Discussion Groups

Simulation Mechanical and Multiphysics

Reply
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012
Accepted Solution

Mass Transfer - Average Concentration in a Volume

250 Views, 7 Replies
08-03-2012 04:25 PM

Hi guys,

 

Hopefully this is an easy question...  Once I run the simulation and I have my results, how do I get to the average concentration withinin a volume?

 

I can select a bunch of elements and get the mean, but this doesn't take into account the volume of the elements, so the mean is skewed.  If I try to use Mean Weighted by Volume or Mean Weighted by Mass, I get an error that says I should use the Weight and Center of Gravity Tool.  When I run that, it says "No current analysis type can be determined for this model!"

 

Thanks!

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012

Re: Mass Transfer - Average Concentration in a Volume

08-07-2012 01:42 PM in reply to: tfjield

The more I look at this, the more it seems like I'm going to have to calculate the volume of the various elements and the average concentration in the volumes myself, based on the data in the database files (elements and nodes).

 

Hopefully someone can show me an easier way???

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Employee
INACTIVE_AstroJohn
Posts: 492
Registered: ‎03-25-2010

Re: Mass Transfer - Average Concentration in a Volume

08-07-2012 02:21 PM in reply to: tfjield

T,

 

The "Mean Weighted by Volume" needs the volume calculation from the "Analysis > Weight and Center of Gravity" calculation, but the Wt&CG calculator does not work with mass transfer. (Not sure why, other than it doesn't work.)

 

But the file that is needed is the file DS.NCG that Wt&CG creates, and that file is independent of the analysis type. So the work around is:

  1. copy the design scenario and set the analysis type to something else (linear stress, steady heat transfer, it really doesn't matter).
  2. Enter the material properties.
  3. "Analysis > Check Model".
  4. "Analysis > Weight and Center of Gravity".
  5. Using Windows Explorer (My Computer), copy the file DS.NCG from the new design scenario folder to the mass transfer-related design scenario. So if your simulation was names "saturn.fem", the mass transfer analysis was in design scenario 2, and the weight and center of gravity calculation was done in design scenario 4, you would copy "saturn.ds_data\4\ds.ncg" to "saturn.ds_data\2\ds.ncg".

Presto. The "Mean Weighted by Volume" command will now work in the mass transfer results.

 

Sincerely,
John Holtz, P.E.
Senior User Experience Designer, Simulation
Autodesk, Inc.

Current version of Mechanical & Multiphysics: 2013 SP1 (2013.01.00.0012 28-Jun-2012)
Please use plain text.
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012

Re: Mass Transfer - Average Concentration in a Volume

08-07-2012 03:40 PM in reply to: INACTIVE_AstroJohn

John,

 

YOU'RE THE MAN!!  Works great!

 

Do you happen to know, is it possible for me to pull this information out of the files programmatically in the same way that I pulled fluid velocity through faces out of the .sfrate or .ufrate files?  (I wrote a program in C# to do that.)  I'm now pulling concentration averages out load case by case manually, but when I have lots of load cases pulling it out of the files would be much faster.

 

Thanks again!

Todd

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Employee
INACTIVE_AstroJohn
Posts: 492
Registered: ‎03-25-2010

Re: Mass Transfer - Average Concentration in a Volume

08-08-2012 08:53 AM in reply to: tfjield

Hi Todd,

 

(I tried to attach an old (but latest) version of some documentation that describes the binary file formats, etc. The full documentation is 1.82 MB which is slightly too large to fit under the 1.5 MB limit of the Discussion Group. If you can provide your email address, either in a post to the general public here or by clicking AstroJohn link and sending a private message, I will provide the complete compiled help (.CHM) file. Otherwise, these PDF printouts will do.)

 

I think the general process would be as follows:

 

  • read the element database (Elements.dbf) to determine which nodes are on which element. (see "ALGOR Database Definition Guide > Technical Details > Element Files"). I do not remember if degenerate brick elements (elements with 4, 5, or 6 nodes) only list n node numbers or whether they duplicate some of the node numbers so that the brick element has "8 nodes". So be aware of that possibility.
  • read the concentration at those nodes from the DS.UMN file (see "Files Used by ALGOR > Other Output Files > Mass Transfer Nodal Results")
  • read the volume of the element from the DS.NCG file (see "Files Used by ALGOR > Other output Files > Weight & Center of Gravity Output").
  • Calculate the concentration in the element's volume. The Results environment may take the element's shape into account (a shape function), or it may just be the average concentration from all of the nodes and the volume. So there may  be some discrepancy between your calculation and that reported from the "Results Inquire > Inquire > Current Results" command.

The binary result files probably have not changed since the documentation was updated. The documentation for the databases (.dbf located in the .mod folder) may not be up-to-date. I usually use Excel to open them as a dBase file to see what those contain.

 

Sincerely,
John Holtz, P.E.
Senior User Experience Designer, Simulation
Autodesk, Inc.

Current version of Mechanical & Multiphysics: 2013 SP1 (2013.01.00.0012 28-Jun-2012)
Please use plain text.
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012

Re: Mass Transfer - Average Concentration in a Volume

08-08-2012 09:32 AM in reply to: INACTIVE_AstroJohn

Hi John,

 

Excellent, that sounds about like what I expected I would need to do.

 

Yeah, from what I've seen the Elements.dbf duplicates nodes for tetrahedrons, for example, so that there are still 8 nodes listed, but some are duplicates.

 

I was wondering the same thing about the average concentrations, but I'm glad that I don't need to calculate element volumes based on nodal coordinates!

 

I'll PM you my email address for the .chm file.

 

Thanks again!

Todd

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012

Re: Mass Transfer - Average Concentration in a Volume

08-09-2012 10:29 AM in reply to: INACTIVE_AstroJohn

Hey John,

 

I have a working program now to read out all of the load cases and calculate the average concentrations in each of my parts.  In my current model, I have 5 parts with 96 load cases, so you can imagine that it would take forever to do manually!

 

A couple things I've noticed.  First, just FYI, there is another column in the Elements.dbf table.  It is surf_id, located between prop_id and n01.  This isn't documented, but it was easy enough to find by connecting to the database via Excel.  Oh, and what fun it was figuring out how to connect to the database from my C# application!  There's 64 bit vs. 32 bit fun with FoxPro database drivers on a 64 bit Windows 7 install...  Since I have 32 bit Office installed, I can only use the 32 bit OLE DB drivers, which means I have to compile my application for x86.  Grr!  But at least it works now!

 

The other thing is a question/assumption that I've made.  In the Elements.dbf table, the parts (groups) are listed by part number, i.e. the number shown in the GUI.  I have several parts, and some of them are suppressed, so the part numbers listed in my table are 1,3,5,7, and 11.  But in the various output files, like the .umn and .ncg files, the parts (groups) are listed sequentially and not called out by part number.

 

So my question is:  In what order are the groups listed in the output files?  I assumed they would be listed in numerical order, and that works for this particular model.  In other words, part #1 was listed first in the output files, followed by part #3, followed by part #5, etc.  Can I rely on this, i.e. will they ALWAYS be sorted in numerical order?

 

Thank you,

Todd

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Employee
INACTIVE_AstroJohn
Posts: 492
Registered: ‎03-25-2010

Re: Mass Transfer - Average Concentration in a Volume

08-10-2012 06:43 AM in reply to: tfjield

Your assumption about the part ordering in the results is correct. The parts are listed in numerical order.

 

Sincerely,
John Holtz, P.E.
Senior User Experience Designer, Simulation
Autodesk, Inc.

Current version of Mechanical & Multiphysics: 2013 SP1 (2013.01.00.0012 28-Jun-2012)
Please use plain text.