低速工况下感应电机转子磁链在线辨识方法

On-Line Identification Method for Rotor Flux of Induction Motor Based on Subtractive Clustering

感应电机在按转子磁场定向的矢量闭环控制系统中,转子磁链及其空间角度需要在线辨识。针对电机参数在低速工况下电动机温度升高而发生变化,不能保证磁链辨识精度以及M轴与转子磁链始终重合的问题,提出了基于减法聚类转子磁链在线辨识的方法。在确定转子磁链、空间角度和各输出量之间的映射关系的基础上,把电机参数变化当作干扰项来处理,利用局部模型网络构建磁链和空间角度的全局模型,在工况发生变化时,对模型结构参数和局部模型参数进行在线调整,保证模型输出与实际输出动态拟合,解决了低速时电机参数变动影响磁链准确观测的问题,仿真结果验证了该方法的准确性和有效性。

In the vector closed loop control system for the induction motor based on rotor flux-orientation, rotor flux and its space angle need to be identified online. According to motor parameter changed with the increase of motor temperature under the low-speed condition, identification precision of rotor flux linkage and its space angle would be affected. The problem about the coincidence between M axis and rotor flux could not be ensured, so an online identification method for rotor flux identified online based on subtractive clustering was put forward. On the foundation about definited the mapping relation between rotor flux, its space angle and the output, the change of motor parameter were regarded as the interference term; the global model for flux and space angle were constructed by the local model networks; model construction parameters and local model parameters were adjusted online when the condition was changing; model the dynamic fit of output and actual output was ensured ; the problem that the observation accuracy of rotor flux was affected by the change of motor parameters was solved. The simulation results verified that this was an accurate and effective method.