You are on page 1of 42
Monday, December 14, 2015 International Short Course on Design and Assessment of Mine Waste Structures (Tailings Dams and Waste Dumps) University of Alberta, Edmonton, Alberta Deformation Analyses Bill G. Chin, M.Eng,, P.Eng. Principal, Kohn Crippen Berger Ltd @)xom Crippen Berger @ acaamateo “Deformation Analyses” — A fortuitous and befitting subject matter for me. Ina lot of ways, it has helped shape my career..... assess @mbarked on a life-long journey of continual learning! 2015-12-03 2015-12-03 ome Early Experiences Late 1970's - foundation design for pulp and paper mills: + Structural engineer had a One question | asked on my “standard” design-basis booklet, _very first assignment: which specified allowable + Have you ever measured actual differential settlements for total and differential settlements ifferent buildings. to validate criteria? + Allowable diferential setiements {at the time) typically range from 1:1200 (high sensitivity to 1:40, (warehouses, ete) 1979 to Early 1980's - foundation design for expansion at Alcan aluminum smelter in Kitimat?, B.C. + Upto 100+ m of normally consolidated marine silts and clays + Original development in 1950's required up to 10 m of site grading fils + Site grading and preload/surcharge fils placed hydraulically + impractical to eliminate post- construction settlements + Settlements upto 2m (~1883 to the ‘Seen mean ane time of expansion project ~1978-1980) + Provisions for settlement monitoring and allowance for re-levelling footings and column bases Eg eee © rem crinpen tocar Early Experiences, For expansion project: + Site had been used for waste stockplles up until~1978 + Ces and 6's calibrated to settlement data + The issue: + can we take advantage of efectve preloading ofthe foundation sols? + how do the new building stresses compare to the effective preload stresses from waste ple? Callorotes parameters used to estate ‘Approach: lective stresses + Airphoto of different vintages used sn to estimate time-rate of loading history from waste pile. + Relative heights ofthe stockpile were estimated from air photos by the "paralax” method rote cibwen sae rly Experiences + The project was one of my first exposures to using instrument data to improve design assumptions and predictions + Supplementary thoughts: + Original design settlement predictions (1952) based on: + Stratigraphy from boreholes + Lab consolidation tests to determine compressibility parameters + 1956 — trending towards higher settlements than predicted; estimates revised based on settlement gauge data + 1959 trending towards higher settlements than 1956 predictions; revised based on “curve-fitting” actual data An early example of adopting what Dr. Morgenstern has now affectionately termed “a performance-based” approach to the design of oil sands tailings dams “eves ‘ vere crppen terger 2015-12-03 Early Experiences First 9 +/- years of my career, “deformation analyses”, primarily meant: + Vertical deformations (i.e. settlements), magnitudes and time-rate. + Fairly “intuitive” mechanism and theory + Fairly simple calculations Mid to Late 1980's: + Areal “step change" in the types of deformation analyses and comprehension + Fueled by geting my Masters degree from the University of Alberta * Forty Mile Coulee Case Study; specifically East Dam \& eens me > @)ehe crpeen teres Case Study — Forty Mile Coulee East Dam ‘in 3, Oavson, OM, ld, son RG. and Campdel 1.1981} Cosrvton Performance the Fam ise Cause tant am ana Soh Cay Founda. Proceedings, caradian Getechncsl Contre, 2. Gin 1994 Same Consiferaions nthe Design and Performance of ath Dams on Sot ly Found, ae Taio Project was Papeete es ama Hcy constructed in Smee cota ise 1986 and 1987 ))Kionn crippen Berger 2015-12-03 Issues + Dam was built in two stages: + High excess pore pressures at end of Stage 1 exceeded original design predictions by up to 15%; led to design and construction of enlarged toe betm prior to Stage 2. (Overall slope flattened from ~GH:1V to ~10H:1V) + Atthe end of Stage 2: + high excess pore pressures exceeded original design presictions by up to 35% + Slow rates of pore pressure dissipation + Relatively large deformations in foundation clay + Upto 240 mm horizontal movements + uptoxoomm semtements + FOS = 1.5 may not be reached + Design Factor of Safety (FOS) for several years because of slow pore pressure dissipation, en + Will reservoir filling without * Operating Conditions: 1.5 meeting FOS = 15 criterion jeopardize di integrity? eantet pee + rvmctepen terger Dam Section and Foundation