Energy Digest

The proposed method optimizes the capacity of inter-area HVDC tie-lines to address frequency security challenges during HVDC faults in multi-area asynchronously interconnected grids, utilizing an emergency-aware and frequency-constrained approach. It allocates emergency control resources and integrates event-driven emergency frequency control to extract frequency nadir security constraints. The simulation results demonstrate superior performance in balancing economic efficiency with frequency security requirements.

Distributors' use of low-voltage distribution transformers is being optimized by dynamically forecasting thermal protection settings through probabilistic forecasting, which predicts optimal transformer usage and reduces overheating risk. This approach yields a 10-12% additional capacity gain compared to traditional static settings. The method also enables risk-informed operational decisions using predicted percentile values.

Port-Hamiltonian systems are highly structured and energy-based modular frameworks for control systems, exhibiting non-polynomial non-linearities that can be immersed into higher-dimensional polynomial representations. The lifted system preserves important features such as internal interconnection geometry, energy balance relations, and energy dissipation along system trajectories. This allows for the design of stabilizing feedback laws using sum-of-squares optimization and passivity-based control concepts.

Batteries play a critical role in ensuring power balance and reducing costs in microgrid energy management, but extreme temperatures pose significant challenges due to endogenous uncertainty in battery degradation. A new paper proposes an XGBoost-based probabilistic degradation model and parametric model predictive control framework to optimize battery degradation and operational costs. The approach is validated through case studies demonstrating its effectiveness from a full life-cycle perspective.

District heating networks (DHNs) have significant potential to decarbonize residential heating by combining modeling-to-generate-alternatives with power flow simulation techniques to optimize their design. The flexibility in technology choice, sizing, and location enables accommodating different real-world needs while achieving high electrification levels without increasing grid loading. Planners can use a decision-support method developed in this study to explore diverse carbon-neutral DHN designs that balance stakeholders' preferences.

KProxNPLVM is a novel probabilistic latent variable model that relaxes its objective function to improve performance, alleviating an approximation error gap introduced by conventional amortized variational inference. By using the Wasserstein distance as a proximal operator, the algorithm can sidestep this error and achieve improved soft sensor modeling accuracy. Extensive experiments on synthetic and real-world datasets demonstrate the efficacy of KProxNPLVM.