基于模糊 RBF 神经网络的无刷直流电机 PID 控制-《计算机技术与发展》

文章信息/Info

作者:: 顾文斌¹; 2 ; 杨生胜¹ ; 王贤良¹ ; 苑明海¹; 1. 河海大学机电工程学院,江苏常州 213022;
2. 南通河海大学海洋与近海工程研究院,江苏南通 226004

Author(s):: GU Wen-bin1; 2 ; YANG Sheng-sheng1 ; WANG Xian-liang1 ; YUAN Ming-hai1; 1. School of Mechanical and Electrical Engineering,Hohai University,Changzhou 213022,China;
2. Institute of Ocean and Offshore Engineering,Nantong Hohai University,Nantong 226004,China

Keywords:: brushless DC motor; fuzzy radial basis function; improved ant colony algorithm; LM algorithm; PID control

摘要:: 针对无刷直流电机在传统 PID 控制方式下,存在抗干扰能力差、响应速度慢以及控制精度低等问题,提出一种基于模糊径向基函数( RBF) 神经网络的无刷直流电机 PID 控制策略。首先,利用模糊控制不需要精确数学模型的优势,能够克服传统 PID 对数学模型的依赖性,而模糊控制规则的制定主要取决于经验,因此,将 RBF 神经网络与模糊控制相结合,可以提高其自学习、自适应能力。此外,利用改进蚁群算法对模糊神经网络的参数进行初始化,避免了在传统聚类方法下陷入局部最优的困境,同时提高了模糊神经网络的收敛速度,然后将列文伯格-马夸尔特算法融入模糊神经网络,以确定神经网络的权值,并提高神经网络的训练速度。最后,在 Simulink 中通过仿真与其他控制策略进行对比。仿真结果表明,模糊 RBF 神经网络 PID 控制策略相较于其他控制策略,在无刷直流电机控制系统中具有更优异的控制性能。

Abstract:: In order to solve the problems of poor anti - interference ability, slow response speed and low control precision under thetraditional PID control mode of brushless DC motor,a PID control strategy of brushless DC motor based on fuzzy radial basis function( RBF) neural network is proposed. First of all,the use of fuzzy control does not need the advantage of precise mathematical model,canovercome the dependency of traditional PID on mathematical model, and the formulation of fuzzy control rules mainly depends onexperience. Therefore,the combination of RBF neural network and fuzzy control can improve its self-learning,self-adaptive ability. Inaddition,the improved ant colony algorithm is used to initialize the parameters? ? ? ? ? of? the fuzzy neural network,which avoids the dilemma offalling into local optimization under the traditional clustering method and improves the convergence speed of the fuzzy neural network.Then,the Levenberg-Marquardt algorithm is integrated into the fuzzy neural network to determine the weight of the neural network andimprove the training speed of the neural network. Finally,the simulation is compared with other control strategies in Simulink,whichshows that the fuzzy RBF neural network PID control strategy has better control performance than other control strategies in the brushlessDC motor control system.