Ss ein ns we @uerm cippen Berger 2015-12-03 2015-12-03 East Dam Instrumentation eeeeEe aT) » (10m crppen terger Why Finite Element Modeling? - FOS = 1.5 may not be reached * Design Factor of Safety (FOS) for several years because of + End-of-construction: 1.3 slow pore pressure dissipation, “ : + Will eservoir filing without jeopardize dam integrity? * Objective: ~ + Provide better understanding of factors affecting foundation clay behavior + Provide a more confident appraisal of dam stability under reservoir loading * Approach: * Calibrate finite element model (FEM) to “match” measured behavior during construction + Use calibrated model to assess future behavior upon reservoir filling nee ne esas (remcieoenteser 2015-12-03 2015-12-03 Historical Interlude The finite element work was carried in the field office during the latter part of construction. Following is extracted from Report to the Alabome Highway Research Centre, dated September 1958 (review of + FEADAM 84 was the finite element avaiable geotechnical engineering program used software}: eee RA ee coun geaany $400 * Finite element analysis of dams. Analyzes stresses ond deformations inthe Incremental constuction ofan eorth ‘embankment + Horaware Requirements: 192K; Dos 2.1 0F ote; coprocesser; 2s dives ‘recommended; 132 calunm printer gr0phies monitor ) ora cippen eae & woot mums “ypcal computer infield office Items shown are NOT the ones actually used, but are reasonable Floppy dlsk for data storage facsimiles Printout of output | ee teed “ rer cipen terger Constitutive Models * Constitutive Models: + Both dam fill and foundation clays are quite ductile and were simulated by a hyperbolic model + Starting parameters derived fom ‘ndrained compression vial ests + Varied to “match? measures ‘behavior during construction @)ner» ceppen serge Typical Results and Comparison: ee a ee eee 2015-12-03 2015-12-03 ‘+ Reservor-induced incremental “immediate” movements are not significantly disproportionate to inereased loading, even ifthe reservoir pressures are assumed to be doubled. + Importantly, the “yielded” zone is adequately contained under the reservoir loads @ter cieoon sage 10 2015-12-03 me from East Dam FEM Study, + Dr. Morgenstern was retained to provide external review of the results + Although limit equilibrium FOS < 1.5, the results of the FEM led to the conclusion that reservoir filling can proceed on the following basis: + Fill the reservoir in controlled stages * Close monitoring of pore pressures and movements during and after each filling period. * Ensure deformation response displays attenuating movements. following each fling stage. & encase 2 ror cippen seer Many of the learnings from Forty Mile Coulee Project have been applied, and advanced, on other water and tailings dam projects involving soft clay foundations over the past 25 years BUT...Let’s fast forward to 2012 — Alameda Dam Case Study i Case Study — Alameda Dam * Excessive runoff in the spring of 2011: * Reservoir surcharged ~5 to 6 m above FSI. (to near MAFL) to reduce d/s flood flows; + Increased rates of shear displacements in the clay shale foundation * Interim stability analyses (Sept 2011): + Minimum factor of safety near 1.0 @ roma rio, M. and carmel). 2044), Geotechnical Assessment of Alameda anaglan Dam Assocation Conference, Sant Alberta, October 4109, 2018 sue of Concer: Kohn Crippen Berger * Recommendations were: + Lower reservoir (to FSL or lower) ASAP + Carry out site investigations ASAP * Conduct detailed assessment ASAP 1 Qnern cece 2015-12-03 12 Phased Approach to Study Phase 1: Preliminary Stability Analysis * Use existing information * Establish operating restrictions for 2012 if required Phase 2: Background Review and Knowledge Gap Identification Phase 3: Field and Lab Investigations Phase 4: Detailed Dam and Structure Evaluation Phase 5: Develop Remedial Options Notes: 1 2 Project commenced eatly 2012 and was completed eaty 201. ‘Work inctuded evaluating impact of engoing deformations on performance ofthe structures (U0 and spliway ‘This lecture is focused onthe main dam Phase 5, develoing option to improve dar stably, was not necessary )) kone Crippen Berg Alameda Dam = Main Features ecu s Q) Wm ctopon Base 2015-12-03 13 2015-12-03 Foundation Conditions © Braggerated Seale (H=5¥) hn Crippen Berger 14 Foundation Conditions — Glacial Till Korn crppen serge 2015-12-03 Instr Overview of Historical Performance of Dam * Excellent account is provided by Mittal and Rahman (2000)*. Overview summary: + Mid-1991 to 1995 construction period + Unanticipated