arXiv subjects

Vladimir Dvorkin

Publications and source records attributed to Vladimir Dvorkin.

At least 19 recordsLinked to original sources

arXiv2026

GPU-to-Grid: Voltage Regulation via GPU Utilization Control

While the rapid expansion of data centers poses challenges for power grids, it also offers new opportunities as flexible loads. Existing power system research often abstracts data centers as aggregate resources, while computer system research focuses on GPU energy efficiency and largely ignores grid impacts. To bridge this gap, we develop a GPU-to-Grid framework that couples device-level GPU control with power system objectives. We study distribution-level voltage regulation enabled by LLM inference flexibility, using batch size as a data-center-side control knob that trades off GPU power consumption, inference latency, and token throughput. We first formulate the problem as an optimization problem and then realize it as an online feedback optimization controller, implemented by the data center operator using its own empirical GPU power-performance model and real-time measurements from both the GPU and grid systems. Our key insight is that reducing GPU power alleviates lower-voltage violations, while increasing GPU power mitigates upper-voltage violations; this challenges the common belief that minimizing GPU power is always beneficial to power grids.

eess.SY↗

arXiv2026

Differentially Private Obfuscation of Power Grid Dynamics

Dynamic models of power systems are critical for analyzing grid response to disturbances and blackouts, but the release of real-world dynamic models is hindered by privacy and cybersecurity concerns, as such models carry sensitive information about transmission, generation, and load parameters. We develop an algorithm for synthesizing dynamic grid models from real-world power grids balancing two objectives: the privacy of the source grid, quantitatively measured using the notion of differential privacy, and the fidelity of the synthesized model. The algorithm applies privacy-preserving noise to obfuscate the original grid parameters, but then optimizes the perturbed parameters to ensure that the resulting model dynamics are statistically consistent with those observed in the source grid. Application to the frequency dynamics of the IEEE 30-bus system reveals the inherent privacy-fidelity trade-off: stricter privacy requirements degrade modeling fidelity, yet optimization significantly improves the quality of the synthesized models.

eess.SY↗

arXiv2026

OpenG2G: A Simulation Platform for AI Datacenter-Grid Runtime Coordination

AI's growing compute demand and new datacenter buildouts present major capacity and reliability challenges for the electricity grid, leading to multi-year interconnection delays for new datacenters and bottlenecking AI growth. To ease this strain, datacenters increasingly offer rapid power flexibility in response to grid signals, where the datacenter can increase or decrease its power consumption by adapting its workload in real time. In order to understand the impact of large datacenters on the grid and to facilitate the design of effective coordination strategies, we build OpenG2G, a simulation platform for AI datacenter-grid runtime coordination. We show that OpenG2G is capable of answering a wide range of coordination questions by allowing users to implement and compare various control paradigms (including classic, optimization, and learning-based controllers), and quantify how AI model and deployment choices affect datacenter flexibility and coordination outcomes. This versatility is enabled by OpenG2G's modular and extensible architecture: a datacenter backend driven by real measurements of production-grade AI services, a grid backend built on high-fidelity grid simulators, and a generic controller interface that closes the loop between them. We describe the design of OpenG2G and demonstrate its usefulness through realistic grid scenarios and AI workloads.

cs.LG↗

arXiv2026

Economic Valuation and Optimal Deployment of Static Synchronous Series Compensators for U.S. Power System Expansion

Flexible AC Transmission Systems (FACTS), particularly Static Synchronous Series Compensators (SSSC), can improve network transfer capability and complement restricted transmission expansion. Evaluations of FACTS within large-scale, real-world power system planning are currently lacking. This paper develops a capacity expansion model for the contiguous U.S. power system toward 2050, incorporating SSSC-modified linear power flow equations and accounting for impedance feedback in transmission expansion. Cost-optimal system expansion leverages widespread nationwide SSSC deployment on small-to-medium capacity lines and reduces the number of corridors to be reinforced. Overall, SSSCs reduce annualized system costs by $1.9 billion or decrease transmission expansion requirements by 20%. The most advantageous deployments achieving benefit-cost ratios of 59 concentrated in the Midwest, facilitating the delivery of central U.S. wind power to eastern load centers. The value proposition of SSSCs is robust to cost sensitivities and potential competition from HVDC network expansion, and increases under higher demand growth and more stringent decarbonization policies. These findings provide a blueprint for leveraging SSSC deployment in the U.S. power system.

