多电机同步系统的多模型辨识  

Multi-model identification to multi-motor synchronous system

作　　者：张今朝[1,2] 刘国海[1] 潘天红[1] 

机构地区：[1]江苏大学电气信息工程学院,江苏镇江212013 [2]嘉兴学院机电工程学院,浙江嘉兴314001

出　　处：《电机与控制学报》2009年第1期138-142,共5页Electric Machines and Control

基　　金：国家自然科学基金(60874014);教育部博士点基金项目(2005029909);江苏省自然科学基金项目(BK2007099);江苏大学高级专业人才科研启动基金项目(08JDG046)

摘　　要：针对多电机同步系统,建立了按转子磁场定向的物理模型,分析了其多变量、强耦合、非线性的复杂关系。由于该物理模型难以应用于控制器设计,为此以三台电机为研究对象,对系统进行分解,建立了速度和张力输出与各输入量之间的映射关系,进一步将系统简化为速度和张力模型。利用数据驱动原理,对输入输出样本数据进行满意c均值聚类,建立了不同工况下的局部线性子模型及相应的调度函数,系统的全局模型是各个子模型的加权综合,从而得出速度和张力的全局模型。仿真结果表明,采用所建立的简化模型能准确地拟合复杂多电机的非线性特性。A first principle model of multi-motor synchronous system was built based on field orientation and the system present complex characteristics such as multivariable, nonlinear and coupling. According to this complex system which can not be controlled effectively by the first principal model, a three motor asynchronous system was researched and decomposed into the models of speed and tension. The relationship between inputs and outputs of speed and tension was found. Furthermore, sample of system was clustered by Satisfactory Fuzzy e-Mean （SCFM） Clustering Algorithm, and the parameters of local models and scheduling function were identified. The global model of speed and tension were calculated by those local models. The results of application with the identification algorithm to simulation illustrate the performanee of the proposed algorithm.

关 键 词：多电机同步系统 满意模糊聚类 局部模型网络 线性化 

分 类 号：TP291[自动化与计算机技术—检测技术与自动化装置]