Simulation Forums

Find Simulation answers, share expertise, and connect with your peers

Simulation CFD


Model study Check: A room with radiant cooling and heating surfaces

93 Views, 13 Replies
02/13/2018 11:45 AM

Hello all,

 

I have done experiments with a test-room which has 5 thermally activated surfaces (3 surrounding walls, floor, and ceiling) and a facade. In our experiment, I cooled down the walls and floor at the relatively constant temperature of 20°C and heated the ceiling at the constant temperature of 40°C. I measured the cooling power cold surfaces (walls and floor) and the warm surface (ceiling), and also I measured the inner surface of the facade (e.g. 17°C) which has a known U-value of 1.3 W/m2K. 

 

In my CFD simulation, I created the same geometry. In the section of boundary conditions, I set the same experimental temperatures at walls, floor, and ceiling, and I set the same negative heat flux at cold surfaces (floor and walls) and positive heat flux at the ceiling. I chose the film coefficient (W/m2K) with the reference temperature of cold ambient (5°C) at the outer surface of the facade.

 

In the Physics, I chose the radiation calculations. Since I do not have forced convection in the room, the forced convection is unchecked but the Buoyancy is there. Turbulence model as K-E was chosen. 

 

I would appreciate having your comments if I have done something wrong or can I do something for its improvement such as turbulence model.

 

Best regards,

Reza


Report Inappropriate Content
Message 1 of 13 ( Views: 93 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/13/2018 1:47 PM

Hi Reza,

 

Thanks a lot for posting. You are going in the right direction and its great that you have experimental data to compare with.

 

It sounds to me as though you have overconstrained your model with too many BCs. Use either temperature or heat flux conditions on the external boundaries. Having both will confuse the solver since they affect one another.

 

I think you are using the film coefficient correctly on the facade and have suitable solver settings.

 

 

Hope that helps a bit. Feel free to upload a .cfz file so that I can check your setup.

 

All the best,

David

 

Note: Attach a [.cfz] file rather than the [.cfdst]. The [.cfz] file is similar to a zip folder and contains all the relevant files for your simulation setup. The [.cfdst] file is a stand alone file that cannot be used without the rest of the design study folder contents.


David Short
Technical Support Specialist, Simulation

Report Inappropriate Content
Message 2 of 13 ( Views: 83 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/13/2018 2:53 PM

Hello David,

 

thank you very much for your prompt reply.

 

Indeed, I thought also that I am overconstraining the BCs by giving two values (Temperature and heat flux). I tried only temperatures at all surfaces (including the window) without heat flux, however, the results seemed incorrect with much temperature stratification compared to the reality and no natural convection. The second option I had was that I set experimental temperatures and heat fluxes at all inner surfaces (except facade), and give the film coefficient to the facade.

 

The problem is that I do not have the measured heat flux through the window. But I have the U-value and outside air temperature (5°C). Is the selection of outer surface of the facade for film coefficient correct?

 

If both abovementioned ways are wrong, can I do the followings:

 - I set the experimental temperatures at all surfaces (except the facade) and set the film coefficient (1.3 W/m2K and 5°C) at the "outer surface of facade".

 

The cfz file including the two simulations is attached for your consideration.

 

https://drive.google.com/file/d/1iKEA3YQfYZmP5MgxqDZEY-RpDQsKGu_Q/view?usp=sharing

 

 

Best regards,

Reza

 

 


Report Inappropriate Content
Message 3 of 13 ( Views: 47 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/14/2018 8:50 AM

Hello David,
thank you very much for your prompt email and kind support.
Indeed, therefore I over constrained the BCs by giving two surface values (temperature and heat flux), however, I tried one simulation with measured surface temperatures at all surfaces including inner facade surface (17 ° C).
The simulation converged within 150 iterations but I had a higher temperature and MRT and higher temperature stratification compared to the reality and it had no convection.
Therefore, I added the heat flux to the walls, ceiling and floor and also film coefficient to the facade (with outer cold temperature) and
I had more reasonably temperatures in the room. However, the simulation was not converged within 1500 iterations because of fluctuated Vx,Vy,Vz. On one hand, I thought selecting the heat flux is correct because I had better temperature distribution and on the other hand I suspected the BCs because of fluctuated velocities and convergence.

