12.07.2015 Views

Numerical Modelling of Instability in Strain-Softening Soils - NGI

Numerical Modelling of Instability in Strain-Softening Soils - NGI

Numerical Modelling of Instability in Strain-Softening Soils - NGI

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

1<strong>Numerical</strong> <strong>Modell<strong>in</strong>g</strong> <strong>of</strong> <strong>Instability</strong><strong>in</strong> Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>Hans Petter JostadAnders Samstad GyllandSte<strong>in</strong>ar Nordal<strong>NGI</strong>/NTNUNTNUNTNUWith special thanks to:Gustav Grimstad (HIOA), Vikas Thakur (NPRA), Francesco Bonadies (Univ. Salerno)ICG Symposium Geohazards and Society November 2012


2Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>LoadPeakReduction <strong>in</strong> resistance for<strong>in</strong>creas<strong>in</strong>g deformationLoadStra<strong>in</strong>s<strong>of</strong>ten<strong>in</strong>gResidualSensitiveclayDisplacementICG Symposium Geohazards and Society November 2012


3KaareHöeg1972


4Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>• Smårød, 20.12.06ICG Symposium Geohazards and Society November 2012


5Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>• Kattmarkveien 13.03.09ICG Symposium Geohazards and Society November 2012


6Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>• Esp 01.01.12Photo:Ned Alley/ScanpixICG Symposium Geohazards and Society November 2012


7Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>• Esp 01.01.12Photo: KRISTOFFER FURBERGICG Symposium Geohazards and Society November 2012


8Sensitive clay• Deposited <strong>in</strong> salt water• Landrise• Fresh water <strong>in</strong>filtration– House-<strong>of</strong>-cards without glue• Liquefies when remoulded– Quick clay: s r < 0.5 kPawww.forskn<strong>in</strong>g.nowww.ngu.noICG Symposium Geohazards and Society November 2012


9Sensitive clay• CIUc triaxial tests, block samples Tiller quick clayICG Symposium Geohazards and Society November 2012


10Sensitive clay• CIUc triaxial tests, block samples Tiller quick clayICG Symposium Geohazards and Society November 2012


(ICSMFE Mexico, 1969)ICG Symposium Geohazards and Society November 2012


12Stig Bernander:Surte slide 1950,Tuve slide 1977PhD 2011ICG Symposium Geohazards and Society November 2012


RIGID SPRINGSCOMPRESSIBLE SPRINGSΔΔNNFTF1Tδ 1 -δ 22δ 2 -δ 33TδTδTδTTδ 1δ 2Tδ 3FFΔΔICG Symposium Geohazards and Society November 2012


ICG Symposium Geohazards and Society November 2012τ 0 > c R


ICG Symposium Geohazards and Society November 2012


16Downward progressive failureLong natural slopeEABar on weak layerForceWeak layerICG Symposium Geohazards and Society November 2012


17Downward progressive failureEAForceABWeak layerδ Aδ A ≠ δ Bδ BxICG Symposium Geohazards and Society November 2012


19Downward progressive failureEAForceABWeak layerStra<strong>in</strong>s<strong>of</strong>ten<strong>in</strong>gτAδ Aδ A ≠ δ BτBδ Bxτδδτ 0Initial shear stressxICG Symposium Geohazards and Society November 2012


20Downward progressive failureEAForceABWeak layerStra<strong>in</strong>s<strong>of</strong>ten<strong>in</strong>gτAδ Aδ A ≠ δ BτBδ Bxτ 0δ δInitial shear stressxICG Symposium Geohazards and Society November 2012


21Downward progressive failure• Slope resistance depends on:StiffnessPeak strengthS<strong>of</strong>ten<strong>in</strong>g behaviourτ 0Initial shearstress level• Higher resistance for:– Higher stiffness– Higher peak strength– Lower rate <strong>of</strong> stra<strong>in</strong> s<strong>of</strong>ten<strong>in</strong>g (perfect plastic = zero stra<strong>in</strong> s<strong>of</strong>ten<strong>in</strong>g)– Lower <strong>in</strong>itial stress levelxICG Symposium Geohazards and Society November 2012