eess.SY↗

arXiv2026

Online Electricity Pricing from Frequency Measurements

Frequency dynamics in power systems reflect active power imbalance in real time, thereby providing an instantaneous signal to inform electricity pricing. However, existing real-time markets operate on much slower timescales and fail to exploit this signal. In this letter, we develop integrated market--frequency dynamics that enable online pricing directly from frequency measurements. Representing the real-time market as a dynamic price-discovery process, and integrating this process with the grid frequency dynamics, we derive an explicit price formation mechanism from frequency measurements. This mechanism manifests as a distributed PID-like controller for each generator, where frequency response is driven and remunerated by electricity prices derived solely from local frequency measurements.

eess.SY↗

arXiv2026

Online Feedback Optimization of Energy Storage to Smooth Data Center Grid Impacts

The growing electricity demand of AI data centers introduces significant voltage variability in power networks, affecting not only their own operation but also the experience of all users sharing the network. To smooth data center impacts on power networks, we develop an online feedback optimization approach that controls distributed battery energy storage systems to mitigate voltage issues induced by data center operations. The controller adjusts the active and reactive power setpoints of distributed battery systems in response to voltage measurements, with a two-fold objective: managing voltage to minimize the magnitude of constraint violations and smoothing voltage profiles. Control performance is evaluated in a high-fidelity simulation environment that integrates a three-phase distribution feeder and a detailed battery system model, and benchmarked against a local control approach with similar objectives but without optimality guarantees and constraint enforcement. We show that the proposed controller delivers consistent voltage regulation in the long term, while the local control approach pursues the objectives more aggressively but quickly hits the storage limits.

eess.SY↗

arXiv2026

DiffOPF: Diffusion Solver for Optimal Power Flow

The optimal power flow (OPF) is a multi-valued, non-convex mapping from loads to dispatch setpoints. The variability of system parameters (e.g., admittances, topology) further contributes to the multiplicity of dispatch setpoints for a given load. Existing deep learning OPF solvers are single-valued and thus fail to capture the variability of system parameters unless fully represented in the feature space, which is prohibitive. To solve this problem, we introduce a diffusion-based OPF solver, termed \textit{DiffOPF}, that treats OPF as a conditional sampling problem. The solver learns the joint distribution of loads and dispatch setpoints from operational history, and returns the marginal dispatch distributions conditioned on loads. Unlike single-valued solvers, DiffOPF enables sampling statistically credible warm starts with favorable cost and constraint satisfaction trade-offs. We explore the sample complexity of DiffOPF to ensure the OPF solution within a prescribed distance from the optimization-based solution, and verify this experimentally on power system benchmarks.

eess.SY↗

arXiv2026

Outage Identification from Electricity Market Data: Quickest Change Detection Approach

Power system outages expose market participants to significant financial risk unless promptly detected and hedged. We develop an outage identification method from public market signals grounded in the parametric quickest change detection (QCD) theory. Parametric QCD operates on stochastic data streams, distinguishing pre- and post-change regimes using the ratio of their respective probability density functions. To derive the density functions for normal and post-outage market signals, we exploit multi-parametric programming to decompose complex market signals into parametric random variables with a known density. These densities are then used to construct a QCD-based statistic that triggers an alarm as soon as the statistic exceeds an appropriate threshold. Numerical experiments on a stylized PJM testbed demonstrate rapid line outage identification from public streams of electricity demand and price data.

eess.SY↗

arXiv2025

Probabilistic Dynamic Line Rating with Line Graph Convolutional LSTM

Dynamic line rating (DLR) is an effective approach to enhancing the utilization of existing transmission line infrastructure by adapting line ratings according to real-time weather conditions. Accurate DLR forecasts are essential for grid operators to effectively schedule generation, manage transmission congestion, and lower operating costs. As renewable generation becomes increasingly variable and weather-dependent, accurate DLR forecasts are also crucial for improving renewable utilization and reducing curtailment during congested periods. Deterministic forecasts, however, often inadequately represent actual line capacities under uncertain weather conditions, leading to operational risks and costly real-time adjustments. To overcome these limitations, we propose a novel network-wide probabilistic DLR forecasting model that leverages both spatial and temporal information, significantly reducing the operational risks and inefficiencies inherent in deterministic methods. Case studies on a synthetic Texas 123-bus system demonstrate that the proposed method not only enhances grid reliability by effectively capturing true DLR values, but also substantially reduces operational costs.

