船用永磁同步电机是一种非线性、强耦合、多变量的复杂系统，使用常规的PID对其进行速度控制时难以达到理想的效果。为了改善船用永磁同步电机调速系统的性能，设计了一种新的速度控制器-径向基（RBF）神经网络分数阶PIαDβ控制器。利用径向基神经网络的自学习和自训练的功能，对控制器的参数进行在线优化，以便使控制器在未知的系统中能够具有快速的适应能力和较好的控制性能。将设计的控制器应用于船用永磁同步电机的速度控制回路中，并在高速度、大负载扰动的条件下对其进行仿真实验。结果表明，使用了RBF神经网络分数阶PIαDβ控制器的电机控制系统，具有良好的动态响应能力和较强的扰动抑制能力。
Marine permanent magnet synchronous motor is a kind of nonlinear, strong coupling and multi variable complex system. It is difficult to achieve the desired effect when using conventional PID to control the speed. In order to improve the performance of the speed control system of marine permanent magnet synchronous motor, a new speed controller-radial basis function (RBF) neural network fractional PIαDβ controller is designed. The parameters of the controller are optimized online by the self-learning and self-training functions of radial basis function neural network, so that the controller can have the fast adaptability and good control performance in an unknown system. The designed controller is applied to the speed loop of marine permanent magnet synchronous motor, and the simulation experiment is carried out under the conditions of high speed and large load disturbance. The results show that the motor control system with RBF neural network fractional order PIαDβ controller has good dynamic performance and strong disturbance rejection ability.
2018,40(4): 94-98 收稿日期：2016-11-03
DOI：10.3404/j.issn.1672-7649.2018.04.019
分类号：TM351
作者简介：庞科旺(1963-),男,副教授,研究方向为大功率电力电子装置及电机拖动
参考文献：
[1] 薛定宇, 赵春娜. 分数阶系统的分数阶PID控制器设计[J]. 控制理论与应用, 2007, 24(5):771-776.XUE Ding-yu, ZHAO Chun-na. Fractional order PID controller design for fractional order system[J]. Control Theory and Applications, 2007, 24(5):771-776.
[2] 薛薇, 李永丽, 路鸦立. 永磁同步电机分数阶智能积分调速控制[J]. 电机与控制学报, 2015, 19(5):67-73.
[3] 吴振宇, 赵亮, 冯林. 基于分数阶PID控制器的智能车控制[J]. 控制工程, 2011, 18(3):401-404.
[4] BAI Duan-yuan, WANG Chun-yang. Parameter calibration and simulation of fractional PID controller for hydraulic servo system[C]//In Proc. of the 28st the Chinese conference Decision and Control conference (CCDC), Yinchuan, China in May 28-30, 2016:1704-1708.
[5] 梁涛年, 陈建军, 王媛, 等. 分数阶系统模糊自适应分数阶PIλDμ控制器[J]. 北京工业大学学报, 2013, 39(7):1040-1045.
[6] 毛书军, 盛贤军. 一种神经网络分数阶PID控制器的实现[J]. 计算机应用, 2014, 34(s2):166-168.MAO Shu-jun, SHENG Xian-jun. Implementation of an neural network fractional order PID controller[J]. Journal of Computer Applications, 2014, 34(S2):166-168.
[7] 王春阳. 分数阶控制系统设计[M]. 北京:国防工业出版社, 2014:4-21.
[8] 李姗姗, 赵春娜, 关永, 等. 分数阶微积分定义的一致性在HOL4中的验证[J]. 计算机科学, 2016, 43(3):23-26.LI Shan-shan, ZHAO Chun-na, GUAN Yong, et al. Formalization of consistency of fractional calculus in HOL4[J]. Computer Science, 2016, 43(3):23-26.
[9] 周品. MATLAB神经网络设计与应用[M]. 北京:清华大学出版社, 2013:232-238.
[10] 阮毅, 陈伯时. 电力拖动自动控制系统:运动控制系统[M]. 北京:机械工业出版社, 2009:243-246.
[11] 王瑞萍. 基于分数阶控制器的永磁同步电动机速度控制研究[D]. 广州:华南理工大学, 2012.
