Pressure Drop in Turbulent Flow Through a Pipe

  • 1.5K Views
  • Last Post 05 June 2018
  • Topic Is Solved
adrianwojtas posted this 29 May 2018

Hello everyone,

I am introducing myself to AIM for future use needs (hydraulic equipment design) and decided to start with some simple ball valve models that can be obtained online from their producers. Yet my simulations couldn't get close to the official results of pressure drop/volume flow characteristics given on technical datasheets for these valves.

So I tried to look for verification cases for ANSYS and found them in some documentation files on ANSYS site. I wanted to start with something completely easy and decided to go : Pressure Drop in Turbulent Flow Through a Pipe

Test Case

I designed simple pipe 2 m length, 2 mm diameter and set physics in AIM as stated:  inlet vel = 50m/s, outlet press = 0 Pa, also I didn't forget to set right material properties as they differ a little bit from presettings in AIM for air.

Results I get are again way over what I'm expecting, in comparison to pressure drop from verification model: ~21,5 kPa, my results vary from 80 kPa to 50 kPa depending on mesh level (results read as geometric average of total pressure variable for inlet and outlet faces).

Results from AIM

Can someone help me get to the reason of my constantly wrong results? I feel like I'm doing some basic mistake on and on..

Thanks in advance,

AW

 

BTW: File extension of AIM project: ".wbpj" was not allowed to be uploaded together with this post

 

Attached Files

Order By: Standard | Newest | Votes
raul.raghav posted this 29 May 2018

Not sure what went wrong in your setup. I tried downloading your attachment but its just a zip file of the .wbpj file without the contents of the workbench project. Look at the attached screenshots to create the workbench archive file .wbpz which would include all the workbench project files. You can attach the .wbpz file directly to your post.

Rahul

  • Liked by
  • adrianwojtas
adrianwojtas posted this 30 May 2018

Thank you Rahul for your response and tip for exporting the project. I attach archive as instructed

 

Edit: or actually I can't attach because somehow the file is 300MB+ size big... so temporarily I put there a link to download it from elsewhere, while I'll look for the reason of that project size - for now it looks like in "results" folder file 1.1.1.h5 is 243MB big ...

https://www110.zippyshare.com/v/liFx4vzu/file.html

 

Edit2: after deleting almost all plotted stuff in result's page and saving the project, the mentioned h5 file did not change it's size

peteroznewman posted this 30 May 2018

Adrian,

1) There is a 120 MB file size limit on attachments to these posts.

2) The Mesh is what makes the archive file size large. Go into Mesh and make some edit that will delete the current mesh, but don't mesh. Save that project archive and it should be smaller.

3) Please reply with what version of AIM you are using: 18.2 or 19.0

4) I have not been successful opening archives of AIM after transferring them from another computer. You might need to export the Geometry and describe the boundary conditions so a model can be built from scratch by someone else.

 

adrianwojtas posted this 30 May 2018

Hello Peter!

 

I am using AIM 19.0, the geometry is exactly what has been explained in a mentioned verification case model so it's a simple 2mm diameter pipe, which is 2m long and boundaries are inlet velocity = 50m/s and outlet pressure = 0 Pa, that makes this model extremely uncomplicated yet I can't get to results posted in verification case (and it's also what happens with much more detailed but still simple valve models that I download from websites of well known companies like Danfoss, Oventrop, Heimeier etc..)

I've reduced mesh to smallest level and it's about 60MB now, thanks for the tip!

I also attached geometry model exported from SpaceClaim

Attached Files

raul.raghav posted this 30 May 2018

I looked into your files and would like to point out the following issues:

1. Geometry: The verification model geometry has a radius of 0.002 [m] while your model has a diameter of 0.002 [m]. So clearly the pressure drop would not match the verification case. The pressure drop in your case (with D = 0.002 [m]) should have been 52.4 kPa, which is in the range you were getting (50 to 80 kPa). You can use an online calculator to do a quick calculation of this number: Pressure drop calculator

2. Mesh: I notice that the "Element Shape" is set as tetrahedrals elements which is not required for this simple geometry. So delete the "Element Shape" option (see screenshot below). Just the default sizing and a boundary layer would enable you to create a good hexa mesh for this geometry.

Rahul

peteroznewman posted this 31 May 2018

Rahul,

Did you open this archive using the AIM Student 19.0 installation or a full-license for AIM 19.0?  I'm having difficulty opening this archive using Student license.

 

Peter

raul.raghav posted this 31 May 2018

Peter, I opened it with the Research license. For some reason, I have a very difficult time with AIM in general. Like with Adrian's files, the software wouldn't complete the mesh. It would literally get stuck at some point during meshing. That's why I couldn't show him the final mesh in this case and just suggested him what to do.

Rahul

adrianwojtas posted this 05 June 2018

Rahul,

Many thanks for pointing my mistake out! That did the job and now my results are comparable with verification case!

Strangely Element Shape that you've pointed out was created and pre-set like that by default and when I changed it to hexa type and later deleted it, then results I was getting were a little bit worse than model ones.  

Now considering that two verification cases that I've designed are good (and that was my geometry design mistake, not mistake in the way I set up physics), I am getting more and more concerned about these simple valve models that I am trying to simulate but still shoot with results about 5x higher than expected considering experimental valve characteristics given by their producers.

Close