbase Subgrade

A is generally assumed to be dominant. Even if it isn’t always true it makes life simpler! It can be calculated relatively easily using multi-layer linear elastic analysis .

We now need a relationship between calculated tensile strain and life. Here are two: Nf [in millions] millions] = 4.17 10-10 (1/εt)4.16 [εt in microstrain] – UK Transport Research Laboratory Nf [in millions] = 0.00432k 1'C (1/εt)3.9492 (1/E)1.281 [εt in microstrain] – AASHTO, US where: k 1' = ' = fn(h fn(h); C = C = fn(mixture volumetrics); E volumetrics); E = = asphalt stiffness modulus; h = asphalt thickness. Putting it all together:

Carry out traffic assessment

Multi-layer linear elastic analysis

εt

Buses 56 2-axle trucks 562 3-axle trucks 401 5 or 6-axle 6-axle artic articulate ulated d trucks trucks (with (with semi-tra semi-trailer iler)) 268 Equivalent standard wheel wheel loads/day 2633 20-year design traffic 2 107

Select a fatigue characteristic

h1

E1, ν1

h2

E2, ν2 E3, ν3

h1

µstrain 2000 1000 800 200 100 50 20 10

εt = 50µstrain εt = 70µstrain εt = 100µstrain εt = 150µstrain 1

10

102

103

104

105

106

107

108 N

41

h2


DEPARTMENT

OF


Design Temperature This is all OK, but we will get a different answer depending on the temperature of the pavement since asphalt stiffness modulus depends on temperature. Its stiffness stiffness changes by a factor of about three for every 10°C temperature change, and stiffness is a key input to any calculation of tensile strain. The problem is far too complex to analyse fully. So, select a design temperature and calibrate your prediction based on experience. In the UK, the temperature selected is 20 C. The equations for Nf on the previous two pages assume an asphalt temperature of 20°C. Typical Stiffness Moduli and Poisson’s Ratios Modulus -

Asphalt surface course: Asphalt concrete, 125 pen (= DBM125): Asph Asphal altt con concr cret ete, e, 50 pen pen (= (= DBM DBM50 50//HDM5 HDM50) 0):: Asphalt concrete, 15 pen (= DBM15): Weak HBM: Crushed rock base: Crushed rock subbase: Capping:

Poisson’s ratio -

asphalt and unbound material: H BM

2500MPa 3100MPa 5000 5000MP MPaa 7800MPa 500MPa 250-500MPa 150MPa 75MPa 0 .3 5 0. 2

Issues in Real Design 800 Thin asphalt layers 000000 f 1000000 M odelling propagation The problem here is that ) 700 o n r ) F or practical desig design n i e a deformation (and therefore Computed b s r e t m s s 600 100000 u s curvature) of pavements with o n a r ( c 500 p i thin asphalt layers is dominated e f i m ( L 10000 by the stiffness of underlying n 400 i 0 50 100 150 200 a r support, which means that it Asphalt Thickness Thickness (mm ) t 300 S only increases slightly as e l i 200 s asphalt thickness reduces; and n Computed e T 100 strain, which is proportional to For practical design 0 curvature but inversely 0 50 100 150 200 250 proportional to thickness, Asphalt Thickness (mm) actually reduces at low thickness. The trouble is that though calculated strain may reduce as asphalt gets very thin, experience is that pavement life does not start increasing!

Suggestion: just extrapolate the design curve derived at greater thicknesses.

42


DEPARTMENT

OF


Load Groups You don’t often get a truly isolated wheel, in which case you might have to take account of effects from neighbouring wheels.

In general: don’t worry about it for asphalt strain ; take it into account for subgrade strain . Dynamic Effects Wheel load fluctuates as the vehicle body oscillates vertically. Some suspension systems are more effective than others at avoiding high dynamic load, but you can’t avoid it entirely. Usually we ignore it in design , assuming it is included in the calibration. But let’s take a look at what might really be happening. Maximum dynamic Mean

Load

Divide into time steps Work out relative damage in each time step [ = (W/Wmean)n ]

20 e 18 g a 16 m a 14 d 12 n o 10 r e i l 8 p i t 6 l u 4 M 2 0

Exponent = 4 Exponent = 8 Exponent = 12 Exponent = 4; repeatable load pattern pattern

1

Average damage over whole cycle

1.2 1.4 Ratio: maximun dynamic/mean

1.6

The first three curves make the assumption that the distribution of dynamic load is random. However, there is plenty of evidence to show that this is often not the case and that similar vehicles, having similar suspension system characteristics, will tend to apply peak loads in roughly the same locations. The fourth curve makes the assumption that the loading pattern is perfectly repeated, in which case the computed damage would occur at regular intervals along the pavement – and you can sometimes see this in reality. Cornering

M

r v

Centre of gravity h P1

P2 d

Sideways force due to cornering =

Mv2/r

Vertical force due to gravity

Mg

=

[g = 9.81m/s2]

Balance vertically:

P1 + P2 = Mg

Moments about inner wheel path:

P1d = (Mv2/r)h + Mg(d/2)

Com Combine bine:: P1 = M (g/2 (g/2 + v2h/rd) P2 = M (g/2 (g/2 – v2h/rd) Example: v = 25m/s (approx 60mph); r = 500m; d = 2m; h = 2m; M = 8T  P1 = 49.24kN; P2 = 29.24kN

43


DEPARTMENT

OF


Vehicle Speed Effects Vehicle speed is important for three reasons:  dynamic loads will be higher at high speeds;  asphalt stiffness varies with loading rate;  soils and granular materials at high levels of saturation may suffer from positive pore pressures (and therefore low strength and stiffness) at high loading rates.

But we still usually ignore it ! Lateral Wander This refers to the fact that not every wheel follows the same path. On highways, the distribution across a wheel path commonly has a standard deviation of around 150mm; on airport runways this is likely to be a metre or more. We could therefore reduce the design traffic slightly to account for this. On roads this is rarely done; on airfields it is. Designing with Cold-Mix Asphalt The trouble with cold mix is that it is neither one thing nor the other. It is partially bound.

Choices: a)

trea treatt it it as as a hothot-m mix Hot-mix binder course + surface course ∼ 3500MPa Cold-mix base ∼ 2500MPa Calculate asphalt tensile strain at base of cold-mix layer; use a standard fatigue characteristic. characteristic. Assumption: the material achieves full curing without damage [optimistic assumption]

b)

treat it as a very superior granular material

Hot-mix binder course + surface course ∼ 3500MPa Cold-mix base ∼ 1000MPa Calculate asphalt tensile strain at base of binder course; use a standard fatigue characteristic. characteristic. Assumption: the material deteriorates to the equivalent of an excellent granular material.

What stiffness to use?

500MPa if you are really quite unsure; 750MPa for most cases; 1000MPa for well controlled construction in UK; 1500MPa under ideal conditions.

44


DEPARTMENT

OF


Anal Analyt ytic ical al Pave Paveme ment nt Desi Design gn – Com Comp posit ositee Pav Pavem emeents nts

The first issue here is: what is a composite pavement? Answer: one with a relatively strong HBM layer in it. A weak HBM is really just like a good granular material because it will end up broken into small pieces So, reviewing the design principles listed on p38, the key differences are: a) Subgrade protection becomes of secondary importance since the subgrade can never rut until the HBM has broken. b) We need to make sure the stress in the HBM isn’t enough to break it, or at least not enough to break it too much. Thermal Stress All HBMs expand and contract with temperature changes. They are solids and so this imposes stresses, mainly in the longitudinal direction. In many climates there is likely to be at least a 20°C difference in HBM temperature between setting and the coldest time of year, leading to a typical strain of 2 × 10-4, and this is around the failure strain for most HBMs. The HBM will therefore crack at intervals transversely across the pavement. In fact it will usually crack during the first few night of its life when the strength is still low. What will the crack spacing be? Angle of friction

Temperature << temperature at time of set

L

HBM layer Crack

h

Shear stress τ =

∴

ρgh × tanφ

Force across centre of slab = τL/2 =

ρghL/2 × tanφ (per metre width of slab)

If slab is intact:  Tensile stress stress at slab centre ≈

ρghL/2 × tanφ ÷ h = ρgL/2 × tanφ

If this is more than the tensile strength of the HBM, it will break in two Example : say tensile strength = 0.2MPa when the first relatively cold night occurs, density ρ = 2300kg/m3 and friction angle φ is 35°.

