Scribd Logo
Upload

Welcome to Scribd!

  • 12 views

    351 Earthquake Resistant Deisng Tied Back Retaining Structures... McManus PDF

    Uploaded by

    Tareq Al Shyoukhy

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content

    You might also like

    351 Earthquake Resistant Deisng Tied Back Retaining Structures... McManus PDF

    Uploaded by

    Tareq Al Shyoukhy
    12 views180 pages

    Original Title

    351-earthquake-resistant-deisng-tied-back-retaining-structures...McManus.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    12 views180 pages

    351 Earthquake Resistant Deisng Tied Back Retaining Structures... McManus PDF

    Uploaded by

    Tareq Al Shyoukhy

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    of 180
    AN Pacific e=== Geotech EARTHQUAKE RESISTANT DESIGN OF TIED-BACK RETAINING STRUCTURES EQC RESEARCH REPORT 06/517 Report to EQC Research Foundation ‘August 2003 Pacific Geotech Ltd Box 6080 Upper Riccarton Christehureh Principal Investigator: Kevin McManus PHO FIPENZ(Geotechnical & Stuctual) CPEng ABSTRACT This report considers design procedures fr ted-back retaining walls under carthquake loading. Tied-back retaining walls are becoming widely used in NZ.to ‘support permanent excavations on sloping ste in order to provide level building platforms for residential and commercial developments. They re also widely used to support excavations for roadways and oer key infrastructure, ‘Very lite guidance is available for the design oftied-back retaining wall to resist earthquake shaking. Lil observational data onthe behaviour of ted back walls during earthquakes has been published, but, what there is suggests that they behave wel, ‘A survey of New Zealand practice has showed that there ino consistency of approach and that most designers ar relying on a range of different “lack box” ‘computer software wih earthquake loading input simpy as an additional horizontal {orc applied dizetly to the wall. The appropriateness ofthis approech is {questionable because the fll range of difereat failure modes is not necessarily auldessed bythe software nor ist always obvious what the software does. In this study, a seismic design procedure for tied back retaining walls was synthesized based on an existing, widely used, semi-empirical design procedure for gravity design of ted-back walls. "The design procedure doesnot depend on specialist computer software, ‘The design procedure was testo by designing a range of case study walls and then subjecting them to simlated earthquakes by numerical time-history analysis sing PLAXIS finite element software for soil and rock. The response ofthe walls to 4 variety of real earthquake rcords was measured including deformations, wall bending ‘moments, and anchor forces. From the results ofthese analyses, it was observed that ll ofthe wall designs were robust and performed very well, ineluding those designed only to resist gravity loads. In some cases large permanent deformations were observed (ap to 400 mm) But hese were for very large earthquakes (scaled peak ground acceleration of 06g) In all’ cases the Walls remained stable with anchor forces safely below ultimate tensile strength. Wall bending moments reached yield in some cases forthe extreme earthquakes, ht his is considered acceptable provided the wall clement are detailed for ductility. ‘Walls designed to resist fow levels of horizontal sceslration (0.1 g and 0.2 g) showed Significant improvements in performance over gravity only designs i terms of | ‘permanent displacement for relatively modest inereases in cos, ‘Walls designed to resist higher levels of horizontal acceleration (03g and 0.4) shovwed edlitional improvements in performance but at much greater increases in cos, Even when walls were designed to resist 100 perent ofthe peak ground aocelerstion ofa particular earthquake record, significant permanent deformations wer still observed, {A tentative, detailed design procedure is provided based on the results ofthe study. ACKNOWLEDGEMENT This project was funded by the EQC Research Foundation under research grant EOC. 06/517. The support and patience of the Foundation is gratefully acknowledged Support ofthe University of Canterbury, Department of Civil Engineering i also ‘gratefully acknowledged, especially for acess to libeary facilites and forthe Senior Fellowship ofthe Principal Investigator. DISCLAIMER ‘This report describes a research project carted out nto the behaviour ofid-back retaining walls under seismic lotding, The conclusions and recommendations Contained within thi report are based on a limited investigation as described n detail inthe report. Pacific Geotech Limited and the Principal Investigator do not make any representations, express or implied, as othe acuraey, completeness, or appropriateness for use in any particular circumstances, of any ofthe information provided, requirements identified, or recommendations made inthis report. Pacifi ‘Geotech Limited and the Principal Investigator donot accept any responsibility for the use or application of, or reliance on any procedures or othe formation, fr any Pupose or reason whatsoever. CONTENTS ABSTRACT. eeoconnnnnnnn ACKNOWLEDGEMENT. os | DISCLAIMER... ‘CONTENTS. 