TRIITY COLLEGE DUBLI Department of Civil, Structural ISSMGE ERTC 0 and Proceedings of the International Workshop on the EVALUATIO OF EUROCODE 7 Cost: 50.00 including postage Order Form ame Affiliation Address E-mail Telephone To order a copy of the Proceedings, please send this form and payment confirmation to: Ms Linda McHugh, Executive Officer, Department of Civil, Structural and Environmental Engineering, Museum Building, Trinity College, Dulin, Ireland, fax: +5 677 07 PAYMET METHODS: (Payment is accepted either electronically, y ank draft or y cheque drawn on major international anks. Please tick ox: Cheque or Draft Electronic funds transfer Electronic payments should contain the following details: Reference: Dept of Civil Engineering, Credit Code: 80094.07.07.06 (Please quote this credit code reference with your lodgement.) Domestic (within Repulic of Ireland) Bank Account ame: Trinity College o. Account ational Sort Code 90007 Bank Account umer 00795 International Bank Account ame Trinity College o. Account BIC/Swift Code: B O F I I E D IBA: IE9BOFI 9000700795
Eurocode 7 Workshop Retaining wall examples 5-7 B. Simpson Arup Geotechnics, London, UK. ABSTRACT Three examples of retaining walls were prepared for comparison of designs: a gravity wall, a cantilever and an anchored emedded wall. An analysis of the designs sumitted is presented here. Despite efforts to avoid unintentional differences, there is a considerale disparity in the results, even among contriutors who have apparently used the same Design Approaches. These may due to differing interpretations of the prolems, differing calculation methods, or perhaps to errors. Where detail has een provided to the author, it appears that relatively small variations in earth pressure coefficients sometimes lead to significant differences in results, proaly outweighing the effects of choice of Design Approach in EC7. ITRODUCTIO On March to April 005, a workshop was arranged in Dulin under the joint auspices of European Technical Committee 0 and Technical Committee of ISSMGE, and GeoTechet Working Party. A total of ten designs were considered, of which three were retaining walls: a gravity wall (Example 5), an emedded cantilever (Example 6) and an anchored emedded wall (Example 7). Calculations had een sumitted in advance for these designs y 5 contriutors and an initial comparison was presented at the workshop. Following this, the examples were re-defined, slightly more tightly, and contriutors were asked to re-sumit their calculations, showing their assumptions, methods and results in a proforma style which facilitated comparisons. In total, contriutors have sumitted calculations, of which provided full details. The graphs in this paper show the results which have een otained. A numer in the range 0 to 7 or a letter A to G has een allocated to each document received, as an anonymous identifier. Some documents contain several calculations for the same prolem, and in some cases several contriutions have een received from the same country. The points on the graphs are annotated as follows to represent the various design approaches: EC7 Design Approach, taking the worst case of Cominations and. EC7 Design Approach EC7 Design Approach an existing ational method In the case of Design Approach, results for the less severe comination are denoted y if it is Comination (formerly Case B) and y c for Comination (formerly Case C). Results are shown in old for which details of methods and assumptions have een studied, from which most of the conclusions in this paper have een drawn.
