About Keentel Engineering

Who We Are

Keentel Engineering is a specialized power system studies and modeling firm serving Resource Entities, Interconnecting Entities, OEMs, developers, and utilities operating in ERCOT and across North American grid jurisdictions.

Keentel Engineering supports the renewable energy transition by delivering rigorous, defensible, and timely dynamic models, model quality test reports, verification reports, and grid interconnection studies. Our team blends deep simulation expertise across PSS/E, TSAT (DSATools), and PSCAD with practical knowledge of ERCOT Planning Guide and Nodal Operating Guide requirements.

From early Feasibility Interconnection Studies (FIS) through Quarterly Stability Assessment (QSA) entry, commissioning Part 3 approval, and the recurring MOD-026/027 obligations, Keentel Engineering acts as a single point of accountability for the technical artifacts ERCOT requires.

Service Areas

• Dynamic stability modeling in PSS/E and TSAT for synchronous, inverter-based, and hybrid resources.
• Electromagnetic Transient (EMT) modeling in PSCAD for IBRs, including weak-grid and subsynchronous studies.
• Model Quality Test (MQT) preparation with combined PSS/E + PSCAD + TSAT overlay plots.
• Model Parameter Verification reports for new commissioning, periodic refresh, MOD-026/027 cycles, and post-tuning field changes.
• User-Defined Model (UDM) review, debugging, and TSAT template (.tudm) preparation.
• Generic versus UDM model trade-off analysis with OEM coordination.
• RIOO-RS submittal preparation, file packaging, and dispute resolution support.
• Specialized weak-grid studies: Short Circuit Ratio (SCR) analysis, sub-synchronous resonance, REGC_A and REGC_B model selection and tuning.
• Advanced Grid Support (AGS) and Grid-Forming (GFM) inverter modeling, including hybrid AGS-ESR + non-AGS configurations.
• Hybrid facility modeling: wind plus solar, solar plus storage, IBR plus conventional, and resource plus large load arrangements.

Why Keentel Engineering

ERCOT model rejection — or worse, an undetected modeling defect that surfaces in commissioning — can delay a project by months and cost a developer millions of dollars in deferred revenue. Keentel Engineering invests in the test rigs, scripts, and quality checks that make first-time-right submissions the norm rather than the exception. Every report we deliver is structured around the exact rubric ERCOT applies during review.

---

Executive Summary

The Electric Reliability Council of Texas (ERCOT) operates one of the most renewable-intensive grids in the world. To safeguard reliability against the dynamic and transient phenomena introduced by hundreds of inverter-based resources, ERCOT enforces a uniquely rigorous framework of dynamic modeling, model quality testing, parameter verification, and EMT analysis. The framework is codified in the ERCOT Planning Guide, the Nodal Operating Guide, and the Dynamics Working Group (DWG) Procedure Manual.

This white paper distills the entirety of ERCOT's current dynamic model submission framework — current to the April 2026 release of the Dynamic Model Submittal Guideline (v1.11) — into a single, practical document. It is written for Resource Entity engineers, Interconnecting Entity project managers, OEM modelers, and senior leadership who need to understand both the technical requirements and the project-management implications of getting these submissions right.

Key Takeaways

---

Section 1: The ERCOT Dynamic Modeling Regulatory Landscape

1.1 Why ERCOT Cares So Much About Dynamic Models

ERCOT operates an islanded interconnection of approximately 90,000 MW peak load, with renewable penetration that regularly exceeds 50 percent on a system-wide energy basis and frequently 100 percent during low-load hours. Unlike conventional synchronous generation, inverter-based resources (IBRs) do not provide natural inertia or fault-current contribution. Their dynamic behavior is entirely a function of firmware: ride-through logic, phase-lock loop behavior, current-limit priorities, and reactive support strategies are all programmable, and they vary substantially across manufacturers, firmware revisions, and even commissioning settings.

Misrepresenting that behavior in a planning model is not an academic concern. Two documented events — the 2021 Odessa Disturbance and its 2022 sequel — saw approximately 1,100 MW and 1,700 MW of solar generation respectively trip off the grid in response to single normally-cleared faults, well within the Voltage Ride Through (VRT) envelope the facilities were obligated to withstand. The root cause in both cases traced back to inverter protection settings that did not match the dynamic models on file with ERCOT and NERC.

