Research Of Strategy Surplus Force Of Aircraft Rudder Electro-hdraulic Load Simulater

Aircraft rudder electro-hydraulic load simulator simulate the flight aerodynamic load of the aircraft rudder in the process of flying under laboratory conditions. with the advantages of good controllability, without damaging. The simulator is a passive force servo control system, the system produce surplus torque when the aircraft rudder run. As a result of the existence of surplus torque seriously affect the control performance of the system. So how to suppress and eliminate the surplus torque disturbance, has become urgent research subject in the field of China's civil aviation aircraft flight control.Based on the present domestic research situation on the basis of rudder electro-hydraulic loading system, firstly analyzing the structure and working principle of the aircraft rudder electro-hydraulic loading system, determining the model of system main components according to the technical index of the load simulator. Secondly, building the mathematical model of electro-hydraulic servo valve, hydraulic cylinder, force sensor, and the of buffer spring, so building the whole mathematical model of the system, and study the mechanism of surplus torque qualitatively. thirdly, proposing the compound structure that combine hydraulic cylinder speed feedforward signal and conventional PID feedback controller, solving the problem that inhibit the surplus torque under the low frequency and large load gradient, but the inhibition effect is poor at high frequency and small load gradient. So proposing compound controller structure that combine hydraulic cylinder speed feedforward signal and the single neuron PID controller based on RBF neural network on-line identification. Finally, from the point of view of the control system, the identification effect of RBF neural network is poor, so it is necessary to research the nearest neighbor clustering algorithm which have a high clustering quality, make its application in RBF neural network. And because of the clustering center of the nearest neighbor clustering algorithm and the clustering radius is not easy to get, and thus the clustering effect, so need to improve the algorithm. Using the ant clustering algorithm to optimize the RBF neural network. Validating the feasibility and effectiveness of verify the conclusion through the computer simulation.