A Novel Partial Discharge Detection Algorithm in Power Transmission Lines Based on Deep Learning

Abstract

The insulation condition of the power transmission line directly affects whether the power system can operate safely. Partial discharge (PD) is one of the main reasons for the deterioration of power line insulation. Therefore, the partial discharge detection of power transmission lines is of great significance. However, partial discharge detection is not easy to achieve for many other sources of noise could be falsely attributed to PD. Here we propose a novel altorithm based on deep learning model. We make an attempt to combine LSTM with textCNN to improve the performance of partial discharge detection. The algorithm mainly includes signal denoising processing, feature extraction, and then inferring whether there is a partial discharge phenomenon in the circuit through a deep learning model. Experimental results show that the proposed method can effectively detect partial discharges and has excellent performance.