Please find the cfz file via the following link for your consideration. It has the two above models.

 

https://drive.google.com/file/d/1iKEA3YQfYZmP5MgxqDZEY-RpDQsKGu_Q/view?usp=sharing


The 3rd solution I have is that I select the temperatures at ceiling, floor, and walls and only I give the film coefficient with the outer surface to the facade. Is it correct?

FYI, I could not measure a constant heat flux through the facade (as I did the same for other surfaces) because it was varying with outside and indoor temperatures. Therefore, I decided to choose film coefficient with constant outside temperature.

Thank you very much and best regards,
Reza

PS. I replied you yesterday, but the text was not available.


Report Inappropriate Content
Message 4 of 13 ( Views: 63 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/14/2018 12:37 PM

Hi Reza,

 

I have been having a go with your model and results are looking promising with a simpler setup. 

 

Test room.png

 

 

The main thing to think about is to replicate your inputs to the experiments and let CFD take care of the rest!

 

I applied temperature BCs to the outer surfaces of the walls, floor and ceiling. You can then measure the heat flux using the wall calculator to see if the results are matching the experiments.

 

For the facade/window, as we know the heat transfer coefficient there is no need to include the window material (you can suppress it if you like) therefore I applied a film coefficient on the internal surface of the window of 1.3 with a reference temp of 5 degrees.

 

In terms of solver settings, use ADV5 and K-epsilon as a starting point. Maybe consider using SST k-omega further down the line if you are not satisfied with the heat transfer in the model. Ensure auto forced convection is disabled if you wish to consider natural convection effects. 

 

With an optimised mesh (maybe think about using mesh adaptation) I think you will achieve the results you are looking for.

 

I have attached the .cfz of the setup i used.

 

Hope that helps.

 

All the best,

David


David Short
Technical Support Specialist, Simulation
Report Inappropriate Content
Message 5 of 13 ( Views: 53 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/14/2018 1:43 PM

Hello David,

 

thank you once again for your time.

 

The additional information may help. In our experiment, we should keep internal surfaces of the walls and the floor at 20C and the ceiling at the 35C. The surfaces are cooled and heated by water circulation through the pipes. Therefore, our interest is to pay attention to the temperature,  not the heat flux. 

 

I am a bit confused that why do you choose the external surfaces as the boundary conditions as 20C and 40C and why do choose the internal surface of the window as film coefficient and with the temperature of 5C (The outside air temperature was about 5C).

 

Regarding your reply, I am thankful if you can explain your following comment

"For the facade/window, as we know the heat transfer coefficient there is no need to include the window material (you can suppress it if you like) therefore I applied a film coefficient on the internal surface of the window of 1.3 with a reference temp of 5 degrees."

 

 

Additionally, as attached, I chose the gave the 20C to internal surfaces of walls and floor and 35C to the internal surface of the ceiling, and I chose film coefficient to the outer surface of the window. However, the wall calculator gives my the heat flux as "0" if select the internal surface of the ceiling (which has the surface temperature of 35C). I am surprised that why it does not give the heat flux as zero! Any physics behind of that?

 

Best regards,

Reza

 

ATTACHMENTS:
aaaa.png
bbbb.png

Report Inappropriate Content
Message 6 of 13 ( Views: 48 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/15/2018 8:55 AM

Hi Reza,

 

The issue with applying the temperatures on the internal surfaces of the walls means that there are internal boundary conditions in your simulation. This is generally considered bad practice as the solver struggles to treat the BCs correctly in the solution matrices.

 

If you wish to apply the temperatures to the internal surfaces then suppress the solids external to them (i.e. the walls floor and ceiling). But then the effects of the wall, floor and ceiling materials will not be considered.

I thought applying the BCs to the external surfaces was actually more representative of the experiment as you heated /cooled the walls and not the fluid volume directly. With the temps on the external surfaces you also seem to get sensible heat fluxes through the internal surfaces (about 24W on each wall and 14 W through the internal surface of the ceiling). Not sure why you are getting 0 W, that doesn't sound right. 

 

Reagarding the window:

 

The U valve that you measured in experiments represents the performance of the window so my thinking was that there is no need to model the window its self. Therefore suppress the window (i forgot to do this in the setup i sent.. apologies!) and apply the film coefficient (U value) to the internal surface (interface between are and window). 

