Construction of a deep network-based insulation status assessment system for oil-paper capacitor-type bushings

The insulation performance of oil-paper capacitor-type bushings directly affects the safe operation of power equipment. With the increasing complexity of the operating environment of power equipment, traditional evaluation methods for oil-paper capacitor-type bushings are no longer sufficient to meet the requirements of real-time performance and accuracy. Therefore, this study proposes a deep network-based insulation state evaluation system for oil-paper capacitor-type bushings. Experimental results show that the Recurrent Neural Network-Convolutional Neural Network (RNN-CNN) model achieves an accuracy of 96.5%, precision of 95.8%, recall of 97.1%, and an F1 score of 96.4%, significantly outperforming traditional CNN, Long Short-Term Memory (LSTM), and CNN-LSTM models. Regarding training and inference time, the RNN-CNN model has an inference time of 8.5 milliseconds, demonstrating excellent performance in bushing insulation performance evaluation. The deep learning-based RNN-CNN model can effectively evaluate the insulation state of oil-paper capacitor-type bushings, exhibiting excellent performance in accuracy, real-time performance, and computational efficiency, and has strong application prospects.