Partial Discharge Detection Method for High Voltage Converter System Using Wavelet Transform and Convolutional Neural Network

Since partial discharges that occur in insulation in converter equipment can cause equipment failure, technology to detect them is important. However, converter devices generate high switching noise, which makes it difficult to detect partial discharges that have small signal strength. In this paper, we propose a new method using wavelet transform and a Convolutional Neural Network (CNN) to detect the partial discharges in converter devices. By using wavelet transform and learning the partial discharge signals and noise signals with a CNN, we have succeeded in constructing a system that can detect partial discharge with a high accuracy of more than 90%.