Partial discharge (PD) is the initial stage of a complete failure in some power systems’ components. If left without repair, PD in components like electrical machines, cables, and covered conductors can eventually lead to substantial power outages and damages. Methods for PD detection rely on statistical feature extraction and conventional machine learning methods. However, the performance of these methods decreases in the presence of noise.
In a study presented at the 2022 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE), engineers from EPE investigated a solution for PD fault detection in Medium Voltage Covered Conductor Overhead lines (MVCCO). The study used a deep learning method based on Long-Term Short Memory (LSTM) and Attention layers. A k-fold stratified cross-validation method was used for training and validation. The impact of some hyperparameters on the deep learning model and the classification result was also investigated.
The method proposed in the study was applied to a large open-source dataset of signals with PD fault provided by VSB’s ENET center. The results were then compared with some traditional machine learning methods, which proved the superiority of the proposed method over the conventional techniques in terms of detecting a faulty signal.
Visit IEEE’s website to read more about the study.
