This technical paper from Dominion Energy examines how scenario-based system-load forecasts at different time granularities can support long-term assessment of planned transmission outages. It is set against rapid growth in hyperscale data-center and AI load in Northern Virginia, which requires new and reinforced grid infrastructure. The authors compare the conventional use of monthly or seasonal peak demand with coordinated daily, weekly, and monthly forecasts, finding that the latter can enable more outage coordination without weakening the N-1 reliability criterion.
Key Takeaways
Growth in AI and hyperscale data-center load in Northern Virginia makes timely reinforcement of the transmission grid critical, while the construction itself requires coordinated outages of existing grid assets.
Assessing every day against the monthly peak load is systematically overconservative: it does not capture intra-month variability and particularly constrains short-duration outages.
Coordinated daily forecasts generally deliver the largest and most consistent improvement in outage-day approvals and full outage approvals; weekly forecasts also capture a significant share of the benefit.
The improvement is achieved without changing N-1 requirements: each scenario is tested for AC power-flow convergence, thermal limits, and acceptable voltages before and after a contingency event.
Integrating daily and weekly forecasts into planning processes can reduce avoidable work rescheduling and accelerate grid expansion and reinforcement projects needed to connect large loads.
Key Figures
Large-load interconnection requests
70 GW
Approximate volume of requests at Dominion Energy.
Loads with assigned interconnection dates through 2031
25 GW
Approximate volume of requests that have already received projected interconnection dates.
Data-center load served by Dominion Energy
5 GW
Approximate current data-center load.
Share of Dominion Energy system demand attributable to data centers
15–20 %
Share of system demand stated in the paper.
Growth in large loads relative to the historical system peak
почти 3 раза
Scale of the projected growth in large loads.
Reliability criterion for outage approval
N-1
Post-contingency thermal constraints, voltages, and AC power-flow convergence are assessed.
Strategic outage-planning horizon
1–5 лет
Assessment period for major outages associated with maintenance, construction, and grid reinforcement.
Long-term outage-planning horizon
45 дней – 1 год
Period preceding outage execution.
ANODE automated study horizon
10 дней
ANODE supports automated outage-reliability studies over this horizon.
Historical SCADA load-recording interval
примерно 4 секунды
Frequency at which historical system-load data are recorded.
Study period
декабрь 2026 – май 2027
Six-month evaluation of three forecast granularities.
Number of study cases
18
Six months assessed across three load granularities.
LightGBM MAPE improvement versus the baseline model
42.4 %
Result of the forecasting-accuracy comparison.
January ODAR: monthly, weekly, and daily
69.65 / 82.14 / 90.56 %
Average Outage-Day Approval Rate across the three load representations.
February ODAR: monthly, weekly, and daily
70.22 / 85.12 / 91.28 %
Average Outage-Day Approval Rate; the daily representation improved on the monthly representation by the stated percentage-point amount.
January full outage approval: monthly and daily
21.65 / 64.95 %
Outage Approval Rate for monthly and daily load representations.
May short-outage approval: monthly, weekly, and daily
55.26 / 76.32 / 81.58 %
Outage Approval Rate for outages lasting 1–10 days.
Practical Value for Data Center Owners
For data-center owners and teams responsible for site interconnection or expansion, the paper highlights the risk that conservative utility outage planning can delay grid reinforcements. When discussing an interconnection schedule and grid works, it is useful to request scenario-based load assessments at daily or weekly granularity rather than relying only on the monthly peak: Dominion’s example showed that this can materially improve the ability to schedule short outages while maintaining N-1. For grid operators, the work also provides a reference architecture for integrating ML-based load forecasting into outage planning and contingency analysis.
A special World Energy Outlook report by the International Energy Agency examining the relationship between AI development and energy. It analyses growth in data-centre electricity consumption and capital expenditure, the physical constraints on scaling AI infrastructure, power-supply options, and implications for energy systems, prices and emissions. It also considers AI’s impact on energy and industrial efficiency, as well as demand scenarios through 2030.
AI data centersData center power supplyGrid interconnectionLiquid cooling
This AVK and Wärtsilä Energy white paper examines constrained grid capacity as a major barrier to the expansion of European data centres, particularly facilities serving AI workloads. It proposes on-site microgrids combining dispatchable generation, BESS and renewables to accelerate site energisation before—or independently of—a grid connection. The paper compares power-system scenarios and four supply configurations for a representative AI data centre in terms of cost, reliability, land use and carbon footprint.
Data Centre Power SupplyAI Data CentresMicrogridsDispatchable Generation
A Deloitte analysis of the growth in data center infrastructure for AI workloads and its implications for the U.S. power system. It uses DC Byte and Wood Mackenzie data, estimates of AI data center penetration, and projections through 2035 to assess capacity. The focus is on interconnection and generation shortfalls, the impact on utility capital expenditures, and potential sources of and approaches to accelerating power supply for large AI sites.
AI data centersdata center power supplygrid interconnectionbehind-the-meter generation