1.2 The Governing Documents

Document	Authority	Primary Subject	Update Cadence
ERCOT Nodal Protocols	Binding tariff	Resource obligations, settlement	Continuous (NPRRs)
Nodal Operating Guide	Binding tariff	VRT, frequency response, AVR, PFR	Continuous (NOGRRs)
ERCOT Planning Guide	Binding tariff	Sections 5.5 and 6.2: model submission, MQT, verification	Continuous (PGRRs)
DWG Procedure Manual	Stakeholder procedure	Detailed MQT test specifications, plot criteria	Periodic
Dynamic Model Submittal Guideline	Reference	Practitioner workflow, deliverables, file naming	Annual (v1.11 Apr 2026)
NERC MOD-026/027	NERC standard	Excitation and turbine-governor model validation	10-year cycle

1.3 The Lifecycle of a Model Submittal

Lifecycle Stage	Required Submission	Planning Guide Reference
Feasibility Interconnection Study (FIS)	PSS/E and PSCAD models; recommended MQT and AVR/PFR pre-check	PG 6.2(5)(a)
Entering Quarterly Stability Assessment (QSA)	Complete PSS/E, PSCAD, MQT overlay; Unit Model Validation (Hardware Benchmark)	PG 6.2(5)(c), (d)
Before Part 3 of Commissioning (IBRs)	MQT overlaying QSA model vs. as-built model response	PG 5.5(4)
Within 30 days of Part 3 Approval	Final PSS/E, PSCAD, TSAT models; combined MQT overlay; Verification Report	PG 5.5(6)(c)
12 to 24 Months Post-Commissioning	Updated Verification Report; refreshed models if any drift	PG 5.5(5)
MOD-026/027 Cycle (every 10 years)	Re-submission of all models; Verification update; MQT overlay	PG 5.5(6)(b)
After Any Field Change	Pre-change MQT comparison; updated full package within 30 days	PG 5.5(6)(a)

1.4 Where Submissions Land: The RIOO-RS Portal

All dynamic model submissions are uploaded to ERCOT's Resource Integration and Ongoing Operations — Resource Services (RIOO-RS) portal as zip-file attachments to an Attachment-Only Change Request. Compressed files larger than approximately 60 MB must be split. File naming follows a strict convention:

• (SITECODE)_DYNAMIC_(YYYY-MM-DD).zip — for the PSS/E and TSAT model package.
• (SITECODE)_PSCAD_(YYYY-MM-DD).zip — for the PSCAD model package.

---

Section 2: PSS/E and TSAT Stability Models

2.1 The Role of PSS/E in ERCOT Planning

Siemens PTI PSS/E is the workhorse of ERCOT Transmission Planning. Its current required versions are PSS/E 35 and PSS/E 36 — both must be supplied because ERCOT is mid-transition. For user-defined models (UDMs), separate .dll files compiled against each version are required.

Required PSS/E Deliverables

• .raw or .sav case file representing the plant up to the Point of Interconnection (POI).
• .dyr dynamic data file containing the dynamic model parameters.
• .dll file for User-Defined Models, compiled against the appropriate PSS/E version.
• Excel-format Dynamic Model Template (Universal template supports any UDM type).
• Python initialization scripts for switched shunts and on-load transformer taps.
• Model manual describing simulation setup, tunable parameters, and control block diagrams.

2.2 The Role of TSAT in ERCOT Operations

Where PSS/E serves Planning, PowerTech Labs TSAT (part of the DSATools package) serves ERCOT Operations. TSAT runs near-real-time stability analysis in the control room and supports Critical Clearing Time (CCT) scans, transfer analysis, and PMU data streaming. TSAT is required for any user-defined PSS/E model.

TSAT-Specific File Requirements

File Type	Purpose	Notes
.dyr	Dynamic data file	Must interoperate with the PSS/E .dyr file
.dll	Compiled UDM code	Separate from PSS/E DLL
.tudm	TSAT template UDM	Defines model structure only; no project-specific data
.raw	Power flow case (bus-number format)	One package using bus numbers
.pfb	Power flow case (equipment-name format)	Second package using equipment names with spaces
.mon, .dat, .swi	Monitor, data, switching files	For test scenarios

