摘要
用支持向量机回归函数来逼近转子位置角与相磁链、相电流之间存在的非线性关系,通过对离线获取的磁化曲线的学习,得到作为转子位置估计器的支持向量回归估计函数。并将其应用于在线仿真试验中进行验证,取得了比较理想的试验结果,证明了该方法的有效性。
In this paper,Support vector regression （SVR） formed a very efficient mapping structure for the nonlinear SRM with flux linkage and current as inputs and the corresponding position as output. Given a sufficiently large training data set, the SVR built up a correlation among flux linkage, current and position for an appropriate regression function. An SVR - based position estimator was used in the online simulation test. The results proved the validity of this scheme.
出处
《微特电机》
北大核心
2007年第10期10-13,17,共5页
Small & Special Electrical Machines
基金
航空科学基金项目(04F53036)
