Research on Electronic Circuit Fault Diagnosis Method Based on SWT and DCNN-ELM

The increase in the complexity of modern electronic products has brought significant challenges to the fault diagnosis of electronic circuits, and current fault diagnosis methods have problems such as long fault identification time, inaccurate positioning, and low diagnostic efficiency. In response to these situations. This paper proposes a fault diagnosis method for electronic circuits combining synchronous synchrosqueezing wavelet transform (SWT), deep convolutional neural network (DCNN), and extreme learning machine (ELM). First, the original fault signal is noise-reduced and converted into a higher resolution two-dimensional time-frequency image using SWT. Then, the improved and optimized DCNN model is used to extract the advanced features of the time-frequency image, and the extracted advanced features are further input into the ELM classifier for fault classification. Finally, the fault diagnosis and validation are performed by experiments. The experimental results show that, compared with other methods, the electronic circuit fault diagnosis method based on SWT and DCNN-ELM ensures the diagnosis accuracy while shortening the diagnosis time, significantly improving the efficiency of electronic circuit fault diagnosis.