Case A: validation
Contents
Engine status: OpenFOAM 12 scored · OpenLB scored · FluidX3D scored · ML surrogate scored
Nomenclature: symbols, columns and scores
Geometry, reference scales and release
- \(H\), \(b\), \(D\) (m) — height of the building, cube, tower or tree belt; width of the building; edge of the cubes of Case C: the lengths the cases are drawn and normalised by
- \(U_H\) (m/s) — reference speed: the mean approach-flow speed at the height \(H\) (each Data page gives the value used)
- \(U_R\), \(H_R\), \(Z_R\) (m/s, m) — reference speed and height of a workbook: Case E, \(U_R\) at \(Z_R\) = 250 m full scale (1 m in the model); Case M, \(U_R\) and \(H_R\) at the reference point of the tunnel and of each field run
- \(U_{\mathrm{ref}}\), \(L_{\mathrm{ref}}\) (m/s, m) — reference speed and length of a case’s normalisation: \(U_H\) and \(H\) (\(U_R\) and \(H_R\) in Case M); for the velocity ratios the approach speed at 15.9 m (Case E) or at 500 m (Case F)
- \(U_b\) (m/s) — Case G: mean approach-flow speed at 9 m, the normalisation of its approach-flow table
- \(q\), \(c_{\mathrm{gas}}\) (m³/s, ppm) — tracer release of the experiment: volume flow of the released gas, and its tracer content (\(10^6\) ppm when pure); \(Q\) is the source strength given to a solve
- \(R_b\) (dimensionless) — bulk Richardson number of the heated-floor cases J and L (negative: unstable stratification)
- \(\mathrm{Re_H}\), \(\mathrm{Re_D}\) (dimensionless) — Reynolds numbers \(U_H H/\nu\) and \(U_H D/\nu\) (\(\nu\): kinematic viscosity of air)
- \(u_\ast\), \(z_0\), \(\kappa\) (m/s, mm, dimensionless) — friction velocity and roughness length of the log law fitted to the approach flow, and the von Kármán constant (0.41)
- \(\mathrm{WD}\) (column
wind_direction_deg; °) — wind direction (each frame says how it is counted); \(\mathrm{WD_R}\) is the mean direction of a field run at its reference point (Case M) - \(x\), \(y\), \(z\) (columns
x,y,z; m) — coordinates in the case’s frame (\(+x\), \(+y\) and \(+z\) are the positive directions; each Data page says which way \(+x\) points), \(z\) up from the floor of the tunnel or from the ground - \(\mathrm{d}x\) (m) — lattice spacing of the OpenLB and FluidX3D runs, in metres or as a fraction of \(H\) (\(H/64\))
- \(L_r\) (m) — reattachment length behind the building, read from a wake line at ground level; reported as \(L_r/H\)
Velocity and turbulence
- \(U\), \(V\), \(W\) (columns
U,V,W; m/s) — mean velocity along \(+x\), \(+y\), \(+z\); \(V\) is positive toward \(+y\) in every case - \(\lvert\mathbf{U}\rvert\) (m/s) — magnitude of the mean velocity vector, \(\sqrt{U^2+V^2+W^2}\)
- \(V_{\mathrm{s}}\) (column
Vs; m/s) — mean omnidirectional scalar speed of a thermistor (never a vector magnitude) - \(\overline{u'u'}\), \(\overline{v'v'}\), \(\overline{w'w'}\) (columns
uu,vv,ww; m²/s²) — velocity variances (overbar: time average, prime: fluctuation about it) - \(\overline{u'v'}\), \(\overline{u'w'}\), \(\overline{v'w'}\) (columns
uv,uw,vw; m²/s²) — Reynolds shear stresses - \(k = \tfrac{1}{2}\left(\overline{u'u'}+\overline{v'v'}+\overline{w'w'}\right)\) (column
k; m²/s²) — turbulent kinetic energy - \(U/U_{\mathrm{ref}}\) (column
velocity_ratio; dimensionless) — mean speed over the case’s reference speed (\(V_{\mathrm{s}}/U_{\mathrm{ref}}\) for a thermistor) - \(\sigma_{U/U_{\mathrm{ref}}}\) (column
velocity_ratio_sd; dimensionless) — standard deviation ofvelocity_ratioover the observations behind a field value (Case F) - \(k/U_{\mathrm{ref}}^2\) (column
