Machine Learning-Based Predictive Fault Detection and Explainable Decision Support for Smart Energy Infrastructure
International Journal of Scientific Research in Computer Science and Engineering · 30.06.2026 · 14 (3) · 1-16
DOI: 10.26438/ijsrcse.v14i3.867 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Scientific Research in Computer Science and Engineering, online erschienen am 30.06.2026. Verfasst von Olaniyi Habib ALIU und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
In this study, a machine learning-based predictive fault detection system in a PV–motor hybrid smart energy infrastructure was designed. The amounts of sensor data in the dataset were divided into 3000 labelled sensor readings, and nine classification algorithms were evaluated for six operating states, five of which were fault states, and one was the normal operating state. According to the performance assessment, six algorithms were able to get a perfect classification result with 100% accuracy, F1 score, recall and precision. Gradient boosting and decision tree models also gave us very good results with an accuracy of 99.83% and 99.67%, respectively. AdaBoost, on the other hand, had merely 73.83% accuracy because some categories of faults were misclassified to a very great extent; therefore, it is not suitable for safety-critical applications. To improve the model interpretability, feature importance analysis was carried out, and it was found that battery current, motor temperature, and PV voltage are the most important factors for fault discrimination. The Random Forest algorithm was chosen because of its great predictive power and robustness and was deployed in the web app FaultGuard ML Detection v1.0. The platform offers real-time manual fault prediction, batch inference via CSV file upload, class probability distributions and maintenance recommendations. The results illustrate that it is possible to achieve very reliable classification of faults in smart energy systems using ensemble and kernel-based machine learning techniques.
Zitieren
APA
ALIU, O. H., OLASINA, J. R., & AJETUNMOBI, D. T. (2026). Machine Learning-Based Predictive Fault Detection and Explainable Decision Support for Smart Energy Infrastructure. International Journal of Scientific Research in Computer Science and Engineering, 14(3), 1-16. https://doi.org/10.26438/ijsrcse.v14i3.867
Vancouver
ALIU OH, OLASINA JR, AJETUNMOBI DT. Machine Learning-Based Predictive Fault Detection and Explainable Decision Support for Smart Energy Infrastructure. International Journal of Scientific Research in Computer Science and Engineering. 2026;14(3):1-16. doi:10.26438/ijsrcse.v14i3.867
BibTeX
@article{ALIU2026,
author = {ALIU, Olaniyi Habib and OLASINA, Jamiu Rotimi and AJETUNMOBI, David Tunmise},
title = {Machine Learning-Based Predictive Fault Detection and Explainable Decision Support for Smart Energy Infrastructure},
journal = {International Journal of Scientific Research in Computer Science and Engineering},
year = {2026},
volume = {14},
number = {3},
pages = {1--16},
doi = {10.26438/ijsrcse.v14i3.867}
}
EUVIMED