无量纲指标趋势分析法在风电机组齿轮箱故障预警诊断中的应用  

Non-dimensional parameters trend analysis method in the early warning and diagnosis of wind turbine gearbox failures

作　　者：顾煜炯[1] 贾子文[1] 尹传涛[1] 任玉亭 

机构地区：[1]华北电力大学新能源电力系统国家重点实验室,北京102206 [2]国华能源投资有限公司,北京100007

出　　处：《振动与冲击》2017年第19期213-220,共8页Journal of Vibration and Shock

基　　金：神华集团科技创新项目(SHJT-12-24);华能集团科学技术项目(HNKJ-H27);中央高校基本科研业务专项基金(2016XS27)

摘　　要：针对风电机组运行工况复杂多变,早期故障特征提取十分困难的问题,提出无量纲指标趋势分析法以实现风电机组齿轮箱故障的早期诊断。通过阶比重采样方法对齿轮箱振动信号进行预处理,将非平稳时域信号转化成平稳角域信号,引用无量纲因子指标反映机组早期故障程度变化趋势;利用指标趋势分析法建立风电齿轮箱故障预警模型,并通过实验室数据和现场数据进行分析验证。结果表明,该方法能够比较准确的对风电机组齿轮箱早期故障进行诊断,并能初步估测故障发生时间,达到故障预警目的。Wind turbine’s feature of early failure is hard to extract on account of its fluctuant working conditions, an non-dimensional parameters trend analysis method is utilized to deal with turbine gearbox early failure diagnosis. Equal time interval time-domain signals can be transformed into uniform angle domain signals by order resampling, and the dimentionless parameters’ tendencies are used to sense the level of turbine gearbox’s early faults. A model of early fault diagnosis, which is used for turbine gearbox, is created based on parameters trend analysis, the lab data and actual failure data are analyzed by it. The results show that the method is qualified in turbine gearbox early failure diagnosis analysis, and the moment of failure can be estimated, which is achieve early warning’s aim.

关 键 词：风电齿轮箱 阶比重采样 无量纲指标 趋势分析法 早期故障诊断 

分 类 号：TK83[动力工程及工程热物理—流体机械及工程]