| The mine sweeping plough electro-hydraulic servo system is a typical nonlinear time-varying system, so traditional control methods are very difficult to meet the demands of control accuracy. Neural internal model control is a control method based on inverse control, which includes the advantages of internal model control and neural adaptive control and fully applies the strong function approximation ability of neural network. Meanwhile, it's simply designed and easily adjusted, the difficulty of nonlinear system modeling and controlling is solved, too. In the control process, this method has good dynamic response, as well as the response error is small. So it is often used in the control of nonlinear systems to achieve good results.This paper includes the following main work:First of all, the theory, hardware structure and component parameters of mine sweeping plough system are analyzed, the hydraulic circuit model of mine sweeping plough is designed by the software of AMESim. Then, a Simulink model of data collection is built. Finally, with the co-simulation achieved, the input and output datas of coulter lift hydraulic circuit are acquired.Because BP neural network has slow speed of convergence and it's easy to cause local extremum, GA-BP neural network is used to construct the system model. By analyzing the error curve of system output and network output and the track of network training, it's found that GA-BP neural network can improve the identification efficiency and the probability of finding the global optimum. Also, it has a strong ability of macro-searching.Traditional PID and neural internal model control systems are adopted for mine sweeping plough to control depth. By analyzing the error signals comparison of sinusoidal and step signal response curves, it can be seen that, compared with traditional PID controller, neural internal model controller has better control effect and robustness, the results and comparisions clearly show its validity. |
PDF Full Text Request