2.3 Standard Library Models vs. User-Defined Models

Attribute	Standard Library / Generic Model	User-Defined Model (UDM)
Source	Built into PSS/E and TSAT	Compiled .dll provided by OEM
Fidelity	Lower — approximated controls	Higher — closer to actual firmware
Setup effort	Low to medium	Medium to high
Weak-grid performance	Often fails at SCR below 2	Usually well-behaved if OEM-tuned
Software-version maintenance	Minimal	DLL must be recompiled per version
MOD-026/027 acceptance	Yes, with caveats	Yes — preferred
VRT compliance risk	Higher (Odessa-class failures)	Lower if maintained
TSAT model required	No	Yes — must be supplied

---

Section 3: PSCAD and Electromagnetic Transient Modeling

3.1 Why PSCAD is Required for Inverter-Based Resources

Positive-sequence stability tools like PSS/E and TSAT do not faithfully represent the inner current-control loop of a voltage-source converter, nor the dynamics of the Phase Lock Loop (PLL), nor the millisecond-scale interaction between inverter protection and grid disturbances. These are precisely the phenomena that drove the Odessa events. PSCAD is an Electromagnetic Transient (EMT) simulator that represents the network in full three-phase detail at time steps of 10 to 20 microseconds.

3.2.1 Accuracy Requirements (Highlights)

• The model must represent the full detailed inner control loop of the power electronics — no transient-stability approximations.
• Validation requires either use of actual hardware code or comparison with measured device response, including a small-signal frequency sweep over 5 to 55 Hz.
• Models assembled from PSCAD master library standard blocks are explicitly flagged as introducing unacceptable approximations.
• PLL loss-of-synchronism, phase-angle-jump, and anti-islanding protections must either be disabled or have their necessity justified by the OEM.
• DC bus dynamics, including capacitor discharging and protection, must be modeled.

3.2.2 Usability Requirements (Highlights)

• The model must initialize itself and ramp to full output without external simulation engineer input.
• It must accept external Pref, Qref, and Vref values via single-parameter adjustment.
• A 10-second simulation must complete in under approximately 300 seconds on a typical workstation.

3.2.3 Efficiency Requirements (Highlights)

• Model must be embedded in the ERCOT PSCAD Template.
• It must compile with Intel Fortran v15+ and Microsoft Visual Studio 2015+, including support for the Intel OneAPI HPC compiler.
• Both 32-bit and 64-bit Fortran libraries must be provided.
• Compatibility with PSCAD v5.0.1 and higher is required.

3.3 Weak-Grid Studies and Subsynchronous Resonance

As ERCOT's renewable fleet has grown in the Panhandle, South Texas, and West Texas, more interconnection points have moved into SCR territory of 2 or below. For any plant with an interconnection SCR of 2 or lower, PSCAD studies — using OEM-provided source-code-based models — are required, and Keentel Engineering routinely performs them.

---

Section 4: The Model Quality Test (MQT)

4.1 Purpose and Test Catalog

The Model Quality Test is the engineering rubric by which ERCOT determines whether a submitted model is fit for use in system studies. The MQT report must overlay PSS/E, PSCAD, and TSAT model responses on the same plot axes whenever multiple platforms apply.

Test	Initialization	What It Demonstrates
Flat-start	100% dispatch, steady state	Correct model initialization, no drift
POI Voltage Step Down 3%	100% dispatch, 0 MVAr at POI	Reactive injection on voltage dip
POI Voltage Step Up 3%	100% dispatch, 0 MVAr at POI	Reactive absorption on voltage rise
HVRT Leading	100% dispatch, 0.95 pf leading	High-voltage ride-through with leading initialization
HVRT Lagging	100% dispatch, 0.95 pf lagging	High-voltage ride-through with lagging initialization
HVRT Preferred (NOGRR-245)	Per NOGRR-245	Updated VRT envelope compliance
LVRT Leading	100% dispatch, 0.95 pf leading	Low-voltage ride-through, leading
LVRT Lagging	100% dispatch, 0.95 pf lagging	Low-voltage ride-through, lagging
Voltage Dip Test (Preferred LVRT)	Per NOGRR-245	Updated LVRT envelope compliance
Freq Step Down 0.3 Hz, No Headroom	80% dispatch	Correct non-response when uncurtailed
Freq Step Down 0.3 Hz, With Headroom	80% dispatch, curtailed flag set	Power increase when curtailed
Freq Step Up 0.3 Hz	80% dispatch	Power reduction
Short Circuit Ratio (SCR) Test	100% dispatch	Stable behavior down to SCR = 1.2
Phase Angle Jump (IBR PSCAD)	100% dispatch	Ride-through angle jumps without trip

