DRNN神经网络逆控制方法在DDV伺服系统中的应用  

Application of DRNN Inverse Control in DDV Server System

作　　者：杜海[1] 任杰[1] 

机构地区：[1]飞行自动控制研究所,西安710065

出　　处：《机械工程师》2007年第5期45-47,共3页Mechanical Engineer

摘　　要：对于直接驱动阀(DDV)式作动器伺服系统,由于很难找到最佳的参数匹配以及回路中存在死区非线性因素,使得传统的PID控制无法取得令人满意的控制效果。为此提出一种动态对角回归神经网络(DRNN)逆控制方法,该逆系统控制方法是将对象的逆模型作为控制器加入到控制回路里面,并与控制对象串联构成一个伪单位环节,使得系统的输出能够跟踪系统的输入,经过仿真实验,可以看出该神经网络逆控制比单一PID控制器的动态性能要好。Because it is difficulty to match the parameters using the conventional PID control, and there is nolinear dead zone in valve loop of the servo system which is named DDV, so it is difficulty to achieve a satisfied result using the conventional PID control.A kind of DRNN inverse controller for this servo system is put forward in this paper. through adding the inverse model of the servo system into the control, it is forming a fake part with the object so as to making the output tracking the input. From the emulator result we can see that the dynamic performance of the compound inverse control is better than that of the conventional PID control.

关 键 词：直接驱动阀 对角回归神经网络 逆控制 

分 类 号：TH164[机械工程—机械制造及自动化]