large shear displacements in clay shale (max. 400 mm @ end of construction) and higher-than-expected excess pore pressures in glacial till (B of 0.7 to 0.85) and clay shale (B of 0.9 to 1.0) + Stopped construction for 18 months: + Dam section was re-designed -> upstream and downstream buttress added + Remainder of dam (EI. 558 to 568.5 m) was completed in a number of carefully controlled and closely monitored stages + Reservoir filled to FSL in 1999. Kiohn Crippen Berger 2015-12-03 16 2015-12-03 Foundation Till Pore Pressures B Hi U *+ High excess pore pressures during construction + Very slow post-construction dissipation; still high pore pressures today + Responds to total stress change from reservoir— response attenuates rapidly in downstream direction )) Kohn crapen Berger 7 CPamoiociaeniee ) worn cipwon + High excess pore pressures during construction + Very slow post-construction dissipation; still high pore pressures today + Responds to total stress change from reservoir ~ response attenuates rapidly in downstream direction REEEREGEDSS ORES eS cence inant » 2015-12-03 18 2015-12-03 Upstream Dam Fill Pore Pressures. + Steady increase over time > slow advancement of saturation front due to reservoir (common for earthdams with wide cores after first impoundment) ape Tey": 19 (Yo Mar ven a Melk TOL Downstream Dam Fill Pore Pressur + Deeper piezometers > Generally a steady decrease over time as construction-induced excess pore pressures slowly dissipate + shallower piezometers > Generally “dry”; no excess pore pressures generated due to low fill height above tip. @ ee ° ) ohn cippen tee 2015-12-03 20 Foundation Movements “Grr crepensese Foundation Movements + Movements dominated by horizontal displacements along discrete shear plane in clay shale + Minimal straining in glacial till above bedrock a 2015-12-03 21 + Shear displacements continue after construction at gradually decreasing rates + “immediate” response to total stress change from reservoir surcharge in 2011. ‘Response sttenvates rapidly in downstream direction + Aso similar response noted in 1999 oc os (U)Kehn Crippen Berger 2015-12-03 22 Slog isplacements in Clay Shale 2015-12-03 23 Limit Equilibrium Stability Analyses ‘alsin secton Previous (September 20 Stabilty Analyses BOY) etenreeot cn | Feton ang of ay reer Anlof Coy Smee Nae ce FEL 962m as 366m 036 sessing storey) er Femmsceaaeed 056 1995 endo corso) ~roresenoit 2005 reser aS 562°} 100 2011 ~tesero aE. 562m) 201 -esercr surg E5685) iS Sa 2015-12-03 24 Limit Equilibrium Stability Analyses + Observations/Comments RE: 2D Stability Results: + Dam has not, and did not, fail: + Actual factor of safety (FOS) was, andi, higher than caleuated + Changes in FOS between the various reservoir scenarios are small: + Within the range of uncertainties (e.g. change in residual angle fcom 7° to 9° > similar change in FOS for reservoir from 562 to 567 m) * For all intents & purposes, the stability of the dam has NOT changed over its past 16+ years of operation + FOS of critical oblique section ~0.94 in 1995 vs ~0.96 in 2012 2 Qrernciepentese ity Analyses ne | een 2015-12-03 25 1 Orem ctppen terse 3D Limit Equilibrium Stability Analyses + Observations/Comments RE: 3D Stability Results + Minimum 3D Factor of Safety ~2.0 + 100% increase over minimum 20 FOS + Physical features of site and patterns of foundation deformations suggest 3D analysis is more representative of the actual conditions than 2D analysis. + However, 3D contribution is greater than normally expected based ‘on KCB’s experience Note: 3D FOS for a 3D model generated perpendicular to the dam axis is also about 2.0 eB 208 * (rer crepe 2015-12-03 26 2015-12-03 SYM aneriare-tutey ays) + Additional site investigations to gather information for deformation analysis (completed in 2012): + Field Program: + Two rotary core holes (Faste Multicril XL rig from Mobile Augers and Research of Edmonton) + Two mini-Sonic holes (Boart Longyear of Calgary) + Two Pressuremeter test holes (In Situ Engineering of Snohomish, Washington) + Three shallow test trenches on left abutment + Lab Program + Index tests + Glacial Til -9 CIUP Triaxials, 4 CID Triaxials, 3 CIUEP Triaxials + Clay shale ~3 direct shear tests (shear box) 2 Qin crppen serge Borehole Location Plan )) Kone Crippen Berger 27 Used late Offence Program FLAC (ast (eayangien analysis of Continua) by tasca le Consulting Group ne + Modeled progression fom orginal ground tivougha sequence of 19 construetion stages using the folowing materi modes ‘+ Dams ner ele + Ghai hyserbal foeus of eben) + snesesone 9 Mote Cover + chystte> tinea ease Initial properties determined from: 1 ale bbten dn: follwes eon uhen 3 ec ware Used effective stress analsis > pore pressures from actual pezometer data + 2D AAC euilent 1D FLAC > earpaation method by de Anes t 5 2015-12-03, 28 2015-12-03 Dam Construction History. 