Е-Публикации
Технически университет - София

Abstract: Reliability of renewable energy assets represents an essential factor for achieving operational stability while lowering maintenance expenses.The traditional rule-based and physics-based reliability models fail to produce accurate results when used in complicated operational settings. The paper shows a predictive maintenance system which depends on the Predictive Maintenance dataset and EDA and ML techniques and methods to solve class imbalance problems. The predictive maintenance framework tests baseline models Logistic Regression and KNN and Random Forest against advanced gradient boosting methods XGBoost and LightGBM, which were optimised through Bayesian hyperparameter tuning. The two solutions for handling imbalance used resampling methods, which included upsampling and SMOTE and class weight adjustments. The LightGBM models which underwent hyperparameter tuning with SMOTE produced the best predictive results because they maintained both precision and recall balance and achieved high ROC AUC scores. The predictive analysis of feature importance indicated that temperature changes together with torque data and tool wear stood out as essential indicators. The data-based methods showed they could enhance energy asset reliability, which would bring about a change in how condition monitoring and maintenance planning work today.

References

1. Hastie T. Tibshirani R. Friedman J. The Elements of Statistical Learning: Data Mining, Inference, and Prediction Springer New York 2009. https://doi.org/10.1007/978-0-387-84858-7
2. Predictive Maintenance Dataset 2018
3. Cover T M. Hart P. E. Nearest Neighbor Pattern Classification IEEE Transactions on Information Theory 13 1 1967 21 27. https://doi.org/10.1109/TIT.1967.1053964
4. Friedman J. H. Greedy Function Approximation: A Gradient Boosting Machine Annals of Statistics 29 5 2001 1189 1232. https://doi.org/10.1214/aos/1013203451
5. Fisher A. Rudin C. Dominici F. All Models Are Wrong, but Many Are Useful: Learning a Variable's Importance by Studying an Entire Class of Prediction Models Simultaneously Journal of Machine Learning Research 20 2019 1 81. https://doi.org/10.48550/arXiv.1801.01489
6. Hosmer D. W. Jr. Lemeshow S. Sturdivant R. X. Applied Logistic Regression: Third Edition Wiley 2013. https://doi.org/10.1002/9781118548387
7. Breiman L. Random Forests Machine Learning 45 1 2001 5 32. https://doi.org/10.1023/A:1010933404324
8. Chen T. Guestrin C. XGBoost: A Scalable Tree Boosting System 2016. https://doi.org/10.1145/2939672.2939785
9. Akiba T. Sano S. Yanase T. Ohta T. Koyama M. Optuna: A Next-Generation Hyperparameter Optimization Framework 2019. https://doi.org/10.1145/3292500.3330701
10. Clarke J. A. Energy Simulation in Building Design Butterworth-Heinemann / Routledge London 2001
11. Sheiretsky K. Tzenova Z. Small Parameter Method Application to the Energy Integral of an Autonomous Oscillating System 2939 Melville NY 2023
12. Zhou J. Tong X. Zhou J. Liu R. Bai S. Feature Selection of Power System Data for Frequency Prediction with BO-LightGBM-Boruta Energy and AI 21 2025 100581. https://doi.org/10.1016/j.egyai.2025.100581
13. King G. Zeng L. Logistic Regression in Rare Events Data Political Analysis 9 2 2001 137 163. https://doi.org/10.1093/oxfordjournals.pan.a004868
14. Xuexuan L. Dejian Y. Minhui Q. Fenghe J. A Hierarchical Optimize Scheduling Strategy for Orderly Charging of Electric Vehicles Based on PTM and CNN-Transformer-LightGBM Model Sustainable Energy, Grids and Networks 44 2025 101944. https://doi.org/10.1016/j.segan.2025.101944
15. Strobl C. Boulesteix A.-L. Zeileis A. Hothorn T. Bias in Random Forest Variable Importance Measures: Illustrations, Sources and a Solution BMC Bioinformatics 8 2007 25. https://doi.org/10.1186/1471-2105-8-25
16. Jardine A. K. S. Lin D. Banjevic D. A Review on Machinery Diagnostics and Prognostics Implementing Condition-Based Maintenance Mechanical Systems and Signal Processing 20 7 2006 1483 1510. https://doi.org/10.1016/j.ymssp.2005.09.012
17. Sheiretsky K. Antonov S. Method of Exponentially Varying Amplitudes for Finding Periodic Solutions of Autonomous Equations 2023. https://doi.org/10.1109/ICEST58410.2023.10187238
18. Sheiretsky K. Tzenova Z. Finding an Approximate Solution for a Non-Autonomous Dynamical System by Direct Using an Asymptotic Method to the Energy Expression 2939 Melville NY 2023
19. Tsanas A. Xifara A. Accurate Quantitative Estimation of Energy Performance of Residential Buildings Using Statistical Machine Learning Tools Energy and Buildings 49 2012 560 567. https://doi.org/10.1016/j.enbuild.2012.03.003

Issue

Full text of the publication