智能优化PI参数的永磁同步电机控制  

Intelligent Optimization PI Parameters Tuning for Servo Systems with Permanent Magnet Synchronous Motor Driving

作　　者：崔业兵[1] 向方明[2] 朱遵义[2] 许敬[1] 

机构地区：[1]南京理工大学,江苏南京210094 [2]南京工程学院,江苏南京211102

出　　处：《电机与控制应用》2013年第8期23-28,共6页Electric machines & control application

基　　金：南京理工大学自主科研专项计划资助项目(ZDJH02)

摘　　要：针对永磁同步电机控制系统控制参数较难整定问题,提出一种基于神经网络和遗传算法极值寻优的PI控制器参数离线整定方法。利用电机控制系统的试验数据训练BP神经网络,并选择预测精度最高的神经网络模型作为遗传算法的优化目标函数,然后利用遗传算法搜索使电机控制系统的动态响应性能最好的PI控制器参数组合。通过永磁同步电机控制系统的仿真,表明系统采用优化后的PI参数能够快速、平稳而又准确地跟踪输入指令变化,抗干扰能力强。The control parameter tuning for servo systems is difficulty, a new method of off-line tuning proportional integral（PI） coefficients for a permanent magnet synchronous motor （PMSM） drives was proposed, which based on error back propagation （ BP ） , artificial neural network （ ANN ） and genetic algorithm （ GA ）. Artificial neural network was used to identify the whole system using test data, and the highest prediction precision neural network model was chosen as optimal objective function of the genetic algorithm, then optimal values of PI controller coefficients were obtained through using genetic algorithm to research, which ensuring the perfect dynamic response performance of the motor system . The servo systems driving by PMSM was used to carry out motor speed control experiment. The results adopted the optimized parameters showed that the output could follow the input instruction change rapidly, steadily and accurately, and it had strong anti-interference ability as well.

关 键 词：遗传算法 神经网络 永磁同步电机 优化 

分 类 号：TM351[电气工程—电机]