Probabilistic seismic demand analysis of steel moment frames by utilising Bayesian statistics

Abstract

Seismic reliability of a structure at a designated site is quantified using Probabilistic Seismic Demand Analysis (PSDA). In PSDA, the seismological and structural uncertainties are decoupled using an intermediate variable called the ground motion Intensity Measure (IM). In this research, the efficiency of various IMs was considered based on the record to record variability in seismic demands of frames with different heights. The results show that IM_oc, defined based on the optimal combination of spectral displacements, is the most appropriate IM regarding practicality and efficiency. Seismic demands under 90 far and near-fault ground motion records were used to consider the possibility of introducing the more efficient formula for the intensity measure IM_oc. Then, PSDA was conducted in the context of Bayesian interpretation to make a direct statement about drift hazard combining randomness with uncertainty. The PSDA results demonstrate that using the common intensity measure of S_a(T₁) for frames with more than six stories leads to considerably conservative drift hazard estimation with respect to other IMs. It is shown that at a given value of demand, hazard value decreases with increasing the first-mode period of vibration. The amount of reduction in moderate and long-period frames is more considerable than short-period ones.

Keywords:

• Probabilistic seismic demand analysis
• Bayesian statistics
• intensity measure
• variability
• drift hazard curve

Disclosure statement

No potential conflict of interest was reported by the authors.