ERCOT Model Quality Tests (MQT): What Engineers Need to Know

ERCOT Model Quality Tests (MQT): What Engineers Need to Know

Based on ERCOT DWG Procedure Manual, Section 3.1.5

Accurate dynamic models are now a baseline requirement for connecting any inverter-based resource (IBR), wind or solar facility, synchronous generator, STATCOM/SVC, or advanced storage plant in ERCOT. Before a model can be accepted into planning studies, ERCOT requires the full Model Quality Test (MQT) suite.

1. What is the Model Quality Test (MQT)?

ERCOT relies on dynamic simulations to understand how the system behaves during faults, voltage dips, frequency events, and weak‑grid conditions. Those simulations are only as good as the models behind them.

The Model Quality Test (MQT) is a standardized set of simulations that every dynamic model must pass before it’s accepted into ERCOT’s planning and operations studies. Resource owners are required (via ERCOT Planning Guide Section 6.2) to:

• Run a defined list of tests on their PSS®E, and PSCAD models, and
• Submit simulation cases + plots showing the response for each test.

The goal is simple: prove that your model behaves like a real, grid‑connected facility over a wide range of credible disturbances.

 Below is a technical summary of the exact tests, expectations, and performance criteria defined in Section 3.1.5 of the ERCOT DWG Procedure Manual.

2. Simulation Framework

All MQT tests use a standardized testbench:

• Resource → POI → controllable infinite bus
• Voltage and frequency can be stepped or ramped
• POI initialized at 1.0 p.u. unless test requires otherwise
• Full real-power dispatch (and full charging for storage)
• All step-up transformers, collector impedance, plant controllers, and reactive devices must be included
• Minimum simulation time: 20 seconds, with stable post-disturbance settling

This setup ensures the model is evaluated independently of ERCOT system topology.

3. Flat Start Test

Objective: Verify numerical stability and correct initialization.

Criteria:

• MW, Mvar, voltage, and frequency remain essentially flat
• No slow drifts, hidden limit interactions, or internal oscillations

This test exposes initialization or controller-interaction problems before larger disturbances.

4. Small Voltage Disturbance Test (+3% / –3%)

What ERCOT checks:

• AVR or plant controller must immediately move toward:
• Lagging (reactive injection) for voltage step down
• Leading (reactive absorption) for voltage step up
• Well-damped behavior; no overshoot or oscillatory instability
• Real power should remain constant

Unacceptable: no AVR action, incorrect Q direction, or tripping.

5. LVRT Test (Low Voltage Ride-Through)

Required for IBRs, WGRs, IBTEs, and AGS-ESR.

ERCOT requires testing against two voltage profiles:

1.    Legacy LVRT curve

2.    Voltage-dip profile (multiple dips, each returning to 1.0 p.u.)

Tests must be run twice:

• nitial 0.95 lagging
• Initial 0.95 leading

Performance requirements:

• Strong reactive injection during voltage recovery
• Active current injection for dips ≥ 0.5 p.u.
• Real power must begin recovering before voltage reaches 0.9 p.u.
• Full MW recovery within 1.0 second after voltage returns to 0.9–1.0 p.u.
• No momentary cessation unless allowed by exemption

Any model that absorbs reactive power during recovery is considered incorrect.

6. HVRT Test (High Voltage Ride-Through)

Also required for IBRs, WGRs, IBTEs, AGS-ESR.

Profiles:

• First test the Preferred HVRT curve
• If failed → test Legacy HVRT as well

Performance expectations:

• Fast transition into deep reactive absorption within ~0.5 seconds
• At sustained 1.1 p.u., facility should operate near maximum leading capability
• Real power should remain essentially constant (≤5% reduction allowed)
• Full recovery when voltage returns to 0.95–1.05 p.u.

Weak reactive response during the high-voltage period is unacceptable.

7. Small Frequency Disturbance Test (±0.3 Hz)

Required for all technologies.

Setup:

• Resource dispatched at 80% of maximum
• For storage → also test at 80% charging
• For wind/solar → two cases:
• Curtailed (with headroom)
• Power-availability state (no headroom)

Criteria:

• Real power must follow droop/deadband
• If no headroom → no false real-power increase
• IRRs must demonstrate correct behavior in both curtailed and uncurtailed modes

This test ensures the frequency response implementation is realistic.

8. System Strength (SCR) Test

Required for IBRs and IBTEs.

SCR values to be tested:

• 5
• 3
• 1.5
• 1.2

Method:

• Change Thevenin impedance during simulation
• Apply a 4-cycle three-phase fault before each SCR transition

Acceptance thresholds:

• Stable and well-damped at SCR ≥ 3 (mandatory)
• SCR = 1.5 acceptable with explanation if unstable
• SCR = 1.2 optional stability (informational)

This test exposes issues in current-limited control, PLL behavior, and voltage-control loops.

9. Phase Angle Jump Test (PSCAD only)

Required only for PSCAD models of inverter-based resources.

Purpose: Evaluate stability when subjected to sudden voltage-angle steps (e.g., ±10°, ±25°).

Criteria:

• Active power must oppose the phase jump (increase when angle decreases, and vice-versa)
• Adequate peak response (≥0.2 p.u. for ±10°, ≥0.5 p.u. for ±25° if within current limits)
• No tripping; stable recovery

This verifies PLL/inner-loop performance during severe phase disturbances.

10. AGS-ESR Additional Tests

Advanced Grid Support Energy Storage Resources must also run:

• AGS-specific small voltage disturbance test
• AGS frequency-change + inertia estimation test (H ≥ 2.5 s)
• AGS system strength tests (SCR 10 → 1.2)
• AGS phase-angle jump test
• AGS loss-of-synchronous-machine test

These tests ensure grid-forming or grid-supporting capabilities.

Why This Matters

MQT has evolved into a strict compliance and model-quality requirement in ERCOT. The results directly affect:

• Interconnection approval
• Accuracy of grid studies (LVRT, HVRT, N-1-1, voltage stability, etc.)
• How the resource behaves in weak-grid regions
• Integration of large clusters (solar, wind, BESS, STATCOMs)

A model that fails any of the above tests can be rejected, forcing re-tuning or re-submission. For developers and consultants, understanding these tests is essential to avoid delays and ensure correct dynamic behavior.