4.2 MQT Scenarios by Facility Type (v1.11 Clarifications)

Facility Type	Required MQT Scenarios
Standalone Non-ESR	100% generation
Standalone ESR	ESR charging (max withdrawal); ESR discharging (max injection)
Hybrid Non-ESR + ESR	ESR charging alone; combined max injection
Hybrid Wind + Solar	Solar only; wind only; combined solar + wind
Hybrid IBR + Conventional	IBR only; IBR + conventional
Resource + Large Load	Large Load only; max withdrawal; max injection (Load offline)
Self-Limiting Facility (SLF)	All scenarios must respect SLF total output limit
AGS-ESR Standalone	Flat-start, LVRT/HVRT (charge and discharge); AGS-specific tests at 0 MW and other dispatches
AGS-ESR Hybrid	All standard tests in three columns: AGS online only, non-AGS online only, both online

4.4 The Headroom / No-Headroom Distinction

IRR models must default to no-headroom behavior. To represent a curtailed (headroom-enabled) state, the model must expose a flag or switch — not require manual adjustment of PMAX. For generic PSS/E models, the suggested approach is to set Dup = 0 in the plant controller model.

---

Section 5: Model Parameter Verification Reports

5.1 Purpose and Cadence

Verification Reports are required at four points: within 30 days of Part 3 commissioning approval; between 12 and 24 months after Part 3 approval; a minimum of every 10 years; and within 30 days of any setting change affecting dynamic response.

5.2 What a Verification Report Looks Like

ERCOT now provides a Verification worksheet directly inside the Dynamic Model Template package. The artifact must include for each tunable parameter: the model parameter name and value, the actual field equipment value in the same engineering units, the verification method, a yes/no match indication, and an explanation when values do not match.

5.3 Synchronous Plant Verification Parameters

Parameter Class	Best Verification Method	Notes
Inertia (H)	Commissioning / delivery reports	Sum all spinning masses; typical 3 to 9 MWs/MVA
Generator Electrical Parameters	Nameplate or delivery report	Generally do not change after delivery
Generator Saturation S(1.0), S(1.2)	Confirm if outside NERC ranges	S(1.0) 0.03 to 0.18; S(1.2) 0.20 to 0.85; ratio 2 to 8
Protection Parameters	Engineer attestation, relay settings	Verify most-limiting setting across all relays
Exciter Gains	MOD-026 plus direct check	MOD-026 alone insufficient
Governor Parameters	MOD-027 or direct check	Standard
Power System Stabilizer	Direct equipment query	MOD-026/027 generally insufficient for PSS

5.4 Inverter-Based Resource Verification

IBR Verification Element	Acceptable Evidence
Protection Settings (Terminal)	OEM firmware documentation; parameter dump
Protection Settings (External Relays)	Engineer attestation comparing model to relay settings
Site-Specific Tunable Parameters	OEM attestation; commissioning report; parameter dump
Non-1:1 Parameters	Qualitative explanation of derivation method
PSCAD Model (Required If Generic PSS/E)	Verification must be performed on the PSCAD model

---

Section 6: User-Defined Model (UDM) Guidelines

6.1 Why ERCOT Publishes a UDM Specification

ERCOT and its DWG observed enough recurring failure modes — models that crashed when islanded, that depended on hidden initialization scripts, that wrote diagnostic files to disk during execution, or that could not run multiple instances in the same case — that the DWG published a formal UDM Guideline.

6.2 The Core UDM Requirements

Documentation: Detailed model manual including powerflow setup, tunable parameters, and a high-level control block diagram. Inter-model communication documentation showing data exchange between Plant Power Controller (PPC), WTG, and STATCOM models.

Software Behavior: Parameter error checking; module-dependency enforcement; multiple-instance support; no file writes under default parameters; validation at multiple operating points (100%, 50%, and 10% dispatch).

PSS/E Implementation: DLL-only — no .OBJ or .LIB files. No hard-coded unit ID requirement. No artificial Qmin/Qmax constraints.