Stress = strength when:

0.2 106 = 2300 × 9.81 × L/2 × tan(35°) i.e. L = 38m

So a 40m length between cracks will break into two 20m lengths, but a 35m length will remain intact. In fact eventual spacing should be between 19m and 38m.

45


DEPARTMENT

OF


Conclusion: expect transverse cracks, but the spacing will depend critically on day-night temperature changes during the first few days and weeks of life, and these cannot reasonably be foreseen. Solution: crack deliberately (i.e. form joints) at much closer spacing, usually 3m.

Recently paved HBM

Form slots to 2/3 depth

Roll HBM

Fill slots with bitumen emulsion

This process creates weaknesses. Hopefully cracks from at eac h weakness but because they are quite close together, each crack should remain very narrow (a hairline crack). Design against Traffic Loading This time we need to calculate the tensile stress at the bottom of the HBM layer, again using multi-layer linear elastic analysis . We then need a relationship between calculated tensile stress and life. Going back to p18, the key quantity is the ratio of tensile stress to tensile strength – well actually flexural strength, i.e. strength from a realistic test arrangement. We can just apply the fatigue equation suggested on p18, namely:

Hot-Mix Asphalt HBM Subbase Subgrade

σt / flexural strength = 1.064 – 0.064 log10 (N) So, if the design is for 50 million standard axles and the calculated tensile stress at the bottom of the layer is 0.8MPa, then the required long-term flexural strength is 1.4MPa, which equates to a compressive strength of 10-15MPa. If we want to use a weaker material we must make the layer thicker, or maybe make the foundation found ation stronger. Reflective Cracking Even if there are no traffic-induced cracks in the HBM, we know that there will be thermally induced transverse cracks (or joints – which amounts to the same thing). These represent discontinuities in the support given to the asphalt layers, which means we are very likely to find a crack appearing through the asphalt at those points. This is reflective cracking . a) Reflective Reflective cracks cracks are a nuisanc nuisancee not a real failure failure;; could just just keep re-seal re-sealing ing them. them. b) Could use Highways Agency rule and always have at least 180mm of asphalt. c) Could Could check check asphal asphaltt tensi tensile le str strain ain εt – but if so then reduce EHBM to 500MPa. This represents the effect of a discontinuity in support to the asphalt.

46


DEPARTMENT

OF


Analytical Pavement Design – Rigid Pavements (jointed unreinforced)

Westergaard Analysis Since the 1920s, the equations developed by H.M. Westergaard have represented the most widely used approach to analysis of concrete pavements under load. The original equations were derived assuming a concrete pavement to act as a slab in pure bending, and several subsequent modifications have been made over the years to increase the accuracy with which a real pavement can be modelled. The equations are in terms of the maximum tensile stress in the concrete due to slab bending and they are for three load locations: i) internal (i.e. distant from a joint); ii) edge; and iii) corner . The load is assumed to consist of a uniformly stressed circular area. Here are commonly applied versions of the equations.

Joints (with zero load transfer) Corner Internal Edge

Plan View

Westergaard equations for stress in a concrete slab:

Internal loading; stress at base of slab: σTensile = [0.275 p / h2].[1 + ν].[4 log10(Ls/b) + log10(12 (1- ν ν2)) – 0.436] Edge loading; stress at base of slab: σTensile = [0.529 p / h2].[1 + 0.54 ν].[4 log10(Ls/b) + 0.359] Corner loading; stress at top of slab: σTensile = [3 p / h2].[1 – (√2 a / Ls)1.2] where: p = load; a = radius; h = slab thickness; ν = Poisson’s ratio; Ls = ‘radius of relative stiffness’ = [E h 3 / (12 k (1- ν2))]0.25; E = stiffness modulus; k = ‘modulus of subgrade reaction’; b = ‘radius of equivalent pressure distribution’ = √(1.6 a2 + h2) – 0.675 h, if a > 0.72 h; = a, if a < 0.72 h.

Estimating modulus of subgrade reaction k: P r

Use multi-layer linear elastic analysis: E1,

1



predict deflection δ

E2,

2



k = (P/πr 2)/δ

E3,

3

Note: the result will depend on the value of r of r .

47


DEPARTMENT

OF


Once you have calculated a worst-case tensile stress you could then apply the fatigue equation suggested above for HBM. However, it is usual to be on the safe side and use a more conservative equation, known as the Packard line.

σt / flexural strength = 0.96 – 0.0799 log10 (N) So, if you calculate a worst case stress of 1.6MPa for example, and you want to use a concrete with a flexural strength of 4.5MPa then the design traffic is about 37 million load applications. Problem : real life isn’t just corners, edges and places far from an edge. Joints usually transfer load, so they aren’t really edges. But then not all joints are the same – look back to p7. Expansion joints are not far from being edges; contraction joints should have pretty good load transfer; warping joints should have excellent load transfer. Solution : Usually ignore the corner case. Often ignore the edge case. Use the internal case but apply a factor depending on how good you think the joints are; e.g. × 1.2 for good joints, × 1.5 for poor joints.

Multi-layer Linear Elastic Analysis The same multi-layer linear elastic analysis as has been introduced for flexible pavements can also be used here to calculate tensile stress. Advanta Advantage: ge:

load load combin combinati ations ons can can be includ included, ed, such such as dual dual or tand tandem em wheel wheel sets. sets.

Disadvantage: all layers have to be infinite infinite in extent; there is no way of analysing an edge or corner situation. Limit State Analysis A problem with both Westergaard and multi-layer linear elastic analysis is that concrete cannot really crack at a single point. If cracking is to occur, then there must be a mechanism of cracks. What will actually happen is that the point of theoretical failure will simply reduce in stiffness locally as the first inter-particle fractures begin to occur. Conclusions based on an elastic analysis will therefore be conservative . The alternative is a limit state analysis . Key equation:

Need to:

WORK done by loads = ENERGY absorbed by foundation + ENERGY dissipated at cracks

a) arrange loads on pavement; b) propose a failure mechanism; c) calculate work done + energy to foundation foundation based on an assumed set of deflections and angles; d) derive the resisting bending moment moment (per metre) at crack locations; locations; e) relate this bending moment moment to a required slab thickness for a given concrete strength using the equation M = σh2/6 (refer back to p17).

