Energy Digest

Advanced Metering Infrastructure (AMI) is vulnerable to denial-of-service and time-delay attacks, which can compromise data availability. The attacks result in increased latency, higher communication unavailability rates, and reduced data reliability, with estimated economic impacts of $1,009/day under normal market conditions and up to $5,097/day during high-price events. A 500ms delay or 50% packet loss scenario can cause significant network degradation.

A novel approach using reinforcement learning is proposed to construct scenario trees for multistage stochastic model predictive control, focusing on its impact on downstream decisions rather than matching underlying probability distributions. The method consistently achieves the highest profit in a risk-averse battery arbitrage problem and outperforms classical methods, with a learned policy exhibiting greater robustness and tail-risk characteristics. The resulting scenario trees are compact and selectively branching, capturing high-impact events while keeping most trajectories nearly deterministic.

Digital Twins (DTw) in High Voltage Direct Current transmission systems rely on standardized data exchange for effective operation. The lack of standard protocols leads to vendor lock-in and interoperability issues among different system vendors. Standardizing the Common Information Model (CIM) defined in IEC 61970 can enable seamless data exchange and harmonization with other standards, such as IEC 61850.

A hierarchical ReAct Energy Management System uses agentic AI to coordinate schedules for shiftable appliances, EV charging, and thermal control, with physical authorization separated from language generation. The system successfully generated feasible single-appliance mixed-integer schedules in 83.3% of runs, with accepted costs slightly higher than the mathematical optimization optimum. The results establish a fail-closed workflow for agentic energy scheduling under the declared physical model.

A life-cycle optimization model is developed for deep-sea co-located energy farms, capturing multi-layer marine-space complementarity, wake effects, and output variability. The adaptive piecewise linearization method simplifies the original mixed-integer quadratic programming problem into a more computationally efficient form. Case studies show that incorporating multi-energy complementarity significantly enhances economic performance of deep-sea CEFs.

Large-scale renewable power-to-hydrogen (ReP2H) plants in cold regions suffer from prolonged startup times and repeated thermal stress due to lack of coordinated heat management. A new framework presents a plant-wide thermal topology, enabling preheating, thermal standby, and waste heat recovery through bidirectional thermal coupling between stacks and the plant's thermal utility system. The proposed framework improves energy efficiency by 0.99%, increases hydrogen yield by 1.50%, and reduces the levelized cost of hydrogen by 3.22%.