| The electro-hydraulic servo system of a hydraulic experiment table is mainly used to control the displacement of hydraulic cylinder. Electro-hydraulic servo system has the advantages of high power, fast response, strong anti-jamming capability and high control precision, which is widely used in the areas of defense and industrial production. Electro-hydraulic servo system is a multi-coupled complex system consists of mechanical, electronic and hydraulic, which has the characteristics of time-varying, dead zone, nonlinearity and so on. So it’s hard to set up an accurate model. For these features of the electro-hydraulic servo system of hydraulic experiment table, we choose wavelet neural network PID as the main controller and applies grey prediction in the feedback loop of hydraulic experiment table. Grey prediction is used to compensate system’s feedback and reduce the time-varying and nonlinear characteristics. It can help main controller achieve better control effect and realize accurate control of hydraulic cylinder position.This paper consists of the following work:Firstly, analyzing the structure and working process of the hydraulic experiment table. The parameters of hydraulic components are introduced and the transfer function of hydraulic experiment table is derived. We build AMESim model of hydraulic experiment table and design interface module for united simulation.Secondly, introducing the grey system theory. The we analyze how to build the grey prediction model to predict the output of hydraulic experiment table’s system. What’s more, self-tuning prediction step is realized in different control stages to improve prediction accuracy. And its predictive effect is proved.Furthermore, switching grey prediction wavelet neural network PID controller is designed in Simulink and it unit AMESim model of hydraulic experiment table for united simulation. We compare its control effect with fixed grey prediction wavelet neural network PID controller and wavelet neural network PID controller, so we can analyze control effect of the controller we designed for system of a hydraulic experiment table.Finally, the semi-physical simulation platform is built to check control effect in actual hydraulic system. Through data acquisition card, DAQ toolbox and real time simulation, we control the displacement of hydraulic cylinder in hydraulic experiment table by controller built in Simulink.Through the simulation results of united simulation and semi-physical simulation, compared to fixed grey prediction wavelet neural network PID controller and wavelet neural network PID controller, switching grey prediction wavelet neural network PID controller achieve better control effect both in response speed and precision of system control and it meet the system’s performance requirements. |
