The Poll Schumann model requires airframe and engine parameters, but published tables cover only part of the fleet. Applications must obtain missing parameters, substitute another aircraft type or exclude unrepresented types. We reconstruct parameters using geometry from certification documents and manufacturer manuals, together with ICAO engine records. Published equations supply the physical parameters, regression estimates the propulsion efficiency scale. The inventory covers 90 type designators, including 22 absent from the direct pycontrails table and 18 without a listed substitute. Across 68 types, mean absolute percentage error (MAPE) against pycontrails was 7.38% for drag and 3.14% for design Mach. Across 66 types in 21 families, efficiency regression achieved 3.91% MAPE with the evaluated family excluded from input selection and fitting. On 100 recorded A320 flights, a type already represented in the published tables, cruise fuel flow MAPE for reconstructed and published parameter sets was 4.50% versus 4.68%, while mean absolute percentage error in integrated fuel over covered airborne intervals was 7.19% versus 6.82%. Public technical records can therefore extend Poll Schumann parameter coverage, with quantified agreement against published coefficients and comparable fuel prediction errors on the evaluated A320 flights.
Figures & tables
Symbol
Physical role
ψ0
Zero-lift drag scale
ψ1
Range scale, including propulsion efficiency
ψ2,ψ3
Lift and lift-to-drag scales, respectively
ψ4=MDO
Design-optimum Mach number
ψ5,ψ6
Reynolds-number and lift scales, respectively
ψ7
Optimum-mass scale at the tropopause
TABLE I: Main dimensionless parameters and their physical roles [ 6 , 7 ] . Parameters are calculated from public inputs using the model equations, except η1 , which is fitted, and ψ1 , which also depends on that fit. DO denotes the design optimum.
Fig. 1: Inputs to the reconstruction. Solid arrows show the calculations and their inputs or outputs. Dashed arrows carry published values used for fitting or comparison. The regression estimates the efficiency parameter η1 from public predictors, using published efficiency values as training labels.
Inventory or baseline comparison
Count
ICAO types
90
Product families
82
Available airframe variants
260
Complete parameter records
928
Unresolved parameter records
2
TABLE II: Parameter availability and comparison with the pinned pycontrails PS tables. Available variants distinguish airframe versions; complete records link variants to engine entries.
ICAO
Product family
ψ0
ψ2
ψ3
ψ4
ψ5
ψ6
ψ7
τ
η1
η2
A19N †
A319neo
7.70
8.25
0.530
0.712
5.96
0.758
0.881
0.179
0.361
0.404
A20N
A320
7.81
8.22
0.522
0.715
5.98
0.779
0.860
0.181
0.361
0.389
A21N
A321neo
8.41
8.23
0.484
0.725
6.07
0.914
0.752
0.190
0.368
0.387
A306
A300-600
8.73
8.40
0.465
0.768
9.36
0.693
0.872
0.149
0.358
0.534
A30B
A300B2-B4
8.07
8.10
0.483
0.783
9.53
0.641
0.881
0.139
0.353
0.552
A310
A310-200
9.17
8.02
0.421
0.787
8.79
0.658
0.957
0.176
0.345
0.536
TABLE III: Characteristic parameters for all 90 types: componentwise medians over available installation records, which need not describe a single installation. Definitions: Sections III-B – III-C . All values are dimensionless; ψ5 is divided by 107 . Types absent from the direct pycontrails table: † has a substitute; ‡ has none.
Parameter
MAPE
Bias
Max.
pycontrails, 68 types
CL,DO
5.56
+4.74
24.68
η2
0.62
+0.09
4.75
J1
2.62
-2.39
8.47
J2
0.23
-0.19
0.97
MEC
0.37
-0.10
2.22
TABLE IV: Agreement of type-average reconstructed parameters with direct references. Errors are percentages, with equal weight per aircraft type. Configuration differences remain. Max. denotes the largest absolute percentage error. Fitted quantities are shown separately as descriptive agreement.
Method
MAE
MAPE (%)
Bias (%)
Distinct input combinations (primary)
Nested selection
0.01373
3.91
-0.21
Global selection
0.01170
3.40
+0.30
Fixed formula
0.01170
3.40
+0.30
Training-only mean
0.03839
11.62
+1.75
Median inputs per type (sensitivity)
TABLE V: Retrospective efficiency errors on 66 types in 21 aircraft families. Errors are averaged within types, within families, then equally across families. All methods exclude the evaluated family from fitting; nested selection also excludes it from choosing inputs. MAE is in η1 units; MAPE and signed relative bias are percentages.
