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 The information provided may be out of date.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.

problem in meshing

Please login with a confirmed email address before reporting spam

Hi Ivar,

I am trying to solve a model where i get this error:
Failed to respect boundary element edge on geometry face.
although i mesh the geometry in parts and using finer mesh in critical areas but still it doesn't work? Do you know what is causing this error in meshing?

Thanks,
Rabia

8 Replies Last Post 24 mag 2012, 06:21 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 mag 2012, 08:33 GMT-4
Hi

I do not know much nore to do than to locate the region and try meshing locally finer, also if your domain loops around (dognut shapes) it's often worth to cut it into two, and mesh the inner parts first

--
Good luck
Ivar
Hi I do not know much nore to do than to locate the region and try meshing locally finer, also if your domain loops around (dognut shapes) it's often worth to cut it into two, and mesh the inner parts first -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 mag 2012, 08:49 GMT-4
Second Ivar's contribution.

I find that separating the structure into parts helps a lot.

Take your time and think it through as a number of sub sections which you mesh part by part.
Second Ivar's contribution. I find that separating the structure into parts helps a lot. Take your time and think it through as a number of sub sections which you mesh part by part.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mag 2012, 09:30 GMT-4
Hi,

Thanks. I have simplified my geometry to go about this. Now i have completed the meshing however I am using
Joule Heating and thermal expansion physics which seems quite appealing however i can't figure out where to put which boundary conditions, i have :

temperature
electric potential(this i know)
ground(this i know)
heat flux
fixed constraint

i want to see what effect passing an electric current has on my part thermally.

Rabia
Hi, Thanks. I have simplified my geometry to go about this. Now i have completed the meshing however I am using Joule Heating and thermal expansion physics which seems quite appealing however i can't figure out where to put which boundary conditions, i have : temperature electric potential(this i know) ground(this i know) heat flux fixed constraint i want to see what effect passing an electric current has on my part thermally. Rabia

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mag 2012, 10:17 GMT-4
Hi

normally your object is fixed w.r.t. some outside parts at a "fixed" boundary, mostly this is at a given temperature too (that means COMSOL adapts the flux through this boundary to keep the temperature constant to the value you define)

Heat flux is if you have some other means of heating or cooling via any surface. If part is cooled by natural air convection you can use a Convection BC with i.e. h=5W/m^2/K and T0 at your room temperature

Have you tried the model library examples on Joule heating and thermal stress ?

--
Good luck
Ivar
Hi normally your object is fixed w.r.t. some outside parts at a "fixed" boundary, mostly this is at a given temperature too (that means COMSOL adapts the flux through this boundary to keep the temperature constant to the value you define) Heat flux is if you have some other means of heating or cooling via any surface. If part is cooled by natural air convection you can use a Convection BC with i.e. h=5W/m^2/K and T0 at your room temperature Have you tried the model library examples on Joule heating and thermal stress ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mag 2012, 12:13 GMT-4
Hi,

thanks. I have tried the example on micro resistor beam but for my problem there is one issue:

I am now trying to solve for temperature and for the boundary conditions i have:

heat flux and then heat source

heat source is electromagnetic power dissipation....

do i need anything else for just temperature solving in the thermal part ? It doesnt solve with these boundary conditions.

Also I don't want to see the thermal expansion of my model but do you think there will be any effect if i give in the fixed constraint boundary condition or not.?

Thanks,
Rabia
Hi, thanks. I have tried the example on micro resistor beam but for my problem there is one issue: I am now trying to solve for temperature and for the boundary conditions i have: heat flux and then heat source heat source is electromagnetic power dissipation.... do i need anything else for just temperature solving in the thermal part ? It doesnt solve with these boundary conditions. Also I don't want to see the thermal expansion of my model but do you think there will be any effect if i give in the fixed constraint boundary condition or not.? Thanks, Rabia

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mag 2012, 12:59 GMT-4
Hi

you should select the physics according to your need: you can select them individually or jointly as COMSOL has multiple physics combinations pre-cooked for us.

