I am modeling a drill bit with several nozzles with specific size. The estimated pressure drop for these nozzles is about 600 psi according to available analytical equation. However, the computed pressure drop is about 900 psi using Autodesk Simulation CFD 2013 . I ran the same model with aaproximately equal number of elemnts with Fluent and the pressure drop is close to 600 psi. Can somebody comment about the accuracy of results in Autodesk Simulation CFD 2013?
Solved! Go to Solution.
Solved by apolo_vanderberg. Go to Solution.
Hi Ali,
This is likely down to a few possible factors:
The mesh is not sufficient (unlikely as it sounds like you know what you are doing)
The model is compressible but you are running incompressible (what are the flow speeds?)
You need to use a different advection scheme or maybe material model.
Are you able to share a cfz file?
I have struggled with this as well but got great help from the support team (Marco Mueller).
What could help you is to set boundary conditions as pressure to pressure (dont forget to switch to Adv 2)
Disable automatic layer adaption, manually reduce layers and montor one of the wall model varables , Y+.
It should be within 35 to 350. At least in the areas where most pressure drop occurs
In my case reducing to only one layer and layer factor to 0,2 took me within 1,5% compared to later lab tests (oil with a pressure drop of approx 1200 psi).
Of course the normal mesh refinement was done too but didnt take me anywhere near these results
Do a search for "Y+" in the help files and read all sections, that will explain this far better than I can
Thanks akarlsso for your response. I used Adv5 which is improved version of Adv2 but I am re-running my model with Adv2 to see if I get any different result. My inlet boundary condition is constant flow rate and the outlet is gauge(0) pressure. I don't know about "pressure to pressure" boundary condition and am not sure how to appy that, can you explain more about that?
Regarding reducing the mesh layers in bondary layer, having more layers should increase the accuracy of flow prediction. How does reducing the mesh layer can help it? I know the same thing has been suggested in the manual for wall model function but I am just curious how?
Thanks
Hi Jon,
The mesh is very fine near the nozzles so I don't think the mesh refinement can solve this issue. Plus, I ran the same model with some extended nozzles and the pressure drop was about 700 psi. The nozzle extension should not affect the pressure drop that much.
I use water as working fluid and the model is incompressible. I tried to share the cfz file here but it said it's too large. Its size is about 13 Mb..
Thanks
Ali
I ran my model with Adv2 and I even got higher pressure drop(1100 psi). I reduced the layers to one and layer factor to 0.2 and the result got even worse! now the pressure drop is 1173 psi. I am really lost....
I am happy to take a look if we can find a way to share the file, Dropbox?
Hi Ali, I have it, thanks.
I am going to move this to a proper case to work on, but we can report the findings here also once it is solved
Jon,
I increased the outlet pressure drop to 2000 psi according to Appolo and your recommendations. However, I found that I could fail the solver with just doing that! The pressure residuals started acting weird and became constant in some random intervals. Even, I noticed the pressure changed at the outlet boundary condition which should be fixed to 2000 psi. I would appreciate your comment.
Hi Ali,
It is best we stick to using the case we have, as both Apolo and myself will receive notifications when you message us.
We will respond from there.
To update for other users that may follow this.
There were a couple reasons for the mismatch in Pressure results.
Namely, once we used the same model between CFD and Test lab, and used the same conditions with an appropriate mesh, we were able to achieve much better correlation with the pressure results.