Discussion Groups

Simulation CFD

Reply
Contributor
mikel_29_mikel
Posts: 13
Registered: ‎03-20-2013

Problem with analysis in CFD

296 Views, 10 Replies
06-05-2013 12:02 AM

Hello everybody,

 

I have a problem and I need help. I have a design that I try to simulate it. I define the materials of parts, I create two parts of fluid (one water and one air) but when I start the analysis on Flow and Thermal with Autoforced convection, gravity and radiation, the message  "The Solver closed unexpectedly..." appear me.

 

Then I try to simulate the same example without Radiation and I get a good result, the problem is that I need use Radiation also. Why is it happening??

 

I need answers please!!

Thank you!!

Please use plain text.
Product Support
apolo.vanderberg
Posts: 299
Registered: ‎08-31-2011

Re: Problem with analysis in CFD

06-05-2013 07:34 AM in reply to: mikel_29_mikel

Solver exit is a generic error message.

You can check the status file to see if there is more specific detail in there.

 

I'm assuming taht the air and water fluids are not touching as this is not supported by SimCFD

 

If it is only failing with radiation, but not when you have radiation off, it could be a resource issue as Radiation will increase the resource use of the solver, and if you have a large element count and insufficient RAM we could be erroring/exiting due to lack of ram.

Please use plain text.
Contributor
mikel_29_mikel
Posts: 13
Registered: ‎03-20-2013

Re: Problem with analysis in CFD

06-07-2013 03:32 AM in reply to: apolo.vanderberg

Yes, the fluids are not in contact. They flow through differents parts but these parts belong to the design.

 

Maybe the problem can be the RAM, but my computer is:

AMD A4-5300 APU with Radeon(tm) Hd Graphics 3.4Ghz

8GB RAM, Windows 8 x64

 

And yes, only it happens with Radiation, if I use convection it works good, if I don't use anything it works good, only when I use Radiation the problem appear.

However, I did other model more simple and I put Radiation, convection, flow and gravity and this scenario works very fine.

My design (the first) is big, but I have simulated other designs without problems with the same size.

I don't understand anything. :mansad:

Please use plain text.
Product Support
apolo.vanderberg
Posts: 299
Registered: ‎08-31-2011

Re: Problem with analysis in CFD

06-07-2013 05:50 AM in reply to: mikel_29_mikel

With 8GB of ram that would mean you could solve at most ~4M elements. If dealing with Radiation this number would actaully be smaller as the radiation solver can be notably more resource intensive than the standard thermal solver.

 

Since this model runs without radiation I would assume that you have the setup correct, you may just need to find a method of strategically coarsening the mesh so that with radiation there are enough resources for the solver.

 

Apolo

Please use plain text.
Contributor
mikel_29_mikel
Posts: 13
Registered: ‎03-20-2013

Re: Problem with analysis in CFD

06-11-2013 03:29 AM in reply to: apolo.vanderberg

Ok Apolo, I will try to do a big mesh. But my design has 20 diferent parts, for this reason I don't understand the radiation problem.

I start to think that this software is very limited, because an example sooooooo easy with an only part issues hot, the rest are steel, air and water and always stop the analysis at 99% of radiation.

 

It is maddening!!!!!!!!!!!!!

Please use plain text.
Product Support
apolo.vanderberg
Posts: 299
Registered: ‎08-31-2011

Re: Problem with analysis in CFD

06-11-2013 04:56 AM in reply to: mikel_29_mikel

Mike,

I'm sure you understood my previous post. I'm not stating that you need to do a bigger mesh. I'm stating that you might not have enough resources for the mesh that you have at the moment.

 

I would suggest seeing if you can do a coarser mesh, however the problem with coraser mesh sizes is that you may lose fidelity in some of the flow solution as you try to fit the model (with radiation) in to your memory footprint.

 

Feel free to attach a support sharefile (*_Support) so i can see what you're running.

The 99% is while we are computing view factors, as we are not in to solving the solution yet which would lead me to believe its a resource issue. Whether this is a simple 2part assy or a complex 100part assy doesn't always make a difference in resource use as it will be a function of the mesh size.

 

 

Please use plain text.
Contributor
mikel_29_mikel
Posts: 13
Registered: ‎03-20-2013

Re: Problem with analysis in CFD

06-11-2013 07:01 AM in reply to: apolo.vanderberg

Can you say me the steps to upload my archive?? Because i try to add a attachment and the platform say me the archive is so high. I try to compress and now the platfornm say me the extension .zip and .rar are not accepted.

Please use plain text.
Contributor
mikel_29_mikel
Posts: 13
Registered: ‎03-20-2013

Re: Problem with analysis in CFD

06-12-2013 08:26 AM in reply to: apolo.vanderberg

Or can you give me your e-mail? So, I can send you my archive and you can see the problem.

I have to say that tomorrow I have been doing other design very similar and the problem with Radiation continue. I don't understand anything, with 8GB I have enough memory.

 

If I had taken into account the specifications about software (2GB RAM is enough) I don't run the software.

I expect your answer please!!

Please use plain text.
Product Support
apolo.vanderberg
Posts: 299
Registered: ‎08-31-2011

Re: Problem with analysis in CFD

06-12-2013 08:48 AM in reply to: mikel_29_mikel

Mike,

  No the specifications of 2GB does not mean that you can run full production level models with Radiation or complex physics. It does mean that you can run some models and get use out of the tool. We do have users that are in that 2-4GB range and leverage the tool quite a bit, however they do have to take in to account that they can easily run in to resources limitations and account for that accordingly.

 

As I mentioned in one of my previous posts. Our typical memory usage is ~2GB per 1million elements. So for 8GB machine at most you would be able to run a 4M element flow model. Quite possilby you would have to be less than that considering the Interface and Windows takes from that 8GB. As you add complexity (such as the Radiation solver, which is memory intense) this will also limit the model size even further.

 

How large is the model that you are currently running (from the output bar or the status file not the Approx element Count on the mesh dialog)?

 

Have you tried the suggestion of coarsening the mesh to try and fit the model within your resources?

 

Apolo

 

Please use plain text.
Contributor
mikel_29_mikel
Posts: 13
Registered: ‎03-20-2013

Re: Problem with analysis in CFD

06-13-2013 08:00 AM in reply to: apolo.vanderberg

Hello Apollo!

Today I uninstalled all the Autodesk sotfware, then I restored the computer to factory conditions and went to install the software, but the problem persists. This is crazy desperate.

I do not understand how two months ago with the CFD 2013 I could do these things and now with the CFD 2014 I can not do anything. To top before installing the CFD CFD 2014 I installed the 2013 and now with him I can not. I'm sick of the famous Solver.

I need you to tell me how to get on the file so that you can help me, because that's coarsing I do not see, that is, I do not find where it is, nor how. On Mesh Size -> Edit, there is a bar to define "Coarse-Fine" but she is disabled.

Please tell me how I can send youthe archive, for mail, Dropbox or whatever, but I need to know why it fails and you appear to be you know look for the problem.

I await your reply, thanks.

Please use plain text.