EXAMPLE 5 GRAVITY RETAIIG WALL. Basic method The design prolem set as Example 5 is shown in Figure, and the results sumitted are summarised in Figures to 4. The asic steps of a calculation using EC7 Design Approach (DA) are shown in Figure 5; the notations C and C is used to indicate cominations and (formerly Cases B and C in EV 997). The force on the virtual ack of the wall (the vertical plane through the rear of the ase) is generally taken to e parallel to the ground surface, ie at an inclination of 0 ; in this particular case this happens to e very close to the value of φ, which is used at interfaces etween soil and concrete not cast against the ground. By all design methods, the equilirium of the wall has to e estalished, incorporating factors of safety, and this is used to fix the width of the ase, B. This requires an iterative calculation in which a value of B is proposed and then equilirium is checked. It is therefore possile that some of the results sumitted do not have B completely minimised.. Eccentricity and inclination of the ase force The final part of this calculation is the check on earing capacity for the resultant force, marked as F in Figure 5. This acts centrally on an effective area of the ase and is inclined to the vertical. Its eccentricity and inclination are therefore critical to the calculation of earing capacity. Derivation of the force F includes factors of safety, and two approaches have een adopted y contriutors in considering eccentricity and inclination when deriving earing capacity: (a) calculating eccentricity and inclination using the force F derived from factored components, and () calculating eccentricity and inclination from the characteristic force F, unfactored. In either case, partial factors of safety are incorporated in the calculation of earing capacity, and this is then compared with the design value of force F derived from factored components. The final check is made y comparing the design value of the vertical component, V d, of force F with the design vertical resistance R d, so the requirement is R d /V d. For typical values related to the footing of the gravity wall, Figure 6 shows how R d /V d varies for width B=.75m. For the characteristic case, R k /V k is.0, and this falls to R d /V d =.04 if characteristic eccentricity and inclination are used in otaining R d, indicating a near-optimum design. However, significantly lower, unacceptale values of R d /V d are otained if the horizontal and vertical force components are factored efore calculating inclination and eccentricity, particularly since the vertical component is favourale to inclination and eccentricity, and the horizontal component unfavourale. Hence, if approach (a) aove is followed the footing has to e made igger, though it is acceptale according to approach (). This issue was discussed during the workshop in relation to spread foundations generally; it is particularly relevant to gravity retaining walls. There appeared to e a majority view that the calculation of eccentricity and inclination should take account of the factored forces, and the author supports this view. In this case, the vertical component depends mainly on the density of the ackfill, which is usually fairly well controlled, whereas the horizontal component depends on oth the density of the ackfill and its strength, which is more uncertain. These features are accommodated in Design Approach Comination, in which density and strength are treated separately, ut it is less clear how to treat this question in the other design approaches. For retaining walls, the author recommends that this prolem is avoided for DA Comination y applying the factors to the derived ending moments and shear forces only; this is considered acceptale provided Comination is also checked (see EC7.4.7..()).. Reasons for the range of results The calculated ase widths shown in Figure suggest that smaller ases have een otained using DA than with DA, at least for the cases which can e checked in detail (old symols). This is thought to e mainly a result of the way eccentricity and inclination are calculated, as discussed aove. Values for ending moment and shear force in the wall are
relatively consistent. Some contriutors have used a coefficient of earth pressure equivalent to ½(K a +K o ) for calculation of action effects in the wall. In the author s view, this is appropriate for the SLS check on the wall, checking deflections and crack widths ased on earth pressures using unfactored parameters, ut proaly not to the ULS strength check. This could well mean that the SLS check governs the thickness of the wall and its reinforcement. Some contriutors used Coulom s equation for derivation of coefficients of earth pressure, in contrast to the methods proposed y E997- Annex C. In this example, this made little difference since only coefficients of active pressure are important. Some contriutors used the passive resistance in front of the wall whilst others ignored it; this made little difference. In general, the results of calculations using EC7 are within the range of results otained using other national methods..4 Serviceaility limit states Some contriutors checked that the resultant force F for the characteristic state passed through the middle third of the footing. This was the only serviceaility check noted; no contriutors attempted to calculate displacement no targets were set in the instructions. EXAMPLE 6 EMBEDDED SHEET PILE RETAIIG WALL The design prolem set as Example 6 is shown in Figure 7, and the results sumitted are summarised in Figures 8 and 9. All contriutors used simple active and passive earth pressure diagrams, of the type illustrated in Figure 0. Results for the various design approaches of E997- are fairly consistent in this case: the differences etween contriutors using, nominally, the same method were greater than those of single contriutors using several methods. The majority of contriutors who provided details used δ= φ, consistently with E997-, 9.5.(6) and allowed for 0.m overdig, following E997-, 9...(), ut there were exceptions to oth of these. Some contriutors considered vertical equilirium and as a result reduced the availale angle of wall friction sustantially on the passive side; the author has not een ale to confirm the need for this. Some variations in result were caused y differing methods of deriving coefficients of earth pressure, including the difference etween the charts and formulae of E997- Annex C. As noted y Simpson and Driscoll (998), for high values of φ and δ/φ the charts give higher values for coefficient of passive pressure than do the formulae. The charts are ased on the work of Kerisel and Asi (990), whereas the formulae give results consistent with Lancellotta (00). According to the author s calculations (Contriutor 0 on the plots), the effect of 0.m overdig is to increase the required emedment y 0.76m. This is shown y the first two results for DA at the left hand side of Figure 8: the point marked * has no allowance for overdig. It has a similarly significant effect on calculated ending moment, as shown in Figure 9. Overall, the results otained using EC7 appear to e similar to the results of calculations using other national methods. 4 EXAMPLE 7 ACHORED SHEET PILE QUAY WALL The design prolem set as Example 7 is shown in Figure, and the results sumitted are summarised in Figures and 4. In this case, calculations could use either simple activepassive diagrams, as illustrated in Figure 5, or a method of soil-structure interaction which allows for arching and redistriution of earth pressures. Similar comments to those of Example 6 apply to this example, and again the point marked * shows the effect of no allowance for overdig.
