Energy Digest

A real-time control policy is proposed for cascaded hydroelectric systems to capture decision-dependent uncertainty, modeling exogenous forecast errors and endogenous uncertainty propagation through a heteroskedastic variance model. The framework ensures reliable system operation under uncertainty using a joint chance-constrained optimization problem, which improves energy generation and reservoir reliability by accounting for uncertainty. The method also enables adaptive risk allocation to enhance resilient hydropower operations.

Physics-informed machine learning surrogates are being explored to accelerate dynamic simulation of power system components. A finite-horizon bound is derived linking surrogate accuracy to algebraic coupling sensitivity, dynamic error amplification, and simulation horizon. Good stand-alone surrogate accuracy does not guarantee accurate in-simulator behavior, with discrepancies concentrated in stressed operating regions.

A data-driven predictive control framework has been proposed to handle non-causal dynamics in stochastic descriptor systems, which can arise in applications like power networks and chemical processes. The framework accommodates algebraic constraints and impulsive modes without explicit system identification by using a causal innovation representation with a noise buffer. This approach was successfully tested on a direct-current (DC) microgrid, demonstrating its effectiveness in predictive control for stochastic descriptor systems.

Dynamic pricing and admission control can maintain equitable access when imposed as a hard state constraint, revealing that price reduction under heavy load may be necessary to prevent unfair outcomes. Conventional monotonic pricing can disproportionately exclude price-elastic users, particularly under high or uncertain demand. A robust dynamic pricing framework is developed to address this issue by integrating safety and fairness constraints with revenue optimization.

Defending the power grid against load-altering attacks using electric vehicle charging proposes to segment cyber infrastructure used by charging station operators to prevent successful hacks. A threat analysis shows that two hacked CSOs can overload two transmission grid branches, exceeding security margins and requiring defense measures. Segmenting CSOs evenly based on installed capacity results in only 23% more segments compared to heuristic optimization results, suggesting potential relevance as a regulatory measure.

The proposed method, PCR-TA, eliminates computational overhead in electromagnetic analysis by directly integrating inter-tape current sharing and radial current bypass behaviors into a finite-element framework, achieving a speedup of approximately 2.4 over field-circuit coupled modeling. A multi-scale approach further increases this speedup to roughly 5.8 for large-scale parallel-wound no-insulation HTS coils. This method allows efficient electromagnetic modeling under both driven and closed-loop operating conditions.

The solvability of real power flow equations is a crucial issue in operational planning and control, particularly with increasing renewable generation that changes steady-state operating points frequently. A new cycle-based solvability condition has been proposed to guarantee the existence and uniqueness of solutions for a given set of power injections. This condition can be used as a foundation for developing more conservative solvability conditions for fully coupled power flow equations.

Tabular foundation models can detect outliers without training but often lack operational context for safety-critical decision-making due to their opacity. A new framework called FoMo-X adds lightweight diagnostic capabilities to these models by leveraging pretrained backbone embeddings and attaching auxiliary diagnostic heads. This approach recovers ground-truth diagnostic signals with high fidelity and negligible inference overhead, making it a scalable path toward trustworthy outlier detection.

A distributed adaptive controller for regulating DC bus voltage in hybrid-electric aircraft propulsion systems has been designed, using back-stepping, adaptive, and passivity-based control techniques. The proposed method enables proportional sharing of electric load among multiple sources, reducing the risk of over-stressing individual sources. Experimental validation demonstrates stable, convergent, and accurate voltage regulation and load-sharing with consideration of unknown power line resistances and inductances.

A multi-period optimization framework is developed to design a voluntary renewable program (VRP) for an electric utility company, aiming to maximize total renewable energy deployments. The framework models the demand function as an exponential decay function based on survey data and derives the optimal pricing policy as a function of grid carbon intensity. Voluntary renewable programs can extend renewable penetration but cannot achieve net-zero emissions or a fully renewable grid.