betaFeedback

WorksEngineering

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

Abstract

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.

Introduction

The rapid growth of artificial intelligence (AI) infrastructure and large computational loads (LCLs) is reshaping modern power system dynamics. Unlike traditional industrial facilities modeled via static or induction motor (IM) representations, AI data centers (AIDCs) present highly nonlinear, fast-varying electronic loads sensitive to grid disturbances [1]. The Electric Reliability Council of Texas (ERCOT) defines LCLs as facilities with at least 75 MW of peak demand, 50% or more of which comprises power-electronic computational equipment [2], a classification encompassing most AIDCs. Accurately capturing AIDC–grid interactions is therefore critical for reliable network planning and operation.

With single-site computational demand exceeding 1,037 MW at facilities such as New Carlisle [3], abrupt AIDC disconnections can trigger severe voltage and frequency excursions. During a net 3800 MW load transfer event on Virginia’s grid, an initial 2970 MW data center disconnection induced voltage and frequency spikes that precipitated a further 1099 MW drop (see Figure 1) [4]. Consequently, grid operators are now mandating or proposing strict voltage and frequency ride-through (VRT/FRT) standards for LCLs [2, 4].

Grid frequency during a load transfer event

Figure 1. Grid frequency during a load transfer event [4].

Evaluating ride-through compliance and converter-driven transients requires electromagnetic transient (EMT) simulations, as phasor-domain root mean square (RMS) models neglect sub-cycle switching dynamics and high-frequency nonlinearities, [5, 6]. Transmission operators now require EMT studies for LCL interconnection requests [7]. Despite this need, only four open-source PSCAD EMT AIDC models are publicly available as of writing [8, 9, 10, 11]. To support model benchmarking and development, this paper introduces a 41-point, software-agnostic EMT AIDC fidelity framework and applies it to these four open-source models to identify their strengths and priority areas for extension.

AI Data Center Topology

Figure 2 illustrates an AIDC comprising four primary subsystems: information technology (IT) loads (CPUs, GPUs, and networking hardware), cooling systems, an uninterruptible power supply (UPS) with battery energy storage (BESS), and auxiliary facility loads (lighting and site management). Below, we outline each subsystem’s characteristics and explain why EMT modeling is required to capture its grid interactions.

AIDC design with a centralized UPS

Figure 2. AIDC design with a centralized UPS.

Information Technology Loads

IT equipment is the dominant power consumer in an AIDC and exhibits highly volatile demand across two primary workloads: AI model training and inference [12]. Training maintains high average utilization punctuated by abrupt, facility-wide power swings during periodic checkpointing, whereas inference fluctuates with real-time user requests of varying length and complexity. Because GPU power supply units (PSUs) rely on high-frequency switch-mode converters [13], EMT modeling is necessary to resolve sub-cycle load ramps, converter current harmonics, and terminal voltage/frequency sensitivity that phasor-domain RMS models smooth out.

Uninterruptible Power Supply and Storage

Coupled with a BESS, the UPS isolates critical IT loads from grid disturbances across three operating modes [14]: double-conversion mode, where grid AC power is rectified and re-inverted into tightly regulated AC voltage; bypass mode, where IT loads connect directly to the grid during maintenance or high-efficiency operation; and battery backup mode, where the BESS supplies the IT loads during severe grid outages. EMT simulation is essential to capture transfer switch transitions, DC-link capacitor voltage transients, converter switching dynamics, and VRT/FRT control responses during faults.

IT Load Power Electronics Configuration

AIDCs primarily employ three power electronics architectures: centralized UPS, distributed rack-level storage, and emerging DC distribution systems [13]. Centralized topologies are most common and isolate multi-rack IT loads, but their sub-cycle transfer logic and converter switching require EMT simulation to resolve high-speed mode transitions. In contrast, distributed designs eliminate double-conversion stages to reduce costs, directly exposing IT power supplies to transient voltage events. Emerging direct-current distribution eliminates output inverters to enhance power density [13], introducing fast DC dynamics, resonance, and control interactions that RMS models cannot capture. Capturing these fast switching dynamics, transfer latencies, and disturbance vulnerabilities makes EMT modeling indispensable.

