优选初值-微变搜索算法及其在双馈风力发电机参数辨识中的应用  

Optimal initial value and micro-variation search algorithm and its application in parameter identification of doubly-fed wind-power generator

作　　者：张仰飞[1,2] 袁越[1] 陈小虎[2] 辛龙彪 吴博文[1] 

机构地区：[1]河海大学电气工程学院,江苏南京210098 [2]南京工程学院电力工程学院,江苏南京211167 [3]大唐漳州风力发电有限责任公司,福建厦门361004

出　　处：《电力自动化设备》2010年第2期19-21,共3页Electric Power Automation Equipment

基　　金：国家科技支撑计划(2008BAA14B02);江苏省自然科学基金项目(BK2008366;BK2009353);江苏省六大高峰人才资助项目(07-E-026KXJ07113)~~

摘　　要：为解决现有辨识方法初值不稳定和辨识结果不精确的问题,提出了优选初值-微变搜索算法。该算法包括2个步骤:首先优选初值,把测量数据对应于不同运行状态进行分类,结合相应的简化模型直接求解参数,或组合起来求解参数,得到参数向量的值,需要时,把测量数据和已知参数向量的值作为已知条件,采用搜索类算法搜索其他参数向量值;再进行微变搜索,把所有测量数据作为已知条件,以已求得的参数向量值作为初值点,进行微变搜索,即在已知参数向量各分量上作微小变化,按照排列组合的方法得到其周围的若干点,分别计算其目标评价函数值。比较和保留最好的点,继续缩小步长进行微变,直至精确化定位到最优解。仿真算例显示,该方法能有效地精确辨识双馈风力发电机的参数。Aiming at the instable initial value and inaccurate identification result,the OIVMSA （Optimal Initial Value and Micro-variation Search Algorithm） is proposed. It consists of two steps：1. Initial value optimization,which classifies the field measurements according to operating states,calculates the initial value of parameter vector with simplified model or grouped models, and if necessary, takes the measurements and the known values of some parameter vectors as known conditions to search the initial values of other parameter vectors with search-based algorithm;2. Micro-variation search,which takes all measurements as known conditions and the known initial values to calculate the relevant values of parameter vector objective evaluation function by slightly changing the values of the parameter vector components to get several surrounding points based on the permutation and combination method,compares and keeps the better points and repeats the micro-variation search with smaller step-length until the optimized solution is obtained. Simulation indicates that the parameters of doubly-fed wind-power generator can be precisely and effectively identi6ad with OIVMSA.

关 键 词：辨识算法 双馈风力发电机 参数辨识 优选初值-微变搜索算法 初值 目标点 

分 类 号：TM614[电气工程—电力系统及自动化] TM315