1. Liquefaction: Behavior Evidence, Prediction, and Prevention Richard P. Ray, Ph.D, P.E.

5. Today’s Menu What is Liquefaction ?What is Liquefaction ? How is it expressed ?How is it expressed ? Can we estimate liquefaction potential ?Can we estimate liquefaction potential ? Can we prevent it from happening or reduce its impact ?Can we prevent it from happening or reduce its impact ?

6. What is it Worm’s eye view - Physical BehaviorWorm’s eye view - Physical Behavior Professor’s eye view - Laboratory and Analytical ApproachProfessor’s eye view - Laboratory and Analytical Approach

7. abc Worm’s eye view

8. Triaxial Test on Sand

9. Triaxial Test on Saturated Sand

12. Expression of Liquefaction Sand boils and flows – free-fieldSand boils and flows – free-field Landslides – static lateral stressLandslides – static lateral stress Foundation Failures – surface loadingsFoundation Failures – surface loadings

23. Evaluating Liquefaction Potential Empirical Methods SPT, CPT Empirical Methods SPT, CPT Stress or Strain-Based Simplified Stress or Strain-Based Simplified Effective Stress Time History Models Effective Stress Time History Models Probabilistic Methods Probabilistic Methods

24. Evaluating Liquefaction Potential Stress Based Strain Based Seed & Idriss One Dimensional Response Analysis One Dimensional Response Analysis Laboratory Stress Cyclic Testing Laboratory Stress Cyclic Testing Dobry et al One Dimensional Response Analysis One Dimensional Response Analysis Laboratory Strain Cyclic Testing Laboratory Strain Cyclic Testing Compare Earthquake Load to Soil Resistance

25. Seed & Idriss Simplified Method FS=Resistance/Load Resistance =  res = fn( SPT Blow Count, Earthquake Mag, Percent Fines) Load =  cyc = fn(Earthquake Accel, Vertical Stress)

26. Load Equation, Seed & Idriss Simplified Method Divide profile into segments (~2ft.)Divide profile into segments (~2ft.) Compute  v, r d for each segmentCompute  v, r d for each segment Compute  cyc for each segmentCompute  cyc for each segment

27. Stress Reduction Factor (r d )

28. Seed & Idriss Load Profile,  cyc

29. Resistance Equation, Seed & Idriss Simplified Method Use same depth segments (~2ft.) Use same depth segments (~2ft.) Compute  ’ v for each segment Compute  ’ v for each segment Compute CSR for each segment Compute CSR for each segment Compute  res for each segment Compute  res for each segment

30. Resistance Equation, Seed & Idriss Simplified Method CSR = Cyclic Stress Ratio  res /  ’ v CSR = Cyclic Stress Ratio  res /  ’ v CSR based on SPT N 1 60 values CSR based on SPT N 1 60 values CSR influenced by Percent Fines CSR influenced by Percent Fines Therefore Must correct N blow countsMust correct N blow counts Must account for Percent FinesMust account for Percent Fines

31. Empirical CSR Based on Blow Count, N 1 60

32. Load-Resistance and F.S. Profiles

33. Resistance Equation, Seed & Idriss Simplified Method (Alternatives) CSR = Cyclic Stress Ratio  res /  ’ v CSR = Cyclic Stress Ratio  res /  ’ v CSR based on CPT =>SPT N 1 60, or CSR based on CPT =>SPT N 1 60, or CSR based on Vs => SPT N 1 60 CSR based on Vs => SPT N 1 60

34. Loading Conditions, Dobry et al Simplified Strain Method Divide profile into segments (~2ft.) Divide profile into segments (~2ft.) Compute  v, r d for each segment Compute  v, r d for each segment Compute  cyc for each segment (iterative procedure) Compute  cyc for each segment (iterative procedure)

35. Liquefaction Resistance, Dobry et al. Simplified Strain Method Perform strain-controlled laboratory tests at several stages of  cyc-lab to measure number of cycles to liquefaction N (l,  ) At  cyc =  cyc-lab compare N eq to N (l,  )At  cyc =  cyc-lab compare N eq to N (l,  ) If N eq > N (l,  ) liquefaction possibleIf N eq > N (l,  ) liquefaction possible If N eq < N (l,  ) liquefaction not possibleIf N eq < N (l,  ) liquefaction not possible

36. Settlement During Liquefaction, Dobry et al-Castro Method Uses correlation of  cyc to  vol for dry and saturated sands (SP, SP-SC, SC) Uses correlation of  cyc to  vol for dry and saturated sands (SP, SP-SC, SC) For every profile segment use  cyc to find  vol For every profile segment use  cyc to find  vol Multiply  vol times segment height to get segment settlement contribution Multiply  vol times segment height to get segment settlement contribution Add all segment contributions for settlement estimate at surface. Add all segment contributions for settlement estimate at surface.

37. Settlement Correlation - Castro

38. Alternative Methods for Liquefaction Potential Estimation Empirical Approach, Cone Resistance Empirical Approach, Cone Resistance Empirical Approach, Shear Wave Velocity Empirical Approach, Shear Wave Velocity MOC 1-D Analysis MOC 1-D Analysis

39. Empirical Approach, Cone Resistance

40. Empirical Approach, Shear Wave Velocity

41. MOC 1-D Analysis

42. MOC 1-D Analysis Output

44. Prevention/Control Soil modification Reduced Loads Drainage (pore pressure relief)