航空科学技术2025，Vol.36Issue(5)：81-88,8.DOI:10.19452/j.issn1007-5453.2025.05.010
基于朴素贝叶斯网络的航空发电机轴承故障诊断方法
Naive Bayesian Network Based Fault Diagnosis Method for Aero-generator Bearings
摘要
Abstract
As a kind of aviation generator,the reliability of permanent magnet synchronous motor(PMSM)is an important prerequisite to guarantee the performance of aircraft and flight safety.Aiming at the fault diagnosis problem of the rolling bearing of permanent magnet synchronous motor in the harsh environment of aviation,this paper proposes a fault diagnosis method based on envelope spectral analysis and Bayesian network.Firstly,the bearing vibration mechanism,fault characteristic frequency and the characteristics of the vibration signal are studied,and the rolling bearing fault time domain signal is subjected to Hilbert transform to extract the fault characteristics and obtain the envelope spectrum,and then compared with the theoretically calculated frequency to realize the diagnosis of the fault category;in order to further accurately locate the dimensions of the rolling bearing faults,the rolling bearing fault dataset of a university is used to train and test the extracted feature value set through the plain Bayesian network.The set of eigenvalues is trained and tested to realize fault diagnosis and identification;finally,compared with neural network and support vector machine diagnosis methods,the method improves the accuracy of fault diagnosis by 6.93%,which confirms the validity of the method,and provides theoretical and technological references to enhance the operational reliability of aviation generators.关键词
永磁同步电机/滚动轴承/故障诊断/贝叶斯网络/包络谱/希尔伯特变换Key words
PMSM/rolling bearing/fault diagnosis/Bayesian network/envelope spectrum/Hilbert transform引用本文复制引用
葛乐飞,宋佳赫,李梓童,刘英..基于朴素贝叶斯网络的航空发电机轴承故障诊断方法[J].航空科学技术,2025,36(5):81-88,8.基金项目
航空科学基金(20220040051002) Aeronautical Science Foundation of China(20220040051002) （20220040051002）