48


DEPARTMENT

OF


The limit state approach opens the door to solving problems of multiple loads and complex joint geometry that would otherwise be impossible. Example:

d

Load P

Assump Assumpti tions ons:: a) b)

load load unifor uniformly mly distri distribut buted ed over over area area of of founda foundati tion on enclosed by cracks square wheel pattern with 0.5m offsets everywhere

Work done by loads = 4Pδ × (√2d/2–0.5×√2)/(√2d/2) = 4Pδ (1–1/d) 0.5m

0.5m

Moment to cause cracking = M per linear metre

Energy dissipated at cracks = 4√2dM × (δ /(√2d/2)) = 8Mδ Energy absorbed by foundation = 4P × (δ /3) Energy balance: 4Pδ (1–1/d) = 8Mδ + 4Pδ /3 

M ≈ P/3 for large d

Since M also equals σf h2/6, therefore:

σf = 2P/h2

Warping Stresses Thermally-induced warping stresses result from a temperature difference between the top and the bottom of a slab; they should not be ignored. Approach 1: assume that the safety margin in using the Packard line (previous page) is enough to cover warping stresses – the usual approach in practice. Approach 2: use equations that were developed to predict warping stresses directly. These are the Bradbury equations for maximum warping stress in a concrete slab, one for internal stress and the other for edge stress, mirroring two of the conditions covered by the Westergaard equations. Internal loading ; stress at base of slab: σTensile = ½.E.α.T × (Cx + ν Cy)/(1 - ν2) Edge loading ; stress at base of slab: σTensile = ½.E.α.T × Cx (or Cy)

where: E = stiffness modulus; α = coefficient of thermal expansion; T = temperature difference top-bottom; Cx,Cy depend on the ratio of slab dimensions x and y respectively to radius of relative stiffness Ls – see inset.

x/Ls=0, Cx≈0; x/Ls=2, Cx≈0.05; x/Ls=3, Cx≈0.18; x/Ls=4, Cx≈0.50; x/Ls=5, Cx≈0.73; x/Ls=6, Cx≈0.89;

You can then add ad d the maximum warping stress to the stress caused b y traffic to give a real maximum value for design – although you then have to make some difficult decisions about the number of likely combined stress applications during the life of the pavement.

49


DEPARTMENT

OF


Joint Spacing The main factor here is the amount of thermal expansion and contraction – not warping. Taking a typical coefficient of thermal expansion of 10-5 per °C and a 30°C difference between maximum summer and minimum winter pavement temperatures, then a nominally 10m long slab of concrete c oncrete will vary in length by 3mm. So, if joints are placed in a road at 10m spacing, there will be a gap of 3mm or more in the winter. This is usually too much. Even with dowel bars (refer back to p 7) there won’t be enough load transfer across joints in the winter because there won’t be any aggregate interlock, which means the design case becomes almost an edge condition. In fact, experience suggests that for most climates c limates joint spacings between 3.5 and 6m represent a reasonable compromise between the cost and nuisance of joint construction and maintenance and the need to maintain load transfer efficiency. Less and the cost becomes too great; more and joint problems become increasingly likely.

6.

Reinforced Concrete Pavements

Lightly Reinforced Reinforcement is often not economically justified. However a light reinforcement mesh is sometimes included near the top of the slab as a means of controlling shrinkage cracking. It can also be used to cut back on the number of joints. The argument goes like this: 1) 2) 3) 4)

joints joints are a nuisance nuisance so so let’s let’s have have less of them; them; great greater er joi joint nt spac spacin ing g  increased warping stresses  less load transfer at joints; therefore therefore there there is a much greater greater likeli likelihood hood of cracking cracking;; but if reinforc reinforcement ement is present present cracks cracks are ‘controll ‘controlled’; ed’; they will will remain remain narrow and the slab will not break up.

In this sort of pavement hairline cracking is accepted; but there is enough reinforcement to hold the slab together, giving plenty of aggregate interlock across each crack and so plenty of load transfer. Continuously Reinforced But why have any joints at all? After all, if we can accept hairline cracking then what putting so much reinforcement in that it never fractures? This is continuously reinforced concrete (CRC), and it sits right at the top of the range of concrete pavement options. The principle is simple. There has to be enough reinforcement to resist the forces generated when the concrete contracts due to cooling. Note: a) b) c) d)

the concrete concrete will will still still crack – but these these will will be hairline hairline cracks, cracks, typically typically every every 1m; reinforcement quantity will typically be 0.6-0.8% of the concrete area; slab thicknes thicknesss is usually usually less less than in the the unreinforce unreinforced d case – but not by much; much; ther theree need needss to to be an anchorage (into the ground) at each end.

50


DEPARTMENT

OF


SURFACE PROPERTIES 1.

Ride Quality

Excellent ride quality is not always needed – but it is on high-speed roads. Typical tolerance limits: ± 3mm in 3m So how can this be achieved? Pavement Quality Concrete No problem achieving the tolerance in a machine-laid wet-formed concrete pavement. Probl Problem em::

concr concret etee is is har hard d and and unabl unablee to to abs absor orb b much much energ energy y fro from m the the tyre tyres. s.  high tyre vibration  relatively high noise  not so pleasant to drive on [joints just make things worse; good texture, e.g. longitudinal grooving or exposed aggregate finish , can help]

Asphalt To get a really good go od finish, the surface course must be relatively thin (say ≤ 50mm); otherwise the paver operator will not be able to control levels well enough. But then the underlying layer must also be reasonably even too – and this principle applies right the way through the pavement. The evenness of the surface of each layer can be constructed slightly better than that of the one below, but only slightly. UK Highways Agency tolerances (absolute maxima) at each level: Pavement surface Binder course Base Subbase

± 6mm ± 6mm ± 15mm + 10mm – 30mm

Impact of different types of surface: Surface Type

Asphalt concrete Hot rolled asphalt + chippings Stone mastic asphalt Porous asphalt Surface dressing Concrete (PQC) Block paving

Vibration generation medium medium low low high fairly high very high

51

Energy absorption medium low high very high low very low low

Ride quality ranking 3 4 2 1 5= 5= 7


DEPARTMENT

OF


Material Strength, Durability etc

A PQC surface is of exactly the same strength as the rest of the concrete slab  not an issue. Asphalt Surfaces Asphalt surface course has to have relatively small stone size due to its low layer thickness and, consequently, a relatively high bitumen content . This leads to a less stiff material but with high fatigue resistance and good durability . Asphalt surface structural properties: Mixture Type

Stiffness

Asphalt concrete Hot rolled asphalt (+ chippings) Stone mastic asphalt Porous asphalt BASE

medium medium medium-low low HIGH

Deformation resistance high low high medium-high HIGH

Fatigue strength medium high medium-high medium-low MEDIUM

Block Paving Although blocks themselves are high-stiffness, the effective layer stiffness is a function of rotation and shear at joints, which depends on how well the joints are filled. It is common practice to assume a stiffness of 500MPa for a combined block-bedding sand layer.

3.

Skid Resistance

Microtexture This term describes the intrinsic frictional properties of the surface. In an an as aspha phalt: In a PQC: In bloc block k pav paviing: ng:

it re relates to to th the ag aggregate pa particles at at th the su surface. it relates to the cementitious mortar. it relat elates es to the the surf urface ace of of the the bloc block. k.

The microtexture represents the ultimate skid resistance potential of a surface, the level applying in dry conditions and without any intervening dirt, bitumen or ice lens. It is logical therefore to insist on improved microtexture at sensitive locations such as approaches to pedestrian crossings and roundabouts, and this approach is adopted by highway authorities all over the world. Certain aggregate types such as gritstones therefore take on a premium value because of their excellent microtexture. The problem is that the frictional properties of a surface change under the action of traffic. In dry weather they are polished by the relative motion of tyre and surface,

52


DEPARTMENT

OF


activated partly by tyre vibration. This reduces the microtexture. It is standard practice therefore to assess the so-called Polished Stone Value (PSV) of an aggregate by first subjecting it to an accelerated polishing regime before measuring the frictional properties using the pendulum test .

Rubber pad

Rubber-tyred wheel

Pendulum Test

Surface aggregate Polishing Test

Microtexture is also seasonal . Polishing occurs mainly in dry weather; wet weather Specimen Accelerated Polishing Polishing Machine restores frictional properties to some extent due to the ab rading effect of small particles of grit which are present in surface water. For this reason, skid resistance should preferably be assessed in summer or during the dry season. Macrotexture If it never rained you would need no macrotexture (the visible texture due to the arrangement of stones or the presence of grooves etc). Neither tyre tread nor visible surface texture make the smallest contribution to basic skid resistance; they are only present to ensure that surface water has somewhere to go. Direct contact is needed between tyre and surface in order for friction to be activated; if a water film remains in between, the vehicle will aquaplane as soon as brakes are applied. app lied. An optimised macrotexture therefore ensures that there is only a short distance between individual contact points and regions where water can be accommodated without danger.

