Open AccessA high-fidelity polynomial chaos modified method suitable for CFD uncertainty quantification
Author(s) -
Jiangtao Chen,
Chao Zhang,
Wei Zhao,
Xiaojun Wu
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1985/1/012042
Subject(s) - polynomial chaos , fidelity , high fidelity , polynomial , computer science , uncertainty quantification , airfoil , process (computing) , algorithm , dimension (graph theory) , mathematical optimization , mathematics , engineering , machine learning , aerospace engineering , statistics , monte carlo method , mathematical analysis , telecommunications , electrical engineering , pure mathematics , operating system
There are many sources of uncertainty in the process of numerical simulation for engineering configure. Evaluating and quantifying the uncertainty of simulation results is very important for the design and evaluation process of the industrial department. The non-intrusive polynomial chaos method is a commonly used uncertainty quantification method, but a polynomial chaos with high accuracy requires a sufficient number of high-fidelity samples, which causes “dimension disaster” for high-dimensional problems. This paper proposes a modified method. First, a low-fidelity calculation model is used to generate a low-fidelity polynomial chaos, and then a small amount of high-fidelity calculation data is used to modify the low-fidelity polynomial chaos. This method is used to analyze the parameter uncertainty of the SA model in the RAE2822 airfoil calculation. Under the premise of effectively ensuring the accuracy, the calculation time is reduced by 60% compared with the method of using the full high-fidelity calculation model.