k_ratio; dimensionless) — turbulent kinetic energy over the square of the reference speed, where only normalised values are published (Case G)
Concentration and temperature
- \(C\) (column
C; ppm) — mean tracer-gas concentration as published (volume fraction \(\times 10^6\) of the released gas diluted in air) - \(c_{\mathrm{rms}}\) (column
c_rms; ppm) — root mean square of the concentration fluctuation - \(c_0\) (ppm) — reference concentration, defined for each case on its Data page: \(c_0 = \dfrac{c_{\mathrm{gas}}\thinspace q}{U_H H^2}\) (\(U_R H_R^2\) in the denominator in Case M)
- \(c^\ast\) (column
concentration_ratio; dimensionless) — dimensionless concentration \(c^\ast = C/c_0\); the value of a solve is \(c\thinspace U_H H^2/Q\) for its mean concentration \(c\) and its source strength \(Q\) - \(\mathrm{Sc_t}\) (dimensionless) — turbulent Schmidt number of a solve (0.3, 0.5 and 0.7)
- \(T\), \(T_{\mathrm{rms}}\) (columns
T,T_rms; °C, K) — mean air temperature as published (Cases J and L) and the root mean square of its fluctuation - \(\overline{u'c'}\), \(\overline{v'c'}\), \(\overline{w'c'}\) (columns
uc,vc,wc; ppm·m/s) — turbulent concentration fluxes - \(\overline{u'\theta'}\), \(\overline{v'\theta'}\), \(\overline{w'\theta'}\) (columns
uT,vT,wT; K·m/s) — turbulent heat fluxes (\(\theta'\): the temperature fluctuation) - \(\theta = \dfrac{T_f - T}{T_f - T_H}\) (dimensionless) — floor-referenced temperature the solves of Cases J and L are scored on: 0 at the floor, 1 in the approach air at \(H\); \(T_f\) is the floor temperature and \(T_H\) the temperature of the approach flow at \(H\)
Scores
- \(n\) — number of pairs of a measured value \(o_i\) and a modelled value \(p_i\) at one point
- \(R\) — Pearson correlation coefficient of the pairs
- \(R^2\) — square of \(R\): share of the measured variance a linear fit to the model explains
- \(\mathrm{FAC2}\) — fraction of pairs with \(0.5 \le p_i/o_i \le 2\)
- \(\mathrm{FB}\) — fractional bias \(2(\overline{o}-\overline{p})/(\overline{o}+\overline{p})\), positive when the model is low
- \(\mathrm{HR}\) (hit rate) — fraction of pairs with \(\lvert p_i-o_i\rvert/\lvert o_i\rvert \le \delta_{\mathrm{rel}}\) or \(\lvert p_i-o_i\rvert \le \delta_{\mathrm{abs}}\), where \(\delta_{\mathrm{rel}} = 0.25\) and \(\delta_{\mathrm{abs}} = 0.05\thinspace U_H\) for a velocity component
- \(\mathrm{NMSE} = \dfrac{\overline{(o-p)^2}}{\overline{o}\thinspace \overline{p}}\) — normalised mean square error
- \(\mathrm{NAD} = \dfrac{\overline{\lvert o-p\rvert}}{\overline{o}+\overline{p}}\) — normalised absolute difference
- \(\mathrm{MG} = \exp\overline{(\ln o-\ln p)}\) — geometric mean bias
- \(\mathrm{VG} = \exp\overline{(\ln o-\ln p)^2}\) — geometric variance
- The ratio of \(\mathrm{FAC2}\) keeps its sign, so a model of the other sign than the measurement is a miss; \(\mathrm{FB}\), \(\mathrm{NMSE}\) and \(\mathrm{NAD}\) divide by the means, so they are undefined (shown as —) for a signed component unless the measured and the modelled mean are both positive
- VDI 3783-9 calls the hit rate \(q\) and the tolerances \(D\) and \(W\); on this site those letters keep their other meanings (\(q\) the tracer release rate, \(D\) the block size of Case C, \(W\) the vertical velocity)
- The detection limit of a concentration is one value per case, \(10^{-3}\) of the largest measured value. It sets \(\delta_{\mathrm{abs}}\) of the hit rate and the both-below rule of \(\mathrm{FAC2}\), and floors the values in \(\mathrm{MG}\) and \(\mathrm{VG}\).
