Application of artificial intelligence for optimal control of reactive power compensation in AC traction power supply systems

Maksym Shpilievyi
Received 15.07.2025, Revised 09.11.2025, Accepted 24.12.2025
Abstract

The study aimed to quantitatively assess the effectiveness of neural network-based control of dynamic reactive power compensation in the traction network of Ukrzaliznytsia. The methodology involved high-frequency monitoring of 118 substations (≈16 trillion measurements), aggregation of 52 million Supervisory Control and Data Acquisition records, Theil-Sen trend estimation, Seasonal-Trend decomposition using Loess, the Brown Forsythe test, and a deep neural network with 64-32-16 layers optimised through a Tree-structured Parzen Estimator, validated in 120 Monte Carlo scenarios in Simscape. The results demonstrated that the average daily Q increased from 309.8 to 367.2 Mvar (18.5%) with a trend of 0.95 Mvar/month (95% CI 0.82-1.08; p < 0.001). Winter peaks reached 520 Mvar, while variance after February 2022 increased 1.5-fold (F = 34.9; p < 0.001). The neural network achieved Mean Squared Error = 184 Mvar² and R² = 0.982; in testing, R² = 0.975, Mean Absolute Error = 11.2 Mvar, and the median error of winter peaks did not exceed 3%. In simulations, the algorithm maintained cos φ between 0.985 and 0.991 for loads ranging from 0.3 to 1.2 p.u., reduced daily losses by 6.1-9.4% (≈1.2 MWh/day), and decreased the number of Static Var Compensator switchings by 38%. Under failure of 20% of Insulated-Gate Bipolar Transistor modules, the duration of cos φ < 0.98 was 0.7%, compared with 19% for the Proportional Integral-Derivative controller. Quantitative comparison of compensation strategies confirmed that the simulated STATCOM-DL system achieved the lowest average cos φ deviations (0.015 vs. 0.030-0.048 for competitors) and minimal annual losses (≈ 430 MWh), outperforming all alternative approaches. The findings confirmed that the combination of deep learning and statistical validation enhances network stability and improves energy efficiency. The practical significance lies in the fact that the resulting algorithms may be employed by railway dispatch centres and energy integrators for real-time optimisation

Keywords:
deep learning; neural network; energy efficiency; phase balance; daily switching; voltage fluctuations

