Artificial Intelligence-based Recognition and Classification of PRPD Patterns in Power Cable

This paper explores the application of AI (Artificial Intelligence) to improve diagnostic accuracy in identifying and classifying PRPD (Phase-Resolved Partial Discharge) patterns in power cables, which is crucial for the maintenance and reliability of electrical infrastructure. Partial discharges, indicating defects like corona, surface discharges, and voids, pose significant challenges in power cable management. By simulating these discharges and capturing the resulting signals, this study transforms them into PRPD images, creating a comprehensive dataset for AI analysis. The focus of this research is on utilizing advanced machine learning techniques to process and analyze these images to classify the unique visual characteristics associated with each defect type. A notable achievement of this research is the development and training of a robust AI model that achieved an accuracy rate exceeding 93.8%, demonstrating the potential of AI to revolutionize power cable diagnostics. This approach offers a new path for predictive maintenance, allowing power utilities to prevent potential failures and enhance operational efficiency. Our methodology involves an in-depth analysis of PRPD images using CNN (Convolutional Neural Network) to extract and learn feature representations specific to each type of discharge. The results not only demonstrate the effectiveness of AI in diagnosing electrical equipment but also pave the way for future innovations in the field, potentially setting new industry standards for maintenance strategies and reliability assessments.