ERCOT’s Inverter-Based Resource Working Group (IBRWG) used its August 16 meeting to get pragmatic about how plants move through voltage ride-through (VRT) and back into primary frequency response (PFR), and to tighten expectations around reactive capability, modeling fidelity, and post-commissioning validation. The session covered ERCOT updates, OEM perspectives, and a quick tour of NERC’s Order 901 milestones.
Clarifying reactive capability at low MW output
ERCOT Operations reported an ongoing legal review to align Operating Guide §2.9.1 (added via NOGRR-245 and referencing IEEE 2800 §5) with Protocol §3.15. The crux is whether inverter-based resources must provide reactive support at all MW outputs, including 0 MW, versus the legacy “≥10% of capacity” framing. ERCOT will return with definitive language and, importantly, how this will be evaluated in interconnection studies and commissioning. Developers should assume the standard is converging toward “reactive at any output” and check contracts, PPC logic, and test plans accordingly.
Making VRT and PFR play nicely together
The centerpiece discussion: plants can misbehave when a PPC changes active-power setpoints during a VRT event based on noisy, short-window frequency measurements at the POI. When inverters exit VRT and hand control back to the PPC, they may “snap” to that transient setpoint (often downward), which is the opposite of what grid operators expect after a fault.
Participants walked through why this happens; phase jumps and non-fundamental content distort zero-crossing-based frequency estimates; PLLs aren’t immune either when voltages are unsettled. Designers typically balance filtering/averaging, temporary control freezing, deadbands, and hysteresis: heavier filtering and a short freeze window tame false detections but slow true frequency response. Typical figures cited in discussion: frequency filtering on the order of hundreds of milliseconds, and freeze/hold behavior on the order of seconds.
Participants also added implementation color: LVRT/OVRT behavior and exits are configurable; inverters often ignore PPC setpoints while in LVRT and then resume following commands upon exit. PPC-based droop tends to cycle at roughly ~1–2 seconds, whereas true FFR must live in the inverter to hit sub-quarter-second targets. Across cases, the room gravitated toward a simple, vendor-agnostic principle: finish VRT cleanly, then check frequency and engage PFR. That implies freezing PPC commands to the pre-disturbance setpoint during VRT, using status flags so PPCs “know” the plant is in LVRT/OVRT, and applying controlled active-power ramps as voltage stabilizes.
Models, validation, and the PPC wrinkle
Participants highlighted persistent gaps between internal OEM models, customer EMT models, and fielded controls, especially when PPC and inverter are from different vendors. The recommended path is layered validation: internal simulation → controller-hardware-in-the-loop → customer EMT/PSCAD models → post-event checks with real data. Several voices called for explicit PPC-model validation checkpoints (e.g., at QSA) rather than focusing solely on unit models. ERCOT reminded owners to update models whenever control behavior changes.
Order 901 progress and adjacent workstreams
On the standards front, PRC-029 has FERC approval (effective Oct 1), which triggers implementation plans for PRC-028 and PRC-030. NERC’s milestone work on generator model/data verification is moving ahead, with additional comment/ballot windows open on model validation topics and a uniform IBR modeling framework. The message for owners and developers: be ready to demonstrate performance and model fidelity, not just attest to it.
Why this matters even more with Large-Loads (including crypto)
ERCOT’s growing cohort of **Large-Loads, crypto facilities and other high-ramp data/industrial loads,**tightens the timing margins around frequency behavior. When a fault occurs near large, fast-changing demand, clean VRT/PFR sequencing and trustworthy frequency measurement windows become essential for avoiding over-reactions, unnecessary curtailments, and second dips in voltage. Getting the plant’s measurement, freeze, and handoff behavior right is a reliability and economics issue—full stop.
What good looks like from here
Expect ERCOT to clarify the reactive-at-low-MW requirement and to sketch best-practice guidance for VRT→PFR sequencing (measurement windows, freeze logic to pre-disturbance values, ramp expectations, and handover criteria). In parallel, owners and developers can pre-empt surprises by confirming PPC–inverter coordination (status flags, freeze semantics), aligning POI meter frequency settings with PPC logic, and proving behaviors with integrated HIL or FAT cases that include frequency-during-VRT and frequency-after-VRT scenarios—then reflecting that logic in customer EMT models and post-commissioning monitoring.
If you want this squared away before the next disturbance, ZEG can run a focused VRT↔PFR Coordination Readiness Check for your ERCOT sites. We’ll review control parameters (freeze/hysteresis/deadbands, pre-disturbance capture, ramps), verify POI frequency-measurement and PPC filtering settings, align your EMT/PSCAD models with the real controller logic, and supply a compact HIL/FAT playbook that exercises VRT-plus-frequency cases with clear pass/fail criteria—plus templates for post-event evidence under PRC-028/030.
Ready to make your plant “VRT/PFR-ready”? Contact us today, and we’ll tailor a scope that fits your timeline and risk profile.
Learn more about large load infrastructure, explore our Comprehensive Guide on Large Load Infrastructure here, or view the IBRWG meeting materials here.
