基于最小二乘支持向量机的开关磁阻电动机建模

Modeling of Switched Reluctance Motors Based on LS-SVM

开关磁阻电动机(SRM)的磁化曲线族是电机建模及性能分析的基础,文中探讨了利用最小二乘支持向量机处理磁化曲线族,建立电机模型的方法。在分析电机非线性磁特性的基础上,运用最小二乘支持向量机(LS-SVM)的回归理论,通过对磁路有限元分析法(FEM)得到的样本数据集进行学习,建立了电机的最小二乘支持向量机模型。与以往的神经网络建模方法相比,该模型在小样本情况下有更好的泛化能力及更快的学习速度。仿真实验表明,该模型比较准确地反映了电机的磁特性,可用来进行SRM其它性能指标的分析。

Considering nonlinear magnetization characteristics of a switched reluctance motor （SRM）, the paper first presents least squares support vector machine（LS-SVM） as a new tool to develop the model of the SRM. The basic premise of the LS-SVM regression is that it forms a very efficient mapping structure for the nonlinear SRM. The sampled data set obtained from the experimental SRM by the finite elements method（FEM）, is comprised of magnetization data with position and current as inputs, and the corresponding flux-linkage as output. Given a sufficiently large training data set, the LS-SVM can build up a correlation among position, current and flux-linkage. Compared with the models based on artificial neural networks（ANN） methods, the proposed model has better capability of generalization and better convergent speed if a small data set is available. The simulation results show that the proposed model can be used to predict accurately the relationship of flux-linkage and current, and can analysis other functional indexes of the SRM.