华电技术 ›› 2017, Vol. 39 ›› Issue (10): 6-9.

• 研究与开发 • 上一篇    下一篇

基于高斯混合模型的汽轮机转轴故障诊断方法

  

  1. 1.大唐华银攸县能源有限公司,湖南株洲 412307; 2.湖南大唐先一科技有限公司,长沙 410007
  • 出版日期:2017-10-25 发布日期:2017-11-22

Gaussian mixture modelbased steam turbine shaft failure

  1. 1.Datang Huayin Youxian Energy Company Limited, Zhuzhou 412307, China; 2.Hunan Datang Xianyi Technology Company Limited, Changsha 410007, China
  • Online:2017-10-25 Published:2017-11-22

摘要:

基于K均值聚类算法与高斯混合模型,通过对某电厂320MW机组历史运行数据的训练,建立汽轮机转轴的高斯混合模型,计算实时状态信息与模型中各工况中期望值的相似度,进行工况隶属分类,再结合汽轮机转轴故障征兆知识库中的故障模式进行故障类型匹配,最终实现了汽轮机转轴的故障诊断。

关键词:

Abstract:

Based on  K means clustering algorithms and Gaussian mixture model, and a power plant’s 320MW unit historical operation data, Gaussian mixture model of steam turbine shaft was built. By analyzing the similarity of realtime operation and expectation value from model, combined with steam turbine shaft failure database, ultimately realized steam turbine shaft failure diagnostic.

Key words: