Polynomial Chaos Expansion Based on Dimension-adaptive Algorithm

• 1、College of Science, Inner Mongolia University of Technology, Hohhot 010051

Abstract：To alleviate the dimension curse during the projection solving of polynomial chaos expansion (PCE), this paper presents a PCE technique based on Dimension-adaptive algorithm, and verifies the feasibility of the proposed method through taking solid rocket motor ignition under low temperature as an example. The main contributions of this work are as follows: (1) Presenting the Dimension-adaptive algorithm based on global sensitivity analysis (GSA). (2) Through establishing a non-uniform grid based on GSA, developing non-invasive solution method based on Galerkin projection. (3) Applying the proposed method to uncertainty quantification (UQ) of solid rocket motor ignition under low temperature. The result indicates that by means of comparing with some conventional non-invasive method , the proposed method is able to raise the computational efficiency significantly and meet the accuracy requirement for high-dimension uncertainty problem.

Keywords： Polynomial chaos expansion Global sensitivity analysis Uncertainty quantification (UQ) Dimension-adaptive