1 Introduction, LL Overview. 2 Design Procedures ose 2A Ove RIEW nnn 22 Gravity Design nnn 22.1 Possible modes offre 2.2.2 Design procedure fo $80 snr 23. Seismic Design. 9 231 OVEEVW sso en 23.2 Mononobe-Okabe Equations. - see 23.3 Wood Procedure. ss 2.34 Comparison between M-O and Wo0d f1088 cweeeenssnnonnne 2 235 Practice in New Zealand 2 23.6 Synthesized Design proce. sent 3 Numerical Modelling of Case Std... 15 3.1 tntodeton. - sen 32. Methodology sn sence 33 Time Histories sen 16 331 Overview. ool 332 Sealing factors... a 1s 34 Case Study 1: Single Row of Anchors in Sand. ssenwnennnnedd B41 Case study description. iv 342 343 345 346 347 348 349 3410 341 3412 3413 344 341s 3416 3417 3418 3419 5 Case Study 2 Two Rows of Anchors i 351 352 383 354 355 356 357 358 (Case Ie: Gravity design Performance of Case 1a under gravity and pseudo-satic loading valuation of ase 1a under gravity loading Performance of sravity design Case 1a unde seismic onding (Case It: MO based design 0. g. Performance of Case Ib under gravity and peeudo-statc loading. Perforrance of Case 1b unde seismic loading Case Le MO bated design 0.2 5 Performance of Case Le under gravity and pseudo-statc loading, Perfomance of Case Le under ssismic oaing Case 1 M-O based design 003 g Performance of Case 1d under gravity and pseudo-satic loading Performance of Case Id under sssmic loading (Case le M-O based design to 04 g Performance of Case 1e under gravity and pecudo-siatic loading, Performance of Case le under scismic loading... Comparson of design cases. Conelusons Case Study Description Case 2a: Gravity design Performance of Caso 2a under gravity and pseudo static loading. valuation of Case 2a under gravity lading, Performance of Case 2a under seismic loading (Case 2: M-O based design 00.1 g. Performance of Case 2b under gravity and pseudo-sttic loading. Performnce of Case 2 under seismic loading 26 26 30 31 33 36 38 Al 42 a7 48 49 32 3s 56 56 37 38 a “1 6s Co 359 Case 2e:M.O based design to 02g 3.510 Performance of Case 2c under gravity and pseudo-tat 35.11 Performance of Case 2c under seismic loading 3.5.12 Comparison of design cass. - 3513 Conclusions. 3.6 Case Study 3: Two Rows of Anchors in Sand with Extended Anchors 3.6.1 Case Study Description. - 362 Case 3a: Gravity design 3.63 Performance of Case 3a under gravity and pseudo-static loading 364 Evaluation of Case Sa under gravity loading 3.65 Performance of Case 3a under seismic loading, 3.66 Case 3b: M-O based design to 0.1 g 3.6.7 Performance of Case 3b under gravity and pseudo static loading. 3.68 Performance of Case 36 under seismic loading, 3.69 Case 3e: M-O based design 100.2 g 3.6.10 Performance of Case 3e under gravity and pscudo-static loading 3.6.1 Performance of Case 3 under seismic loading 3.6.2 Comparison of design cass 3.6.13 Conctusions. 4 Design Guidetins. - 41 Overview. 42 Seismic Design Accelerations 43. Proposed Design Guidelines for *Sand” soils Summary and Conclusions... 7 6 Recommendations for Future Reseach 7 References. 7 Appendix A. vo - o n 14 78 7” » 9 8 82 82 85 86 87 on 9s 97 98 98 100 104 108 109 |A.1_ Design calculations for ease study Sand 1a ~ Gravity based designs. 11 [AZ Design calculations tor case study Sand Tb —M-O based design Dl gou119 [AB Design calculations for ease study Sand 1e—M-O based design 2 g..126 ‘AA Design calculations for cas study Sand 1d-~ MLO based desig 3g... 133, ‘AS. Design calculations for case study Sand 1e~M-O based design 4g. 140 Appendix B - M47 B.1 Design calculations for case study Sand 26 Gravity based design nu 147 B.2 _Dasign calculations for ease study Sand 2 —M.O based design .1 gu. 154 1B _Dssign calculation for case stidy Sand 25 ~M-O based design 02 g...161 APPEDEIK Conn 168 CA Design calculations for case study Sand 38 ~ Gravity based design. 