Axial Flow Pump CFD - Meshing: Non-orthogonality Issue

[jshakar]

I am having an issue with meshing an axial flow pump I have designed. This is the gen 2 design where I have setup a boundary box to represent incoming flow (think of a jet ski on water, the box is the free stream water).

Link to gen 2 project: SimScale Login

I have ran CFD on a gen 1 design design without the incoming flow boundary box and it meshed fine.

Link to gen 1 project: SimScale Login

Issue: When meshing my gen 2 setup, I am getting non-othrogonality issues:

I have tried 3 different meshes to see if I could improve it.

Mesh 3: everything kept the same as gen 1 mesh settings
Result: CFD simulation diverged.
Max. Non-orthogonality: 89.4
Number of cells with Non-orthogonality > 88 deg: 512

Mesh 1: everything kept the same as gen 1 mesh setting, but increased mesh fineness to 7
Result: worse
Max. Non-orthogonality: 89.4
Number of cells with Non-orthogonality > 88 deg: 543

Mesh 2: everything kept the same as gen 1 mesh settings, but created more local element refinements (the impeller blade tips and sides, the stator leading edge, the pump inlet where it meets the water surface, and the area between the impeller and stator since there is a fine gap there.
Result: even worse
Max. Non-orthogonality: 89.3
Number of cells with Non-orthogonality > 88 deg: 787

Unsure of what is causing the issue. After reviewing the mesh using the mesh quality checker, I can see the non-orthogonal cells. The aspect ratio and edge length inspections look fine, however the logger shows really high maximum values.

I am trying to understand what is causing this. I know sharp edges and gaps are an issue, but the gen 1 model has the same gaps and sharps corners in the same spots. The only thing that has really change (aside from some slight geometry differences) is the addition of the water box boundary.

I can “clean up” the CAD model if needed, but unsure of what to clean up. Any help regarding this would be much appreciated.

Thanks!

Pictures of the CAD Assembly as well as the flow simulation model to see clearances/gaps. The rotating zone is 0.2 mm off the impeller as taught in the centrifugal pump tutorial.

[igaviano]

Hi @jshakar, thanks a lot for posting at our forum

I apologize for the delay in the reply!

Your model looks very interesting but also complex and suited for a professional application maybe. I think one of SimScale’s professional engineers should be able to have a closer look and provide better help concerning the issues you’re facing.

A Sales Manager could get you in contact with them in no time!
Request a demo over at: Plans & Pricing | Professional, Team & Enterprise | SimScale

Best,
Igor

[jshakar]

Any advice on the non-orthogonality issue though?

[jshakar]

Screen shots of the mesh quality

[jshakar]

@igaviano I’ve attached pictures of the issue above

[beredis802]

Meshing issues, particularly non-orthogonality problems, can be quite challenging in CFD simulations, especially when dealing with complex geometries like axial flow pumps. Non-orthogonality refers to the angle between the face normal of a cell and the line connecting the centers of two adjacent cells. High non-orthogonality can lead to numerical inaccuracies and convergence issues in simulations.

[jshakar]

Understood. Just not sure how to resolve the issue. Especially as a very similar geometry had meshed previously.