The Design of Temporary Excavation Support to EC7

Geotechnical Engineering Volume 165 Issue GE1 The design of temporary excavation support to Eurocode 7 Markham

Proceedings of the Institution of Civil Engineers Geotechnical Engineering 165 February 2012 Issue GE1 Pages 3–12 http://dx.doi.org/10.1680/geng.2012.165.1.3 Paper 900096 Received 17/12/2009 Accepted 20/01/2011 Keywords: codes of practice & standards/excavation/temporary works

ICE Publishing: All rights reserved

The design of temporary excavation support to Eurocode 7 Paul David Markham MSc, CEng, FICE Director, RNP Associates Limited, Sutton Coldﬁeld, UK

This paper examines how Eurocode 7 relates to the design of temporary excavation support. It is shown that total stress design and the net available passive resistance method described in Ciria Report 104 can produce inconsistent results, which must be carefully checked for sensitivity to the soil parameter values used. Many temporary excavations are designed using moderately conservative soil parameter values and a limit equilibrium method of analysis with a lumped factor of safety of 1.5. It is concluded that design in accordance with Eurocode 7 produces higher propping forces than design using this approach, but that Eurocode 7 produces propping forces over 60% lower than design in accordance with Ciria Report C580.

Notation cu cu;d ka kac kp kpc Psls ua up ªF ªG;dst ªG;stb a p va vp v9a v9p

1.

undrained cohesion design value of undrained cohesion active pressure coefficient active pressure coefficient passive pressure coefficient passive pressure coefficient value of the prop force derived from the SLS load case pore water pressure on active side of wall pore water pressure on passive side of wall partial factor for an action partial factor for a permanent destabilising action partial factor for a permanent stabilising action active earth pressure passive earth pressure total vertical stress on the active side of the wall total vertical stress on the passive side of the wall effective vertical stress on the active side of the wall effective vertical stress on the passive side of the wall

Where frames or props are used, proprietary equipment available for hire will usually be used (Figure 1). The major differences between the design of temporary and permanent excavation support include the following. Temporary works are required for a comparatively short time, thus in fine-grained soils it can be appropriate to use total stress parameters for the design rather than effective stress parameters j The design has to be undertaken based on the ground investigation report for the permanent works. Often this report has not been conceived or written for the design of temporary excavations; the necessary information for design j

Introduction

Eurocode 7 (EC7) (BSI, 2004a) is now the sole national and European standard for the design of geotechnical works and its use is mandatory on publicly funded work. There are several design guides to EC7 (Bond and Harris, 2008; Department for Communities and Local Government, 2007; Frank et al., 2004), which state the general manner in which EC7 is applied. This paper will specifically refer to the design of routine excavation support, usually carried out by a contractor as part of the construction process, and is based on the author’s experience of the design and checking of temporary works schemes. These excavations are supported by the use of steel trench sheets or steel sheet piles with or without support from frames or props.

Figure 1. Typical small temporary excavation supported by trench sheets and proprietary frames

3

Geotechnical Engineering Volume 165 Issue GE1

The design of temporary excavation support to Eurocode 7 Markham

of such structures is absent and there is not time to conduct a supplementary site investigation. Where suitable site investigation data are not available, then it would not be possible to do a design in accordance with EC7. j The excavation is in the upper layers of soil. The ground in this area often consists of fill, made ground, disturbed ground, head or recent deposits that do not form coherent beds. Owing to a lack of suitable case histories, it is usually difficult to base the design on case histories from previous excavations in similar material. j These temporary excavations are usually less than 6 m deep.

usually treated as unfavourable throughout. However, towards the toe of the wall the soil on the retained side becomes favourable because it prevents rotation of the toe of the wall and reduces bending moment in the wall and reduces prop loads. It is not clear where the soil changes from being unfavourable to favourable and using the principles of EC7 different partial factors should be applied to each of these actions, which would make the calculations unnecessarily complex. When favourable and unfavourable actions come from the same source, EC7 allows the same partial factor, ªF, to be applied to all those forces; this reduces the complexity of the calculations and has become known as the single source principle.

BS 8002:1994 (BSI, 1994) has not been widely adopted by designers of sheet pile walls (Puller and Lee, 1996). Therefore, temporary works design of retaining walls uses a hybrid of several methods. Gaba et al. (2003) recognised that there was a multitude of guidance for designers and that this advice was often contradictory, and therefore aimed to summarise both best practice and the alternative design methods in Ciria Report C580. This was published before the publication of EC7 and is not compatible with EC7 on some issues, such as factors of safety for stability and strut design.

2.

For retaining walls, in cohesionless soils, the active and passive earth pressures are calculated from the formulae active pressure

a ¼ k a v9a þ ua

passive pressure

p ¼ k p v9p þ up

Eurocode 7

2.1 Application of the partial factors Eurocode 7 is a limit state and partial factor code which adopts the same philosophy as the other structural Eurocodes, with actions subject to different partial factors depending on whether they are favourable (or stabilising) or unfavourable (or destabilising). For example, when treated as an action, earth pressure is subject to the partial factors ªG;fav ¼ 1.0 for a favourable action or the partial factor ªG ¼ 1.35 for an unfavourable action. On a propped wall (Figure 2) the soil on the retained side of the wall is

Substituting the usual formula for effective stress and rearranging to separate the effects of earth and water pressure, the formulae become:

a ¼ k a va þ ua ð1 k a Þ

and

Unfavourable permanent geotechnical action, γG ⫽ 1·35 Actual earth pressure distribution (indicative)

Earth pressure according to active earth pressure theory Actual earth pressure distribution Favourable permanent geotechnical action, γG ⫽ 1·00 Surcharge, γQ ⫽ 0

Earth pressure according to passive earth pressure theory

Figure 2. Propped retaining wall showing that it is not clear where the earth pressure changes from being an unfavourable action to a favourable action

4


p ¼ k p vp þ up ð1 k p Þ

Since ka is generally less than 1.0, these formulae show that the vertical total stress on the active side of the wall, va , and the water pressure on the active side of the wall, ua , are both unfavourable, since they increase the pressure on the active side of the wall. On the passive side of the wall, as the vertical total stress, vp , increases so the passive pressure increases, hence the vertical total stress is favourable. However, since kp is greater than 1.0, the term (1 kp ) is negative and water on the passive side reduces the passive resistance; hence it is unfavourable. To avoid applying different partial factors to the earth and water pressures it is simplest to use the single source principle. 2.2 Comparable experience Eurocode 7 places considerable value on comparable experience, which is defined in clause 1.5.2.2 as follows. Documented or other clearly established information related to the ground being considered in design, involving the same types of soil and rock and for which similar geotechnical behaviour is expected, and involving similar structures. Information gained locally is considered to be particularly relevant. (BSI, 2004a)

Therefore, to use comparable experience, the soil and the scale of the structure must be similar. However, there is a shortage of case histories for shallow excavations, which limits the circumstances in which comparable experience can be applied (Driscoll et al., 2008). Consequently, an alternative method is required (Section 3). 2.3 Deformation Eurocode 7 requires the engineer to assess deformation when this could have an effect on neighbouring structures. Where there are sufficient geotechnical data available, sufficient time for the analysis and appropriate engineering expertise, then deformation can be calculated using the finite-element (FE) method of analysis. However, the data and time are not usually available for FE analysis for these routine excavation supports. Therefore, deformation is best assessed on the basis of case history data such as the charts by Clough and O’Rourke (1990). 2.4 Guidance Previous standards, such as CP2:1951 (Institution of Structural Engineers, 1951) and BS 8002:1994 (BSI, 1994), attempted to provide the designer with rules for design and the geotechnical information necessary to do the design. EC7 provides only the rules for design to ensure safety and economy (Simpson and Driscoll, 1998) and allows the engineer considerable scope regarding choice of method. Therefore, the engineer is required to consult other references to non-contradictory complementary information (NCCI) such as Ciria Report C580 (Gaba et al., 2003). The British Standards Institution is also publishing NCCI for structures subject to