168 C2 Design calculations for case study Sand 3b —M-O based design 0.1 g..170 (C3___Design calculations for case study Sand 3e~ MO based design 0.2 g...172 ac 00477 Tied Back Retsning Wal 1 Introduction ‘Kramer [1996] has sunmatised the limited rescarch available on this topic. Very few reports of the behaviour of ied back walls during earthquakes are available, Ho el [1990] surveyed ten anchored walls inthe Lot Angeles area following the Whitir earthquake of 1987 and concluded that they performed very well with litle or no loss of interty. ‘Numerical analyses of tiod-back walls have been performed by Siler and Frawley [1992] and Siler and Dolly {1992} who found that walls with sf, more closely spaced anchors develop smaller and more uniform permanent displacements than walls with softer anchors and greater vertical spacing of anchors. Walls designed for higher static earth pressures were al found to develop smaller permanent Aisplacement than walls designed to lower static pressures, Walls with higher initial anchor preloads were found to develop smaller permanent displacements than walls ‘with lower preloads ‘ragaszy etal. [1987) found that wal lements thet extend int the foundation soils ‘may be subjected to very high bending moments atthe base because of phase sierences in movements between the top and bottom ofthe wall, Inclined anchors extending below the base of the excavation may become highly stressed when the bonded end ofthe anchor embedded in soil moves out of phase with the wall face. Daaaited design guidance as been provided by Sabatini et.al. [1999] within a general design manual for tied-back walls prepared forthe US Department of Transportation, Federal Highway Administration, This manval is i wide use within the US and is ‘ining increasing acceptance within New Zealand. They recommend the use of the seudo-static socalled Mononobe-Okabe method [Okabe, 1926; Mononobe and “Matsuo, 1929} to ealeulat earthquake induced active earth pressures acting agunst the back face ofa tied-back wall. seismie coefcient from between one-half to ‘wothids ofthe peak horizontal ground acceleration (0.5 PGA to 0.67 PGA) is recommended to provide a wall design that will Knit deformations to small values acceptable for highway facilis Sebati et.al. [1999] recommends that brite elements ofthe wall system (the ‘grouttendon bond) should be governed bythe peak ground acceleration “adjusted to account for the effect of local soil conditions andthe geometry of tae wall” and a ‘actor of salty of 1.1 applied. Design of ductile elements including the tendon, should be govemed by the cumulative permanent seismic deformation. They recommend that, based on studies using Newmark type sliding wedge analyses, ductile elements shouldbe designed using forces calculated by pseudo static analysis using a seismic coefiient of 0.5 PGA with a factor of safety of Il applied, The length ofthe ground anchors may need tobe increased beyond that calculated for static design with the anchor bond zone located outside of the Mononobe-Okabe active wedge of soll The use ofthe Mononobe-Okabe method to calculate earth pressure for design of ted back walls has the advantage of being straightforward and is widely used for design of| gravity retaining wall. However, itis based on limiting equilibrium and the ac ower? ‘led Back Retaining Walls ‘August 2008 evelopment ofan active failure wedge of soil that tat odds with the design procedure for satc loads for tied-back walls. The recommendation to plae the bond Zone ofthe anchors behind the active soil wedge means that the wall Isat ree to ‘move withthe vedge, as assumed by the Mononobe-Okabe procedure 1.4 Overview This project has studied the performance of ied-back retaining walls by us of| ‘numerical time-istory analysis using PLAXIS finite element software for sil and rock [Brinkgreve & Vermeer, 1988). Too few field studies frm actual earthquakes are available to make meaningful conclusions and testing of sealed down models on shaking table is of limited use because ofthe impossibility of satisying scaling laws without increasing the gravity field ina centrifuge, Numerical analysis of problems in _geomechanies his become a recognised to! Tor exploring soil-stracture infraction problems and isprobably the only practical way to investigate the complexity of tied- ‘ack wall behaviour during earthquake shaking The project has ‘oeussed on developing rational and practical design procedure then ‘verifying the prcedure by considering different casestudies oftied-back walls. "The case study walls were designed using the proposed procedure and then subjected to Afferent cartnquaketime-histories using PLAXIS, The performance of each wall design was assessed foreach earthquake by monitoring various key paramere including displacement, wall ending moments, and anchor forces. Alter assessing the performance ofthe various wall designs, the proposed design