用支持矢量机建立电励磁双凸极发电机的非线性模型  

Nonlinear Modeling of Fielding-Winding Doubly Salient Generator by Support Vector Machine

作　　者：王莉[1] 宋晓峰[1] 孟小利[1] 严仰光[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016

出　　处：《电工技术学报》2004年第9期1-5,共5页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金重大资助项目(503370300)

摘　　要：作为一种新型的交流无刷电机,电励磁双凸极发电机输出经过桥式整流即可输出直流电,结构简单,可靠性高。但它是高度非线性、时变的对象,其磁链和电感均为励磁电流、相电流、转子位置的非线性函数,无法采用解析方法建立精确数学模型。支持矢量机(SVM)摒弃经验风险最小化(ERM)原则,而采用结构风险最小化(SRM)原则,可有效地应用于非线性建模。本文采用SVM建立了电励磁双凸极发电机磁链y(cbaf,,,,iiiiq)的非线性模型,所建模型拟合精度较高,其推广泛化能力也较强。所建模型为飞机高压直流发电系统的设计、动静态特性分析奠定了良好的基础。Fielding-winding doubly salient generator is a new type of brushless electrical machine, which offers an excellent balance between cost, reliability, power density, and high speed capability. So there is an excellent potential in aircraft and automobile. But it always operates within the saturation region. The phase flux linkage and inductance are a nonlinear variation of the rotor position and current. So it is difficult to derive a comprehensive mathematic model to exactly describe the behavior of the generator. Based on Structure Risk Minimization principle other than Empirical Risk Minimization principle, support vector machine can be applied in nonlinear modeling. By support vector machine, a nonlinear model of fielding-winding doubly salient generator is firstly presented in this paper. Simulation results show that this model is highly precise and less time consuming for convergence. It establishes the base of analyzing high voltage direct current aircraft electrical power generating system.

关 键 词：发电机 励磁电流 磁链 高压直流 桥式整流 转子位置 发电系统 支持矢量机 SVM 结构风险最小化 

分 类 号：TM31[电气工程—电机] TP391[自动化与计算机技术—计算机应用技术]