Figures 6 and 7 show two methods used to represent redistriution of earth pressures. Figure 6 shows the results of a calculation of soil-structure interaction y the Oasys FREW program, which models the soil as an elastic continuum limited y active and passive pressures, which are allowed to redistriute within certain theoretical limits. Figure 7 shows redistriution rules typical of German practice (EAU 996). These methods tend to give lower ending moments, possily shorter walls, ut higher anchor forces than are otained y simple active-passive diagrams, depending on the stiffness assumed for the anchor. In this context it is relevant that the additional specifications given after the workshop include the requirement the length of the wall is to e the minimum allowale. It would have een equally possile to specify the length of the wall is to e x metres, the anchor force is to e the minimum allowale or the ending moment is to e the minimum allowale, each of which could have led to different designs. 5 STRUCTURAL DESIG Examples 6 and 7 required the ULS design ending moments to e calculated and compared with national methods. However, the most relevant comparisons etween methods would e ased on the final outcome: in this case the lengths of the walls and the steel sections to e adopted or, more generally, the cost of construction. Designs to EC7 should e completed y reference to EC Part 5, which allows for plastic design of most roust sheet pile sections, hence providing a saving compared with some traditional methods, even for the same design ending moment. This is particularly significant for the propped cantilever since some limited rotation is allowed to occur at the first hinge to form, allowing further redistriution of earth pressures. E997- notes the importance of potential forms of rittle ehaviour ( reduction in strength with deformation ) in retaining wall design at 9.4.(4) and 9.7.6(4). either of these paragraphs refer specifically to struts or anchors, ut it is ecause of fear of rittle ehaviour in strutting, in particular, that some national codes apply large factors of safety to calculated strut forces. A distinction might e made etween forces calculated for struts and anchors or etween rittle and ductile situations, ut, as in EC7, this is sometimes not done. (An example of this is shown y the national result from Contriutor A in Figure 4 which had a factor of.85 on the anchor force.) In the author s opinion, designers should e particularly aware of the use of rittle components in support of retaining walls, especially if simple active-passive diagrams, like that of Figure 5, are adopted, since these tend to underestimate strut or anchor forces. 6 HUMA ERROR The results show considerale scatter, even when calculated y different contriutors using nominally the same method. They clearly represent a considerale range of safety, and it seems likely that some of the less conservative designs would actually e unsafe if used in practice. If this is not so, then most of the designs are grossly uneconomic. When the results were first presented to the workshop, the range was even greater. It has een reduced partly y the additional specifications noted on Figures, 7 and, and partly y contriutors correcting their own calculations. It is apparent that misunderstandings and calculation errors can have significant effects on engineering designs. In the author s opinion, factors of safety have an important role in covering a certain degree of human error. For this reason, the results of new design processes should e checked against traditional methods. Reductions in overall safety levels, and related economies in designs adopted for construction, should only e accepted in small increments, and tested in extensive practice efore further reductions are considered. 7 COCLUSIOS The following conclusions are drawn.
