Consequences of Modeling Choices in Seismic Performance Assessment of Buildings

Performance-based approaches utilizing nonlinear analyses have become increasingly popular for seismic evaluation of buildings. Nonlinear simulations of building response require various assumptions and modeling decisions-from choice of software to model parameters-which opens the door to differences in demand assessments from what essentially could be very similar computational models. This study examines nonlinear response sensitivity of three steel moment frame structures to variations in basic nonlinear modeling parameters using three different software platforms: OpenSees, Perform-3D, and SAP2000. The building models were analyzed in the inelastic range using a suite of near-fault and farfault ground motions, and response sensitivity was assessed using interstory drift and plastic rotation demands. Findings from the study indicate that sensitivity to modeling assumptions and choice of software are more pronounced at the local/element level than at the global/system level and can have an impact in performance-based seismic assessment.