traffic loading. BSI has also published NCCI for structures subject to traffic loading in PD 6694-1 (BSI, 2011) which says very little about embedded retaining walls and refers the reader to Ciria Report C580 (Gaba et al., 2003). This leaves the engineer in a difficult position as illustrated by an international workshop (Orr, 2005: p. 4) which was held to review ten geotechnical examples that had previously been distributed to members of the European Technical Committee 10 (ETC10). The committee members prepared solutions to the design examples and the solutions showed a considerable spread in the range of results. It was concluded (Simpson, 2005) that some of the spread was due to contributors applying additional measures on top of the requirements of EC7, for example using values of =9 lower than required or applying additional penetration beyond that required by the equilibrium equations. However, it is necessary to apply additional measures that are not covered by EC7 to deal with the likes of passive softening, arching, minimum effective fluid pressure, eccentricity of load on props or accidental load on props. Therefore, owing to the lack of definitive NCCI, engineers will apply existing guidance in an inconsistent manner, which could result in unsafe or uneconomic designs (Simpson, 2005). 2.5 Selection of soil parameter values from test results Historically, there has been little guidance to the designer on how parameter values should be selected (Simpson and Driscoll, 1998) and engineers have always had to use ‘engineering judgement’ to obtain parameter values. The precise method of doing this has not been well defined, but EC7 attempts to provide some guidance to the engineer on this subject by suggesting that parameter values can be obtained using statistical techniques. The guidance is to be welcomed, but in practice the statistical techniques are difficult to apply and a large number of test results is required to produce meaningful results. Geotechnical engineers, by definition, have put their time and effort into the study of geotechnics and few geotechnical engineers will also have studied statistics to a sufficiently high level to be able to apply these statistical techniques (Simpson and Driscoll, 1998); consequently, the use of statistical methods is not appropriate on the majority of projects (Department for Communities and Local Government, 2007). The exception is on large projects where there is an abundance of high-quality ground investigation data and a team can be built up to include statisticians. This does not apply to the small, temporary excavations under consideration. So, although these statistical methods cannot be used directly, they can be used to show the way in which engineering judgement should be heading (Driscoll et al., 2008). 2.6 Accidental overdig It is usual to make some allowance for accidental overdig in design. This is commonly 10% of the depth of excavation below the bottom strut, up to a maximum of 0.5 m (see BS 8002:1994 (BSI, 1994) and BSC (1997)). Control of dig level during 5



excavation of a cofferdam is very good. Overdig costs money to excavate, more money to dispose of the arisings, and the overdig has to be replaced with expensive concrete, hence site management has incentives to minimise overdig. Where 75 mm of blinding is specified, the target dig level can be 10 mm high so that there is never any overdig.

monitoring frequency must be stated and reflect the level of risk and mode of failure.

EC7 allows the depth of accidental overdig to be determined by risk assessment, in which case the following factors should be taken into consideration. j j j j j j j j

Total depth of excavation. Depth below bottom frame. Is it possible to get a digger into the excavation to bottom up? (Figure 3) Is the ground easy to trim to the required level, or is it likely to come out in irregular lumps (e.g. weak sandstone)? Will the excavator driver work from outside the cofferdam? Will the excavator driver be able to see where he is digging or will he rely on a banksman? Will there be experienced supervisors on site throughout the construction period? Can remedial action be taken in the event that distress to the cofferdam is noted?

Where risk assessment shows that it is not necessary to consider accidental overdig in the design, then a sensitivity analysis should be undertaken to examine the effects of accidental overdig and care must be taken to ensure that the result of analysis is still reasonable. Additionally, the requirement for supervision and construction control measures must be specified in the geotechnical design report (Bond and Harris, 2008). Construction control measures would include a monitoring system with clear trigger levels. The

Figure 3. Trimming formation in an excavation where there can be good control of the dig level and little or no accidental overdig

6

2.7 Design in fine-grained soils The indicative design working life of a temporary structure is stated in Eurocode: Basis of Structural Design, BS EN 1990: 2002 (BSI, 2002) as 10 years. This is only indicative and designers can use other values for the design life, but the designer has to be able to justify the adoption of a shorter design life. The significance to a temporary retaining wall is that a small cofferdam can be installed, excavated and removed within a few weeks or months. For the design of excavation support in London clay, undrained conditions are commonly assumed for durations up to six months (Gaba et al., 2003). However, clause 9.6(3) of EC7 states that for silts and clays ‘water pressures should normally be assumed to act behind the wall. . .[corresponding] to a water table at the surface of the retained material’. This produces the same water pressures as assuming a water-filled tension crack (Figure 4) for the full depth of the wall. Where water is not expected, EC7 does not contain any rule regarding minimum effective fluid pressure (MEFP) of the retained soil, but the UK National Annex (BSI, 2004b) does refer to Ciria Report C580 (Gaba et al., 2003) as a source of NCCI, which does require MEFP. Where there is experience of excavations in similar soil, mixed total and effective stress design is recommended in Ciria Report C580 (Gaba et al., 2003) in conjunction with pore water pressures obtained from a flow net and passive softening. Where there is no potential for recharge, either at excavation level or within the soil, on the passive side of the wall, the recommended depth of passive softening is 0.5 m. Further recommendations are made in Ciria Report C580 (Gaba et al., 2003) where there is potential for recharge below formation, possibly by sand or silt layers. However, in this case, the appropriateness of design assuming undrained conditions is questionable, which leads to the conclu-

Figure 4. Tension crack adjacent to an excavation



sion that the depth of passive softening should be taken as 0.5 m whenever the design is based on undrained parameters; alternatively, undrained parameters should not be used. This should be used with the assumption that there is zero adhesion between the pile and the soil so kac ¼ kpc ¼ 2.0.

