Grid Utilization
Draft for public commentPublished 2026-11-11
This methodology is a draft in its public comment period. Send comments to the Alliance’s Google Group.
Draft version 0.1, revised 2026-11-08.
Foreword
This document has been prepared by the OpenEAC Alliance, a collaborative group of energy engineering and measurement and verification professionals focused on creating transparent and standardized methods to measure the environmental and grid impacts of demand side interventions. These methodologies are designed to align with best practices and reflect input from industry experts. The OpenEAC Alliance aims to enhance accountability and consistency in documenting energy, carbon, and grid utilization outcomes from distributed energy resources (DERs).
About the OpenEAC Alliance
The OpenEAC Alliance is an organization of professionals dedicated to standardizing the measurement and verification of DERs. By focusing on clarity and rigor, the Alliance aims to overcome challenges that have hindered the proper valuation of DERs for their climate benefits. Central to this effort is the Energy Attribute Certificate (EAC), a standard that provides a detailed, auditable record of energy project impacts, including attributes such as type, timing, and location of energy production or reduction.
The OpenEAC Alliance collaborates to develop open-source methodologies, ensuring that DER projects can be accurately measured, verified, and valued in a way that promotes trust and drives market adoption.
1 Introduction
1.1 An electric grid is built to serve its peak load and runs well below that peak for most hours of the year. The share of the grid’s capacity that is used over a year is its utilization. When a grid delivers more energy without a higher peak, its fixed costs are spread over more kilowatt-hours, which can put downward pressure on rates (see the Bibliography).
1.2 This document defines a reference grid signal, in the way a marginal emissions rate is a reference signal in carbon accounting. A Publisher computes the signal once for each grid and year, and a resource methodology that values a DER’s timing cites the published figure rather than computing its own.
1.3 The definition of utilization follows Beyond Peak: Defining, Measuring, and Optimizing Grid Utilization (Utilize Coalition, September 2026): energy delivered divided by energy deliverable, reported as a nominal and an operational figure together. This document applies that definition to public data for United States utilities.
1.4 This document also defines two companion figures, Load Factor and Peak Loading, which the same paper names as distinct measures, and an Hourly Signal, which is an addition to the paper’s framework. The Hourly Signal shows in which hours of the year a grid has the most and the least headroom.
2 Scope
2.1 This document outlines the definitions, data sources, screening rules, and labels used to publish, for a calendar year, a grid’s Utilization, Peak Loading, Load Factor, and Hourly Signal.
2.2 This methodology supports:
2.2.1 Utilization and Peak Loading, nominal and operational, for a utility that reports its distribution substations in the Substation Schedule. In practice these are investor-owned utilities.
2.2.2 Load Factor for a utility, from the utility’s own hourly load curve where one is published, and otherwise from its annual EIA-861 report, with its energy taken from FERC Form No. 1 where EIA-861 reports the utility’s peak but no energy.
2.2.3 Load Factor and the Hourly Signal for a balancing authority, from EIA-930 hourly demand.
2.2.4 The Hourly Signal for a utility, from its own hourly load curve or, labeled as a Proxy, from the hourly load of its balancing authority.
2.3 This methodology does not currently support:
2.3.1 Utilization or Peak Loading for a balancing authority, because no distribution capacity is published for one.
2.3.2 Utilization or Peak Loading for a utility that reports no distribution capacity in the Substation Schedule, such as most municipal utilities and cooperatives.
2.3.3 Utilization of a single feeder, substation, or transmission path.
2.3.4 A composite figure for an area served by more than one utility or balancing authority.
2.3.5 The effect of a single DER on any of these figures. A resource methodology that applies the Hourly Signal to a DER’s hourly change in load is a separate document.
2.4 Where a grid falls outside the supported scope for a figure, the Publisher reports that figure as unavailable, with a note that states the reason. The Publisher does not substitute a figure from another grid, except for the Hourly Signal under 6.5.3, where the substitution is labeled.