Procedure was citically assessed and final guidelines and recommendations made Every wall design case in practice is different in some way fom every previous design. Ttwas impossible within the constraints of time and budget to consider every possible wall circumstance. Instead, the ease studies were based onthe simple, ese ‘off deep uniform sand sil deposit with suitably reneric properties. This, simplification is both necessary and desirable because it allows the basic trends in wal performance tbe observed without “lutier" fom a myriad of different parameters, Atte commencement ofthe project survey was undertaken to identify available published design procedures and to identify current New Zealand practice. This information was ised to identify the most rational design procedure and to clarify and refine such a procedare as necessary. The case study designs and analyses then were Undertaken to prove or otherwise the efficacy and safety ofthe desi procedure, 00 06477 Td Bock Retaining Was ‘August 2008 2 Design Procedures 24 Overview Tied-back retaining walls were usod originally asa substiue for braced retaining alsin deep excavations. Ground anchor te-backs were used to replace braving struts that caused congestion and constuction difficulty within the excavation, Design procedures evolved ftom those developed for braced excavations and are ‘typically based onthe so-called "apparent earth pressure” diagrams of Terzaghi nd ‘Peck [1967] and Peck [1969]. These diagrams were deve oped empirically from ‘measurements of loads imposed on bracing strus during cep excavations in sands it ‘Bedin, Munich, and New York; in sotto medium insensive glacial clays in CChieago; and in sof to medium insensitive marine clays ia Oslo, ‘These original “apparent earth pressure diagrams” were nit intended by the authors to be a realistic representation of actual earth presses agaits wall butt be :merely an artifice for calculating values ofthe strut lous that wil not be exceeded ‘nny real strut ina similar open cut. In general, the bencing moment in the sheeting ‘oF soldier piles, and in wales and lagging, will be substantally smaller than those calculate from the apparent earth pressure diagram suggested for determining strut loads."Terzaghi & Peck, 1967], Since 1960, remarkably few significant modifications to this original work have been adopted in practice. More recently, Sabatini et. l. [1999] proposed a more detailed design procedure based on the apparent earth pressure apyroach intended specifically for pre-tensioned,ted-back retaining walls in comprehensive manual prepared for the US Department of Transporation, Federal Highway Administration. ‘This manual is in wide use within the US and is gaining increasing acceptance within New Zealand. ‘A detailed and well proven design procedure for walls und gravity loading is given in this maowal which wall be referred to doughout this report asthe “FHWA. procedure". The manual also makes suggcstions for desigh of tied-back wall to resist ‘earthquake loading although a detailed procedure isnot given, ‘Increasingly, practitioners are relying on computer “bleckbox” software to design ticd-back walls with methodologies that range from fully dlastc "beam-on-elastic foundation” approaches o limiting equilibrium spprosches. Caution is required when using "black box” software to ensure that all posible failure modes have been considered 22. Gravity Design 22.1 Possbie modes offre Possible modes of failure for ted-back retaining walls re lusrated in cartoon fashion in Figure 22.1 (a), A complete design procedure neds to address each of. these modes of failure ac oar Tied Back Retaining Wate ‘August 2008 8) Tensile faiture of tendon: The range tendon loads mast be established ‘with suitable margins for saety ') Grout/ground bond failure: Generally this should always be established on site by proof testing given the diffenlty in prdicting the capacity and the ‘dependence on installer skill and technique ¢) Tendon/grout bond failure: anchor details Prevented by reference o provenlcommercial 6 Wall bending failure: Actual wall moments are very dificult to predict, ‘because ofthe interaction betwoen sil and structure stiffness and he non Tinearity of sll stiffness. However, wall hinging does not necessarily create a :mechanism provided the wal element i ductile, €) Passive failure at foot of wall: Insufficient embedment depth fr cotimious als or soier piles leads to pasive failure of the soi immediately infront of| the wall and instability ofthe wall and soil mass, 1). Forward rotation of wall: Staging of excavation i necessary to prevent forward rotation of wall