Cooling Systems

Because rack power densities in AIDCs exceed 100 kW [13], conventional fan-based air cooling is paired with chiller-based liquid cooling, such as direct-to-chip cold plates or liquid immersion, driven by variable frequency drives (VFDs) to limit motor inrush currents. From an EMT perspective, VFD front-end rectifiers inject non-sinusoidal harmonics, while IM fans and chillers exhibit distinct slip, stalling, and post-fault re-acceleration dynamics during voltage sags. Furthermore, coupling electrical cooling demand with rack thermal inertia is critical for evaluating sustained disturbance recovery.

Grid Interface, Onsite Generation, and Site Controls

Step-down transformers interface the AIDC with the utility grid at the point of interconnection (POI) and are safeguarded by protective relays. Facilities also integrate medium-voltage (MV) utility-scale gas turbines to offset grid demand alongside low-voltage (LV) backup diesel generators [10]. Finally, a supervisory site control system coordinates workload throttling, cooling curtailment, and generator dispatch. EMT modeling of these elements is required to capture transformer core saturation and inrush currents, protective relay trip coordination, generator synchronization transients, and control communication latencies during severe grid events. Together, these dynamics define the baseline architectural features examined across the open-source PSCAD models in Section III and the 11 evaluation categories established in § 4.

Overview of Open-Source EMT AIDC Models

Power System Computer Aided Design (PSCAD/EMTDC) is a graphical EMT simulation platform developed by Manitoba Hydro International that solves time-domain differential equations via its EMTDC engine [15]. As of October 25, 2026, a survey of publicly available repositories identified four open-source PSCAD EMT AIDC models (Table I), referred to in our paper as Models 1–4 [8], [9], [10], [11].

Model #Brief DescriptionInstitutionLast UpdatedTest Cases
1Dynamic Modeling of AI Data Center Load in PSCAD [8]ERCOT/Texas A&M05/20263
2Electromagnetic Transient Modeling of Data Centers [9]NLR08/20264
3Electromagnetic Transient Modeling of Large Data Centers for Grid-Level Studies [10]PNNL08/202610
4Generic EMT Modeling of Data Center Load [11]Dartmouth College10/20262

Table I. Open Source EMT AIDC Models in PSCAD

Model 1 from ERCOT and Texas A&M University uses a modular, bottom-up architecture with averaged power electronic converter models to represent a double-conversion UPS facility, VFD-driven motor cooling, and aggregated AC/DC computing loads. Its three test cases demonstrate UPS operating mode transitions, voltage ride-through, and time-varying compute load fluctuations. Model 2 from the National Laboratory of the Rockies (NLR) represents a data center built from scaled 33 kW IT rack power shelves (six 5.5 kW power supply units), a multi-mode AC UPS (double-conversion, eco, and single-conversion), cooling loads, and a 260 MW onsite gas turbine generator across four grid-connected and off-grid use cases for converter-driven stability and power quality studies. Model 3 from the Pacific Northwest National Laboratory (PNNL) provides the Data Center Model Library, comprising 10 test cases across centralized UPS, distributed rack-level UPS, and mixed distribution topologies. Designed for bulk grid-level interconnection studies, it scales up to 1.54 GW total active power (1,050 MW IT load), matching the largest operational AIDCs, and separates cooling loads into aggregated 0.2 MW chiller and 0.6 MW fan units. Model 4 from Dartmouth College implements a modular centralized-UPS data center model offering average, switching, and reduced-order converter representations alongside motor-driven cooling and synthetic workload profiles. Its two test cases evaluate full-facility grid coupling and trade-offs across converter modeling fidelity levels. To evaluate component parameters on a consistent architectural basis, we selected one centralized-UPS test case per model: TestCase_ITLoadFluctuation_Ave (Model 1), UseCase3_DataCenter (Model 2), DML_DC2_Central_UPS (Model 3), and DC_EMT_demo (Model 4). Table II summarizes the base and current-scaled power ratings of each test case, all of which exceed ERCOT’s 75 MW LCL threshold. Because the models represent different facility ratings and test scenarios, we audit their component-level representations rather than comparing raw simulation waveforms directly.

Model #LoadBase RatingScaled Power
1IT1.5 MW150 MW
1Cooling0.05 MW5 MW
1AuxiliaryN/AN/A
1Total Active Power155 MW
2IT1.65 MW132 MW
2Cooling0.5 MW40 MW
2Auxiliary0.1 MW8 MW
2Total Active Power180 MW
3IT1 MW1050 MW
3Cooling0.2 & 0.6 MW200 & 200 MW
3Auxiliary100 MVA100 MVA (0.9 lag)
3Total Active Power1540 MW
4IT0.15 MW60 MW
4Cooling0.5 MW20 MW
4Auxiliary11.18 MVA11.18 MVA (0.89 lag)
4Total Active Power90 MW

Table II. Power Ratings of Selected Test Cases

Proposed Evaluation Framework

Table III presents our software-agnostic, 41-point evaluation framework across 11 categories for auditing EMT AIDC models. We apply this framework to evaluate the component-level completeness of the four open-source PSCAD models. Categories A-J (37 points) are synthesized from [1], which define required component representations across ten modeling layers alongside recommended PSCAD implementations.

Cat.IDSpecification
Utility & MV BusesA.1Dynamic voltage-/frequency-sensitive POI with impedance
Utility & MV BusesA.2MV protection relays & utility coordination schemes
Utility & MV BusesA.3Simulation of voltage sags/swells, frequency excursions, & AI-load-driven POI oscillations
TransformersB.1Inrush current & core saturation phenomena
TransformersB.2On-load tap changer (OLTC) control behavior
TransformersB.3Protective relaying schemes for overcurrent, differential protection, & under/overvoltage trips
TransformersB.4Interaction with fast-varying DC loads & grid disturbances
ITC.1Time-varying load transitions & rapid ramp capabilities
ITC.2Load sensitivity to voltage & frequency variations
UPSD.1Dual-mode operation (grid-following & grid-forming)
UPSD.2Transition logic for choosing utility power & backup
UPSD.3Compliance with VRT requirements
Cooling LoadsE.1Voltage & frequency sensitivity of motor-driven components
Cooling LoadsE.2Staged startup, load shedding, & curtailment control
Cooling LoadsE.3Thermal inertia effects influencing disturbance recovery times
Cooling LoadsE.4Coordination with facility management systems
BESSF.1Fast transient response & dynamic control during grid events
BESSF.2Compliance with IEEE Std 1547 [16]
BESSF.3State-of-charge management & site controller interactions
Generator ModelsG.1Startup delays & ramp rates
Generator ModelsG.2Load acceptance dynamics & backup power connection times
Generator ModelsG.3V, I, & f protective relays
Generator ModelsG.4Synchronization & reconnection mechanisms
Generator ModelsG.5Automatic generation control
Generator ModelsG.6Automatic voltage regulation
Generator ModelsG.7Isochronous control, droop control, & load-sharing schemes
Protection SystemsH.1Protective relay logic with thresholds for V, I, & f
Protection SystemsH.2Single- & three-pole circuit breaker models with trip time delays
Protection SystemsH.3Relay coordination & cascading trip sequences
Protection SystemsH.4Instrument transformer saturation effects & harmonic filtering
Higher-Level ControlI.1Custom control sequences for breaker operation, load/DER dispatch, & demand response
Higher-Level ControlI.2Real-time communication & monitoring interfaces
Higher-Level ControlI.3Fault detection, system isolation, & black-start control strategies
Transfer SwitchesJ.1Fast transfers between independent utility & backup sources
Transfer SwitchesJ.2Sensing- & logic-driven transfer initiation
Transfer SwitchesJ.3Realistic transfer time, transient response, & interruption characteristics
Transfer SwitchesJ.4Voltage synchronization, inrush control, & load acceptance
OursK.1Compliance with FRT requirements
OursK.2Separate modeling of fan- & chiller-based cooling loads
OursK.3Temperature-dependent cooling power response
OursK.4DC-side IT load representation

Table III. EMT AIDC Model Evaluation Framework

To address emerging trends identified in Sections I and § 2, we introduce four additional requirements in Category K: (K.1) FRT Compliance—Incorporating frequency ride-through (FRT) modules into UPS controls to reflect evolving LCL interconnection mandates; (K.2) Disaggregated Cooling Loads—Separating fan- and chiller-based cooling subsystems to capture their distinct electrical and operational characteristics; (K.3) Temperature-Dependent Cooling—Coupling cooling power consumption to ambient or rack operating temperatures; and (K.4) DC IT Load Representation—Modeling IT equipment as DC loads rather than aggregated AC equivalents that abstract power electronic converter stages.