29 Selected Model for Glacial T! Selected the simplified Cap-Vield (Cysoil) soil model, which was used as a nonlinear elastic equivalence of the conventional Duncan and Chang! hyperbolic model: bg Young modulus number at reference pressure) © Poisson ratio © Cohesion % Unite tition angle timate iatiorargte 2 Tramsition rictorrangie foe eration tam} — mo) Modulus exponents A, Failure ratio eet es cc Quern ciiopen berger 2015-12-03 Triaxial and Pressuremeter Calibrations| 2015-12-03 31 2015-12-03 Final “Calibrated” Parameters Used in FLAC 3D ism 0m) cwen vite) ghy . S - (ern cibpen sesso ie” Tet cams @)) Rote Gippen tert 32 3D FLAC — Shear Displacements * Good match to sear splacement magnitudes + Matching esplacement = rections was not as seeeseul | “etnies eC 2S « \Kiohn Crippen Berger 3D, Model — Failure Assessment + Having ‘matched’ the model response to the inclinometer data, it was used to generate failure scenarios for comparison with the monitoring results: * ‘Conventional strength reduction’ + Strength of all materials progressively reduced until failure (defined by model non-convergence) occurs * Consistent with conventional definition of limit equilibrium FOS + ‘Till strength reduction only’ + Strength of till reduced until fallure occurs 2015-12-03 33 2015-12-03 ntional” vs “Till Only Uysaastste Vested Vy ee 34 2015-12-03 eee e “wer cisco eae Tr} fener 35 2015-12-03 of High Shear Strain from Till-Only Reduction Method? Canvetonsl sense asicion Gry Sen Reson |S mee 2 Orem cicpen seas: 36 Theoretical Displacements versus Strength spre egonsoew a crease |jcmet™ [ooman | eee” ae Scemiencars| mee’ | Sorin | omens SSS fete nene Setiacicans [owtrrtmlrae eat and fe | AT | poy Sonate | cay =| Comparison of the ‘failure assessment’ with the monitoring data, indicates there is a significant strength reserve in the glacial till within the passive wedge area. ‘The deformations to date are well within acceptable levels This is in good agreement with and, therefore, provides greater confidence in the high 3D factor of safety (~2.0) calculated from limit equilibrium method, The main conclusion was that the dam is stable nt mms om Q)tetn cpp eae 2015-12-03 37 Applications to Oil Sands Embankments (and, indeed, embankment’s of any other industry) + Burland’s 1987 Nash Lecture? > presented the view that geotechnical engineering practice is comprised of three parts: a) establishing ground profile b) Defining ground behavior ) modeling + Referred to as the Burland Triangle: Allthree are intetinked and all three are underpinned by experience consisting ofempiricism and precedent _ @romcippensoaee 2015-12-03 38 + One of the strongest endorsements on the value of numerical modeling for oil sands embankments was given by Dr. Morgenstern in his 2000 Inaugural Lumb Lecture*, where he described the following experience with the Syncrude tailings dam: Seioemitohaec at ce nS jer en cea gk meu @)xer» crppen berger 2015-12-03 39 Many Other Si Oil Sands Examples Increasingly Imp: t * Deformation modeling for oil sands embankments, both 2D and 3D, will continue to increase in importance and necessity: Having to deal with more complex geology + Trending towards “performance-based” design )) xiomn Crippen Berger 2015-12-03 40 Complex Geology am | t Ba ' é Trending to Performance-Based Design + Upto now, “performance” and “behavior” are terms that are often used interchangeably. + Moving towards a “performance-based” design approach, it will be very important to have clear definitions to avoid inadvertent misunderstandings. wacom 8 Klohn Crippen Berger 2015-12-03 41 Trending to Performance-Based Design + favor my colleague, Bryan Watts’s, view that: + Measured behavior is manifested by instrumentation readings end visual observations + Expected behavior is more judgmental but is normally predicated ‘on experience from similar case histories and/or representative ve modeling i.e, Evaluating the performance of a dam based on measured behavior alone is almost meaningless without a rational benchmark/standard + Deformation modeling will play an important and integral part of any performance-based design approach ne Bern cpr sae Thank You @ ——— eee “ (xen cappen berger 2015-12-03 42

You might also like