基于BP网络横向磁场永磁电机调速系统的设计  

A Novel and Better Design of TFPM(Transverse Flux Permanent-magnet Motor) PI Speed Control System Using BP Neural Network

作　　者：杨宁[1] 史仪凯[1] 袁小庆[1] 黄磊[1] 

机构地区：[1]西北工业大学机电学院,陕西西安710072

出　　处：《西北工业大学学报》2011年第5期824-828,共5页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(50275125);陕西省科学研究发展计划(2009K9-16)资助

摘　　要：文章针对横向磁场永磁电机非线性、强耦合的特点以及传统PI调节器存在超调和控制参数无法自适应的缺陷,提出一种基于BP神经网络的横向磁场永磁电机PI控制方案。该方案在分析电机矢量控制的基础上,建立了四相横向磁场永磁电机的数学模型,利用BP神经网络对PI调节器参数进行在线整定,通过自调整学习速率改进BP神经网络学习能力,实现电机跟踪性能和抗负载扰动性能的提高。实验结果表明,所采用的控制策略可行,在参数突变和突加负载时,均能够达到跟踪额定转速的效果,使系统具有良好的鲁棒性。Aim. The introduction of the full paper reviews some papers in the open literature and then proposes the novel and better design mentioned in the title. Sections 1 through 4 explain our design and its evaluation. Section 1 briefs the TFPM mathematical model. Fig. 2 in section 2 is a schematic showing the structure of TFPM control system structure. In section 3, eq. （4） gives the PI （proportional integral） basic control algorithm. Section 4, utili- zing formulas mentioned in section 3, gives simulation PI speed control response curves in Figs. 6 through 9. The core of sections 1 through 4, stated in a different way, consists of： （1） Fig. 3 in section 3 is a schematic demonstrating the BP （backward propagation） network topology structure, and then eq. （8） is used in forward algorithm of BP neural network to ensure the exact neural point number of hide level; （2） eq. （14） is introduced in reverse algorithm to accelerate the convergence rate and improve the anti-interference performance of PI adjuster; （3） Table 1 in subsection 3.3 indicates the initial weighted value of the whole BP neural network; （4） simulation results show preliminarily that the proposed strategy not only makes the motor track the scheduled speed well when the load is suddenly added on but also strengthens the systemg robustness.

关 键 词：横向磁场永磁电机 PI调节器 BP神经网络 矢量控制 

分 类 号：TM351[电气工程—电机] TP273[自动化与计算机技术—检测技术与自动化装置]