Simulation Mechanical and Multiphysics

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

Load Cases

371 Views, 5 Replies
10-19-2012 02:36 PM

I think I might not be understanding the load cases correctly in my analyses.

 

I'm doing a steady state, coupled thermal and cfd analysis.  In the first setup, I have the following fluid load cases:

 

Initial Multiplier:  1.0

 

Step     Pseudo Time        Multiplier

1                  1                     0.1

2                  2                     0.2

3                  3                     0.4

4                  4                     0.6

5                  5                     0.8

6                  6                     1

7                  7                     2

8                  8                     4

 

In the Coupled Analysis Parameters, I have the steps set to 8.

 

Now, in the second setup, I have:

 

Initial Multiplier:  1.0

 

Step     Pseudo Time        Multiplier

1                  1                     4

2                  2                     8

3                  3                     16

4                  4                     32

 

In the Coupled Analysis Parameters, I have the steps set to 4.

 

Now here's what I don't understand...  Shouldn't the results from setup one, load case #8 be the same as setup two, load case #1?  They're not.  They're drastically different.

 

Am I missing something fundamental???

 

Thanks,

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: Load Cases

10-23-2012 09:30 AM in reply to: tfjield

OK, as another experiment I tried to run 4 load cases in two different orders in order to compare the results.

 

In the first case, I had:

 

"Ascending"

Initial Multiplier:  4

 

Step     Pseudo Time        Multiplier

1                  1                     4

2                  2                     8

3                  3                     16

4                  4                     32

 

In the second case, I had:

 

"Descending"

Initial Multiplier:  4

 

Step     Pseudo Time        Multiplier

1                  1                     32

2                  2                     16

3                  3                     8

4                  4                     4

 

I already confirmed that changing the Initial Multiplier does not affect the results.

 

Now in my model, I'm looking at the average temperature in one of the parts, looking for trends.  Here are the results I got:

 

Order.png

 

You can see the temperature magnitudes are different as is the trend.

 

Any idea what's going on here?  What am I missing??


Thanks!!

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Valued Mentor
AstroJohnPE
Posts: 503
Registered: ‎08-30-2012

Re: Load Cases

10-23-2012 01:36 PM in reply to: tfjield

Hi Todd,

 

I think I understand your description. If all of the multipliers are the same, the steady state results in two identical models should be equal regardless of which step the equal multipliers occur in.

 

I did a test model, and my results are close. For my coupled analysis parameters, I had gravity values of 0 so that the buoyancy effect would not affect the analysis. I setup 4 time steps in one second. Just like you, I changed the fluid flow multipliers. In one case they go from 0 to 10, and in the other case, they go from 10 to 0. The attached images show the fluid flow Parameters dialog and the average temperature results. In summary:

 

Fluid multiplier | Average temperature from 2 models

2.5                      | 72.2 versus 72.1

5.0                      | 80.4 versus 80.2

7.5                      | 84.8 versus 84.8

10.0                   | ??? versus 87.6

 

In my model, the fluid is heating the container. So as the velocity (multiplier) increases, the fluid is able to heat more of the container, so the temperature should increase. (And it does.)

 

I believe the temperatures do not match exactly because fluid flow is an iterative solution, and each model is approaching the theoretical / numerical solution from opposite directions. With more steps or finer tolerance, the two results should converge.

 

In your "descending / ascending" graph, the results are so close to 0 that it may be hard to draw a conclusion.

 

To elaborate on a few things (in case someone reads this at a later time), the initial multiplier CAN affect the analysis. Although the calculation is not performed at time 0, the multiplier after time 0 is interpolated from the initial multiplier and the  input on the first row of the spreadsheet. If the first time interval has only one time step, then the initial multiplier has no effect. If the first time interval has multiplie time steps (as in my example), the initial multiplier is used in the interpolation.

 

The Report tab gives three "Log" files.

