A new aerodynamic identification technology for short-time hypersonic wind tunnels while considering inertial force interference

Abstract

When aerodynamic force tests are performed in hypersonic wind tunnels, the test model in the force measurement system (FMS) will be subjected to transient impact and induced transient vibration due to the high-speed airflow generated during start-up. The output signal of the FMS will contain the real aerodynamic signal and the inertial force signal caused by model vibration. The inertial force signal completely overcomes the aerodynamic signal, strongly influencing the test accuracy of the aerodynamic force and even making the aerodynamic force signal completely unusable. Therefore, this study develops a novel intelligent aerodynamic identification method based on deep residual shrinkage network (DRSN) deep learning technology and applies this method to aerodynamic force tests in hypersonic wind tunnels. This method can achieve intelligent identification and filtering of inertial force interference signals and instrument noise signals, and improve the test accuracy of aerodynamic forces. Also, this study performed a numerical simulation experiment and a test-bed experiment; verified the results through hypersonic wind tunnel test signals; and compared the identification results of the DRSN network method with the identification results of the traditional CNN network method, the mean method and the FFT method. From comparison results, the aerodynamic force identification accuracy of the DRSN network method proposed in this paper is shown to be nearly ideal and always remains above 95%. Thus, the proposed method effectively reduces the interference of instrument noise and inertial force, markedly improving the accuracy and reliability of the hypersonic wind tunnel aerodynamic force test.