基于小波神经网络的电动负载模拟器的复合控制  

Compound Control of Electric Load Simulator Based on Wavelet Neural Network

作　　者：杨瑞峰[1,2] 付梦瑶 郭晨霞[1,2] 张鹏[1,2] 

机构地区：[1]中北大学仪器与电子学院,山西太原030051 [2]中北大学仪器科学与动态测试教育部重点实验室,山西太原030051

出　　处：《液压与气动》2016年第3期14-18,共5页Chinese Hydraulics & Pneumatics

基　　金：国家自然科学基金(51375462);国家国际科技合作项目(2014DFR70650);高等学校博士学科点专项科研基金(20121420110003)

摘　　要：为了提高电动负载模拟器的信号跟踪精度和多余力矩抑制能力,在分析系统结构和工作原理的基础上建立了电动负载模拟器系统的完整数学模型。针对电动负载模拟器中存在的力矩跟踪精度问题,提出了一种前馈补偿和基于小波网络的PID控制相结合的复合控制方法。利用改进的前馈补偿法抑制多余力矩,基于小波网络的PID控制器可以在线调整PID参数补偿系统的非线性环节,提高系统动态性能。仿真结果表明,复合控制器对多余力矩有良好的抑制效果,跟踪精度满足要求,和传统PID控制相比,系统鲁棒性得到显著提高。To improve the signal tracking accuracy and suppression of surplus torque of electric load simulator system, the structure and principle of electric load simulator were analyzed, on the basis of which a complete mathematical model of the system was established. A compound control method based on the combination of feed-forward compensation and PID control based on wavelet neural network was proposed to solve the torque tracking of the e- lectric loading system. The improved feed-forward compensation method was used to restrain the surplus torque. The PII） controller based on wavelet neural network was used to adjust the PID parameters on-line and to improve the dynamic performance of the system. Simulation results demonstrate that the compound controller has a good inhibito- ry effect on the surplus torque and the tracking accuracy meets the requirements. The system robustness is improved significantly comparing with the traditional PID control.

关 键 词：电动负载模拟器 多余力矩 小波神经网络 复合控制 

分 类 号：TH137[机械工程—机械制造及自动化] TP273[自动化与计算机技术—检测技术与自动化装置]