Why do prism layers collapse, disappear or become irregular?

Problem: the requested prism layers are not the prism layers you get

Common symptoms include fewer layers in tight regions, sudden termination, pinched prisms, highly stretched transition cells, gaps with no near-wall layers, or local areas where y+ is much worse than expected.

A global layer count can therefore be misleading. What matters is whether the regions controlling wall shear or heat transfer retain a coherent near-wall mesh.

Check the near-wall geometric budget

1. Total prism thickness versus local gap.

   Two opposing walls each growing a thick prism stack can consume the entire passage. Leave enough core region for a valid transition unless the method intentionally spans the gap.

2. Curvature radius versus surface cell size.

   Coarse surface cells around tight curvature can force prism normals to intersect or rotate too quickly.

3. Concave corners.

   Layers growing from intersecting walls compete for the same space and may require local truncation or corner-specific treatment.

4. First layer and growth ratio.

   A tiny first cell combined with many layers and aggressive growth can create a surprisingly thick outer layer and difficult transition.

Likely cause groups

Diagnose where and why layers are being lost

1. Inspect actual layer count spatially.

   Do not rely only on requested settings. Visualize cross-sections through corners, gaps and high-curvature areas.

2. Compare local gap to twice the requested total thickness.

   This simple geometric check often explains opposing-wall collapse immediately.

3. Reduce one parameter at a time.

   Test fewer layers, lower growth, lower total thickness or finer surface mesh separately so the root cause stays visible.

4. Check whether the lost layers matter physically.

   A collapse at a stagnant symmetry corner is different from loss on a heated wall, blade surface or high-shear passage.

5. Re-evaluate y+ after the repaired mesh solves.

   Geometry determines the mesh; solved wall shear determines the actual y+.

Fix order

1. Repair geometry and surface topology.

   Eliminate small defects before tuning extrusion controls.

2. Refine the surface where curvature or gap resolution demands it.

   Prism quality depends on a credible surface mesh.

3. Reduce total thickness in constrained regions.

   Use local controls rather than weakening the whole domain when possible.

4. Moderate growth rate and layer count.

   Preserve useful wall-normal resolution while smoothing the transition to the core.

5. Recheck first-cell requirement and y+ target.

   Do not preserve a layer recipe that no longer matches the turbulence/wall-treatment strategy.

Common mistakes

• Choosing first-cell height, layer count and total thickness independently.
• Trying to force identical prism settings onto every wall despite very different local geometry.
• Increasing surface refinement everywhere instead of only where prism normals conflict.
• Accepting the requested layer count without checking the actual generated layers.
• Reducing prism thickness so aggressively that the boundary layer transitions into coarse isotropic cells too early.
• Treating a y+ target as sufficient proof of good near-wall resolution.