On the relative performance of spectrum superposition methods considering modal interaction effects

Abstract

The contribution of modal interaction in the various available spectrum superposition methods is accounted via the modal cross-correlation coefficient, which has been defined in several different approximate ways. Further, in these methods, to define the final expressions directly in terms of the response spectrum amplitudes, the peak factors for all the modal responses are approximated to be equal to the peak factor for the total structural response. However, these assumptions have been found to be violated significantly in many cases and do not hold good in general. Therefore, some recent studies have attempted to improve upon these assumptions. In this paper, detailed investigations are made to study the relative performance of the various available methods considering the modal interaction effects. To find out which of the available methods, in general, gives the better results, the response of a five-storey asymmetric hypothetical building, characterized by significant interaction effects, has been computed from different methods for several widely differing input excitations and the results have been compared with the exact time-history solution.