- The scores \(\mathrm{FB}\), \(\mathrm{FAC2}\), \(\mathrm{NMSE}\), \(\mathrm{NAD}\), \(\mathrm{MG}\) and \(\mathrm{VG}\) follow Chang & Hanna (2004); the hit rate follows VDI 3783-9.
Quantities scored
- \(U\), \(V\) and \(W\) (mean velocity components) over \(U_H\)
- \(k\) (turbulent kinetic energy) over \(U_H^2\)
- the speed \(\lvert\mathbf{U}\rvert\) and \(\sqrt{\lvert\mathbf{U}\rvert^2+2k}\), against a thermistor’s \(V_{\mathrm{s}}\)
- \(c^\ast = C/c_0\) (dimensionless concentration) at turbulent Schmidt numbers \(\mathrm{Sc_t}\) of 0.3, 0.5 and 0.7
- \(\theta\) (floor-referenced temperature)
Measurement locations
Explore the points
OpenFOAM 12
Status: scored
Measured against modelled
- Normalised by \(U_H\). FB is positive when the model is low.
- Reattachment length behind the building: \(L_r/H\) = 1.49.
- \(V\) and \(W\) change sign. The mean measured \(W\) is ~0.01 m/s.
- The scores table shows the served steady run (RANS) beside the same solver run transient (URANS: SST k-omega, PIMPLE, 1000 s of flow time on the measured inlet, scored on the time means with \(k\) as the modelled mean plus the resolved part). The URANS run reattaches at \(L_r/H\) = 1.29 and resolves 27 % of the wake’s \(k\) (3 % upstream); its inlet is the tunnel’s measured profile, the RANS run’s the fitted log law.
Interpretation: Case A scores and profiles
- The streamwise field is right: \(R = 0.96\), FAC2 0.82.
- \(k\) is under-predicted (FB +0.48, \(R = 0.48\)); Cases B and H show the same under-prediction.
- \(V\) and \(W\) change sign, so factor-of-two agreement is a harsh measure near zero and FAC2 is low; R and hit rate carry the comparison there.
- \(W\)’s FB of −1.50 and its NMSE follow from a near-zero mean; the correlation is \(R = +0.71\).
- The model follows the measured profile closely above roof height and in the recovering wake; the near-ground deficit immediately behind the building is where the closure does least well.
Field maps
The solved velocity on the measured planes, every point drawn on it as a dot coloured by its measured value on the same scale: a dot that stands out from the colour around it is a point the model misses.
Scores
Reattachment length \(L_r/H\) = 1.49.
| Score | Run | R | \(R^2\) | FAC2 | FB | Hit rate |
|---|---|---|---|---|---|---|
| \(U\) | RANS | +0.955 | 0.912 | 0.817 | +0.185 | 0.586 |
| URANS | +0.953 | 0.908 | 0.763 | +0.181 | 0.683 | |
| \(V\) | RANS | +0.961 | 0.923 | 0.349 | -0.207 | 0.892 |
| URANS | +0.953 | 0.909 | 0.360 | — | 0.817 | |
| \(W\) | RANS | +0.708 | 0.502 | 0.398 | -1.501 | 0.801 |
| URANS | +0.675 | 0.456 | 0.414 | -1.352 | 0.699 | |
| \(k\) | RANS | +0.480 | 0.230 | 0.667 | +0.478 | 0.177 |
| URANS | +0.626 | 0.392 | 0.694 | +0.453 | 0.392 |
Every score is over the same 186 points.
