基于等值网络模型的电子设备故障智能分析诊断技术

Intelligent analysis and diagnostic technology for electronicdevice faults based on equivalence network model

随着电力电子设备电路的日益复杂,故障可能性也在增加。为了及时捕捉电路故障,提出了一种检测电子电路问题的系统。通过分析从故障电路中获取的特征信号,该系统能够检测和定位电路故障。故障检测系统的信号采集模块具备8个通道,用于数据采集。系统还借助闭环控制的DC-DC变换电路建立仿真模型,使用峰值电压信号作为故障特征参数。通过BP神经网络模型,提取关键信息并生成新特征向量,实现了数据维度的减少。实验结果显示,本研究的故障诊断技术在硬故障诊断方面达到100%的成功率,在软故障诊断方面,诊断率高达100%,最低为93%。

With the increasing complexity of power electronic equipment circuits,the possibility of failure is also increasing.In order to catch circuit faults in time,a system for detecting electronic circuit problems was proposed.By analyzing the characteristic signal obtained from the faulty circuit,the system could detect and locate the circuit fault.The signal acquisition module of the fault detection system had 8 channels for data acquisition.The simulation model was also established with the closed-loop control DC-DC conversion circuit,and the peak voltage signal was used as the fault characteristic parameter.Through the BP neural network model,the key information was extracted and new feature vector is generated,which realized the reduction of data dimension.The experimental results showed that the fault diagnosis technology in this study had 100%success rate in hard fault diagnosis,and the diagnosis rate was as high as 100%in soft fault diagnosis,and the lowest was 93%.