3 Normative References
3.1 The following documents are essential for the application of this specification. For references with no date, the latest edition applies:
3.1.1 Utilize Coalition, Beyond Peak: Defining, Measuring, and Optimizing Grid Utilization, September 2026.
3.1.2 U.S. Energy Information Administration, Form EIA-930, Hourly and Daily Balancing Authority Operations Report: https://www.eia.gov/electricity/gridmonitor/
3.1.3 U.S. Energy Information Administration, Form EIA-861, Annual Electric Power Industry Report: https://www.eia.gov/electricity/data/eia861/
3.1.4 Federal Energy Regulatory Commission, Form No. 1, Annual Report of Major Electric Utilities, Substations schedule (pages 426–427): https://www.ferc.gov/general-information-0/electric-industry-forms/form-1-1-f-3-q-electric-historical-vintage
3.1.5 Federal Energy Regulatory Commission, Form No. 1, Annual Report of Major Electric Utilities, Electric Energy Account schedule (page 401a): https://www.ferc.gov/general-information-0/electric-industry-forms/form-1-1-f-3-q-electric-historical-vintage
4 Terms and Definitions
4.1 Publisher: The party that computes and publishes Grid Utilization figures under this methodology.
4.2 Grid: A utility, identified by its EIA utility id, or a balancing authority, identified by its EIA-930 code.
4.3 Year: A calendar year in the Grid’s Local Standard Time.
4.4 Local Standard Time: The Grid’s time zone at its standard offset from UTC, with no daylight saving time. Every Year then has exactly 8,760 hours, or 8,784 in a leap year, and no hour repeats.
4.5 Energy Delivered: The energy the Grid delivered in the Year, in MWh.
4.6 Energy Deliverable: The energy the Grid’s distribution capacity can deliver in the Year, in MWh: the capacity, times the Power Factor, times the hours in the Year.
4.7 Power Factor: The ratio used to convert a rating in MVA to a deliverable power in MW. This methodology fixes it at 0.95.
4.8 Substation Schedule: The Substations schedule of FERC Form No. 1 (pages 426–427), in which a major electric utility reports each substation’s capacity, voltages, number of transformers, and whether it serves transmission or distribution.
4.9 Nominal Capacity: The sum of the rated capacity, in MVA, of the Grid’s kept distribution substation rows.
4.10 Operational Capacity: The Nominal Capacity after a contingency reserve at each substation, under 6.1.4.
4.11 Utilization: Energy Delivered divided by Energy Deliverable. Nominal Utilization uses the Nominal Capacity, and Operational Utilization uses the Operational Capacity.
4.12 Peak Loading: The Grid’s annual peak load, in MW, divided by the capacity times the Power Factor. It can exceed 100%.
4.13 Load Factor: The mean hourly load in the Year divided by the highest hourly load in the Year.
4.14 Load Curve: An hourly series of a Grid’s load, in MW.
4.15 Published Curve: A Load Curve that covers the Grid itself: a balancing authority’s own EIA-930 demand, or a curve that the Publisher’s curve map assigns to a utility with the label published.
4.16 Proxy: A Load Curve of a wider area than the Grid, used only for the Hourly Signal and always labeled.
4.17 Glitch Hour: An hour that the screen in 6.4 removes.
4.18 Coverage: The kept hours of a Load Curve divided by the hours in the Year, or in the elapsed part of a Year in progress.
4.19 Relative Load: The position of an hour’s load between the Year’s lowest and highest kept load, from −1 to +1, with 0 at the mean load, computed under 6.5.1.
4.20 Hourly Signal: The Relative Load of every kept hour of a Grid’s Year.
4.21 Final: A figure whose inputs the Publisher no longer expects to change. A figure that is not Final is Provisional.
4.22 Energy Account: The Electric Energy Account schedule of FERC Form No. 1 (page 401a), in which a major electric utility reports the sources and disposition of its energy for the year, including its sales to ultimate consumers and its energy losses.
4.23 Respondent: A company that files FERC Form No. 1, identified by its FERC respondent id.
4.24 Respondent Map: The Publisher’s assignment of each Respondent, for each report year, to the EIA utility id whose distribution system its filing describes.
5 Data Requirements
5.1 To publish Utilization and Peak Loading for a utility, the Publisher shall use:
5.1.1 The utility’s EIA-861 retail sales, delivery-only sales, losses, summer peak, and winter peak, from the latest report year not later than the Year that has retail sales and a seasonal peak greater than zero.
