energy · family: the rulebook was written for a world that no longer exists
standards written for machines that retire
Grid-forming inverters may be inadvertently limited by interconnection standards written for grid-following inverters
Problem statement
Power system operations and controls were designed around the physical characteristics of synchronous machines, and as more inverter-based resources are integrated, systems with high levels of them lack the rotating inertia and synchronizing torque those machines supply. Grid-forming (GFM) inverters hold a nearly constant internal voltage and are expected to resist changes in frequency, though they are not expected to supply the high fault current of rotating machines; but the two interconnection standards — IEEE 1547-2018 (distribution) and IEEE 2800-2022 (transmission) — focus primarily on grid-following (GFL) inverters. UNIFI Consortium simulation tests of a generic GFM model found that it responds faster than the IEEE 1547-2018 minimum allowable response times, and that the generic model had to be modified for the tests. The gap analysis leaves open whether the existing standards "have a gap or inadvertent barrier to GFM technology"; the UNIFI specifications say some requirements "may not be appropriate for or may even inadvertently limit the use of GFM resources."
Why this matters
The UNIFI specifications — written by authors from national laboratories, EPRI, universities, utilities, and equipment manufacturers — say that GFM resources with black start capability "may be essential in high IBR penetration grids," because synchronous machines may no longer provide it. They also say that the requirements written for GFL plants in IEEE 1547-2018 and IEEE 2800-2022 apply to GFM plants too unless explicitly identified as inapplicable, that in some circumstances a requirement of those standards "is not practical for a GFM IBR," and that "exceptions may be warranted until these standards can be updated to fully account for GFM inverters." Until then, GFM resources are held to requirements that, in the specifications' words, are "generally not designed to ensure acceptable power system operation with grid-forming (GFM) resources." Interconnection queue delays are a separate problem, covered in existing brief `energy-grid-connection-queue-bottleneck`. The UNIFI Consortium — co-led by NREL, the University of Texas at Austin, and EPRI, and supported by the DOE Solar Energy Technologies Office — is addressing the challenges of integrating GFM inverters alongside rotating machines and other inverter-based resources.
What’s been tried and why it hasn’t worked
The UNIFI Consortium published a preliminary gap analysis (NREL/TP-5D00-88609, 2024) testing generic GFM models in PSCAD simulation against IEEE 1547 and 2800 requirements. To run the tests the authors had to modify the generic model: the frequency droop gain was changed to 20 (5%) because oscillations in active power were observed with the default values during frequency disturbances. In the volt-var test, for a voltage step inside the continuous operation region, the model responded in under 1 second — the fastest response IEEE 1547-2018 allows — and so fell outside the allowable range. In the frequency-droop test the model was again faster than the minimum of the allowable ranges, but a footnote under Table 24 of IEEE 1547-2018 permits response times lower than the minimums shown, so the report concludes that this response is allowed by the standard. Against IEEE 2800-2022 the model "mostly aligns" with the requirements tested, except that the rise and settling times of its currents during low voltage ride-through were longer than the standard requires. The report says its evaluation metrics "are not the pass/fail criteria for compliance" and that it did not tune the model to meet the tests. The UNIFI Consortium, with DOE support, also published specifications for GFM resources (Version 1, December 2022), which establish functional requirements and performance criteria for GFM resources at any scale. The specifications describe the existing standards as "in different stages of adoption"; neither source says how long it would take to update them.
What would unlock progress
An accelerated, narrowly scoped amendment to IEEE 1547 and 2800 that explicitly accommodates GFM behavior — rather than a full revision cycle. This requires validated test procedures for GFM-specific characteristics: voltage source behavior, faster-than-minimum frequency response, and operation in islanded mode. Certification test procedures would also need a GFM-specific protocol. The UNIFI preliminary gap analysis identifies "many more tests still yet to be done" — completing this testing program would provide the technical basis for standards amendments.
Entry points for student teams
A team could build a hardware-in-the-loop simulation of a GFM inverter connected to a simplified grid model and systematically test which IEEE 1547 requirements the GFM passes, fails, or triggers ambiguous results. This extends the UNIFI gap analysis, which was run in PSCAD simulation on a generic model. A power systems simulation (e.g., in PSCAD or OpenDSS) comparing grid stability under GFL-only vs. mixed GFL/GFM scenarios would also produce useful data. Relevant disciplines: electrical/power engineering, control systems, simulation/modeling.
Genes
Tags marked “under review” were questioned by a later calibration pass; they stay visible here but are left out of filters until re-adjudicated.
Structural cousins: stuck the same way, in other fields
Sources
Ramasubramanian, D., et al., "Preliminary Gap Analysis of Existing IEEE 1547 and IEEE 2800 Standards Towards GFM Technology," NREL/TP-5D00-88609, UNIFI Consortium / National Renewable Energy Laboratory, February 2024, Kroposki, B., et al., "UNIFI Specifications for Grid-forming Inverter-based Resources—Version 1," UNIFI-2022-2-1, UNIFI Consortium, December 2022, go to source 1 ↗ go to source 2 ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
- Related to existing brief `energy-grid-inertia-loss-frequency-instability` (which covers the physics problem) but this brief focuses specifically on the standards barrier that blocks the solution.
- Related to existing brief `energy-grid-connection-queue-bottleneck` (which covers queue delays) but this is about a specific technical standards mismatch, not the queue process.
- The gap analysis authors are from EPRI, the University of California, Berkeley, and GE Energy Consulting, plus a student intern, with an NREL technical monitor; the UNIFI Consortium is co-led by NREL, the University of Texas at Austin, and EPRI, and is supported by the DOE Solar Energy Technologies Office.
