CS Paper Radar

AI Agents Executing an RMS and EMT Ride-Through Workflow on a 2,000-Bus Synthetic Texas Grid: A Case Study with Open-Source Tools

Tam Kemabonta, Brighton Ombuki, Vladimir Abdelnour

eess.SY ·submitted 2026-11-17 04:00 UTC · v1
code/data

Abstract

This paper reports a case study of engineer-supervised AI agents executing a predefined RMS and EMT ride-through workflow on the 2,000-bus ACTIVSg2000 synthetic Texas grid with open-source tools. The test case is a 133 MW battery with a co-located 130 MW data-center load, assessed against ride-through criteria fixed before simulation and for loss of synchronism elsewhere in the grid. One agent ran 44 RMS cases and 15 without-project comparisons in ANDES; the other ran a 220-row EMT intake bench and an 84-row system study in DPsim and ParaEMT on a Kron-reduced area. Across the 44 RMS cases, 127 criterion cells fail, including loss of synchronism in all 44. Two findings concern the project plant. The battery keeps absorbing reactive power during sags on the light-load schedule (B3, as low as -19.8 Mvar on S4 T1), and the data-center load returns only at 5.00-5.56 s after long bolted faults (L3), which misses the ERCOT load-return requirement. In 13 of the 15 comparisons, which the authors selected after slips or speed-gate stops, the same machines slip without the project; the other 29 runs have no without-project comparison. RMS and EMT values differ in network, plant model, controls, operating point, and base, and so they are reported as separate, unaligned observations. The paper reports the errors the authors identified in the agent logs, 63 author messages, and 3.8 days from agent start to the integrated study report. The agents executed and documented the predefined ride-through and stability workflow, which covers one part of an interconnection study.

Comment
5 pages, 1 figure, 4 tables. Submitted to the 2027 IEEE PES General Meeting

Closest papers

Nearest neighbours by meaning across the whole index.

Low-Voltage Ride-Through in AI Data Centers via Coordinated Battery and Computing Load Control

eess.SY ·2026-11-14 ·Xiaoyang Wang, Xin Chen, Fang Chen Lin, Prasad Enjeti, Ali Yazdanpanah +2 more
hot direction similarity 0.59
Abstract · 5 pages, 7 figures

The ability of large AI data centers to ride through grid voltage sags is critical to power system stability and reliability. This paper frames data center low-voltage ride-through (LVRT) as a DC-link power balance problem. Once the rectifier reaches its current limit, the power drawn from the grid decreases in proportion to the grid voltage. The DC-link capacitor can supply the resulting power deficit only briefly before the DC-link voltage falls low enough to trigger undervoltage protection tripping. To address this issue, this paper proposes effective control strategies for coordinating uninterruptible power supply (UPS) battery and computing load curtailment to offset the power deficit and enable LVRT at minimum control cost. Three control scenarios are studied: UPS battery support alone, computing load curtailment alone, and the coordinated use of both resources. Electromagnetic transient (EMT) simulations using the TAMU-ERCOT AI data center model demonstrate that battery support and the coordinated strategy achieve LVRT for all tested sags, including a near-zero-voltage fault, while load curtailment alone rides through moderate voltage sags.

0.22

Mitigating the Grid Impact of Gigawatt-Scale AI Data Centers via High-Fidelity Model Flexibility

eess.SY math.OC ·2026-11-18 ·Eihab E. E. Ahmed, Saeed Manshadi
hot direction similarity 0.59
Abstract

Gigawatt-scale artificial intelligence data centers (AIDCs) are being connected at a single point of interconnection (POI) faster than transmission can be reinforced, yet studies of their flexibility rely on DC power flow and simplified load envelopes that omit voltage, cooling dynamics, and workload deadlines. This paper addresses this gap by embedding a high-fidelity AIDC model, with separate inference and training pools, hall thermal mass, PV, a battery energy storage system (BESS) with reactive capability, and a STATCOM, in a five-minute multi-period AC optimal power flow. Simulations on the RTS-GMLC test system show that active-power flexibility mitigates local price spikes and relieves congestion, whereas reactive support from the BESS inverter or a STATCOM is needed to keep the POI voltage away from sagging around its 0.95~pu limit. Sensitivity analyses show that cost-driven charging of large batteries can induce new morning congestion and that storage largely substitutes for wider workload deferral.

