永磁球形步进电机转子位置检测的全局优化  

Global Optimization on Detecting the Rotor Position of Permanent Magnetic Spherical Stepper Motor

作　　者：王群京[1] 雍爱霞[1] 张胜虎[2] 

机构地区：[1]合肥工业大学,安徽省合肥市230069 [2]解放军电子工程学院,安徽省合肥市230037

出　　处：《中国电机工程学报》2007年第33期11-16,共6页Proceedings of the CSEE

基　　金：国家自然科学基金项目(05377010);安徽省自然科学基金项目(03044103)。~~

摘　　要：介绍了永磁球形步进电机模型,在转子球面上进行随机编码,光电传感器的输出用来表示转子的位置。对转子球体的三维旋转进行了群描述;针对96个传感器的输出所构成的位置空间进行搜索,而传统的单点搜索会陷入局部极小值。定义目标函数,根据目标函数值的二维关系图的特点,采用遗传算法对函数进行全局寻优,并嵌入最速下降算子来加强遗传算法的局部搜索能力,得出当前转子的空间位置。对于远距离目标节点,仿真表明这种多点并行搜索方法能够快速地收敛到目标位置,精确度取决于传感器的分辨率和搜索步长。The permanent magnetic spherical stepper motor model is introduced in this paper, random encoder is painted on the surface of the rotor, and outputs of optoelectronic sensors are used to express the rotor position. This paper gives group representation of the spherical rotor in three-dimension space; it searches the rotor orientation in state space that consists of outputs of 96 sensors, but traditional single-nod-search method may get into local minima. Defining target function, the paper uses genetic algorithm to seek global optimum according to two-dimension plot of target function, and sets in steepest decent operator to strengthen genetic algorithm＇s local search, thus gets the current rotor orientation. For the remote target node, simulation shows parallel search method of more-nods could fast convergence to the target orientation, the precision lies on the resolution and search step length.

关 键 词：球形电机 光电传感器 分辨率 遗传算法 最速下降算法 

分 类 号：TM35[电气工程—电机]