Accidental torsion within the frame of nonlinear dynamic analysis using code accidental eccentricities and Monte Carlo simulations

Abstract

A comparative analysis of responses obtained with typical design procedures that consider accidental eccentricities to provide torsional strength to buildings is presented. Design accidental eccentricities are assessed in terms of both ductility demands and lateral distortions, using dynamic nonlinear analyses and Monte Carlo simulations. Two models representing generic symmetrical steel buildings of four and eight stories are considered. Results suggest that designs based on floor torsion moments lead to similar values of both ductility demands and inter-story drifts, regardless of the fraction (0.05, 0.10 or variable) of the building plan dimension b used as accidental eccentricity ea. The procedures to provide buildings with accidental-torsion resistance based on either floor torsion moments or shifting the centers of mass are not equivalent; the former leads to larger ductility demands. For the dynamic step-by-step (nonlinear) analyses used by some codes for building revision, shifting the centers of mass a distance equal to 0.05b of their nominal locations is a conservative device to take into account the accidental torsion. Moreover, shifting the masses is a simpler method than applying floor torsion moments.