Cartesian mesh with inflation layer (first thickness layer)

[a20050237]

Hi everyone,

I'm simulating cavitation inside a Venturi 3Dchannel. My benchmark paper says that to couple turbulence, cavitation, and interface tracking correctly, I need a homogeneous volume mesh to avoid numerical diffusion variations.

The paper shows Inflation layers: pure hexahedrons (quads on surface) to control Y⁺ (≈1.5) with a specified First Layer Thickness.
Core flow: uniform Cartesian/hex grid.

The problem:

    Ansys Meshing (Cartesian method): When I use the Cartesian method, it kills my inflation layers – reduces 10 layers to just 3, or fails completely. It can't grow structured layers on a thin 3D body with this algorithm.

    Fluent Meshing (Watertight Geometry): With Hexcore or Poly‑Hexcore volume mesh, the surface mesh is forced to be triangles. Then the boundary layers are extruded as triangular prisms (wedges), not pure hexahedrons.

I want strictly hexahedral inflation layers, no wedges. The rest of the core can be Cartesian/hex. My geometry dimensions: Length X = 350 mm, Y = 25 mm, thickness Z = 6.35 mm. I'm using high‑order discretization schemes.

My questions:

    Is there a setting in Fluent Meshing (maybe TUI commands or a different workflow) to force a quad/hex surface mesh before volume generation?

    How can I force Ansys Meshing (Workbench) to respect a 10‑layer First‑Layer‑Thickness inflation without breaking or suppressing the Cartesian / MultiZone hexa‑core mesh?

    Any workaround or alternative workflow to get this specific topology (hexahedral inflation + homogeneous hex core)?

[Essence]

Hello,

You can use Multizone mesh method in Fluent meshing. Alternatively, you can also use ICEM CFD for generating 100% hex mesh.

[a20050237]

Many thanks for your answer. I'm simulating with multizone and inflation, and I will see how it works. I haven't tried to use ICEM yet.

[Rob]

When was the paper written to require a full hex mesh? Where do you expect to see cavitation? If that occurs in the near wall region, what do you think the inflation aspect ratio will do relative to the gradients? 

[a20050237]

Hello,

The paper I'm referring to is: "CFD Turbulence Models Assessment for the Cavitation Phenomenon in a Rectangular Profile Venturi Tube" – Fluids (2024, 9, 71). It was published in March 2024.

The cavitation occurs in the near‑wall region – specifically at the Venturi throat, where the pressure drops below vapor pressure and bubbles start to form. That's exactly why controlling the boundary layer (Y⁺ ≈ 1.5) with a well‑defined first layer thickness is so critical.

Regarding the inflation aspect ratio: a high aspect ratio (e.g., first layer = 0.015 mm, growth rate = 1.15, 10 layers) is actually desirable here. The velocity and pressure gradients are much steeper in the wall‑normal direction than in the streamwise direction within the boundary layer. A high aspect ratio helps me resolve those steep gradients efficiently without using too many cells. The challenge is not the final cell shape, but getting the mesher to respect the requested number of layers and the first layer thickness.

I'm currently testing the Multizone method, but I had to reduce the first layer thickness from 1.5e‑5 m to 5e‑6 m to get a Y⁺ around 2.

Thanks again for the help!

[Rob]

You're welcome.

Agreed re the velocity gradients, but if you're also cavitating in the near wall region the phase/pressure gradient is significant in the streamwise direction. It's all very well focussing on y+ but you can't neglect streamwise in this scenario.