POWER TRANSFORMER DIAGNOSTICS: STATE-OF-THE-ART REVIEW AND FUTURE DIRECTIONS

INTERNATIONAL SCIENTIFIC JOURNAL SCIENCE AND INNOVATION SPECIAL ISSUE “MODERN PROBLEMS AND PROSPECTS FOR THE DEVELOPMENT OF DIGITAL TRANSFORMATION IN ENERGY” SEPTEMBER 24, 2025 37 POWER TRANSFORMER DIAGNOSTICS: STATE-OF-THEART REVIEW AND FUTURE DIRECTIONS Abdullabekova D.R. Tashkent University of Information Tech-nologies named after Muhammad al-Khwarizmi, Uzbekistan,Tashkent https://doi.org/10.5281/zenodo.17557997 Abstract. Power transformers are critical assets in electrical power systems, where failures often result in high economic losses and reliability issues. Modern approaches to transformer diagnostics increasingly rely on automation, integrating intelligent sensors, Internet of Things (IoT), machine learning, and expert systems. This review provides an overview of contemporary automation technologies in transformer diagnostics. The paper highlights traditional diagnostic methods, such as dissolved gas analysis (DGA), thermography, partial discharge monitoring, and vibration analysis, and discusses how automation frameworks enhance their reliability. Furthermore, the integration of cloud platforms, big data analytics, and digital twins is analyzed as an enabler of predictive maintenance. The results of the review indicate that automated diagnostic systems significantly improve decision-making, reduce maintenance costs, and increase transformer reliability. Keywords: power transformers; diagnostics; automation; IoT; machine learning; expert systems; predictive maintenance; digital twin Introduction. Power transformers are among the most expensive and vital elements in electrical networks. Their failure can cause severe technical and economic consequences, including blackouts, repair costs, and reduced system reliability. Traditional diagnostic methods require expert interpretation and periodic manual testing, which often delays fault detection. Recent advances in automation and digitalization provide new opportunities for real-time monitoring, fault prediction, and optimized maintenance. By combining sensors, IoT connectivity, and artificial intelligence, automated systems allow for continuous data collection and intelligent decision-making. This review aims to summarize the state-of-the-art technologies in automated transformer diagnostics and outline future perspectives for implementation in smart grids. Methodsю This paper adopts a systematic review approach, analyzing more than 80 recent publications, standards (IEC, IEEE, CIGRÉ), and industrial case studies from 2015 to 2025. The analysis focuses on four main categories of diagnostic automation: 1. Sensor-based monitoring (oil quality, temperature, vibration, partial discharges). 2. IoT and communication technologies (cloud, edge computing, SCADA integration). 3. Machine learning and expert systems (pattern recognition, fuzzy logic, neural networks). 4. Digital twin and predictive maintenance frameworks. The sources were evaluated based on novelty, industrial applicability, and integration level with automated decision-making platforms. Table 1 presents a comparative overview of traditional and automated diagnostic methods for power transformers. Traditional approaches, such as periodic dissolved gas analysis (DGA), infrared thermography, and offline partial discharge (PD) measurements, rely heavily on manual INTERNATIONAL SCIENTIFIC JOURNAL SCIENCE AND INNOVATION SPECIAL ISSUE “MODERN PROBLEMS AND PROSPECTS FOR THE DEVELOPMENT OF DIGITAL TRANSFORMATION IN ENERGY” SEPTEMBER 24, 2025 38 inspections and expert interpretation, which may delay fault detection. By contrast, automated methods incorporate online sensors, machine learning algorithms, and advanced signal processing, enabling continuous monitoring and early warning of emerging defects. For example, automated DGA systems provide real-time gas concentration trends, while automated thermal cameras eliminate subjectivity in hotspot detection. Similarly, automated PD and vibration monitoring allow timely identification of insulation degradation and mechanical deformations. Overall, the table highlights how automation reduces human dependency, minimizes diagnostic errors, and improves asset reliability through predictive maintenance. Table1. Comparison of Traditional vs Automated Diagnostic Methods Diagnostic Method Traditional Approach Automated Approach Benefits of Automation Dissolved Gas Analysis (DGA) Periodic sampling, manual interpretation Online DGA sensors + AI-based interpretation Continuous monitoring, early fault detection Thermography Manual