53 Water movement away from contact points


DEPARTMENT

1. Measure out exact volume of sand


Macrotexture is generally expressed as a texture depth in millimetres. The basic measure comes from a procedure known as the sand patch test , although there are also laser based pieces of equipment on the market for rapid, sometimes trafficspeed, measurement.

2. Pour onto pavement surface

3. Spread out level with tops of aggregate particles

OF

4. Record diameter of sand patch

Sand Patch Test

Macrotexture can also deteriorate under traffic loading. The same wet-season abrasion that restores microtexture also reduces the height of individual aggregate particles, eventually reducing the texture depth excessively. For this reason, it is necessary to specify abrasion resistance, for example the Los Angeles Abrasion value.

4.

Spray

Spray from surface water is a safety hazard . If water cannot easily flow across the surface of a pavement then it will be available to form spray. The issue is not texture depth but barriers to lateral flow . Traditional UK Hot Rolled Asphalt (HRA) with rolled-in chippings has a particularly bad reputation for spray since each individual chipping sits in its own small indentation (negative texture) into the asphalt surface, allowing a small ‘pond’ of water to remain around it until it either evaporates or is dispersed in the form of spray. Most other surfaces consist of protrusions (positive texture) from a more general surface level and water can flow around these protrusions and make its way sideways. Asphalt concrete and SMA therefore generate much less spray than HRA. Grooved concrete is also good. However, porous asphalt is undoubtedly the premier material. Porous asphalt allows water to drain straight into the pavement itself and then to pass laterally through it, below the level of the tyre-surface contact. The result: virtually no spray at all. Of course, the pavement has to be able to cope with the presence of water within the porous asphalt. Usually the porous asphalt surface course has to overlie a dense, impermeable binder course; otherwise pavement durability problems are likely.

54


DEPARTMENT

OF


Noise

This can be an important issue in urban areas. It is a highly complex field and it is not necessary for the pavement engineer to appreciate the exact acoustic mechanisms involved. 108 106

Blocks

104

Porous Porous as phalt

Other materials

102

) A 100 ( B d e 98 s i o 96 N 94

Stone Stone mas tic asphalt

SMA

As As phal t concrete

Surface dressing

Slurry se al Expos Expos ed aggreg ate concrete

Porous asphalt

92

Surface Surface dress ing

Textured Textured con crete

90

Blocks

88 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Texture de pth (mm) (mm)

As expected, noise level generally depends on texture depth , i.e. roughness. However, the picture is clearly more complicated than this , with a 10dB(A) difference – a factor of about 3 in actual sound pressure magnitude – between block paving and porous asphalt for the same texture depth. Of the more common asphalt surfaces, SMA is evidently the quietest at normal texture depths (around 1mm). 1 mm). Conceptually: 

Noise is caused by vibration, principally of the tyre tread elements.  Surface type affects both the amplitude and frequency of tyre tread vibration. [a rough surface will induce a high amplitude of tyre vibration and therefore high no ise] Some of this noise will be absorbed by b y the surface, and this will depend on the  hardness of the surface material. [concrete has poor ability to absorb any sort of vibration energy including noise and this means that it is difficult to produce a low-noise con crete surface] Porous asphalt has very low stiffness and therefore causes little excitation to the tyre tread elements; it also has excellent noise absorption properties – an ideal low-noise material.

55


DEPARTMENT

OF


PAVEMENT EVALUATION 1.

Visual Condition Surveys

A visual survey is the most basic and yet often the most useful survey type of o f all. Put simply, not all cracks are the same; nor are all ruts or surface defects. An e xperienced engineer can deduce a great deal about the internal health or otherwise of a pavement just by inspecting the surface. Cracking in asphalt pavements a) Is there more cracking in the wheel path than elsewhere? Yes: traffic is responsible, whatever that cracking may look like. No: traffic is irrelevant and the cause is environmental or due to a general material defect.

b) Are there transverse cracks right across the pavement? Yes: either low-temperature cracks or reflective cracking from an HBM or PQC base.

c) Are there more transverse cracks in the wheel paths? Yes: either traffic-induced reflective cracking or defects built in during construction.

d) Is there a single well-developed longitudinal crack in the wheel path? Yes: traffic-induced fatigue of a thick flexible/composite pavement (cracks usually top-down).

e) Is there multiple cracking (crazing) in the wheel-path? Yes: shallow failure; either either thin asphalt or the the upper layer has become become debonded.

f) Is slurry pumping up to the surface through cracks? Yes: water has become trapped, either in the bound materials or else in the foundation.

g) Are there localised wheel path depressions d epressions where more than one crack is present? localised damage  water ingress, loss of support, settlement. Yes: probably a HBM base; localised

Cracking in PQC Concrete Pavements a) Is cracking (of a jointed pavement) largely restricted to transverse cracks? Yes: thermally-induced, assisted by traffic; initial joint spacing was excessive.

b) Are significant longitudinal cracks present in or around the wheel path? Yes: traffic-induced damage; cracks will propagate rapidly along the pavement.

c) Are longitudinal cracks narrow, relatively close-spaced and straight? Yes: lightly reinforced concrete; minor defects at the time of construction.

d) Are there corner cracks at joint intersections? Yes: lack of slab support close to joints; damage is limited and will extend no further.

e) Are there regular transverse cracks at 1-2m spacing but no joints? Yes: continuously reinforced.concrete; reinforced.concrete; should be hairline; if wide then pavement is too weak.

Chipping Loss: loss of adhesive properties in the b inder due to bitumen ageing. Ravelling: widespread chipping loss, leading to the development of pot-holes. Bleeding: excess bitumen in the pavement  shiny surface  significant safety hazard.

56


DEPARTMENT

OF


Rutting

Surface course problem

2.

Binder course - base problem

Foundation problem

Profile Surveys

International Roughness Index (IRI) IRI is defined by the amplitude of motion of a vehicle suspension system as it travels along the road, measured in cumulative metres of suspension system movement per kilometre of travel (m/km or mm/m). The vehicles that measure IRI are known as bump integrators . +ve

Suspension movement

–ve

Typical Data 7 6

IRI < 2 m/km 2-3 m/km 3-4 m/km 4-5 m/km > 5 m/km

excellent satisfactory moderately bumpy bumpy very bumpy

) 5 m k 4 / m ( I 3 R I

2 1 0 0

1

2 Distance (km)

3

4

Laser Profile Surveys Laser-based systems are now very commonly used, usually with an array of lasers pointing down at the road surface. Reflective waves are monitored. They can be used for profile measurement , texture depth and rut depth . These types of survey can be carried out at normal traffic speed.

57


DEPARTMENT

OF


Skid Resistance Surveys

Skid resistance is a property that is directly related to the safety of users, applying to both highways and airfield runways. In the case of airfield runways there are international standards and it is necessary for airport authorities to check skid resistance regularly, particularly under adverse weather conditions (rain or snow). Highways are governed by standards set by individual countries, regions and cities. The Sideways Force Coefficient Routine Investigation Machine (SCRIM) Water tank

Trailer-mounted alternatives

Sideways force

Separation force

20°

Drag force

Plan Views Plan View

4.

Wheel under braking

Cores and Trial Pits

Surface course

Bottom-up reflective crack beginning to grow

Binder/Base course debonding

Crack in HBM base possibly reflected from subbase

HBM base

Crack; reflected from HBM base through through asphalt surfacing

Crack; top-down, penetrating through about 50% of the asphalt

Serious crack in disintegrating HBM sub-base

58


DEPARTMENT

OF


Coring, using cutters of 100mm or 150mm in diameter, is a relatively non-destructive method of sampling. Trial pits represent an alternative, labour-intensive method of sampling. They are suitable i) where bound layer thickness is low; ii) where samples of unbound foundation material are required; or iii) where specific information is needed which demands a larger area than can be afforded by a core.

