Case K: validation

Contents
  1. Measurement locations
  2. Explore the points
  3. OpenFOAM 12
    1. Measured against modelled
    2. Vertical profiles
    3. Field maps
    4. Scores
  4. Not run
  5. Leaderboard

Engine status: OpenFOAM 12 scored · OpenLB pending · FluidX3D pending · ML surrogate pending

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 of velocity_ratio over 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

Measurement lines and planes, Case K
Figure 1. Case K, measurement lines and planes: A vertical profiles, \(y/H = 0\); B vertical profiles, \(y/H = \pm 1\); C vertical profiles, \(y/H = \pm 2\); D vertical profiles, \(y/H = \pm 3\). Positions in multiples of \(H\).

Explore the points

model lowmodel high
Loading the points… Hover a point to see where it sits in 3D and how the model differs from the measurement; click to pin it. This needs JavaScript; the figures on this page show the same scores.

OpenFOAM 12

Status: scored

An Eddy3D production run solved on 2026-10-18, realizable k-ε at numerics level 3 and 3000 iterations, on one mesh. The solve ran to the iteration limit with the SIMPLE residual control unmet; the initial residual of epsilon at the last iteration was \(7 \times 10^{-3}\).

  • The tunnel at 1:100 (\(H = 6\thinspace\mathrm{m}\)). Measured and modelled velocities are both divided by the dataset’s stated \(U_H\) = 2.33 m/s, and \(k\) by \(U_H^2\). The model is driven by the measured approach profile (3.23 m/s at 0.2 m), which gives 1.97 m/s at \(z = H\), 0.85 of the stated \(U_H\).
  • The concentration is the measured \(C/c_0\) (\(c_0\) = 429.2 ppm) against the model’s \(c^\ast = c\thinspace U_H H^2/Q\) at the turbulent Schmidt numbers 0.3, 0.5 and 0.7, three species on one flow; all three are published.

Measured against modelled

Case K, measured vs modelled
Figure 2. Case K, measured vs modelled \(U\), \(W\), \(k\) and \(c^\ast\) at three turbulent Schmidt numbers; the dots take the colour of their line in Figure 1. The concentration panels show the positive pairs within five decades; the four points within one cube height of the source are open.
Quantity R FAC2 FB Hit rate n
\(U\) +0.958 0.586 +0.594 0.123 203
\(V\) −0.151 0.054 — 0.901 203
\(W\) +0.573 0.143 — 0.685 203
\(k\) +0.330 0.044 +1.201 0.005 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.3 +0.953 0.793 −0.690 0.458 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.5 +0.951 0.685 −0.875 0.163 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.7 +0.950 0.394 −0.983 0.099 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.3, without the 4 near-source points +0.923 0.799 −0.218 0.467 199
\(c^\ast\), \(\mathrm{Sc_t}\) 0.5, without the 4 near-source points +0.941 0.698 −0.430 0.161 199
\(c^\ast\), \(\mathrm{Sc_t}\) 0.7, without the 4 near-source points +0.923 0.402 −0.561 0.101 199
\(U\), y-symmetric ceiling (control) +1.000 1.000 +0.000 1.000 203
\(c^\ast\), y-symmetric ceiling (control) +0.997 0.995 +0.000 0.793 203

FB is positive when the model is low, and is left out (—) for a signed component whose measured and modelled means are not both positive.

