AI-Driven Optimization and Fault Detection in Electrical Power Systems: A Comparative Study

The increasing complexity and integration of renewable sources in modern electrical power systems necessitate advanced methodologies for condition monitoring and fault detection to ensure reliability and safety. Traditional maintenance strategies are often insufficient for handling the intricate dynamics of these evolving systems. This study presents a comparative analysis of machine learning models for engine fault detection using multisensory data. A comprehensive dataset of 10,000 samples, featuring vibration, temperature, acoustic, and pressure sensor data, was utilized to train and evaluate a Neural Network and a Random Forest classifier. The experimental results reveal that while both models demonstrate competence in multi-class classification, the Random Forest model exhibits superior performance in identifying fault instances, a critical aspect for predictive maintenance. This research highlights the significant potential of AI in enhancing fault detection and provides critical insights into the comparative efficacy of different machine learning algorithms in real-world engineering applications, addressing challenges such as class imbalance and feature overlap.