6.3 Practical UDM Review Process

Review Dimension	Typical Findings
Documentation completeness	Missing control block diagrams; undocumented parameters
Initialization behavior	Power flow conditions outside the model's stable operating range
Switched-shunt and tap initialization	Missing Python init scripts; wrong starting position
Multi-instance behavior	Crashes when used at two sites simultaneously
TSAT dual-format equivalence	Equipment-name results diverge from bus-number
MQT at full operating range	Acceptable at 100%, unstable at 10% dispatch
Protection-model behavior	Instantaneous protection trips on numerical transients
Headroom toggle	Frequency response not implementable without recompiling

---

Section 7: The Generic vs UDM Decision

7.1 The Case For and Against Generic Models

Generic (WECC standard library) models offer real benefits — they ship with the simulation tool, do not require recompilation, and are well-documented. But ERCOT's October 2019 stakeholder presentation documented that of approximately 27 IRR sites that had migrated from UDM to generic: roughly 80% showed significant differences in dynamic response, many failed Voltage Ride Through, and generic models showed numerical instability at SCR below 2.

7.2 Why Generic Models Fail

WECC generic models encode a simplified abstraction of inverter behavior. The internal current control loop is not modeled. The PLL is not modeled. When a consultant constructs a generic equivalent purely from MOD-026/027 small-signal measurements, the resulting parameter set captures only slow response. Fast transient behavior — exactly the behavior that determines VRT compliance — is essentially guessed.

7.4 Keentel Engineering's Decision Framework

Condition	Recommended Path
Plant has SCR > 5 at POI, OEM supports UDM	Maintain UDM
Plant has SCR between 2 and 5, OEM supports UDM	Maintain UDM; consider REGC_B if generic also kept
Plant has SCR < 2	UDM required; PSCAD studies mandatory
OEM out of business, original UDM unsupported	Carefully tune generic with measured-response validation
MOD-026/027 study uses external consultant only	Validate any proposed generic against original UDM
Plant exhibits oscillatory behavior in operations	PSCAD investigation; do not migrate to generic
Hybrid facility with shared controller	UDM strongly preferred for the combined model

---

Section 8: REGC_A and REGC_B — Choosing the Right Converter Interface Model

8.2 REGC_A — The Current-Source Model

REGC_A represents the generator-to-network interface as a controllable current source. In weak-grid conditions, a small change in injected current produces a large change in voltage, and the network solution can oscillate within an integration step — producing simulated voltage spikes with no physical meaning.

8.3 REGC_B — The Voltage-Source Model

REGC_B uses a voltage-source-behind-impedance interface, parameterized by Re, Xe, and a small firing-delay time constant Te. The Low Voltage Power Logic block of REGC_A is removed entirely; voltage-dependent current limiting is handled upstream via VDL tables.

8.4 Performance Comparison

Short Circuit Ratio	REGC_A Behavior	REGC_B Behavior
~30 (strong grid)	Stable, similar to REGC_B	Stable, similar to REGC_A
~2 (moderate)	Stable but with unrealistic voltage spikes during fault	Stable, no spurious spikes
< 1 (very weak)	Numerically unstable; results meaningless	Converged; results usable

---

Section 9: Common Pitfalls and How to Avoid Them

9.1 Procedural Pitfalls

Pitfall	Consequence	Mitigation
Missing PSCAD in updated submission	Rejection or rework	Submit all models every time, dated
Wrong filename convention	Difficulty tracking, possible rework	Use (SITECODE)_DYNAMIC_(date).zip strictly
Oversized zip > 60MB	Upload fails	Split into multiple zips
No Python init scripts for shunts	Wrong initial conditions, MQT fails	Always provide initialization scripts
Unit count or sizes do not match registration	Direct rejection	Cross-check RIOO-RS Unit Attributes
Different aggregation in PSS/E vs TSAT	Rejection	Match aggregation exactly

9.2 Technical Pitfalls

Pitfall	Consequence	Mitigation
IRR model not set to no-headroom default	Incorrect frequency response, rejection	Default to no-headroom; expose flag for curtailment
PSCAD using PSCAD master library blocks	Inaccurate fault response	Use OEM hardware-code-based PSCAD model
Anti-islanding / PAJ protection left active	Spurious trips, VRT failure	Disable unless OEM confirms necessity
Generic model substituted without OEM endorsement	80% chance of response divergence, VRT failure	Maintain UDM or get OEM-validated generic
Hybrid components submitted separately when coupled	Rejection	Combined MQT for shared transformer or controller
TSAT bus-number and equipment-name results differ	Rejection	Verify dual-format equivalence pre-submission
AVR step test not run during planning	Tuning required at commissioning, model rework	Test AVR/PFR on the model before FIS

