泵控缸电液位置伺服系统建模研究

摘要

针对泵控缸电液位置伺服系统存在非线性和时变性,从而难以对其精确建模的问题,研究了该系统机理建模和智能建模的方法.通过实验仿真表明,机理建模精度低、泛化能力差；模糊建模和BP神经网络建模可较好的拟合系统固有的非线性和时变性特性；基于遗传算法的BP神经网络较好的解决了BP神经网络易陷入局部最小的问题,具有建模精度高、泛化能力好的特点.%An exact mathematic model is difficult to establish for the electro-hydraulic position servo systems of pump-controlled cylinder because of its nonlinearity and time variability. The mechanism modeling and intelligent modeling techniques were proposed to order to solve the problem. The simulation results show that the mechanism modeling technique has poor generalization and low modeling accuracy,otherwise, the fuzzy modeling and BP neural-network modeling technique can fit the nonlinearity and time-variability of the system. BP neural network modeling based on genetic algorithm offers a better solution to the local minimum in BP neural network, so the neural network modeling based on genetic algorithm has better generalization and high modeling accuracy.