Advanced Lateral Spread Modeling

Transcription

1 Adv. Liquefaction Modeling Page 1 Advanced Lateral Spread Modeling Reading Assignment Lecture Notes Other Materials Homework Assignment 1. Complete FLAC model 10a.pdf 2. Modify the example in FLAC model 10a.pdf to do the following: Degrade the shear modulus from a peak to residual value Degrade the shear strength from a peak to a residual value Start the degradation when the ground acceleration first exceeds 0.15g Assume that liquefaction is fully developed at t = 28 s Assume that the clay has a strength of V Assume that the residual strength of the sand is equal to the average undrained strength for an N160 value of 10. Plot the same results that are requested in FLAC model 10a.pdf

2 Adv. Liquefaction Modeling Page 2 Accelerations and Pore Pressure Generation During Liquefaction

3 Adv. Liquefaction Modeling Page 3 Flow Failure versus Deformation Failure Stable Slope Deformation Failure Flow Failure

4 Adv. Liquefaction Modeling Page 4 Flow Failures Sheffield Dam Pasted from < San Fernando Dam

5 Adv. Liquefaction Modeling Page 5 Deformation Failures Pasted from < ~liquefaction/selectpiclique/rivers/motagua.jpg> Pasted from <

6 Adv. Liquefaction Modeling Page 6 Counting Cycles to Liquefaction

7 Adv. Liquefaction Modeling Page 7 Equivalent Stress Cycles Versus Earthquake Magnitude Earthquake magnitude, M Number of representative uniform cycles at 0.65τ max Seed et al., (1975)

8 Adv. Liquefaction Modeling Page 8 Number of Cycles to Liquefaction

9 Adv. Liquefaction Modeling Page 9 Pore Pressure Buildup Versus No. of Cycles For = 1

10 Adv. Liquefaction Modeling Page 10 Pore Pressure Generation Scheme for Modeling When r u reaches 1.0, then complete liquefaction has occurred. Functions to degrade residual strength and shear modulus according to r u = ( max - residual ) (1-r u ) + residual G = (G max - G residual ) (1-r u ) 1/2 + G residual

11 Adv. Liquefaction Modeling Page 11 Relating Residual Strength with Residual Shear Modulus Sr/Gr ratio = shear strain (decimal fraction)

12 Adv. Liquefaction Modeling Page 12 Strain - Strain Loops Beaty and Byrne, 1999 Soft reloading curve = 10 percent of stiff unloading curve

13 Adv. Liquefaction Modeling Page 13 Model Verification Input motion Model Geometry

14 Adv. Liquefaction Modeling Page 14 Model Verification (cont.) Hysteresis loops for site soil with low (5 k Pa) residual strength Flat top part of loop shows perfectly plastic yielding Loading curve is soft Reloading curve is stiffer (10 x modulus of loading curve)

15 Adv. Liquefaction Modeling Page 15 Model Verification (cont.) Comparison with Kobe Site Note that liquefaction has caused a significant decreases in the surface ground motion

16 Adv. Liquefaction Modeling Page 16 Model Verification (cont.) FLAC model for Kobe Site

17 Adv. Liquefaction Modeling Page 17 Model Verification (cont.) Comparison of surface response spectra for predicted vs. measure motions Comparison of strain-strain loops at 8 m

18 Adv. Liquefaction Modeling Page 18 Model Verification (cont.) Comparison of pore pressure generation plot

19 Adv. Liquefaction Modeling Page 19 Model Verification (cont.) Wildlife site - liquefied sand

20 Adv. Liquefaction Modeling Page 20 Model Verification (cont.) Measured down hole vs. surface acceleration FLAC model for Wildlife site

21 Adv. Liquefaction Modeling Page 21 Model Verification (cont.) Predicted vs. measured surface acceleration time histories

22 Adv. Liquefaction Modeling Page 22 Model Verification (cont.) The difference may due to: The relatively low permeability of the liquefied silty around piezometer. Pore pressure need to migration to reach the piezometer. Thus the pore pressure records at the WLA may not indicate when liquefaction (r u = 1) was reached. Possible that this area reached the liquefied state later, on average, than typical liquefied site. FLAC modeling approach does not consider pore pressure migration or redistribution, also it based on average soil properties.

23 Adv. Liquefaction Modeling Page 23 Model Verification (cont.) Stress-Strain loops for Wildlife

24 Adv. Liquefaction Modeling Page 24 Model Calibration Case history sites used in the calibration process For each case history without recorded ground motion, 7 synthetic strong motions were selected from 30 synthetic strong motion recorded generated by the SGMSV5 program (Papagorgiou, 2004). These 7 motions were selected at the following according to the spectral acceleration at the fundamental period of the liquefied soil column. i. mean value ii. maximum value iii. minimum value iv. +1/2 standard deviation v. 1/2 standard deviation vi. +1 standard deviation vii. 1 standard deviation

25 Adv. Liquefaction Modeling Page 25 Model Calibration Results Back-calculated values of S r (residual strength) normalized to the effective vertical stress v ' Regression equation from Meng (2011) to predict the normalized residual strength: Sr / v' = N160CS * D50 (mm)

26 Adv. Liquefaction Modeling Page 26 Correlation with In situ Properties Mesri and Stark (1992) Note that the data from this study (red diamonds) suggest that the correlation with N160CS (N160 adjusted to a clean sands value) is approximately between the mean value and lower bound value determined by Stark and Mesri (1992).

27 Adv. Liquefaction Modeling Page 27 Determining the Residual Shear and Bulk Modulus Sr = residual shear strength Gr = residual shear modulus