This paper presents and compares the applications of three numerical procedures, Shake, Dysac2 and Sumdes , to the evaluation of earthquake-induced damages. Shake, a total stress-based equivalent-linear procedure, and Dysac2, an effective stress-based, fully coupled, nonlinear, ยฎnite element method-based procedure, were applied to the evaluation of liquefaction-induced building damage in the San Francisco Marina District which occurred during the 1989 Loma Prieta earthquake. Sumdes , an effective stress-based, fully coupled, nonlinear, ยฎnite element method-based numerical procedure, was applied to the evaluation of liquefaction-induced deformation which occurred at the Port Island during the 1995 Kobe earthquake. Three sites in the Marina District were selected for this study and analyzed using numerical procedures Shake and DYSAC2. In-situ and laboratory tests were used to obtain the initial state parameters and constitutive model constants for the liquefaction-induced building damage analysis during the Loma Prieta earthquake. The results of this analysis are presented and compared with the observed behavior in the Marina District during the 1989 Loma Prieta earthquake. Differences between the predictions of the Dysac2 and Shake computer programs are discussed.
The laboratory formation factor tests and shear wave velocity results were used to obtain the initial state parameters and constitutive model constants for the analysis of the liquefaction-induced deformation of an instrumented site at Port Island, Kobe. A summary of the results of the analyses performed using Sumdes are presented and compared with observed behaviors. Close agreements are shown to exist between the evaluated and observed results. The two effective stress-based numerical procedures, Dysac2 and Sumdes , were veriยฎed during the VELACS (Veriยฎcation of Liquefaction Analysis using Centrifuge Studies) and VELACS-Extension projects.