LaminaroStudy quality
Hot and cold water mixing in a tee.ipt · Example · Hot and cold water mixing in a tee · Run 2 · 2026-10-29

Study quality and what to do next

21 checks on this run: 17 fine, 4 worth a look.

What to do next

1Check the mesh: one run on a coarser mesh shows how much the cell size moves the result.Check the mesh
2.5× fewer cells than this run, about 11 min; this run stays as it is. Then optionally a finer one for the full check (Verify ×3).
2.5× fewer cells than this run, about 11 min; this run stays as it is. Then optionally a finer one for the full check (Verify ×3).
2Change the near-wall cell size (finer to get below 5, or coarser to get above 30).
y⁺ is about 17.22: between the two ranges the wall model is least accurate.
y⁺ is about 17.22: between the two ranges the wall model is least accurate.
fineworth a lookact on it

Convergence

Precision ?PrecisionGood enough for engineering use. The pressure drop is steady within ±0.2 %, the flow through inlet 1 within under ±0.1 %, the flow through inlet 2 within under ±0.1 % (last 200 iterations); mass balance 0.00 %. The remaining unsettledness is mostly next to the walls.good enough for engineering use
StoppedStopped at iteration 387: the pressure drop is steady within ±0.2 % over 200 iterations; the residuals levelled off at about 2 × 10⁻², so parts of the flow keep moving (likely unsteady regions): fine for the pressure drop, not for the detailed flow pattern.
Pressure drop ?The ± on a resultHow much it moved over the last 200 iterations. ≤ ±1 %: good enough for engineering use · ±1–3 %: trends and comparisons · above ±3 %: not settled yet. The mesh's own error is not in it: Check the mesh for that.±0.22 %
Flow through inlet 1 ?The ± on a flowHow much it moved over the last 200 iterations. ≤ ±1 %: good enough for engineering use · ±1–3 %: trends · above ±3 %: not settled yet.How much it moved over the last 200 iterations. ≤ ±1 %: good enough for engineering use · ±1–3 %: trends and comparisons · above ±3 %: not settled yet.< ±0.1 %
Flow through inlet 2 ?The ± on a flowHow much it moved over the last 200 iterations. ≤ ±1 %: good enough for engineering use · ±1–3 %: trends · above ±3 %: not settled yet.How much it moved over the last 200 iterations. ≤ ±1 %: good enough for engineering use · ±1–3 %: trends and comparisons · above ±3 %: not settled yet.< ±0.1 %
Total flow ?Total flowHow much it moved over the last 200 iterations. ≤ ±1 %: good enough for engineering use · ±1–3 %: trends and comparisons · above ±3 %: not settled yet.< ±0.1 %
Mass balance (inflow vs outflow) ?Mass balanceInflow against outflow: below 0.5 % means the flow field is consistent. A larger gap: the run has not settled, or the mesh leaks at an opening.below 0.5 % = the flow field is consistent0.001 %
Where it is still settling ?Where it is still settlingThe part of the flow that still changes between iterations; the 3D view colours it (Colour → Where still settling).the 3D view colours it: Colour → Where still settlingThe remaining unsettledness is mostly next to the walls.
Residuals at the end ?ResidualsHow much the solver's equations are still out of balance, scaled. Low (below 10⁻³) or no longer falling, with steady key results, means the run is done. A hint; the ± on the key results is the measure that counts.Residuals measure how far the solver's equations are from being exactly satisfied in each cell; they shrink as the solution settles. Here they levelled off at about 1 × 10⁻² instead of below the usual 10⁻⁴, so the flow field may still change a little. The key result moved by ±0.2 % over the last 200 iterations, so it is settled.pressure 7.7 × 10⁻³ · velocity 1.5 × 10⁻²
10⁻³10⁻²10⁻¹10⁰100200300iterationpressurevelocity

Mesh

Cells ?CellsThe number of cells the flow was computed on. More cells: finer detail and less mesh error, but a longer run. Draft is for set-up checks and trends, Standard for numbers you report, Fine to confirm them.530 900
Mesh level ?Mesh levelDraft and quick look give trends; Standard is for numbers you report; Fine confirms them.Standard
Max non-orthogonality ?Non-orthogonalityHow far the line between two cells' centres leans away from the face between them (0° = square on). Below 65°: good. 65–75°: fair, the solver corrects for it. Above 75°: poor: less accurate gradients, slower to settle, sometimes a blow-up. If poor: a finer mesh where it happens (sharp edges, thin gaps), or fewer wall layers there.below 65° good, above 75° poor (less accurate, slower to settle)55°
Max skewness ?SkewnessHow far the face between two cells sits from the midpoint of the line between their centres, relative to the cell size. Below 4: good. 4–8: fair. Above 8: poor: values on the faces are interpolated wrongly. If poor: a finer mesh near curved walls and small features, or smooth tiny details away in Inventor.2.391
Mesh checksall passed
Wall layers ?Wall layersThin cells stacked along the walls, where the velocity changes fastest. Fewer than planned show up at sharp edges and complex corners: the wall friction there is less accurate.thin cells along the walls; fewer than planned where the surface is sharp or complex3.0 of 3 on average
y⁺ on the walls (mean) ?y⁺ on the wallsThe first cell's distance from the wall in viscous units. The wall functions used here want about 30–300; below 5 also works with k-ω SST; 5–30 is the least accurate band.the wall functions used here want about 30 to 300; below 5 also works with k-ω SST; 5 to 30 is the least accurate band17.22 (min 2.062, max 68.27)

Fluid region

Backflow at outlet ?Backflow at an outletFlow coming back in over part of an outlet: the outlet is too close to a change of section or to a wake, and its condition distorts the result.flow coming back in at the outlet means the outlet is too close to a change of section or to the wake0 % of its area

Physics

Simulated as ?Simulated asyour choice in 6 Run: the solver looks for one time-averaged flowsteady
Flow found ?Flow foundfrom the run: the residuals levelled off at about 2e-02, mostly next to the walls: the key result is fine; a transient run would show the movementpartly unsteady
Turbulence model ?Turbulence modelLaminar below Re ≈ 2 000 in a duct; k-ω SST the general choice for turbulent flow; k-ε realizable a robust older model. Between 2 000 and 4 000 every model is uncertain.Re ≈ 5·10⁴: turbulentk-ω SST (automatic)
Inlet turbulencesuggested by Laminaro (5 % in ducts, 1 % in free air)

Every check on this page is made by Laminaro's code from this run's own files (solver logs, the mesh report, the saved fields); the guidance values are common CFD practice.