基于模糊物元分析的变桨系统状态评估方法  

Method of Condition Assessment of Pitch System Based on Fuzzy Matter-Element Analysis

作　　者：张磊[1] 朱希盼 张琨[1] 车璐 

机构地区：[1]河北工业大学控制科学与工程学院,天津300130

出　　处：《中南民族大学学报（自然科学版）》2017年第4期88-94,共7页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：河北省自然科学基金资助项目(F2015202231)

摘　　要：针对现有风电机组状态评估方法实时可靠性差和过多人为因素影响的问题,提出了模糊物元分析评估方法,以实现对机组状态的准确评估.选择故障率较高的变桨系统为研究对象,分两步构建了变桨系统状态评估模型:1)分别以3σ准则和四分位分析法对变桨参数分类处理,避免对参数分布的主观评判;用ANFIS算法对数据进行训练来减小极端值影响,获得多特征参数故障检测结果.2)基于模糊物元分析理论,将上一步多特征参数检测结果作为模糊量值代入物元评估模型中,实现了检测结果模糊值与等级评价指标的统一.应用该方法对风电机组实际运行状态进行了测试,结果表明:与传统二元决策方法相比,能够明显反映变桨系统的运行状态,具有更好的评估效果.从定性角度对比分析,该方法较模糊综合评判方法、传统物元分析方法在变桨系统状态评估方面更有优势.In order to improve the real-time reliability of wind turbine and to solve the effect of artificial factors,fuzzy matter-element analysis is proposed.The pitch system,which has high failure rate,is researched in this paper.The condition assessment model of pitch system is built in two steps：（1） To avoid the subjective judgment of the parameter distribution,this thesis uses 3σ rule and quartile analysis method to get the boundary data sets.The ANFIS algorithm is used to train the data to reduce the extreme value,and results of multi-feature fault detection are obtained.（2） Results from fault detection are applied in condition assessment model based on fuzzy matter-element analysis to realize the unification of fuzzy value of detection result and grade assessment.The method is applied to test the actual operation state of wind turbine.The results show that compared with the traditional two-element method,it can obviously reflect the pitch system,and has better assessment effects.From the qualitative point of view,this method is more advantageous than the fuzzy comprehensive evaluation method and the traditional matter-element analysis method in the condition assessment of the pitch system.

关 键 词：变桨系统 故障检测 模糊物元分析 状态评估 风电机组 

分 类 号：TM743[电气工程—电力系统及自动化]