Application of Artificial Intelligence Tools for Partial Discharge Classification

Partial Discharges (PD) are a critical phenomenon in high-voltage insulation systems, and their accurate classification is essential for effective diagnostics and maintenance. In this study, various electrode configurations were designed and experimentally tested to generate and measure PD signals. The collected data was then used to evaluate different machine learning models for classifying electrode configurations based on signal characteristics. The classification process involved analyzing PD signals using MATLAB tools, specifically utilizing the Classificator Learner for model training and validation. Various AI-based classification models, including decision trees, support vector machines, neural networks, and ensemble methods, were tested to determine their effectiveness in distinguishing between different PD patterns. The results highlight the advantages and limitations of each approach, providing insights into the most suitable AI models for PD classification and improving diagnostic accuracy in insulation systems.