Where the Error Lives: Separating Lid-Edge and Body-Induced Error in Neural Pressure Prediction for 3D Cavity Flow
Organizations: School of Computing, National University of Singapore, Singapore · EXDYNA Pte. Ltd., 81 Ayer Rajah Crescent, #02-49, Singapore 139967 · Tandon School of Engineering, New York University, Brooklyn, NY, USA · Institute of Advanced Intelligence and Computing (IAIC), Agency for Science, Technology, and Research (A*STAR), Singapore, 138632
Abstract
Can a pressure surrogate score much better globally without better resolving the pressure around immersed bodies? We test this in a three-dimensional lid-driven cavity with an unregularized lid and centered pressure at native resolution. On 60 cases from 12 unseen geometries, with predictions sealed before the CFD truth existed, adding a body-free, Reynolds-conditioned reference correction to the same frozen surrogate lifted to lowers the mean full-volume relative error from 20.6% to 7.5%. In the post hoc regional analysis, the lid-edge junction occupies 1.5% of the fluid but holds 69% of the truth pressure energy and 95% of the uncorrected surrogate's squared error. Interior error changes by just +0.16% relative, and object-local peak error and recall change little on average. Retrospective ablations locate the learned term's contribution in the interior and near the bodies; rebasing on the reference lowers global error further with little change in cohort-mean near-body metrics. Reference-relative body-response error remains about 25% in the interior near-body band. Grid-refinement, lid-regularization and convergence checks delimit the numerical scope. In this benchmark, a large global gain chiefly reflects correction of a shared boundary field, not better resolution of the body response. We recommend reporting regional truth-energy and model-error shares alongside near-body and reference-relative body-response errors.
Figures & tables
| Method | Solution information available | Learned vs. deterministic operations |
|---|---|---|
| Reference Neural Operators ( Cheng et al., 2024 ) | reference-domain solution, reference and query geometry, a constructed deformation | transport the reference; learn the query-minus-transported correction |
| Multifidelity DeepONet ( Lu et al., 2022 ) | low- and high-fidelity training data; low-fidelity predictions as inputs | learn a low-fidelity operator and a high-fidelity residual |
| Hybrid wind interpolation ( Shaa et al., 2026 ) | CFD snapshots for a fixed object and sampled inlet angles | interpolation ansatz inside a trained neural model |
| Ours | immersed-geometry training fields; an offline body-free library; query geometry and only at inference | freeze the surrogate; add the deterministic residual ; center |
| Metric | Definition |
|---|---|
| Equation 2 on the full fluid volume | |
| Equation 2 on the near-body band | |
| Relative error of axiswise central differences where the center and both neighbours are fluid | |