- Per-point values:
/api/cases/A/pairs/openfoam12-realizableKE.csv - RANS: realizable k-ε, log-law inlet (the served run). URANS: SST k-ω run transient, measured inlet, the submission
eddy3d-of12-sst-urans, scores/api/submissions/eddy3d-of12-sst-urans.json - Reproduce: Eddy3D fixture TestAijCaseAContainerE2E; figures validation/plot_cases_ab_openfoam.py
OpenLB
Status: scored
OpenLB (Eddy3D’s eddy3dWind: Smagorinsky, third-order regularised D3Q19, image 1.9.3-r6, CPU build) solved the experiment with the same fitted inflow as OpenFOAM 12 and probed the same measured points, at \(\mathrm{d}x = 1.6\thinspace\mathrm{m}\) (\(H/10\)), 1.3 million cells, 13 min on 24 threads of one node (54 MLUPS). \(H/20\) costs about 16 times the cell-steps.
| Quantity | n | FAC2 | FB | hit rate | R | OpenFOAM FAC2 / hit rate / R |
|---|---|---|---|---|---|---|
| \(U/U_H\), all points | 186 | 0.47 | +0.38 | 0.38 | 0.78 | 0.82 / 0.59 / 0.96 |
| \(U/U_H\), points ≥ 2 cells above the floor | 121 | 0.65 | +0.22 | 0.56 | 0.80 | — |
| \(V/U_H\) | 186 | 0.25 | — | 0.66 | 0.77 | 0.35 / 0.89 / 0.96 |
| \(W/U_H\) | 186 | 0.18 | — | 0.56 | 0.57 | 0.40 / 0.80 / 0.71 |
- OpenLB meets the ground with a plain no-slip (bounce-back) wall at \(z = 0\) and carries the roughness only in the inlet profile. In the undisturbed column at the windward face its \(U\) is 0.60 of the inlet log law at the first node (half a cell up), 0.88 one cell up, 1.14 at 1.5, 1.10 at 2, 1.04 at 3 and 1.01 from 4. Case A’s lowest plane sits 0.6 cells up.
- The building’s wall faces sit at \(z = 0\) and \(16\thinspace\mathrm{m}\) (250 solid nodes). Until image
1.9.3-r6the solver’s lattice started at the domain’s floor, which put the wall half a cell up and the roof a row short (14.4 m); this run replaces the one made with that lattice. - The resolved \(k\) is 1 % of the measured one; \(k\) is not compared.
- Reattachment length: \(L_r = 2.80\thinspace H\) against OpenFOAM’s 1.49; the field fits best sampled \(H/2\) downstream (R 0.81 against 0.78).
- Control: with the lateral sign flipped, \(V\)’s R becomes −0.77; \(U\) scrambled across the points drops R to 0.10.
Interpretation: OpenLB on Case A
- The near-zero model \(U\) at the floor in the OpenLB row of Figure 7 is the no-slip wall; above two cells FAC2 rises to 0.65 and FB falls to +0.22.
- The solve is nearly steady: a steady log-law inflow at \(H/10\) resolves almost no turbulence, so \(k\) is not compared, and with nothing to entrain the wake runs long.
- Its frame is right.
Scores
Reattachment length \(L_r/H\) = 2.80.
| Score | R | \(R^2\) | FAC2 | FB | Hit rate | n |
|---|---|---|---|---|---|---|
| \(U\) | +0.782 | 0.611 | 0.468 | +0.378 | 0.376 | 186 |
| \(U\), points ≥ 2 dx | +0.800 | 0.640 | 0.653 | +0.219 | 0.562 | 121 |
| \(V\) | +0.767 | 0.588 | 0.253 | -0.162 | 0.661 | 186 |
| \(W\) | +0.574 | 0.329 | 0.177 | -1.711 | 0.559 | 186 |
| \(k\) (resolved) | +0.162 | 0.026 | 0.000 | +1.958 | 0.000 | 186 |
| \(V\), lateral sign flipped (control) | -0.767 | 0.588 | 0.124 | -24.677 | 0.554 | 186 |
- Per-point values:
/api/cases/A/pairs/openlb.csv - Reproduce: Eddy3D fixture TestAijCaseAOpenLbContainerE2E (OpenLB eddy3dWind, image 1.9.3-r6, CPU build, \(\mathrm{d}x = 1.6\thinspace\mathrm{m}\)); figure validation/plot_cases_ab_engines.py
FluidX3D
Status: scored
FluidX3D (Eddy3D’s GPU lattice-Boltzmann engine: D3Q19 with Smagorinsky LES) solved the experiment at \(\mathrm{d}x = H/64\) (2.5 mm, 117 million cells) with the same fitted log-law inflow, averaging over 200 \(H/U_H\) after 40 of spin-up. The run has an empty-domain control on the same lattice, and a point is scored only at a height where that control holds the imposed log law within 5 %: 116 of the 186 points. Its row is in Figure 7.