2.10 Factors applied to water pressure Eurocode 7 is not definitive regarding the factors that should be applied to water pressure, and the engineer is left with considerable scope regarding the factors to be used and the water levels to be used. This is discussed by Bond and Harris (2008) and they show five different ways in which EC7 can be interpreted regarding water levels and partial factors. Some of the interpretations make the calculations very difficult or lead to unrealistic situations and Bond and Harris (2008) go on to suggest the following approach, which is both realistic and reliable. When earth pressures are factored, the same partial factor (ªG ¼ 1.35) is applied to the earth and water pressure, which is calculated using the highest normal water level (i.e. the characteristic value). Alternatively, when the factor ªG ¼ 1.0 is applied to the effective earth pressure, the same factor is applied to the water pressure but the water pressure is calculated using the highest possible water level. This level can be obtained by applying a margin to the characteristic water level.

2.8 Thermal effects Props are affected by temperature because the ground, on the retained side, restrains the props against thermal expansion. This paper is concerned only with comparatively shallow sheet pile walls and, for a frictional soil, passive resistance is proportional to depth. Therefore, for a frame near ground level, supporting a sheet pile wall, there will not be much restraint to temperature effects, so prop loads should not be greatly affected by temperature. For example a 10 m long prop, with a 258 rise in temperature would lengthen by 3 mm (1.5 mm at each end). This would result in the force in the prop being slightly higher than the active force, but still, probably, less than that due to the earth pressure at rest. This leads to the conclusion that temperature changes have little effect on prop loads supporting flexible walls (Twine and Roscoe, 1999). However, using the values suggested in Ciria Report C580 (Gaba et al., 2003), a 10 m long, 305 mm 3 305 mm 3 97 kg/m universal column propping a flexible wall in stiff soil would have a temperature effect of 300 kN in addition to the prop force from the stability analysis. The method described in Ciria Report C580 (Gaba et al., 2003) was based on measurement of prop loads in strutted excavations and the method does not take account of the depth of the prop below ground level. For deeper props there will be more restraint, but for shallow props supporting sheet pile walls there is less restraint from the ground and the method in Ciria C580 could be conservative. 2.9 Workload Eurocode 7 has been criticised for requiring additional work from the engineer. However, for the relatively simple excavations under consideration, the amount of additional work is small in comparison to the overall amount of work required (Table 1).

3.

Reliability of existing methods

The difficulty of obtaining accurate predictions of design effects using numerical methods is illustrated by the benchmarking of the software Plaxis (Schweiger, 2009). Several engineers used Plaxis to analyse a strutted retaining wall where the constitutive model, the soil parameters and the geometry were predefined. Users were then free to make decisions regarding the type of elements to use, the tolerance settings, groundwater modelling and the extent of the grid. The range of calculated maximum displacement varied between 12 and 28 mm. The calculated bending moment ranged from 25 to 51 kN m and the calculated strut force ranged between 84 and 124 kN/m. Where engineers have to make decisions regarding the software to use, the soil parameter values, the constitutive model and the mesh, then there would be a greater range of calculated values. Software for FE analysis is complex, and incomplete understanding of the constitutive model was at least partially responsible for the Nicoll Highway collapse (Karlsrud and Andresen, 2007). Schweiger (2009) concluded that the results of a FE analysis need to be

Existing requirements based on Ciria Report C580 (Gaba et al., 2003)

Requirement of EC7-2

Understand the requirements Assessment and selection of soil parameter values The assessment of the load in the support system at an overall factor of safety of unity The assessment of the design wall depth at a factor of safety of 1.2–2.0 depending on the method of analysis chosen Where required, the serviceability load case Check sensitivity to accidental overdig and variation in soil parameter values Production of written calculations Production of drawings

As existing As existing DA1-C1, DA1-C2 As existing As existing Production of geotechnical design report As existing

Table 1. Comparison of the analysis required by EC7-1 and existing requirements

7



checked very carefully and that experience is necessary for obtaining reliable results from such an analysis.