| Equation 2 on the truth’s top-5% cells of (global peaks) | |
| Equation 2 on the truth’s top-5% cells of , excluding three cells at the outer walls (object peaks) | |
| Fraction of those truth object-peak cells also in the prediction’s same-size top-5% set |
| Held-out Re (16 values, 11–864) | Mean | p90 | Worst | Re 11–37 / 48–205 / 274–864 |
|---|---|---|---|---|
| of (%) | 0.072 | 0.164 | 0.198 | 0.039 / 0.051 / 0.131 |
| of (%) | 0.071 | 0.162 | 0.244 | – |
| , field correlation ( ) | 0.37 | 1.34 | 1.87 | – |
| Cohort A (primary): 60 cases, 12 unseen geometries | ||||||
|---|---|---|---|---|---|---|
| Lifted surrogate | 20.6 | 10.1 | 53.5 | 21.8 | 8.4 | 86.0 |
| Reference-corrected (ours) | 7.5 | 9.6 | 12.5 | 7.2 | 8.4 | 86.0 |
| Change | 63.5% | 5.5% | 76.6% | 66.9% | 0.1% | 0.01 pp |
| 95% CI (geometry) | [ 64.4, 62.6] | [ 15.8, 6.6] | [ 76.8, 76.3] | [ 67.2, 66.6] | [ 0.6, 1.1] | [ 0.02, 0.03] |
| Geometries improved | 12/12 | 7/12 | 12/12 | 12/12 | 8/12 | 8/12 |
| Arm | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Cohort A | |||||||||
| Lifted surrogate | 20.6 | 10.1 | 53.5 | 21.8 | 8.4 | 86.0 | 24.4 | 6.9 | 73.1 |
| Reference-corrected (ours) | 7.5 | 9.6 | 12.5 | 7.2 | 8.4 | 86.0 | 7.8 | 6.9 | 10.1 |
| Reference alone, | 8.9 | 43.9 | 3.6 | 3.3 | 39.3 | 55.8 | 1.4 | 20.5 | 3.3 |
| Rebased on the native reference | 3.2 | 9.4 | 2.5 | 1.2 | 8.2 | 86.0 | 0.7 | 6.7 | 2.6 |
| Cohort B | |||||||||
| (%) | Ours (%) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Cohort | Re | Lifted | Ours | Change | Worst | Worst | |||
| A | 10 | 24.8 | 8.2 | 66.8 | 8.8 | 8.6 | 92.2 | 17.1 | 86.8 |
| A | 30 | 24.6 | 8.2 | 66.9 | 8.7 | 8.2 | 91.9 | 14.1 | 85.7 |
| A | 100 | 22.8 | 7.6 | 66.7 | 9.6 | 8.9 | 88.0 | 15.9 | 75.0 |
| A | 300 | 17.4 | 6.9 | 60.5 | 10.7 | 7.4 | 87.8 | 13.5 | 81.0 |
| A | 900 | 13.5 | 6.8 | 49.5 | 10.0 | 9.0 | 69.9 | 22.8 | 41.5 |
| Share (%) | Regional error (%) | Change of (%) | ||||||
| Truth | Cohort A | Cohort B | ||||||
| Region | Cells | energy | Lifted | No-learning | Ours | +Reference | +Learned | +Reference |
| Junction | 1.5 | 68.6 | 24.4 | 7.9 | 7.8 | 67.8 | 0.3 | 67.7 |
| Rest of wall shell | 31.7 | 18.0 | 5.7 | 9.1 | 5.3 | 6.7 | 42.2 | – |
| Interior | 66.9 | 13.4 | 6.9 | 17.3 | 6.9 | 0.2 | 59.9 | 0.2 |
| Near body | 12.3 | 2.1 | 10.0 | 40.0 | 10.0 | 0.0 | 75.0 | – |
| Truth | Share of sq. error (%) | Regional error (%) | Median ratio | ||||
| Metric | Region | share (%) | Lifted | Ours | Lifted | Ours | ours/lifted |
| Wall shell | 98.7 | 99.7 | – | 53.8 | 12.3 | – | |
| Junction | 94.5 | 99.4 | 90.1 | – | – | – | |
| Interior | 1.3 | 0.3 | 5.3 | 22.6 | 22.6 | – | |
| Junction | 17.2 ∗ | 98.9 | 92.5 | – | – | 0.32 | |
| Interior | 32.1 ∗ | 0.3 | 2.2 | – | – | 1.01 | |
| Truth energy | Regional error (%) | |||||
|---|---|---|---|---|---|---|
| Subregion | Cells | share (%) | Lifted | No-learning | Ours | alone |
| Perpendicular edges ( ) | 14,336 | 59.9 | 24.4 | 7.7 | 7.7 | 1.1 |
