中国机械工程

• 机械基础工程 • 上一篇    下一篇

基于数据挖掘的风电机组齿轮箱运行状态分析

贾子文;顾煜炯   

  1. 华北电力大学能源动力与机械工程学院,北京,102206
  • 出版日期:2018-03-25 发布日期:2018-03-21
  • 基金资助:
    中央高校基本科研业务费专项资金资助项目(2016XS27)
    Fundamental Research Funds for the Central Universities(No. 2016XS27)

Wind Turbine Gearbox Operation State Analysis Based on Data Mining

JIA Ziwen;GU Yujiong   

  1. School of Energy,Power and Mechanical Engineering,North Electric Power University,Beijing,102206
  • Online:2018-03-25 Published:2018-03-21
  • Supported by:
    Fundamental Research Funds for the Central Universities(No. 2016XS27)

摘要: 针对风电机组齿轮箱运行状态识别困难的问题,基于数据挖掘思想,提出改进的非线性状态估计方法,得到了机组参数与运行状态之间的关联机制;通过因子分析,有效减小样本数据空间,提高了样本数据的解释能力;根据机组实际运行特点,运用滑动窗口与异常率的方法对预测残差进行分析,保证了最终分析结果的准确性与可信性。结合实际案例分析可知,该方法较传统分析方法能够更及时、准确地察觉风电机组齿轮箱异常状态。

关键词: 风电机组齿轮箱, 数据挖掘, 非线性状态估计方法, 因子分析

Abstract: Aiming at the problems of wind turbine gearbox operation conditions which were difficult to be identified, a correlation mechanism between turbine parameters and operating status was established by an improved NSEM, which was on the concept of data mining. Using factor analysis method, the sample data space was reduced and the interpret ability of sample data was improved. By the actual operation characteristics of the unit, sliding window and exception rate method were used to analyse predictive residual, which ensured the accuracy and credibility of the final analysis results. Combined with actual case analysis, the method may be more timely and accurately than traditional one to find out abnormal states of wind turbine gearboxes.

Key words: gearbox of wind turbine, data mining, nonlinear state estimate method(NSEM), factor analysis

中图分类号: