Intelligent aerodynamic identification method based on wavelet transform and deep learning

Wind tunnel force testing is particularly crucial for the aerodynamic characteristics of hypersonic vehicles, and the intelligent and high-precision aerodynamic identification technology serves as the key supporting technology. Due to the short effective test time of the pulse combustion wind tunnel, the inertial force vibration generated by the force measurement system is difficult to be attenuated completely at the start of the wind tunnel. At the same time, there are many interference signals in the output signal during the test, which makes it difficult to achieve high-precision aerodynamic load identification. Therefore, in order to eliminate the influence of interference signals on aerodynamic identification accuracy and obtain accurate aerodynamic loads, this paper introduces an intelligent aerodynamic identification method based on wavelet transform and deep learning. Firstly, meyer wavelet is selected as the wavelet basis, and signal processing is performed on the output signal through wavelet transform to eliminate interference signals and improve signal quality. Then, through RNN-GRU network model to identify the load. The reliability of the method is verified by the rod balance test platform. The results show that the method can effectively reduce the influence of interference signal, and the relative error between the identified load and the real load is about 2%. At the same time, the identification results of the method are compared with those of the mean value method, which further verifies the accuracy of the proposed method.