Evaluating thermal flow through window frame

  • Last Post 21 February 2020
HCR_student posted this 16 February 2020

Hi everyone, I am working on a project to evaluate the thermal flow through a window frame. I started out using FLUENT but even after countless of video tutorials, I am struggling to understand how to set the conditions etc.

When I tried using Fluent, this was my setup:

- Inlet on the side that was exposed to the outside environment

- Outlet on the side within the "room"

- Temperature was set at both inlet and outlet.

I tried running the simulation where my frame was set at solid but there was the error and was only able to run when I changed it to fluid. However, I would like to run it as solid as I want to assign the material.

Will ANSYS Aim or Steady State Thermal Analysis be better for the analysis of my project?

Attached is an image of the window frame I would like to evaluate.

Please help, thank you!!

Order By: Standard | Newest | Votes
peteroznewman posted this 16 February 2020

The simplest model would be a Steady State Thermal model in ANSYS Mechanical, where simple BC assumptions would be convection on the inside face to the inside temperature and convection on the outside face to the outside temperature and just allow conduction to be the only mode of heat transfer through the frame, since by default, all the other faces will be insulated. This would give you a first approximation answer to the heat flow.  You will have to make an estimate of the convective film heat transfer coefficient to get that though.

HCR_student posted this 17 February 2020

Thank you so much for your reply! May I know how do I derive the film coefficient for natural convection? 

Is it possible to assume this? --> free convection to air at 30 degrees C (5 W/m^2-K), or forced convection at low speed of airflow over the surface (10 W/m^2-K)

peteroznewman posted this 17 February 2020

You look film coefficients up in reference books, you don't derive them and yes, they depend on free or forced convection and vertical or horizontal orientation.

HCR_student posted this 18 February 2020

Alright thank you! 

If I want to find out the thermal efficiency of the sash design, how should I go about doing this? 

For example, under steady-state thermal do I set Convection on the outside face with the ambient temperature as the outside temperature THEN Convection on the inside face and select tabular data for the ambient temperature? 

Is there also any way where I could see the distribution of heat through the different cavities?  

HCR_student posted this 18 February 2020

Aside from the above query, 

I followed your assumptions for the BC but how do I "just allow conduction to be the only mode of heat transfer through the frame"? Or is that the case by default? 

For my environmental settings, the outside temperature is 30 degree celsius while the inside room temperature is 25 degree celsius. Thus, I assumed these settings for the convection on the outside and inside face respectively : free convection to air at 30 degrees C (5 W/m^2-K), and forced convection at low speed of airflow over the surface (10 W/m^2-K) at 25 degree Celsius. 

Attached is the result found from the above setting 



peteroznewman posted this 18 February 2020

The two Convection BCs under Steady State Thermal are correct.

Delete the Temperature BC.  What is that on?  You don't have anything that has a fixed temperature.

Yes, the material has a defined thermal conductivity, and by default, that is being used to conduct heat from the hot side to the cool side.

The walls of the metal that don't have convection are assigned perfect insulation, so the cavity is completely separated from the metal frame. No heat is transferred  into the cavity.  You could fill the cavities with another material called air, which has a very low thermal conductivity, but in Mechanical, there would be no air flow.

What is the definition of thermal efficiency?

I think you want to calculate the U-value or the R-value of the window according to this site.

HCR_student posted this 21 February 2020

I have deleted temperature BC. However, what is the function of the initial temperature and how will it influence my analysis?

The material of the window frame is UPVC. The aim of my project is to increase the thermal efficiency of the window frame by decreasing the thermal conductivity, where the chamber design in the frame is able to reduce the amount of heat from the outside face from escaping into the room (inside face). I am hoping to realise this goal by showing that the cooler temperatures are covering the majority of the frame structure and this will mean an increase in thermal efficiency of the window frame. However I am still quite unsure if this will be sufficient or correct to conclude my project aim or this other idea. The other idea is where I only set a temperature on the outside face and find the eventual temperature on the inside face, and I am not sure if this is doable so i will need your kind advice.

Also as for the film coefficient, I am still confused by it even after doing some research. Is it the same as heat transfer coefficient? If it is then can I derive the film coefficient through the math equation for heat transfer coefficient? Will it be affected by the temperature I want to set? If I were to change the outside face temperature to 36.4 degree Celsius (free convection), will my setting of 5 W/m^2-K(film coefficient) still be the same?

Too make sure, for an air conditioned room, do you know if I will be able to assume low speed of airflow over the surface (10 W/m^2-K) at 25 degree Celsius? http://thermopedia.com/content/841/

I noticed a feature which is the fluid solid interface. May I know how will the inclusion of this setting influence the results or what is the function of it?

May I also ask if you are familiar with the design in the window frame and if you know why are there slants and curvatures in certain designs? (or to whoever else reading this thread!!!)

HCR_student posted this 21 February 2020

I also have this problem while trying to rerun the simulation. All I have changed from previous simulation was to delete the galvanised steel reinforcement.

Error message: Your product license has numerical problem size limits, you have exceeded these problem size limits and the solver cannot proceed.


peteroznewman posted this 21 February 2020

You should learn the basic physics of convection before building any models. You can start on Wikipedia.

The free Student license has a node or element count limit of 32,000. So if you use too many elements in your mesh, the solver refuses to run.

You won't have any problem if you run this as a 2D problem. Does the solid body pass through the XY plane? If so, create a plane and slice the solid body to get faces on the XY plane. In SpaceClaim, copy and paste those faces to get surfaces on the XY plane. Delete the solid body.  Save the SpaceClaim file with the surfaces in it.

In Workbench, drag out a new Steady State Thermal analysis and on the Geometry cell, in the Properties window on the right, set the Analysis Type to 2D, then import the file saved above.  Now you will be able to mesh with good detail without exceeding the node count limit.