5.1.2 The rows of the utility’s Substation Schedule, from the latest report year not later than the Year whose rows kept under 6.2 give a Nominal Capacity and an Operational Capacity greater than zero. The Publisher shall read, for each row, its name, its label (transmission, distribution, or other), its capacity in MVA, its high-side and low-side voltages in kV, and its number of transformers in service. The label does not decide whether a row is counted; 6.2 decides that from the voltages.
5.1.3 For a report year in which the utility’s EIA-861 report gives a seasonal peak greater than zero but gives neither retail sales nor total disposition greater than zero, the sales to ultimate consumers and the losses in the utility’s Energy Account for the same report year, in place of the EIA-861 sales and losses. The Publisher shall take the seasonal peaks from EIA-861. A report year with Energy Account sales to ultimate consumers greater than zero is then usable under 5.1.1.
5.2 Where a utility has no usable EIA-861 report, or no usable Substation Schedule, the Publisher shall report Utilization and Peak Loading as unavailable, with a note that names the missing source.
5.3 Where the EIA-861 delivery-only sales are missing, the Publisher shall count them as zero and shall state that in a note. Where the losses are missing, the Publisher shall leave them out and shall state that in a note.
5.4 To publish Load Factor and the Hourly Signal, the Publisher shall use one of the following Load Curves, chosen under 6.3:
5.4.1 For a balancing authority, its EIA-930 hourly demand.
5.4.2 For a utility, the Load Curve that the Publisher’s curve map assigns to it, or otherwise the EIA-930 hourly demand of its balancing authority, as a Proxy, under 6.5.3.
5.5 The Publisher shall publish its curve map, which assigns a Load Curve and a label to a utility, and shall state the map’s revision with every figure that uses it.
5.6 Where a Load Curve is built from several areas, the Publisher shall keep an hour only when every area reported a load for that hour.
5.7 The Publisher shall not use the Energy Account for a report year in which the utility’s EIA-861 report gives retail sales or total disposition greater than zero. A utility that sells at wholesale reports total disposition, and its Energy Account sales to ultimate consumers do not describe a distribution system.
5.8 Where the Energy Account reports no losses, the Publisher shall leave them out of Energy Delivered and shall state that in a note.
5.9 The Publisher shall take a utility’s Substation Schedule and Energy Account from the Respondents that the Respondent Map assigns to the utility’s EIA utility id for the report year.
5.10 The Publisher shall assign a Respondent to no EIA utility id for a report year in which its filing describes more than one EIA utility, or no distribution system. The Publisher shall not assign two Respondents to one EIA utility id in the same report year.
5.11 The Publisher shall publish the Respondent Map, with the reason for each assignment that differs from a public crosswalk, and shall state the map’s revision with every figure that uses it.
6 Methodology
6.1 Utilization
6.1.1 The Publisher shall compute Utilization for a utility as Utilization = E / (C × PF × H) × 100%, where E is Energy Delivered, in MWh; C is the Nominal Capacity or the Operational Capacity, in MVA; PF is the Power Factor, 0.95; and H is the hours in the report year that supplies E, 8,760 or 8,784.
6.1.2 The Publisher shall take Energy Delivered as the sum of EIA-861 retail sales, delivery-only sales, and losses. For a utility under 5.1.3, the Publisher shall take Energy Delivered as the sum of the Energy Account’s sales to ultimate consumers and losses.
6.1.3 The Publisher shall take the Nominal Capacity as the sum of the rated capacity of the rows of the Substation Schedule kept under 6.2.
6.1.4 The Publisher shall take the Operational Capacity as the sum, over the kept rows, of: the rated capacity, less 0.7 times the rated capacity divided by the number of transformers in service, at a substation with two or more transformers in service; and the full rated capacity at a substation with one transformer, or with no count reported.
6.1.5 A row with no rating adds 0 MVA to either capacity.
6.1.6 The Publisher shall report Nominal Utilization and Operational Utilization together, and shall not publish one without the other.
6.1.7 Where the Substation Schedule report year and the report year of Energy Delivered differ, the Publisher shall state both years in a note. Where both are earlier than the Year, the Publisher shall state the year used.
6.1.8 The Publisher shall compute Peak Loading, nominal and operational, as P / (C × PF), where P is the higher of the EIA-861 summer peak and winter peak, in MW, from the EIA-861 report of the same year as Energy Delivered. The Publisher shall publish Peak Loading above 100% as computed.
6.1.9 Where Energy Delivered divided by the product of H and P is below 15% or above 100%, the Publisher shall report Utilization and Peak Loading as unavailable, with a note that gives that ratio. The energy and the peak then describe different systems, as where a generation cooperative reports a wholesale peak beside a small retail energy.
6.1.10 Where the nominal Peak Loading is below 25% or above 130%, the Publisher shall report Utilization and Peak Loading as unavailable, with a note that gives the capacity, the peak, and the Peak Loading. The Substation Schedule then does not describe the utility’s distribution system: it lists only part of the system, or it counts some capacity more than once.
6.2 Screening the Substation Schedule
6.2.1 The Publisher shall decide whether a row is counted from its voltages and transformer counts, whatever its label. Utilities choose the labels transmission and distribution by their own practice, and the same kind of station carries either label at different utilities.
6.2.2 The Publisher shall keep a row whose low-side voltage is below 30 kV, unless 6.2.4 to 6.2.9 leave it out.
6.2.3 The Publisher shall read the class of rows whose low-side voltage is from 30 kV up to, but not including, 46 kV as follows. Over the rows with one or more transformers in service and not labeled generation, let S be the capacity that steps down to the class, and F the capacity whose high-side voltage is in the class and whose low-side voltage is greater than 0 and below 30 kV. Where F is at least 0.5 times S, the class supplies other stations, and the Publisher shall leave its rows out. Otherwise the class is a distribution voltage, and the Publisher shall keep its rows, unless 6.2.4 to 6.2.9 leave them out. A reviewed exception under 6.2.13 replaces this reading.
6.2.4 The Publisher shall leave out a row whose low-side voltage is 46 kV or more, as a transformer that supplies other stations.
6.2.5 The Publisher shall leave out a row with no transformer in service, as a spare.
6.2.6 The Publisher shall leave out a row labeled generation.
6.2.7 The Publisher shall leave out a row with no label, or with no low-side voltage. The Publisher shall keep a line that groups stations under 10 MVA and reports no low-side voltage, because FERC Form No. 1 permits a filer to report its substations under 10 MVA as one grouped line. A line groups such stations where its name says the stations are under 10 MVA, or where its name begins with a number of stations and its capacity divided by that number is 10 MVA or less.
6.2.8 The Publisher shall leave out a row that otherwise qualifies under 6.2.2 or 6.2.3 and whose capacity per transformer in service is more than 100 MVA, as a generator step-up transformer or a capacity entered in kVA rather than MVA.
6.2.9 The Publisher shall leave out a row whose name begins with a number of stations and whose capacity divided by that number is more than 10 MVA, as a count of stations listed elsewhere in the filing. The Publisher shall also leave out a row whose name marks it as a summary, subtotal, or total, and whose capacity is within 0.5% of the sum of the other kept rows with the same high-side and low-side voltages, as a subtotal of those rows. The Publisher shall keep a row so named that reports no voltages, or whose capacity matches no such sum.
6.2.10 The Publisher shall state, in a note, the number of rows and the MVA left out under each of 6.2.3 to 6.2.9, and, where the class under 6.2.3 is kept, that it was counted as distribution.
6.2.11 Where the Operational Capacity differs by more than 40% from that of the utility’s latest earlier filing whose rows kept under this screen give a capacity greater than zero, the Publisher shall flag the change in a note, with both values and their years. The Publisher shall screen the earlier filing with the same reading of the class under 6.2.3 as the current filing. The note shall give a station-by-station account of the difference: the stations counted in the current filing and not in the earlier one, and the reverse.
6.2.12 The Publisher shall not withhold a figure because of a flag under 6.2.11.
6.2.13 The Publisher shall keep a list of reviewed exceptions, each for one utility. An entry shall do one of the following: withhold the utility’s Utilization and Peak Loading, with a reason that the Publisher publishes as the note; or set the reading of the class under 6.2.3. The Publisher shall make an entry only on evidence that a reader can check, such as the utility’s own published substation ratings, or a filing that visibly omits stations. The Publisher shall publish the list with the evidence for each entry, and shall state its revision with every figure that uses it. Appendix 4 gives the list for this version.
6.3 Load Factor
6.3.1 Where a balancing authority’s Load Curve has Coverage of at least 90%, the Publisher shall compute its Load Factor as the mean of the kept hourly loads divided by the highest kept hourly load. Below 90%, the Publisher shall report the Load Factor as unavailable, with the Coverage in a note.
6.3.2 For a utility with a Published Curve whose Coverage is at least 90%, the Publisher shall compute the Load Factor from that curve, under 6.3.1.
6.3.3 For any other utility, the Publisher shall compute the Load Factor from EIA-861 as E / (H × P), with E, H, and P as in 6.1, and shall label the source as EIA-861, with its report year. Where E comes from the Energy Account under 5.1.3, the Publisher shall label the source as FERC Form No. 1 and EIA-861. Where the utility has no usable EIA-861 report, the Publisher shall report the Load Factor as unavailable, with the reason.
6.3.4 The Publisher shall not compute a utility’s Load Factor from a Proxy, because a wider area’s Load Factor is not the utility’s.
6.3.5 The Publisher shall state the source of every Load Factor (hourly; EIA-861; or FERC Form No. 1 and EIA-861) and its year.
6.3.6 The Publisher shall not compute a Load Factor for a balancing authority on its exclusion list, and shall publish that list with the reason for each entry. Appendix 2 gives the list for this version.
6.3.7 An EIA-861 Load Factor applies the same definition as an hourly one. It is the same figure at an older vintage, because EIA-861 is published about a year after the year it reports.
6.3.8 Where a Load Factor computed under 6.3.3 is below 15% or above 100%, the Publisher shall report the Load Factor as unavailable, with a note that gives the computed value.
6.4 Screening the hourly load
6.4.1 The Publisher shall remove a Glitch Hour from a Load Curve before it computes a Load Factor or an Hourly Signal. A Glitch Hour is an hour whose load is lower than the load of each neighboring hour by more than 30% of the median hourly load of the Year, or higher than each by more than that amount.
6.4.2 The Publisher shall take the neighbors of an hour as the nearest hours that have a load. The first and last hours of the Year have one neighbor. The Publisher shall judge each hour against the unscreened loads.
6.4.3 The Publisher shall not estimate a value for a Glitch Hour or a missing hour. Each counts against Coverage.
6.4.4 The Publisher shall publish the number of Glitch Hours removed with every figure computed from the Load Curve.
6.4.5 The screen catches a single bad hour. A block of several bad hours passes it, which is why 6.3.6 provides an exclusion list.
6.5 The Hourly Signal
6.5.1 The Publisher shall assign each kept hour a Relative Load. Where the hour’s load L is at or above the mean load M of the kept hours, the Relative Load is (L − M) / (Lmax − M). Where L is below M, the Relative Load is (L − M) / (M − Lmin). Lmax and Lmin are the highest and lowest kept loads.
6.5.2 Where the highest and the lowest kept loads are equal, the Publisher shall assign every hour a Relative Load of 0.
6.5.3 For a utility, the Publisher shall take the Hourly Signal from the Load Curve that its curve map assigns, with the label of that entry. Where the curve map has no entry for the utility and its latest EIA-861 report lists exactly one balancing authority, the Publisher shall take the Hourly Signal from that balancing authority’s EIA-930 demand, labeled as a Proxy. Where the report lists no balancing authority or several, the Publisher shall report the Hourly Signal as unavailable.
6.5.4 The Publisher shall label every Hourly Signal as published, proxy, or unavailable, and shall name the area its Load Curve covers.
6.5.5 The Publisher shall summarize the Hourly Signal as a 12 × 24 table of the mean Relative Load for each month and hour of day in Local Standard Time, with an empty cell where no kept hour falls.
6.5.6 The busiest hour of the Year has a Relative Load of +1, an hour at the mean load has 0, and the quietest hour has −1. Each side of the mean is scaled on its own, so a Relative Load compares an hour with the Year’s own extremes. For example, ISO New England’s mean load in 2025 was 13,160 MW and its peak 25,898 MW, so an hour at 19,529 MW, halfway from the mean to the peak, has a Relative Load of +0.5.
6.5.7 The exact effect of one hour’s load on the Load Factor is zero in every hour except the peak hour. That derivative cannot value timing, so this methodology defines Relative Load as the signal.
6.5.8 A Relative Load is not comparable between Grids, because each Grid is scaled to its own Year. A Relative Load of +0.5 in two Grids corresponds to different loads in MW.
6.5.9 This methodology does not score a resource. A resource methodology that scores a change in load against the Hourly Signal takes the score as −Σ(ΔLh × rh) / Σ|ΔLh|, where ΔLh is the change in load in hour h (added load positive, removed load negative) and rh is that hour’s Relative Load. The sign makes a positive score mark a change that helps the Grid: load added in quieter hours, or removed from busier hours. The Relative Load itself keeps +1 for the busiest hour, because it describes hours, not resources.
6.6 Time basis
6.6.1 The Publisher shall run each Year from midnight on January 1 to midnight on the next January 1, in the Grid’s Local Standard Time.
6.6.2 The Publisher shall take the standard offset as the Grid’s offset from UTC on January 15.
6.6.3 Where a Grid has no mapped time zone, the Publisher shall report the Hourly Signal and the hourly Load Factor as unavailable.
6.7 Finality
6.7.1 The Publisher shall treat an EIA-930 Load Curve as Final 90 days after the Year closes in Local Standard Time.
6.7.2 The Publisher shall treat an ISO load zone Load Curve as Final only when the ISO has marked every hour of the Year as verified.
6.7.3 Where a figure uses a Load Curve, the Publisher shall label it Provisional until that curve is Final. Where a figure uses no Load Curve, the Publisher shall label it Provisional until the Year closes in Local Standard Time.
6.7.4 Where no Year is requested, the Publisher shall publish the latest Year that closed at least 90 days earlier.
6.7.5 The Publisher shall not publish a figure for a Year before 2019.
6.7.6 A Provisional figure for a Year in progress measures Coverage against the hours elapsed. It leaves out at least one season, so it is not comparable with a Final figure.
6.8 Reporting requirements
6.8.1 Every published figure shall state the methodology name and version, the Grid, the Year, and whether the figure is Provisional.
6.8.2 Every Utilization figure shall state the source and report year of its Energy Delivered (EIA-861, or the Energy Account), its Substation Schedule report year, the Power Factor, both capacities in MVA, the rows left out, every flag under 6.2, and any exception under 6.2.13 that applies.
6.8.3 Every Load Factor shall state its source and year.
6.8.4 Every Hourly Signal shall state its label, the area of its Load Curve, the Coverage, and the number of Glitch Hours removed.
6.8.5 Where a figure is unavailable, the Publisher shall state the reason.
7 Assumptions
7.1 The following assumptions shape how closely the published figures describe the grid’s real use of its capacity, listed roughly from largest to smallest expected effect:
7.1.1 Utilities rate their transformers on different bases. The Substation Schedule does not say which rating a filer reports. Most utilities checked for this draft file within about 15% of the normal ratings they publish elsewhere, but some file the self-cooled base nameplate, about half the rating a bank runs to. A base-nameplate filing biases Utilization and Peak Loading high, by up to about two times, and stays inside the screens in 6.1.9 and 6.1.10. The Publisher withholds the known cases under 6.2.13. An unknown case is not detected. Across all filers, the basis is the largest remaining source of scatter between utilities.
7.1.2 The voltage rule reads some stations wrongly. A row is counted from its low-side voltage, not its label. A transformer that steps down below 30 kV to feed other stations is counted, and so is the station it feeds, which biases Utilization low and overstates headroom. A 30–46 kV class that the test in 6.2.3 reads the wrong way either counts supply stations or leaves out distribution stations. Against the substation ratings that 19 utilities publish, the rule read the class correctly in every case where the published data decided it, and placed counted capacity within about 25% of the published ratings at all but the base-nameplate filers.
