Incipient fault diagnosis of analog circuits based on wavelet transform and improved deep convolutional neural network

To enhance the reliability of analog circuits in electrical systems, this letter proposes a novel incipient fault diagnosis method by integrating wavelet transform(WT) and improved convolutional neural network. Different from traditional methods, where feature extraction classification are separately designed performed, aims to automatically learn features classify type faults simultaneously. An network named multi-channel compactness (MC-CNN) is proposed, which can obtain complementary rich information multi-scale components extracted transform. Moreover, we adopt center loss as an auxiliary function maximize interclass separability intraclass samples. The proposed fully evaluated with Sallen-Key bandpass filter circuit four-opamp biquad high-pass circuit. experimental results demonstrate that very effective for diagnosis, has higher accuracy than other typical methods.