基于BP神经网络的伺服速度控制参数自整定研究  

Study on Speed Control Parameters Auto-tuning of Servo System Based on BP Neural Network

作　　者：周佳[1] 卢少武[1] 周凤星[1] 

机构地区：[1]武汉科技大学信息科学与工程学院,武汉430081

出　　处：《组合机床与自动化加工技术》2016年第7期70-72,77,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金项目(51405349)

摘　　要：针对PMSM伺服系统PI控制参数自整定方法控制效果不理想、整定效率低的问题,提出一种基于BP神经网络的伺服系统速度控制参数自整定方法,该方法利用BP神经网络具有以任意精度逼近任意非线性函数的能力,构造BP神经网络速度PDFF控制器,对伺服系统速度控制参数进行在线整定,改善常规速度PI控制器的控制效果,最后通过仿真实验进行了验证。仿真结果表明:与常规PI控制方法相比,该方法稳定有效,控制精度高,收敛速度快,控制效果更好。引入了BP神经网络的伺服系统,提高了动态性能,增强了系统稳定性和快速性。In order to solve the problem of PI control parameters auto-tuning of PMSM servo system about being low efficiency, this paper presents a speed control parameters auto-tuning algorithm of servo system based on BP neural network. The neural network was used because it could change parameters itself on line. The aim of the algorithm is to improve the control effect of PI controller. It is verified through simulation fi-nally. The simulation results show that, compared with PI control algorithm, the proposed algorithm is sta-ble and effective, and has high precision, fast convergence rate and better control effect. Based on the BP neural network, the dynamic performance, the study speed and the stability of the servo system is improved obviously.

关 键 词：伺服系统 永磁同步电机 BP神经网络 PDFF参数整定 

分 类 号：TH166[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]