推进技术 ›› 2009, Vol. 30 ›› Issue (3): 342-346.

• • 上一篇    下一篇

转速波动状态下涡轮泵典型故障诊断方法

夏鲁瑞,胡茑庆,秦国军   

  1. 国防科学技术大学机电工程与自动化学院;国防科学技术大学机电工程与自动化学院;国防科学技术大学机电工程与自动化学院
  • 发布日期:2021-08-15
  • 基金资助:
    国家自然科学基金(50675219);高等学校全国优秀博士学位论文专项资金(200434);新世纪优秀人才支持计划资助项目(NCET-05-0904)

Diagnosis method for turbopump typical fault under conditions of speed fluctuation

  1. Coll.of Mechatronics Engineering and Automation,National Univ.of Defence Technology,Changsha 410073,China;Coll.of Mechatronics Engineering and Automation,National Univ.of Defence Technology,Changsha 410073,China;Coll.of Mechatronics Engineering and Automation,National Univ.of Defence Technology,Changsha 410073,China
  • Published:2021-08-15

摘要: 利用涡轮泵振动信号的变换域信息可有效地检测与诊断故障。针对涡轮泵转子叶片断裂与脱落这种典型故障,首先分析其出现的原因,并从动力学的角度研究其振动特征,选择可有效反映该故障的特征频率。然而,涡轮泵转速波动会造成这些特征频率提取的困难,为此提出一种解决此难题的新思路,通过一系列变换域处理来消除转速波动对振动频率的影响,在变换域中提取出稳定的特征频率,从而解决了涡轮泵转速波动状态下该型故障诊断问题。通过涡轮泵历史试车故障数据的验证表明,通过跟踪变换域中这些特征频率的幅值变化,可以有效检测与诊断涡轮泵转子叶片断裂与脱落故障。

关键词: 涡轮泵;故障诊断;变换域处理;特征频率

Abstract: To effectively detect and diagnose its faults,it is necessary to take full advantage of the turbopump information of vibration signal in transform domain.The reasons of the typical fault of turbopump rotor blade abruption and abscission were analyzed.The vibration features of the fault were studied by the method of dynamic analysis to select the feature frequency revealing the fault effectively.However,turbopump speed fluctuation will result in the difficulty of the feature frequency extraction.So,a new way to solve the difficulty was shown.By a series of transform processing,the speed fluctuation effects on the vibration frequency was driven out to extract the stable feature frequency in transform domain.Thus,the problem of the typical fault diagnosis was solved under conditions of speed fluctuation.With the fault data in turbopump historical test,it is shown that the fault of turbopump rotor blade abruption and abscission can be detected and diagnosed effectively by tracking the amplitude variation of the feature frequency in transform domain.

Key words: Turbopump;Fault diagnosis;Transform processing;Feature frequency