You would need something like Joule heating solving for V (Volts) and T (Temperature) but if you select JH and TS you get also thermal stress, but you will then get twice the temperature T, one too much, not a good idea.

First make up your mind what you want to see, your inputs your outputs, then select the physics according to the dependent variables solved.

You need your geoemtry defined, a material with conductivity and dielectric constant for the conduction part and GND, and voltage, or current for the BC, that will do for solving the Voltage correctly, hence the power dissipated. Then you need your materials (also the non electrric conducting ones, their density, heat conduction and heat capacity. The electrically dissipated power should enter as a heat source (domain) in the electrically heated part, and then probably your parts are held somewhere with a fixed reference temperature (maybe not, but if you do not dissipate any heat the temperature will only rise and newer cool down (inergy balance in-out.

If you want to see the expansion (TS) then you ned also the material data such as Young modulus, Poisson factor and thermal expansion cefficients (for linear maerials, or tensors for anisotropic materials)

Then be sure that the joule heating physics apply only to the conducting part and the linear maerial to the non conducting part, no need to solve for the joule heating if no current flows.

if you go through your model step by step you will find it easy, but when you start with everything in the same time its often confusing

--
Good luck
Ivar
Hi you should select the physics according to your need: you can select them individually or jointly as COMSOL has multiple physics combinations pre-cooked for us. You would need something like Joule heating solving for V (Volts) and T (Temperature) but if you select JH and TS you get also thermal stress, but you will then get twice the temperature T, one too much, not a good idea. First make up your mind what you want to see, your inputs your outputs, then select the physics according to the dependent variables solved. You need your geoemtry defined, a material with conductivity and dielectric constant for the conduction part and GND, and voltage, or current for the BC, that will do for solving the Voltage correctly, hence the power dissipated. Then you need your materials (also the non electrric conducting ones, their density, heat conduction and heat capacity. The electrically dissipated power should enter as a heat source (domain) in the electrically heated part, and then probably your parts are held somewhere with a fixed reference temperature (maybe not, but if you do not dissipate any heat the temperature will only rise and newer cool down (inergy balance in-out. If you want to see the expansion (TS) then you ned also the material data such as Young modulus, Poisson factor and thermal expansion cefficients (for linear maerials, or tensors for anisotropic materials) Then be sure that the joule heating physics apply only to the conducting part and the linear maerial to the non conducting part, no need to solve for the joule heating if no current flows. if you go through your model step by step you will find it easy, but when you start with everything in the same time its often confusing -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 24 mag 2012, 04:43 GMT-4
Hi,

Thanks. I am not interested in the Thermal Expansion part but only the Joule Heating, I want to look at how much my model heats up due to my voltage application, but there is this new physics which couples electric thermal and solid mechanics. If I dont give in the solid mechanics boundary condition comsol doesnt solve the problem. I am dealing with all conductors and not dielectric material. I want to use a heat source and a heat flux only to see the temperature rise due to application of voltage. Is there any way i can just use the joule heating part.

Rabia
Hi, Thanks. I am not interested in the Thermal Expansion part but only the Joule Heating, I want to look at how much my model heats up due to my voltage application, but there is this new physics which couples electric thermal and solid mechanics. If I dont give in the solid mechanics boundary condition comsol doesnt solve the problem. I am dealing with all conductors and not dielectric material. I want to use a heat source and a heat flux only to see the temperature rise due to application of voltage. Is there any way i can just use the joule heating part. Rabia

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 24 mag 2012, 06:21 GMT-4
Hi

if you choose a combined physics you must give enough BC to make all PDE solutions unique. if its only Joule heating you need only to select a physics that solves for V and for T not for "u"

TEM solves for u, V and T, (I'm ignoring J Radiosity for radiative exchanges)
but JH alone solves for V and for T and that should then be enough for your case then

--
Good luck
Ivar
Hi if you choose a combined physics you must give enough BC to make all PDE solutions unique. if its only Joule heating you need only to select a physics that solves for V and for T not for "u" TEM solves for u, V and T, (I'm ignoring J Radiosity for radiative exchanges) but JH alone solves for V and for T and that should then be enough for your case then -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.