The Application Of Neural Networkinverse System Method To Electro-Hydraulic Force Control System

The nonlinear problem is ubiquitous and is difficult to solve in electro-hydraulic servo system. It is caused by the throttling property, because of the existing of the servo valve, proportion valve and throttling valve, by the dead band, the back lash, the saturation. The nonlinear caused by the latter (usually called Nonlinearity in essence) has been solved by the describing function technique. But to the former, there is no satisfying method.The inverse system approach, using the feedback method to transform nonlinear systems for linear systems, then making use of the complete linear system theory to integrate system, is a major breakthrough of the nonlinear system control theory. Because of its clear physical concept is straightforward and easy to understand, it has been widely used in the theory of nonlinear control research. However, the inverse system approach, which requires charged with the mathematical model and the specific parameters of the system accurately known, practical application is very difficult; neural network control method of intelligent control theory is independent on the contrary, or not entirely dependent upon charged with the mathematical model, convenient, theoretical research difficult. Now combining inverse system method and neural network control methods ,we will receive a successful effect. Based on this thought, this paper presents a neural network-integral inverse system and its application in the area of electro-hydraulic driven system control.This paper mainly does research on electro-hydraulic driven system designing based on neural network-integral inverse system. We focus on the band integral inverse system and reverse theories in the course of introducing basic theory of neural network-integral inverse system; through the electro-hydraulic system mathematical model, which indicates that the electro-hydraulic system is equivalent to a single-input and single-output system, and the system is reversible; using neural network-integral inverse system control method, combining inverse system with the original systems together, we can not only realize the original system linear, but through the rational designing of neural networks inverse system, and will enable the pseudo-linear system of the composite pole on the complex plane arbitrary configuration, which has more desirable frequency characteristics.Experimental results show that the using of the control strategy to achieve the stability control of electro-hydraulic system, which gets a good target tracking signal, strong enough to withstand the outside interference and parameters of the load, has a strong robustness.