prior to anchor installation. Wall needs sufficient ‘bending strength to resist cantilever moments for aged excavation. Anchors need tobe of sufficient capacity and length to prevent forward rotation, 8) Bearing failure underneath wall: Cansed by downwards component of anchor fore, Check axial eapacity of solder piles, or, bearing capacity of foot of continuous wall. Bearing loads may be reduced by reducing the anchor inclination a mush as possible (15 degres isa practical minima), 1) Failure by overturning: Essentially same as (9), Anchors need tobe of sufficient capacity and length to prevent forward ration, i) Failure by siding: Possible mode for cohesionless soils. Factor of safety controlled by increasing depth of embedment of wall andlor soldier piles. Fector of safety calculated using limiting equilibrium “wedge” analysis, J) Failure by rotation: Possible mode for cohesive oils. Factor of safety controled by increasing depth of embedment of wall andor soldier piles Factor of safety calculated using limiting equibirium “Bishop” analysis similar E06 0877 “Tied Back Retaining Walls Aoust 2008 t+ ts 3 — ioarer 0c 06477 Tied Back Retaring Was August 2008 Figure ??1 (a) Possible modes of failure for ed-back retaining walls [Sabatini et. al, 1999}, 222 esgn rececuretor sana ‘The following procedure addresses ach ofthe above failure modes systematically (forthe gravity load caso) and is based on the FHWA procedure with minor ‘modifications and clarifications where noted, Its assumed herein that the all and retained solar fully drained. Ths procedure is intended tobe readily calculated by hand, although use of calculation sofware such as Mathcad ot Excl wil be useful for design iterations. Example calculations using Mathcad forthe casestudies are inelude inthe appendices, 8) Initial trial geometry: The depth of excavation and depth to each row of anchors needs tobe estimated asa first step, based on experience or tral and errox. Typically for stronger soils, the frst row will beat «depth of 2m with subsequent rows at Sm intervals © gt pnp gs Aston ise 22200 Nate oie, =b6-8] ‘ein ae rss va a se ey ‘CaneePic ta tape etfs on Sih ta paneer set yc ee a on ) Calealate wail base reaction, R: As shown in Figure 2.2.2 (2) ©) Calculate wall setion bending moment: From the apparent crth pressure diagram as shown in Figure 2.22 (b). These methods are considered to provide conservative estimates of the calculated bending moment, but may ‘ot sccuratly predic the exact locations ofthe maxima, FHWA, document recommends an allowable sess of Fy » 0.55 Fy fr stel soldier ples. For "New Zealand design procedures using load and resistance factor design (LRED) principles and fora strength reduction factor for stel sections of 0.8, an equivalent load factor of @ ~0.8/0.55 = 1.45 is implied, However, for Consistency with NZS 4203 (sce discussion elsewhere) s load factor of 16 was !sopted fr this study forthe parpose of sizing wall structural elements 1, Determine depth of embedment: Calculate required depth of embedment for soldier ples to resist wall base reaction (R) using Broms (1963) or similar, or, or continnous walls using pasive resistance fom Coulomb theory or log spiral theory such as NAVFAC DM-7. FHWA document recommends a factor of safety of 1.5 for these calculations. For this stdy, a strength reduction factor of3 is applied tothe Broms [1965] formulation because ofthe lange plastic strains required to mobilise the fll passive resistance. Use ofthis ‘eduction factor was found to give realistic embodment depts consistent with avoidance of wedge failures and better contol of displacements coc oar? Tied Back Retaining Wal ‘August 2008 £8) Check internal stability of the wall: A possible intemal failure mechanism is shown in Figure 2.2.2 (c), with an active failure wedge immediatly behind the wall, a passive wedge immediately in font ofthe embeded toe of the all, andthe anchors) developing their ultimate capacity (aken tobe the proven ext pacity, normally 1 3 times the design load or 80 percent ofthe anchor tensile eapaciy), Te tie factor of safety shouldbe determined by reducing the assumed soil strength progressively inthe calculations until the driving and resisting forces are just equa, ie Active force ~ Passive force + anchor ultimate force when the factor of safety against ling is given by wan") 1" Praa) FS Aniterative procedure i required to make this calculation s shown in ‘Appendix A using Mathcad, [No specific guidance on suitable factor of safety is given in the FHWA. document but FS> 1.3 for gravity loading would seem tobe a sensible value, a ‘Check external stability ofthe wall: Extemal