7.1.3 Energy Delivered includes customers whose stations are screened out. Energy Delivered counts every customer the utility bills. A large customer served from a station that 6.2.4 leaves out adds energy but no counted capacity, which biases Utilization high, and losses include transmission losses. The direction is not conservative. The effect depends on the voltage that serves the customer, not on the type of customer: for 2025, Appalachian Power’s rate schedules bound it between 3% (five named contract customers) and 30% (all large-power schedules) of Energy Delivered, while a data center station that steps down below 30 kV is counted, so its energy and its capacity are both in the figure. A utility’s own circuit-level figures do not have this bias. A later version will subtract the energy of customers served from screened-out stations where public filings identify them.
7.1.4 The contingency reserve assumes equal transformers. The Operational Capacity removes 0.7 of the average transformer at a substation, not the largest one. Where units differ in size, the reserve is understated, and Operational Utilization is biased low.
7.1.5 The Power Factor is fixed. A real power factor varies by utility and hour. Energy Deliverable scales in proportion to it, so a true power factor of 0.90 or 1.00 moves Utilization by about 5% of its value.
7.1.6 The capacity year and the energy year can differ. The two sources are taken independently, at their latest usable report years. A gap is stated in a note, and the error grows with the load growth between the two years.
7.1.7 Energy Account energy can leave out losses. For a utility under 5.1.3, Energy Delivered is the Energy Account’s sales to ultimate consumers. These utilities deliver energy that retail providers sell, as in the ERCOT market, and their filings do not always report losses. Without losses, Energy Delivered is low by the loss share, typically 3% to 6%, which biases Utilization and Load Factor low and overstates headroom. The direction is not conservative for a claim of headroom. The note under 5.8 states when losses are missing.
7.1.8 A Proxy curve describes a wider area. A utility’s Hourly Signal from its balancing authority reflects that wider area’s timing. The signal is labeled, and this methodology never takes a utility’s Load Factor from it.
7.1.9 The glitch screen misses a block of bad hours. Such a block can set the peak and lower the Load Factor. The exclusion list in 6.3.6 removes the known cases. An unknown block is not detected.
7.1.10 Local Standard Time differs from the clock in summer. In daylight saving time, an hour of the Hourly Signal sits one hour earlier than the wall clock. The 12 × 24 table states its clock.
Appendix 1: Worked Illustrations
Appendix 1.1: Utilization of a utility (Ohio Power Company, 2025)
Ohio Power Company (EIA utility id 14006) reports its distribution substations in the Substation Schedule. For 2025, the inputs and results are:
| Quantity | Nominal | Operational |
|---|---|---|
| Energy Delivered, EIA-861 2025 (MWh) | 55,865,244 | 55,865,244 |
| Capacity, Substation Schedule 2025, kept rows (MVA) | 14,554.18 | 12,463.14 |
| Energy Deliverable = capacity × 0.95 × 8,760 h (MWh) | 121,119,886 | 103,718,266 |
| Utilization | 46.1% | 53.9% |
| Annual peak, higher of summer and winter (MW) | 10,224 | 10,224 |
| Peak Loading = 10,224 / (capacity × 0.95) | 73.9% | 86.4% |
The screen in 6.2 left out 154 rows: 104 rows (18,505 MVA) that step down to 46 kV or more (6.2.4), 32 spares (1,986 MVA, 6.2.5), 15 rows labeled generation (855 MVA, 6.2.6), 1 row with no low-side voltage (12 MVA, 6.2.7), and 2 rows over 100 MVA per transformer (598 MVA, 6.2.8). The filing feeds few stations from its 30–46 kV class, so 6.2.3 counts that class as distribution. The label rule of the earlier draft gave 53.4% and 59.2%: it counted spares and generation rows, and left out distribution-voltage stations labeled transmission.
Ohio Power has no Published Curve of its own, so its Load Factor comes from EIA-861 under 6.3.3: 55,865,244 MWh / (8,760 h × 10,224 MW) = 62.4% (EIA-861, 2025). Its Hourly Signal is a Proxy, the PJM AEP zone.
