CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   triangular mesh vs. quadrilateral mesh (http://www.cfd-online.com/Forums/fluent/45043-triangular-mesh-vs-quadrilateral-mesh.html)

kim June 15, 2007 11:39

triangular mesh vs. quadrilateral mesh
 
hi!

first of all, i am very sorry for my poor english.

i receive review comment of my journal paper from reviewer.

a reviewer ask me as this.

"Is there a particular reason why the quadrilateral mesh was chosen for the Fluent simulations? Further explanation may be helpful to readers who are not familiar with Fluent."

Frankly, i have not particular reson of quadrilateral mesh,

but i can't write this.

So, I want know advantage of quadrilateral mesh against triangular mesh.

thank you very much. ^^;

Phil June 15, 2007 11:48

Re: triangular mesh vs. quadrilateral mesh
 
kim,

You should try using Gambit and creating a tetrahedral mesh and comparing it with your quad mesh. By examining the mesh you'll see that the quad mesh has far far lower skewness which improves quality and convergence as well as avoiding errors.

I'm currently running a ~570k cell gas turbine model and have had a very very frustrating few days. With denser grids closer to important areas, the tet grid gave me lots of problems. Turbulent viscosity would grow and grow until Fluent crashed. Look at my recent post called 'second order convergence problems'.

I needed alot of cells close to the middle of cylinder model that is a 45 degree segment so tri elements seemed good. This was I reckon what caused my problems. I've ended up with all quad elements except the fuel injection where I have some tri elements but carefully done with the worst cell skewness 0.74. It's iterating as we speak so fingers crossed!

Phil

Sham June 17, 2007 20:18

Re: triangular mesh vs. quadrilateral mesh
 
To my understanding (or at least to the model that I am investigating), quad mesh gives me better control of the mesh and also will reduce numerical error so to say that the results obtained will mostly due to the floe behaviour not the numerical error. Well, that of course depends on your model.


All times are GMT -4. The time now is 07:48.