  • Velocity. \(U\): \(R = 0.96\), FAC2 0.59; the mean \(U/U_H\) is 0.272 modelled against 0.502 measured (FB +0.59). FB by region: +0.35 over the roofs (31 points, FAC2 1.00, R +0.06), +0.79 in the streets along the wind (44 points, R +0.997, FAC2 0.34) and +0.78 in the streets across it (128 points, R +0.95, FAC2 0.57). Both sides are divided by 2.33 m/s; the solved \(U\) at \(z = H\) is 0.85 of that.
  • \(k\). The mean \(k/U_H^2\) is 0.027 modelled against 0.107 measured; R is 0.33 and FAC2 0.04.
  • \(V\) and \(W\). \(V\) has a constant offset in the tunnel and almost no antisymmetric part. A model with \(V\) = 0 everywhere hits 0.91 of the points (\(\delta_{\mathrm{abs}} = 0.05\thinspace U_H\)); \(V\)’s hit rate is 0.90 and its R is −0.15. \(W\): \(R = +0.57\), FAC2 0.14.
  • Concentration. R is 0.95 at every Schmidt number. FAC2 falls from 0.79 to 0.39 and FB from −0.69 to −0.98 as the Schmidt number rises; the mean \(c^\ast\) is 0.374 measured against 0.768, 0.956 and 1.097 modelled. The modelled peak at the measured points is 44, 57 and 66 against a measured peak of 8.73. The four points within one cube height of the source (the first point is the source location) are 3.7, 4.5 and 5.1 times the measured value on average. Without those four points: FB −0.22, −0.43 and −0.56, FAC2 0.80, 0.70 and 0.40, R 0.92, 0.94 and 0.92. Case H has FB −0.60, −0.83 and −0.93 at the same three Schmidt numbers (Case H).
  • Symmetry. The measured plume is not y-symmetric: C at \(-y\) exceeds C at \(+y\) at 74 of 74 mirror pairs, and Eddy3D puts the measured plume’s lateral centroid 0.011 to 0.017 m toward \(-y\). The solve is y-symmetric. Scored against the mean of each mirror pair, the measurement has \(R = 1.00\) for \(U\) and \(R = 0.997\), FAC2 0.995 for \(c^\ast\) (the ceiling-control rows).
  • First ground cell. 47 of the 203 points (the \(z = 4\thinspace\mathrm{mm}\) plane) are inside the first ground cell; their modelled values are that cell’s, pulled toward the wall’s zero by the interpolation.
  • Flow structure. The air over the roofs outruns the street along the wind, which outruns the cavity across it; the plume is clean 1 to \(1.5\thinspace H\) ahead of the array, decays through four cavities and is cleaner over the roofs than in them; the solved velocity is y-symmetric to twice the tunnel’s own mirror-pair difference.
Interpretation: Case K velocity, k and concentration
  • The modelled velocity is under the measured mean in every region. Part of the deficit is the inflow.
  • \(k\) is a factor of four low and uncorrelated: the model starts with more \(k\) than the tunnel (the inlet \(k\) of the log-law inflow is about 1.7 times the tunnel’s) and has less at the points.
  • \(V\) and \(W\) carry little. \(V\)’s hit rate of 0.90 scores nothing, because a model with \(V\) = 0 everywhere hits almost as many points.
  • The concentration pattern is correlated and the level is high. The excess at the four near-source points follows from the model’s source: the four cells (0.75 m across) that a 0.44 m square cuts, where the tunnel’s outlet is 5 mm.
  • The ceiling scores of a y-symmetric model are far above the model’s, so symmetry does not limit the scores. In Eddy3D’s own results a lateral shift of the model by +0.01 m raises R of \(c^\ast\) at \(\mathrm{Sc_t}\) 0.5 from 0.951 to 0.975 (not recomputed here).

Vertical profiles

Where along the height the model meets the measurement, station by station.

Vertical profiles, Case K
Figure 3. Case K, vertical profiles of \(U/U_H\) on the centre plane \(y = 0\) at four stations along the measured lines, measured points against the modelled line, offset by 1 per station; the dashed line is the roof height.

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.

Solved field and measured points, Case K, plan
Figure 4. Case K, the horizontal plane \(z = 0.01\thinspace\mathrm{m}\): (a) \(u\) over \(U_H\); (b) the same with iso-lines every 0.1 and the edge of the reversed flow, \(u = 0\), dashed; (c) the in-plane speed with streamlines.
Solved field and measured points, Case K, section
Figure 5. Case K, the centre plane \(y = 0\): (a) \(u\) over \(U_H\); (b) the same with iso-lines every 0.1 and the edge of the reversed flow, \(u = 0\), dashed; (c) the in-plane speed with streamlines.

Scores

Quantity: \(U\), \(V\), \(W\) over the stated \(U_H\) = 2.33 m/s and \(k\) over \(U_H^2\) (split fibre, 203 points); \(c^\ast = C/c_0\) (\(c_0 = \dfrac{q}{U_H H^2}\) = 429.2 ppm) against the model's \(c\thinspace U_H H^2/Q\) for turbulent Schmidt numbers 0.3, 0.5 and 0.7; the four points within one cube height of the source are scored with and without.

Score R \(R^2\) FAC2 FB Hit rate n
\(U\) +0.958 0.917 0.586 +0.594 0.123 203
\(V\) -0.151 0.023 0.054 — 0.901 203
\(W\) +0.573 0.329 0.143 — 0.685 203
\(k\) +0.330 0.109 0.044 +1.201 0.005 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.3 +0.953 0.908 0.793 -0.690 0.458 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.5 +0.951 0.905 0.685 -0.875 0.163 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.7 +0.950 0.903 0.394 -0.983 0.099 203
\(c^\ast\), \(\mathrm{Sc_t}\) 0.3, without the 4 near-source points +0.923 0.853 0.799 -0.218 0.467 199
\(c^\ast\), \(\mathrm{Sc_t}\) 0.5, without the 4 near-source points +0.941 0.886 0.698 -0.430 0.161 199
\(c^\ast\), \(\mathrm{Sc_t}\) 0.7, without the 4 near-source points +0.923 0.851 0.402 -0.561 0.101 199
\(U\), y-symmetric ceiling (control) +1.000 1.000 1.000 +0.000 1.000 203
\(c^\ast\), y-symmetric ceiling (control) +0.997 0.994 0.995 +0.000 0.793 203
  • Normalisation: \(U\), \(V\), \(W\) by the stated \(U_H\) = 2.33 m/s and \(k\) by \(U_H^2\), the same divisor on both sides; \(c^\ast\) is the measured \(C/c_0\) (\(c_0 = \dfrac{q}{U_H H^2}\) = 429.2 ppm) against the modelled \(c\thinspace U_H H^2/Q\) (full scale, \(Q = 1\thinspace\mathrm{kg/s}\)) for turbulent Schmidt numbers 0.3, 0.5 and 0.7, with one detection limit of \(10^{-3}\) of the largest measured value; hit rate \(\delta_{\mathrm{rel}} = 0.25\), \(\delta_{\mathrm{abs}} = 0.05\) in \(U_H\)
  • Per-point values: /api/cases/K/pairs/openfoam12-realizableKE.csv
  • Reproduce: Eddy3D fixture Eddy3DCli.Tests/Validation/TestAijCaseKContainerE2E (9 x 9 cubes, floor point source, 1:100; OpenFOAM 12 realizable k-ε, numerics level 3, 3000 iterations, 972,637 cells, solved 2026-10-18); figures validation/plot_cases_gk_openfoam.py