Appendix 1.2: Load Factor of a balancing authority (ISO New England, 2025)
ISO New England (EIA-930 code ISNE) reported a load for every hour of 2025, and the screen in 6.4 removed none, so Coverage is 100%.
| Quantity | Value |
|---|---|
| Mean hourly load (MW) | 13,160 |
| Highest hourly load (MW), 24 June 2025, 18:00 EST | 25,898 |
| Load Factor = 13,160 / 25,898 | 50.8% |
Utilization is unavailable for a balancing authority (Scope 2.3.1). In the Hourly Signal, the busiest hours are July afternoons and evenings, from about 16:00 to 21:00 EST. The quietest are April and May nights and middays. The midday trough follows from behind-the-meter solar, which ISO New England’s demand series records as lower load.
Appendix 2: Balancing Authorities Excluded from Load Factor
| Code | Balancing authority | Reason |
|---|---|---|
| NWMT | NorthWestern Energy (Montana) | 2025 demand carries a multi-hour data error that the one-hour glitch screen does not catch. |
The exclusion applies to the Load Factor for every Year. The balancing authority keeps its Hourly Signal.
Appendix 3: Balancing Authority Aliases
A Publisher can accept a common ISO label for a balancing authority and answer with the EIA-930 code, stating the substitution:
| Label | EIA-930 code |
|---|---|
| CAISO | CISO |
| ERCOT | ERCO |
| NYISO | NYIS |
| ISO-NE | ISNE |
| SPP | SWPP |
| LADWP | LDWP |
Appendix 4: Reviewed Exceptions
This list is normative for this version. It applies 6.2.13.
| EIA utility id | Utility | Entry | Evidence |
|---|---|---|---|
| 13214 | Rhode Island Energy (The Narragansett Electric Company) | Withhold | The filing rates each transformer at its self-cooled base nameplate. The utility’s published summer-normal bank ratings are 1.98 times the filed ratings at the median of 73 stations matched by name, and many banks peak above the filed rating. |
| 13573 | Niagara Mohawk Power Corporation | Withhold | The filing rates each transformer at its self-cooled base nameplate. The utility’s published summer bank ratings are 1.5 times the filed ratings at the median of 375 stations matched by name, and 1.9 times in total. |
| 14154 | Orange and Rockland Utilities | Withhold | The filing lists about 30 distribution stations. The utility publishes ratings for 96 substation banks at about 50 stations. |
| 24590 | Unitil Energy Systems | 30–46 kV class read as distribution | The utility’s hosting capacity map lists 34.5 kV as the low-side voltage of feeders it serves directly. |
Bibliography
- Shea, R., Chen, X., Reaves, C., and Webster, B. The Opportunity for Heat Pumps to Put Downward Pressure on Electricity Rates. RMI, 15 October 2026. https://rmi.org/resources/the-opportunity-for-heat-pumps-to-put-downward-pressure-on-electricity-rates/. Cited in 1.1 for the effect of improved utilization on rates. Its own utilization measure (average load against installed generating capacity) differs from the one defined here.
Version History
- 0.1, 2026-10-29: First draft. Not yet circulated for public comment.
- 0.1, revised 2026-11-08: Draft revision, before public comment. Adds the Energy Account (FERC Form No. 1, page 401a) as the energy source for a utility that reports a peak but no energy to EIA-861 (5.1.3, 5.7, 5.8, 6.1.2, 6.3.3), and the Respondent Map that assigns FERC Form No. 1 filings to EIA utilities (5.9 to 5.11). Adds the subtotal screen, the plausibility screens (6.1.9, 6.1.10, 6.3.8) and assumption 7.1.7. Replaces the label-based count of the Substation Schedule with a voltage-based one (5.1.2, 6.2): rows are counted by the voltage they step down to, the 30–46 kV class is read from the filing (6.2.3), spares, generation and step-up rows are left out, and the subtransmission flag is removed. Adds reviewed exceptions (6.2.13, Appendix 4) and assumptions 7.1.1 and 7.1.2. States the sign of a resource score against the Hourly Signal (6.5.9).