. In general, the results of the EC7 calculations were within the range of results from national methods. The variation due to differences of approach to earth pressure coefficients, for example, are as great as those due to the differing design approaches of EC7.. The method of calculating eccentricity and inclination of loading to e used in deriving earing capacity of spread foundation has a significant effect; agreement is needed.. Allowance for overdig has a ig effect on the resultant design. 4. EC7 should perhaps differentiate, in terms of methods or safety factors, etween those struts or anchors which are ductile and those which fail in a rittle manner. 5. Virtually no attention was paid to serviceaility. In principle, this could control the geometry of the designs, ut this is unlikely in these cases. 6. It is necessary that factors of safety are large enough to provide some protection against human error. REFERECES EAU 996 Recommendations of the Committee for Waterfront Structures, Harours and Waterways. Ernst & Sohn. Kerisel, J and Asi, E. 990. Active and passive earth pressure tales. Balkema. Lancellotta, R 00. Analytical solution of passive earth pressure. Géotechnique 5, 8 67-69. Simpson, B & Driscoll, R 998. Eurocode 7 - a commentary. Construction Research Communications Ltd, Watford, UK.
Surcharge 5kPa 0 o 6m 0.4m Fill 0.75m Sand B =? Design situation - 6m high cantilever gravity retaining wall, - Wall and ase thicknesses 0.40m. - Groundwater level is at depth elow the ase of the wall. - The wall is emedded 0.75m elow ground level in front of the wall. - The ground ehind the wall slopes upwards at 0 o Soil conditions - Sand eneath wall: c' k = 0, φ' k = 4 o, γ = 9k/m - Fill ehind wall: c' k = 0, φ' k = 8 o, γ = 0k/m Actions - Characteristic surcharge ehind wall 5kPa Require - Width of wall foundation, B - Design shear force, S and ending moment, M in the wall Additional specifications provided after the workshop: The characteristic value of the angle of sliding resistance on the interface etween wall and concrete under the ase should e taken as 0º. The weight density of concrete should e taken as 5 k/m. The earing capacity should e evaluated using to the EC7 Annex D approach. 4 The surcharge is a variale load. 5 It should e assumed that the surcharge might extend up to the wall (ie for calculating ending moments in the wall), or might stop ehind the heel of the wall, not surcharging the heel (ie for calculating staility). Figure Cantilever Gravity Retaining Wall
Example 5 - Gravity wall 6.0 BASE WIDTH m 5.0 4.0.0.0 = =.0 0.0 0 0 5 5 5 8 8 6 6 7 G C C C C C C C Figure Gravity retaining wall - calculated required ase width Example 5 - Gravity wall 00 BEDIG MOMET km/m. 000 800 600 400 00 = = 0 0 0 5 5 5 8 8 6 6 7 G C C C C C C C Figure Gravity retaining wall calculated ULS design ending movements
Example 5 - Gravity wall BEDIG MOMET km/m. 00 000 800 600 400 00 0 = = 0 0 5 5 5 8 8 6 6 7 G C C C C C C C C:\BX\BX-C\EC7\Dulin\[Dulin-results.xls] Figure 4 Gravity retaining wall calculated ULS design shear force Surcharge 5kPa 0 o 6m 0.4m Fill 0 o 0.75m K a γz Sand B =? Figure 5 Gravity retaining wall calculation to DA
Data γ (k/m ) 9 φ' k ( o ) (c' k = 0) 4 Overurden depth (m) 0.75 Overurden pressure(kpa) 4.5 Column no. 4 4 Base width.75.75.75.75.75 Eccentricity (m) 0.57 0.57 0.57 0.79 0.79 Effective width B' (m).6.6.6.7.7 Vertical force k/m 690 94 690 94 690 Horizontal force k/m 07 85 85 85 85 Inclination H/V 0.0 0.0 0.4 See note 0.4 q 9.4 9.4 9.4 9.4 9.4 c 4. 4. 4. 4. 4. g 8.4 8.4 8.4 8.4 8.4 iq 0.49 0.49 0.4 0.4 0.4 ig 0.4 0.4 0.0 0.0 0.0 ic 0.47 0.47 0. 0. 