Construction Information Cores Cores and and tri trial al pit pitss rev revea eall the the fol follo lowi wing ng::

Mate Materi rial alss pre prese sent nt;; Layer thicknesses; Visual quality; Inter-layer bond.

Samples can also be taken back to the laboratory for testing. In-situ Tests The relatively small size of most core holes means that there is a limit to the types of test that can be carried out. In fact, there is only one in-situ test device that is commonly used and that is the Dynamic Cone Penetrometer (DCP – see p14). Trial pits also allow the portable Dynamic Plate Test (DPT – also p14). Laboratory Tests These tests include:  compressive strength of HBM (height-diameter ratio of at least 1.0 required);  uniaxial stiffness modulus of HBM;  indirect tensile strength (ITS), either of HBM or asphalt;  indirect tensile stiffness modulus (ITSM) of asphalt;  indirect tensile fatigue test (ITFT) for asphalt;  repeated load axial test (RLAT) for asphalt deformation;  inter-layer bond strength tests . P Also, density and void content can be obtained on specimens of any convenient shape. Asphalt specimens can also be broken down into their constituents, by means of a centrifuge, with solvents used to extract the bitumen. Aggregate gradation and binder content can be checked. Binder quality can also be measured using a Dynamic Shear Rheometer (DSR ). ).

Torque Test

Leutner Test P

P

Peak shear for ce ce

Shear slip at failure

59


DEPARTMENT

OF


Ground Penetrating Radar (GPR) Transmitter / receiver

Indi Indiv vidu idual Rec Recei eive ved d Sign Signal al

Longi ongittudin dinal Si Signal gnal Pro Profi file le

Surface reflection Asphalt/HBM reflection HBM/Granular reflection

Radar wave pulses: Frequency 0.2-1.5MHz

Time

Interpreted Thickness Profile Asphalt Hydraulically-bound material Granular material

Depth

Thickness determination is the main reason for doing a radar survey and pretty good data is usually obtained, accurate to around 1cm.However, 1cm. However, it’s all down to image recognition software, so mistakes are possible. Radar can also give data on moisture (water molecules become excited at radar frequencies), voids (because of the strength of a solid-air interface), and steel reinforcement (steel interferes with wave propagation).

6.

Deflection Surveys

The Benkelman Beam This is the oldest and simplest form of deflection test device and it is successfully used throughout the world. 6350kg axle load Twin tyres

Benkelman Beam Pivot Arm

Distance travelled, x travelled, x

Dial Gauge Measurement, y Measurement, y

x Deflection

Adjust for temperature Approximate construction

Life in years [using an empiricallydetermined set of equations]

60

Annual Traffic

Deflection δ due to wheel load

y


DEPARTMENT

OF


The Benkelman Beam is a simple frame with an arm on a hinge, the rotation of which is read from a dial gauge. The equipment is placed on the ground immediately behind the twin rear tyres on one side of a goods vehicle loaded to a standard weight, the arm resting on the pavement surface between the twin tyres. When the operator is ready, the goods vehicle is slowly driven forward and a maximum reading is taken as the tyres pass pa ss the end of the arm; when it has driven forward some metres, a minimum reading is also taken. The difference relates to the deflection caused by the loaded wheel. The Lacroix Deflectograph The Deflectograph is an extension of the Benkelman Beam idea, initially developed in France. However it allows the vehicle to travel continuously along the road. The reference frame is in front of the measurement axle, and it is repeatedly dragged forward relative to the body of the vehicle and then released. As soon as the rear tyres of the vehicle have drawn level with the tip of the measurement arm, the frame is winched forward toward the front of the vehicle ready for the next reading. The result is that a measurement is taken every 3-4m and that the vehicle can travel continuously at a speed of 2-3km/hr. Readings are taken in both wheel paths.

Twin tyres 6350kg axle load

Winching mechanism Reference frame

Distance travelled, x travelled, x Arm

Adjust to equivalent Benkelman Beam deflection

Deflection

x Adjust for temperature Approximate construction

Life in years [using an empiricallydetermined set of equations]

Annual Traffic

Deflection δ due to wheel load Deflection, y

The problem with both the Benkelman Beam and the Deflectograph is that they are relatively low-resolution measurements and rely on empirical interpretation . They also do not give good data on PQC pavements . They are fine for estimating structural condition of asphalt pavements for network-level management but they are not really reliable enough for project-level design. For this we need something a bit more sophisticated. The Falling Weight Deflectometer (FWD) The FWD gives a very precise value of absolute deflection (accuracies of ± 2 microns commonly quoted), and that opens the door to a much more sophisticated method of interpretation.

61


DEPARTMENT

FWD – Longitudinal Section

Tow bar

Offset (m)

2

1

Measured load

0


Load: 10-200kN Duration: 25-50msecs Platen radius: 150mm

Falling Weight Rubber buffers Loading Plate

Deflection sensors

OF

Calculate deflections using multi-layer linear elastic analysis

Known layer thicknesses Layer stiffnesses

Deflection Bowl

Adjust layer layer stiffnesses

Deflection (microns)

no

Good match to measured deflections?

yes Finish

The machine is usually trailer-mounted. Tests are performed with the equipment stationary and with the loading plate and deflection sensors lowered onto the surface. The load pulse is then generated by the action of a falling weight onto a set of rubber buffers. Key advantages:  the load magnitude can be selected to match a typical wheel load;  the pulse duration is similar to that from a moving vehicle;  the deflections are absolute and highly accurate (using velocity transducers);  measurements are taken not only at the load location (through a hole in the centre of the loading plate) but also at selected distances from it. The full set of readings describes a deflection bowl (or basin) which can be backanalysed (or back-calculated) to deduce the combination of layer stiffnesses present.

d1 d2 d3 d4 d5 d6

Load

Upper Pavement - affec affects ts curv curvat atur uree of of cen centr tral al part part –d3 - typical indicator: d1 –d2 or d1 –d Base and Sub-base - affe affect ctss slo slope pe in next next regi region on –d5 - typical indicator: d2 –d4 or d3 –d Subgrade - affe affect ctss defl deflec ecti tion on at dis dista tanc ncee ical indic ndicaator: tor: d6 or d7 - t ical

d7

Asphalt over good foundation Concrete over poor foundation

Back-analysis is done by computer, with the following assumptions:  all layers are of uniform thickness and of infinite lateral extent;  all materials are linear elastic and homogeneous;  the load consists of uniform stress on a circular area;

62


DEPARTMENT

OF




dynamic effects due to inertia are negligible. How How man many y lay layer erss can can be anal analys ysed ed?? Two: Two: no prob proble lem m Three: Three: should should be OK OK Four: Four: be carefu careful; l; don’t don’t just just beli believe eve the the result result Another advantage over the Benkelman Beam and Deflectograph is that the FWD is equally useful on concrete or o r asphalt. It is particularly well suited to measuring load transfer efficiency across a joint. The loading plate is positioned one side of the joint and an d deflections are measured either side. Load Transfer Efficiency - affe affect ctss ‘st ‘step ep’’ acr acros osss joi joint nt –d3 or d3/d2 - typical indicator: d 2 –d (note: ≠ 0 or 100% even if perfect, due to distance between sensors 2 and 3) Slab Support - affe affect ctss ang angle le of load loaded ed slab slab –d2 / L12 - typical indicator: d 1 –d where L = distan distance ce between between sensors sensors

L12 d1

d2 d3

Poor load transfer Moderate load transfer Load Joint

Good load transfer

Rolling Wheel Deflectometers The ultimate deflection test device would be one that measured a full deflection bowl like the FWD, but which travelled at traffic speed along a highway, thus combining both quality and quantity of information. Such a device is the rolling wheel deflectometer , and several versions have been developed over the years, achieving their goal to varying degrees. One is currently being trialled by the Transport Research Laboratory. There is an inevitable trade-off between measurement accuracy and travel speed. Several companies and research organisations have used lasers to measure either distance from a datum of vertical velocity of the surface. So: not used much yet – but watch this space.