0.30

TuneBench: Benchmarking AI Agents for Power System Dynamic Model Calibration

eess.SY ·2026-11-03 ·Andrea Pomarico, Salvatore Tessitore, Giorgio Giannuzzi, Alberto Berizzi
benchmarkcode/data similarity 0.59
Abstract

Large Language Model (LLM)-based agents can automate complex engineering workflows by interacting with tools, interpreting feedback, and making iterative decisions. This paper introduces TuneBench, a benchmark for evaluating Agentic AI in power system dynamic model calibration, where controller parameters are adjusted to reproduce target dynamic responses. The benchmark reflects a common Transmission System Operator (TSO) practice, in which equivalent models used for Dynamic Security Assessment (DSA) are calibrated against measurements from grid-code compliance tests when proprietary models are unavailable. Two cases of increasing complexity are considered: the governor and Automatic Voltage Regulator (AVR) of a synchronous generator and the control system of a Battery Energy Storage System (BESS), both tuned against active-power and voltage step tests within a budget of 15 simulations. LLM-based agents are compared with random search and Genetic Algorithm (GA) baselines over 10 independent runs. Only the most advanced models, Opus 5.5 and GPT-6, accurately reproduce all target trajectories, with median R2 above 0.995 in every test and tracking errors up to two orders of magnitude lower than the baselines. The agent logs show that these agents autonomously extract dynamic performance indices, such as settling time and damping, and relate them to specific control parameters, decomposing the tuning problem into physically interpretable sub-problems. The benchmark is publicly available at https://github.com/Power-Agent/PowerAgentBench/tree/main/benchmarks/dynamic/level3.

0.42

Comparison of Open-Source PSCAD EMT Models for Large-Scale AI Data Centers in Smart Grids

eess.SY ·2026-11-17 ·Alex Chen, Ahmad Mohammad Saber, Deepa Kundur
similarity 0.58
Abstract · 5 pages, 2 figures, submitted to PESGM 2027

The rapid deployment of artificial intelligence data centers (AIDCs) introduces gigawatt-scale, highly nonlinear, and fast-varying electronic loads that challenge grid voltage regulation, frequency stability, and disturbance ride-through. Capturing their transient behavior during grid faults and rapid computational workload swings requires high-fidelity electromagnetic transient (EMT) simulations rather than root mean square models. However, limited methodologies exist to audit the component-level adequacy of EMT AIDC models. This paper proposes a software-agnostic, 41-point fidelity framework spanning 11 categories that synthesizes 37 criteria from an existing modeling guideline with four novel requirements: frequency ride-through compliance, disaggregated fan and chiller cooling loads, temperature-dependent thermal response, and DC information technology (IT) load representations. Applying this framework to the four publicly available AIDC models in PSCAD/EMTDC reveals varying degrees of completeness: while the models capture IT load transitions, uninterruptible power supplies, and transfer switches for bulk stability studies, critical gaps remain in electro-thermal cooling coupling, internal protective relaying, and site-level controls. Addressing these gaps will enable comprehensive protection coordination and ride-through compliance studies.

0.22

Local AI Agents in Power Electronics Simulation: Deployment, Measurements and Costs

cs.AI eess.SY ·2026-11-18 ·Yuzhuo Li, Pasan Gunawardena, Yunwei Li
similarity 0.58
Abstract · 27 pages, 23 figures, 12 tables

Agentic artificial intelligence (AI) is emerging in power electronics, where it drives the simulation loop that engineers used to close largely by hand. In this loop, a converter design, or a study of the network it connects to, is built, run, measured and revised until it satisfies its specification. Such an agent can be served from a cloud API or from an inference device on the premises. For organizations whose designs, customer data or export-controlled material cannot leave the site, and whose simulators are licensed there, local deployment is the natural choice. In such a deployment, the agent becomes an electrical load on the organization's own equipment, i.e., every run consumes power and energy on a local device, generates tokens and takes time. However, existing work focuses mainly on what the agent can achieve with cloud-hosted models. What a local deployment costs in power, energy, tokens and time on real simulation tasks, and what its choices trade against one another, has received little quantitative investigation under comparable conditions. To address this gap, a systematic evaluation framework is presented in this paper. It applies locally served models to agentic power electronics tasks, from the converter level to the system level, and measures their power, energy, token and time costs. Based on these measurements, the capability boundary of a local agent on such tasks and the cost per solved task are characterized. The results are expected to provide practical insight for the use of agentic AI in power electronics and the power systems it serves.

