基于趋势分析的发动机振动故障识别

Vibration Fault Diagnosis for Aero-Engine Based on Trend Model

关键词:  航空发动机  故障诊断  振动  趋势模型  识别函数

Key words:  Aeroengine  Fault diagnosis  Vibration  Trend model  Criterion function

摘要:

总结十种航空发动机整机故障的振动特征发现,单纯从频率角度分析几乎所有故障模式都表现为转子基频,但是不同故障的振动幅值发展趋势却有所不同。基于此提出了基于趋势模型的发动机振动故障识别方法。讨论了发动机整机振动趋势的数学描述及不同故障模式的趋势发展特征；建立了平稳波动模型、周期摆动模型、线性发展模型和阶跃突起模型分别描述发动机振动正常状态和三种典型故障模式；为了有效识别上述四种模型,提出了模型识别准则并实现了识别算法。最后以四组发动机实测趋势数据作为案例进行验证,结果表明该方法可以有效地区分故障模式,验证了该方法的正确性和工程实用性。 

Abstract:

By summarizing characteristics of body vibration faults for ten aero-engines, frequency analysis indicates that most of fault mechanisms are related with rotor fundamental frequency. However, different fault mechanism has different vibration trend. Thus, using trend model can effectively identify a certain fault mechanism. Firstly this paper discussed the mathematical description of engine vibration trend and the characteristics of each trend development. Secondly steady vibration model, shimmy model, linear vibration model and step change vibration model were set to describe the normal and three typical fault situations. In order to efficaciously distinguish the four trend models mentioned above, recognition criteria was set up and recognition algorithm was realized. Finally, four actual groups data for engines were used to verify the correctness and engineering practicability of this method. The results demonstrate that vibration fault diagnosis for aero-engine based on trend model can distinguish the fault mode. 