I am carrying out a simulation of fluid flow using helical pipes using ANSYS Fluent. I am having problems with how to get the optimum mesh and also with how to carry out a mesh independent study. I generate the default mesh and then start by changing the edge sizing to get a new mesh.
Also, from the attached file, how do I know which one the best mesh generated is?
Try to observe the flow and pressure values at some designated points. Then change the mesh size and obtain the converged solution and compare the vaues with previouly recorded. At some mesh size you will find no change in those values and this could be considered as Grid nIdpendent Solution
I just went through your results which you are provided in doc file so from just visual comparison I can say that the best mesh size is Mesh-3 in which around 1.3 million nodes because after that the results are almost constant (not much changes predicted ) while in the mesh -5 the mesh size is double but the consequence results are not much changes so there is no need to use that mesh for further study hence the optimum size of the mesh is mesh-03.
On behalf of provided data i assumed that every result were calculated well which leads the optimum mesh size is mesh-3 although i feel there is uncertainty in your results it might be because you did mistakes to calculate because you didnt describe here might be you averaging the quantity or something else so i cant comment about that but you should plot the graphs of properties for different mesh size and from that graph you will able to judge the independent grid size and this is the best way to trace the changes.
good luck
regards
S Shukla
Hi,
The mesh independence has been well explained above by both the researchers. I would like to further comment to what has been mentioned by Sushil. The discrepancy is in velocity which I assume you may have taken to be massFlowave at outlet. Since it is highly dependent on two parameters (Velocity and density), I would suggest you take some parameter which is individually varying.
Secondly, reg ideal mesh, I would suggest you may use Multizone and slicing feature to get mesh of high orthogonality (and low skewness).
In case you are not using Ansys Workbench, then you may refer to ICEM CFD for high quality mesh.
Hope this helps
your velocity and Average Nusselt number are behaving differently. You have not mentioned y+ values which affects wall effects . If there was a domain layout with boundary conditions it would be easy to diagnose the problem correctly. Many times mesh misguides the results because of some poor quality cells. Avoid automatic grid and plan a structured meshing.Bring down the skewness and check smoothness of grid. Instead of trying different grids, use a coarse grid and then do 10% grid adaptation based on the gradients of the desired property . This will give you better grid than trying different sizes. Grid independence study is half the story which takes care of only numerical errors
For grid generation of complex moving and sliding numerical grids of helical structures like screw compressors, pumps etc... You can use SCORG(c) http://www.staff.city.ac.uk/~ra600/DISCO/DISCO%20Intro.htm
Regards
I think you should have another criterion like determinant 2*2*2 to judge better.
Grid independence test is not always a method to say that your result is near to the experiment results....I highly agree with Mahesh K. Varpe. Check firsth Y+ with the turbulence model with mesh size.
Hi everyone, I have similar question but I am using comsol. My results are mesh dependent, but I want to achieve same maximum velocity... How can I do this, Also, I want to check for y+ , but I don't know how to do this in comsol
Hi, we know the problems students facing in ICEM CFD. We have helped many students in this regard.
Here is our course published on udemy on ICEM CFD Hexa. Hope this will help students in learning meshing.
https://rebrand.ly/ICEMCFD
- Badges
- Science topic