  • The one with the header "Simulation Steady-State Heat Transfer" is from the heat transfer analysis.
  • The one without a header if from the fluid analysis. It is the same file that is shown when performing a pure fluid flow analysis and therefore gives additional information about the convergence of the fluid analysis which can help debug a model. (Specifically, the convergence rates of the three velocities and the pressure, etc. This log file is described on the page  "Performing Fluid Flow Analyses".) Unfortunately, this log file only gives the last time step. It looks like it gets overwritten on each step instead of appended.
  • The one labeled "Simulation Steady Coupled Fluid Flow and Thermal Analyses" in the header gives the convergence rate of the velocity and temperatures from one iteration to the next.

I'm running out of time for now, so I did not have time to look this up. For regular fluid low, the convergence type can be changed (somewhere) from a relative value to an absolute value. In low velocity situations -- cases where they buoyancy is important -- I find it helpful to change the tolerance type from the default of relative to absolute (or incremental).  I have seen cases where the calculated velocity should be small (let's say 1), the relative norm converges to its tolerance (let's say 1E-4), but the absolute/incremental velocity norm is large (let's say 0.5). A large absolute/incremental norm indicates that the velocity is varying by plus or minus 0.5 units. This is insignificant if the velocity is 1000 but very significant (a large error) if the velocity is supposed to be 1!

 

Fifteen years and I'm still learning about the software! I hope that doesn't discourage anyone. :-)

 

 

 

 

John Holtz, PE
Mechanical Engineer
Pittsburgh, PA

16 years experience with Simulation Mechanical
Please use plain text.
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012

Re: Load Cases

10-23-2012 03:16 PM in reply to: AstroJohnPE

Thanks, John.  As a followup, you did your timesteps differently than I did mine.  For the ascending, you went with an initial of 0 and in your load case you specified a multiplier of 10, and I'm assuming in the combined analysis parameters window you told it that you wanted 4 steps?

 

I set mine up as follows:

 

Fluid Window.png

 

and

 

Combined Window.png

 

I was under the impression that by setting the number of steps to 4 my load cases that were output would be the 4 load cases I had set up in the fluid parameters window, i.e. 32, 16, 8, and 4.  Is that correct, or am I getting some weird interpolation between them?

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.
Valued Mentor
AstroJohnPE
Posts: 503
Registered: ‎08-30-2012

Re: Load Cases

10-24-2012 09:51 AM in reply to: tfjield

Hi,

 

My understanding is that your example analysis should have simulated 1 second of time because the analysis duration is controlled by the Coupled analysis parameters. You can confirm or disprove this by checking the times shown in the Results environment.

 

But with your setup, you have asked for 4 time steps between time 0 and 1 sec (the first row in the Coupled analysis parameters spreadsheet). The coupled load (buoyancy) is constant because the initial multiplier and the multiplier at 1 second are 1.0. The Fluid Flow analysis parameters indicates to ramp the fluid loads (velocity, pressure, etc) from 4 to 32 during the first second. The rows after 1 second should have no effect on the analysis.

 

So your results should be

Time  |   Multiplier

0         |        4

0.25   |       11

0.50   |       18

0.75   |       25

1.00   |       32

 

Hope this helps.

John Holtz, PE
Mechanical Engineer
Pittsburgh, PA

16 years experience with Simulation Mechanical
Please use plain text.
Valued Contributor
tfjield
Posts: 98
Registered: ‎06-19-2012

Re: Load Cases

10-24-2012 10:33 AM in reply to: AstroJohnPE

Hi John,

 

Your explanation helps a lot!  Yeah, I was completely confused by how the load cases in the two different windows overlapped and I couldn't find documentation for the multiphysics load cases that explained it.

 

Turns out I was not understanding something fundamental about how the load cases were being processed and that's what resulted in my weird results.

 

I don't suppose that there is anywhere in the results environment where it actually states what my multipliers are?

 

Thanks!

Todd

Current version of Mechanical & Multiphysics: 2013 SP2 (2013.02.00.0010 2-Oct-2012)
Please use plain text.