Model Evaluation Results

Table IV summarizes the evaluation of the four open-source EMT models against our 41-point framework, with total scores ranging from 13 to 21. Subsections V-A through V-K correspond to Categories A-K in Table III, while Sub-section V-L synthesizes facility-wide disturbance ride-through capabilities. Across the suite, the models provide strong EMT representations of IT load profiles, UPS, and transfer switches. Because these releases were scoped for bulk grid-level stability studies, thermodynamic cooling feedback, internal protection relaying, and site-level communications remain outside their current scope, highlighting clear priorities for future model extensions. Models can also benefit from detailed disturbance ride-through control implementation for disturbance studies.

CategoryA (/3)B (/4)C (/2)D (/3)E (/4)F (/3)G (/7)H (/4)I (/3)J (/4)K (/4)Sum (/41)
Model 11213120002214
Model 21222127002120
Model 31213117003121
Model 41223110002113

Table IV. Evaluation Scores Across Models

Utility Interface and Medium-Voltage Bus

All four models establish POIs with configurable grid impedances and short-circuit ratios. Models 1-3 include standard North American HV bus voltages, whereas Model 4 starts at the MV level; all models include standard MV and 480 V LV buses. Model 3 represents the grid via a scaled synchronous generator, step-up transformer, and series impedances, while Models 1, 2, and 4 use non-ideal three-phase voltage sources. Adding MV protective relays, utility coordination schemes, and frequency excursion test cases would enable broader compliance studies.

Transformers

All four models use PSCAD three-phase, two-winding transformer blocks equipped with inrush and core saturation representations (B.1) that capture EMT interactions (B.4). Although saturation is toggled off in the default test cases, likely due to limited nameplate data, it can be readily enabled once parameters are specified.

Information Technology Loads

Rapid IT load transitions are modeled via rack-level variable resistances driven by empirical cluster telemetry in Model 1 (MIT Supercloud Dataset v1 [17]), synthetic AI profiles in Models 2 and 3 (ON.energy and EdgeTunePower [18]), and square, sinusoidal, and ramp waveforms in Model 4. Models 2 and 4 retain power electronic converter interfaces that capture voltage and frequency sensitivity, whereas Models 1 and 3 abstract the converter stage. All models use current scalers to scale single-rack loads to multi-megawatt ratings; future studies can verify aggregation linearity during fast transients.

Uninterruptible Power Supply

All four models provide detailed dual-mode UPS representations with automated utility-to-backup transition logic. Models 1, 3, and 4 support programmable VRT curves, and Model 1 additionally incorporates configurable FRT settings.

Cooling Loads

Cooling systems are represented using voltage- and frequency-sensitive IMs. Models 1 and 3 used electrical parameters identical to PSCAD IM library defaults without rescaling for base voltage and power, while Models 2 and 4 aggregate cooling into 0.5 MW IM equivalents. Because cooling demand is currently decoupled from IT utilization, integrating VFD curtailment, thermal inertia, and facility coordination will improve post-disturbance recovery fidelity.

Battery Energy Storage System

All models capture fast BESS transient response and dynamic power control. Models 1 and 2 track SOC dynamically, whereas Models 3 and 4 hold SOC constant, a practical simplification over short EMT windows. Additional communication modeling will allow models to meet IEEE Std 1547 [16].

Generator Models

Models 2 and 3 incorporate onsite generation—a 260 MW gas turbine (300 MVA, 13.2 kV synchronous machine) in Model 2 and a reciprocating engine with an autosynchronizer in Model 3—whereas Models 1 and 4 represent grid-supplied facilities. As noted by Model 3’s developers [10], more sophisticated generation models are needed for plant-level studies at the design stage. This could include adding governor droop, automatic generation control, and load-sharing logic.

