摘要
风电机组的故障诊断是大型风电场运行中亟待解决的问题。由于风力发电的特殊性,大部分风电场都是在边远地区,风机与风机之间的距离更高,不能像火电或水电等设施,以方便检查,因此如何实现故障诊断的风力发电机组设备的状态是特别重要的。本文利用多变量状态估计(Multivariate State Estimation Technique)方法对齿轮箱的温度进行状态监测,通过对设备正常工作状态下的历史数据进行学习,对系统各个参数之间的关系进行定义,通过相关性分析来建立正常运行状态下多个相关变量间的内在非线性模型。然后,利用滑动窗口的统计方法,计算残差均值,当平均值曲线超出阈值范围时,设备运行异常。
The prediction of wind turbine faults becomes a urgent problem in large-scale wind farm operation.Becauseofthep articularity of wind power, most wind turbines are built in remote areas and the distance between wind turbines is very huge. As-well as the height of the wind turbines is very high, it can not be repaired conveniently like other coal-fired power plants.So it is urgenttofindthediagnosticmethodofwindturbineswhichisbasedonothertheory. The method of Multivariate State Estimation Technique is used to predict the temperature of the gearbox in this paper. Through the studying of the history data, the inherent nonlinear model is conducted. An appropriate threshold is set to measure the condition of the wind turbine.
出处
《仪器仪表用户》
2015年第6期58-61,共4页
Instrumentation
关键词
滑动窗口统计
齿轮箱温度
残差
MSET
MSET
statistics sliding window
gearbox temperature
residuals
作者简介
张艳霞（1988-），女，山东济宁人，在读硕士研究生，主要从事风电机组状态监测与故障预测方面的研究。
