PREDICTIVE MAINTENANCE SYSTEM FOR DIESEL GENERATORS USING MACHINE LEARNING

This project focuses on the design and design and development of a predictive maintenance system for diesel generators, aimed at forecasting potential failures before they occur. Diesel generators play a crucial role in providing backup power in critical role in providing backup power in critical sectors such as the University. However, existing maintenance strategies are inadequate: corrective maintenance leads to costly unplanned downtime, while preventive maintenance often results in unnecessary repairs and higher operational costs. Literature highlights the limitations of traditional maintenance approaches and underscores the advantages of predictive maintenance and described predictive maintenance as a key enabler of industry 4.0, this demonstrates its effectiveness in enhancing reliability and reducing costs through machine learning. These studies confirm the need for more intelligent and data-driven solutions. The methodology adopted in this project is the CRISP-DM, and it involved data collection and pre-processing of generator performance parameters like fuel consumption, work load and running hours. Key features were selected, and multiple machine learning algorithms including Random Forest, LSTM, Logistic Regression, Gradient Boosting networks were trained and evaluated. The best performing model was implemented into a user-friendly interface that accepts generator input data and provide real-time prediction and decision support. Result shows that the proposed system achieved a high prediction accuracy of 92%, enabling timely detection of potential faults and reducing unplanned downtime. By optimizing maintenance schedules, the system minimizes costs while extending generator lifespan.In conclusion, this study contributes to maintenance engineering by developing a tailored, datadriven predictive maintenance framework for diesel generators. It demonstrates that predictive analytics can outperform traditional strategies, offering a practical solution for improved reliability, cost efficiency and operational continuity in critical sectors x