2025-04-10https://repository.covenantuniversity.edu.ng/handle/123456789/47583Unplanned downtime in industries poses significant challenges, affecting production efficiency and profitability. To address this issue, companies strive to optimize operations and minimize disruptions that hinder meeting customer demands and financial targets. Predictive maintenance, utilizing advanced technologies such as data analytics, machine learning, and IoT devices, enables real-time monitoring and analysis of equipment data. This study focuses on training an adaptable machine-learning model for predicting faults in induction motors in industrial settings. By implementing such a model, proactive maintenance can be facilitated, leading to reduced downtime in industrial operations. A dataset containing healthy and faulty conditions of four 3 phase induction motors, along with relevant features for fault prediction, was obtained. Multiple machine learning algorithms were trained using this dataset, and they demonstrated promising performance. The RF model achieved the highest accuracy of 0.91, followed by the Ann and k-NN models with an accuracy of 0.9. The DT model achieved the lowest accuracy of 0.89. Further evaluation of the models was conducted using a confusion matrix, which provided a detailed breakdown of the model's performance for each class, indicating the number of correctly and incorrectly classified induction motor conditions. The outcome of the confusion matrix demonstrated that the models successfully classified the different states or conditions of the induction motors. To enhance the performance of the models, future work should involve refining the ANN and RF models, exploring transfer learning or ensemble methods, and incorporating diverse datasets to improve generalization.application/pdfT Technology (General), TK Electrical engineering. Electronics Nuclear engineeringThesis