7.

Diagnosis

Pavements with an Asphalt Surface Rutting: Check the visual condition. If ruts are narrow with shoulders, the problem is near the surface (surface course or binder course probably); the wider the rut, the deeper the problem. Inspect cores carefully. If an asphalt layer appears rich in binder, especially if that binder is soft, that is likely to be the cause of the problem. Consider Co nsider carrying out repeated load axial tests (RLAT) to check whether materials are deformation susceptible. Also look at DCP data (if it exists). This should relate to rut resistance of foundation materials. Check FWD data. A subgrade stiffness of 50MPa or less indicates potentially deformable material.

63


DEPARTMENT

OF


Transverse Cracks : Look at the detailed crack shapes. If they are straight and regularly spaced, they are reflective cracks from joints in an underlying concrete pavement. If the shape is less regular, they may either be reflective cracks over a hydraulically-bound base or else low temperature cracking of the asphalt. If reflective cracking is suspected, check the crack distribution. If there is a concentration in the wheel paths then traffic is clearly playing an important part; if not, then the effect is almost entirely thermally driven. driven. Short transverse cracks may also have originated as construction defects and they may be progressing due to binder embrittlement . If FWD data is available, check for loss of asphalt layer stiffness in the wheel paths. This provides evidence of crack severity. Longitudinal Cracking in the Wheel Path : The fact that this cracking is in the wheel path proves that it is the traffic which is doing the damage. If there are several cracks within the zone of the wheel path then the effect is almost a lmost certainly shallow, possibly associated with debonding between asphalt layers. If there is only a single crack, cores through cracks are advised. In many cases, crack depth is shallow. Cores will also indicate where debonding between layers is associated with cracking. If an FWD survey has been carried out, check the asphalt layer stiffness . If it is low or variable then this implies significant damage and/or debonding. Compare FWD-derived stiffnesses with those from laboratory testing of recovered samples. If the two measures agree then the asphalt layers are likely to be intact and well bonded. Check evidence for binder hardening , e.g. from unusually high stiffness of laboratory specimens, or poor binder adhesion , e.g. unusually low stiffness, even of undamaged material. Also evaluate the insitu stiffness of any HBM layer from FWD evidence. If it is less than expected, this is evidence that cracking is present. Using the best available evidence for the stiffness of each layer, carry out multi-layer linear elastic pavement analysis and compute pavement life. Compare the theoretical life with past traffic numbers – and with current general pavement condition. Ravelling : Ravelling (and associated pot-holes) occurs when adhesion between binder and aggregate breaks down. Consider checking penetration of recovered surface course binder; or consider measuring surface course stiffness in the ITSM. These should identify binder hardening . Also check binder and filler contents since excess filler can contribute to poor adhesion.

p resent. Inspect the cores. Multiple layers of Bleeding : There is too much bitumen present. surface dressing are one common source of excess binder. Otherwise consider determining the void content of the surface course. Bleeding should only occur at void contents of 2% or less. Also check the visual condition for rutting since low void content also leads to asphalt deformation. Pavements with a Concrete Surface Transverse Cracks (jointed PQC) : Transverse cracking, either at mid-bay or a metre or so from joints, is common; it implies that the joint spacing was too large for the thermally-induced stresses and strains which have occurred. If the joint spacing is greater than about 20 times the slab thickness, joints cannot be expected to function properly.

64


DEPARTMENT

OF


Also find out whether the cracking occurred soon after construction. If so, shrinkage cracking may be the cause. Transverse Cracks (CRC) : Transverse cracks are expected in CRC. They should form at a spacing of 1-2m but should remain narrow. If there are more or they are no longer narrow, then the pavement is not functioning as intended and will continue to deteriorate. Check concrete strength, slab thickness and foundation stiffness. The combination should reveal whether the pavement as it was constructed should have lasted longer than it has. If so then the problem lies elsewhere, e.g. e.g . reinforcement defects, overloading. Longitudinal Cracking : This is a sign of overloading and is a very serious mode of distress. Check concrete strength, slab thickness and foundation stiffness. If the computed life is less than the current condition suggests, then this implies that one of the input parameters was more favourable in the past. Possibly the foundation used to be stiffer and has deteriorated over the years. Check FWD data for poor slab support . Also check evidence for subgrade softening, from DCP, unbound material samples and/or a drainage survey. If the computed life is greater than the current condition suggests, then something has happened which the computation doesn’t take into account. This could be shrinkage cracking during initial concrete curing. Also investigate the crack distribution. If it is localised then this suggests other areas may have much longer life. Faulting across Joints : This has a serious effect on ride quality. If dowels or tie-bars are present, there should be no faulting. If faulting is present, this can only mean serious corrosion of the bars and disintegration of the surrounding concrete. Even without dowels or tie-bars, faulting implies poor load transfer , possibly due to excessive joint spacing, poor durability aggregate, or a deformable foundation. Surface Deterioration : Concrete relies on the presence of a balanced combination of cement mortar and aggregate throughout. Excess cement mortar at the surface results in a relatively weak surface layer. Once trafficking has removed this excess mortar, there will be a decrease in ride quality. Another possibility is scaling , which means the loss of discrete areas of surface. This is usually caused b y frost action. 8.

Prognosis

Statistical Treatment of Data Since condition inevitably varies it is usual to work in terms of statistics, for example: 50 perc percen enti tile le (e.g. (e.g. of FW FWD D bac backk-ca calc lcul ulat ated ed stif stiffn fnes ess, s, or of thic thickne kness ss)) 15 percentile (i.e. 85% is better)

= aver averag agee = for design

Pavement Life Prediction Here you usually need an analytical computation, typically taking 15 percentile stiffness moduli. You must include consideration of the realistic long-term properties of materials. Sometimes it is possible to consult a design guide, but most are not flexible enough to cope with a deteriorated pavement.

65


DEPARTMENT

OF


Result = predicted lives to failure. The next step is to take account of the fatigue damage that has taken place already. Miner’s Law : This law states that relative damage is cumulative.

= 30 × 106 axle loads = 12 × 106 axle loads = 40% (18 × 106 axle loads remaining)

For example: predicted fatigue life past traffic therefore relative damage

The concept of relative damage becomes important for strengthening designs since, whatever the predicted life of the strengthened pavement, up to 40% of it may have to be discounted straight away due to past damage. For example: required future life = 30 × 106 axle loads  design traffic = 50 × 106 axle loads (since 40% has to be discounted) Rutting is a bit different. You have to look at what has happened in the past and use that to calibrate your future prediction. For exa mple, if you predict rutting due to excess subgrade strain, but there is no sign of rutting in the past – then ignore your prediction!

Calculations should never be believed without question, especially when so many assumptions are being made. This is especially true of concrete pavements , where predictions can change dramatically with a small change in one of the input parameters. The Effect of Debonding Debonding between bound pavement layers significantly decreases the overall apparent stiffness of the bound layers. 1.8

debonded interface

Strain at base of asphalt

s 1.6 s e n 1.4 f f i 1.2 t s / n 1 i a r t 0.8 s e 0.6 v i t a 0.4 l e R 0.2

Apparent stiff ness

Zero Bond

0 0

0. 2 0.4 0.6 0. 8 Depth of interface/total thickness

1

In the figure the debonded interface is assumed to transfer no shear stress at all, which is an extreme case. However, it is common to find that the apparent ap parent stiffness of the bound pavement layers is no more than half that expected when debonding is present. Note the tensile strain at the bottom of the asphalt. It can be up to 60% 6 0% higher than it would have been in a non-debonded pavement.

