推进技术 ›› 2020, Vol. 41 ›› Issue (10): 2308-2315.DOI: 10.13675/j.cnki.tjjs.190325

• 结构 强度 可靠性 • 上一篇    下一篇

基于粒化模糊熵的机载燃油泵故障诊断

戴邵武1,陈强强1,4,戴洪德2,李娟3   

  1. 1.海军航空大学 岸防兵学院,山东烟台 264000;2.海军航空大学 航空基础学院,山东 烟台 264000;3.鲁东大学 数学与统计科学学院,山东 烟台 264000;4.海军92728部队,上海 200040
  • 发布日期:2021-08-15
  • 基金资助:
    山东自然科学基金面上项目(ZR2017MF036);国防科技项目基金(F062102009)。

A Fault Diagnosis Method of Airborne Fuel Pump Based on Granulation and Fuzzy Entropy

  1. 1.College of Coastal Defense,Naval Aviation University,Yantai 264000,China;2.College of Basic Sciences for Aviation,Naval Aviation University,Yantai 264000,China;3.College of Mathematics and Statistics,Ludong University,Yantai 264000,China;4.Naval 92728,shanghai 200040,China
  • Published:2021-08-15

摘要: 由于机械系统的复杂性,机载燃油泵振动信号的随机性表现在不同尺度上,因此需要对振动信号进行多尺度分析。为了实现机载燃油泵的故障状态特征提取,以模糊熵作为机载燃油泵振动信号的基本特征,提出了基于模糊信息粒化和模糊熵的机载燃油泵故障诊断方法。首先,采用模糊信息粒化方法对振动信号进行粒化处理,得到包含最小值、中值、最大值三组模糊信息粒;其次,计算模糊信息粒的模糊熵值;最后,将熵值作为特征向量,输入基于粒子群优化支持向量机建立的分类器。将该方法应用于机载燃油泵及轴承实验数据,分析结果表明,该方法可有效实现故障诊断。

关键词: 特征提取;模糊信息粒化;模糊熵;故障诊断;机载燃油泵

Abstract: Due to the complexity of mechanical systems, the randomicity of the vibration signal behave on different scales, making it necessary to analyze the vibration signal in a multi-scale way. In order to realize fault feature extraction of airborne fuel pump, Fuzzy entropy is used as the basic feature of vibration signal of airborne fuel pump, a new method of airborne fuel pumps fault diagnosis based on fuzzy information granulation (FIG) and fuzzy entropy (FE) is put forward. Firstly, the FIG method is used to granulate the vibration signal and the FIG particles made up of minimum, median, and maximum are obtained. Secondly, the FE of the particles is calculated. The entropies are accordingly seen as the characteristic vector, then input to the particle swarm optimization (PSO) - support vector machine (SVM) based classifier. Finally, the proposed method is applied to the airborne fuel pumps and bearing experimental data. The analysis results show that the proposed approach can effectively achieve fault diagnosis.

Key words: Feature extraction;Fuzzy information granulation;Fuzzy entropy;Fault diagnosis;Airborne fuel pump