Real-Time Data Imputation for Low Inertia Nanogrid Digital Twins

We introduce data-driven, scalable digital twins (DTs) and real-time data imputation to improve inverter synchronization in low-inertia microgrids. The DTs act as cyber-physical replicas, enabling real-time monitoring and data-driven control. To address data acquisition inconsistencies, we implement real-time data imputation based on the rate of missing data in a rolling window. To enhance the data stream close to real-time, we employ a variety of data imputation techniques to fill missing values based on the rolling window model. The rolling windows continuously measure data size and the number of missing values, utilizing the most suitable imputation model to reduce computational latency. Numerical case studies validate the efficacy of this approach.