66


DEPARTMENT

OF


MAINTENANCE & REHABILITATION 1.

Localised Cracking

First option = just seal the cracks to stop water getting in.

Second option = patch repairs Cracks Debonded

interfaces Break out to required depth

Tack Coat

Problem: it is impossible to get the patch to be as good as the original pavement. It is difficult to compact, especially in the corners and it never bonds perfectly to surrounding materials.

2.

Seal

Surface De Deterioration

Surface problems are: Ravelling = pieces of stone becoming detached from the surface; Bleeding = excess bitumen coming to the surface (of an asphalt surface course); Polishing = stones losing their friction properties due to tyre action; Loss of texture = surface high-points wearing away or being pushed down into an asphalt.

Possible solutions: Gritting = spreading fine stone (e.g. 3mm) over the surface– counters bleeding – counters polishing Bush-hammering = abrade the surface stones Jet-blasting = eroding bitumen-filler mortar – counters loss of texture Grooving = cutting slots in the surface – also counters loss of texture – counters everything SURFACE DRESSING (see next page)

67


DEPARTMENT

OF


Surface Dressing (also known as Chip Seal) This is really a brand new thin surface layer (sometimes put straight onto a new road).

a) spray a layer layer of bitumen bitumen (usuall (usually y in the form form of emulsion) emulsion) over over the surface; surface; b) spread a single layer of aggregate particles, usually 10-15mm in size; c) watch watch the the traf traffic fic compac compactin ting g it. it. Single surface dressing

A ‘racked-in’ surface dressing includes a second a pplication, of much smaller aggregate size, designed to fill the gaps between be tween larger particles. Racked-in surface dressing

Design:

Surface hardness Skid resistance

Durability requirements Aggregate selection and spread rate

Binder selection and spread rate

Local economics

Surface dressing type

Season/ weather

Problems: Can’t be done in cold, wet weather; The surface isn’t particularly nice to drive on.

3.

Reflective Cracking

This is the phenomenon of cracks appearing in a surface course directly over cracks in the layer underneath. It happens for 2 reasons: a) b)

Thermal exp expansion/contraction  opening and closing of joints/cracks; Traffic loads causing high stress/strain stress/strain over a joint/crack.

What can you do about it?  At least 180mm of asphalt overlay (UK Highways Ag ency);  A geogrid or strong geotextile, providing actual enhanced strength (A in next figure);  A standard geotextile, acting as a separation layer (B in figure);  A high-durability asphalt layer (C in figure);  An open-graded asphalt layer (D in figure); 68

U NIVERSITY NIVERSITY OF NOTTINGHAM Pavement Engineering - Module H23P01  

DEPARTMENT

OF


A granular interlayer (E in figure); Crack and seat an underlying concrete layer (F in figure). A

B

C

D

E

C1: <0.5m

Is existing pavement visibly cracked?

Yes

Typical crack spacing?

C2: 0.5-2.0m C3: >2.0m

No

F

Geogrid

C1-2, T2, F1 Geotextile

C1, T1, F2 Traffic level?

T1: <5 106 T2: any

High-durability asphalt

C3, T2, F2

Open-graded asphalt

There is no need for an interlayer Is foundation water-sensitive?

F1: no F2: yes

C2, T2, F1 Granular interlayer

C2-3, T2, F1

4.

More serious cracking or rutting

We are now thinki thinking ng about: about: a) b) c) d) Question: how deep is the problem?

overlay overlays; s; inlays; deeper reconstruction; some combination.

0

10

20

Years

0 40mm inlay; A

Rutting might be entirely due to asphalt 10 deformation – but then it depends which layer is responsible. 20

Rutting in original pavement

A: mainly in surface course B: mainly in binder course

Rut depth (mm)

Rehabilitation

69

Rutting in rehabilitated pavement

40mm overlay; B

40mm inlay; B

40mm overlay; A


DEPARTMENT

OF


Cracking also may not penetrate too far from the surface. In thick pavements it is common to find cracks to 50-100mm depth only. 

just replace the surface course? Or maybe the binder course as well?

Question: how do we deal with broken materials? Generally, for any material that is going to be left in the pavement we have to: assign a long-term stiffness – for analytical  design;  estimate an equivalent thickness of new intact material – if using design manuals.

Overlay/Inlay Intact asphalt Failed asphalt

Downrate stiffness, e.g. to 500MPa

Damaged HBM

Take realistic longterm in-situ stiffness

We can also use geogrid reinforcement or geotextiles, especially if the problem is likely to involve reflective cracking.

Overlay/Inlay Failed asphalt

Predict cracking of geogrid-reinforced layer

Cracked Concrete

Take realistic stiffness

Subgrade

Take realistic stiffness

If a PQC pavement is in reasonable condition we can just bond an extra thickness onto the surface – used in USA particularly. pa rticularly.

Geotextile

Failed asphalt

Overlay/Inlay

Sub-base

Occasionally, usually on airfield pavements, it makes sense to put an entire new PQC slab on top of an existing cracked layer.

Damaged HBM

Overlay/Inlay

Treat as intact; predict cracking of of combined layer

PQC overslab Thin-bonded overlay

Predict cracking of geogrid-reinforced layer

Interlayer

Intact PQC

Failed PQC

HBM Sub-base

HBM Sub-base

Capping

Capping

Subgrade

Subgrade

Question: what if the problem is in the subgrade? Well, it may be OK to use an overlay to increase the thickness of the pavement. That will reduce the stress on the subgrade. However it may also be possible to improve its strength by adding/repairing drainage .

70


Has the upper subgrade softened?

OF


Yes Has it affected pavement peformance?

No

DEPARTMENT

No Don’t worry about the drainage

Yes

Yes

Is the drainage defective?

No

Is improvement technically feasible?

Yes

Think about drainage improvement

No

5.

Recycling

Most pavement materials can be recycled off site. For example, concrete can be crushed to generate new aggregate. Recycled asphalt planings (RAP) can be added to new asphalt – depending on quality and consistency of the source. Probl Problem em:: you you can’ can’tt hea heatt the the RAP RAP dire direct ctly ly becau because se it it wou would ld give give off off tox toxic ic fume fumes. s. Solut Solutio ion: n: supe superh rhea eatt conv conven enti tion onal al aggr aggrega egate te;; mix mix with with RAP; RAP; hea heatt is trans transfe ferr rred ed.. Consequence: there is a limit to the proportion of RAP that can be used – 15-30% say. Hot In-situ Recycling

Hopper for fresh asphalt

Preheater Units

Milling unit

Mixing drum

Remixer

Screed

Compactor

Limitations: Be careful with the heating – toxic fumes [flame, infra-red, superheated gases]  Impractical to heat more than say 50mm, 5 0mm, usually less   Needs a reasonably consistent pavement with shallow damage only Cold In-situ Recycling a) Brea Break k eve everrything hing up to a dept depth h of of up up to to 350 350mm mm;; b) Mix in a binder of some sort [emulsion, foamed bitumen, cement & water]; c) Compact; d) Pref Prefer erab ably ly leav leavee exp expos osed ed to the the air air for for a few few day days; s; e) Apply a new surface course.

71


DEPARTMENT

OF


Milling unit

Recycler

Compactor

Limitations: The ingredients will be relatively uncontrolled, including the water content.  Mixing will be imperfect, meaning there may be less binder at depth.   This means that properties will be poor relative to a plant-mixed material.  The surface will not be particularly smooth. Consequences: You generally need a plant-mixed, paver-laid surface course.  If emulsion or foamed bitumen are used, the resulting material will be like a high quality granular material; long-term stiffness may be around 1000MPa – but there is a large uncertainty margin on this. If cement and water are used then the material will be like a fairly weak (and  highly variable) HBM.

