Time-variant reliability analysis using the parallel subset simulation

Time-variant reliability problems commonly occur in practical engineering applications due to deterioration in material properties, dynamic load and other causes. Since this kind of problem is usually a small probability event, subset simulation is more efficient than Monte Carlo simulation (MCS). However, subset simulation can only focus on a single limit function when propagating the conditional samples. Parallel subset simulation is applied to deal with time-dependent reliability analysis in this paper. A new method is proposed to construct a function called "principal variable". The "principal variable" can represent limit state at each time instant to generate conditional samples. In addition, the update procedure of "principal variable" should be set at each simulation stage to keep the correlations between "principal variable" and n(t) limit states strong. Two numerical examples are used to demonstrate the effectiveness and accuracy of the developed parallel subset simulation for time-variant reliability analysis.