Comparative Evaluation of Machine Learning Models for Fault Detection in Induction Motors

References

[1] Bahgat, B. H., Elhay, E. A., & Elkholy, M. M. (2024). Advanced fault detection technique of three-phase induction motor: comprehensive review. Discover Electronics, 1, Article 9.

[2] Yousuf, M., Alsuwian, T., Amin, A. A., Fareed, S., & Hamza, M. (2024). IoT‑based health monitoring and fault detection of industrial AC induction motor for efficient predictive maintenance. Measurement & Control, 57(8), August 2024. https://doi.org/10.1177/00202940241231473

[3] Samiullah, M., Ali, H., Zahoor, S., & Ali, A. (2024). Fault Diagnosis on Induction Motor using Machine Learning and Signal Processing. arXiv.

[4] Abdulkareem, A., Anyim, T., Popoola, O., Abubakar, J., & Ayoade, A. (2025). Prediction of induction motor faults using machine learning. ResearchGate.

[5] Sehri, M., Ertagrin, M., Yildirim, Ö., Orhan, A., & Dumond, P. (2025). Deep Learning Approach to Bearing and Induction Motor Fault Diagnosis via Data Fusion. arXiv.

[6] Montejano Leija, A. B., Ruiz Beltrán, E., Orozco Mora, J. L., & Valdés Valadez, J. O. (2025). Performance of Machine Learning Algorithms in Fault Diagnosis for Manufacturing Systems: A Comparative Analysis. Processes, 13(6), 1624.

[7] Scornet, E. (2020). Trees, forests, and impurity-based variable importance. arXiv preprint arXiv:2001.04295.

[8] Rahmani, N., Doostmohammadi, M., & Emadi, M. (2023). Exploring the variable importance in random forests under correlations. BMC Medical Research Methodology, 23, 2023.

[9] Halder, R. K., Singh, K. R., & Jain, P. (2024). Enhancing K-nearest neighbor algorithm: A comprehensive review. Journal of Big Data, 11(1), 113.

[10] Pagan, M., Zarlis, M., & Candra, A. (2023). Investigating the impact of data scaling on the k-nearest neighbor algorithm. CSIT, 4(2), 135–142.

[11] Raschka, S. (2023). Large Language Models and Nearest Neighbors. The ML Magazine. https://magazine.sebastianraschka.com

[12] Aumüller, M., Bernhardsson, E., & Faithfull, A. (2018). ANN-Benchmarks: A benchmarking tool for approximate nearest neighbor algorithms. arXiv:1807.05614.

[13] Saradhi, T. V. (2025). A Study on Hyperparameter Tuning in Support Vector Machines and its Impact on Model Accuracy. GJEIIR, 5(1).

[14] Chang, Y. J., Lin, Y. L., & Pai, P. F. (2025). Support Vector Machines with Hyperparameter Optimization Frameworks for Classifying Mobile Phone Prices in Multi-Class. Electronics, 14(11), 2173.

[15] Wang, W., Sun, L., & Yu, H. (2024). Interpreting XGBoost performance in medical data: A comprehensive view. Scientific Reports, 14, Article 60173.

@article{1710649,
      author = {Damfebo Franklin Ayebagbalinyo, Inanumo Emmanual, Idongha Charles Godspower},
      title = {Comparative Evaluation of Machine Learning Models for Fault Detection in Induction Motors},
      journal = {Iconic Research And Engineering Journals},
      year = {2025},
      volume = {9},
      number = {3},
      pages = {764-773},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1710649.pdf},
      abstract = {This study, titled Comparative Evaluation of Machine Learning Models for Fault Detection in Induction Motors presents a comparative evaluation of machine learning (ML) models for fault detection in three-phase induction motors, which are essential components in industrial applications. Despite their robust construction, these motors are prone to faults such as stator winding failures, rotor defects, and overvoltage conditions that can cause unexpected breakdowns and operational losses. Traditional fault detection techniques often lack the sensitivity and consistency required for early fault detection. To address this gap, this research investigates the effectiveness of four supervised ML algorithms?Random Forest (RF), K-Nearest Neighbors (KNN), Gradient Boosting Machine (GBM), and Support Vector Machine (SVM). A 7.5?kW induction motor was modeled using MATLAB/Simulink, simulating six operating conditions to generate relevant datasets. These were preprocessed and evaluated in Pytorch using standard classification metrics: accuracy, precision, recall, and F1-score. Among the models, Random Forest delivered the best performance with an average accuracy of 94.7%, precision of 93.7%, recall of 92.7%, and F1-score of 93.2%. GBM followed closely with an accuracy of 92.0% and F1-score of 90.5%, while SVM achieved moderate results. KNN showed the lowest performance across all metrics. The results confirm Random Forest as the most robust and reliable model for industrial motor fault detection.},
      keywords = {Fault Detection, Induction Motors, Machine Learning, Random Forest, Classification Metrics},
      month = {September},
      doi = {https://doi.org/10.64388/IREV9I3-1710649-1267}
  }