22Downward progressive failure• Initial shear stress level– Highly sensitive– 10% change <strong>of</strong> the <strong>in</strong>itialshear stress level gives a40-50% change <strong>of</strong> thecapacity• Stiffness– Sensitive for low values• S<strong>of</strong>ten<strong>in</strong>g– Sensitive for low values– Uncerta<strong>in</strong> parameterICG Symposium Geohazards and Society November 2012


24Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong>LoadPeakReduction <strong>in</strong> resistance for<strong>in</strong>creas<strong>in</strong>g deformationLoadStra<strong>in</strong>s<strong>of</strong>ten<strong>in</strong>gResidualSensitiveclayDisplacementICG Symposium Geohazards and Society November 2012


25Downward progressive failureICG Symposium Geohazards and Society November 2012


26Localized failure - shear bandsICG Symposium Geohazards and Society November 2012


27S<strong>of</strong>ten<strong>in</strong>g gives mesh dependencyδΤδΤγ 1t 1γ 2t 2W 1 W 1 > W W 22LLττ 1τ 2γ 1 γ 2γΤ21δICG Symposium Geohazards and Society November 2012


28Regularisation technique <strong>in</strong> order to obta<strong>in</strong>mesh <strong>in</strong>dependent solution4Need a procedure that gives a capacity orsafety factor that is mesh <strong>in</strong>dependent.L o a d p / s u C32Shear band thickness:tsb = 150 cmtsb = 50 cmtsb = 1 cmThis means that the solution shouldconverge upon mesh ref<strong>in</strong>ements and theshear band thickness should be larger thangiven by the element size.10(Andresen /Jostad)0 0.01 0.02 0.0300 D i s p l a c e m e n t δ m a xICG Symposium Geohazards and Society November 2012


Regularization – non-local stra<strong>in</strong>p*p pα( ) ( ) ( ) (p∆ ε x =∆ε x −α⋅∆ ε x + ∫ w( x) ⋅∆ε ( x)) dVi i iVδhτγ pt sbControl <strong>of</strong> the shear band thicknessγ


1D shear columnNormalized shear stress (τ/su)1.00.90.80.70.60.50.40.30.20.1α = 1.58, l/L = 0.1 => t/L =π/101 el10 el20 el50 elMesh <strong>in</strong>dependent solutionwhen the element size issmaller than the shearband thickness!0.00.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0Normalized displacement δ/L [%]τδ1 elThe shear band thickness isdef<strong>in</strong>ed by the s<strong>of</strong>ten<strong>in</strong>g zone!10 el20 el50 el100100100100Lγt sbNormailzed position (x/L) [%]806040Normailzed position (x/L) [%]806040Normailzed position (x/L) [%]806040Normailzed position (x/L) [%]8060402020202000 10 20 3000 10 20 3000 10 20 3000 10 20 30Normalized displacement (δ/L) [%]Normalized displacement (δ/L) [%]Normalized displacement (δ/L) [%]Normalized displacement (δ/L) [%]


Slope stability problem“Almost” mesh <strong>in</strong>dependent results!


32Problem:The shear band is mm scale, elements are m scaleTrick for FEM simulations:Increas<strong>in</strong>g the <strong>in</strong>ternal length by reduc<strong>in</strong>g the s<strong>of</strong>ten<strong>in</strong>g stra<strong>in</strong>, ∆γ s<strong>of</strong>t∗Shear stress, τPeak∆γ s<strong>of</strong>t∗ ∙l <strong>in</strong>ternal * = ∆γ s<strong>of</strong>t ∙ l <strong>in</strong>ternalResidualOnly really OK for 1D problems∆γ s<strong>of</strong>t∗Shear stra<strong>in</strong>, γICG Symposium Geohazards and Society November 2012


Material model – <strong>NGI</strong>-ADPS<strong>of</strong>tτs uCf( σ,γp)= τ −su( α,γp)s uDSSα = 0 os urCσ zzσ 14 5 oτ xzτo9 0 oγzασ xxs uE