3.1 Problems with total stress design There are problems with the gross pressure method where, in a total stress analysis, the factor of safety reduces as the pile length increases. With the strength factor method, a related problem can arise in which there is extreme sensitivity of the calculation when, on the retained side, the vertical effective stress at formation is about four times the undrained cohesion below formation (assuming ka ¼ kp ¼ 1.0, kac ¼ kpc ¼ 2.0 and the same soil on both sides of the wall) (Figure 5). This is because the active pressure, a ¼ k a v9a –k ac cu and the passive pressure, p ¼ k p v9p –k pc cu are about equal. Therefore, for the conditions stated, the net pressure below formation, p a 0, which makes the results of calculations oversensitive to small variations in the parameter values used. This type of problem is always present with total stress design; the inclusion in the calculation of MEFP, tension cracks and passive softening complicates the matter, but the same oversensitivity occurs for some combinations of circumstances. For example, a propped wall with a 5 m retained height has been analysed using the factor on strength method (Table 2) with allowance for passive softening, a MEFP of 5 kN/m3 and a water-filled tension crack. The driving depth, d, increases from 5.88 m to 12.26 m as the undrained cohesion is decreased from 38 kPa to 36 kPa. This is due to the net pressure below formation being close to zero when the factor of 1.4 is applied to the strength of the clay. Therefore, total stress design must be used with caution with the factor on strength method and the results should be checked for sensitivity to the parameter values used. This problem also occurs when using the net available passive resistance method as described in Ciria 104 (Padfield and Mair, 1984). The undue sensitivity only affects the driving depth of the piles; the prop load and moment in the sheets are only affected in proportion to the parameter values (Table 3).

These excavations are usually designed by the contractor as part of the construction process using the site investigation report commissioned for the permanent works design. On most small schemes the only laboratory test results available are the results of index tests and quick undrained triaxial tests. Accurate values of the stiffness parameters for numerical design are not generally available and the engineer either has to use approximate input data or, preferably, use a simple method of design with the soil parameters that are available (Gaba et al., 2003). Therefore, a simple robust method is needed for temporary works design. EC7 allows design by calculations; adoption of prescriptive measures; experimental models; load tests; or by an observational method. Numerical methods may not give realistic results and can give a false impression of accuracy; the distributed prop load method (DPL; Twine and Roscoe, 1999) is not appropriate for excavations less than 6 m deep, neither is the observational method. The remaining option, use of a limit equilibrium method, provides a simple, robust design method with an established history of successful use. Several comparisons have shown that calculation methods do not predict actual behaviour well for either limit equilibrium methods of analysis or for numerical methods of analysis (Day and Potts, 1989; Gaba et al., 2003; Kort, 2002; Lambe and Turner, 1970). The comparisons also show that the high factors of safety recommended by Gaba et al. (2003) in Ciria Report C580, are necessary for a safe design; however, these factors are not always used in temporary works design (see Section 4).

Surcharge, w

5·0 m

Prop

γ ⫽ 20 kN/m3 ka ⫽ kp ⫽ 1·0 kac ⫽ kpc ⫽ 2·0 cu ⫽ 25 kPa Active pressure

d

1·0 m

σ⬘v ⫽ 100 kPa

pp ⫽ kp γ d ⫹ kac cu ⫽ 50 ⫹ 20d

Figure 5. Extreme sensitivity when using the strength factor method

8

pa ⫽ ka σ⬘v ⫺ k ac c u ⫽ 100 ⫹ 20d ⫺ 2 ⫻ 25 ⫽ 50 ⫹ 20d



cu : kPa

Factor on strength

cud : kPa

Driving depth, d: m

Maximum moment: kN m

Prop load: kN m

38 36

1.4 1.4

27.14 25.71

5.88 12.26

145 207

81 99

Table 2. Oversensitivity when using total stress design with the gross pressure method of analysis

cu : kPa

Factor on strength

cud : kPa

Driving depth, d: m

Maximum moment: kN m

Prop load: kN m

56 49 45 43

1.4 1.4 1.4 1.4

40 35 32 31

1.04 1.83 4.02 6.78

36 48 44 46

29 30 32 33

Table 3. Oversensitivity when using the net available passive resistance method of analysis as described in Ciria Report 104 (Padﬁeld and Mair, 1984)

4.

