Heat source doesn't seems to be propagating heat 1

[bernardirafael]

I managed to solve the problem of my last thread here, but now I am having another one. The Workbench is plotting the heat source and it is clearly heating the plate, but it seems that the results of the last substep are not being considered on the next one. As the heat should be propagated through a great part of nodes/elements, it is only next to the heat source (see image below for better understanding).

It should be working like this:

As you can see, the heated zone should remain with a higher temperature over time, until it cools down entirely. Unfortunately, it is not happening on my analysis, since the heat source moves along with your effect area but it seems that it's not accounted for the next substep.

Here is my code:

CMSEL,ALL
*GET,EMAX,ELEM,,NUM,MAX
*GET,EMIN,ELEM,,NUM,MIN
ALLSEL

AUTOTS,ON

TIME_WELD=200
DT=1
A=0.003
B=0.004
C1=0.004
C2=0.016 !A, B AND C ARE JUST THE HEAT SOURCE DIMENSIONS
TAU=0 !TIME DELAY
FF=0.4 !FRACTION OF HEAT ON THE FRONT PART OF THE HEAT SOURCE
FR=1.6 !FRACTION OF HEAT ON THE REAR PART OF THE HEAT SOURCE
Q=3840 !WELDING ENERGY J/S (W)
VEL=0.002 !WELDING SPEED
EE=2.71 !I HAD TO DEFINE THE EULER NUMBER AS 2.71 INSTEAD OF USING exp() DUE TO UNSTABILITIES ON MY RESULTS

NPT=TIME_WELD/DT

*DO,ii,1,NPT,1 !TIME INCREMENT
WTIME=(ii/2)
TIME,WTIME

HCENTER=VEL*WTIME !HEAT SOURCE CENTER

*DO,jj,EMIN,EMAX,1 !SELECTS THE CENTROID OF EVERY ELEMENT

X=CENTRX(jj)
Y=CENTRY(jj)
Z=CENTRZ(jj)

CSI=Z+(VEL*(TAU-WTIME)) !MOVING COORDINATE (IT MOVES THROUGH Z AS TIME PASSES

*IF,Z,GT,HCENTER,THEN !DEFINES WHETHER IT'S THE FRONT PART OF THE HEAT SOURCE OR THE REAR PART

C=C1
F=FF

*ELSE

C=C2
F=FR

*ENDIF


PART1=(6*(3**0.5)*Q*F)/(A*B*C*3.14*(3.14**0.5))
PART2=(EE**(-3*(X/A)**2))*(EE**(-3*(Y/B)**2))*(EE**(-3*((CSI/C)**2)))

QF=PART1*PART2 !HEAT SOURCE EQUATION

BFE,jj,HGEN,,QF

*ENDDO

SOLVE

*ENDDO

CMSEL,ALL
BFEDELE,ALL,ALL
ALLSEL

Any ideas on how to solve this problem? What am I doing wrong?

 

[hendersdc]

I did not go through your code but it looks to me like the last time step is being taken into account because the temperature gradient is skewed to one side from the heat source. If the simulation were looking at a blank slate for each time step the source would be centered.

One obvious difference between the two analyses you show is the thickness of the material. If your heat load is kept the same but the thickness increased I would expect to see temperatures lower as there is more cross sectional area for the thermal conduction to occur and more material volume to soak up the energy. It will take more energy to raise the large block temperature uniformly by 1 degree than it will for the thin sheet.

A quick way to verify if the thickness is the source of your different results is to leave everything else the same and re-run your simulation with a much lower thickness.

 

[bernardirafael]

@hendersdc

Thanks for your reply!

Actually the heat source is moving so the "heat shape" shown was expected.

I managed to fix the problem, now I am looking for a similar work to validate my model... thanks for your help!