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.

Modeling a Lamp Heat source to heat up a specimen

Please login with a confirmed email address before reporting spam

Hi!

I am trying to model two bodies in comsol, one is the heat source and the other body represents a specimen. The heat source is going to heat up the specimen. The model is explained in the attached image file. I am using Heat Transfer in solid module.

I have two Questions:

1) As I am transferring heat from one body to another, so I have to take into account surface to surface radiation but I tried different options and am unable to heat up the top surface of the specimen. And how I have to define in Comsol that in the surrounding we have air.

2) Once I am able to heat up the specimen top surface using the heat source, my next step is to move the heat source along the specimen in the x-direction. I want to know that which physics I have to use for that and how can I do it?

Please I am new to comsol and will really appreciate any input which can help me in solving my problem. If my problem is not properly explained, let me know I will try to elaborate more.

Thanks


4 Replies Last Post Oct 2, 2017, 1:54 p.m. EDT
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: 7 years ago Aug 16, 2017, 2:10 a.m. EDT
Hi Numi

Well even if your model seems simple it can, by all the physics involved rapidly grow quite complex. So first of all, my advise would be to follow the no 1 rule: "start simple, one physics at the time"

I always start by scanning the numerous Application Library models related to the subject, among other I see here: www.comsol.eu/model/free-convection-in-a-light-bulb-303
witch covers many of your subjects.

Another "trick" I always do when I have such a problem is to sketch down on a piece of paper the main COMSOL model tree constituents:

- Geometry: (you have by your drawing there, but I see only 2 components the source and the receiving slab, what about some air container around ? how big ...) But in my opinion for your case you should start in 2D with a simple cut view and then later, once you master and have checked (the VV&C process see (this one is free): www.nafems.org/publications/browse_buy/browse_by_topic/qa/verification_and_validation/ ) all physics for your model only go to a full 3D

- Materials: and the materials properties required, for each specific part

- Physics: involved, you state HT with Temperature as dependent variable but also radiation exchange ... but then you talk about convection => as an analytical boundary equation or as a full CFD mode (becomes heavy) and what about air conduction? (it might be ignored on larger models, but can become important for small (MEMS) devices ...)

- BCs: which, where, do we have enough to constrain correctly our physical equations

Once you have this you select the simplest case and start to model (do not forget to check your mesh as the diffusion equation behind HT is rather sensitive to regions with high temperature gradients, here you need a fine mesh in these critical regions, or you need to adapt your initial conditions to reduce the gradient steepness)

In your case with radiation you need to master this subject rather well as it's tricky (i.e. "View Factors" this is a good semester course by itself to cover all subtleties) you need to add the boundary emissivities of your radiative exchanging surfaces (these are mostly NOT already defined in the DB of bulk material properties of COMSOL).

For your HT case you must ensure you turn on with the tick mark the Radiation properties in the main HT node in COMSOL, as well as turn on the grouping feature further down in the same node, these are off by default.

I have added a simple stationary radiative exchange alone model hereafter but I have only V5.3 available (sorry for that, and COMSOL is not backwards compatible). Convection losses are set up with an analytical equation with an exchange ration h0[W/(m^2*K)] defined as a parameter applied to external surfaces, radiation group and diffuse emissivity values are set on related boundaries. I have invented sizes in the meter domain so they might not fit yours, but that is easy to adapt.

Be sure you get your grey body emissivity values correct as thermal radiation exchange depend strongly on these and the 4th power of the ABSOLUTE surface Temperatures IN KELVIN.
Emissivity is the same as absorption for grey bodies in the thermal regime. Mostly one talk about absorption in the visible (i.e. 0.5 um wavelength region in direct sun light) and emissivity in the thermal 10um wavelength regimes. But for grey bodies these are exactly the same physical values, we tend only to give different names to them depending on which radiation/light wavelength we are working in, this is often highly confusing. Check the HT documentation its all explained :)

Once you want to move the source you need to find the best way as that will continuously change the view factors telling how much of the energy is transferred by each surface to any other surface (see the HT user manual under radiation).
My first approach would be to move the source stepwise and remesh/restart the time series analysis for each new position, by using the previous temperature values as new initial T values. i.e if you have a source velocity of v0 and a source diameter of D0, then a time step for source stepping and remesh would be typically Dt=~D0/(2*v0) as radiative exchanges are "soft" exchanges.