eess.SY↗

arXiv2025

Constrained Diffusion Models for Synthesizing Representative Power Flow Datasets

High-quality power flow datasets are essential for training machine learning models in power systems. However, security and privacy concerns restrict access to real-world data, making statistically accurate and physically consistent synthetic datasets a viable alternative. We develop a diffusion model for generating synthetic power flow datasets from real-world power grids that both replicate the statistical properties of the real-world data and ensure AC power flow feasibility. To enforce the constraints, we incorporate gradient guidance based on the power flow constraints to steer diffusion sampling toward feasible samples. For computational efficiency, we further leverage insights from the fast decoupled power flow method and propose a variable decoupling strategy for the training and sampling of the diffusion model. These solutions lead to a physics-informed diffusion model, generating power flow datasets that outperform those from the standard diffusion in terms of feasibility and statistical similarity, as shown in experiments across IEEE benchmark systems.

cs.LG↗

arXiv2025

Synthesizing Grid Data with Cyber Resilience and Privacy Guarantees

Differential privacy (DP) provides a principled approach to synthesizing data (e.g., loads) from real-world power systems while limiting the exposure of sensitive information. However, adversaries may exploit synthetic data to calibrate cyberattacks on the source grids. To control these risks, we propose new DP algorithms for synthesizing data that provide the source grids with both cyber resilience and privacy guarantees. The algorithms incorporate both normal operation and attack optimization models to balance the fidelity of synthesized data and cyber resilience. The resulting post-processing optimization is reformulated as a robust optimization problem, which is compatible with the exponential mechanism of DP to moderate its computational burden.

eess.SY↗

arXiv2025

Optimization over Trained Neural Networks: Difference-of-Convex Algorithm and Application to Data Center Scheduling

When solving decision-making problems with mathematical optimization, some constraints or objectives may lack analytic expressions but can be approximated from data. When an approximation is made by neural networks, the underlying problem becomes optimization over trained neural networks. Despite recent improvements with cutting planes, relaxations, and heuristics, the problem remains difficult to solve in practice. We propose a new solution based on a bilinear problem reformulation that penalizes ReLU constraints in the objective function. This reformulation makes the problem amenable to efficient difference-of-convex algorithms (DCA), for which we propose a principled approach to penalty selection that facilitates convergence to stationary points of the original problem. We apply the DCA to the problem of the least-cost allocation of data center electricity demand in a power grid, reporting significant savings in congested cases.

math.OC↗

arXiv2025

Probabilistic Dynamic Line Rating Forecasting with Line Graph Convolutional LSTM

Dynamic line rating (DLR) is a promising solution to increase the utilization of transmission lines by adjusting ratings based on real-time weather conditions. Accurate DLR forecast at the scheduling stage is thus necessary for system operators to proactively optimize power flows, manage congestion, and reduce the cost of grid operations. However, the DLR forecast remains challenging due to weather uncertainty. To reliably predict DLRs, we propose a new probabilistic forecasting model based on line graph convolutional LSTM. Like standard LSTM networks, our model accounts for temporal correlations between DLRs across the planning horizon. The line graph-structured network additionally allows us to leverage the spatial correlations of DLR features across the grid to improve the quality of predictions. Simulation results on the synthetic Texas 123-bus system demonstrate that the proposed model significantly outperforms the baseline probabilistic DLR forecasting models regarding reliability and sharpness while using the fewest parameters.

eess.SY↗

arXiv2025

Regression Equilibrium in Electricity Markets

In two-stage electricity markets, renewable power producers enter the day-ahead market with a forecast of future power generation and then reconcile any forecast deviation in the real-time market at a penalty. The choice of the forecast model is thus an important strategy decision for renewable power producers as it affects financial performance. In electricity markets with large shares of renewable generation, the choice of the forecast model impacts not only individual performance but also outcomes for other producers. In this paper, we argue for the existence of a competitive regression equilibrium in two-stage electricity markets in terms of the parameters of private forecast models informing the participation strategies of renewable power producers. In our model, renewables optimize the forecast against the day-ahead and real-time prices, thereby maximizing the average profits across the day-ahead and real-time markets. By doing so, they also implicitly enhance the temporal cost coordination of day-ahead and real-time markets. We base the equilibrium analysis on the theory of variational inequalities, providing results on the existence and uniqueness of regression equilibrium in energy-only markets. We also devise two methods to compute regression equilibrium: centralized optimization and a decentralized ADMM-based algorithm.

eess.SY↗

arXiv2024

Privacy-Preserving Convex Optimization: When Differential Privacy Meets Stochastic Programming

