Propulsion Credibility Board - Quarterly Review Minutes
Date: March 18, 2026
Chair: Dr. A. Patel
Program: Advanced Turbine Engine Program
Subject: HPT Blade CHT Analysis - Take-off Transient COU (COU1)

Attendees: Dr. A. Patel (Chair), J. Kim (CFD Lead), R. Fernandez (Test Lead),
           S. Thompson (Independent Reviewer), M. Chen (Configuration Mgmt)

Distribution: Chief Engineer, Program Manager, Propulsion Chief Engineer

---

Item 1 - Verification

MMS benchmarks: all 5 conjugate heat transfer coupling cases passed with errors
below 0.01 percent. Board concurs: Level 3 achieved on Numerical Code Verification.

SQA: ANSYS CFX 2025 R1 has ISO 9001 certification and vendor regression suite.
Board concurs: Level 2 achieved on Software Quality Assurance.

Item 2 - Cascade Validation

48-point TC rake at cascade exit: mean error 1.8 percent, max 4.2 percent at
endwall station 18 (span 36.2 percent). Hot spot location within 2mm of thermal
paint indication. Board noted: max error at endwall NOT at tip; cascade rig does
not represent engine tip conditions but acceptable for take-off screening.

R. Fernandez: cascade Re = 1.20e6, engine Re = 1.26e6 (4.8% delta). Within
band for take-off representativeness. Cascade was NOT designed for cruise
or off-design conditions; future COUs requiring those regimes will need
dedicated validation.

Item 3 - Mesh Convergence

GCI 0.8 percent at mid-span using refinement ratio r = 1.26 (cbrt(2)). Board
noted refinement ratio falls below Celik et al. 2008 / Roache recommended
r >= 1.3 threshold. J. Kim to document caveat in final report; board accepts
for MRL 3 given fine mesh is above the asymptotic range demonstrated by
monotonic convergence.

Board FLAGGED: tip region mesh convergence NOT assessed. Peak temperatures
during take-off occur at the blade tip. This is a significant gap.

ACTION 3.1 (J. Kim): Complete tip region mesh convergence study with
r = 1.30 refinement ratio before MRL 4 readiness review.

Item 4 - UQ Status

Monte Carlo uncertainty propagation completed for cascade boundary condition
inputs (15 uncertain parameters, 500 samples, LHS). NOT extended to engine
COU. Probabilistic UQ on peak temperature prediction is REQUIRED at MRL 3
per NASA-STD-7009B Factor 5.4 but NOT completed.

ACTION 4.1 (J. Kim): Complete probabilistic UQ (Monte Carlo, min 1000
samples) on engine COU peak temperature before MRL 4 readiness.

Board accepted status with condition. Cascade UQ demonstrates methodology
works; extension to engine COU is a pedigree extension, not a methodology
development.

Item 5 - Film Cooling and Cascade Applicability

S. Thompson (Independent Reviewer) raised concern: cascade rig is NOT film
cooled, but engine blade has 47 film cooling holes (showerhead + pressure
side + suction side rows). Cascade validation therefore does not constrain
the film cooling model performance. Narrative §4.3 acknowledges this but
it is a material applicability gap for QoI (peak metal temperature) where
film cooling effectiveness dominates the local hot-spot prediction.

ACTION 5.1 (J. Kim): Document film cooling validation status explicitly in
credibility narrative. Identify which QoI prediction locations are most
sensitive to film cooling model uncertainty.

ACTION 5.2 (R. Fernandez): Propose cascade upgrade (film cooled) or
alternate validation path (engine ground test PT locations) for MRL 4.

Item 6 - Downstream Gate and Use Constraint

Board clarified scope of Model Acceptance:

ACCEPTED for PRELIMINARY SCREENING at MRL 3. Specifically:
 - Blade designs predicted to exceed the 1150K limit by more than 50K are
   REJECTED at concept review (no physical test).
 - Blade designs predicted to be within +/-50K of the 1150K limit are
   FLAGGED as marginal and MUST proceed to engine ground test before any
   design-freeze or certification-track activity.
 - Model output is NOT used for final blade life assessment, creep margin
   sign-off, or any certification-facing analysis at this MRL.

Decision consequence thus limited to concept-stage down-selection. Any
expansion of use requires re-assessment at the new MRL.

Item 7 - Numerical Solver Error

J. Kim reported numerical solver error is bounded within the discretization
error budget via residuals of 4.2e-7 to 8.8e-7 (target 1e-6) and double
precision arithmetic. Board accepts this as a partial assessment of Factor
5.1.2 at Level 1 rather than full scoping-out. Language in the narrative
has been updated accordingly (see Item 8).

Item 8 - Editorial

ACTION 8.1 (M. Chen): Update credibility narrative language for Factor 5.1.2
to read "bounded within numerical error budget via residual history" rather
than "scoped out."

Item 9 - Decision

ACCEPTED for preliminary screening at MRL 3, with conditions per Action
items 3.1, 4.1, 5.1, 5.2.

Model use bounded to concept-stage down-selection per Item 6.

Next Review: June 2026 (MRL 4 readiness review). All actions must be closed
by May 31 2026 to support June review.

Minutes approved: Dr. A. Patel, March 20, 2026
