Article Details

Abstract

To address delays and high false-alarm rates in oil condition monitoring at power plants, this paper proposes an intelligent oil degradation diagnostic system integrating multimodal sensing, dynamic baseline modeling, and self-evolving digital twin technology. The system collects six types of oil parameters in real time, achieves high-accuracy fault identification through a dynamic LSTM baseline model and a mirror diagnostic engine, and uses a reinforcement-learning-based twin platform to simulate degradation paths and optimize operation and maintenance. Results from an actual deployment test at a 300 MW coal-fired power plant show that the system controls the viscosity measurement error within ±1.5% (R2 = 0.998) and demonstrates good monitoring accuracy and diagnostic stability in the fire-resistant oil system of the steam turbine unit, the lubricating oil system of the auxiliary turbine, and the hydraulic oil system of the coal mill. The system can effectively promote the transition of power plant operation and maintenance from “reactive maintenance” to “predictive maintenance” and provide reliable technical support for the safe and stable operation of equipment.

Keywords

power plant main equipment; oil degradation; multimodal sensing; dynamic baseline model; Digital Twin

References

1. [1] Zhong Jie, Liu Wei, Long Guojun, etc Research and application of online monitoring technology for chloride ions in desulfurization systems of power plants [J]. Thermal Power Generation, 2024, 53 (2): 170-176
2. [2] Xu Kai The influence of oil viscosity on the performance of dual field coupling separation device [J]. Fluid Machinery, 2024, 52 (3): 41-48
3. [3] Wang Zehe, Xing Lei, Sun Yanan, etc Research on Oil Viscosity Time Model and Braking Characteristics of Electro Hydraulic Brakes in Cold Regions [J]. Hydraulic and Pneumatic, 2024, 48 (7): 111-119
4. [4] Xu Feng, Zhang Qianqian, Ji Wenlong, etc Oil Spectral Data Diagnosis and Comprehensive Transmission Device Abnormal Wear Localization Method [J]. Spectroscopy and Spectral Analysis, 2024, 44 (5): 1398-1404
5. [5] Chen Yang, Shi Dawei, Li Xuejun, etc Method for fault location and diagnosis of motor bearings by integrating multimodal rotor vibration characteristics [J]. Journal of Vibration Engineering, 2026, 39 (5): 1279-1290
6. [6] Zhou Deng, Li Xuefeng, Yan Gang, etc Research on Impact Damage Identification of Honeycomb Structures Based on Electrical Imaging and Deep Learning [J]. Aerospace Manufacturing Technology, 2024, 67 (13): 84-91
7. [7] Zhang Sihui, Song Wensheng, Tang Tao, etc A modeling and health parameter monitoring method for gray box single-phase PWM rectifier based on digital twin [J]. Journal of Electrical Engineering, 2025, 40 (2): 463-476
8. [8] Liu Xiaodong, Huang Shengwen Development and application of rotating equipment vibration situation awareness platform based on edge computing [J]. Modern Information Technology, 2023, 7 (3): 164-167
9. [9] Li Yongqiu, Xu Jin, Wang Keyou, etc Digital Twin Method for State Monitoring and Single Tube Open Circuit Fault Diagnosis of Small Solid State Transformers [J]. Power System Automation, 2023, 47 (5): 153-161
10. [10] Li Yang, Qiu Jianpei, Song Kun Design and Application of Intelligent Inspection System Based on Multi modal Data Perception of Audio and Video [J]. Modern Information Technology, 2025, 9 (3): 189-193

About the authors

Chen Jianxun (1973-), male, Han ethnicity, from Yueqing, Zhejiang, assistant engineer, research direction: steam turbine maintenance management; Zhou Baiwei (1986-), male, Han nationality, from Yueqing, Zhejiang, engineer, bachelor's degree, research direction: coal mill maintenance management; Yan Xiangye (1972-), male, Han nationality, from Yueqing, Zhejiang, senior engineer, bachelor's degree, research direction: steam turbine maintenance management; Rong Chengbo (1986-), male, Han nationality, from Yueqing, Zhejiang, engineer, bachelor's degree, research direction: steam turbine maintenance management; Liu Liqiang (1978-), male, Han ethnicity, from Yueqing, Zhejiang Province, senior technician, bachelor's degree, research direction: steam turbine maintenance management; Wang Jing (1986-), male, Han ethnicity, from Yueqing, Zhejiang Province, assistant engineer, bachelor's degree, research direction: water, coal, and oil analysis in electric power chemical industry.