BP神经网络预测模糊控制液压马达性能研究  

Research on Performance of hydraulic Motor Based on BP Neural Network Prediction Fuzzy Control

作　　者：王晓晶[1] 刘美珍 陈帅 刘昊 WANG Xiao-jing;LIU Mei-zhen;CHEN Shuai;LIU Hao(School of Mechanical and Power Engineering,Harbin University of Science and Technology,Harbin 150080)

机构地区：[1]哈尔滨理工大学机械动力工程学院,黑龙江哈尔滨150080

出　　处：《控制工程》2020年第8期1394-1400,共7页Control Engineering of China

基　　金：国家自然科学基金面上项目(51975164);黑龙江省普通本科高等学校青年创新人才培养计划(UNPYSCT-2017096)。

摘　　要：针对连续回转马达高度非线性和摩擦、泄漏等不确定性,严重影响系统的跟踪性能,提出了一种基于BP神经网络预测的模糊控制策略。该控制策略采用了BP神经网络进行伺服系统的输出预测,并计算了预测输出与给定输入间的未来误差和误差变化率,根据专家经验编制电液伺服系统模糊控制规则,设计模糊控制器,从而调整电液伺服系统的控制量,实现对连续回转电液伺服马达位置状态的实时跟踪。通过Simulink仿真表明,基于BP神经网络预测的模糊控制和传统PID控制相比,有效的提高了电液位置伺服系统的位置跟踪精度和抗干扰能力,缩短了响应时间,拓宽了系统响应频带。Due to the uncertainties caused by high non-linearity,friction,leakage in continuous rotary motor,the tracking performance of the system is affected seriously,the strategy of fuzzy control based on BP neural network prediction is proposed.BP neural network model is adopted to predict the output of servo system and calculate the future error between the original input signal and predictive output,as well as the error change rate.According to expert experience,the fuzzy control rule of electro-hydraulic servo system is formulated and then fuzzy controller is designed,so that the control variable is adjusted to realize the real-time tracking of continuous rotary electro-hydraulic servo motor’s position.By comparing with the traditional PID control’s simulation result in Simulink,the result shows that by adopting the fuzzy control algorithm based on BP neural network prediction,the position tracking performance and anti-interference ability of electro-hydraulic position servo system are improved extremely,and the response time is decreased,the system response frequency band is widen as well.

关 键 词：连续回转电液伺服马达 BP神经网络控制 模糊控制 

分 类 号：TP137[自动化与计算机技术—控制理论与控制工程]