Heat transfer abaqus

[anoop_1989]

Dear all
I am trying to simulate welding heat transfer in a simple plate of dimensions 0.15*.2*.002 m with 304ss. Heat input was given by the DFlux subroutine with a voltage of 25v current 280A and velocity of .0025m/s. Heat flux is given as body heat flux. The heat transfer coefficient is given as 50W/m2. Radiation effects are not considered. As per the journal paper temperature of 667 degree Celcius is observed at a distance of 10mm from the weld centerline. But I am getting at a range more 1200 degree Celcius. I have checked all the units in properties (thermal conductivity, specific heat, density)(All are temperature dependent properties).Step time is given as 120 for heat with intial size of.1 and increment of .0001.Cooling time for 300 seconds. Along the weld center line, I am getting the temperature in 4000-degree Celcius...Can any suggest the correction I have to make so that I will get the temperature in 600-700 range.?
This was the subroutine used
SUBROUTINE DFLUX(FLUX,SOL,KSTEP,KINC,TIME,NOEL,NPT,COORDS,
1 JLTYP,TEMP,PRESS,SNAME)
INCLUDE 'ABA_PARAM.INC'
DIMENSION FLUX(2),TIME(2),COORDS(3)
CHARACTER*80 SNAME
REAL t,x,y,z,Q,cf,cr,b,a,Amp,volt,n
REAL z1,x2,y2,z2,v,ff,fr,d1,d2,l
REAL FLUX1

t = time(2)

c welding arc
cf =.007
cr =.003
b = .003
a = 0.004
ff=1.4
fr=.6

c speed of welding in Z direction
v=.0025
z1 = v*t
c coordinate
x = COORDS(1)
y = COORDS(2)
z = COORDS(3)

c
x2=x
y2=y
z2=z+z1
c Q=n*V*i
n=0.5
volt=25
Amp=280
Q=n*volt*Amp
if(JLTYP.EQ.1) then
if(KSTEP.EQ.1) then
c Goldak's volumetric heat source model
if(z2>=0) then
FLUX(1)=((1.8663*ff*Q)/(a*b*cf))*exp(-3*((x2)**2/a**2+(y2)**2/b**2+(z2)**2/cf**2))


else if (z2<0) then
FLUX(1)=((1.8663*fr*Q)/(a*b*cr))*exp(-3*((x2)**2/a**2+(y2)**2/b**2+(z2)**2/cr**2))


end if
end if
end if

return
end

 

[anoop_1989]

cae file

 

[FEA way]

When you attach cae file you should also include jnl, But could you attach inp file too ? And most importantly - which scientific paper you are talking about (title and authors) ?

Is element activation used in your analysis ? Are you sure that all settings are the same as in the article ? Maybe its authors have done something differently.

Also, you can try with Abaqus Welding Interface plugin.

 

[anoop_1989]

Sir,

This was the paper I referred to.I didn't include element deletion

Comparison between different heat sources types in thin-plate welding simulation
M. Hashemzadeh, B.-Q. Chen & C. Guedes Soares
Centre for Marine Technology and Engineering (CENTEC), Instituto Superior Técnico,
Technical University of Lisbon, Lisbon, Portugal

Attaching the jnl file

 

[anoop_1989]

Sir

Attaching the inp file

 

[Erik Panos Kostson]

In the paper there is no mention of the efficiency used (as far as I read), and the dimensions of the weld spot.
Also they use half symmetry thus Q needs to be divided by two (you have done that, but then you assume an efficiency (ita) of 1 which might not be what they had)
According to the experiments (Choobi 2010), they used an efficiency (ita) of 60% - thus you need to have 0.6*Q or Q=0.5*0.6*Current*Voltage. They also mention a current of 96 A and voltage of 10 V (seems low), so perhaps you want to contact some of the authors of these two papers and ask about these settings.

Also just to be sure I would specify DT/DX (symmetry plane)=0, in order to be sure on the thermal symmetry.
I think though if you do not specify anything there then the flux is zero so it should be OK

 

[anoop_1989]

Thank you @ Erik Panos Kotson,

Yes, the parameters for the ellipsoidal model are missing in this paper.

I will try with half heat input and I believe that may be reason

Also, I would like to do residual stress analysis after this along with phase transformations.If you have any examples for writing phase transformation subroutine please share the link.

 

[Erik Panos Kostson]

I have never done anything like that - I am sure that if you search on the internet that you will find a lot of info.

All the best

 

[FEA way]

You can find comprehensive information about simulations with phase transfotmation if you search for Print to Perform on Dassault Systemes website. There you will find several papers about additive manufacturing process simulations including those focusing on phase transformations. Residual stresses are also discussed in some articles (eigenstrain approach can be interesting for you too).

 

[anoop_1989]

Thank you