| Parallel edges ( ) | 14,336 | 0.8 | 5.7 | 7.5 | 5.4 | 8.1 |
| Vertex blocks | 2,048 | 7.9 | 25.4 | 8.9 | 8.9 | 1.2 |
| Junction (all) | 30,720 | 68.6 | 24.4 | 7.9 | 7.8 | 1.4 |
| Width | Cells | Truth energy | Lifted sq. error | Change inside | Change outside |
|---|---|---|---|---|---|
| (cells) | (%) | share (%) | share (%) | (%) | (%) |
| 1 | 0.02 | 18.5 | 88.4 | 85.0 | 6.8 |
| 2 | 0.10 | 31.5 | 91.3 | 75.3 | 3.9 |
| 4 | 0.38 | 49.3 | 93.2 | 69.7 | 9.0 |
| 8 | 1.47 | 68.6 | 95.1 | 67.8 | 3.3 |
| 12 | 3.20 | 77.9 | 95.7 | 67.3 | 0.8 |
| Arm (cohort A) | Status | ||||||
| Background only | r | 21.8 | 38.3 | 53.6 | 22.1 | 36.4 | 61.9 |
| Lifted surrogate | p | 20.6 | 10.1 | 53.5 | 21.8 | 8.4 | 86.0 |
| No-learning control | r | 10.1 | 38.0 | 12.9 | 7.9 | 36.4 | 61.8 |
| Reference-corrected (ours) | p | 7.5 | 9.6 | 12.5 | 7.2 | 8.4 | 86.0 |
| Ours, 3 retrained seeds (mean) | r | 7.4 | 9.6 | 12.5 | 7.2 | 8.5 | 85.6 |
| Matched residual FNO, 3 seeds (mean) | r | 7.9 | 12.6 | 12.7 | 7.3 | 11.4 | 81.7 |
| Run (cohort A) | ||||||
| Historical (sealed; Table 4 ) | 7.54 | 9.57 | 12.53 | 7.23 | 8.41 | 85.97 |
| Retrained, seed 42 | 7.44 | 9.66 | 12.52 | 7.19 | 8.37 | 85.46 |
| Retrained, seed 314159 | 7.43 | 9.70 | 12.53 | 7.21 | 8.59 | 85.55 |
| Retrained, seed 271828 | 7.36 | 9.35 | 12.52 | 7.20 | 8.40 | 85.83 |
| Range, 3 retrainings | 0.08 | 0.34 | 0.01 | 0.02 | 0.23 | 0.37 |
| Range, all 4 runs | 0.18 | 0.34 | 0.02 | 0.04 | 0.23 | 0.51 |
| Lid gap | Change when adding the reference | ||||||
|---|---|---|---|---|---|---|---|
| Geometry | Placement | (box side) | (% of ) | (%) | (%) | (%) | (pp) |
| capsule-00 | Interior | 0.242 | 0.00 | 63.3 | 0.5 | 0.6 | 0.04 |
| capsule-01 | Near | 0.328 | 0.00 | 65.1 | 0.3 | 0.4 | 0.02 |
| capsule-02 | Near / face | 0.078 | 1.12 | 62.2 | 47.4 | 4.6 | 0.01 |
| capsule-03 | Near | 0.516 | 0.00 | 62.8 | 0.2 | 0.1 | 0.01 |
| pair-00 | Interior | 0.492 | 0.00 | 62.7 | 0.1 | 0.2 | 0.01 |
| 8 cells | width | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Lid | ||||||||||
| sharp | 10 | 83.9 | 72.5 | 62.7 | 66.6 | 72.5 | 76.6 | 5.48 | 4.51 | 3.85 |
| sharp | 100 | 85.1 | 71.4 | 58.6 | 66.4 | 71.4 | 74.9 | 4.71 | 3.79 | 3.24 |
| sharp | 997 | 73.1 | 65.0 | 44.8 | 63.0 | 65.0 | 66.1 | 2.72 | 1.11 | 0.68 |
| regularized | 10 | 74.1 | 48.9 | 24.2 | 47.2 | 48.9 | 49.4 | 0.36 | 0.06 | 0.01 |
| regularized | 100 | 78.0 | 51.4 | 24.1 | 50.0 | 51.4 | 51.8 | 0.30 | 0.04 | 0.01 |
| Lid | Arm | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| sharp | lifted | 68.6 | 20.6 | 10.1 | 53.5 | 21.8 | 8.4 | 86.0 | 24.4 | 6.9 | 95.1 |
| + reference | 7.5 | 9.6 | 12.5 | 7.2 | 8.4 | 86.0 | 7.8 | 6.9 | 75.4 | ||
| regularized | lifted | 51.5 | 31.4 | 12.7 | 126.5 | 34.6 | 9.8 | 85.5 | 43.3 | 9.1 | 93.5 |
| + reference | 32.6 | 12.8 | 131.8 | 36.0 | 9.8 | 85.5 | 45.1 | 9.2 | 93.9 |
Appendix figures & tables10 assets
Supplementary material from the paper’s appendix.
