Re-evaluation of equivalent lateral force procedure for prediction of displacement demand in seismically isolated structures

• Recent studies evaluating the ELF procedure resulted in conflicting conclusions.
• The accuracy of the ELF procedure depends on ground motion selection and scaling.
• Revisions of the seismic design provisions of ASCE 7 have significant effects.
• The ELF procedure provides good estimates of demands for spectrum-matched motions.
• Predictions of the procedure could be un-conservative for weighted-scaled motions.

The accuracy of the equivalent lateral force (ELF) procedure described in codes and specifications for the analysis and design of seismically isolated structures is of concern, because this procedure is always used either directly under certain conditions or to establish a lower-bound limit on the results of response history analysis (RHA). A number of recent studies have resulted in conflicting conclusions with some showing the ELF procedure to produce accurate results and others showing the opposite. A review of these studies shows that they primarily differ in the selection and scaling of motions for RHA, and in the interpretation of the results. Accordingly, this study re-evaluates the accuracy of the ELF procedure based on RHA of a range of parameters of isolation systems subjected to bidirectional excitations and investigates the effects of (a) the selection of seed earthquake motions for RHA (near-fault pulse-like versus non-pulse-like), (b) the procedure for scaling the seed motions for RHA (weighted scaling versus spectral matching), (c) the specification of the response spectrum (geometric mean versus maximum direction spectrum), and (d) the measure of central tendency (average versus median). It is found for the isolation systems considered, the ELF procedure always provides reasonably good or conservative estimates of the isolation system displacement demands provided that non-pulse-like motions are selected and spectral matching is used for scaling. In other cases, the demands predicted by the ELF procedure could be un-conservative, especially for large effective period and low effective damping isolation systems.