摘要
针对风电机组变工况运行造成设备故障诊断困难的问题,提出了主成分-灰色关联分析方法,解决风电机组齿轮箱故障诊断问题。通过阶比重采样方法对信号进行预处理,消除原始数据非线性带来的不良影响;考虑到信号能量变化会对分析带来误差,用无量纲参数作为故障诊断的特征数据;应用主成分-灰色关联分析法,对各特征参数赋予权重,增强了分析数据与故障特征间的关联性,提高了故障诊断精度。试验及实际应用结果分析表明,文章所提出的方法能够较准确地对风电机组齿轮箱故障进行诊断。
In view of the difficulties in failure diagnosis of wind turbine, which is because of equipment varying duty, a PCA(principal component analysis)-GRA(grey relational analysis)method is proposed to deal with it. Utilizing the method of order resampling, the nonlinearity data bad influence can be eliminated; in consideration of analytic error by signal energy fluctuation, non-dimensional parameters are regarded as characteristics of the fault diagnosis data; the weight of every characteristic parameter can be given by PCA-GRA, which not only promote the relationship between failure mode and analyze data,but also improve the accuracy of fault diagnosis. The test and actual analysis results shown that the improved method could diagnosis wind turbine gearbox failure effectively, which has a good practical significance.
出处
《可再生能源》
CAS
北大核心
2017年第4期508-514,共7页
Renewable Energy Resources
基金
神华集团科技创新项目(SHJT-12-24)
中央高校基本科研业务专项基金(2016XS27)
关键词
风电齿轮箱
阶比重采样
无量纲参数
主成分分析
灰色关联分析
wind turbine
order resampling
non-dimensional parameter
principal component analysis
grey relational analysis