[jpc]

in the help file it has an impeller example
it looks as though the impeller has been subtracted from the rotating region
http://wikihelp.autodesk.com/Simulat...1-CCD16D349875
however the sample files don't include that subtraction and i don't see that in the recommendations.

does anyone know if the help file is just not correct?

[jpc]

for anyone running this type of analysis, this is an interesting find:
http://usa.autodesk.com/adsk/servlet...inkID=18154757

[fisichel]

Hi JPC,

If I remember that example correctly, the impeller is subtracted in CAD, but then Simulation CFD detects the closed volume and fills it for you. In Simulation CFD, you define the material for that new volume to be something other than a fluid and the code will calculate the flow around it.

It's sort of a back-door way of getting geometry into the model, but I can see why they do it; the captured volumes exactly match the fluid that surrounds it.

HTH,
Chris

[jpc]

if you're using the cut-out method, would the cut-out be from the main fluid volume or from the rotating region?

it seems like the rotating region should be a full cylinder and then the geometry that is rotating within it either solid within the main volume (and the rotating region) or a cut-out from the main fluid volume (and within the rotating region). but that isn't what i'm seeing in this screenshot

http://wikihelp.autodesk.com/Simulation_CFD/enu/2013/Help/0407-Learning407/0726-Guidelin726/0765-Rotating765#GUID-982DE07C-29A7-4056-A40F-B4911028CFBE

i used a solid rotating component, full cylinder rotating region and it seemed to work ok. that was closer to the models that came with the software but different than how it shows in the tutorial. in the tutorial it seems as though there was a cutout, that was filled and it gets selected as the solid. http://wikihelp.autodesk.com/Simulat...9-39A86207D30D

that being said, i think one of my original key failures was having a cutout AND having the solid which caused crashes during meshing.

[fisichel]

Hi jpc,

When I did that example, I was using Pro/Engineer and saw how they used a boolean subtract operation to remove the impeller from the rotation region volume. I also originally would explicitly put in a solid impeller, but that turns out to be unnecessary.

With regards to your first question, the impeller cutout should be from the rotation region. Or, at least that is what I have had success doing.

I'm glad you got it to mesh. Good luck with your analysis. It sounds like you are almost there.

[jpc]

any chance you have the CAD files that you could send me? i can pm a dropbox. i'd just like to take a look at the pro/e and solidworks files. (i just deleted the software from my computer) i swear that there were no boolean ops in either the solidworks or pro/e files.

in case someone finds this, it sounds like the options are:
1. keep the solid that will rotate, generate a rotating cylinder which will also be a solid that overlaps both the main fluid volume and the rotating component. apply rotation to the rotating cylinder. (i had success with this)
2. subtract the rotating component from the rotating cylinder. the rotating cylinder will only overlap with the main fluid volume. autodesk sim will fill in the solid (and it must fill in the rotating cylinder). apply rotation to the rotating cylinder. (fisichel had success with this)

[fisichel]

Hi jpc,

I can't post the CAD. I'm sorry. I grabbed a screen shot, however, that shows how they built their model. I noticed there were two features on the tree. The first was the protrusion and the second was a cutout. I edited the cutout so you could see they selected the impeller to be subtracted from the rotating region.

HTH,
Chris

[marktm250xc]

If an airtight void is detected, it is filled and assigned a new volume during launch. Some versions of ProE in the past were very, very finicky ... if you left the impeller in, the launch or meshing would at times fail. But if you took it out and let CFdesign create a new volume, the launch would succeed.

As far as meshing the impeller, there used to be an issue in that if you did not mesh it, you would not get torque values. Pretty sure that is fixed now. However, if you choose not to mesh the impeller, animations may not appear correct.

[jpc]

hi fishchel, well that shows it. i wonder if i was just looking at something incorrectly. if i get a copy of the software or try the trial again, I'll check again.

any idea what you're supposed to do if you have a solid? if you should cut the solid from the rotating region so that it is equivalent to what you'd have if you had a void cut out of the rotating region?

[fisichel]

Hi jpc,

Yes, that would be a great first start. I started with a more real impeller. I then created a rotating region around it and I used the exact same subtract operation that AutoDesk uses.

When I brought it into SimCFD, I did not bring in the impeller, since I already had it cut out. The software filled it in for me. It looked exactly the same as the impeller I originally started with, as you would expect.

Good Luck,
Chris