34Smårød, 20.12.06ICG Symposium Geohazards and Society November 2012


Slides by Francesco Bonadies, Univ. SalernoVolume [m 3 ] 750.000Length [m] 200Width [m]500Slip surface [m] 15Three ma<strong>in</strong> phases:- area A first- area B next,- area C f<strong>in</strong>alA fill <strong>of</strong> about 7 meters is the trigger<strong>in</strong>g agent ?Simulated by <strong>NGI</strong>-ADPS<strong>of</strong>t as part <strong>of</strong> the SVV-NVE project :Effekt <strong>of</strong> progressive failure on physical development <strong>of</strong> areas with quick clay


Slides by Francesco Bonadies, Univ. Salerno• dry crust• s<strong>of</strong>t clay layer• firm bottom123123


37SmårødICG Symposium Geohazards and Society November 2012


38SmårødICG Symposium Geohazards and Society November 2012


Slides by Francesco Bonadies, Univ. SalernoTrigger<strong>in</strong>gembankmentOldembankmentRiver190 mPr<strong>in</strong>cipal total stra<strong>in</strong>s directions


Slides by Francesco Bonadies, Univ. SalernoInput


Slides by Francesco Bonadies, Univ. Salerno


Effect <strong>of</strong> updated geometry <strong>in</strong> analyses <strong>of</strong> progressive failureUpdated geometry is important for captur<strong>in</strong>g the f<strong>in</strong>al slide configurationGylland, A. & Jostad, H.P (2010) NUMGEICG Symposium Geohazards and Society November 2012


Initial stresses are difficult to evaluate:Masterthesis Magne Mehli, vår 2010ICG Symposium Geohazards and Society November 2012


How important is stra<strong>in</strong> s<strong>of</strong>ten<strong>in</strong>g forevaluat<strong>in</strong>g the stability <strong>of</strong> a slope?Key project:Effekt <strong>of</strong> progressive failure on physicaldevelopment <strong>of</strong> areas with quick clayThanks to:


Effect <strong>of</strong> s<strong>of</strong>ten<strong>in</strong>g:Karlsrud/<strong>NGI</strong>Handbook 016standard psamples?Prelim<strong>in</strong>ary results from <strong>NGI</strong> project:Effekt <strong>of</strong> progressive failure on physicaldevelopment <strong>of</strong> areas with quick clay


46Modified triaxial cell• Mov<strong>in</strong>g, low friction base sled• Instrumented• Planar shear bands• Why triax?– Optimal sample handl<strong>in</strong>g and quality– Ma<strong>in</strong>ta<strong>in</strong> relevance <strong>of</strong> tested materialICG Symposium Geohazards and Society November 2012


47Excess pore pressure• Stra<strong>in</strong> s<strong>of</strong>ten<strong>in</strong>g by excess pore pressure• Characteristic consolidation time ≈ lab test time• Internal pore pressure gradientsPlastic shear<strong>in</strong>gGeneration <strong>of</strong> excesspore pressureElastic unload<strong>in</strong>gReciever <strong>of</strong> excesspore pressureICG Symposium Geohazards and Society November 2012


48Micro CT, PhD work A. GyllandICG Symposium Geohazards and Society November 2012


49Conclud<strong>in</strong>g remarks• <strong>Numerical</strong> Model<strong>in</strong>g <strong>of</strong> <strong>Instability</strong> <strong>in</strong> Stra<strong>in</strong>-S<strong>of</strong>ten<strong>in</strong>g <strong>Soils</strong> is still difficult,but may now be done <strong>in</strong> a consistent manner• The <strong>NGI</strong> – ADPS<strong>of</strong>t with non-local stra<strong>in</strong> is a powerful tool• A pragmatic «effect <strong>of</strong> s<strong>of</strong>ten<strong>in</strong>g factor» is studied, 10% ?• Old message repeated: Prevent the <strong>in</strong>itial slide!• More research needed:– Understand<strong>in</strong>g material behaviour– Scal<strong>in</strong>g the s<strong>of</strong>ten<strong>in</strong>g curve only exact for 1D– Initial stresses are hard to determ<strong>in</strong>e, but has high <strong>in</strong>fluence– Effective stress based simulations wanted Ge<strong>of</strong>uture• Design codes and guidel<strong>in</strong>esICG Symposium Geohazards and Society November 2012


50Thank you ICG !ICG Symposium Geohazards and Society November 2012

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!