基于小波变换的电子设备故障诊断技术研究  

Research on Electronic Equipment Fault Diagnosis Technology Based on Wavelet Transform

作　　者：赵玉刚[1] 鞠建波[2] 杨兵兵[2] 

机构地区：[1]第二炮兵工程大学士官学院,山东青州262500 [2]海军航空工程学院,山东烟台264001

出　　处：《兵器装备工程学报》2016年第5期49-52,共4页Journal of Ordnance Equipment Engineering

基　　金：国家自然科学基金资助项目(60874112);军队科研专项资助项目(41512321);军队科研专项资助项目(415C173)

摘　　要：基于BP神经网络的电子设备故障诊断,以其不需要建立故障诊断模型,易于智能化实现等优势,发展较为迅速,而采用此方法的故障诊断大部分以故障树获得的特征参数为依据;由于电子设备往往由大量的工作模块组成,各组成模块之间相互耦合、联系紧密,仅仅依靠故障树获得的特征参数难以较好地反映设备状态;所以基于这些特征参数的BP神经网络故障诊断往往诊断率较低;将小波分析的方法运用于特征参数的提取中,利用BP神经网络模式识别完成电台故障诊断;最后通过获得的实测数据进行了实例分析,验证了该方法对于提高电子设备故障诊断率,是可行有效的。The electronic equipment fault diagnosis based on BP neural network, with dvantages that it does not need to establish a fault diagnosis model and is easier to implement intelligent, develops very quickly. However, electronic devices often consists of a lot of work module, the mutual coupling between the each composition module, closely linked, just rely on the fault tree to obtain the characteristic parameters and are hard to better reflect the state of equipment. So, based on the characteristic parameters of the BP neural network, the diagnosis rate of fault diagnosis is often low. The wavelet analysis method was applied to the extraction of characteristic parameters of complete station fault diagnosis using BP neural network pattern recognition. Finally through the test data for the instance analysis, we verified that the method can improve the diagnostic rate of electronic equipment and is feasible and effective.

关 键 词：小波变换 BP神经网络 故障诊断 

分 类 号：TJ01[兵器科学与技术—兵器发射理论与技术]