Analysis of a Polynomial Chaos-Kriging Metamodel for Uncertainty Quantification in Aerospace Applications

Metamodeling can be effective for uncertainty quantification in computational fluid dynamics simulations. In this research, we introduce modifications to our existing metamodel1 that combines a reduced polynomial chaos expansion approach and universal Kriging (RPCK) and evaluate the new metamodel for aerospace applications. Focus is given to determine which metamodel parameters most effect the solution accuracy by measuring the errors. Additionally, a new adaptive refinement algorithm is explored and the methodology is presented. Results show the metamodel’s need for robustness in aerospace engineering applications, including the non-smooth output of separated airflow.

[1]  Soshi Kawai,et al.  Kriging-model-based uncertainty quantification in computational fluid dynamics , 2014 .

[2]  H. Wackernagle,et al.  Multivariate geostatistics: an introduction with applications , 1998 .

[3]  Sonja Kuhnt,et al.  Design and analysis of computer experiments , 2010 .

[4]  V. Dubourg.  Adaptive surrogate models for reliability analysis and reliability-based design optimization , 2011 .

[5]  Bruno Sudret,et al.  Adaptive sparse polynomial chaos expansion based on least angle regression , 2011, J. Comput. Phys..

[6]  Sourajeet Roy,et al.  Fast Uncertainty Quantification in Engine Nacelle Inlet Design Using a Reduced Dimensional Polynomial Chaos Approach , 2016 .

[7]  Joe Wiart,et al.  Combining Polynomial Chaos Expansions and Kriging , 2014 .

[8]  G. Gary Wang,et al.  Review of Metamodeling Techniques in Support of Engineering Design Optimization , 2007 .

[9]  Olivier Dubrule,et al.  Cross validation of kriging in a unique neighborhood , 1983 .

[10]  François Bachoc,et al.  Cross Validation and Maximum Likelihood estimations of hyper-parameters of Gaussian processes with model misspecification , 2013, Comput. Stat. Data Anal..

[11]  Phillip Colella,et al.  A fourth-order boundary treatment for viscous fluxes on Cartesian grid finite-volume methods , 2014 .

[12]  Sourajeet Roy,et al.  Combining A Reduced Polynomial Chaos Expansion Approach with Universal Kriging for Uncertainty Quantification , 2017 .

[13]  Stefano Marelli,et al.  Rare Event Estimation Using Polynomial-Chaos Kriging , 2017 .

[14]  J. Wiart,et al.  Polynomial-Chaos-based Kriging , 2015, 1502.03939.

[15]  Marko Wagner,et al.  Geostatistics For Environmental Scientists , 2016 .

[16]  Yvan Saeys,et al.  An alternative approach to avoid overfitting for surrogate models , 2011, Proceedings of the 2011 Winter Simulation Conference (WSC).

[17]  G. Gary Wang,et al.  Review of Metamodeling Techniques in Support of Engineering Design Optimization , 2007, DAC 2006.