9.3 Documentation Pitfalls

Pitfall	Consequence	Mitigation
Verification Report citing only MOD-027 for IBR	Insufficient evidence, rejection	Add OEM attestation and PSCAD comparison
Missing model manual	Reviewer unable to interpret, rejection	Provide manual or PSS/E reference for std lib
No combined MQT overlay plot	Rejection (required since June 2023)	Always overlay PSS/E, TSAT, PSCAD on one axis
Sample test files not included	Reviewer cannot reproduce, rejection	Include all .sav, .dyr, .dll, PSCAD project

---

Section 10: How Keentel Engineering Supports Your Project

10.1 Service Tiers

Tier	Scope	Typical Engagement
Advisory	Strategy, decision frameworks, regulator interface	Discrete consults; 1-4 weeks
Standard	Single-site dynamic model, MQT, verification	8-16 weeks per submission cycle
Comprehensive	Multi-site fleet, MOD-026/027 program management	Ongoing retainer; 6-24 months
EMT Specialty	PSCAD weak-grid, SSR, AGS modeling	Targeted; 4-12 weeks per study
Hybrid Resource Specialty	Wind+solar+ESR or resource+large load combined modeling	12-20 weeks per facility

10.2 What a Typical Keentel Engagement Looks Like

1. Project intake: review of facility one-line, OEM manuals, prior submissions, ERCOT correspondence.
2. Gap analysis: comparison of current model package against the ERCOT Guideline v1.11 checklist.
3. Model build or audit: PSS/E, TSAT, and PSCAD models built or revised in their respective templates.
4. MQT execution: DMVIEW and PMVIEW runs, combined overlay plot generation, anomaly investigation.
5. Verification: parameter-by-parameter comparison with field equipment, OEM attestation gathering, documentation.
6. Submission package assembly: RIOO-RS zip preparation, filename validation, attachment list.
7. Reviewer-correspondence support: response to ERCOT comments, model updates, resubmission.
8. Post-acceptance maintenance: 12-24 month refresh, MOD cycle planning, change-management for field updates.

---

Frequently Asked Questions

Q1. When can I use a generic (WECC standard library) model instead of the OEM's User-Defined Model?

When the OEM has confirmed in writing that the proposed generic parameter set accurately represents their equipment, and you have validated that the generic produces equivalent response across the full MQT test catalog. Generics are most defensible when: (1) the OEM endorses the generic for a brand-new project; (2) the original UDM is no longer maintained; (3) the interconnection SCR is greater than 5. Generics are most dangerous when used as a labor-saving substitute by a third-party MOD-026/027 consultant building from small-signal data alone.

Q2. What is the practical difference between REGC_A and REGC_B, and when should I use each?

REGC_A is a current-source model; REGC_B is a voltage-source model. REGC_A works well in strong-grid conditions but becomes numerically problematic in weak grids — ERCOT's benchmarks show it becoming unstable at SCR below 1, with unrealistic voltage spikes at SCR around 2. REGC_B is well-behaved at low SCR, but is a numerical improvement, not a fidelity improvement. At SCR ≤ 2, vendor-specific PSCAD analysis is still required regardless of which model is used.

Q3. What goes wrong most often during commissioning, and how do I prevent it?

AVR and PFR field tuning during Part 2 of commissioning conflicts with the performance objectives demonstrated in the MQT, forcing model updates and resubmission — typically costing four to twelve weeks. Prevention: during the FIS and QSA stages, run an AVR voltage-step test and a PFR frequency-step test on the model. If either fails in simulation, fix the model before any concrete is poured.

Q4. What is required for a hybrid facility, and how does v1.11 change things?

If dissimilar components share a main power transformer or a master plant controller, a combined MQT and combined model package are required. For wind plus solar: solar-only, wind-only, and combined scenarios. For IBR plus ESR: standalone ESR charging and combined maximum injection. For resource plus large load: large-load-only, maximum withdrawal, and maximum injection (load offline). For AGS-ESR hybrid: three scenario columns across AGS-online-only, non-AGS-online-only, and both-online.

Q5. How often do I have to submit a Verification Report, and what does ERCOT actually want to see in one?

