Solving for globally optimal line switching decisions in AC transmission grids can be intractably slow. Machine learning (ML) models, meanwhile, can be trained to predict near-optimal decisions at a fraction of the speed. Verifying the performance and impact of these ML models on network operation, however, is a critically important step prior to their actual deployment. In this paper, we train a Neural Network (NN) to solve the optimal power shutoff line switching problem. To assess the worst-case load shedding induced by this model, we propose a bilevel attacker-defender verification approach that finds the NN line switching decisions that cause the highest quantity of network load shedding. Solving this problem to global optimality is challenging (due to AC power flow and NN nonconvexities), so our approach exploits a convex relaxation of the AC physics, combined with a local NN search, to find a guaranteed lower bound on worst-case load shedding. These under-approximation bounds are solved via MathOptAI.jl. We benchmark against a random sampling approach, and we find that our optimization-based approach always finds larger load shedding, by an average margin of 33% in the largest test case. Test results are collected on multiple PGLib test cases and on trained NN models which contain more than 10 million model parameters.
DC Optimal Power Flow (DCOPF) is widely utilized in power system operations due to its simplicity and computational efficiency. However, its lossless, reactive power-agnostic model often yields dispatches that are infeasible under practical operating scenarios such as the nonlinear AC power flow (ACPF) equations. While theoretical analysis demonstrates that DCOPF solutions are inherently AC-infeasible, their widespread industry adoption suggests substantial practical utility. This paper develops a unified DCOPF -> ACPF pipeline to recover AC feasible solutions from DCOPF-based dispatches. The pipeline uses four DCOPF variants and applies AC feasibility recovery using both distributed slack allocation and PV/PQ switching. The main objective is to identify the most effective pipeline for restoring AC feasibility. Evaluation across over 10,000 dispatch scenarios on various test cases demonstrates that the structured ACPF model yields solutions that satisfy both the ACPF equations, and all engineering inequality constraints. In a 13,659-bus case, the mean absolute error and cost differences between DCOPF and ACOPF are reduced by 75% and 93%, respectively, compared to conventional single slack bus methods. Under extreme loading conditions, the pipeline reduces inequality violations by a factor of 3 to 5.
Managing reservoir pressure is critical in large-scale geologic carbon sequestration (GCS), as pressure interference among multiple projects targeting the same geologic formation can significantly influence injection performance and long-term storage efficiency. This study evaluates how interference affects the ability of individual wells to access and utilize reservoir pore space. Using basin-scale dynamic reservoir simulations of the Williston Basin and Rate Transient Analysis (RTA) methods, we systematically assessed the role of well spacing, reservoir properties, and injection strategies on pressure evolution and storage outcomes. The results demonstrate that inter-well interference at distances of up to 40 km can reduce cumulative CO2 injection capacity by as much as 56% compared to isolated well scenarios. Decreasing well spacing from 40 km to 20 km exacerbates this effect, reducing the accessible pore volume by up to a 70%, a relationship captured by a second-order polynomial fit. Conversely, optimizing injection rates is a powerful mitigation strategy that can enhance pore volume accessibility by up to 170%, following a power-law relationship. Sensitivity analyses reveal that inherent reservoir properties (porosity and permeability) determine the magnitude of interference, whereas operational choices (well-spacing and injection strategy) govern the trend and temporal evolution of these effects. Furthermore, normalization across reservoir realizations showed consistent interference patterns for identical well configurations, reinforcing the generality and robustness of the findings. These results underscore the necessity of coordinated basin-scale planning among operators to minimize interference, consequently maximizing storage efficiency, and improving the economic viability of large-scale geologic carbon storage.
Power system operators need tools for rapid, real-time counterfactual assessments of grid security under fast-changing conditions. Traditional N-1 contingency analysis lacks dynamic evaluation, especially of frequency swings from common faults. This paper introduces a real-time dashboard framework to screen dynamic contingencies. It assumes: (a) the grid starts in a balanced state; (b) faults can occur randomly on any transmission line, temporarily de-energizing and then reconnecting it within about one second; and (c) contingencies are flagged if post-fault transients cause line flows to exceed safety thresholds. The key contributions are: (1) Overload Indicator: a system-wide metric quantifying integrated N-1 dynamic risk from a given state; (2) Scalable Fault Evaluation Algorithm: a linear-scaling method to assess dynamic fault impacts without brute-force simulations; and (3) Risk Estimation: a Cross Entropy Adaptive Importance Sampling method estimating the likelihood of low probability by high risk events, e.g. associated with potential transformer over-current. We demonstrate the framework on the Israeli transmission power grid (IG).
The increasing interconnection of power systems through AC and DC links enables energy storage units to access multiple electricity markets, yet most existing arbitrage models remain limited to single-market participation. This gap restricts understanding of the economic value and operational constraints associated with cross-border storage operation. To address this, an optimal multi-region energy storage arbitrage model is developed for a grid-scale battery located at one end of an interconnector linking two distinct day-ahead markets. The formulation incorporates battery capacity and ramping limits, converter and interconnector losses, and market-specific buying and selling prices. Using disjunctive linearization of nonlinear terms, this work exactly reformulates the multi-region energy arbitrage optimization as a mixed-integer linear programming problem. The proposed formulation ensures that the battery either charges or discharges from all participating energy markets simultaneously at any given time. Case studies using eight years of Belgian-UK price data demonstrate that multi-region participation can increase arbitrage revenue by more than 40% compared to local energy arbitrage operation only, while also highlighting the negative impact of interconnector congestion on achievable gains. The results indicate that cross-border market access substantially enhances storage profitability while considering the cycle of battery and that the proposed formulation provides a computationally efficient framework for evaluating and operating storage assets in interconnected power systems. Finally, a pseudo-efficiency term is introduced to improve battery utilization by discarding less profitable charging and discharging battery cycles.