0. R (k/m) 9 7 879 659 659 γ(r).4.4.4.4 Rd (k/m) 9 98 68 47 47 Rd/Vd.0.04 0.9 0.50 0.68 Column no. Column no. Column no. Column no. 4 Column no. 5 Characteristic values of all parameters. Characteristic eccentricity and inclination; forces and resistance factored. Characteristic eccentricity; unfavourale (horizontal) force and resistance factored. Favourale (vertical) force not factored in deriving inclination or for comparison with resistance. Unfavourale (horizontal) force and resistance factored. Favourale (vertical) force not factored in deriving inclination or eccentricity, ut factored for comparison with resistance. Unfavourale (horizontal) force and resistance factored. Favourale (vertical) force not factored in deriving inclination or eccentricity, or for comparison with resistance. Figure 6. Base of gravity wall alternative approaches to eccentricity and inclination
0kPa.5m Sand.0m D=? Design situation - Emedded sheet pile retaining wall for a m deep excavation with a 0kPa surcharge on the surface ehind the wall Soil conditions - Sand: c' k = 0, φ' k = 7 o, γ = 0k/m Actions - Characteristic surcharge ehind wall 0kPa - Groundwater level at depth of.5m elow ground surface ehind wall and at the ground surface in front of wall Require - Depth of wall emedment, D - Design ending moment in the wall, M Additional specifications provided after the workshop: The surcharge is a variale load. The wall is a permanent structure. Figure 7 Example 6 Emedded cantilever Example 6 - Emedment 8 7 = = EMBEDMET DEPTH D m. 6 5 4 * = 0 0 0 0 A 5 5 5 5 7 7 7 7 8 9 C C C C E E E E 9 9 9 0 0 G 5 5 5 5 5 5 6 6 7 Figure 8 Emedded cantilever calculated required emedment
Example 6 - Bending moment 00 BEDIG MOMET km/m. 50 00 50 00 * = 50 0 0 0 0 A 5 5 5 5 7 7 7 7 8 9 C C C C E E E E 9 9 9 0 0 G 5 5 5 5 5 5 6 6 7 C:\BX\BX-C\EC7\Dulin\[Dulin-results.xls] Figure 9 Emedded cantilever calculated ULS ending moment kpa 600.0-00.0 0 Pressure [kpa] Figure 0 Emedded cantilever calculation to DA, Comination
0kPa.5m Tie ar anchor GWL.m Sand 8,0m Water.0m D =? Design situation - Anchored sheet pile retaining wall for an 8m high quay using a horizontal tie ar anchor. Soil conditions - Gravelly sand - φ' k = 5 o, γ = 8k/m (aove water tale) and 0k/m (elow water tale) Actions - Characteristic surcharge ehind wall 0kPa - m depth of water in front of the wall and a tidal lag of 0.m etween the water in front of the wall and the water in the ground ehind the wall. Require - Depth of wall emedment, D - Design ending moment, M in the wall Additional specifications provided after the workshop: The surcharge is a variale load. The wall is a permanent structure. The length of the wall is to e the minimum allowale. Figure Example 7 Anchored sheet pile quay wall
Example 7 - Emedment depths 8 EMBEDMET DEPTH m. 7 6 5 4 * 0 0000000AA555777789D466BCCCCC 55555 Figure Example 7 Anchored sheet pile quay wall Example 7 - Bending moments BEDIG MOMET km/m. 600 500 400 00 00 00 0 * 0 0 0 0 0 0 0 A A 5 5 5 7 7 7 7 8 9 D466 B C C C C C 55555 Figure Anchored sheet pile quay wall calculated ULS ending moment
Example 7 - Anchor force 00 ACHOR FORCE k/m. 50 00 50 00 50 * 0 0000000AA555777789D466BCCCCC 55555 Figure 4 Anchored sheet pile quay wall calculated ULS anchor force.000 0.0 kpa.0 [] 0.7k/m -.000 Reduced Level [m] -4.000-6.000-8.000-0.00 -.00 [] -8.500 [] Toe -.74m -4.00 Water Pressure Actual Pressures -75.0-5.0-75.00-5.00 5.00 75.00 5.0 75.0 Scale x :60 y :60 Pressure [kpa] Figure 5 Anchored sheet pile quay wall design without redistriution
4.000.000.0 -.000 0.0 [] kpa 5.05 k/m -4.000-6.000 [] -8.000-0.00-8.500 [] -.00 Water Pressure Actual eff. Pressures -4.00 Passive Limit Active Limit -40.0-00.0-60.00-0.00 0.00 60.00 00.0 40.0 Scale x :54 y :78 Pressure [kpa] STAGE : Dig to full depth Figure 6 Redistriution of earth pressures Oasys FREW Figure 7 German practice for sheet pile design EAU (996)