基于模糊神经网络的无刷直流电动机能量回馈控制  

Energy Feedback Control of Brushless DC Motor Based on Fuzzy Neural Network

作　　者：费蓝冰[1] 崔方[2] 魏玉平[2] 

机构地区：[1]江苏大学,江苏镇江212013 [2]江苏科技大学,江苏镇江212003

出　　处：《微特电机》2014年第1期50-53,共4页Small & Special Electrical Machines

基　　金：国家自然科学基金项目(5037507);江苏省数字化制造技术重点实验室基金项目(HG-DML-0909)

摘　　要：为了提高无刷直流电动机能量回馈系统的能量回馈效率,针对无刷直流电动机能量回馈模型非线性、时效性的特点,设计了一个基于模糊神经网络控制的无刷直流电动机能量回馈控制系统。根据影响能量回馈效率的关键因素,以制动电流与反馈电流的偏差、当前转速作为控制的输入,占空比作为控制的输出,为了提高控制精度,通过一个五层的神经网络,用误差反向传播网络学习算法,调整模糊逻辑控制器的输入和输出参数,使得控制系统具备自适应能力。最后对所设计开发的无刷直流电动机能量回馈控制系统进行了仿真和实物试验,试验结果表明,基于模糊神经网络的控制系统,回馈效率高,鲁棒性强。According to the characteristics of nonlinear and timeliness of energy feedback for brushless DC motor rood el ,a brushless DC motor control system which based on fuzzy neural network to control energy feedback was proposed to improve brushless DC motor regenerative energy feedback system efficiency. According to the key factors influencing the energy feedback efficiency, the deviation between brake and feedback current, the current speed were taken as the input of control ler,and the output of controller was the adjustment of the pulse width of registers. In order to improve the control precision, a neural network of five layers with BP learning algorithm was found,the control system had the adaptive ability by adjusting the relationship between input and output parameters of fuzzy logic controller. Tile proposed and developed energy feedback control system of brushless DC motor simulation and physical test results show that the control system found by fuzzy neural network has high energy feedback efficiency and strong robustness.

关 键 词：无刷直流电动机 能量回馈 模糊神经网络 误差反向传播网络 自适应 

分 类 号：TM33[电气工程—电机]