Fast performance uncertainty estimation via pushover and approximate IDA

Abstract

Approximate methods based on the static pushover are introduced to estimate the seismic performance uncertainty of structures having non-deterministic modeling parameters. At their basis lies the use of static pushover analysis to approximate Incremental Dynamic Analysis (IDA) and estimate the demand and capacity epistemic uncertainty. As a testbed we use a nine-storey steel frame having beam hinges with uncertain moment rotation relationships. Their properties are fully described by six, randomly distributed, parameters. Using Monte Carlo simulation with Latin hypercube sampling, a characteristic ensemble of structures is created. The Static Pushover to IDA (SPO2IDA) software is used to approximate the IDA capacity curve from the appropriately post-processed results of the static pushover. The approximate IDAs allow the evaluation of the seismic demand and capacity for the full range of limit-states, even close to global dynamic instability. Moment-estimating techniques such as Rosenblueth's point estimating method and the first-order, second-moment (FOSM) method are adopted as simple alternatives to obtain performance statistics with only a few simulations. The pushover is shown to be a tool that combined with SPO2IDA and moment-estimating techniques can supply the uncertainty in the seismic performance of first-mode-dominated buildings for the full range of limit-states, thus replacing semi-empirical or code-tabulated values (e.g. FEMA-350), often adopted in performance-based earthquake engineering. Copyright (C) 2009 John Wiley & Sons, Ltd.