Using Ansys Fluent Mosaic Meshing Technology

September 24, 2020 Alex Lefebvre

For some odd reason I have always liked meshing. I remember the good old PhD days when my supervisor told me: “if you can mesh that thing in a week, I’ll open a bottle of Champagne”. In retrospect, I should have known he meant sparkling wine at best but still, the challenge was on! Using Pointwise structured mesh topology, I set to produce the mesh starting with generating a point repartition on a couple thousand edges to produce the surface mesh. Only after checking the surface mesh both visually and through quality metrics could I start working on the volume grid. I had two best friends at the time - patience and the elliptic solver. You guys remember the elliptic solver, right? Well, as I said before, those were the good old days, and I have a new pal now: the mosaic enabled Fluent mesher!

Picture1

A 2-D multi-block structured grid for a multi-element airfoil generated with Pointwise. The software generates high quality structured grid at the cost of significant engineering time. Image Source: Pointwise website.

But let’s back up a little. After the PhD, I was incredibly lucky to land a first job in Formula 1 racing where I worked as a CFD expert for four years. During that time, I was introduced to both PowerFLOW and Star CCM which I used daily to, you know, launch 10’s of simulations. No big deal! Coming from an in house CFD code programmed in Fortran and Pointwise mesher, those codes seemed incredible to me. In PowerFLOW, I did not have to specify any edges at all. I would extrude car parts to generate “VR” regions (volume refinement) that would be used by the code to locally refine the cartesian grid. This was still a time-consuming process – trust me, there are quite a few parts on an F1 car! –  and extruding those parts was difficult to automate. But still, it was much faster than what I was used to. With STAR things got even better, for the first time I could set only a couple of global mesh settings and a couple of volume refinement boxes to generate a decent F1 mesh. The mesh quality and speed was decent enough with the exception of the poly cells which were quite slow to generate and some areas on the car suffered from mesh quality issues at the attachment between the prism layer and the hex layer. I want to make it clear that none of the software cited above are bad. On the contrary, Pointwise is an incredible tool for research purposes when the last 0.1% accuracy matters. Star CCM and PowerFLOW brought to the table meshing process enhancement that made those packages competitive, at the time, for F1 simulation despite limitation in the time it took to generate a mesh and the final quality of the mesh.

Enter the Ansys Fluent mesher.

The first time I utilized the mosaic-enabled Fluent poly-hex mesher from Ansys, was the last time I used any other mesher in a production case. The mosaic mesher links a poly-prism layer to a hex layer conformally using a poly layer in between. The connection utilizes a low cell count and has high quality. The task-based mesh generation is intuitively easy, surprisingly fast and the default settings are usually spot on for a (coarse) production case. Exit the bad prism to hex connection. Exit the slow poly generation. Finally a simple tool that produces quality mesh fast and has the added bonus of producing meshes gorgeous to look at. I’ll let you be the judge of that last statement with the image below showing the hexagonal surface discretization in grey, the poly-prism layer in purple, and the transition poly layer and hexahedra layer in light blue.

Picture2A 3D airfoil grid is generated with minimal user input utilizing the mosaic enabled Fluent Mesher.

If you too want to have a shot at mosaic meshing, launch Fluent 2019R3 or later and select Mesh in the launch window.

About the Author

Alex Lefebvre

CFD Simulation Specialist<br><br>Alex collaborates with clients to find solutions that meet their needs, whether it be new software, technical training, or consulting services. He completed his Ph.D. in Mechanical and Aerospace Engineering at the University of Miami, with a focus on aircraft design. After the completion of his Ph.D., Alex worked in Formula 1 as a CFD expert where he helped the team produce a car that finished 5th in the world's championship.

Follow on Linkedin More Content by Alex Lefebvre
Previous Article
Supersonic Turkey Cooking (STC) using Ansys Fluent
Supersonic Turkey Cooking (STC) using Ansys Fluent

Come Thanksgiving one can find a variety of ways to cook a turkey. The most traditional way to cook a turke...

Next Flipbook
Ansys Fluent
Ansys Fluent

Get help from CFD experts

Learn More