Predictive Modeling of Engine Health using Machine Learning Techniques

• Harini S V,
• Harish J G,
• Thanvitha Nandakumar,
• Prakash P,
• Sakthivel V,

Vol. 19, No. 10, October 31, 2025
10.3837/tiis.2025.10.005, Download Paper (Free):

• Data Resampling
• Engine Health Prediction
• Ensemble Methods
• Explainable AI (XAI)
• Machine Learning
• predictive maintenance

Abstract

Predictive maintenance is critical for keeping mechanical systems reliable and efficient. This research creates a machine learning system to categorize the condition of engines as either healthy or defective, overcoming the drawbacks of conventional diagnostic methods. Six distinct machine learning methods, including Random Forest Classifier, Gradient Boosting Classifier, XGBoost Classifier, Logistic Regression, Voting Ensemble, and Stacked Ensemble models, were assessed. The procedure involved scaling and resampling the dataset, then training and optimizing the hyperparameters of the classifier. The assessment standards included accuracy, ROC-AUC, the confusion matrix, and the classification report, containing precision, recall, and F1-score. The Stacked Ensemble model outperformed the rest with an accuracy of 88.66%, an ROC-AUC of 0.941, and an F1-score of 0.86. Various techniques for resampling data greatly impacted the results, with ENN and SMOTEENN standing out as the most effective strategies for improving predictions of minority groups. The findings suggest that combining ensemble methods with advanced data-balancing techniques is highly effective for predictive maintenance assignments. This study demonstrates the potential for machine learning to improve diagnostic accuracy and prevent unexpected errors. Future work will focus on integrating IoT technology for real-time data collection with edge computing for immediate processing. The cloud will serve as the platform for model deployment, generating predictions and sending alert notifications to drivers. Additionally, Explainable AI (XAI) methods will be implemented to enhance transparency in decision-making, ensuring better-informed drivers and more efficient vehicle maintenance. This combination of IoT, edge computing, cloud infrastructure, and XAI will improve the system's effectiveness, reliability, and overall user experience.

Statistics

Cite this article