Alexandre Lucas

Abstract
The application of Artificial Intelligence (AI) in the energy sector offers new opportunities for developing flexible, efficient, and sustainable infrastructures. Nevertheless, real-world deployment is still constrained by the lack of large-scale, integrated environments that can evaluate advanced algorithms under realistic operating conditions while ensuring regulatory compliance. This paper presents EnerTEF (which stands for Energy Testing and Experimentation Facility), a federated platform for testing and experimentation in the energy sector designed to address this gap. We introduce a unified TEF architecture that enables full-stack evaluation of intelligent systems, including predictive modeling, optimization, learning under data distribution shifts and federated learning across geographically distributed sites. The framework integrates high-fidelity digital twins, a privacy-preserving data exchange framework and regulatory sandboxing to support transparent, explainable and robust AI development. EnerTEF demonstrates how such a framework can be deployed in critical energy domains through three real-world scenarios including short-term hydropower generation forecasting, coordination between distribution network operators and distributed energy resources and real-time optimization of self-consumption for municipal buildings. Results show that EnerTEF effectively enables the development of novel AI models, improves cross-context generalizability and supports innovation for complex energy infrastructures, ultimately creating a practical, scalable path for addressing different energy-related problems and heterogeneous data.

Abstract
Interoperability among diverse devices, from traditional substation control rooms to modern inverters managing components like Distributed Energy Resources (DERs), is a primary challenge in modern power systems. It is essential for streamlining decision-making and control processes through effective communication, ultimately enhancing energy management efficiency. This paper introduces the open-source Legacy Protocol Converter (LPC) grounded in the IEEE 2030.5 standard, which incorporates advanced features for improved adaptability. The LPC bridges legacy equipment using standard protocols such as Message Queuing Telemetry Transport (MQTT) and Modbus with a light-weight asynchronous Neural Autonomic Transport System (NATS) communication system. In light of the limitations inherent in traditional synchronous RESTful systems-specifically those compliant with IEEE 2030.5 that are incapable of facilitating multiple endpoints-the adoption of asynchronous NATS is implemented. This approach can notably enhance communication flexibility and performance. The implementation is containerized for efficient service orchestration and supports the reusability of solutions. The LPC is engineered for seamless integration of DERs with Energy Management System (EMS), aggregation platforms, and Hardware-in-the-loop (HIL) testing environments. In this paper, the LPC has been tested and further developed in various use cases such as multi-physics optimization involving HIL and fast frequency services, e.g., virtual inertia and load shedding, each in a different architectural setup. The findings validate the applicability of LPC not only for devices within modern power systems, but also for heat pumps in the thermal energy sector, facilitating sector coupling. Moreover, the paper provides additional insights into LPC's functionality, reaffirming its efficacy as a scalable, robust, and user-friendly solution for bridging legacy systems through the enhanced IEEE 2030.5 standard designed for the monitoring and control of DERs.

Abstract
Accurately imputing missing data is critical in time series analysis. The present work compares Foundation Model Chronos against Linear Interpolation, K-Nearest Neighbor Imputer, and Gaussian Mixture Model Imputer with three types of missing data patterns: random, short sequential chunks, and a long sequential chunk. These results confirm that for random missing values, KNN and interpolation yield the highest performance, while Chronos outperforms these on sequences. Indeed, however, for longer sequences of missing values, Chronos starts suffering from cascading errors which eventually allow the simpler imputation methods to outrank it. Another test with limited quantities of training data showed different tradeoffs for the different methods. Unlike KNN and interpolation, which smooth out the gaps, Chronos generates variable synthetic data. This can be beneficial in tasks which require control or simulation. The results highlight the strengths and weaknesses of the imputers and, therefore, offer practical insights into trade-offs between computational complexities, accuracy, and suitability for time series imputation scenarios.

Abstract

Abstract
The rapid rise of Renewable Energy Communities (REC) offers unique opportunities for decentralizing and decarbonizing energy systems but also brings challenges in designing fair mechanisms for distributing the benefits of collective self-consumption. This paper evaluates three approaches for benefit-sharing based on the Shapley value, direct marginal contributions, and system marginal cost. A case study compares these methodologies in terms of practicality, fairness, and impact on financial returns. Additionally, this paper proves that settling local transactions using system marginal costs ensures that all REC participants incur equal or lower costs compared to operating independently.