双种群教与学算法在磁轴承PID控制器中的应用  

Application of Double Populations Teaching-Learning-Based Optimization in PID Controller of Magnetic Bearings

作　　者：张建生[1,2] 吴璇[1] 王一夫[1] 马啸宇[1] 卜平凡 

机构地区：[1]南通大学电气工程学院,江苏南通226019 [2]常州工学院电子信息与电气工程学院,江苏常州213002

出　　处：《轴承》2017年第8期32-36,共5页Bearing

基　　金：国家自然科学基金项目(51175052)

摘　　要：针对主动磁轴承系统的开环不稳定性,提出采用改进双种群教与学算法与PID结合的方法实现对系统的闭环控制。该算法在教与学算法对PID参数的优化中加入教师反向学习和学生变异学习,解决基本算法的早熟问题,提高解的精度。运用MATLAB/Simulink搭建系统模型进行仿真试验,结果表明:与基本教与学算法和遗传算法相比,该算法收敛到最优参数值的速度更快,PID控制磁轴承系统的响应速度更快,具有更好的动态性能和稳态性能。The open-loop instability of active magnetic bearing system is studied,and the method of improved double populations teaching-learning-based optimization combined with PID is proposed to realize closed-loop control of the system. The algorithm integrates teacher backward learning and student variation learning during optimization of PID parameters in teaching-learning optimization,which solve premature problem about basic algorithm and improve precision of solution. The system model is built to simulate experiment in MATLAB/Simulink. The results show that compared with basic teaching-learning-based optimization and genetic algorithm,the algorithm faster converge to optimal parameter values,the PID controller faster response and has better dynamic and steady-state performances.

关 键 词：主动磁悬浮轴承 教与学优化算法 双种群 反向学习 PID参数优化 

分 类 号：TH133.3[机械工程—机械制造及自动化] TP273.2[自动化与计算机技术—检测技术与自动化装置]