MATHEMATICAL MODELS FOR PROCESSING DIAGNOSTIC INFORMATION TO SOLVE PROBLEMS OF ASSESSING THE TECHNICAL CONDITION OF ELECTRICAL EQUIPMENT AT POWER PLANTS
Abstract
This article examines approaches to developing mathematical models for processing diagnostic information to assess the technical condition of electrical equipment at power plants. It is shown that traditional monitoring methods based on individual parameters and threshold values no longer provide the required reliability and efficiency as energy facilities become more complex and data volumes grow. Based on an analysis of modern works on statistical diagnostics, the theory of random processes, and machine learning methods, a generalized structural diagram of a mathematical diagnostic model is proposed. It includes a model of an object as a stochastic dynamic system, a measurement model, a feature extraction subsystem, and a decision-making subsystem based on statistical criteria and classification algorithms. Formal expressions are provided for generating a diagnostic residual, calculating integral condition criteria (e.g., the Hotelling T² statistic, the Mahalanobis distance ), and generating diagnostic features in the time and frequency domains. It is shown that the combined use of state models, multivariate statistical control methods, and machine learning algorithms can improve the sensitivity and noise immunity of electrical equipment diagnostics, as well as implement early detection of defects.
Keywords
diagnostic information, technical condition, electrical equipment of power plants, mathematical model, stochastic processes, statistical diagnostics, machine learning, residual signal.
References
- Babak, V. P., et al. Information Provision of Diagnostic Systems for Energy Facilities. Kyiv: Akademperiodyka , 2019.
- Babak, VP Statistical Diagnostics of Electric Power Equipment. Springer, 2020.
- Babak, V. P., et al. "Models and Measures for the Diagnosis of Electric Power Equipment." E3S Web of Conferences , 2019.
- Maican , C. A., et al. "Review of Fault Detection and Diagnosis Methods in Power Generation Plants." Applied Sciences, 2025.
- Patel, H.R. "Fault Detection and Diagnosis Methods in Power Plants." PDPU Journal of Energy and Management, 2018.
- Katser , I., et al. "Machine Learning Methods for Anomaly Detection in Nuclear Power Plant Power Transformers." arXiv:2211.11013, 2022.
- Samiullah, M., et al. "Fault Diagnosis on Induction Motor using Machine Learning and Signal Processing." arXiv:2401.15417, 2024.
- Nguyen, B., et al. "Spatial-Temporal Recurrent Graph Neural Networks for Fault Diagnostics in Power Distribution Systems." arXiv:2210.15177, 2022.
- "PRANA Predictive Analytics and Remote Monitoring System." Official materials from ROTEC.