同步发电机参数辨识的模拟进化方法

Simulated Evolutionary Approaches to Parameter Estimation of Synchronous Machines

目前解决同步电机参数辨识最有效的途径是在线辨识。但人们经常会遇到不同试验或同一试验时辨识所得参数相差较大， 即所谓参数辨识不稳定。本文认为其原因有两方面： 一是参数是否可辨识； 二是参数辨识方法是否合适。本文在前文献［５］ 分析了同步发电机模型参数可辨识性的基础上， 以模拟进化方法为工具， 提出了全局性好、鲁棒性强的辨识方法。针对浙江衢化热电厂进行的实际应用表明， 进化策略法具有较高的精度和较少的计算量，

On line identification is a available way for synchronous machine parameter estimation A difficulty one often meets is unstable problem,i e different parameter sets may be obtained during a test or tests According to authors′ opinion,this may due to two causes,i e whether the parameters are identifiable and whether the parameter estimation approach is suitable Based on the identifiability analysis made in[5],the global and robust identification approaches to parameter estimation of synchronous machines are developed in this paper. The application results to a power plant in Zhejiang Province indicate that evolutionary strategy approach is better for synchronous machine parameter estimation