• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

电力建设 ›› 2014, Vol. 35 ›› Issue (8): 125-129.doi: 10.3969/j.issn.1000-7229.2014.08.022

• 发电技术 • 上一篇    下一篇

基于相对主元分析的风电机组塔架振动状态监测与故障诊断

周进,房宁,郭鹏   

  1. 新能源电力系统国家重点实验室(华北电力大学),北京市 102206
  • 出版日期:2014-08-01
  • 作者简介:周进(1985),男,硕士研究生,研究方向为新能源发电技术,E-mail:285503103@qq.com; 房宁(1986),男,硕士研究生,研究方向为新能源发电技术; 郭鹏(1975),男,博士,教授,研究方向为新能源发电技术。
  • 基金资助:
    国家自然科学基金(51207049);新能源电力系统国家重点实验室开放基金项目(LAPS13011)。

Tower Vibration Fault Diagnosis and Monitoring for Wind Turbines Based on RPCA

ZHOU Jin, FANG Ning, GUO Peng   

  1. State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources, North China Electric Power University, Beijing 102206, China
  • Online:2014-08-01

摘要: 振动信号是风电机组数据采集与监控系统(supervisory control and data acquisition system, SCADA)中一类重要变量。以风电机组SCADA运行数据为基础,首先结合风机运行原理详细分析了导致塔架振动的主要因素。进而采用相对主元分析(relative principal component analysis,RPCA)和某风电机组2011年3~5月份的SCADA运行数据,建立了覆盖塔架正常工作状态的RPCA振动模型,计算得出监控统计量Hotelling T2(简称T2)和平方预测误差(squared prediction error,SPE)。采用塔架振动RPCA模型,准确检测出风电机组变桨系统故障,验证了所研究方法的有效性。

关键词: 风电机组, 塔架振动, 状态监测, 相对主元分析(RPCA), 数据采集与监控系统(SCADA), 建模

Abstract: Vibration signal is one kind of important variables in supervisory control and data acquisition system (SCADA) for wind turbines. Based on SCADA data and wind turbine operating theory, this paper firstly analyzed the factors that had great influence on tower vibration. Then relative principal components analysis (RPCA) was used to model the tower vibration during the normal work of wind turbine combined with the SCADA data of a wind turbine between March and May 2011. Two statistic variables Hotelling T2 and SPE (squared prediction error) were calculated. The RPCA tower vibration model then was used to accurately detect the blade angle asymmetry fault, which could prove the effectiveness of this method.

Key words: wind turbines, tower vibration, condition monitoring, relative principal component analysis(RPCA), supervisory control and data acquisition system(SCADA), modeling