摘要:磁链特性是开关磁阻电机系统建模的重要基础,文中探讨了利用极限学习机处理磁化曲线簇,建立电机模型的方法。在详细分析电机非线性磁链特性的基础上,运用极限学习机的非线性回归理论,通过对开关磁阻电机进行有限元分析得到样本数据集进行训练学习,建立了电机的非参数模型,与传统神经网络、支持向量机相比,该电机模型具有更高的回归精度与更快的回归速度。仿真实验表明,该模型比较准确地反映了电机的磁链特性。
关键词:开关磁阻电机;极限学习机;非线性磁特性;建模 Abstract: Flux-linkage characteristics were vital for modeling of switched reluctance motors, and a novel approach was proposed in this paper that modeling of flux linkage in switched reluctance motors was achieved by utilizing extreme learning machine. With flux-linkage characteristics were analyzed in details, sample data were obtained by finite element analysis for non-parametric modeling using non-linear regression theory of extreme learning machine. Compared with traditional artificial neural network (such as BP neural network, RBF neural network, etc.) and support vector machine, extreme learning machine demonstrated a higher precision and a faster speed of regression. Finally, it was proved that the trained non-parametric model indicated flux-linkage characteristics accurately.
Key words: switched reluctance motors;extreme learning machine;flux-linkage characteristics;modeling
|