Protection Systems

Current protection modeling focuses primarily on the UPS interface. Model 3 extends coverage to the wider facility using configurable single- and three-phase breakers with undervoltage and voltage-unbalance relays. Building on Model 3 by adding overcurrent and frequency relays, time-graded coordination, and instrument transformer saturation will support internal fault and cascading trip analysis.

Higher-Level Control Systems

Site-level supervisory controls remain an area for development across all four models. Extending existing utility-fault test cases with coordinated load dispatch, demand response, communication latencies, and internal fault isolation will enable full facility-level disturbance studies.

Automatic/Static Transfer Switches

All models link transfer switches to UPS sensing and transition logic. Model 3 further incorporates voltage synchronization, inrush control, and staged load acceptance for onsite generation. Adding realistic mechanical/static transfer delays and arc-interruption dynamics across all models would refine sub-cycle transfer accuracy.

Additional Evaluation Points

Across the four proposed criteria in Category K, Model 1 incorporates UPS FRT settings (K.1), Model 3 disaggregates fan and chiller cooling loads with separate scaling factors (K.2), and Models 1, 2, and 4 include DC-side IT load representations (K.4). None of the models currently couple temperature dynamics to cooling power consumption (K.3), leaving electro-thermal feedback as an important next step.

Disturbance Ride-Through Control

While Models 1, 3, and 4 capture local UPS VRT behavior (D.3) and Model 1 includes FRT compliance (K.1), the absence of site-level supervisory control (Category I) prevents end-to-end simulation of coordinated IT throttling, cooling curtailment, generation dispatch, and UPS transfers. Integrating facility-wide ride-through controls is therefore essential for studying and mitigating large-scale load transfer events such as those described in Section I.

Conclusion

This paper introduced a software-agnostic, 41-point fidelity framework spanning 11 categories to audit the component-level completeness of EMT AIDC models, integrating emerging requirements for frequency ride-through, disaggregated and temperature-dependent cooling, and DC IT load representations. Applying this framework to the four publicly available PSCAD/EMTDC AIDC models shows that current releases provide strong foundational representations of IT load fluctuations, UPS dynamics, and transfer switches for bulk grid-level stability studies. Extending these open-source libraries with electro-thermal cooling feedback, internal protective relaying, and supervisory site-level controls, alongside standardized cross-model output and runtime benchmarking, will enable comprehensive protection coordination and facility-wide ride-through compliance studies.