Reconciliation 2026-09-30: The whole write-up was read against both sources it names, each in full text, fetched 2026-09-30. The gap analysis (52 pages) was read at https://docs.nlr.gov/docs/fy24osti/88609.pdf, because the address on the Source line, docs.nrel.gov, no longer resolves; the file at the new address is the same report, and its record was also read at https://www.osti.gov/biblio/2318959. The second source had no link; it was found and read (18 pages) at https://www.energy.gov/sites/default/files/2023-09/Specs%20for%20GFM%20IBRs%20Version%201.pdf. Changes: (1) Long title. It read "Grid-Forming Inverters Can't Interconnect at Scale Because Standards Were Written for the Technology They're Replacing." Neither source says grid-forming inverters cannot interconnect; the gap analysis finds its model mostly within the requirements, and the specifications say the standards "may even inadvertently limit the use of GFM resources." The title now reads "Grid-Forming Inverters May Be Inadvertently Limited by Interconnection Standards Written for Grid-Following Inverters." The Display title was left as it was. (2) Source line. The report number was given as NREL/TP-88609; the cover gives NREL/TP-5D00-88609, February 2024, and the first author is Ramasubramanian. The second source was given as a DOE document dated 2023; the document itself names the UNIFI Consortium (Kroposki and others) as author, carries the identifier UNIFI-2022-2-1 and a release date of December 15, 2022, and is titled "UNIFI Specifications for Grid-forming Inverter-based Resources—Version 1." DOE funds the work and hosts the file. The Source line now carries these details, copied from the documents, and the link for the second source. (3) The opening sentence said the grid loses synchronous machines "as coal and gas plants retire." Neither source mentions plant retirements. The sentence now gives what the specifications say: power systems were designed around synchronous machines, and systems with many inverter-based resources lack their rotating inertia and synchronizing torque. (4) The text said grid-forming inverters "can replicate" voltage formation, inertia and fault current. The specifications say a grid-forming inverter holds a nearly constant internal voltage, is expected to resist changes in frequency, and is not expected to inject the high fault current of a rotating machine; the sentence now says that and no more. (5) The text said the standards were designed for grid-following inverters "that depend on the grid they're supposed to be replacing." The specifications say only that the standards "focus primarily on grid-following (GFL) technologies"; the rest was removed. (6) The text said twice that the model changes made for the tests "wouldn't be needed for GFL," and called them "test procedure failures." The report says only that the frequency droop gain was changed to 20 (5%) because oscillations in active power were seen with the default values. The comparison with grid-following inverters and the word "failures" were removed. (7) The phrase "inadvertent barriers," in quotation marks, was presented as a finding, with the claim that the requirements "block GFM deployment." The report uses the phrase once, to say more tests are needed "to further verify if the existing standards have a gap or inadvertent barrier to GFM technology." The text now quotes that, and the specifications' own wording. (8) The quotation "not immediately obvious, creating a de facto barrier" is not in either source. The report says the opposite about the footnote under Table 24: because of it, the fast frequency response "is allowed by" IEEE 1547-2018. The quotation was removed and the paragraph now reports what the tests found: one response faster than the standard allows (the volt-var test, inside the continuous operation region), and current rise and settling times longer than IEEE 2800-2022 requires during low voltage ride-through. (9) "DOE published GFM specifications (Version 1) as a reference, not a standard" now names the UNIFI Consortium as publisher, with DOE support, and describes the document in its own terms. (10) Removed because neither source contains them: that needing grid-forming inverters is "the consensus view of DOE, NREL, EPRI, and major utilities"; that they "cannot be interconnected at scale until standards are revised"; that the interconnection backlog is severe and standards uncertainty compounds queue delays (the pointer to the queue brief is kept); that DOE created an "Essential Grid Operations from Solar (EOS)" project for this; that the IEEE 1547 revision ran from 2003 to 2018 and took 15 years; and that UL 1741 SA/SB certification testing lags by another year or more. The matching Source Notes lines on the 15-year cycle, the UL 1741 certification gap, and the EOS program were removed, and the line on who the technical authorities are now gives the authors' institutions and the funder from the report. In the unlock section, "The UL 1741 certification test procedure" became "Certification test procedures," and "island-to-grid transition" became "operation in islanded mode," which is the report's wording. (11) Entry points. "This directly replicates and extends the UNIFI gap analysis at lower cost" was changed: the analysis was a PSCAD simulation of a generic model, so a hardware-in-the-loop test extends it and does not replicate it, and neither source says anything about cost. Confirmed in the sources and left as written: the report title; IEEE 1547-2018 for distribution and IEEE 2800-2022 for transmission; tests of generic grid-forming models against both standards; the droop gain of 20 (5%); the Table 24 footnote; PSCAD as the simulation platform; the quotation "many more tests still yet to be done." Not in either source and left as the write-up's own proposal: the accelerated, narrowly scoped amendment; the list of test procedures needed; OpenDSS as a tool. The three tagging notes above dated 2026-08-18 quote the text as it stood then, including the 15-year figure and UL 1741, and are left as written. Genome tags untouched. A second reader went over each of these repairs on the same day, 2026-09-30, after reading both sources again in full at the same addresses (the gap analysis at docs.nlr.gov, its OSTI record 2318959, and the specifications at energy.gov); each repair was confirmed in the sources. One sentence was reworded: the repaired opening had said grid-forming inverters "can take on part of that role" (supplying rotating inertia and synchronizing torque), which neither source states, and item (4) above now gives the final wording. The other changes stand as written, and nothing was undone.