| Quantity, comparable heights | n | FAC2 | FB | hit rate | R | OpenFOAM on the same points: FAC2 / hit rate / R |
|---|---|---|---|---|---|---|
| \(U/U_H\) | 116 | 0.60 | +0.41 | 0.57 | 0.84 | 0.85 / 0.73 / 0.95 |
| \(V/U_H\) | 116 | 0.26 | — | 0.75 | 0.85 | 0.27 / 0.90 / 0.95 |
| \(W/U_H\) | 116 | 0.22 | — | 0.41 | 0.59 | 0.49 / 0.79 / 0.93 |
- The run meets the ground with FluidX3D’s bounce-back wall; it predates Eddy3D’s opt-in ground wall function. In the empty domain the flow at the lowest plane (\(z = 0.01\thinspace\mathrm{m}\), 4 cells up) reads 0.52 of the imposed log law on average, and 0.86 at \(z = 0.04\thinspace\mathrm{m}\). Those points are scored separately (
velocity.*.excluded) and drawn hollow in Figure 7. - Empty-domain control on the comparable points, \(U\): FAC2 0.77, FB −0.26, hit rate 0.59, R 0.51. The solved field: FAC2 0.60, FB +0.41, hit rate 0.57, R 0.84.
- Control: with the lateral sign flipped, the \(V\) correlation becomes −0.85.
Interpretation: FluidX3D on Case A
- The excluded points are the floor again.
- The approach flow alone scores better on \(U\) (FAC2, FB and hit rate); only the correlation shows the building’s effect.
- Its frame is right.
Scores
Reattachment length \(L_r/H\) = 0.05.
| Score | R | \(R^2\) | FAC2 | FB | Hit rate | n |
|---|---|---|---|---|---|---|
| \(U\), comparable heights | +0.840 | 0.705 | 0.603 | +0.405 | 0.569 | 116 |
| \(V\), comparable heights | +0.845 | 0.714 | 0.259 | -0.692 | 0.750 | 116 |
| \(W\), comparable heights | +0.591 | 0.349 | 0.224 | -1.678 | 0.414 | 116 |
| \(U\), all points | +0.806 | 0.650 | 0.597 | +0.372 | 0.452 | 186 |
| \(U\), approach flow alone (control) | +0.510 | 0.260 | 0.767 | -0.256 | 0.586 | 116 |
- Per-point values:
/api/cases/A/pairs/fluidx3d.csv - Reproduce: Eddy3D fixture TestAijCaseAFluidX3DE2E (Eddy3DCli.Tests/Validation/TestAijVelocityFluidX3DE2E; FluidX3D GPU lattice Boltzmann, D3Q19 Smagorinsky LES, \(\mathrm{d}x = H/64\), 40 + 200 \(H/U_H\) spin-up + averaging, with an empty-domain control on the same lattice); figure validation/plot_cases_ab_engines.py
ML surrogate
Status: scored
Eddy3D’s Wind Predictor (the public Yel-2.0 network) predicts \(\lvert\mathbf{U_{\mathrm{mean}}}\rvert/U_{\mathrm{ref}}\) at one height, 1.8 m, so the case is scaled ×180 to put its measurement plane there; the prediction is read at the tunnel’s own sensor coordinates. How the surrogate is run is on the ML surrogate page.
| Quantity | n | FB | FAC2 | R | hit rate | OpenFOAM, same points and quantity |
|---|---|---|---|---|---|---|
| \(\lvert\mathbf{U}\rvert/U_H\), \(z = 0.01\thinspace\mathrm{m}\) plane | 65 | +0.77 | 0.26 | −0.13 | 0.09 | FB +0.27, FAC2 0.85, R +0.90, hit rate 0.45 |
- FB is positive when the model is low. Reversing the wind changes the predicted field by at most \(1 \times 10^{-4}\) \(U_{\mathrm{ref}}\): the prediction does not depend on the wind direction.
