How to plot DSFLUX in Abaqus viewer?

[UssM]

Hi everybody,

first of all congratulation for this very good forum. I want to ask a simple question: is possible to plot non-uniform DSFLUX distribution in Abaqus viewer in a way similar to how I can do with FLUXS output (it works only with uniform flux)? Let me explain my problem further.

I did an heat transfer analysis on a component. On one face heatflux is applied by means of "*DSFLUX, SNU" command and DFLUX user subroutine, because I need to map a NON-UNIFORM flux reading values from a csv table.

The simulation finishes and I need to check if subroutine worked well or if coordinates in csv table are correctly alligned with model. In other words, I want a feedback plotting the fluxes on the surface, like FLUXS output (but it works only with uniform flux) or FILMCOEF (for FILM subroutine). Is this possible, maybe writing a Python postprocess script?

A possible workaround is plotting node output RFLE (total flux at the node [J/s]), but I need the nodal area to return to applied flux [J/s/m2]. Is possible to retrieve nodal area from the output database, maybe activating some special output?

Thank you,
M

 

[Mustaine3]

Writing data in a text file and reusing that with a python script to write it onto an .odb is always possible.

But you should first check, if you can use a SDV to store the data defined in the subroutine and request element output into the .odb of the SDV.
Use a small and simple example to test that.

 

[UssM]

Thanks for reply.

Your first option is, in my opinion, a little bit redundant because require a double pass: first run the simulation with DFLUX subroutine, writing a text file containing association between node id and flux value; second execute a python script that read that text and update the .odb.

Probably your second option is better. I know a little bit about state dependent variable used with UMAT. But I can't figure out how to store a SDV value for EACH node within DFLUX subroutine. Is it possible? How?

Regards,
M

 

[Mustaine3]

I've never done that, but theoretical you could add a simple USDFLD to your DFLUX subroutine and transfer the flux value with a common block. In the USDFLD you could then save the values in a SDV. It would require to use a dummy field variable in the model, so the USDFLD is used for that.

But as mentioned, I don't know if that would actually work.