Methodology for Training a Neuro-Fuzzy Control System for a Diesel-Generator Unit Under Variable Operating Conditions

Andrii Zalata

Authors

Andrii Zalata Ukrainian State University of Railway Transport

Keywords:

diesel-generator unit, intelligent control, neuro-fuzzy systems, machine learning, autonomous rolling stock

Abstract

This paper presents a methodology for constructing and training a neuro-fuzzy control system for a diesel-generator unit operating under variable railway conditions. Modern traction power units encounter significant fluctuations in operational factors such as train mass, track profile, and section length, which necessitate adaptive regulation of power output. Traditional control systems are limited in their ability to respond to complex multifactor dynamics, motivating the use of hybrid intelligent systems. The proposed approach integrates Fuzzy C-Means (FCM) clustering to determine the initial structure of the fuzzy rule base and to form Gaussian membership functions based on cluster centers. A hybrid learning strategy is implemented, combining backpropagation and stochastic gradient descent to adjust both the fuzzy and neural components of the model. This enables the system to refine membership parameters, optimize rule interactions, and adapt to nonlinearities in the operational data. The developed neuro-fuzzy model is validated using test samples not included in the training dataset. The results demonstrate high approximation accuracy and strong generalization capability, with prediction errors remaining within acceptable limits. The model effectively reproduces optimal control actions across diverse operating scenarios. The proposed methodology is suitable for integration into traction energy control systems and provides a foundation for future enhancements through expanded datasets, improved optimization algorithms, and full-scale simulation or field testing.

References

Al-Hadithi, B. M., & Gómez, J. (2025). Fuzzy control and modeling techniques based on multidimensional membership functions defined by fuzzy clustering algorithms. Applied Sciences, 15(8), 4479. https://doi.org/10.3390/app15084479.

Chen, T., & Chen, J. C.-Y. (2020). Decentralized fuzzy C-means robust algorithm for continuous systems. Aircraft Engineering and Aerospace Technology, 92(2), 222–228.

Cheng, Y., Zhang, J., Al Shurafa, M., Liu, D., Zhao, Y., Ding, C., & Niu, J. (2025). An improved multiple adaptive neuro fuzzy inference system based on genetic algorithm for energy management system of island microgrid. Scientific Reports, 15, 17988. https://doi.org/10.1038/s41598-025-98665-x.

Elborlsy, M. S., Hamad, S. A., El-Sousy, F. F. M., Mostafa, R. M., Keshta, H. E., & Ghalib, M. A. (2025). Neuro-fuzzy controller based adaptive control for enhancing the frequency response of two-area power system. Heliyon, 11(4), e42547. https://doi.org/10.1016/j.heliyon.2025.e42547.

Ferdaus, M. M., Anavatti, S. G., Garratt, M. A., & Pratama, M. (2019). Development of C-means clustering based adaptive fuzzy controller for a flapping wing micro air vehicle. Journal of Artificial Intelligence and Soft Computing Research, 9(2), 99–109. https://doi.org/10.2478/jaiscr-2018-0027.

Fekry, H. M., ElDesouky, A. A., Kassem, A. M., & Abdelaziz, A. Y. (2020). Power management strategy based on adaptive neuro fuzzy inference system for AC microgrid. IEEE Access, 8, 192087–192100. https://doi.org/10.1109/ACCESS.2020.3032705.

Fu, J., Li, J., Li, Y., Dai, X., Sun, J., & Xiao, J. (2025). Research on optimization of diesel engine speed control based on UKF-filtered data and PSO fuzzy PID control. Processes, 13(3), 777. https://doi.org/10.3390/pr13030777.

Gorobchenko, O., Holub, H., & Zaika, D. (2024). Theoretical basics of the self-learning system of intelligent locomotive decision support systems. Archives of Transport, 71(3), 169–186. https://doi.org/10.61089/aot2024.gaevsp41.

Gorobchenko, O., Zaika, D., Maliuk, S., Arkhypov, O., & Nevedrov, O. (2025). Research of theoretical basis of implementation of intelligent control systems for locomotive traction transmission. Transport Systems and Technologies, (45), 145-160. https://doi.org/10.32703/2617-9040-2025-45-11.

Hussein, H., Aloui, A., & AlShammari, B. (2018). ANFIS-based PI controller for maximum power point tracking in PV systems. International Journal of Advanced and Applied Sciences, 5(2), 90–96.

Kumar, K., Das, M., & Karn, A. K. (2024). ANFIS robust control application and analysis for load frequency control with nonlinearity. Journal of Electrical Systems and Information Technology, 11, 65. https://doi.org/10.1186/s43067-024-00175-9.

Mantalas, E.-M., Sagias, V. D., Zacharia, P., & Stergiou, C. I. (2025). Neuro-fuzzy model evaluation for enhanced prediction of mechanical properties in AM specimens. Applied Sciences, 15(1), 7. https://doi.org/10.3390/app15010007.

Nasim, F., Khatoon, S., Ibraheem, Urooj, S., Shahid, M., Ali, A., & Nasser, N. (2025). Hybrid ANFIS-PI-based optimization for improved power conversion in DFIG wind turbine. Sustainability, 17(6), 2454. https://doi.org/10.3390/su17062454.

Oladipo, S., Sun, Y., & Adegoke, S. A. (2025). Application of soft computing techniques to load frequency control system of electric power systems: A brief survey. Cogent Engineering, 12(1), 2572297. https://doi.org/10.1080/23311916.2025.2572297.

Panoiu, M., Panoiu, C., & Mezinescu, S. (2023). Modelling and prediction of reactive power at railway stations using adaptive neuro fuzzy inference systems. Applied Sciences, 13(1), 212. https://doi.org/10.3390/app13010212.

Rahmani, M., Arabi Nowdeh, S., & colleagues. (2018). Frequency control of islanded microgrids based on fuzzy controller equipped with STATCOM. Journal of Renewable Energy and Environment.

Selvaraju, R. K., & Somaskandan, G. (2017). ACS algorithm tuned ANFIS-based controller for LFC in deregulated environment. Journal of Applied Research and Technology, 15(2), 152–166. https://doi.org/10.1016/j.jart.2017.01.010.

Sharma, K., Menaka, C., Saraswat, N., & Kulhar, K. S. (2024). Neuro-fuzzy controllers for power quality improvement of grid connected PV-battery-diesel based hybrid supply system. E3S Web of Conferences, 540, 10002. https://doi.org/10.1051/e3sconf/202454010002.

Shynkarenko, V. I., Sablin, O. I., & Ivanov, O. P. (2016). Constructive modelling for zone of recovery energy distribution of DC traction. Science and Transport Progress, 5(65), 125–135. https://doi.org/10.15802/stp2016/84036.

Wang, D., & Dang, G. (2024). Fuzzy recurrent stochastic configuration networks for industrial data analytics. IEEE Transactions on Fuzzy Systems. Advance online publication. https://doi.org/10.1109/TFUZZ.2024.3511695.