stability of iod back retaining alsin sand is controlled by horizontal sliding ofthe wall wit formation of ln active soil wedge behind the wall and a passive wedge in font ofthe wall base, a shown in Figure 2.2.2 (c). The critical future surface i assumed to ‘pass immediately behind the anchor bond zone, a shown, The same procedure was adopted for evaluating the factor of safety as, described in ) above. [No specific guidance on suitable factor of saety is given in the FHWA, ‘document but FS > 13 for gravity loading would seem tobe a sensible value, ‘August 2008 (2) Wate wih one trad wt tpl fra roround ahve Ow crarandaneton Figure 2.2.2 (a) Apparent earth pressure diagram for sand. [Sabatini et a, 1999] My = 2h we Morante nine ane My =2% og pg sor 8 Me (2) Wate wt nave goed chore (2) a ty mle nl round anchors Figure 2.2.2 (b) Method for estimating wall bending moments for sand. [Sibatin et al, 1999), 0c 06477 Tie Back Retaining Wal August 2008 Ineoenal Staby / A Fig 2.22 (¢} Internal and extemal mechanisms for tod back walls 2.3. Seismic Design 23.4 Overview Little euidanceis availabe forthe design of ted-back retaining walls to resist seismic actions. Gravity retaining walls are normally designed using a preudo-static approach: ‘The active wedge of soil immediately behind the wall hasan edition pscudo-statc ferce component equal tothe mass of soil within the wedge maltipied by acceleration. Typically, the resulting forces ae resolved to derive anew critical wedge geometry and necessary wall pressure to achieve equilibrium, a in the Moncnobe-Okabe (M-0) theory [Oksbe, 1926; Mononobe and Matsuo, 1929} For retaining wall that are rigid and unable to move sufficiently 0 allow sol yielding and development of a Rankine condition behind the wall (eg busied basements), ‘theoretical linear elastic solution for soil pressure derived by Wood {1973} is normally used to calculate dynamie soil pressure ‘These two approaches represent, perhaps, an upper snd lower botnd of what the resulting dynamic soil load might be aginst a ied-back retaining wall ‘The only published advice specific to design of tied-back retaining walls was found within the FHWA manual [Sabatini et.al, 1999]. FHWA recommend use ofthe ‘pseudo-static Mononobe-Okabe (M-O) theory to design tid-back retaining walls but ‘donot give a detailed procedure. Nor is such a procedure obvious because the recommended design procedure for tied-back walls under gravity loading is based on empirical “apparent earth pressure” diagrams. 0c 0677 Tied Back Retaong Wall ‘August 2008 ‘The FHWA manual states thatthe design of brite elements (¢. the growttendon ‘bond) shouldbe governed by the peak force (1c. corresponding to peak ground acceleration, PGA). Design of ductile elements (eg tendons, steel shee piles, soldier piles) should be govemed by cumulative permanent seismic deformation, ori ew of such analysis, design shouldbe based on 0.5 times the PGA. However, no advice is given aso haw the “peak force” might be calculated Given that the anchor tendons are, effectively, long springs with litle mass then there Seem no reason wit they should be subject to high peak forces and should respond only to elongation from gross movements within the sil mass. Neither of the formulations (Wood or MO) for calculating wall loads during shaking, take any account of the flexibility ofthe wall and the likely kinematic effect and soil 23.2 Mononabe-Ohabe Equations The M-O equations ae an extension ofthe Coulomb equations based on ‘considerations of equilibrium ofa triangular shaped active (or passive) wedge of soil interacting with sliding wall. The important assumption is made thatthe soli, yielding in shear aloog a planar failure surface at the base ofthe wedge with ‘solution then made ofthe resulting force polygon as shown in Figure 2.3.2 (a). An ‘equivalent equation exists forthe passive ease, but, a fr the Coulomb equation, its inaccurate for walls wth fiction yg =» ———¢-8- _ coxy ora cou-+0+y)[1-+ [SRE snl] v aw Ai Figure 2.3.2 (a) Mononobe-Okabe equation for active case (Kramer, 1996] 239 Wood Procedue ‘Wad [1973] developed a procedure for estimating dynamic load azainst smooth, ‘gid walls based on en assumption tat the soil remains linea elastic and tha the wall is completely rigid. While not intended originally for tied-back retaining walls but for rigid basements and tie lke, this procedure might be considered to given an “upper ue 0647 Tied Back Retaining Wate ‘August 2008 bound” ofthe sil pressure that may develop for any given horizontal aceeleration against the fice ofa retaining wall, ‘The dynamic component of thrust and overturning moment respectively