Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Heat Transfer in Solids Transient behaviour
Posted 10.09.2014, 16:39 GMT-4 Heat Transfer & Phase Change, Structural Mechanics Version 4.3b 3 Replies
Please login with a confirmed email address before reporting spam
Hello Everyone
I am trying to model laser generated acoustic waves in an aluminum disk. Ultimately I plan on using the Thermal Stress physics, but I am starting with just the Heat Transfer portion to make sure the temperature makes sense before running the Thermal Stress Module. The model is 2D axis-symmetric.
To simulate the heating caused by the laser I am applying a body heat source Q = Q(t)*exp(r^2 / s^2)*exp(2z/d). where "s" is the beam radius, "d" is the laser penetration depth and Q(t) is a triangular pulse of 10ns in duration. See the attached PDF for details on the model & boundary conditions.
The resulting temperature profile (see attached pdf) is wrong. The maximum temperature increase is ok (about 180 K at the surface) and the initial time dependence is ok, but then, very large and sharp oscillations appear. These should not be there as there is no more heat being added to the disk and the temperature should keep exponentially falling. Also at some time steps the temperature falls below "ambient" temperature (293.15 K). I have cheked that the source is indeed "dead" at the time the oscillations ocur that the initial temperature and convective fluid temperature are correct, tried smaller relative and absolute tolerances, generalized alpha and BDF time stepping, as well as MUMPS and PARDISO solvers, and all options lead to similar results. I see no physical reason for the sharp temperature increases seen in the response, and I cant seem to figure out what I am doing wrong as this should be a very straightforward model.
Any help is greatly appreciated
I am trying to model laser generated acoustic waves in an aluminum disk. Ultimately I plan on using the Thermal Stress physics, but I am starting with just the Heat Transfer portion to make sure the temperature makes sense before running the Thermal Stress Module. The model is 2D axis-symmetric.
To simulate the heating caused by the laser I am applying a body heat source Q = Q(t)*exp(r^2 / s^2)*exp(2z/d). where "s" is the beam radius, "d" is the laser penetration depth and Q(t) is a triangular pulse of 10ns in duration. See the attached PDF for details on the model & boundary conditions.
The resulting temperature profile (see attached pdf) is wrong. The maximum temperature increase is ok (about 180 K at the surface) and the initial time dependence is ok, but then, very large and sharp oscillations appear. These should not be there as there is no more heat being added to the disk and the temperature should keep exponentially falling. Also at some time steps the temperature falls below "ambient" temperature (293.15 K). I have cheked that the source is indeed "dead" at the time the oscillations ocur that the initial temperature and convective fluid temperature are correct, tried smaller relative and absolute tolerances, generalized alpha and BDF time stepping, as well as MUMPS and PARDISO solvers, and all options lead to similar results. I see no physical reason for the sharp temperature increases seen in the response, and I cant seem to figure out what I am doing wrong as this should be a very straightforward model.
Any help is greatly appreciated
Attachments:
3 Replies Last Post 19.01.2015, 18:36 GMT-5
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Dear Jorge,
This type of non-physical temperature oscillation is common when the time step size is too small. The critical time step size is a function of the element size and the thermal diffusivity. You can avoid it by (i) increasing the time step size if possible, (ii) applying temperature loads more gradually, or (iii) reducing the element size according to this equation: L=sqrt(k*dt/(rho*Cp)) where dt is the time step size. You will find more details on this topic in an older thread: www.comsol.com/community/forums/general/thread/16808/.
Nagi Elabbasi
Veryst Engineering
This type of non-physical temperature oscillation is common when the time step size is too small. The critical time step size is a function of the element size and the thermal diffusivity. You can avoid it by (i) increasing the time step size if possible, (ii) applying temperature loads more gradually, or (iii) reducing the element size according to this equation: L=sqrt(k*dt/(rho*Cp)) where dt is the time step size. You will find more details on this topic in an older thread: www.comsol.com/community/forums/general/thread/16808/.
Nagi Elabbasi
Veryst Engineering
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Dear Nagi
Thank you very much for your response. It seems like I am caught between a rock and a hard place because I am currently using a 3e-7 element size and with the equation you provided I should be using 1.75e-7. The problem is that to ensure that the wave propagation phenomenon is properly sampled I need to ensure that the fastest wave does not cross an element in less that a time step, which means I am probably going to need to reduce the step size, which in turn means a smaller time step...
I am going to try and balance these requirements and see if that results in a good solution.
Thanks again for you help!
Jorge Quintero
Thank you very much for your response. It seems like I am caught between a rock and a hard place because I am currently using a 3e-7 element size and with the equation you provided I should be using 1.75e-7. The problem is that to ensure that the wave propagation phenomenon is properly sampled I need to ensure that the fastest wave does not cross an element in less that a time step, which means I am probably going to need to reduce the step size, which in turn means a smaller time step...
I am going to try and balance these requirements and see if that results in a good solution.
Thanks again for you help!
Jorge Quintero
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Dear Jorge,
This type of non-physical temperature oscillation is common when the time step size is too small. The critical time step size is a function of the element size and the thermal diffusivity. You can avoid it by (i) increasing the time step size if possible, (ii) applying temperature loads more gradually, or (iii) reducing the element size according to this equation: L=sqrt(k*dt/(rho*Cp)) where dt is the time step size. You will find more details on this topic in an older thread: www.comsol.com/community/forums/general/thread/16808/.
Nagi Elabbasi
Veryst Engineering
Hi Nagi,
I read the older thread. You mentioned this inequality which determines the element size is derived from the non-dimensionalizing of the heat transfer equation. Could you please explain this a little bit?
Thanks in advance.