References

  1. [1]S. Talukdar, S. Marti, K. Prabakar, and D. Vaidhynathan, “Modeling Framework for Data Center,” National Laboratory of the Rockies (NLR), Golden, CO (United States), Feb. 13, 2026, NLR/TP-2C00-97716. DOI: 10.2172/3019724 Accessed: Nov. 5, 2026. [Online]. Available: https://www.osti.gov/biblio/3019724DOI
  2. [2]ERCOT. “NOGRR282 - Large Computational Load Ride-Through Requirements,” Board Priority - Large Computational Load Ride-Through Requirements, Accessed: Nov. 5, 2026. [Online]. Available: https://www.ercot.com/mktrules/issues/NOGRR282#summary
  3. [3]Epoch AI, AI data centers, Nov. 2026. [Online]. Available: https://epoch.ai/data/ai-data-centers
  4. [4]PJM, Anchor, July 22, 2026 dominion load transfer event, Presentation to the PJM System Operations Committee, PJM Interconnection, Jul. 31, 2026. [Online]. Available: https://www.pjm.com/ - /media/ DotCom/committees - groups/subcommittees/sos/2026/20260730/20260730 - item - 06 --- july - 22 - dominion - load - transfer - event --- presentation.pdf
  5. [5]A. Shahriar Pramanik, S. Sepasi, B. Griswold, and A. Adegboyega, “Assessing Dynamic Behaviors in Converter-Dominated Power Systems via RMS and EMT Simulations: A Study of Hawaii’s NELHA Microgrid,” IEEE Access, vol. 13, pp. 83 881–83 896, 2025, ISSN: 2169-3536. DOI: 10.1109/ACCESS.2025.3568782 Accessed: Nov. 4, 2026. [Online]. Available: https://ieeexplore.ieee.org/document/11000107/DOI
  6. [6]S. Favuzza, R. Musca, G. Zizzo, and J. A. Sa’ed, “Comparative modeling and analysis of emt and phasor rms grid-forming converters under different power system dynamics,” IEEE Transactions on Industry Applications, vol. 60, no. 2, pp. 3613–3624, 2024. DOI: 10.1109/TIA.2023.3339108DOI
  7. [7]IESO, “Technical requirements for large computational loads connecting to the ontario power system,” Independent Electricity System Operator (IESO), Toronto, ON, Canada. Technical Requirements, May 1, 2026, Version 1.0. [Online]. Available: https://ieso.ca/-/media/Files/IESO/Document-Library/engage/large-computational-loads/LargeComputationalLoads-20260501-Technical-Requirements.pdf
  8. [8]X. Wang et al., “Dynamic Modeling of AI Data Center Load in PSCAD,” TAMU-ERCOT, Apr. 2026, p. 105. [Online]. Available: https://www.ercot.com/files/docs/2026/04/24/Dynamic-Modeling-of-AI-Data-Center-Load-in-PSCAD-Manual.pdf
  9. [9]S. Shah and W. Yan, “Electromagnetic Transient Modeling of Data Centers,” National Laboratory of the Rockies, Golden, CO., Technical Report, Aug. 2026, NLR/TP-5D00-101189, p. 49. Accessed: Nov. 4, 2026. [Online]. Available: https://www.nlr.gov/docs/fy26osti/101189.pdfDOI
  10. [10]B. A. Ross, S. Nekkalapu, X. Lyu, and J. D. Follum, “Electromagnetic Transient Modeling of Large Data Centers for Grid-Level Studies: Beta Release,” Pacific Northwest National Laboratory (PNNL), Richland, WA (United States), Aug. 14, 2026, PNNL-38817 Beta. DOI: 10.2172/3422378 Accessed: Nov. 4, 2026. [Online]. Available: https://www.osti.gov/biblio/3422378
  11. [11]X. Zhao, Y. Tang, and J. Zhao, Generic EMT Modeling of Data Center Load v0 (DC_EMTv0), Dartmouth College, 2026. [Online]. Available: https://github.com/xingyuz-phd/Generic-EMT-Modeling-of-Data-Center-Load
  12. [12]M.-S. Ko and H. Zhu, “Wide-area power system oscillations from large-scale AI workloads,” IEEE Transactions on Power Systems, vol. 41, no. 5, pp. 3908–3921, 2026. DOI: 10.1109/TPWRS.2026.3685506DOI
  13. [13]P. P. Gyang, P. Chakraborty, L. Meegahapola, and X. Yu, “Grid-integrated data centers: Architectures, modeling, interconnection requirements, challenges, and opportunities,” IEEE Open Journal of the Industrial Electronics Society, vol. 7, pp. 1080–1112, 2026. DOI: 10.1109/OJIES.2026.3706771DOI
  14. [14]X. Chen, X. Wang, A. Colacelli, M. Lee, and L. Xie, “Electricity demand and grid impacts of AI data centers: Challenges and prospects,” Nexus, vol. 3, no. 3, p. 100 162, Sep. 15, 2026, ISSN: 2950-1601. DOI: 10.1016/j.ynexs.2026.100162 Accessed: Nov. 4, 2026. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S2950160126000458DOI
  15. [15]Manitoba Hydro International Ltd., PSCAD/EMTDC: Electromagnetic transients program including DC systems, manual, Winnipeg, MB, Canada: Manitoba Hydro International Ltd. [Online]. Available: https://www.pscad.com
  16. [16]“IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces,” IEEE Std 1547-2018 (Revision of IEEE Std 1547-2003), pp. 1–138, 2018. DOI: 10.1109/IEEESTD.2018.8332112DOI
  17. [17]S. Sami et al., “The MIT supercloud dataset,” in 2021 IEEE High Performance Extreme Computing Conference (HPEC), 2021, pp. 1–8. DOI: 10.1109/HPEC49654.2021.9622850
  18. [18]EdgeTunePower, ETP AI Load Profile Generator. [Online]. Available: https://www.edgetunepower.com/copy-of - data - center - and - ibr - assessment

Paper details

vlnc:W_x965tsjv
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Engineering

Not recorded for this paper

View statement
Looking for related papers…
Contents