RADIATOR MAINTENANCE USING MACHINE LEARNING FOR AUTOMATED ELECTRICAL EQUIPMENT FAILURE DETECTION

Predictive maintenance is a proactive strategy designed to anticipate equipment failures before they happen, ensuring the reliable operation of electrical equipment. Existing manual methods for radiator maintenance rely heavily on periodic inspections and reactive repairs, leading to unexpected downtimes, high maintenance costs, and inefficiencies in detecting early failure signs. These traditional approaches often fail to leverage operational data for proactive decision-making, resulting in delayed responses to equipment degradation. This research presents a machine learning-driven application for predictive maintenance of electrical equipment, specifically targeting radiator failure detection, using a Tkinter-based GUI for user interaction. The proposed method addresses these challenges by implementing a systematic pipeline: users upload a CSV dataset, which is preprocessed by handling missing values and applying label encoding to categorical features, followed by an 80-20 train-test split for model training and evaluation. Two classifiers—Passive Aggressive Classifier and DTC—are trained, with the DTC demonstrating superior performance in identifying radiator status. The application enables predictions on new datasets, displaying outcomes in the GUI, and provides performance comparisons through visualizations, empowering users with an automated, data-driven tool for timely maintenance decisions in an Industry 4.0.