Benchmarking of Nonlinear Geotechnical Ground Response Analysis Procedures PEER Lifelines Project 2G

1
Benchmarking of Nonlinear Geotechnical Ground
Response Analysis ProceduresPEER Lifelines
Project 2G02
2
Motivation

• 1D equivalent linear are by far the most commonly
  used method of analysis in the Western U.S.
• Effects of parametric uncertainty on computed
  results are poorly understood, especially with
  respect to nonlinear models

Source Kramer and Paulsen, 2004
3
Project Objectives

• When is NL beneficial with respect to EL?
• Strain ranges (non-liquefiable soils)
• Intensity measures
• Develop consensus guidelines for use of NL codes
• Guidelines for input parameter selection
• Free parameters
• Fixed parameters
• Uncertainty in input parameters and effect on
  analysis results

4
Project Participants

• Management team
• Stewart, Kwok, Bozorgnia, Shantz
• Developers
• Hashash (DEEPSOIL)
• Matasovic (DMOD)
• Pyke (TESS)
• Wang (SUMDES)
• Yang (CYCLIC 1D)

• Advisory Board
• S. Chang
• I.M. Idriss
• S. Kramer
• F. Makdisi
• G. Martin
• L. Mejia
• W. Silva
• J. Sun

5
Project Overview

• Two-year project July 2004 June 2006
• Web page http//cee.ea.ucla.edu/faculty/jstewart/
  groundmotions/PEER2G02/
• Three general tasks
• Develop parameter selection protocols
• Verification studies
• Parametric studies
• Effects of parametric variability
• Benefits of NL relative to EL and application in
  PSHA

6
Project Overview

• Two-year project July 2004 June 2006
• Web page http//cee.ea.ucla.edu/faculty/jstewart/
  groundmotions/PEER2G02/
• Three general tasks
• Develop parameter selection protocols
• Verification studies
• Parametric studies
• Effects of parametric variability
• Benefits of NL relative to EL and application in
  PSHA

7
Schedule
8
Parameter Selection Protocols

• Identify all parameters (free/fixed)
• For free parameters, develop baskets
  (rule-based procedures)
• Should specific best estimate values and
  distribution
• Compiled for generic soil types (unsaturated
  clean sands and fines-dominated soils with
  various plasticities)
• Based on typical soil data FC, PI, Vs, Su, s0?

NL analyses should yield stable and unbiased
median estimates of site amplification as well as
reasonable levels of variability when model
parameters and input motions are simultaneously
varied over their distributions.
9
Todays Agenda

• Introductory remarks (Stewart)
• Presentations on nonlinear codes (Developers)
• 30 minutes each
• General structure and capabilities of codes
• Input parameters used
• Existing parameter selection protocols (if any)
• Discussion on parameter selection (All)
• Verification plan (Stewart)
• Developers Year 1 scope of work (All)