Energy Digest

Tests on the POD-P controller of the INELFE Spain-France VSC-HVDC interconnector showed improvements to its gain and compatibility with angle difference control. Simulation studies and field tests were conducted to ensure correct behavior, aiming to enhance damping of power oscillations and increase the controller's contribution to system stability. The modifications aim to improve the effectiveness of the POD-P controller in Continental Europe's power system.

POD controllers can effectively damp electromechanical oscillations in power systems through supplementary control attached to devices such as synchronous machines or power converters. The correct implementation of POD controllers is strongly linked to their settings, and guidelines for implementation have been proposed using synthetic test systems and compliance criteria. The use of POD controllers with grid-following control in voltage source converters has shown potential in damping inter-area oscillations in large-scale power systems.

Conventional model predictive control designs rely on terminal costs and constraints derived from a steady state but can be limited by this assumption; a new approach called policy-guided MPC overcomes this limitation by using a known sub-optimal control policy to define the terminal region. This method establishes closed-loop performance guarantees relative to the guiding policy for both finite and infinite horizon problems without requiring a steady state or reference trajectory. The framework is demonstrated through numerical simulations on an energy management example.

A two-dimensional geometric Laplace transform is developed to monitor DC lines with multiple power converters, allowing for spatial-temporal analysis of faults along the line. The transformed residual can factorize into separate temporal and spatial components, enabling fault detection, localization, and classification in distributed-converter networks. This representation can be used for distributed control and fault diagnosis, including simultaneous fault detection and distinction between plant outages and cable defects.

A novel continuous-time (CT) aggregation model is introduced to reveal the potential intra-hour flexibility of heating, ventilation, and air conditioning (HVAC) loads, which can provide considerable reserve provision for power systems. The model incorporates distributionally robust chance constrains to handle outdoor temperature uncertainty and uses tailored reformulation techniques for tractable calculation. A customized hierarchical dispatch framework integrates the CT aggregation model into reserve provision in power system dispatch to efficiently schedule massive HVACs.

Seed-anchored graph rendering is a method that prioritizes query-local graph evidence in question answering on industry-standard power-grid models, preserving all local evidence for multi-hop items. It outperforms naive descriptions-first rendering, especially under fixed context budgets, with accuracy rising from 0.450 to 0.970. The approach matches or exceeds state-of-the-art graph representations without using LLM graph-construction tokens.

A dimension-reduced approximate dynamic programming (ADP) method is proposed for real-time microgrid operation with massive air-conditioning loads under multiple uncertainties, using a post-decision value function to characterize future operating costs. To address the curse of dimensionality, a consistency-based value function projection and piecewise linear approximation are employed. The proposed method achieves near-optimal operation performance with low computational cost in case studies on 33-bus and 123-bus systems under both deterministic and stochastic conditions.

A novel algorithm called Time-Causal Aggregation-based Approximate Dynamic Programming (TCA-ADP) is proposed to efficiently schedule massive air conditioners in real-time. The algorithm leverages time-causal state variable aggregation to balance computational efficiency with accuracy, enabling near-optimal scheduling through parallel and closed-form disaggregation. TCA-ADP achieves high aggregation accuracy and uncertainty handling while optimizing economic feasibility.

The rapid deployment of distributed energy resources can trigger prosumer synchronisation, where behaviour becomes coordinated to avoid load peaks, but create new peaks at other times. Adopting time-varying tariffs such as Time-of-Use and Real-Time Pricing can reduce reinforcement investment costs by over 50%, but high adoption rates can reverse this trend. Once prosumer adoption exceeds optimal thresholds of 50% for Time-of-Use and 40% for Real-Time Pricing, reinforcement needs increase by up to 39% and 134%.

A control co-design framework is established to optimize sustainability in energy systems, specifically data centers and microgrids. The framework categorizes sustainability metrics into three lifecycle stages - manufacturing, operation, and disposal - supporting optimization and comparative analysis of the metrics for different design options. It provides insights that GHG-equivalent emissions from manufacturing can be significantly higher than those generated during system operation, highlighting the need for sustainable design practices.