Retrieved from Volume 23, No. 2, 2025

Pages 34–45

References Suggested citation
  1. Badrudeen, T.U., Ariyo, F.K., & Nwulu, N. (2025). Performance evaluation of STATCOM placement in real-world power system through DOA-NVSP algorithm for congestion management and stability improvement. Scientific Reports, 15, article number 26399. doi: 10.1038/s41598-025-11232-2.
  2. Bandura, I., Romaniuk, M., Komenda, N., Hadai, A., & Volynets, V. (2023). Optimisation of energy solutions: Alternative energy, reactive power compensation, and energy efficiency management. Machinery & Energetics, 14(4), 121-130. doi: 10.31548/machinery/4.2023.121.
  3. Barros, L.A.M., Tanta, M., Martins, A.P., Afonso, J.L., & Pinto, J.G. (2021). Evaluation of static synchronous compensator and rail power conditioner in electrified railway systems using V/V and Scott power transformers. EAI Endorsed Transactions on Energy Web, 8(34), article number e9. doi: 10.4108/eai.29-3-2021.169164.
  4. Brown, M.B., & Forsythe, A.B. (1974). Robust tests for the equality of variances. Journal of the American Statistical Association, 69(346), 364-367. doi: 10.1080/01621459.1974.10482955.
  5. Chandrasekaran, S., Arghir, C., Joerg, P., Doerfler, F., & Mastellone, S. (2025). Reactive power flow optimization in AC drive systems. ArXiv. doi: 10.48550/arXiv.2504.10360.
  6. Chen, W., Zhang, X., Liu, H., Song, K., Li, S., & Duan, L. (2022). Reactive power compensation optimization of high voltage transmission network based on artificial bee colony algorithm. In 18th international conference on AC and DC power transmission (pp. 1089-1095). Stevenage: Institution of Engineering and Technology. doi: 10.1049/icp.2022.1353.
  7. Choi, J., Lee, H., Ku, H., & Yoon, M. (2024). Optimization of the operating point of multi-reactive power compensation equipment considering facility characteristics. In TENCON 2024 – 2024 IEEE region 10 conference (pp. 1182-1185). Singapore: Institute of Electrical and Electronics Engineers. doi: 10.1109/ TENCON61640.2024.10902882.
  8. Dan, W., Hanqing, L., Yidong, H., Ke, S., Dong, L., Rui, S., & Xu, G. (2024). Study on optimal configuration of dynamic reactive power compensation for AC and DC power network. In 2024 9th Asia conference on power and electrical engineering (pp. 511-516). Shanghai: IEEE. doi: 10.1109/ACPEE60788.2024.10532291.
  9. Ding, C., Guo, Y., You, H., & Zhang, H. (2023). Optimization of algorithm for solving railroad power conditioner compensation power reference value and system power quality analysis based on optimal compensation model. Energies, 16(20), article number 7073. doi: 10.3390/en16207073.
  10. DSTU EN 50388:2015. (2015). Railway transport – power supply for rolling stock – technical criteria for coordination between power supply (substations) and rolling stock to ensure compatibility. Retrieved from https://online. budstandart.com/ua/catalog/doc-page.html?id_doc=74115.
  11. Ettappan, M., Vimala, V., Ramesh, S., & Kesavan, V.T. (2020). Optimal reactive power dispatch for real power loss minimization and voltage stability enhancement using Artificial Bee Colony Algorithm. Microprocessors and Microsystems, 76, article number 103085. doi: 10.1016/j.micpro.2020.103085.
  12. Feng, D., Zhou, H., & Lin, S. (2023). A reactive power optimization method for AC metro power supply system based on particle swarm optimization algorithm. In 2023 IEEE vehicle power and propulsion conference (pp. 1-6). Milan: Institute of Electrical and Electronics Engineers. doi: 10.1109/VPPC60535.2023.10403259.
  13. Feng, J., Chen, M., & Sun, Z. (2024). Hierarchical comprehensive compensation strategy for multi-voltage level traction power supply system. In 2024 21st international conference on harmonics and quality of power (pp. 306311). Chengdu: Institute of Electrical and Electronics Engineers. doi: 10.1109/ICHQP61174.2024.10768811.
  14. Feng, N., Feng, Y., Su, Y., Zhang, Y., & Niu, T. (2024). Dynamic reactive power optimization strategy for AC/DC hybrid power grid considering different wind power penetration levels. IEEE Access, 12, 187471-187482. doi: 10.1109/ACCESS.2024.3392851.
  15. Guamán, W., Benalcazar, P., Cordova-Garcia, J., & Torres, M. (2025). Machine learning-based projections of longterm electricity consumption: The case study of Ecuador. In T. Guarda, F. Portela & M.F. Augusto (Eds.), ARTIIS 2024 international workshops: Advanced research in technologies, information, innovation and sustainability (pp. 174-187). Cham: Springer. doi: 10.1007/978-3-031-83432-5_12.
  16. Guamán, W.P., Pesantez, G.N., Torres, M.A., Falcones, S., & Urquizo, J. (2023). Optimal dynamic reactive power compensation planning in power systems: Case study of Ecuador-Perú interconnection. Electric Power Systems Research, 218, article number 109191. doi: 10.1016/j.epsr.2023.109191.
  17. Hao, F., Zhang, G., Chen, J., & Liu, Z. (2021). Distributed reactive power compensation method in DC traction power systems with reversible substations. IEEE Transactions on Vehicular Technology, 70(10), 9935-9944. doi: 10.1109/tvt.2021.3108030.
  18. Hu, H., Liu, Y., Li, Y., He, Z., Gao, S., Zhu, X., & Tao, H. (2024). Traction power systems for electrified railways: Evolution, state of the art, and future trends. Railway Engineering Science, 32, 1-19. doi: 10.1007/s40534-02300320-6.
  19. IEC 61557-1:2019. (2019). Electrical safety in low-voltage distribution systems up to 1 000 V AC and 1 500 V DC – equipment for testing, measuring or monitoring of protective measures – part 1: General requirements. Retrieved from https://webstore.iec.ch/en/publication/60891.
  20. IEEE Xplore. (n.d.). About IEEE Xplore. Retrieved from https://innovate.ieee.org/about-the-ieee-xplore-digitallibrary/.