Comparison with existing methods of design

A comparison has been made, for a range of frictional soils, between the output from a design to EC7 with the output of a design to Ciria Report C580 for a propped wall (Figure 6). Design approach 1 (DA1) has been used for the analysis of ultimate limit states (ULSs) to EC7. This requires the analysis of two ULS cases, referred to as combination 1 (DA1-C1) and combination 2 (DA1-C2). In addition, the serviceability limit state (SLS) must be considered. In DA1-C1 the main partial factors are applied to the actions (loads); this is a similar approach to that used in the other structural Eurocodes. The engineer can choose where in the calculation to apply the partial factors and for this comparison the calculations have been carried out using characteristic values and the effects factored. This applies the same partial factor to active pressure, passive pressure and water pressure and is known as the single source principle.

5·0 m

1·0 m

Surcharge ⫽ 10 kPa Prop γ ⫽ 19 kN/m3 γw ⫽ 10 kN/m3 Linear variation of hydraulic head (BS 8002: 1994)

Figure 6. Wall analysed for comparison of design to net pressure method with design to EC7

Using characteristic values for the soil parameters in this manner results in a calculated factor of safety for toe stability of 1.0 for DA1-C1. In DA1-C2 the main partial factors are applied to the material propertie (i.e. the shear strength of the soil). Therefore, the toe depth is almost always determined from the DA1-C2 calculation. The water level for each design situation must be assessed by the engineer, but for this example of temporary excavation support the highest possible water level has been taken as the characteristic value. Therefore, the margin suggested by Bond and Harris (2008) and discussed in Section 2.10 has been taken as zero and the characteristic value of water pressure has been used in the calculations (ªG ¼ 1.0). A factor of safety of 1.0 (using design values) has been used in the equilibrium calculations to determine the pile length. For both the analysis to EC7 and Ciria C580, the excavation depth has been taken as 5.4 m in the ULS load cases, which is a 10% increase on the actual depth below the frame to allow for unplanned excavation. The serviceability load case has been analysed using characteristic values and there is no allowance for unplanned excavation. The analysis shows (Figure 7) that for designs to EC7, the design moment for high-friction soil (9 ¼ 408) is obtained from DA1C1, whereas for the other three soils considered the design moment is obtained from DA1-C2. This shows that it is always necessary to check both load cases. The design prop load (Figure 8) for all the soils considered was obtained from DA1-C1. For design to Ciria Report C580, for the four soils considered, the design moments and propping force are obtained from the ULS analyses; in other words the prop forces from the ULS 9



600

Design approach 1 – combination 1

800

700



550


Ciria C580 ULS load case

500

Ciria C580 ULS load case

Ciria C580 – ULS moment derived from SLS load case (i.e. SLS ⫻ 1·35)

450

Ciria C580 SLS load case 600

Prop load: kN

Moment: kN m

400

500

400

350 300 250 200

300 150 100

200

25

100 25

30 35 Angle of shearing resistance: deg

40


40

Figure 8. Comparison of results of design to EC7 with design to net pressure method – prop load

Figure 7. Comparison of results of design to EC7 with design to net pressure method – moment 14

analysis are greater than 1.35 times the prop force from the SLS analysis (Figure 8). However, this is not the case for all soils and, as with design to EC7, it is necessary to check both load cases.


13

Ciria C580 ULS load case 12

Driving depth: m

The analysis to EC7 design approach 1 always produced higher bending moments in the sheets than the analysis to Ciria Report C580, by between 3% and 9% (Figure 7) and required longer sheets (Figure 9) by between 40 mm and 690 mm (0.5% and 6%). In practice, the sheets are selected from a limited range and are sized for driveability as well as the moment produced from this analysis, so this will not make much difference to the final design. The pile length, in the EC7 calculations, was always obtained from DA1-C2.


11 10 9 8 7

However, the analysis to Ciria Report C580 always produced higher design values for the prop forces than the analysis to EC7 design approach 1 by between 61% and 65% (Figure 8). This is because of the large factors of safety (Table 4) applied at the end of the calculation using the Ciria Report C580 method. Some engineers might consider these factors of safety a model factor also to be applied to the output from a limit equilibrium 10

6 25


40

Figure 9. Comparison of results of design to EC7 with design to net pressure method – driving depth


SLS prop load (unfactored soil parameters) ULS prop load (factored soil parameters)