Notes:

  • The inflow is the measured approach profile (3.23 m/s at 0.2 m, \(z_0\) fitted to 13.5 mm). The stated \(U_H\) = 2.33 m/s is not on that profile: the solved \(U\) at \(z = H\) is 1.97 m/s, 0.85 of the stated \(U_H\). Measured and modelled velocities are both divided by 2.33 m/s.
  • The inlet \(k\) from that log law, \(u_\ast^2/\sqrt{C_\mu}\) = 0.77 m²/s², is about 1.7 times the tunnel's 0.33 to 0.45.
  • The first measurement point is the source location: the four points within one cube height of the floor outlet hold the four largest \(c^\ast\) (8.73, 5.31, 3.02 and 2.86 against 1.03 for the next). The model's source is the 4 cells (0.75 m across) that a 0.44 m square cuts; the outlet is 5 mm. Those four points are scored with and without.
  • 47 of the 203 points (the \(z = 4\thinspace\mathrm{mm}\) plane, 0.4 m at 1:100) lie inside the first ground cell (0.735 m): their modelled values are that cell's, pulled toward the wall's zero by the point interpolation.
  • The measured plume is not y-symmetric (C at -y exceeds C at \(+y\) at 74 of 74 mirror pairs; its lateral centroid lies 0.011 to 0.017 m toward -y at every station) and the solve is. The 'y-symmetric ceiling' controls score the measurement against the mean of each mirror pair.
  • \(V\) has a constant offset and almost no antisymmetric part; a model with \(V\) = 0 everywhere hits 0.91 of the points at \(\delta_{\mathrm{abs}} = 0.05\thinspace U_H\).
  • The solve stopped at the 3000-iteration limit with the residualControl unmet (no 'solution converged' line); the epsilon initial residual at the last iteration was 7e-3.
  • The air over the roofs outruns the street along the wind, which outruns the cavity across it; the plume leaves the source, is clean 1 to \(1.5\thinspace H\) ahead of the array, decays through four cavities and is cleaner over the roofs than in them; the solved velocity is y-symmetric to twice the tunnel's own mirror-pair difference.
Interpretation: Case K scores

The modelled velocity sits under the measured mean. The y-symmetric ceiling is the best a y-symmetric model can do on this file. \(V\) does not score the flow's divergence. \(V\)'s hit rate says nothing.

Not run

  • A second mesh or iteration count.
  • A finer source.
  • OpenLB, FluidX3D and the ML surrogate.
Interpretation: Limits of these runs
  • With one mesh and the residual control unmet, none of these numbers is mesh- or iteration-independent.
  • The concentration near the source depends on a source zone of four cells.
How the scores are computed
  • Scores recomputed from the pairs by validation/dispersion_metrics.py; the SHA-256 of every file the snapshot was read from is in its results file, under sources.

Leaderboard

Rank Run Source Hit rate, \(U\) Hit rate, \(V\) Hit rate, \(W\) Hit rate, \(\lvert\mathbf{U}\rvert\) Hit rate, \(k\) Hit rate, \(c^\ast\) Hit rate FAC2 R \(R^2\)
1 OpenFOAM simpleFoam (Eddy3D Outdoor case writer) (eddy3d-of12-casek-workbook-inflow) community 0.562 0.877 0.670 0.507 0.103 0.399 0.520 0.464 +0.596 0.525
2 OpenFOAM simpleFoam (Eddy3D Outdoor case writer) (eddy3d-of12-nutcap-lifted) community 0.232 0.901 0.680 0.192 0.039 0.478 0.420 0.415 +0.616 0.543
3 OpenFOAM 12 · \(\mathrm{Sc_t}\) 0.3 Eddy3D 0.123 0.901 0.685 0.084 0.005 0.458 0.376 0.363 +0.602 0.530
4 OpenFOAM 12 · \(\mathrm{Sc_t}\) 0.5 Eddy3D 0.123 0.901 0.685 0.084 0.005 0.163 0.327 0.345 +0.601 0.530
5 OpenFOAM 12 · \(\mathrm{Sc_t}\) 0.7 Eddy3D 0.123 0.901 0.685 0.084 0.005 0.099 0.316 0.296 +0.601 0.530

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).