  21. ISO 50001:2018. (2018). Energy management systems – requirements with guidance for use. Retrieved from https://www.iso.org/standard/69426.html.
  22. ISO/IEC 17025:2017. (2017). General requirements for the competence of testing and calibration laboratories. Retrieved from https://www.iso.org/standard/66912.html.
  23. Jaisiva, S., Prabaakaran, K., Kumar, C., Lakshmanan, M., Alwabli, A., Jaffar, A., Alharbi, A., & Miyajan, A. (2023). A novel solution for the optimal reactive power dispatch problem using an artificial neural network integrated with the firefly optimization algorithm. Frontiers in Energy Research, 11, article number 1310010. doi: 10.3389/ fenrg.2023.1310010.
  24. Jia, D., Tao, J., Fan, R., & Lu, J. (2020). Research on coordinated control strategy of ITER power supplies and reactive power compensation system. Journal of Fusion Energy, 39, 491-499. doi: 10.1007/s10894-020-00270-5.
  25. Jumani, T.A., Mustafa, M.W., Hamadneh, N.N., Atawneh, S.H., Rasid, M., Mirjat, N.H., Bhayo, M.A., & Khan, I. (2020). Computational intelligence-based optimization methods for power quality and dynamic response enhancement of AC microgrids. Energies, 13(16), article number 4063. doi: 10.3390/en13164063.
  26. Kumar, L., Kar, M.K., & Kumar, S. (2022). Statistical analysis-based reactive power optimization using improved differential evolutionary algorithm. Expert Systems, 40(1), article number e13091. doi: 10.1111/exsy.13091.
  27. Kumar, L., Kar, M.K., & Kumar, S. (2023). Reactive power management of transmission network using evolutionary techniques. Journal of Electrical Engineering & Technology, 18, 123-145. doi: 10.1007/s42835-022-01185-1.
  28. Lin, S., Mu, D., Xu, L., & He, Z. (2025). Parameter optimization method for AC filters in HVDC considering reactive power compensation effectiveness. IEEE Transactions on Power Delivery, 40(1), 365-375. doi: 10.1109/ tpwrd.2024.3495706.
  29. Nishad, D.K., Tiwari, A.N., Khalid, S., & Gupta, S. (2024). Power quality solutions for rail transport using AI-based unified power quality conditioners. Discover Applied Sciences, 6, article number 651. doi: 10.1007/s42452-02406372-5.
  30. Pan, H., Qin, K., Zhang, J., & Yuan, C. (2022). Fruit fly optimization algorithm multi-objective control method for MMC traction power supply system with unbalanced distribution network. International Journal of Dynamics and Control, 10, 1922-1933. doi: 10.1007/s40435-022-00927-3.
  31. 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), article number 212. doi: 10.3390/ app13010212.
  32. Pivnyak, G., Azukovskiy, O., Papaika, Y., Cabana, E.C., Miedziński, B., Polnik, B., Jamróż, A., & Polyanska, A. (2024). Managing the power supply energy efficiency by means of higher harmonics. Energy Policy Journal, 27(4), 59-80. doi: 10.33223/epj/191813.
  33. Rauch, J., & Brückl, O. (2023). Achieving optimal reactive power compensation in distribution grids by using industrial compensation systems. Electricity, 4(1), 78-95. doi: 10.3390/electricity4010006.
  34. Sayeh, K.F., Tamalouzt, S., Sahri, Y., Belaid, S.L., & Bekhiti, A. (2025). Artificial intelligence-based direct power control for power quality improvement in a WT-DFIG system via neural networks: Prediction and classification techniques. Journal of the Franklin Institute, 362(1), article number 107401. doi: 10.1016/j.jfranklin.2024.107401.
  35. Scopus: A comprehensive abstract and citation database for impact makers. (n.d.). Retrieved from https://www. elsevier.com/products/scopus.
  36. Shokouhandeh, H., Latif, S., Irshad, S., Kamarposhti, M.A., Colak, I., & Eguchi, K. (2022). Optimal management of reactive power considering voltage and location of control devices using Artificial Bee Algorithm. Applied Sciences, 12(1), article number 27. doi: 10.3390/app12010027.
  37. Skalozub, V., Zhukovytskyi, I., Bosyi, D., & Lakhno, V. (2024). Usage of specialized intellectual and artificial intelligence systems to improve railway transport technologies. In A. Sładkowski (Ed.), Using artificial intelligence to solve transportation problems (pp. 27-122). Cham: Springer. doi: 10.1007/978-3-031-69487-5_2.
  38. Utama, C., Meske, C., Schneider, J., & Ulbrich, C. (2022). Reactive power control in photovoltaic systems through (explainable) artificial intelligence. Applied Energy, 328, article number 120004. doi: 10.1016/j. apenergy.2022.120004.
  39. Wang, H., Liu, W., Li, Q., Wu, B., & Liu, T. (2021). Compensation of negative sequence for a novel AC integrated power supply system of urban rail transit. IET Generation, Transmission & Distribution, 15(22), 3188-3203 doi: 10.1049/gtd2.12251.
  40. Xu, Y., Pan, X., Guo, J., Sun, X., & Zhang, H. (2023). Dynamic reactive power characteristics analysis and optimization of synchronous condensers. In 2023 10th international forum on electrical engineering and automation (pp. 484488). Nanjing: Institute of Electrical and Electronics Engineers. doi: 10.1109/IFEEA60725.2023.10429104.
  41. Yagup, V.G., & Yagup, K.V. (2024). Analytical method of determining conditions for full compensation of reactive power in the power supply system. Electrical Engineering & Electromechanics, 2, 75-80. doi: 10.20998/2074272x.2024.2.11.
  42. Yuan, J. (2023). Reactive power compensation optimization of distribution network with distributed power supply. Journal of Physics: Conference Series, 2567, article number 012009. doi: 10.1088/17426596/2567/1/012009.
Shpilievyi, M, (2025). Application of artificial intelligence for optimal control of reactive power compensation in AC traction power supply systems. Journal of Kryvyi Rih National University, 23(2), 34-45. https://doi.org/10.31721/2306-5451-2025-2-23-34-45
en