Application research of deep learning technology in flow field analysis of scramjet engines
The flow field reconstruction method based on deep learning is an effective means for detecting the evolution of wave structures inside the combustion chamber of scramjet engines,and it holds significant importance for controlling the main flow dynamics of the flow field.To accurately capture shock wave structures in complex sce-narios,a deep learning-based algorithm for reconstructing and segmenting flow field wave structures is proposed.Leveraging the powerful feature extraction capabilities of convolution,the algorithm first utilizes pressure data to di-rectly generate flow field shadow images through a flow field reconstruction model.Subsequently,a super-resolu-tion reconstruction algorithm is employed to enhance the image resolution.Finally,a segmentation model is used to extract the wave structures in the flow field.The model is trained and tested on a dataset constructed from experi-ments with a direct-connected supersonic pulse combustion wind tunnel at a Mach number of 2.5,using different equivalence ratios of hydrogen fuel in a scramjet engine.Experimental results demonstrate that the proposed model achieves optimal performance in flow field reconstruction and accurate segmentation of shock waves.
deep learningflow fieldwave structuresegmentation algorithm
NETL
NSTL
万方数据
