电液负载模拟器神经网络辨识器及控制器设计  

Neural Network Identifier and Controller Design on Electro-hydraulic Load Simulator

作　　者：刘晓琳 李卓 陈立东 LIU Xiao-lin;LI Zhuo;CHEN Li-dong(College of Electric Information and Automation,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学电子信息与自动化学院,天津300300

出　　处：《科学技术与工程》2020年第2期834-840,共7页Science Technology and Engineering

基　　金：天津市自然科学基金(17JCYBJC18200);中央高校基本科研业务费项目(3122018C002);中国民航大学研究生科技创新基金项目(201930);大学生创新创业训练计划项目(201910059120,201910059121)。

摘　　要：针对传统控制方法无法解决飞机舵机电液负载模拟器受多余力等非线性因素严重干扰的问题,给出了一种基于神经网络辨识器及控制器的复合控制结构,结合了神经网络系统辨识与自适应实时控制的工作特点。根据电液负载模拟器控制结构及工作原理,采用反向传播(back propagation,BP)神经网络辨识器在线辨识,获得系统辨识模型以替代理论数学模型。然后,采用Adaline神经网络控制器实时控制,利用系统误差信号与BP神经网络反向递归计算Adaline网络权值调整信息,获得系统控制参数,实现复合控制器的有效监督与智能控制。最后利用Matlab进行实验验证,仿真结果表明:该方法能够提高系统控制精度,多余力消扰率达92%;并且可以有效模拟飞机舵机所受力载荷的变化情况,实现系统指令信号快速、准确、稳定的加载。A traditional control method is unable to solve serious disturbance of electro-hydraulic load simulator of aircraft rudder affected by several nonlinear factors,such as surplus force.To solve the problem,a compound control structure was developed based on neural network identifier and controller in combination of the neural network system identification and adaptive real-time control.According to the control structure and working principle of the electro-hydraulic load simulator,a back propagation(BP)neural network identifier was used for online identification system,from which a system identification model was obtained to replace the theoretical mathematical model.Then,the Adaline neural network controller was used for real-time control,the system error signal and BP neural network were used to calculate the weight adjustment information of the Adaline network,the system control parameters were obtained,and the effective supervision and intelligent control of the composite controller were realized.Finally,the experiment results were verified using MATLAB.The simulation results show that this method improves the control accuracy of the system,reaches high disturbance rejection rate(92%)of surplus force,simulates effectively the change of the force load on the aircraft steering gear,and realizes the fast,accurate,and stable loading of the system command signal.

关 键 词：电液负载模拟器 神经网络 系统辨识 自适应控制 多余力 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TP301.6[自动化与计算机技术—计算机系统结构]