Dynamic Thermal Digital Twin of Liquid-Immersed Power Transformer

This paper presents the application of a detailed dynamic thermal-hydraulic network model (DTHNM) to real liquid-immersed power transformers (LIPTs). Unlike existing simple models based on ordinary differential equations (ODEs), such as those described in the standards, the DTHNM is deeply rooted in physics and resolves many key issues, such as distributed heat accumulation, the dependency of losses on tap position, load, and conductor temperature, as well as the influence of ambient temperature and the various cooling equipment operating stages. The DTHNM provides a time-varying global distribution of liquid flow inside the transformer tank and the distribution of flows and temperatures across each transformer part. It is validated here using field data from a three-phase 66-MVA transformer (summer and winter operating periods) and a single-phase 370-MVA transformer (60-hour extended temperature rise test). The DTHNM closely matches the measured temperatures, whereas standard models show larger errors, especially under high-load and/or low-ambient conditions. The main contributions of this paper are: 1) extension of the detailed steady-state thermal-hydraulic network model (steady-state THNM) to a detailed dynamic model for digital twin applications; 2) application of the model to large LIPTs with validation against measurements; 3) comparison of DTHNM accuracy with that of IEC and IEEE loading guide models; 4) analysis of practical issues including data requirements, computation time, real-time feasibility and limitations.