永磁风力发电机的结构参数与定位力矩之间存在复杂关系,难以迅速建立起准确通用的数学模型。针对上述问题,提出了一种新的定位力矩优化方法。首先,通过拉丁超立方抽样(LHS)与有限元分析(FEA)相结合的方法获取支持向量机(SVM)回归的训练样本;其次,利用训练样本构建定子槽口宽度、极弧系数、偏心距、气隙长度以及永磁体厚度等结构参数与定位力矩之间的回归模型;最后,基于此模型,应用布谷鸟搜索(CS)算法对永磁风力发电机的结构参数进行寻优,将优化后的参数输入ANSYS进行仿真分析。仿真结果表明,定位力矩得到了有效削弱,验证了该方法的正确性和优越性。
Considering the complex relationship between structural parameters and cogging torque of the permanent magnet wind generator, it is difficult to speedily establish a precise and universal mathematical model. To tackle this problem, a new method is proposed to optimize cogging torque.Firstly, training samples for SVM are obtained by combining Latin hypercube sampling(LHS) with Finite element analysis(FEA). Secondly, a regression model of the stator slot width, pole-arccoefficient, polar arcs eccentricity, air gap length, as well as the thickness of permanent magnet and cogging torque, is constructed by applying these training samples. Finally, based on this model, cuckoo search(CS)algorithm is used to optimize the structural parameters of permanent magnet wind generator. Optimized structural parameters are input into the ANSYS software for simulation analysis. Final findings show that the cogging torque is effectively weakened. The validity and superiority of this method are thus verified.