Power Quality Classification with S-Transform and Ensemble Learning for Multi-class Complex Analysis

Power quality (PQ) disturbances are major challenges in modern electrical power systems, requiring an accurate classification method to assure reliability system. To address this issue, this paper presents an approach using ensemble machine learning with bagging techniques for the classification of complex PQ disturbances. The methodology involves the application of the Stockwell Transform (ST) for feature extraction from simulated power quality signals, capturing both time and frequency domain information important for disturbance analysis. Through experimentation involving simulated complex signals, we successfully classified 16 distinct power quality disturbance classes with an overall accuracy of 99%. This approach offers a promising solution for real-world application in power systems for the reliable detection and classification of PQ disturbances.