Convex optimization finds many real-life applications, where--optimized on real data--optimization results may expose private data attributes (e.g., individual health records, commercial information), thus leading to privacy breaches. To avoid these breaches and formally guarantee privacy to optimization data owners, we develop a new privacy-preserving perturbation strategy for convex optimization programs by combining stochastic (chance-constrained) programming and differential privacy. Unlike standard noise-additive strategies, which perturb either optimization data or optimization results, we express the optimization variables as functions of the random perturbation using linear decision rules; we then optimize these rules to accommodate the perturbation within the problem's feasible region by enforcing chance constraints. This way, the perturbation is feasible and makes different, yet adjacent in the sense of a given distance function, optimization datasets statistically similar in randomized optimization results, thereby enabling probabilistic differential privacy guarantees. The chance-constrained optimization additionally internalizes the conditional value-at-risk measure to model the tolerance towards the worst-case realizations of the optimality loss w.r.t. the non-private solution. We demonstrate the privacy properties of our perturbation strategy analytically and through optimization and machine learning applications.

math.OC↗

arXiv2024

Agent Coordination via Contextual Regression (AgentCONCUR) for Data Center Flexibility

A network of spatially distributed data centers can provide operational flexibility to power systems by shifting computing tasks among electrically remote locations. However, harnessing this flexibility in real-time through the standard optimization techniques is challenged by the need for sensitive operational datasets and substantial computational resources. To alleviate the data and computational requirements, this paper introduces a coordination mechanism based on contextual regression. This mechanism, abbreviated as AgentCONCUR, associates cost-optimal task shifts with public and trusted contextual data (e.g., real-time prices) and uses regression on this data as a coordination policy. Notably, regression-based coordination does not learn the optimal coordination actions from a labeled dataset. Instead, it exploits the optimization structure of the coordination problem to ensure feasible and cost-effective actions. A NYISO-based study reveals large coordination gains and the optimal features for the successful regression-based coordination.

eess.SY↗

arXiv2024

Evaluating Offshore Electricity Market Design Considering Endogenous Infrastructure Investments: Zonal or Nodal?

Policy makers are formulating offshore energy infrastructure plans, including wind turbines, electrolyzers, and HVDC transmission lines. An effective market design is crucial to guide cost-efficient investments and dispatch decisions. This paper jointly studies the impact of offshore market design choices on the investment in offshore electrolyzers and HVDC transmission capacity. We present a bilevel model that incorporates investments in offshore energy infrastructure, day-ahead market dispatch, and potential redispatch actions near real-time to ensure transmission constraints are respected. Our findings demonstrate that full nodal pricing, i.e., nodal pricing both onshore and offshore, outperforms the onshore zonal combined with offshore nodal pricing or offshore zonal layouts. While combining onshore zonal with offshore nodal pricing can be considered as a second-best option, it generally diminishes the profitability of offshore wind farms. However, if investment costs of offshore electrolyzers are relatively low, they can serve as catalysts to increase the revenues of the offshore wind farms. This study contributes to the understanding of market designs for highly interconnected offshore power systems, offering insights into the impact of congestion pricing methodologies on investment decisions. Besides, it is useful towards understanding the interaction of offshore loads like electrolyzers with financial support mechanisms for offshore wind farms.

econ.GN↗

arXiv2023

Uncertainty-Informed Renewable Energy Scheduling: A Scalable Bilevel Framework

This work proposes an uncertainty-informed bid adjustment framework for integrating variable renewable energy sources (VRES) into electricity markets. This framework adopts a bilevel model to compute the optimal VRES day-ahead bids. It aims to minimize the expected system cost across day-ahead and real-time stages and approximate the cost efficiency of the stochastic market design. However, solving the bilevel optimization problem is computationally challenging for large-scale systems. To overcome this challenge, we introduce a novel technique based on strong duality and McCormick envelopes, which relaxes the problem to a linear program, enabling large-scale applications. The proposed bilevel framework is applied to the 1576-bus NYISO system and benchmarked against a myopic strategy, where the VRES bid is the mean value of the probabilistic power forecast. Results demonstrate that, under high VRES penetration levels (e.g., 40%), our framework can significantly reduce system costs and market-price volatility, by optimizing VRES quantities efficiently in the day-ahead market. Furthermore, we find that when transmission capacity increases, the proposed bilevel model will still reduce the system cost, whereas the myopic strategy may incur a much higher cost due to over-scheduling of VRES in the day-ahead market and the lack of flexible conventional generators in real time.

eess.SY↗