Abstract
The paper is concerned with the development of a methodology for selecting and preparing data for training predictive diagnostic models of electrical equipment using power transformers as an example. Implementing such a monitoring and diagnostic system is a critical aspect of enhancing electrical equipment reliability. Predictive diagnostics enables the early-stage detection of developing defects in power transformers before they escalate. This, in turn, facilitates proactive maintenance planning, reduces unplanned downtime, and minimizes associated economic losses. The study analyzes power transformers failure statistics indicating that 78% of failures can be identified and tracked at an early defect development stage through the use of a predictive diagnostic system. A list of key parameters for the diagnostic model of power transformers is proposed including analytical, electrical parameters and oil condition parameters, ensuring comprehensive monitoring and diagnosis of the main types of power transformer defects. Recommendations are provided for selecting and forming the initial historical data sample, which will be used to create a sample of the power transformer healthy state, necessary for training the predictive diagnostic model. A methodology for expert preparation and filtering of initial data to form the operating condition sample is proposed, including the removal of equipment downtime periods, incorrect data and intervals of equipment operation with observed early signs of defect development. A methodology for calculating parameter deviation limits based on data variability and a confidence coefficient is defined and an algorithm for testing the model on deferred data is described.
Keywords
power transformer, maintenance, reliability improvement, equipment condition, predictive diagnostics, early warning system, condition monitoring, data analysis, machine learning, model training
1. Moubray J. Reliability-Centered Maintenance. NY, Industrial Press, Inc., 1997. 448 p.
2. Karandaev A.S., Yachikov I.M., Khramshin V.R., Nikolaev A.A. Tasks of multiparametric diagnostics of the technical condition of power transformers in on-line monitoring systems. Izvestiya vysshikhuchebnykh zavedeniy. Elektromekhanika [Bulletin of Higher Educational Institutions. Electromechanics], 2016, no. 4, pp. 65-73. (in Russian). doi: 10.17213/0136-3360-2016-4-65-73
3. Kiryanov N.A., Komkov A.N. High-voltage electrical motors predictive diagnostics. Intellektualnaya elektrotehnika [Smart Electrical Engineering], 2024, no. 2, pp. 53-68. (in Russian)
4. Korolev V.I. Methods of predictive monitoring of the electrical machines technical condition. Elektrotekhnicheskie i informatsionnye kompleksy i sistemy [Electrical and data processing facilities and systems], 2023, no. 2, vol. 19, pp. 62-72. (in Russian). doi: 10.17122/1999-5458-2023-19-2-62-72
5. AO «Tekhnicheskaya inspektsiya EES» -Opyt ekspluatatsii system monitoringa vysokovoltnogo oborudovaniya na obektakh OAO «FSK EES». [JSC «UES Technical Inspection» -Experience in operating monitoring systems for high-voltage equipment at FGC UES facilities]. Available at: https://www.ti-ees.ru/fileadmin/f/Conference/guk.pdf. (accessed 14 May 2023).
6. Radionov A.A., Liubimov I.V., Yachikov I.M., Abdulveleev I.R., Khramshina E.A., Karandaev A.S. Method for forecasting the remaining useful life of a furnace transformer based on online monitoring data. Energies. 2023, vol. 16, no. 12, 4630 p. doi: 0.3390/en16124630
7. Gemke R.G. Neispravnosti elektricheskikh mashin [Electrical machine malfunctions]. Saint Petersburg, Energomashizdat Publ., 1989. 334 p. (in Russian)
8. Organization standard 34.01-12-001-2020. Technical requirements for equipping power transformers of 35 kV and more than 35 kV with primary monitoring sensors for automated monitoring and technical diagnostics systems. The ROSSETI Group of Companies. 25 p. Available at: https://www.rosseti.ru/upload/iblock/120/btmjatlcjjq9yrdiuzukq4wca5sb0z1t.pdf (accessed 14 May 2025) (in Russian)
9. State Standard R 30804.4.30. Electric energy. Electromagnetic compatibility of technical equipment. Power quality measurement methods. Moscow, STANDARTINFORM Publ., 2020. 58 p. (In Russian)
10. John J. Power transformers principles and applications. NY, Marcel Dekker, Inc., 2002. 303 p.
11. Harlow J. Electric Power Transformer Engineering. CRC Press, 2003. 496 p.
12. State Standard R 55191-2012. High voltage test techniques. Partial discharge measurements. Moscow, STANDARTINFORM Publ., 2019. 47 p. (In Russian)
13. IEEE Standard for Test Code for Dry-Type Distribution and Power Transformers. IEEE Std C57.113, 2010. 48 p.
14. IEC 60422:2013 Mineral insulating oils in electrical equipment – Supervision and maintenance guidance. Geneva, International Electrotechnical Commission, 2013. 78 p.
15. IEEE Guide for Acceptance and Maintenance of Insulating Oil in Equipment. IEEE Std C57.106, 2015. 72 p.
16. Bishop C. Pattern Recognition and Machine Learning. Springer, 1971. 738 p.
17. Murty N., Devy S. Introduction to pattern recognition and machine learning. World Scientific, 2015. 404 p.
Kiryanov N.A., Komkov A.N. Data Selection and Preparation for Machine Learning Predictive Diagnostic Models Using a Power Transformer as an Example. Elektrotekhnicheskie sistemy i kompleksy [Electrotechnical Systems and Complexes], 2025, no. 2(67), pp. 79-85. (In Russian). https://doi.org/10.18503/2311-8318-2025-2(67)-79-85