Limit equilibrium calculations

Soil–structure interaction calculations

1.85 3 calculated value The greater of: 1.35 3 Psls 1.85 3 calculated value

1.0 3 calculated value The greater of: 1.35 3 Psls 1.0 3 calculated value

Table 4. Factors of safety for the design prop loads to Ciria Report C580 (Gaba et al., 2003)

calculation to EC7. Additionally, it is recommended in Ciria Report 104 (Padfield and Mair, 1984) that the prop loads from a limit equilibrium analysis are increased by 25% to allow for arching. However, neither the model factor nor the 25% increase are written into EC7, so engineers can produce a design in accordance with EC7 without using them; this means that a design to EC7 would not have the same level of reliability as existing design methods, which was one of the fundamental requirements of EC7 (Simpson and Driscoll, 1998). However, the frames for temporary excavations are often designed using lower factors of safety than recommended in Ciria Report C580. For routine excavation, it is common, throughout the hire industry, to use moderately conservative soil parameter values with a limit equilibrium method of analysis and apply a lumped factor of safety of 1.5 to the analysis output to calculate the ultimate frame load. This is equivalent to the analysis of the serviceability loadcase and the calculations show that the ULS prop load calculated to EC7 for these examples varies from between 1.59 and 1.60 times that calculated from the serviceability loadcase. Therefore, analysis of the serviceability load case and a factor of safety of 1.5 on the frame load do not produce a design with the same reliability as a design to EC7. Limit equilibrium methods of analysis are not geotechnically rigorous; usually the movement of the sheets is insufficient to achieve the fully active or fully passive earth pressures on which the methods are predicated. However, they do have the advantage that they are simple to apply and robust in use. That is, the factors of safety that are used with these methods are method dependent and have been based on much experience to produce a safe design. Any limit equilibrium method is permitted by EC7, but the factors of safety (partial factors) for use with design to EC7 have been predefined in the National Annex to EC7 (BSI, 2004b). In the past few years, the net available passive resistance method of design (see Padfield and Mair (1984)) has been widely used. However, its use with partial factors can produce inconsistent results (Table 3) and it may prove necessary to develop another method that can be used with the partial factors of EC7.

5.

Conclusions

Currently, in the UK, a variety of methods are used for the design of temporary retaining walls and, in general, design is at working load whereas the structural codes use partial factors. In that it is a

limit state and partial factor code, EC7 is a major step forward in standardisation, but the lack of prescriptive measures will lead to much disagreement and uncertainty among engineers. In the UK, only design approach 1 is permitted by the National Annex to EC7-1 (BSI, 2004b). Even using a single method of design, different interpretations of EC7 can produce very different results. For example, passive earth pressure can be treated as a resistance, a favourable action or a negative unfavourable action, each of which produces a different result to the calculations. Water pressure can also be treated in several different ways and a consistent approach is required, such as using the single source principle, which removes some of the complexity of, and is easier to use than, applying different partial factors to the different parts of the earth and water pressure diagrams. Some of the opposition to EC7 can be explained by conservatism of geotechnical engineers and the unwillingness to change existing practices that are known to work. Interpretation of EC7 produces disagreement, among engineers, regarding whether numerical modelling is required and interpretation of the individual clauses. Using EC7, it is necessary for the engineer to refer to NCCI to complete the design. In the UK, Ciria Report C580 (Gaba et al., 2003) is cited as NCCI; however, parts of Ciria C580 are contradictory, so a partial rewrite is required urgently, to ensure reliability in design and to ensure that EC7, unlike BS 8002, is widely accepted by temporary works engineers. Currently, for cohesive soils, total stress design or mixed total and effective stress design are often used for temporary excavations. However, whenever total stress parameters are used on the passive side of the wall, the analysis can become ill-conditioned and a sensitivity analysis is necessary. The routine method of design for small temporary excavation support has become a hybrid method based on Ciria Reports 104 (Padfield and Mair, 1984) and C580 (Gaba et al., 2003). When used with the partial factors of EC7, the overall factor of safety is significantly less than recommended in Ciria Report C580 and the results can be sensitive to the parameter values used, which makes it essential to check the results for sensitivity to parameter values and accidental overdig, as well as checking against the output from an existing method of design. 11



