基于小波变换的电子档案查阅系统电路故障自动诊断研究

Research on circuit fault automatic diagnosis of electronic archives retrieval system based on wavelet transform

针对传统电路故障的诊断准确率低下的问题,以电子档案查阅终端中的充电模块中的三项桥式全控整流电路为研究对象,采用小波包改进的小波变换对电路波形特征进行提取,然后采用基于高斯变异和轮盘赌算法改进的人工免疫算法,并将改进后的算法应用到电路诊断中,以此提高电路故障诊断的准确率。结果表明,小波包变换可提升故障波形的特征提取的效果。在此基础上,改进后的免疫克隆变异算法的故障识别准确率趋近于99.8%,且诊断时间为1.62 s,相比较传统的电路诊断算法,诊断准确率提升25%,诊断时间缩短了0.27 s。由此说明,基于小波包变换和改进后的人工免疫算法能显著提升电子档案查阅系统故障诊断的准确率和效率。

Aiming at the problem of low accuracy of traditional circuit fault diagnosis, taking the three bridge fully controlled rectifier circuit in the charging module in the electronic archives access terminal as the research object, the circuit waveform characteristics are extracted by using the wavelet transform improved by wavelet packet, and then the artificial immune algorithm based on Gauss mutation and roulette wheel algorithm is adopted, and the improved algorithm is applied to circuit diagnosis, In order to improve the accuracy of circuit fault diagnosis. The results show that wavelet packet transform can improve the effect of feature extraction of fault waveform. On this basis, the fault recognition accuracy of the improved immune clonal mutation algorithm approaches 99.8%, and the diagnosis time is 1.62 s. Compared with the traditional circuit diagnosis algorithm, the diagnosis accuracy is increased by 25%, and the diagnosis time is shortened by 0.27 s. This shows that the fault diagnosis accuracy and efficiency of electronic archives retrieval system can be significantly improved based on wavelet packet transform and improved artificial immune algorithm.