A critical review of current approaches on the determination of seismic force demands on nonstructural components

During and after an earthquake damaged nonstructural components pose severe threats to human lives and usually account for the majority of economic losses in non-residential buildings. Thus, seismic design of secondary structures should be duly considered, whereat seismic forces acting on components have to be known with appropriate accuracy. To serve this purpose four different approaches to determine horizontal seismic forces on components were reviewed: i) Simplified formulas contained in current code provisions; ii) Simplified formulas enhanced by refined determination of peak floor accelerations by means of modal superposition methods; iii) Decoupled time history analyses (floor response spectrum method) and iv) Coupled time history analyses. All methods were applied to a set of three realistic moment resisting steel frames, revealing large discrepancies between simplified and refined methods. Outcomes in terms of peak accelerations at the supporting structures' floors as well as at the component itself are discussed. It was found that the use of strongly simplified formulas as they are suggested in current codes could lead to unsafe design of nonstructural components. It is further shown that calculation of floor accelerations, allowed as refinement of simplified formulas in some provisions, is not as straightforward as might be expected. Further results show that incorporation of heavy components in a coupled model with its supporting structure can lead to significantly reduced calculated acceleration values. In this context it is discussed why caution should be paid when Rayleigh damping approach is used. Finally recommendations are given regarding the applicability of simple or refined methods.