This study examines why the fine-grid Transition SST case preserves the Run 145 pressure comparison while producing a substantially different thermal field from the SST baseline. Phase A diagnoses the accepted Transition SST solution; Phase B changes the inlet turbulence state one parameter at a time while keeping the mesh, thermal boundary conditions, numerics and model settings fixed.
The case levels are controlled diagnostic perturbations selected independently of the NASA comparison error.
All sensitivity cases restart independently from the accepted Transition SST state at iteration 556.
| Case | Tu_in |
mu_t/mu_in |
Purpose |
|---|---|---|---|
baseline_tu065_vr10 |
6.5% | 10 | accepted baseline |
tu065_vr05 |
6.5% | 5 | viscosity-ratio sensitivity |
tu065_vr01 |
6.5% | 1 | viscosity-ratio sensitivity |
tu083_vr10 |
8.3% | 10 | inlet-intensity sensitivity |
The accepted baseline also uses inlet intermittency 1.0, the built-in Transition SST correlations, and no roughness or curvature correction. Detailed settings are in references/model_inputs/transition_sst_settings.csv.
The computational inlet is about 58.605 mm upstream of the geometric leading edge. The saved inlet faces contain the prescribed Tu = 6.5%, but the local turbulence level decays strongly before the vane.
For the upstream diagnostic, fluid cells ahead of the leading edge and more than 8 mm from a wall are grouped in axial bins. Turbulence intensity is reconstructed from the saved fields as
Tu = 100 sqrt(2 k / 3) / |U|.
In the 2–5 mm bin immediately upstream of the leading edge, the accepted baseline gives:
- median
Tu = 1.2473%; - median
mu_t/mu = 7.851; - median
Re_theta_t = 412.933.
The full decay curve is in baseline_freestream_decay.csv.
The vane wall is traversed from leading edge to trailing edge on the pressure and suction sides. In the baseline case, the strongest concurrent suction-side intermittency and wall-temperature gradients occur near x/Cx = 0.6534, with the wall-shear gradient near 0.6551.
This location is used only as a transition-like response front in the CFD solution. It is not treated as an experimentally measured transition onset. The extracted thresholds and gradient locations are in baseline_transition_signature.csv.
Reducing the inlet turbulent-viscosity ratio produces much stronger freestream decay before the vane and a large thermal response.
mu_t/mu_in |
Near-LE median Tu |
Mean wall temperature | External heat rate | Outlet Mach | Suction-side response |
|---|---|---|---|---|---|
| 10 | 1.2473% |
608.879 K |
28.5483 kW/m |
0.903351 |
about 0.65 x/Cx |
| 5 | 0.8701% |
604.438 K |
27.7758 kW/m |
0.903562 |
about 0.69 x/Cx |
| 1 | 0.3637% |
574.705 K |
22.9198 kW/m |
0.905424 |
trailing-edge region |
The ratio-1 case moves the intermittency and wall-shear response to about x/Cx = 0.962–0.968. For that case, the interior gradient search is restricted to x/Cx < 0.98 so the geometric trailing edge does not dominate the diagnostic.
The three-point result is summarised in b1_three_point_summary.csv; detailed case outputs remain in the corresponding tu065_vr05_*, tu065_vr01_* and comparison CSVs.
NASA-CR-168015 reports Tu as the average inlet turbulence intensity for the C3X cascade. The combustor-induced inlet level was 6.5%, measured with laser Doppler anemometry (LDA), and upstream rods increased it to 8.3%.
These values are therefore treated as inlet-level conditions, not leading-edge targets. The exact axial correspondence between the experimental measurement plane and this reduced computational inlet is not assumed.
Changing the prescribed inlet turbulence intensity from 6.5% to 8.3% produces a very different result from the viscosity-ratio sweep:
Tu_in |
Near-LE median Tu |
Mean wall temperature | External heat rate | Outlet Mach |
|---|---|---|---|---|
| 6.5% | 1.2473% |
608.879 K |
28.5483 kW/m |
0.903351 |
| 8.3% | 1.2370% |
608.678 K |
28.5133 kW/m |
0.903358 |
The two decay curves differ farther upstream but approach and cross before the vane. By the final 2–5 mm bin, the near-vane turbulence levels are almost the same. Mean wall temperature changes by only -0.033%, external heat rate by -0.122%, and the suction-side response shifts by one wall-face station.
Detailed outputs are in tu083_vr10_freestream_decay.csv, tu083_vr10_transition_signature.csv, tu083_vr10_global_checks.csv, tu083_vr10_integral_summary.csv, tu065_vs_tu083_vr10_diagnostic_summary.csv and b2_two_point_summary.csv.
The sensitivity runs use the same 20-iteration report-definition criterion as the accepted case: external heat rate, mean wall temperature and outlet Mach must each remain within a 0.02% maximum relative-change window, with continuity below 1e-3. Mass balance, fluid-solid interface heat mismatch and wall y+ are checked at closure; the retained cases satisfy the study limits.
For tu065_vr01, the final CFF archive does not contain a persisted cooling-hole total Flux Report, so a final solid heat-imbalance value is not reconstructed from that file. Later cases explicitly save the required Flux Reports at closure. The detailed checks remain in the case-specific *_global_checks.csv files.
At fixed Tu_in = 6.5%, changing
mu_t/mu_in: 10 → 5 → 1
changes the near-leading-edge turbulence roughly as
Tu_2-5mm: 1.247% → 0.870% → 0.364%
and produces large changes in the thermal field and transition-like wall response while outlet Mach moves comparatively little.
By contrast, at fixed mu_t/mu_in = 10, changing the documented inlet level from 6.5% to 8.3% leaves the near-vane turbulence and thermal solution almost unchanged.
For the tested setup, inlet-to-vane turbulence decay is therefore strongly sensitive to turbulent-viscosity ratio. The study does not identify an optimal ratio, and no coarse or medium Transition SST cases were run, so the response-front locations are treated as fine-grid model diagnostics rather than mesh-independent transition predictions.