【作者】 马瑞；

【导师】 刘延泉；

【作者基本信息】 华北电力大学（河北） ， 控制理论与控制工程， 2007， 硕士

【摘要】 本文研究小波神经网络与自适应逆控制的结合,对具有惯性和滞后特性的对象控制作出探讨。目的在于提出一种消除大纯滞后特性的神经网络自适应逆控制方法,消除控制对象受到的扰动,根据期望的参考模型确定动态特性。文中使用小波神经网络自适应逆控制系统对各种典型对象:线性对象、非线性对象,时变对象以及具有纯滞后和惯性特性的对象作了仿真,尤其对主蒸汽温度对象做了仿真试验,从理论上证明了小波神经网络自适应逆控制的是切实可行的。更进一步的,针对逆控制系统的不足之处,增加了预测环节,使得对纯滞后对象的控制达到了实时控制。更多还原

【Abstract】 In paper, we focus on combination between wallet neural network and adaptive inverse control, Furthermore,we take a discussion on control for object with inertia and lag. It is our goal to propose a new control method upon adaptive inverse control based on wallet neural network aiming at eliminating noise on object, and to determine dynamic performance. At end of paper, simulation is made by adaptive inverse control based on neural network on steam temperature which has great inertia and lag to verify its practicality. We make simulation experiments on representative system such as linear-system, nonlinear-system, time-varying system and system with inertia and lag and on steam temperature especially, to prove advantages of adaptive inverse control based on neural network theoretically. Against shortages of Adaptive Inverse Control, We add predication part to make real time control on lag system come true.更多还原