@Mohamed Hamdy Check out this blog post which highlights the fundamentals of CFD and books to get started with. It has pros and cons contrasted well. https://cfd.university/blog/how-to-get-started-with-computational-fluid-dynamics-cfd/ Hope this helps!

Hi Umer, Can you give a bit more context about the problem? What is it, what output are you expecting to analyse, what boundary conditions and solution methods are you using?

@Umer Hayyat Mo's method below is a good shout. More often than not it is the boundary conditions that affect this behaviour in my experience. But, I would also maybe try modifying your under relaxation factors and the discretisation of your gradient and k. For temperature, if you are inputting it as a boundary condition, I would try to use it as a heat flux if possible. That is more robust in solving than a direct temperature specification

Can you specify more details? what software are you using? What is the problem, your setup and details of discretisation (if possible) and what you expect to see?

@Sadees PalaniVel Heyo. You have not fully specified your issue, So I cant correctly help you but I'll suggest some general things you can try. Iam assuming your CFL is fixed between 0.3-0.5 (I dont know why you want to fix it that low but OK!). Check your under relaxation factors (try lower values or higher than 1 to over-relax). If your heat transfer is specifed via temperature Boundary condition, I would suggest changing to a heat flux (if possible, this would result in a better convergence). If none of this works, then you have to look at "Solution Methods" and I would recommend with Coupled Solver, 2nd order Upwind for all terms with least square or green gauss node based for spatial terms. If this also is not fully convergent, then you can try Pseduo-transient or go into Multi-grid methods. Before all of this, I would highly recommend checking your mesh and your turbulence model (if its turbulent flow, try K-epsilon (realizable) or K-omega. Spallart Allmaras is fine as well.

Hello Yasin. While I dont have much experience in fine tuning the criterias in OpenFOAM, I found the following resource that might be helpful to your issue. Please have a look and hope this helps! https://cfdmonkey.com/tips-and-tricks-for-significantly-improving-your-snappyhexmesh/ https://cfdmonkey.com/snappyhexmesh/

Hello. I would say no. Primarily because the scales of motion you are resolving in RANS VS LES is highly different. Even within RANS your choice of mesh depends on the RANS model, your boundary wall resolution (y+) and the mesh refinement based on what flow feature you want to analyse. The same criteria also applies to LES and DES to a more strict standard. There are some onlne resources that can aid in determining your mesh criteria's for LES and DES based on the subgridscale model you choose.

@Sadees PalaniVel Heyo. I would advise you to search up your requirement up on the internet as sometimes the advice can vary based on your problem (pipe, combustion, channel etc..). As a general rule, this ANSYS guide is in my opinion, a good starting point. https://www.tfd.chalmers.se/~lada/comp_turb_model/postscript_files/Quick_Guide_to_Setting_Up_LES_version_1.4_for_Lars.pdf Let me know if that helps!