Appendix
| Cohort | Statistic | ||||||
|---|---|---|---|---|---|---|---|
| A | Mean | 7.5 | 9.6 | 12.5 | 7.2 | 8.4 | 86.0 |
| A | p90 (recall: p10) | 8.4 | 12.3 | 14.9 | 8.5 | 13.5 | 67.1 |
| A | Worst | 9.2 | 20.2 | 15.1 | 8.6 | 22.8 | 41.5 |
| B | Mean | 7.4 | 9.8 | 12.5 | 7.3 | 9.7 | 83.3 |
| B | p90 (recall: p10) | 8.3 | 13.7 | 14.9 | 8.5 | 16.2 | 65.3 |
| B | Worst | 8.5 | 15.2 | 15.2 | 8.6 | 19.7 | 43.6 |
| A + | A + vs. A ( ) | A + drift ( ) | B: residual | ||||
|---|---|---|---|---|---|---|---|
| Geometry | iterations | global | near | global | near | at 16,000 | |
| capsule-00 | 10 | 7,096 | 0.32 | 0.65 | 0.18 | 0.36 | |
| capsule-00 | 900 | 3,879 | 1.17 | 1.63 | 1.75 | 3.50 | |
| capsule-01 | 10 | 7,215 | 0.56 | 1.45 | 0.28 | 0.67 | |
| capsule-01 | 900 | 3,856 | 1.67 | 2.29 | 1.70 | 2.93 | |
| pair-00 | 10 | 7,117 | 0.25 | 1.07 | 0.14 | 0.62 | |
| Cohort | Role | Cases/geom. | Grid | Comparison |
|---|---|---|---|---|
| Coarse replication 1 | prospective | 120/24 | volume baseline (text) | |
| Coarse replication 2 | prospective | 120/24 | volume and spectral baselines ( Table 19 ) | |
| Surface cohort | prospective | 120/24 | surface-conditioned vs. volume model | |
| Surface ablation | development | 60/12 | complete vs. null surface input | |
| Composition cohort | prospective | 80/16 | coarse surrogate on new compositions | |
| Freeform cohort | prospective | 80/16 | coarse surrogate on freeform shapes |
| Model | ||||||
|---|---|---|---|---|---|---|
| BR-MGL (residual) | 5.26 | 13.02 | 3.67 | 2.13 | 9.18 | 80.17 |
| Direct MGL | 10.17 | 13.03 | 15.48 | 8.38 | 9.08 | 80.13 |
| Residual FNO | 6.13 | 14.88 | 4.84 | 2.35 | 11.30 | 78.34 |
| Direct FNO | 13.04 | 16.37 | 20.97 | 10.32 | 11.38 | 77.19 |
| Cohort | Predictor | ||||||
|---|---|---|---|---|---|---|---|
| Surface | BR-MGL | 5.18 | 13.00 | 3.79 | 2.43 | 9.31 | 80.29 |
| Surface-conditioned | 4.12 | 9.81 | 3.38 | 1.83 | 6.68 | 83.95 | |
| Composition | Background only | 7.11 | 39.06 | 4.84 | 2.80 | 41.06 | 54.23 |
| Coarse surrogate | 3.11 | 9.68 | 2.63 | 1.24 | 7.35 | 85.94 | |
| Freeform | Background only | 9.01 | 47.59 | 5.52 | 3.78 | 48.74 | 49.62 |
| Coarse surrogate | 3.67 | 9.88 | 2.91 | 1.52 | 7.53 | 84.54 |
| Change of | Median | LOGO | One-sided 95% | Geometries | Sign-flip | ||
|---|---|---|---|---|---|---|---|
| Cohort | Metric | means (%) | geometry (%) | range (%) | bound (%) | improved | |
| A | 63.46 | 63.16 | [ 63.72, 63.28] | 62.74 | 12/12 | 0.0005 | |
| A | 5.53 | 0.08 | [ 6.11, 0.05] | 4.28 | 7/12 | 0.2168 | |
| A | 76.59 | 76.70 | [ 76.69, 76.55] | 76.39 | 12/12 | 0.0005 | |
| A | 66.89 | 66.93 | [ 66.96, 66.82] | 66.66 | 12/12 | 0.0005 | |
| A | 0.08 | 0.09 | [ 0.12, 0.13] | 0.95 | 8/12 | 0.9854 |
| Cohort | Family | ||||||
|---|---|---|---|---|---|---|---|
| A | capsule | 7.6 | 8.1 | 12.5 | 7.2 | 4.8 | 87.7 |
| A | pair | 7.6 | 10.7 | 12.7 | 7.2 | 10.2 | 87.8 |
| A | triplet | 7.5 | 9.9 | 12.5 | 7.3 | 10.3 | 82.4 |
| B | capsule | 7.4 | 8.5 | 12.4 | 7.3 | 7.4 | 81.1 |
| B | pair | 7.5 | 10.8 | 12.5 | 7.3 | 11.8 | 83.7 |
| B | triplet | 7.3 | 10.0 | 12.5 | 7.2 | 10.1 | 85.0 |
Neural Surrogate Modeling
Neural surrogate modeling uses learned models to approximate costly simulations, enabling faster prediction, design, control, and inference.
See all 466 papers