Fault diagnosis of rectifier in aircraft power system
-
摘要: 某型号飞机电源系统现有的机上自检测装置由传统的硬件逻辑电路构成,存在功能扩展性差、可靠性低等缺点.为符合新一代机上自检测装置微机化、智能化的特点,在对其进行故障模式分析的基础上,采用基于小波神经网络的故障诊断方法,通过对整流装置输出电压的实测信号样本的频谱分析,获得对故障敏感的特征频率点,根据小波变换的多分辨率分析理论,确定了与特征频率点相对应的小波母函数和变换尺度.在此基础上,通过定义频带能量特征向量,将小波变换得到的小波系数转换为一组特征向量.将特征向量作为BP神经网络的前端输入,由神经网络完成故障的识别与分类.经验证,故障特征得到了有效地提取,使神经网络可在各种工况下对故障进行故障诊断,符合灵敏性、鲁棒性的要求.Abstract: The built-in test equipment in a type of aircraft power system is composed with traditional logic hardware, so the abilities on expansion and reliability are limited hardly. According to the characters of built-in test equipment in the next generation, such as computerization and intelligence, based on the fault pattern analysis of the rectifier in aircraft power system, the key points on frequency to each fault pattern were gained by the frequency analysis on output voltage. By using the multi-resolution analysis in the wavelet theory, the basic wavelet function and scale corresponding to key points was confirmed. And then, the wavelet coefficient was converted to a character vector that is the input of the BP neural network which fulfill the diagnosis by defining frequency energy character vector. The results show that this method can distinguish each fault efficiently.
-
Key words:
- wavelet /
- neural network /
- rectifier /
- fault diagnosis
-
[1] 温熙森. 智能机内测试理论与应用[M]. 北京: 国防工业出版社, 2002: 67-103 Wen Xisen. Theory and application of intelligent built-in test[M]. Beijing: National Defense Industry Press, 67-103(in Chinese) [2] 徐宗祥, 蒋新华, 刘建成. 小波变换在电力机车主整流器故障检测中的应用研究[J]. 电力机车与城轨车辆, 2003, 26(2): 5-7 Xu Zongxiang, Jiang Xinhua, Liu Jiancheng. Application of wavelet transform on fault detection of electric locomotives main converter[J]. Electric Locomotives & Mass Transit Vehicles, 2003, 26(2): 5-7(in Chinese) [3] 张征平,任震. 基于小波脊线的电动机转子故障检测新方法[J].中国电机工程学报, 2003, 23(1): 97-101 Zhang Zhengping, Ren Zhen. A novel method of motor faulted rotor detection based on wavelet ridge [J]. Proceedings of the SCEE, 2003,23(1): 97-101(in Chinese) [4] Aminian M. Neural-network based analog-circuit fault diagnosis using wavelet transform as preprocessor [J]. Analog and Digital Signal Processing, 2000, 47(2): 47-51 [5] Milimongared J, Kelk H M. A novel approach for broken-rotor-bar detection in cage induction motors [J]. IEEE Trans on Industry Applications, 1999, 35(5):1000-10006 [6] Elkasabgy N M, Eastham A R. Detection of broken bars in the cage rotor on an induction machine [J]. IEEE Trans on Industry Application , 1992, 22(1): 165-171
点击查看大图
计量
- 文章访问数: 3693
- HTML全文浏览量: 86
- PDF下载量: 1042
- 被引次数: 0