are given by the following equations An aL, AM, =P HE inwhich hy ~ brizontal acceleration a8. a proportion ofg, and Fy» Fae factors siven in Figures 2.3.3. (a and (b) below. The ratio /H in the Figures refers to length, Zin the horizontal direction for soil contained within a rigid box of dept, H, that was modeled by Wood. For tied-back etining walls, L/ should be assumed to be infinite The point of effective application ofthe dynamic soi load is ata height above the base ofthe wal given by: Figure 2.33.1 (2) and (6) Dimensionless thrust factor and moment factors, After ‘Wood (1973] ac 00477 Te Back Retain Was ‘August 2008 234. Comparison between MO and Wood factors ‘The addtional nominal wall loading caused by a pscudo-static horizontal acceleration ‘was calculated using ether the M-O or the Wood equations as shown in Teble 2.3.4 (@) forthe case of walls in sand with 435 dogreos Table 2.3.4 (a) Comparison of nominal wal leading caused by pscudo-static ‘scceeration ——_____** — Horizontal Acceleration Wood! Mononobe-Okabe hy a= he v=03 6-35 on oa 0.06 02 02 043 03 03 021 04 04 ost At lower levels of acceleration, the M-O equation gives about ¥ the load ofthe Wood equation, increasing t0% a 04g, The M-O equation is expected to give much lower Toading because it assumes that soil shear strength is fully mobilised to resist the acceleration 238 Pracioe in Mew Zealand Given the paucity of guidance in the literature, it was decided to conduct a survey to find out how practitioners were designing tied-back walls to resist earthquakes in current practic, Current practice in New Zealand was surveyed by conducting a series of personal interviews with senior staf inthe largest practices and also fom the author's ‘experience in numerous design reviews. Lite consistency in approach was evideat, With most espondents relying on “black box” computer software that doesnot specifically consider earthquake loading. ‘The most commonly used software package is “\WALLAP” [Copyright 2002, DL Borin, Geosolve, UK]. This software combines limiting equilibriam analysis to British and European standards to compute factors of safety coupled with I-D “beam on elastic foundation” or finite element analysis to compute wall element stresses and deformations. ‘Parthquake “loads” are typically being input as static Toads applied to nodes. The calculation ofthe pesudo static loads are made sing citer the M-O equations or the ‘Wood [1973] analysis according othe judgement ofthe designe. ‘Typically, the fre length ofthe anchors ae located aecording tothe inclination of the Coulomb, gravity only active soil wedge, with no increase to allow for the flattening ofthe active wedge under acceleration (at last one major consultancy). 2 EOC 06477 Ted Back Retaining Wals ‘August 2008, (Note: A new version of “WALLAP" has recently beca released which allows input ofearthquake accelerations drecly, althouph the methodology for computing ‘earthquake responses not now). 2386 Smibesized Dasign procedure ‘With no detailed procedure for the design of ted-back wall o resist earthquake loading available, it was necessary to synthesize a rial procedure. A procedure was synthesized based onthe FHWA procedure for gravity loading by epplying the following rationale, 1. ‘Sine the apparent earth pressure used for wall design in gravity loading is caleulated based on Ky, the Rankine coefficient of active earth pressure, simply substiute Ky, the M-O coefficient of active earth pressure under earthquake accleraton to ealeulate an equivalent apparent earth pressure for the earthquake design case 2. Anchor free lengths are normally extended to beyond the location ofthe ‘Couomab ative wedge slip plane when designing ted-back walls for gravity Toading. Therefore, extend the anchor free length to beyond the equivalent M- (slip plane for enrthquake loading 3, The MO equations should also be used when checking the extemal stability ofa wal ‘The following detailed procedure was adopted on a tral basis fr the case studies examined inthis project. Based onthe results ofthe time history analyses, additonal ‘minor recommendations and improvements were made and thse are included in the final eeommended procedure of Section 4 1) Initial trial geometry: The depth of excavation and depth to each row of, anchors needs tobe estimated asa fis step, based on experience or trial and eror. Typically, for stronger soils the fest row willbe ata depth of 2m with subsequent rows at 5m interval. b Prepare apparent earth pressure diagram: As shown in Figure 2.22 (a), "Note that Kis calculated using the M-O equation withthe selected design ppseudo-satc acceleration. The wali