Required at four points: within 30 days of Part 3 approval; between 12 and 24 months after; every 10 years minimum; and within 30 days of any dynamic setting change. ERCOT wants: a narrative describing how and when parameters were checked, a table comparing model parameters with field settings in matching engineering units, and supporting evidence (screenshots preferred, OEM attestations accepted).

Q6. What is the no-headroom default, and why does ERCOT insist on it?

ERCOT requires IRR models to default to no-headroom because when PGEN is less than PMAX for a renewable plant, it is because of insufficient wind or solar — not curtailment. The plant has no fuel reserve to draw from in a low-frequency event. If a model treats the PGEN-to-PMAX gap as headroom by default, every system study will over-estimate available primary frequency response — lulling operators into trusting reserves that do not exist.

Q7. Can I just use one PSCAD model across all my similar sites, or do I need a unique model per facility?

The same underlying PSCAD model with site-specific parameterization. Site-specific items include main power transformer impedance and rating, collector system equivalent impedance, number and rating of inverters, plant-controller gain and droop settings, and protection thresholds. A developer with eight sites using the same inverter platform should maintain eight separate PSCAD project files — each with site-specific parameters — but only one underlying inverter component model.

Q8. How do I demonstrate that my IBR can ride through a weak-grid event?

Four artifacts: (1) a PSCAD model built from OEM hardware code including the inner-current-control loop and PLL implementation; (2) an SCR MQT test demonstrating stable behavior from SCR = 8 down to SCR = 1.2; (3) a Phase Angle Jump test in PSCAD; (4) where SCR is at or below 2, a vendor-specific facility study confirming inner-control-loop bandwidth coordination with the PLL. Common failures include PLL loss-of-synchronism, phase-angle-jump, or anti-islanding protection left active.

Q9. What do I have to submit when I change a setting in the field, and what is the timeline?

Before making a field change at an operational IBR: submit a pre-change MQT overlay comparing proposed vs. as-built model response (both PSS/E and PSCAD), with email subject 'IBR Proposed Modification' to dynamicmodels@ercot.com. After the change is approved, within 30 days submit final PSS/E, PSCAD, and TSAT models to RIOO-RS with a combined MQT overlay and updated Verification Report. For synchronous facilities, the pre-change MQT comparison is not required.

Q10. We are a small developer / a first-time Resource Entity. What is the realistic timeline and cost?

A standalone 200 MW solar facility using vendor-supported UDMs typically requires 8 to 12 weeks from FIS through Part 3 commissioning. A hybrid facility with shared main transformer typically requires 14 to 20 weeks. The first-time risk factor that disproportionately impacts new developers is OEM responsiveness. Standard engagements range from the mid five figures to the low six figures USD for a single site. The largest hidden cost is not the modeling work — it is the schedule cost of an avoidable model rejection deferring commercial operations by months.

---

Case Study 1: Generic Model Substitution Triggers VRT Failure

Client Profile

A 300 MW solar PV facility in the West Texas region, commissioned in 2017. Originally interconnected with a vendor-supplied UDM in PSS/E and a corresponding TSAT model.

The Triggering Event

The Resource Entity engaged a third-party MOD-026/027 consultant for the 10-year compliance cycle. The consultant proposed replacing the UDM with a WECC generic model set (REGC_A, REEC_A, REPC_A) on the rationale of reduced ongoing maintenance burden. The generic parameters were derived from MOD-026/027 small-signal test data and submitted to ERCOT as the updated dynamic model.

What ERCOT Found

The Voltage Step Down 3% test showed reactive injection only 40 percent of the UDM magnitude, with response time four times slower. The HVRT Leading test showed the generic tripping at approximately 1.18 per unit voltage, while the original UDM rode through beyond 1.20 per unit. The SCR test showed the generic becoming oscillatory at SCR = 1.8, while the original UDM remained stable down to SCR = 1.2. ERCOT rejected the submission.

Resolution

Keentel Engineering reverted to the original UDM, recompiled for PSS/E v35 and v36, supplemented with a refreshed TSAT model and a PSCAD model built from OEM hardware code. Every test passed, including SCR = 1.2 stability. The resubmission was accepted on first review — 9 weeks from Keentel engagement to ERCOT acceptance. The model was the problem, not the plant.

---

Case Study 2: Weak-Grid Oscillation at a Panhandle Wind Plant

Client Profile

A 350 MW wind facility in the ERCOT Panhandle region, interconnected at a 345 kV point with a Short Circuit Ratio of approximately 1.6 under the most-binding contingency.

