Propulsion Credibility Board - Quarterly Review Minutes
Date: April 9, 2026
Chair: Dr. A. Patel
Program: Advanced Turbine Engine Program
Subject: HPT Blade CHT Analysis - Cruise Steady-State COU (COU2)

Attendees: Dr. A. Patel (Chair), J. Kim (CFD Lead), R. Fernandez (Test Lead),
           S. Thompson (Independent Reviewer), M. Chen (Configuration Mgmt),
           K. Nakamura (Blade Life Engineering — attending as stakeholder)

Distribution: Chief Engineer, Program Manager, Propulsion Chief Engineer,
              Blade Life Engineering

---

Item 1 - Purpose of Review

K. Nakamura requested cruise-condition peak temperature and creep-damage
predictions from the CHT model to support upcoming certification margin
analysis for the 25,000-hour blade life target. The model was accepted
at MRL 3 for take-off transient screening (COU1, March 2026). This review
assesses credibility for the new cruise COU at the higher MRL 4 required
by certification-facing analysis.

Item 2 - Scope Clarification

Dr. A. Patel opened by clarifying scope: this is NOT a re-review of COU1.
COU1 remains accepted for its defined use. This is a NEW COU assessment
for cruise operating conditions at MRL 4. All COU1 V&V evidence remains
valid; the question is whether that evidence transfers to cruise.

Item 3 - Model Form and Numerics Transfer

Board concurs: verification factors (SQA, NCV, discretization at mid-span,
solver error, use error) transfer cleanly. Same code, same numerics, same
mesh, same turbulence model. No re-verification required.

Item 4 - Cascade Validation Transfer (CENTRAL ISSUE)

S. Thompson (Independent Reviewer) raised the central concern:

Quote: "We validated the model at take-off Reynolds number 1.26e6 using a
cascade rig designed to match that condition. We're now applying the model
at cruise Reynolds 0.85e6 — 41 percent below the cascade rig Reynolds.
We've moved from 4.8 percent mismatch (acceptable) to 41 percent mismatch
(not acceptable). The cascade does not constrain model performance at
cruise."

J. Kim response: "The cascade-derived error distribution (1.8 percent mean,
4.2 percent max) is assumed to bound cruise error."

S. Thompson: "That assumption is not validated. We have no measurements at
cruise Re to confirm it, and the flow is in a different regime."

Board discussion: the issue is structural, not a matter of error-bar
magnitude. At cascade Re the flow is fully turbulent throughout. At cruise
Re the flow is in the transitional-turbulent regime where Nu-Re scaling
exponent changes from approximately 0.8 to 0.7. Extrapolating the cascade
error distribution across a regime change is not defensible.

Item 5 - Status Determination for Affected Factors

K. Nakamura asked how many separate factors are affected and what status
they should be assigned.

S. Thompson proposed that five factors cannot be legitimately assessed at
cruise operating conditions with the existing evidence package:

  (a) Factor 2.1 Model Form: model form validated only at take-off Re
  (b) Factor 2.4 Test Conditions: no cruise-Re test conditions exist
  (c) Factor 3.1 Equivalency of Inputs: cannot establish equivalency
      without cruise validation experiment
  (d) Factor 3.2 Output Comparison: no cruise output measurements exist
  (e) Factor 4.2 Relevance of Validation to COU: cascade outside cruise
      operating envelope

J. Kim asked whether these should be recorded as Level 1 achieved (low
assessment) or Not Assessed (no valid assessment).

S. Thompson: "Level 1 would misrepresent the situation. A Level 1 rating
implies an assessment was performed with a low-credibility outcome. What
we actually have is no valid assessment for the cruise COU. Five factors
are Not Assessed pending cruise-regime validation."

Dr. A. Patel concurred: "This is the honest characterization. We are not
saying the model is bad at cruise — we are saying we do not know, because
we have not done the assessment at cruise conditions. Not Assessed is the
correct status. Recording Level 1 would invite the wrong follow-up
questions."

Board voted: five factors recorded as Not Assessed for cruise COU. The
factors may be assessed in a future review once cruise-regime validation
evidence is available.

Item 6 - Discussion on Path Forward

Two options considered:

Option A (SHORT-TERM, NOT RECOMMENDED): Retain the factor assessments at
nominal Level 1 with explicit error bar expansion. Board rejected:
expanding error bars to cover unvalidated regimes is not a credibility
argument. It also misrepresents Not Assessed as Assessed.

Option B (RECOMMENDED): Record the five factors as Not Assessed for cruise,
decline to accept the model at MRL 4, and initiate one of:
  (B1) Cascade rig re-instrumentation for cruise Reynolds matching
       (estimated 9-month campaign, 4 engineers)
  (B2) Engine ground-test instrumented run at cruise power with surface
       thermocouples or thermal paint (estimated 4-month effort but
       requires test engine availability)
  (B3) Scaled rig with matched Re and Nu ranges (estimated 12+ months)

K. Nakamura noted that rejecting the model creates schedule pressure on
certification. Dr. A. Patel responded: schedule pressure is not a
credibility argument. The model must earn credibility through evidence,
not through decision necessity.

Item 7 - Decision

NOT ACCEPTED at MRL 4 for cruise creep-life certification analysis.

Board determines that the existing validation evidence package does not
support the COU at the required maturity level. The systematic
applicability gap manifesting as five Not Assessed factors (2.1, 2.4,
3.1, 3.2, 4.2) must be addressed through new cruise-regime validation
evidence.

The model MAY continue to be used for preliminary comparative studies at
cruise (delta predictions, parametric trends) with explicit
non-certification caveats.

Item 8 - Actions

ACTION 8.1 (J. Kim): Update credibility narrative to record the five
affected factors as Not Assessed for cruise COU. Document the single
root cause in narrative §6.

ACTION 8.2 (R. Fernandez): Evaluate feasibility and cost of options B1,
B2, B3 above. Report to board within 30 days.

ACTION 8.3 (K. Nakamura): Identify certification schedule impact and
alternative life analysis approaches in the interim (e.g., conservative
bounding analysis, prior-generation engine data extrapolation).

Item 9 - Record

This assessment is to be documented in the aero-evidence-cou2 evidence
package and archived with a NOT-ACCEPTED decision record. The assessment
is a legitimate credibility-board output, not a failure of the model.
Five Not Assessed factors accurately represent the evidence state.

Next Review: Upon completion of cruise validation campaign (est. Q3 2026
earliest).

Minutes approved: Dr. A. Patel, April 10, 2026