0.28

Can STATCOM and UPS-Based Voltage Support Mitigate Fault-Induced Reductions in AI Data Center Grid Power Draw?

eess.SY ·2026-11-12 ·Ziang Liu, Yimeng Liu, Francis Yunhe Hou
hot direction similarity 0.58
Abstract · 10 pages, 5 figures

Artificial intelligence data centers (AIDCs) are rapidly increasing in number and capacity. These large electronic loads are interfaced through power electronic devices and can sharply reduce grid power intake during fault-induced voltage sags. Disturbances exceeding equipment ride-through capabilities may also trigger protective disconnection, posing risks to power system security. Static synchronous compensators (STATCOMs) and appropriately controlled AIDC uninterruptible power supply (UPS) rectifiers can provide reactive power support during voltage sags. This paper compares the individual and combined effectiveness of STATCOM and UPS-based voltage support in mitigating fault-induced reductions in grid power intake through electromagnetic transient simulations of a 1 GW AIDC connected to an IEEE nine-bus system. Simulation results show that STATCOM support better preserves short-term intake in milder tested faults, whereas UPS coordination maintains rectifier operation under deeper sags by reducing active intake to reserve reactive current capacity. Combined support improves auxiliary load retention and recovery in selected cases. However, severe tested sags still cause substantial grid-demand reductions even with combined support.

0.29

EMT Modeling and Coordinated Mode-Transition Ride-Through Control of Modern Data Centers

eess.SY ·2026-11-05 ·Junyuan Zheng, Salish Maharjan, Min Lwin, Soummya Roy, Zhaoyu Wang +3 more
similarity 0.57
Abstract

Large data centers are emerging as power-electronic loads whose response to grid voltage disturbances can affect bulk-system stability. This paper presents the composition and control characteristics of a proprietary electromagnetic transient (EMT) model of a modern data center comprising an uninterruptible power supply (UPS) with battery energy storage, induction-motor cooling loads, a computer room air handler (CRAH) driven by a variable-frequency drive (VFD), and an aggregated server load. A voltage-based state machine coordinates three aspects of the facility response: the UPS operating mode (eco, double conversion, or battery), the supply path of the CRAH, and a ramped reconnection of the grid-side UPS converter after the disturbance clears. A 100 MW facility is implemented in PSCAD/EMTDC and benchmarked against the PSSE PERC1 aggregate data center model under a symmetrical three-phase fault and low- and high-voltage ride-through (LVRT/HVRT) events. The two models agree closely in their voltage and active-power trajectories, while the EMT model also resolves the converter and motor-reacceleration transients that the phasor-domain model cannot.

0.50

Communication Delay Robust Control of BESS for AI Training Load Smoothing

eess.SY ·2026-09-18 ·Xue Lyu, Wei Du, Sheik Mohammad Mohiuddin, Brett A. Ross
hot direction similarity 0.57
Abstract · 10 pages, 17 figures

AI training loads can exhibit rapid power fluctuations because their power demand differs significantly between computational and communication phases, creating challenging ramp rates at the data center point of common coupling (PCC). Integrating battery energy storage systems (BESS) in data centers is a promising mitigation option. This paper proposes a hybrid BESS control strategy that combines droop-based grid-forming (GFM) control with instantaneous load current-based compensation to suppress high frequency load fluctuations. The GFM control loop regulates the long-term power exchange of the BESS, while the load-following control provides fast compensation for short-term AI workload fluctuations. Communication delay between the load current measurements and the BESS controller is explicitly modeled, and its impact on smoothing performance is analyzed. To mitigate delay-induced degradation, a predictor-based compensation method is incorporated into the BESS control structure. High-fidelity electromagnetic transient simulations are conducted to validate the proposed approach. Results demonstrate effective smoothing of AI training load fluctuations across different grid strength conditions and under time varying communication delays.

0.28