基于模糊神经网络的电路板故障诊断方法研究  

Research on Fault Diagnosis Method of Circuit Board Based on Fuzzy Neural Network

作　　者：王昌福 WANG Changfu(Yantai Automoblle Engineering Professional College,Yantai 265500,China)

机构地区：[1]烟台汽车工程职业学院,山东烟台265500

出　　处：《通信电源技术》2020年第19期113-114,117,共3页Telecom Power Technology

摘　　要：考虑到传统电路板故障诊断方法在实际应用过程中不能精准定位电路板故障的具体位置,导致故障诊断置信度低的问题,提出了基于模糊神经网络的电路板故障诊断方法。通过构建电路板故障诊断模型,集中化处理电路板故障特征参数,诊断电路板运行过程中的电幅频率,训练采集到的电路板故障信息,提取离散型故障信号的有效值,结合电路板中多种故障模式,进行故障模式的划分。最后利用模糊神经网络技术单独选用中间层的传递函数和模糊神经元数目作为模糊神经元的控制核心,实现基于模糊神经网络的电路板故障诊断定位。设计实例分析表明,设计的故障诊断方法的故障诊断置信度明显高于实验对照组,其故障诊断精度更高。Considering that the traditional fault diagnosis method cannot accurately locate the fault location of the circuit board in the practical application,which leads to the problem of low confidence in fault diagnosis,a fault diagnosis method based on fuzzy neural network is proposed.By constructing the fault diagnosis model of the circuit board,centralizing processing the fault characteristic parameters of the circuit board,diagnosing the circuit board frequency in the process of running,training the fault information of the circuit board collected,extracting the effective value of discrete fault signal,and dividing the fault mode by combining various fault modes in the circuit board.Finally,the transfer function of the middle layer and the number of fuzzy neurons are selected as the control core of fuzzy neurons by using fuzzy neural network technology to realize fault diagnosis and location of circuit board based on fuzzy neural network.The analysis of the design example shows that the reliability of the designed fault diagnosis method is obviously higher than that of the experimental control group,and its fault diagnosis accuracy is higher.

关 键 词：模糊神经网络 电路板 故障诊断 

分 类 号：TN4[电子电信—微电子学与固体电子学]