| For the country, the wind tunnel is an important research infrastructure in the field ofaerospace and aviation. The main task of the wind tunnel control is to improve the quality ofthe wind tunnel flow field and to improve the accuracy of the data of the wind tunnelexperiments. However, as the multidimensionality, the complexity of wind tunnel, and theindirect measurement of the Mach number, the accurate control of the Mach numberbecomes a key point and also a difficult point in the wind tunnel control. In the practicalengineering application, we usually use the traditional PID controller to control the Machnumber in the wind tunnel. The parameters of the controller are determined by theexperience of the technical staff. To obtain the accurate data of the experiments, we need tocarry out the blowing process repeatedly. So it’s a waste of human and material resources.A solution based on the Neural Network PID Control is proposed to control the WindTunnel Mach number. First of all, the whole measure-control system of the FD12windtunnel and the test process of the subsonic and transonic flow field are introduced, in orderto get a general understanding of the whole Mach number control system. Second, thebasics of the Artificial Neural Network are introduced. And then, the math model of thecontrolled object is set up using the Artificial Neural Network identification. Somesimulations are done to verify the correctness of the model. Finally, compared with thetraditional PID controller, the Neural Network PID controller is selected to control thesystem. The whole system is simulated by the software MATLAB, and the controlperformance can be compared by the control of the traditional PID controller and the SingleNeuron PID controller.Through the simulation results, the dynamic performance and the anti-interferenceability of the system is greatly improved by the control of the Single Neuron PID controller.The expected target of the control performance is reached. |
PDF Full Text Request