基于GRNN的永磁力矩电机复杂轨迹进给位置误差预测研究  

Research on Prediction of Complex Trajectory Feed Position Error of Permanent Magnet Torque Motor Based on GRNN

作　　者：于嘉龙 彭宝营[1] 侯明鹏 杨庆东[1] 耿冬冬 YU Jia-long;PENG Bao-ying;HOU Ming-peng;YANG Qing-dong;GENG Dong-dong(School of Mechanical and Electrical Engineering,Beijing Information Science&Technology University,Beijing 100192,China;Beijing National Innovation Institute of Lightweight Ltd.,Beijing 100084,China)

机构地区：[1]北京信息科技大学机电工程学院,北京100192 [2]北京机科国创轻量化科学研究院有限公司,北京100084

出　　处：《组合机床与自动化加工技术》2019年第12期61-64,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金项目(51575056);北京市教育委员会科技计划项目(KM201711232001)

摘　　要：为了提高力矩电机转子位置的精度,考虑影响力矩电机位置精度的主要因素,使用广义回归神经网络(GRNN)建立了力矩电机转子位置误差预测模型。该模型采用实验台运行的正弦轨迹数据为训练样本,三角波轨迹运行数据为测试样本。选取40组正弦波轨迹数据和10组三角波轨迹数据进行仿真预测和验证。以正弦波信号的指令位置和指令速度作为模型的输入,以三角波信号的位置误差作为输出。结果表明建立的力矩电机转子位置误差预测模型的精度要高于其他神经网络(BP、Elman),文中所建立的GRNN模型能够有效预测力矩电机复杂轨迹进给位置误差。In order to improve the accuracy of the rotor position of the torque motor and consider the main factors affecting the position accuracy of the torque motor, a generalized regression neural network(GRNN) is used to establish the rotor position error prediction model of the torque motor. The model uses the sinusoidal trajectory data of the experimental platform as the training sample, and the triangular wave trajectory operation data is the test sample. 40 sets of sine wave trajectory data and 10 sets of triangular wave trajectory data were selected for simulation prediction and verification. The command position and command speed of the sine wave signal are used as the input of the model, and the position error of the triangular wave signal is used as the output. The results show that the accuracy of the rotor motor position error prediction model is higher than that of other neural networks(BP, Elman). The GRNN model established in this paper can effectively predict the complex path feed position error of torque motor.

关 键 词：永磁力矩电机 复杂轨迹 神经网络 误差预测 

分 类 号：TH115[机械工程—机械设计及理论] TG659[金属学及工艺—金属切削加工及机床]