基于径向基神经网络的无位置传感器开关磁阻电机采煤机牵引系统  

Position sensorless control of switched reluctance motor for shearer traction system based on RBF neural network

作　　者：张旭隆[1,2] 谭国俊[1,2] 蒯松岩[1,2] 吴涛[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221008 [2]江苏省电力传动与自动控制工程技术研究中心,江苏徐州221116

出　　处：《煤炭学报》2011年第9期1570-1574,共5页Journal of China Coal Society

基　　金：江苏省自然科学基金资助项目(BK2009526);中国矿业大学青年科研基金资助项目(2009A025)

摘　　要：提出了无位置传感器SRM应用于采煤机牵引系统的可行性,分析了开关磁阻电机非线性的磁链特性及实时计算方法,建立了以磁链和相电流为输入、转子位置角度为输出的径向基(RBF)神经网络模型,以轴编码器实时获得的转子位置角度为学习样本,对SRM的数学模型进行了在线学习,给出了学习算法和训练步骤。以TMS320F2812 DSP为控制芯片,开发完成了1套18.5 kW三相12/8极无位置传感器开关磁阻电机样机,并进行了采煤机牵引实验。实验结果表明,系统运行可靠,具有良好的静动态性能,位置检测误差≤2°。Feasibility of using sensorless SRM for shearer traction system was proposed.Nonlinear characteristics of SRM and its real-time calculation method were analyzed.Radial basis function（RBF） neural network of rotor position estimation for sensorless SRM drive was established,with two input variables：flux linkage and phase current.Real-time rotor position angle obtained from shaft encoder was adopted as learning sample data,on-line learning algorithms and training procedures were also given.Sensorless control of SRM based on RBF neural network was achieved by TMS320F2812 DSP,18.5 kW three-phase 12/8 pole sensorless controller was set up.Traction experimental results of shearer show that the system has a good dynamic performance and high accuracy position detection with maximum error less than 2°.

关 键 词：径向基 无位置传感器 开关磁阻电机 采煤机牵引系统 

分 类 号：TD421.65[矿业工程—矿山机电]