Digital Twins for Power Transformers

This paper presents a computationally effective method for the development of Digital Twins (DTs) for power transformers. The DT continually receives the voltage and current measurements from a transformer and utilizes them for simulating the transformer and providing transformer parameters which are not readily measurable (e.g., the magnetizing currents). Employing the Adaptive Discrete Kalman Filter (ADKF), the proposed method can largely remove the measurement noises. The proposed method improves the simulation inaccuracies arising from the low sampling rates of conventional measurement sources (e.g., digital relays) or communication shortcomings (e.g., low update-rate of SCADA). The BCTRAN model with additional components representing the iron core of transformers is implemented in the proposed DT. The proposed method was evaluated using Matlab and EMTP-RV, and test results demonstrate the effectiveness of the proposed method in the development of DTs for power transformers.