Full Program »

Impact of computational choices on estimated structural collapse capacity under earthquake ground motions

A key premise of the Performance-Based Earthquake Engineering (PBEE) paradigm is the ability to accurately simulate the response of a structure under an earthquake ground motion. Estimating the collapse capacity of a structure, which describes the probability of structural collapse under a given intensity of ground excitation, requires the accurate simulation of structural response under intense ground motions that produce large inelastic deformations. A number of computational choices need to be made when running such simulations, like choosing the type of analysis (e.g. incremental dynamic analysis, multiple stripe analysis), the algorithm used to conduct the analysis (e.g. sequence of scaling ground motions), the time integration scheme used (e.g. implicit schemes like the Newmark average acceleration and HHT schemes, and explicit schemes like the Central Difference scheme), and the criteria used to determine structural collapse. Understanding the impact of these computational choices on the final, estimated structural collapse capacity is key to successfully interpreting the results of the analysis. This study quantifies the influence of these computational choices on the estimated structural collapse capacity, and makes recommendations on procedures to follow when conducting collapse simulations, based on the obtained results. Key among these recommendations is to use explicit, instead of implicit time integration schemes to avoid problems related to non-convergence. ​​Recommendations are also made to speed up collapse analyses using parallel computing to facilitate the use of such analyses more frequently in structural design practice.

Author(s):

Reagan Chandramohan    
Stanford University
United States

Jack Baker    
Stanford University
United States

Greg Deierlein    
Stanford University
United States