Scores
| Score | R | \(R^2\) | FAC2 | FB | Hit rate |
|---|---|---|---|---|---|
| \(\lvert\mathbf{U}\rvert\) (\(z = 0.01\thinspace\mathrm{m}\) plane) | -0.125 | 0.016 | 0.262 | +0.772 | 0.092 |
| \(\lvert\mathbf{U}\rvert\), lattice cut at ±250 m | +0.434 | 0.188 | 0.615 | +0.401 | 0.338 |
| wind reversed (control) | -0.125 | 0.016 | 0.262 | +0.772 | 0.092 |
Every score is over the same 65 points.
- Per-point values:
/api/cases/A/pairs/ml-surrogate.csv - Reproduce: Eddy3D fixture Outdoor.ML.Tests/TestAijSurrogateE2E; figure validation/plot_ml_surrogate.py
Grid uncertainty
OpenFOAM on four meshes (edge ratio \(\sqrt{2}\), 151,210 to 1,411,566 cells), eddy-viscosity cap lifted as in eddy3d-of12-nutcap-lifted. The served run keeps the cap.
- Median band: \(U\) 0.011 \(U_H\) (p90 0.051); \(V\), \(W\), \(k\) about 0.001. 140 of 186 points within the hit rate’s 0.05 \(U_H\); 30 without a band.
- Across the meshes: \(U\) FAC2 0.79–0.82, hit rate 0.58–0.63; \(k\) \(R\) 0.28–0.44.
- \(L_r/H\) 1.49, 1.44, 1.50, 1.51; extrapolated 1.51, measured 0.38–0.75. The long wake is the closure, not the mesh.
- Bands:
grid/openfoam12-realizableKE.csv. Method: Grid uncertainty.
Notes on every run
- Normalisation: velocities by \(U_H\), \(k\) by \(U_H^2\) (the benchmark's own convention; Case E by \(U\) at 15.9 m, as it is published); hit rate \(\delta_{\mathrm{rel}} = 0.25\), \(\delta_{\mathrm{abs}} = 0.05\) in those units
Leaderboard
| Rank | Run | Source | Hit rate, \(U\) | Hit rate, \(V\) | Hit rate, \(W\) | Hit rate, \(\lvert\mathbf{U}\rvert\) | Hit rate | FAC2 | R | \(R^2\) |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | OpenFOAM simpleFoam (Eddy3D Outdoor case writer) (eddy3d-of12-mesh2x-measured-inlet) | community | 0.651 | 0.887 | 0.833 | 0.677 | 0.762 | 0.603 | +0.900 | 0.819 |
| 2 | OpenFOAM simpleFoam (Eddy3D Outdoor case writer) (eddy3d-of12-nutcap-lifted) | community | 0.613 | 0.903 | 0.833 | 0.667 | 0.754 | 0.613 | +0.891 | 0.805 |
| 3 | OpenFOAM simpleFoam (Eddy3D Outdoor case writer) (eddy3d-of12-measured-inlet) | community | 0.624 | 0.876 | 0.828 | 0.667 | 0.749 | 0.610 | +0.894 | 0.810 |
| 4 | OpenFOAM pimpleFoam (Eddy3D Outdoor case writer) (eddy3d-of12-sst-urans) | community | 0.683 | 0.817 | 0.699 | 0.720 | 0.730 | 0.590 | +0.877 | 0.782 |
| 5 | OpenFOAM 12 | Eddy3D | 0.586 | 0.892 | 0.801 | 0.634 | 0.728 | 0.610 | +0.891 | 0.805 |
| 6 | OpenFOAM simpleFoam (Eddy3D Outdoor case writer) (example-openfoam12) | example submission | 0.586 | 0.892 | 0.801 | 0.634 | 0.728 | 0.610 | +0.891 | 0.805 |
| 7 | FluidX3D | Eddy3D | 0.452 | 0.618 | 0.441 | 0.554 | 0.516 | 0.462 | +0.717 | 0.546 |
| 8 | OpenLB | Eddy3D | 0.376 | 0.661 | 0.559 | 0.409 | 0.501 | 0.353 | +0.728 | 0.538 |
All runs scored by validation/score_submissions.py. Ranked by mean hit rate (VDI 3783) over the tasks; bold: best per column. Not ranked: the ML surrogate (no point ids) and negative controls. Community rows are self-reported (Validate your own solver).