best method to design welding in ansys workbench

[shinobi9]

Bye to all,

Since I never happened but now I would need it, I wanted to ask you what is the best way to design a welding with ansys workbench.

We admit that I have 2 welded compoents, it would be lawful to create semparately (in slidworks then using 3d elements in ansys) the 2 bodies and welding and then assemblers in ansys and consider contacts between the parts?

Thank you.

 

[Stan9411]

if you model the soldering you have to assign properties, but in reality a welding cord, despite you know the material, has lost homogeneity, isotropy, part of its ductility and is statistically subject to defects that affect the mechanical properties compared to the ideal condition of a fem model. Moreover the interface with the basic material will never be as perfect as you would represent it in a fem because the materials merge partially among themselves.
in the academic field and I also think in the industrial one, at least for simple carpenters the norms are used with the appropriate methods and coefficients. then so good sense, experience and confidence in the welder. numerical/experimental analysis of welds are still subject to academic studies (my friends are doing the magisterial thesis), so maybe if you are looking for among the various papers published by researchers around the world, some approach to the problem that gives good results according to me you find it.
more so I can't tell you why I went into the world of work I abandoned this aspect of design.

 

[Stan9411]

test on google scholar with stuff like “weldings finite element method” or “weldings numerical analysis”.
Usually the articles you will find describe a generic approach, then it will be you to apply it in ansys. but press the right button in the workbench I think is the least of this kind of problem.

 

[shinobi9]

Thanks for the answers.

It is certainly a complex problem, we say that I was interested in knowing what "practice" was for welding in fem modeling.
I wonder why I have a strange condition to evaluate, it was a welded beam to an extreme doi directly by hand but I have to do with a welded plate to two supports at its extremes with angle cords and perpendicular load to the plate itself as from attached image

 

[shinobi9]

looking around I decided to use the method outlined to the discussion:
https://www.cad3d.it/forum1/threads/saldature-in-ansys-wb.15133/

solo che nella sezione "comment" ho scritto (avendo 6 saldature):
"
! commands inserted into this file will be executed immediately after the ansys /post1 command.
! if a set command is issued, results from that load step will be used as the basis of all
! result objects appearing in the solution folder.

! active unit system in workbench when this object was created: metric (mm, kg, n, c, s)

csys,0

! define the names of the components, the name must match the component names defined
*dim, cm_weld, char, 50,1
cm_weld(1)='weld_a' ! local,12
cm_weld(2)='weld_b' ! local,13
cm_weld(3)='weld_c' ! local,14
cm_weld(4)='weld_d' ! local,15
cm_weld(5)='weld_e' ! local,16
cm_weld(6)='weld_f' ! local,17

/post1
! print out nodal loads for nodal components
*do,i,1,6,1
cmsel,s,%cm_weld(i)%
csnum=11+i
*get,csx,cdsy,csnum,loc,x
*get,csy,cdsy,csnum,loc,y
*get,csz,cdsy,csnum,loc,z
rsys,csnum
spoint,,csx,csy,csz ! change the moment summation point to the local cs origin
/output,%cm_weld(i)%_fsum,lis,, , ! print the nodal forces and moments to a text file
fsum,rsys,,
/output,term
*enddo "

ma i file dei risultati mi riportano la scritta:
"
*** warning *** cp = 20.641 time= 15:31:42
no element nodal forces available. the fsum command is ignored.
"

e cercando nel file "err" ho il seguente errore:
"
*** error *** cp = 1.469 time= 15:50:25
component weld_a is not defined.
/com,ansys release release 16.0 build 16.0 up20141203 15:53:51
"
come posso risolvere?

ps: come si salva la simulazione in un database ( file .wbdb)?

 

[Stan9411]

boh will be that it is a few years that I do not use it ... but the workbench is not the one with the user interface in cad style? This kind of script you posted I don't know and I can't read it.. Maybe it's classic ansys, or a methodology I've never used

 

[shinobi9]

for some workbench things it interfaces with apdl through the "command" tool that allows to write scripts in apdl.

 

[Stan9411]

you never stop learning