Posted Sun, 30 Dec 2007 09:31:24 GMT by Rouven
Hi everybody,

I have another question concerning the IfmBugetComponent... API function.

I am using a modul to set an dirichlet boundary at inner nodes of my model. As a consequence I receive a boundary flux leaving my model at this nodes that I like to detect by using the IfmBudgetComponent... command. Unfortunatly the boundary flux I get from the API function differs completly ( ten times higher) from the value returned by the budget analyzer of the postprocessor file.

I think these values should be the same or it least in the same region?!!!
Could anybody help me with this problem or is there a better way to get the boundary flux  at a single inner node by using a API function?

Thank's Rouven
Posted Fri, 04 Jan 2008 18:03:52 GMT by Denim Umeshkumar Anajwala

Hey Rouven,

those fluxes should indeed be the same. If you model isn't too big, feel free to email it to me with the code and I'll have a look at it.

Cheers, Chris
Posted Sat, 05 Jan 2008 19:45:32 GMT by Rouven
Hi Chris,

the model is quit  big for mailing, but the Wasy support team already helped me out with my problem.
The modul simultaneously adjusts the material properties of elements next to the dirichlet boundaries. For high conductivity values, large boundary fluxes result. In this case the budget analyzer of the postprocessor file returns wrong values (I don't know why). So it's the budget analyzer one schould be carefull with! I compared the flux values of my API function with the results I received for an observation point group defined at the boundary nodes. These values were exactly the same!

Cheers and thank's for your help

Posted Sun, 06 Jan 2008 09:33:39 GMT by Boris Lyssenko
The budget analyzer calculates the boundary flows inserting the resulting heads into the matrix and solving for the boundary flows. As changes in material properties done by the using the programming interface during the simulation are not stored in the postprocessor file, this calculation is done based on the original conductivities in the file. In such a case the boundary flows should always be calculated during the simulation (by using IFM or observation point groups) as then the matrix contains the correct values.

Posted Sun, 06 Jan 2008 15:40:48 GMT by Denim Umeshkumar Anajwala
This is an excellent comment!!! I was wondering about this exact same thing recently, when I was using a PreSimulation IFM to update K data. If I understand the process correctly, this is how FeFlow generates a DAC file:

1) Copy data (incl. K's) from FEM to DAC file
2) Run simulation (next time step)
3) APPEND simulated heads (and velo's) to the DAC file
    (for those time steps specified)

That should explain why K-data manipulated through an IFM during the simulation is not stored in the DAC file. Only simulated heads get appended to the DAC. Correct?

[Edited: Ops, I just noticed that I had made a similar comment before under "IfmBudgetComponents - area flux"  ;D]

Now I'm wondering: Is it possible in FeFlow to specify time-varying K data? I've never tried that (and can't really think about a reason why I ever would), but a quick peek at the FeFlow help suggests, that this is possible by either using a T-List or power-functions. How would that be reflected in the DAC file and how would the Budget Analyzer consider this time-varying K-data in it's computations? ???

Posted Mon, 07 Jan 2008 18:50:25 GMT by Denim Umeshkumar Anajwala

I can confirm that one can specify time-varying K data using the t-list and power funtions. I have used such a feature to simulate the intallation of a slurry wall and to simulate the excavation of contaminated (native soil) with clean fill (of differing hydraulic properties). As for the rest of your questions, I do not know the answer.


You must be signed in to post in this forum.