Search

第53回流体力学講演会/第39回航空宇宙数値シミュレーション技術シンポジウム (2021年6月30日-7月2日. 日本航空宇宙学会 : 宇宙航空研究開発機構(JAXA)オンライン会議), The 53rd Fluid Dynamics Conference / the 39th Aerospace Numerical Simulation Symposium (June 30 - July 2, 2021. The Japan Society for Aeronautical and Space Sciences : Japan Aerospace Exploration Agency (JAXA), Online meeting), Evolutionary algorithms (EAs) have been widely used in design optimization that often requires expensive evaluations (e.g., computational fluid dynamics: CFD). An artificial neural network (ANN) based surrogate model is used in the EA optimization routines to reduce the time for these expensive evaluations. The ANN can model the relationship between many design variables and objective functions in a single surrogate model, unlike other surrogate models (e.g., Kriging). In this study, a genetic algorithm (GA) coupled with a dynamically retrained ANN is proposed and applied to multi-objective transonic airfoil shape optimization where aerodynamic performances are evaluated with CFD. The proposed method is shown to converge more quickly towards the Pareto-optimal front with fewer CFD evaluations compared to a stand-alone GA, proving the efficacy of ANN as the surrogate model in the GA., 形態: カラー図版あり, Physical characteristics: Original contains color illustrations, 資料番号: AA2130027008, レポート番号: JAXA-SP-21-008

International audience; It is sometimes desirable to delve further into how the inputs affect the output in design optimization and uncertainty analysis. Surrogate models such as Gaussian Process Regression and support vector regression are useful for such tasks and can be further enhanced by introducing advanced post-processing methods. This paper investigates Shapley Additive Explanation (SHAP) as a tool to aid surrogate-assisted data-driven analysis. In particular, surrogate-enabled SHAP analysis allows visualization of the input-output relationship in a meaningful way using summary SHAP plots and SHAP dependence plots. Some importantinformation that can be extracted and visualized from SHAP includes the importance of input variables (i.e., global sensitivity analysis), nonlinearity level, and level of interactions. The benefits of Shapley values for engineering analysis using surrogate models are demonstrated in three engineering test problems.

In this paper, we propose the use of the composite kernel function (CKL) technique for the multi-fidelity Kriging (MFK) surrogate model. The principle of MFK-CKL is to automatically learn the best combination of single kernels in both low- and high-fidelity Kriging to create a more accurate Kriging model. The combination is in the form of a weighted sum in which the weights are treated as extra hyperparameters. We implement the CKL into the recursive co-Kriging approach. It is relatively straightforward to do the same for other MFK approaches. Demonstrations on a set of non-algebraic problems show the high efficacy of MFK with CKL, outperforming the single kernel approach in terms of approximation error. Besides, the use of CKL successfully eliminates the non-trivial process of manual kernel selection in multi-fidelity Kriging.