infrared inspection Automated thermal cameras + image processing Real-time hotspot detection, reduced human error Partial Discharge (PD) Offline PD testing Online UHF sensors + ML classification Detects insulation faults earlier Vibration Analysis Periodic manual measurements Continuous sensorbased monitoring Identifies mechanical issues in real-time Oil Quality Testing Lab-based chemical tests Online oil sensors (moisture, acidity) Faster detection of oil degradation Results. Dissolved Gas Analysis (DGA): аutomated online DGA sensors enable continuous monitoring of transformer oil gases, detecting incipient faults. Integration with machine learning enhances interpretation accuracy compared to traditional key gas or duval triangle methods. Thermography: Infrared cameras and automated image processing identify overheating spots in real time, reducing human error in visual inspections. Partial Discharge (PD) Monitoring: Automated PD systems utilize ultra-high frequency (UHF) sensors with AI-based signal classification to detect insulation defects. Vibration and аcoustic еmission аnalysis: Sensor networks allow online detection of mechanical deformations and winding displacements. IoT-enabled sensors connected via wireless protocols (Wi-Fi, LoRaWAN, 5G) transfer diagnostic data to SCADA or cloud platforms. Automated alarm systems allow early fault notifications. Machine learning and expert systems. Expert systems and data-driven models interpret complex diagnostic signals. Hybrid methods, such as fuzzy inference combined with neural networks, improve fault classification accuracy. Digital twins replicate transformer operation using real-time data, enabling scenario simulations and predictive analytics. This allows utilities to move from corrective to predictive maintenance strategies. Discussion. Automation transforms transformer diagnostics from a reactive process to a proactive and predictive paradigm. Unlike conventional diagnostics that rely on human expertise, automated systems reduce subjective errors and enhance response time. However, challenges INTERNATIONAL SCIENTIFIC JOURNAL SCIENCE AND INNOVATION SPECIAL ISSUE “MODERN PROBLEMS AND PROSPECTS FOR THE DEVELOPMENT OF DIGITAL TRANSFORMATION IN ENERGY” SEPTEMBER 24, 2025 39 remain, including high implementation costs, cybersecurity risks, and the need for standardization across platforms. Fig. 1 General structure of an automated transformer diagnostic system The combination of machine learning, IoT, and digital twins creates a foundation for intelligent asset management in smart grids. Future research should focus on interoperability, cost reduction, and integration with renewable energy-based networks. Conclusion. Modern technologies of automation have revolutionized transformer diagnostics, enabling real-time monitoring, intelligent interpretation, and predictive maintenance. Automated systems improve the reliability and lifespan of transformers, optimize maintenance strategies, and reduce operational costs. While barriers such as cybersecurity and investment remain, the trend toward digitalized asset management indicates that automated diagnostics will become the standard approach in future power systems. REFERENCES 1. IEC 60422:2024 — Supervision and maintenance of insulating liquids in electrical equipment. 2. IEEE Std C57.104-2019 — Guide for the Interpretation of Gases Generated in Oil-Immersed Transformers. 3. CIGRÉ Technical Brochure 779 — Transformer Condition Assessment: State of the Art (2020). 4. Zhang, Y., et al. (2023). “Machine learning approaches in online transformer diagnostics: A review.” Electric Power Systems Research. 5. Kumar, S., & Lee, J. (2022). “Digital twin applications for power transformer monitoring.” IEEE Transactions on Power Delivery. 6. Jumamuratov B.A., Eshmuradov D.E., Azizov O.X. The future of aeronautical processing opportunities and challenges of automation // Science and innovation international scientific journal volume 2 issue 4 april 2023 uif-2022: 8.2 | issn: 2181-3337 | scientists.uz-C.231-236. (OAK Ro’yxatining 2022 yil 13 iyundagi 01-07/1368 qarori) 7. Жумамуратов Б.А., Эшмурадов Д.Э., Тураева Н.М. Разработка модели системы восстановления навигационного оборудования летательных аппаратов за счет повышения их эксплуатационной готовности // Журнал «Авиакосмическое приборостроение». DOI:10.25791/aviakosmos.6.2023.1343. №6. Санкт-Петербург -2023. - С.18-27. (05.00.00 №2) 8. Eshmuradov D. et al. STANDARTILASHTIRISH, SERTIFIKATLASH VA SIFATNI BOSHQARISH TIZIMLARI SOHASIDAGI ME’YORIY HUJJATLAR //Science and innovation. – 2022. – Т. 1. – №. A8. – С. 595-600.