Thermo-mechanical simulation related

[johnh]

1. I am working to simulate a laser excited wave propagation in a plate. (The input is attached at this email).
The laser heat is modeled by a short-time flux input, and I found that the solution is too sensitive with c (specific heat, in line 21 of the attached input).
With c=0, the flux decreases after removing the input flux, but with nonzero c values (although it is extremely small), the flux does not.

The maximum temperature is about 7*10^-3 degree with c=920 although the maximum temperature is 10000 degree (C) with c=0.
Those are very different even if I use c=1 (quite quasi-static).

2. The second issue is related with csurface.
I wrote cboundary command in line 175, and it disables the calculation.
The input works rather without the cboundary. It is oppisite to the case for displacement imposition using the csurface command.

Thank you for your help in advance.

Regards,
J

[FEAP_Admin]

I you set c=0, your problem no longer makes sense since your thermal mass is zero.  In this case you should redo your input file to solve a static (quasi-static) problem.

Regarding boundary conditions:

If you set the boundary codes to non-zero values then FEAP will ignore flux values and instead use given temperature values -- assuming zero if you do not provide them.

[johnh]

FEAP Admin,

Thank you for your answer.

With c=0, the problem is solved statically while it is solved transiently with c=nonzero.

The problems take places in the transient simulation (attached input file: ithm_transient):

1. flux does not decrease although it is decreased at 8 nano seconds in the input time profile (using "prop" command)

2. The maximum temperature is very low compared with the results obtained in the static simulation.

For your convenience, I made a simpler case, and attached both of the inputs at this email.

I would appreciate your help and comments.


Best wishes,
John

[FEAP_Admin]

I think your time scales are not what you are expecting; i.e. your diffusivity is not what you think it is.

Attached is an even more simplified version of your problem -- a bar with an transient heat flux imposed on one end.
Using your properties and 40 time steps essentially nothing happens -- the diffusivity is too low.  Then taking 40 more
time steps but with larger dt, one sees evolution of the state.

I also have output the nodal reactions at z=-h where the flux is imposed so that it is clear that the flux rises properly and then goes back to zero -- as it was a bit hard to see well in the plots (with the default color bars).

[johnh]

Dear FEAP Admin,

Thank you so much for your help!

I checked that the flux decreases.

I have one more question.

In the input Itbc, there is no described temperature boundary (e.g. Temperature is zero at the bottom).

Isn't it required for thermomechanical or thermal simulation?


Best wishes,
John

[FEAP_Admin]

The boundary conditions in FEAP default to zero Neuman boundary conditions if you do not specify anything and all values start initialized to zero.

So in this problem,  there is an imposed heat flux at z=-h and at z=h the displacements are fixed to zero.  The initial conditions as zero for all field variables and their derivatives.