Adaptive Control Strategy Based On Neural Network Of The Wind Tunnel Flow Field Mach Number Study

With the development of aerial and air space technology, the design and research of all kinds of aircrafts more and more depends on the Wind-tunnel test. The quantity of data from Wind-tunnel flow test, especially the transonic-supersonic characteristic, influences the aerodynamic character of the aircraft directly .The main task of control system of Wind-tunnel is how to increase the controlling accuracy, to improve the character of flow field of wind-tunnel, and to obtain more precise data. It is also the base of designing aircraft. This thesis has studied the design of controller of the parameter of FL26 Wind-tunnel-Mach, and it's adaptive control system.The studying object of this thesis is intermittent-flow, transonic-supersonic FL26 Wind-tunnel. Based on studying the structure and principle of FL26 Wind-tunnel and the control tactics of Mach in China and in the world, analyzing the tunnel's character and existing problems, the total construct of Mach adaptive-decoupling controller of FL26 Wind-tunnel has been put forward. Then the arithmetic of NN(neuron network) is introduced, the key questions of the FL26 Wind-Tunnel control system are studied, the solving schemas are given, and the characters of system is analyzed by simulation. Finally, the total simulation model of Mach control system is put out and the software is designed.The main completed tasks of the thesis are as following:(1) On the basis of widely studying and summarizing the worldwide existed achievements of theoretic and technologies of Mach control tactics applying in the flow fields of wind tunnel, the model of Mach auto-adaptive decoupling is set up.(2) Mach identification on-line is solved, and the model of Mach predictive schema of NNI (Neural Network Identification) is developed.(3) The question of the parameter coupling of FL26 Wind Tunnel-----total press andmach is solved and the decoupling model is developed.(4) The controller of fence of flow field of FL26 Wind-tunnel--SNSC (Single Neuron Servo Controllers designed and improve servo-control system of the fence system auto-adaptation.(5) To develop the Mach auto-adaptive controller of the outer loop of Mach control system--NNCPID(Neural Network Controller and PID).(6) Based on the former works, the double closed loop control model of Mach control system is developed, and the characters of Mach control system are analyzed.