 

Once you have BCs you are happy with you should concentrate on getting an optimized mesh. Mesh adaptation could work well for your model i think.

 

Hope that helps,

David


David Short
Technical Support Specialist, Simulation

Report Inappropriate Content
Message 7 of 13 ( Views: 38 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/15/2018 4:15 PM

Dear David,

 

thank you very much for your clear explanation.

 

I followed your structure and the results and the model are as attached. Could you please give me your last comments regarding the BCs and if it is correct.

I have two doubts regarding the convergence and the heat fluxes through the surfaces. The simulation converged despite relatively strong fluctuation of temperature, pressure, and velocity. Additionally, the source heat flux through the ceiling is much higher than the sink heat fluxes through other surfaces. Is it reasonable?

 

What type of Turbulent do you suggest for this kind problem (BCs with radiation)?

 

 

Best regards,

Reza

Report Inappropriate Content
Message 8 of 13 ( Views: 31 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/16/2018 8:43 AM

Hi Reza,

 

Your setup looks good if you want to apply BCs on the internal surfaces. My only concern is measuring heat fluxes over surfaces that have temperature BCs applied on them may through erroneous results. I prefer to apply the temps on the external surfaces  then measure heat fluxes on the internal surfaces, you could always increase the external temp BCs such that the internal surfaces are at the desired 20 and 35 degree temps.

 

Having said that I am interested to see how your setup gets on.

 

Once you have a good mesh the energy in the system should balance. To achieve an appropriate mesh either use mesh adaptation or apply a manual mesh study. With a refined mesh you would likely achieve good results with either k-e or SST k-omega (this will require specific boundary layer meshing).

 

Hope that helps,

David


David Short
Technical Support Specialist, Simulation

Report Inappropriate Content
Message 9 of 13 ( Views: 26 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/19/2018 10:04 AM

Dear David,

 

I have two concerns and confusion regarding applying BCs on external surfaces as follows:

1- According to which physics, I cannot apply internal surface? and, why I cannot measure the heat flux on internal surfaces?

2- In other forums, it mentioned that it is suggested to consider the solid in the calculations if the heat transfer through the solid is important. Since the temperatures of surfaces are important for me and heat transfer through the solid materials is not my concern, should I still apply the temperature on external BCs?

3- It is relatively difficult to do trial-and-error to set different temperatures in order to get desirable temperatures on internal surfaces (20 and 35°C). Additionally, I have several Boundary conditions (30, 33, 35, 40 °C) which make it more difficult to repeat trail-and-error for different cases.

 

Regarding aforementioned concerns, what would be your suggestion? Should I still choose outside BCs?

 

Thanks and best regards,

Reza


Report Inappropriate Content
Message 10 of 13 ( Views: 17 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/19/2018 12:27 PM

Hi Reza,

 

Regarding internal condition, it is more a matter of mathematics rather than physics. The matrix solvers can get confused if boundary conditions are applied internally.

 

It is really up to you whether you apply BCs on the exterior surfaces or interior. You have two options:

 

1. Apply BCs on internal surfaces and suppress the solid parts external to them. This will allow you to easily set your desired temperatures but heat flux results may be compromised as the wall calculator may struggle...test it and see!

 

2. Apply BCs externally. With a good mesh through the solids the initial temperatures on the internally surfaces should be fairly similar.

 

As always in CFD ensure you have a good mesh and converged results then play around with the boundary conditions to find which method you prefer and match the experimental results the best.

 

All the best,

David


David Short
Technical Support Specialist, Simulation

Report Inappropriate Content
Message 11 of 13 ( Views: 13 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/19/2018 12:31 PM

Hello David,

 

thank you very much for your prompt reply and your time.

 

I am trying different options as suggested and let you know if I would have problem.

 

Best regards,

Reza


Report Inappropriate Content
Message 12 of 13 ( Views: 11 )

Re: Model study Check: A room with radiant cooling and heating surfaces

02/19/2018 12:37 PM

No problem. Happy to help.

 

Good stuff...let me know!

 

All the best,

David


David Short
Technical Support Specialist, Simulation

Report Inappropriate Content
Message 13 of 13 ( Views: 7 )