NASA-CR-168015 reports an estimated ±3% uncertainty for the radial
cooling-hole heat-transfer-coefficient calculation used in the experimental
thermal reduction. I already had Fluent sensitivity cases at h/h0 = 0.95,
1.00 and 1.05, so I used those solved cases to estimate the response to a
±3% change instead of running another pair of cases.
The source value is transcribed in
references/experimental_data/c3x_experimental_uncertainty_summary.csv
from Hylton et al., NASA-CR-168015 (1983).
For each response Y, I take the slope through the two central sensitivity
cases:
dY / d(+1% h) = [Y(1.05 h0) - Y(0.95 h0)] / 10
and apply it around the baseline:
Y(h0 ± 3%) = Y(h0) ± 3 * dY/d(+1% h)
The existing five-point h sweep is close to linear over
h/h0 = 0.90–1.10 (R2 is about 0.999 for the main thermal responses),
so this local interpolation is adequate for the small ±3% step. The table is
rebuilt by
scripts/postprocess/build_nasa_informed_cooling_envelope.py
and written to
nasa_informed_internal_htc_envelope.csv.
| Quantity | Baseline | h - 3% |
h + 3% |
Half-width |
|---|---|---|---|---|
| Mean external wall temperature | 655.619 K |
657.354 K |
653.884 K |
1.735 K |
| External heat-transfer rate | 35.820 kW/m |
35.267 kW/m |
36.372 kW/m |
0.552 kW/m |
| Wall-temperature bias, pressure side | +8.887 K |
+10.947 K |
+6.827 K |
2.060 K |
| Wall-temperature bias, suction side | +12.999 K |
+14.582 K |
+11.416 K |
1.583 K |
| Wall-temperature MAPE, pressure side | 1.448% |
1.760% |
1.136% |
0.312 percentage point |
| Wall-temperature MAPE, suction side | 2.005% |
2.243% |
1.766% |
0.238 percentage point |
The wall-temperature bias stays positive on both surfaces across this range, so
the ±3% internal-HTC change does not remove the SST/NASA temperature
difference.
This is an interpolation of the existing h sensitivity study, not a new
Fluent run. The Tbulk ±5 K and ±10 K cases remain separate sensitivity
cases because the retained NASA record does not give an equivalent quantitative
uncertainty for those coolant temperatures. NASA Table VI already supplies the
experimental external-HTC uncertainty used for the comparison error bars, so the
internal ±3% term is not added to those intervals again.