Energy Digest
Summary
Summary
Summary
Summary
Summary
Summary
Summary
Summary
Summary
Technical Papers & Research
AI-curated academic research for power system engineers
Grid Operations & Resilience 4 papers
The proposed AI data center power smoothing system uses a Hybrid Energy Storage System (HESS) that combines a Battery Energy Storage System (BESS) and a supercapacitor (SC) to mitigate power fluctuations. The system employs a Bayesian Optimization approach to optimize real-time control parameters, ensuring grid stability and reducing extreme power demand fluctuations. The system successfully smooths out raw demand fluctuations to fit within a tight target band, suppressing resonant excitation and maintaining grid stability.
A transformer-guided variable-projection estimator is proposed to accurately estimate phasor, DDC parameters, and fault inception time from POW samples, achieving low phasor estimation errors in post-fault windows and online operation. The estimator's mean maximum overshoot (MO) is at most 2.7% on post-fault windows and 2.2% during online operation, outperforming a traditional DFT-based estimator. The proposed estimator reduces phasor estimation errors in DDC offset scenarios.
Adaptive Residual Modeling for Scenario Learning (ARMSL) addresses the challenge of capturing rare observations and extremes in power system operations by proposing a two-stage framework. Stage I models environmental responses and load patterns, while Stage II calibrates extreme marginal quantiles using a reference forecast. The framework outperforms baselines in numerical simulations, achieving significant reductions in pinball loss and feeder error rates.
The increasing share of renewable energy sources in power systems, driven by decarbonization efforts, poses challenges for fault protection and reliability in power systems. Existing protection schemes may not provide accuracy and reliability due to limitations in converter controls and the unique characteristics of inverter-based resources, such as limited negative and zero sequence current injection. New understanding and models are needed to address these challenges and ensure the stability and resilience of power systems with high shares of converter-based resources.
Other 4 papers
Dynamic models of a transmission network and power-electronic devices, including grid-following inverters and UPS-based AI data centers, have been developed. The models cover a range of devices, including inverters, transformers, and battery systems, and are represented in the synchronous dq frame. These models aim to improve the efficiency and accuracy of grid-simulation and optimization studies.
A cooperative game-theoretic framework is proposed for real-time deviation settlement in Community-Based Local Energy Markets, modeling the process as a transferable-utility cooperative game. The framework evaluates three game-theoretical allocation mechanisms, finding that two are efficient and individually rational, but only one guarantees coalitional rationality. The proposed framework is demonstrated using real-world data from a residential energy community in Amsterdam.
A GPU-parallel solver for $N\!-\!1$ contingency analysis that uses a single reference Jacobian and low-rank Woodbury corrections to handle topology changes. The method attains competitive solve times with an optimized CPU baseline and has a minimal impact on workload, making it feasible at transmission scale. It achieves this by separating the change in operating state from the change in topology.
A new synchronverter control strategy for wind turbine grid-side converters is proposed, harnessing the energy storage capability of the DC-link capacitor to emulate the behavior of a synchronous generator and provide grid inertia. The strategy utilizes a dual-controller architecture to maintain DC voltage stability and prevent overvoltage and undervoltage excursions. The proposed method shows superior performance compared to conventional grid-following control, improving grid stability and addressing challenges posed by high renewable energy penetration.
Was this digest helpful?