REFERENCES

the International Workshop held in Dublin, Ireland on 31 March–1 April 2005. Trinity College, Dublin, Ireland. Padfield CJ and Mair RJ (1984) Design of Retaining Walls Embedded in Stiff Clay. Construction Industry Research and information Association, London, UK, Ciria Report 104. Puller M and Lee CKT (1996) A comparison between the design methods for earth retaining structures recommended by BS8002:1994 and previously used methods. Proceedings of the Institution of Civil Engineers – Geotechnical Engineering 119(1): 29–34. Schweiger HF (2009) Plaxis Benchmarking. Plaxis, Delft, the Netherlands. See http://www.plaxis.nl/upload/benchmarks/ Benchmark%203%20-%20Embankement%201.pdf (accessed 27/07/2009). Simpson B (2005) Retaining wall examples 5–7 Addendum October 2005. Evaluation of Eurocode 7: Proceedings of the International Workshop held in Dublin, Ireland on 31 March–1 April 2005 (Orr TLL (ed.)). Trinity College, Dublin, Ireland. Simpson B and Driscoll R (1998) Eurocode 7 – A Commentary. Building Research Establishment, Watford, UK. Twine D and Roscoe H (1999) Temporary Propping of Deep Excavations – Guidance on Design. Construction Industry Research and information Association, London, Ciria Report C517.

Bond A and Harris A (2008) Decoding Eurocode 7. Taylor and

Francis, Abingdon, Oxfordshire, UK. BSC (British Steel Corporation) (1997) Piling Handbook, 7th edn.

British Steel Corporation, Scunthorpe, UK. BSI (1994) BS 8002: Code of practice for earth retaining

structures. BSI, London, UK. BSI (2002) BS EN 1990: Eurocode: Basis of structural design.

BSI, London, UK. BSI (2004a) BS EN 1997: Eurocode 7: Geotechnical design –

Part 1: General rules. BSI, London, UK. BSI (2004b) NA to BS EN 1997: UK National Annex to

Eurocode 7: Geotechnical design – Part 1: General rules. BSI, London, UK. BSI (2011) PD 6694, pt 1: 2011: Recommendations for the design of structures subject to traffic loading to BS EN 1997-1: 2004. BSI, London, UK. Clough GW and O’Rourke TD (1990) Construction induced movement of in-situ walls. In Design and Performance of Earth Retaining Structures (Lambe P and Hansen LA (eds)). American Society of Civil Engineers, Reston, VA, USA, Geotechnical Special Publication no. 25, pp. 439–471. Day RA and Potts DM (1989) A Comparison of Design Methods for Propped Sheet Pile Walls. The Steel Construction Institute, Ascot, UK, SCI Publication 077. Department for Communities and Local Government (2007) A Designers’ Simple Guide to BS EN 1997: 2004. Department for Communities and Local Government, London, UK. Driscoll R, Scott P and Powell J (2008) EC7 – Implications for UK Practice. Construction Industry Research and Information Association, London, Ciria Report C641. Frank R, Baudin C, Driscoll R et al. (2004) Designers’ Guide to EN 1997-1 Eurocode 7: Geotechnical Design – General Rules. Thomas Telford, London, UK. Gaba AR, Simpson B, Powrie W and Beadman DR (2003) Embedded Retaining Walls – Guidance for Economic Design. Construction Industry Research and Information Association, London, UK, Ciria Report C580. Institution of Structural Engineers (1951) Earth Retaining Structures. The Institution of Structural Engineers, London, UK, Code of Practice No. 2, CP2. Karlsrud K and Andresen L (2007) Design of deep excavations in soft clay. In Geotechnical Engineering in Urban Environments: Proceedings of the 14th European Conference on Soil Mechanics and Geotechnical Engineering, Madrid, Spain (Cuellar V, Dapena E, Alonso E et al. (eds)). IOS Press, Amsterdam, the Netherlands, vol. 1, pp. 75–99. Kort DA (2002) Sheet Pile Walls in Soft Clay. University Press, Delft, the Netherlands. Lambe TW and Turner EK (1970) Braced excavations. Proceedings of the Speciality Conference on Lateral Stresses in the Ground and Earth Retaining Structures, Cornell University, Ithaca, New York, 22–24 June 1970, pp. 149– 218. Orr TLL (ed.) (2005) Evaluation of Eurocode 7: Proceedings of 12

WH AT DO YO U T HI NK?

To discuss this paper, please email up to 500 words to the editor at [email protected]. Your contribution will be forwarded to the author(s) for a reply and, if considered appropriate by the editorial panel, will be published as a discussion in a future issue of the journal. Proceedings journals rely entirely on contributions sent in by civil engineering professionals, academics and students. Papers should be 2000–5000 words long (briefing papers should be 1000–2000 words long), with adequate illustrations and references. You can submit your paper online via www.icevirtuallibrary.com/content/journals, where you will also find detailed author guidelines.

The Design of Temporary Excavation Support to EC7

Recommend Documents