Hope this helps on the way :)
--
Good luck
Ivar
Hi Numi Well even if your model seems simple it can, by all the physics involved rapidly grow quite complex. So first of all, my advise would be to follow the no 1 rule: "start simple, one physics at the time" I always start by scanning the numerous Application Library models related to the subject, among other I see here: https://www.comsol.eu/model/free-convection-in-a-light-bulb-303 witch covers many of your subjects. Another "trick" I always do when I have such a problem is to sketch down on a piece of paper the main COMSOL model tree constituents: - Geometry: (you have by your drawing there, but I see only 2 components the source and the receiving slab, what about some air container around ? how big ...) But in my opinion for your case you should start in 2D with a simple cut view and then later, once you master and have checked (the VV&C process see (this one is free): https://www.nafems.org/publications/browse_buy/browse_by_topic/qa/verification_and_validation/ ) all physics for your model only go to a full 3D - Materials: and the materials properties required, for each specific part - Physics: involved, you state HT with Temperature as dependent variable but also radiation exchange ... but then you talk about convection => as an analytical boundary equation or as a full CFD mode (becomes heavy) and what about air conduction? (it might be ignored on larger models, but can become important for small (MEMS) devices ...) - BCs: which, where, do we have enough to constrain correctly our physical equations Once you have this you select the simplest case and start to model (do not forget to check your mesh as the diffusion equation behind HT is rather sensitive to regions with high temperature gradients, here you need a fine mesh in these critical regions, or you need to adapt your initial conditions to reduce the gradient steepness) In your case with radiation you need to master this subject rather well as it's tricky (i.e. "View Factors" this is a good semester course by itself to cover all subtleties) you need to add the boundary emissivities of your radiative exchanging surfaces (these are mostly NOT already defined in the DB of bulk material properties of COMSOL). For your HT case you must ensure you turn on with the tick mark the Radiation properties in the main HT node in COMSOL, as well as turn on the grouping feature further down in the same node, these are off by default. I have added a simple stationary radiative exchange alone model hereafter but I have only V5.3 available (sorry for that, and COMSOL is not backwards compatible). Convection losses are set up with an analytical equation with an exchange ration h0[W/(m^2*K)] defined as a parameter applied to external surfaces, radiation group and diffuse emissivity values are set on related boundaries. I have invented sizes in the meter domain so they might not fit yours, but that is easy to adapt. Be sure you get your grey body emissivity values correct as thermal radiation exchange depend strongly on these and the 4th power of the ABSOLUTE surface Temperatures IN KELVIN. Emissivity is the same as absorption for grey bodies in the thermal regime. Mostly one talk about absorption in the visible (i.e. 0.5 um wavelength region in direct sun light) and emissivity in the thermal 10um wavelength regimes. But for grey bodies these are exactly the same physical values, we tend only to give different names to them depending on which radiation/light wavelength we are working in, this is often highly confusing. Check the HT documentation its all explained :) Once you want to move the source you need to find the best way as that will continuously change the view factors telling how much of the energy is transferred by each surface to any other surface (see the HT user manual under radiation). My first approach would be to move the source stepwise and remesh/restart the time series analysis for each new position, by using the previous temperature values as new initial T values. i.e if you have a source velocity of v0 and a source diameter of D0, then a time step for source stepping and remesh would be typically Dt=~D0/(2*v0) as radiative exchanges are "soft" exchanges. Hope this helps on the way :) -- Good luck Ivar


Please login with a confirmed email address before reporting spam

Posted: 6 years ago Sep 28, 2017, 2:44 a.m. EDT

Hi Ivar!

Thank you very much for your detailed reply. I am sorry for getting back to you but I was trying to learn whatever you explained and use it in my model, which worked out very well.

I am using moving mesh to move the rod(heater) over the specimen which seems is working fine. But the only issue I am facing now is the effect due to speed of heating rod. The heating rod diameter is 2mm and the time step for the simulation is Dt = 0.05. I am doing a parametric study with velocity equal to 20mm/s, 40mm/s and 60mm/s. For 20mm/s speed The heat flow on the specimen with the rod consistent. But if I increase speed to 40mm/s or 60mm/s it lags as can be seen in attachment.

I was going through your reply again and I think I should change the time step according to this (i.e. Dt=~D0/(2v0) ). Am I right? and please can you explain the concept behind selecting time step according to this equation?

Also, my simulations are taking very long time i.e. nearly 7hrs. I tried to change mesh and solver from BDF to Generalized alpha but it didn't help. Please guide me in this regard too if you can suggest that how can I speed up my simulations.

I look forward to hearing from you.

Hi Ivar! Thank you very much for your detailed reply. I am sorry for getting back to you but I was trying to learn whatever you explained and use it in my model, which worked out very well. I am using moving mesh to move the rod(heater) over the specimen which seems is working fine. But the only issue I am facing now is the effect due to speed of heating rod. The heating rod diameter is 2mm and the time step for the simulation is Dt = 0.05. I am doing a parametric study with velocity equal to 20mm/s, 40mm/s and 60mm/s. For 20mm/s speed The heat flow on the specimen with the rod consistent. But if I increase speed to 40mm/s or 60mm/s it lags as can be seen in attachment. I was going through your reply again and I think I should change the time step according to this (i.e. Dt=~D0/(2v0) ). Am I right? and please can you explain the concept behind selecting time step according to this equation? Also, my simulations are taking very long time i.e. nearly 7hrs. I tried to change mesh and solver from BDF to Generalized alpha but it didn't help. Please guide me in this regard too if you can suggest that how can I speed up my simulations. I look forward to hearing from you.


Please login with a confirmed email address before reporting spam

Posted: 6 years ago Sep 30, 2017, 2:57 a.m. EDT

I am using moving mesh (ALE) physics for moving the heat source, is it correct or I should be using Multibody dynamics (mbd). Because using both physics I can do translational motion of the heat source but don't know which one is better in my case?

I am using moving mesh (ALE) physics for moving the heat source, is it correct or I should be using Multibody dynamics (mbd). Because using both physics I can do translational motion of the heat source but don't know which one is better in my case?

Please login with a confirmed email address before reporting spam

Posted: 6 years ago Oct 2, 2017, 1:54 p.m. EDT

I can provide tranlation motion to solid ( Right slid solid heat transfer domain node and select transaltion) . Alos fix the temperature boundary on entering end and outflow boudary on the exit end.

I can provide tranlation motion to solid ( Right slid solid heat transfer domain node and select transaltion) . Alos fix the temperature boundary on entering end and outflow boudary on the exit end.

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.