03.4 - Applied Aerodynamics
TRANSFER LEARNING FOR REDUCED-ORDER MODELING OF TRANSONIC FLOWS USING MULTIFIDELITY DATA
J. Kou¹, C. Ning¹, W. Zhang¹; ¹Northwestern Polytechnical University, China
Combining data from multiple sources to construct accurate reduced-order flow models requires accurately capturing the correlation between multi-fidelity data. Therefore we propose multi-fidelity flow reconstruction framework based on transfer learning, which maintains low computational cost, while significantly improve the prediction accuracy over the traditional method.rn