assumed tobe fctonless (ie. the wall is likely to move downvvards with any active soil wedge), 6), Caleutate anchor design load: As shown in Figure 2.2.2 (3), 4) Cateutate wall base reaction, R: As shown in 2.2.2 (a) ©) Calculate wal section bending moment: From th apparent earth pressure diagram as shown in Figure 2.2.2 (6). load factor of 1.6 is recommended for the purpose of sizing wall structural elements using New Zealand standard, 1) Determine depth of embedment: Calculate required depth of embedment for soldier piles te resist wall base reaction (R) using Broms {1965} (bat 1B 0c 06477 ‘ied Back Retaining Walls August 2008 calculating Ky using the M-O equations), or for continuous walls using passive resistance from M-O Okabe theory. strength reduction factor of 3s commended to be applied to these calculation because ofthe large plastic strains required to mobilise the full passive resistance, Use of this reduction factor has been found to give realistic embedment depths consistent with avoidance of wedge failures and better control of displacement, 1) Check internal stability ofthe wall: A possible internal failure mechanism is shown in Figure 2.2.2 (0), with an active failure wedge immeiately behind the wal, passive wedge immediately infront ofthe embeddedtoe ofthe wal, and the anchors) developing their ultimate capacity (taken o be the proven, test capacity, normally 1.3 times the design load or 80 percent ofthe anchor tensile capaci), The true factor ofsefety may be determined by progressively recucing the sssumed sil strength inthe calculations until the driving and resisting forces are just equal, ie Active force = Passive force + anchor ultimate force ‘when the factor of safety against sliding is given by: 2 1" Byated) For the earthquake load case using pseudo-statc design, a minizum factor of safety of 1.1 is recommended, but not less than the factor of safer against extemal stability 1) Set “free” length of anchor tendons: The “sree” length ofthe eachor ‘encons should extend beyond the active soil wedge defined by the M-O theory and originating atthe base ofthe wal or the embedded soldier ples as indicated in Figure 22.2 (@). §) Check external stability of the wall: Extsmal stability ofied-tack retaining ‘walls in cobesionles soil is conrolled by horizontal siding ofthe wall with formation of an ative soil wedge behind the wall anda passive wedge in font ‘ofthe wall base, as shown in Figure 2.2.2). The ential failure surface is ‘assumed to pass immediately behind te anchor bond zone, as shown. For the earthquake load ease using pseudo static design, a tinimam “true” factor of safety of 1.0 based on mobilised soil shear strength is recommended. i) Note: When calculating passive soil resistance, the interface friction angle should beset be no more than #/2. Use of higher values i nat recommended because the resulting values of passive resistance will be unrealistically high, ac osi77 Tied Back et ring Wals ‘August 2008, 3 Numerical Modelling of Case Studies 3 Introduetion [No good case study data is available regarding the performance of ted-back retaining ‘alls in real earthquakes. No data was found for physical model studies for ed-back ‘eiaining walls in simulated earthquakes. Modelling of geotechnical systems is sificut,n any case, because the laws of physical similtude require that model experiments be caried out either at very large sea, or, at small scale under high accelerations in a centfuge Numerical modelling in geotechnical engineerng has become an sccepted research ‘tool and is viewed as a practical substitute to physical modeling for many problems. For study of tied-back retaining walls under earthquake loading, mimerical modeling ‘may be the only practical method fr realistic simulation given the complexities ofthe ‘wall consiuetin. For this study, two representative tied-back wall designs have boca modelled ‘numerically: A simple wall with one level of e-back anchors and a more complex ‘wall with two levels of anchors. Simplified sol eandltions have been chosen tobe representative of real conditions. Obviously, in practice, much mere complex strtigraphies are likely tobe encountered, bute objective herein ito gain understanding ofthe fundamentals of wall performance without introducing confasion fom complex stratigraphy. Detailed design of the walls was made in accordance with the tril design procedure with slight variations andthe performance ofeach under both static gravity and seismic conditions was determined using PLAXIS finite clement software fo soil and rock mechanies [Brinkpreve & Vermeer, 1988). Earthquake performance wae

    You might also like