An efficient method of system reliability analysis of steel cable-stayed bridges

The IHS-EIS method is developed for reliability analysis of steel cable-stayed bridges by integrating the improved Latin Hypercube (IHS) and a proposed effective importance sampling (EIS). All nonlinearity sources of bridges are directly captured by the proposed practical advanced analysis (PAA) using catenary elements for cables and beam-column elements for pylons, girders, and cross beams. The innovation of the proposed method comes from both PAA for performing structural analysis and IHS-EIS for solving reliability analysis. Compared to commercial software ABAQUS, the computational cost of structural analysis is significantly reduced by using the proposed PAA method. IHS-EIS can accurately capture the failure probability of structure and considerably decrease the number of samples in comparison with Monte Carlo simulation, importance sampling, Latin hypercube, and subset simulation methods. Three mathematical examples and two steel frames are first presented to demonstrate the accuracy and efficiency of IHS-EIS. This method is then applied for the semi-harp type of a steel cable-stayed bridge. The reliability sensitivity of the bridge is also investigated.