72


DEPARTMENT

OF


PAVEMENT MANAGEMENT Pavement management refers to the decision-making process as to what maintenance to do and when – which depends on the data available from condition monitoring. So the first question is: what type of monitoring should be carried out and how frequently?

1.

Managing Pavement Monitoring

Network Level A pavement authority needs to monitor performance regularly so as to plan budget allocation. Survey Type Traffic Count Axle weight Visual Condition

Profile Deflectograph SCRIM

Information Trends; design traffic Vehicle damage factors Roughness (approx) Structural condition (very approx) Skid resistance (possibly) Roughness Structural condition (very approx) Structural condition (approx) Skid resistance

Usefulness High Low-medium Very high

High

Medium Medium

Traffic counts can be automatic, using piezo-electric strips buried in the road surface, or manual. The advantage of automatic counts is that they can continue day and night for long periods with little expenditure. The adv antage of manual counts is that traffic can be classified accurately into different vehicle types (cars, buses, light goods, heavy goods – different axle configurations). Both are extremely useful in working out priorities. Axle weight surveys can be carried out either by stopping and weighing wagons on a fixed weighbridge or by installing a weigh-in-motion (WIM) device into the pavement. Such surveys would normally only be carried out at state or country level , in order to monitor trends in goods traffic development through the years, enabling wear factors to be updated as necessary. Visual survey data is essential . The information which results relates to ride quality , structural condition and also safety-related features such as skid resistance . It can be processed to give single numbers (e.g. Pavement Condition Index – PCI) that can be used (approximately) to assign likely remaining life and future maintenance costs. Visual survey data also allows minor maintenance to be programmed. Profile surveys , Deflectograph measurements and SCRIM surveys will all give a more accurate evaluation of individual aspects of condition, namely ride quality , structural

73


DEPARTMENT

OF


condition and skid resistance . However, the high productivity achievable from a profile survey, together with the fact that structural condition tends to correlate very roughly with profile, makes it the favourite for network-level monitoring. On major highways they would be carried out reasonably frequently (e.g. annually or biannually).

Project-Level A pavement management system (PMS) will include ‘triggers’ to indicate when detailed evaluation at project level is needed, based on individual indicators (visual, profile etc.) or a combination of them all. But when is the best time to do it? Condition Initial condition

Time range for optimised major maintenance

Projected performance following major maintenance

Condition for cost-effective cost-effective major maintenance Condition according to coarse network-level surveys

Range of possible actual condition

Time

It is sensible to carry out a detailed d etailed (project-level) evaluation well before the optimum intervention level is reached – according to network-level survey data. For example, if optimum intervention level is assumed to occur when the network-level condition indicator falls to 60% of its original value, it would be worth programming a detailed evaluation when the survey data indicates 70-75% because the real level may already be at 60%! Principle:  Decide on one or more network-level condition indicators.  Estimate what condition (according to the condition indicators) represents the optimum time for major maintenance.  Set triggers for project-level surveys at condition indicator values significantly higher than those for optimum major maintenance. Aims of project-level suvey:  to establish the deterioration mechanism(s); mechan ism(s);  to provide parameters for analysis and rehabilitation design. Possible tools to use:

74

U NIVERSITY NIVERSITY OF NOTTINGHAM Pavement Engineering - Module H23P01 Survey Type Cores; trial pits; laboratory tests

Radar (GPR) Detailed Visual Survey Falling Weight Deflectometer Drainage survey

DEPARTMENT

Information Layer thickness Bound material stiffness modulus Fatigue strength Deformation resistance Foundation strength Layer thickness High moisture locations Crack density Rutting locations/severity Layer stiffness modulus Joint condition Drainage efficiency

OF


Frequency Occasional

Continuous Continuous @10-100m Targeted

Basically, you have to decide what you need to know and choose accordingly. Projectlevel surveys need ‘designing’ rather than following a set procedure. Cores or trial pits are almost essential for every investigation, together with a detailed visual inspection of the pavement. Other surveys depend on situation. If analytical design procedures are to be used then the FWD (or rolling wheel deflectometer) is a key tool, supported b y appropriate laboratory tests and by use of the DCP in core holes or trial pits. Radar is less essential and its use should depend on the likely variability in construction. Drainage surveys are important where water-related problems are suspected.

2.

Managing Maintenance

Practical Constraints Multi-Lane Highways : Clearly one cannot raise (or lower) the surface level of one lane without doing the same to all other lanes on a carriageway. This means that a simple overlay solution may not really be as cost-effective as it looks. Highway Structures : Clearance to overbridges places an upper limit on raising levels. The carrying capacity of underbridges also places a limit; even a thin overlay adds a significant extra load. Kerbs and Barriers : Both kerbs and barriers can be raised – but at a cost. In both cases there is usually a degree of flexibility, allowing small increases in pavement level, and it is therefore important to know just how much flexibility is available in a particular case. Airport Runways: It is generally only the central 20m width which is damaged by aircraft and it is often permissible to ‘ramp down’ from a relatively thick overlay over the central part of the runway to little or zero thickness at the edge.

Project-Level Optimisation

75


DEPARTMENT

OF


This is not always easy in practice. The problem is that budget constraints may mean that the theoretical minimum ‘whole-life cost ’ solution may not be affordable in the short term! There therefore has to be some interaction between project-level design and o verall network-level management. In practice, the usual simplification is for the authority to tell the engineer what design lives to use (e.g. 20-year; 40-year). Usually therefore the designer will be asked to come up with the best b est solution to strengthen the road to take another x x years of traffic. In this case, it is simply a matter of costing up all the practical alternatives, taking into account any add-on costs due to practical constraints. Indirect costs (delays to road users, safety-related safety-related issues) may also be considered. One key question is: what should x be? This is not an easy question to answer. There has been a trend to increase design life over the years so that 40 years is now common, and the key parameter is the so-called discount rate . The idea is that a certain sum of money today is assumed to increase in purchasing power with time due to continuing economic growth (!!). The corollary is that future costs can therefore be ‘discounted’ at a certain annual rate. A debate d ebate rages as to the correctness of this assumption – and discount rates range from zero to about 10%. One further issue is that no-one really knows how long the need for any particular pavement will last. So, just accounting for this uncertainty, some sort of discount is probably justified. Network-Level Optimisation Network-level pavement management forces non-pavement costs to be taken seriously, including costs which are not strictly financial:  User costs (speed restrictions, vehicle wear and tear, fuel consumption, delays);  Accident costs;  Environmental costs (air pollution, energy usage). The most common method of o f dealing with user costs is to link them to International Roughness Index (IRI). The following are examples of equations eq uations used for this. Vehic hicle ope operrating co cost (V (VOC) Time cost (TC)

= VO VOClow IRI × (1+0.06[IRI–3]) = TClow IRI × (1+0.03× IRI)

[IRI > 3]

There is nothing fundamental about these two equations. The economic equation is so different in different countries that calibration is always required. Accident costs can be linked to skid resistance coefficient (µ) and to pavement condition , but only very approximately. Example equations are: Skidding accident cost Accident cost due to pavement

∝ Traffic flow × 10(1-1.8µ) ∝ Traffic flow × (rut depth – 10mm)

76


DEPARTMENT

OF


A good management system will include predictive equations for future deterioration – based on experience, possibly also on the history of each length of road. It is then possible to run ‘ what if?’ scenarios. For example, what if a particular length of road was strengthened? Or just surface-dressed? Then the up-front cost of the treatment can be balanced against the reduction in ongoing and future cost. Does the perfect Pavement Management Man agement System exist? NO; it’s just too complex and there are too many different factors to take account of. In the end, network-level management is often simply a matter of avoiding av oiding foolish mistakes!

77

Pavement Engineering Notes 2012

Recommend Documents