Phase
Set
MAPE
Bias
MAE
Fuel APE
(%)
(%)
(kg/s)
(%)
Climb
P
9.67
-8.34
0.139
9.40
R
9.88
-8.51
0.143
9.70
Cruise
P
4.68
-2.72
0.032
3.14
R
4.50
-2.10
0.030
2.58
Descent
P
26.83
-3.98
0.092
14.97
TABLE VI: Fuel flow and integrated-fuel errors on 100 A320-214 flights, with equal flight weights. P: published 5B4/P; R: reconstructed base 5B4. Bias is signed relative fuel flow error. Climb includes initial climb; descent includes approach. Fuel APE is absolute percentage error in integrated fuel over common covered intervals. Noncruise results use the clean-configuration model; takeoff and landing are excluded.
Aerodynamic surrogate models trained on high-fidelity CFD data reproduce numerical predictions of both scalar outputs and entire fields accurately, yet their predictive fidelity is limited by systematic discrepancies between CFD and experimental observations. We present an experimentally grounded correction framework that adapts a CFD-trained deep learning surrogate using wind-tunnel PSP measurements. A Geotransolver surrogate trained on 2,300 high-fidelity CFD simulations of the NASA CRM wing-body configuration, spanning geometric variation, Mach 0.70-0.85, and angles of attack 0 to 4 degrees, reproduces the CFD integrated aerodynamic forces and pitching moment to R2 > 0.99 but does not match the experimental data. To incorporate experimental information without retraining the surrogate, a correction network is trained on spatially registered PSP measurements at two freestream Mach numbers (0.70 and 0.85) across the same angle-of-attack range, learning the discrepancy between the surrogate-predicted and experimentally measured surface-pressure distributions. At Mach 0.85 the correction substantially improves agreement with PSP, particularly at the wing suction peak, shock location, and subsequent pressure recovery, reducing both the magnitude of the prediction error and the fraction of wetted surface on which it exceeds 0.05 in Cp, and it does so from a limited experimental dataset without modifying the pretrained surrogate parameters. On held-out angles of attack the grounded surrogate agrees with measurement to within 2.3-2.7% of the measured Cp range, and outperforms direct interpolation between the measured conditions at every state tested. Experimental measurements can therefore ground a large-scale simulation-trained surrogate by learning systematic CFD-to-experiment discrepancies while preserving its generalization capability and computational efficiency.
Nitin Nagesh Kulkarni, Dheeraj Vemula, Yin Yu +2
Luminary AI 101 S Ellsworth Ave, San Mateo, CA, USA
We build structure-preserving reduced-order models for both challenge problems of the AePW-4 High-Speed Working Group. They are a clamped thin panel at Mach 1.92 (RC-19) and a cantilever plate at Mach 5.85 (HyMAX). The published data for each case hold no time-resolved response and no repeated runs. The physics must therefore be carried by the form of the model. One construction gives two models per problem. A Ritz-Galerkin reference model describes the plate. A learned model keeps the reference model's exact potential, fits the remaining generalized forces, and trains only a dissipation that cannot add energy. Its energy balance holds for any weights by construction. On the trained model the worst relative residual is 2.2e-15. On RC-19, shock relations predict a pressure ratio of 1.521 across the panel. Pressure-sensitive paint measured the same value. Thirty runs of the learned model place the mean dominant frequency within 1.0% of the measured 236 Hz. At one periodic and one chaotic condition, the learned model reproduces the response regime of a published reduced-order model in all sixteen runs. It does so without that model's 113 Hz spectral peak, which the measurements do not show. Measured displacements can also correct the chaotic response through one more force that respects the energy balance. We tested this on synthetic data at the experiment's 5 kHz sampling rate. The state error is 1.25% of the panel thickness without noise and 3.5% with 10 micrometers of noise. On HyMAX the learned model follows the reference model's trailing-edge displacement for 65 ms with a relative error of 3.7%. These results show that a model whose structure carries the physics can match measured flutter behavior when the available data cannot train it.
Miguel A. Aguilo
Morphorm LLC, 100 Sun Avenue Northeast Suite 650, Albuquerque, NM 87109-4670, USA
Wind profiles in the atmospheric boundary layer (ABL) drive short-range forecasting and wind-energy siting, yet sparse radiosonde coverage forces practitioners to interpolate them from a handful of physical predictors. We benchmark modern tabular learners against analytical parameterizations on 158,097 IGRA soundings. CatBoost, TabM, and FT-Transformer reduce RMSE to 1.3-1.6 m/s, against approximately 3 m/s for power-law and Monin-Obukhov fits; once the predictor set is fixed the three architectures differ by less than 0.05 m/s, so accuracy is driven by predictors rather than architecture. A cohort analysis of within-bin profile dispersion places an empirical floor of approximately 1.77 m/s on any point-wise predictor that uses standard radiosonde features, consistent with prior eddy-viscosity inversions. A one-dimensional Fourier Neural Operator applied as a post-hoc corrector to frozen tabular predictions reduces near-surface RMSE on U.S. stations from 1.5 to 1.2 m/s, while leaving upper-layer error unchanged. Further gains should come from new predictors rather than larger models.