Font Size: 

The present work suggests an efficient method for reliability assessment using sequential algorithm for stochastic response surface method (SRSM) in a sparse grid framework. The stochastic response surface is formed by orthogonal Hermite polynomial basis. The unknown coefficients of the response surface are evaluated using moving least squares technique. For constructing the response surface, collocation points (as in the conventional SRSM) are replaced by the sparse grid scheme which reduces the number of function evaluations. Additionally, the sparse grid is populated sequentially based on the optimization process for finding the most probable failure point. After constructing the sequential SRSM, reliability assessment is conducted using importance sampling. Numerical study shows the effectiveness of the proposed sequential SRSM in terms of accuracy and reduction in time-exhaustive evaluation of the original performance function.

Reliability Analysis, Polynomial Chaos Expansion, Moving Least Squares, Hermite Polynomial, Sparse Grid, Sequential Modelling