PDI-HFP: An intelligent method for heat flux prediction on hypersonic aircraft based on projection depth images

Abstract

We propose a novel intelligent method to predict the heat flux on hypersonic aircraft. This method considers the aircraft shape and the inflow conditions as inputs and directly outputs overall surface heat flux values. Specifically, PDI-HFP first projects the aircraft shape onto two-dimensional (2D) images in six directions and then utilizes well-trained neural networks to predict the corresponding heat flux images. Finally, PDI-HFP reconstructs the predicted surface heat flux from 2D space to 3D space, and an interpolation method is then implemented to obtain the heat flux distribution on the surface of the 3D aircraft. To the best of our knowledge, this is the first work to apply deep learning techniques to 3D heat flux prediction on arbitrary types of surface grid. Extensive experimental results demonstrate that the values of the heat flux predicted by our method are very close to those generated by CFD simulation. More importantly, compared with CFD simulation, the use of PDI-HFP effectively shortens the computational time, achieving a speedup by a factor of 200–1000, depending on the aircraft shape.