Research On Servo System Control Of Two - Axis Turntable With Variable Load

Turntable as precision mechanical-electrical integration products are widely used in aerospace, marine, robotics, artillery and numerical control machine and other fields. Such as control the azimuth and elevation angle of radar, missile and artillery; the simulation for ships and satellites actuator; simulate the yaw, pitch, roll and others attitudes of various hull; control the robot motion and so on. In this paper, taking the weapon servo system as engineering background, study its performance. This servo system requires high precision and must adapt to a wide range of load, so the conventional control methods can not meet its performance requirements. In this paper, presenting a control method that combine Wavelet neural network and adaptive control, the simulation results show that this control method not only can improve the precision of servo system but also has strong robustness for the system parameter perturbation and external disturbance.The main research work of this thesis focuses on the following aspects:First, made depth analysis for the structural characteristics and working principle of the servo system, designed the system hardware circuit that takes TMS320F2812 as the core. Built the mathematical model for permanent magnet DC torque motor, and on this basis to finish the servo system modeling, analyzed the uncertainty and nonlinear factors in the system.Secondly, for the problem that it is difficult to build the precise mathematical model for the servo system because there are many problems like nonlinear and uncertainty and so on. Identification results show that identification based on wavelet neural network can well adapt to rapid changes in the control environment, and has good dynamic performance.Finally, proposed the indirect wavelet neural network self-turning control program that combined with wavelet neural network and adaptive control, the program use wavelet neural network identification that based on genetic algorithm to provide a gradient for wavelet neural network adaptive controller in real time, so it can make the learning correction of the controller performed correctly. Experimental analysis, the indirect wavelet neural network self-tuning control program has better control performance.