Community
CFD Forum
Welcome to Autodesk’s CFD Forums. Share your knowledge, ask questions, and explore popular CFD topics.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Roof Chiller Simulation

9 REPLIES 9
SOLVED
Reply
Message 1 of 10
Q-FGA
1257 Views, 9 Replies

Roof Chiller Simulation

Dear all,

 

We're simulating the operation of four chillers located in an enclosure on the roof of a building. We have modeled the chiller fans as Heat Exchangers as we will be analyzing both flow and temperature increase because of the enclosure.

 

We have set the external volume (air) far apart as the tutorial suggests (7D) and ran an initial flow only analysis with no boundary conditions. Running on advection 1 with turbulent k-epsilon model and volume mesh growth of 1.25. We successfully solve the model with apparently correct values.

 

We then apply the 0 gauge static pressure BC to the top of the external volume to simulate the open space and solve for flow only. In this case the solution diverges.

 

After changing to advection 5 we get solution convergence but the heat exchangers (fans) show flow upwards and downwards as it they had two outputs instead of one input+one output.

 

We have also tried opening the top and the bottom with the 0 gauge static pressure as if it was a free hanging model but get the same behavior.

 

I'm attaching an image of the vector results near the heat exchangers. The flow upwards to the open space seems OK.

 

Any help would be appreciated.

 

Best regards,

 

Francesc

9 REPLIES 9
Message 2 of 10
Jon.Wilde
in reply to: Q-FGA

Hi Francesc,

 

I think it would be best to point you towards the Heat Exchanger limitations shown here: http://wikihelp.autodesk.com/Simulation_CFD/enu/2014/Help/0532-User_s_G532/0698-Process698/0719-Setu...

 

There are a couple here to bear in mind.

One is that they must only be used within closed loop systems

The other is that they must not be within a system but outside of it - or within a suppressed block.

 

Hopefully this helps explain why this is not working. The other question though is how to set this up so that it does work, we can help with that also if you would like, although we would need more detail on the model setup.

 

Kind regards,

Jon

Message 3 of 10
Q-FGA
in reply to: Jon.Wilde

Dear Jon,

 

Thanks for pointing out the limitations of the heat exchangers.

 

We can easily integrate the heat exchangers into an block so the mesh can be suppressed. Only the inlet and the outlet are in contact with the meshed air.

 

We need the system to be in an open space to allow for heat dissipation.

 

It would be great if you could help us with the setup. Let me give you more details on what we've done so far:

 

We have created the model enclosed in walls and roof in accordance to the "bucket" configuration 7D width and depth, and 5D height. The interior of the enclosure is the outdoor air.

 

General Model View.jpg

 

The model itself consists in four chillers located inside four walls (Simulating an acoustic enclosure with no roof) and some concrete structures outside the enclosure.

 

Model View.jpg

 

The chillers themselves have a constant airflow of 52.596m3/h and constant heat dissipation of 150kW. The chillers use the outdoor air to dissipate the heat. The air enters the heat exchanger from the bottom of the chiller, heats up and exits in an upwards direction. The heat exchanger have been modeled as cylinders (Simulating the chiller fans) embedded in a block which is left unmeshed. A couple of resistances have been included in the chiller block to simulate the effect of the cold water coils and filters. The following image is a breakdown of the chiller parts.

 

Chiller Parts View.jpg

 

The ambient temperature is 43ºC. The aim is to check for recirculation of hot air to see if the inlet temperature rises above the 43ºC at steady state. Therefore we need some kind of mechanism for the hot air to be evacuated from the model and fresh air to enter at 43ºC simulation an open space.

 

Please let us know if you need any more details.

 

Best regards,

 

Francesc Galobardes

 

Message 4 of 10
Jon.Wilde
in reply to: Q-FGA

Hi Francesc, 

 

Thank you for this information, it certainly helps understand the problem.

 

I have a suggestion which could help. We certainly do need to leave this as a closed system but how about using film coefficients on one or all of the walls of the air domain? Here we could draw heat out and set a reference ambient temperature too. Something like 20 W/m2/K at 43C should be enough. This is quite a high coefficient but I suggest you test it and see how it performs.

 

Please let me know if this helps you move forward.

 

Kind regards,

Jon

Message 5 of 10
Q-FGA
in reply to: Jon.Wilde

Dear Jon,

 

We have used a closed system and obtained what appears to be a correct airflow through the heat exchangers. The flow moves upwards into the ambient the its velocity decreases. We can also observe some recirculation which was expected.

 

We also applied film coefficients ranging from 10 to 30W/m2/K to analyze the differences in the thermal behavior. According to the Summary File - Fluid Energy Balance Information - (Energy Out - Energy In) value, we get around 570Watt when using 10W/m2/K which should be a good balance.

 

We only see one minor issue in the same fluid energy section of the summary file. MdotIn x Cp x DeltaT = -597kW which is in accordance to the 600kW of the heat exchangers. But then under the Heat Transfer from Wall to Fluid we get -460kW. Then in Solid Energy Balance . Heat Transfer from Fluid to Solid 475kW. 

 

Summay File.jpg

 

Why do the theoretical 600kW decrease to around the -460kW? We assume that the MdotIn x Cp x DeltaT is obtained from the calculations over the while domain and therefore should hold true.

 

Thanks for your support.

 

Best regards,

 

Francesc Galobardes

Message 6 of 10
Jon.Wilde
in reply to: Q-FGA

Pleased to see it is running better now.

 

I would suggest using Advection Scheme 5 here (Solve -> Solution Controls -> Advection) and also you may need to refine the mesh to further improve these values.

 

Kind regards,

Jon

Message 7 of 10
imatoric
in reply to: Jon.Wilde

Hello All,

 

I am doing a similar simulation in Autodesk CFD 2016 SP2 and I have based my model on the instructions I have read here.

 

The issue that happens with my model is that the entire air in the domain has a temperature rise about 5-10C which in my opinion is not possible.

 

It seems like the heat generated by the exchangers is not being removed. I tried apply several different values of film coefficients.

 

Am I missing something with the film coefficients?

 

Thanks,

Message 8 of 10
imatoric
in reply to: Jon.Wilde

Hello All,

 

I am still cant get the simulation to work. Was the film coef. applied to the outter walls while they were suppressed?

 

Thanks,

Message 9 of 10
Jon.Wilde
in reply to: imatoric

Hi - could I suggest you start a new thread? This one is marked as solved 😉

That would be super useful!

Message 10 of 10
imatoric
in reply to: Jon.Wilde

Hello,

 

I oppened a new thread as u suggusted:

 

https://forums.autodesk.com/t5/cfd-general-discussion/roof-chiller-simulation-outdoor-air-temperatur...

 

Thanks and best regards

 

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums  

Autodesk Design & Make Report