This paper compares the Chronos foundation model to established imputation methods such as linear interpolation, KNN, and Gaussian Mixture Models for filling missing values in time series. Results demonstrate Chronos’s strength in maintaining temporal consistency and generating realistic synthetic data, despite higher computational demands — insights that are valuable for improving energy system simulations and resilience.

This paper will be presented by ELES at the 17th Conference of Slovenian Electric Power Engineers (CIGRE-CIRED Slovenia). It explores the use of protection relay data for circuit breaker control. It is highly aligned with Use Case 1 of the TEF TSO and serves as a foundational contribution to the topic. The insights presented are expected to spark a fruitful discussion among system operators and market players on the evolving needs and opportunities in protection relay analytics—critical for shaping the design and functionalities of the Testing and Experimentation Facility for TSOs under EnerTEF.