The Issue and Investigation

During QSA-stage stability assessment, ERCOT observed sustained ~3 Hz oscillations in PSS/E simulations. Switching to REGC_B dampened the oscillations materially but did not eliminate them. A PSCAD case using OEM source-code-based models showed an initial transient that damped within ~400 milliseconds, with real residual oscillation. Frequency-domain analysis identified the residual oscillation as PLL interaction at a specific bandwidth, and the OEM confirmed that enabling a firmware 'weak-grid option' setting would eliminate it.

Resolution

The OEM committed to enabling the weak-grid option at commissioning. Keentel Engineering rebuilt the PSCAD model with that setting active and produced a combined PSS/E (REGC_B), PSCAD, and TSAT overlay showing damped, stable response. ERCOT accepted the QSA submission and the plant proceeded to construction on its original schedule.

---

Case Study 3: A Hybrid Solar-Plus-Storage-Plus-Large-Load Facility

Client Profile

A first-of-kind hybrid facility in South Texas: 250 MW solar PV, 100 MW / 400 MWh battery energy storage (AGS / Grid Forming), and a co-located 75 MW electrolyzer load for green hydrogen production — all sharing a single 345 kV main power transformer and a single master plant controller.

The Modeling Challenge

Because all components share both a main transformer and a master controller, the v1.11 Guideline requires a combined MQT and combined model submission. Three distinct OEMs — solar inverters, battery inverters, and electrolyzer power conversion — each with its own UDM and PSCAD model — had to be integrated into a single combined model. Total individual simulations: 142 across PSS/E, TSAT, and PSCAD.

Approach and Resolution

Keentel Engineering convened a coordination workshop with all three OEMs, built combined PSS/E and TSAT models over six weeks, then spent eight weeks building the combined PSCAD model — requiring OEM license negotiations and resolution of cross-vendor Fortran compilation conflicts. The final Combined Model Submittal Package included a 380-page joint MQT report, three Unit Model Validation reports, and a master plant controller manual. ERCOT accepted the package after one round of clarifying questions.

---

Appendix: Quick-Reference Submission Checklist

A.1 PSS/E and TSAT Package

• PSS/E v35 model (.raw or .sav, .dyr, .dll for UDM).
• PSS/E v36 model (.raw or .sav, .dyr, .dll for UDM).
• PSS/E model in Excel Template format (Universal template for UDMs).
• TSAT model with .raw, .dyr, .dll, .tudm files (if UDM).
• TSAT bus-number example case and equipment-name example case (with space character).
• TSAT .pfb file plus .mon, .dat, .swi files.
• Python initialization scripts for switched shunts and on-load transformer taps.
• Model manual or PSS/E manual reference for standard library models.
• Unit count, size, and names match RIOO-RS registration exactly.
• IRR models default to no-headroom; headroom flag exposed.

A.2 PSCAD Package (IBRs)

• PSCAD model embedded in ERCOT PSCAD Template.
• Completed PSCAD Guideline Checksheet.
• PSCAD model manual.
• 32-bit and 64-bit Fortran libraries (where applicable).
• Unit Model Validation (Hardware Benchmark) report — required if commissioned after 3/1/2021.

A.3 Model Quality Test (MQT) Report

• Combined overlay plot of PSS/E, PSCAD, and TSAT responses (TSAT if UDM).
• All required test scenarios per facility type.
• Test case files (.sav/.raw, .dyr, .dll, PSCAD project).
• Honest disclosure of any test that fails, with corrective-action plan.

A.4 Verification Report (Or Template Worksheet)

• Narrative describing how and when parameters were checked.
• Parameter-by-parameter comparison table with matching units.
• Verification method noted for each parameter.
• Supporting evidence (screenshots preferred, OEM attestation accepted).
• Protection model verification, including PSS/E protection if PSCAD is primary model.

A.5 RIOO-RS Submission Mechanics

• Files named: (SITECODE)_DYNAMIC_(YYYY-MM-DD).zip and (SITECODE)_PSCAD_(YYYY-MM-DD).zip.
• Zip files under approximately 60 MB each (split if larger).
• Attachment-Only Change Request in RIOO-RS.
• All models submitted — even unchanged ones — with dates indicating which are new.
• TSP notified of the change.