Efficient reliability assessment of structural dynamic systems with unequal weighted quasi-Monte Carlo Simulation

An efficient method for reliability assessment of structural dynamic systems is proposed. In the proposed method, the reliability is evaluated by the equivalent extreme value distribution of structural dynamic systems, where an estimator-corrector scheme based on the principle of maximum entropy with fractional moments as constraints is adopted to derive such distribution. The fractional moments, which are of paramount importance to the efficiency and accuracy of reliability assessment, can be obtained by a multidimensional integral. To calculate such an integral, a new approach named unequal weighted quasi-Monte Carlo Simulation is put forward. In this approach, different strategies for the construction of unequal positive weights, the rearrangement of sampling points to be integration points and the determination of required number of points are explored. The integral for the evaluation of fractional moments is then determined efficiently based on the obtained integration points and the weights straightforwardly. Numerical example is studied to verify the proposed method. The investigations indicate that the proposed method is of accuracy and efficiency for reliability assessment of structural dynamic systems, with only a few hundreds of deterministic dynamic response analysis for a practical problem involving multiple random parameters.