Nonlinear, Effective-Stress Ground Motion Response Analyses following AASHTO Specifications for Load and Resistance Factor Design Seismic Bridge Design

Nonlinear, effective-stress ground motion response analyses were conducted for a bridge site north of Seattle, Washington. This site was characterized by silt and sand layers in the upper 90 ft (27.4 m) of the soil profile that were expected to liquefy during a 975-year seismic event. Procedures described in the 2009 AASHTO Guide Specifications for LRFD Seismic Bridge Design were followed when site-specific analyses of ground motion response were conducted. Two one-dimensional, nonlinear effective-stress computer programs, D-MOD2000 and PSNL, were used to perform the analyses. Field and laboratory tests were conducted to define soil parameters for the nonlinear soil models. Results of the analyses were used to identify soil layers that were likely to liquefy as well as changes in design response spectra that could result from liquefaction. These results demonstrated important differences in site-response prediction from subtle changes in soil modeling, including the potential for shielding of upper liquefiable soil layers and the importance of soil dilation when lower layers liquefy. This paper provides a summary of the analyses and discusses observations made from comparison of the two sets of analyses. The paper also highlights uncertainties that can occur when nonlinear, analyses of effective-stress ground motion response are conducted.