Discussion Groups

Simulation CFD

Reply
New Member
aperri
Posts: 2
Registered: ‎11-03-2013

Closed Loop Cooling system

261 Views, 3 Replies
11-03-2013 08:28 PM

Hi all, (first post)

 

This is my first time working with CFD and I have done a few of the tutorials available on the Autodesk site. Up to this point I have been testing and slowing advancing the model of a Stator/sleeve combo for a cooling heat exchanger for a project.

 

I have not had many problems with the individual part, but now I am trying to make a closed loop system and I had a few questions (plus getting the simulation to run would be great :smileyhappy:  ).

 

The system consists of a pump, two stator/sleeve combos ( in series) followed by a radiator all linked by some tubing. We would like to see if the fluid is able to be sufficiently cooled after extended use of the motors.

 

 

 

I had the following questions:

 

The stators/sleeves are fairly detailed for the simulation as they are the focus of the entire simulation. I have the pump and radiator set to be EXTREMELY basic geometries (essentially cubes). I did this as I have the manufacturers data sheets concerning the pumps and radiators concerning properties such as Pressure, Flow Rate, and thermal resistance. 

 

Should I improve the models of these devices despite the fact they are I only need them for their properties ( should I stop treating them like black boxes)?

 

As I said, I have some data from the manufacturer concerning the various properties of the parts, but when when working with the editor I am having difficulties finding how to utilize these values.

 

Is there a way to provide non-static functions for properties such as flow rate vs. pressure change and thermal resistance vs. flowrate?

(the only non-static models I see involve time dependent functions, whereas these are property dependent.)

If there is not, am I forced to try and find comparable pumps?

[I have yet to work on the radiator, but I believe it would be using the Heat exchanger option (which also doesn't seem adjustable to actual specs)]

 

I also did not include the reservoir as it would just be another black box where the pressure would drop before the pump. I was thinking the pressure could be assigned to something in the flow without the need of another black box. If I add it, I also was wondering if I would have to subject it to two fluids (delving into free surfaces) .

 

Should I include a reservoir into my model, if so, how would I go about that?

 

 

 


I'll attach my .cfz in order to aid in my description of the system, [at the moment it doesn't run as I have not yet added my recent epiphany concerning the way the internal fan/pump should be structured, which to my knowledge can be best described as  water<-pump<-water (surrounded by solid) where in this model I assumed water (surrounded by pump) being two literal...]

 

Please forgive any grievous errors (but please do point them out :smileyhappy: )

 

Thank you for any help you can give. (if it is a "Read the Manual" kind of answer, could you throw out a good keyword for the relevant situations?)

 

Tony

 

(edit: the cfz is 4mb and is over the upload limit, is the cfdst acceptable?)

Picture (the outer volume of Air hidden, water is the working fluid):

Current System

Please use plain text.
Product Support
wildej
Posts: 726
Registered: ‎08-25-2011

Re: Closed Loop Cooling system

11-04-2013 01:22 AM in reply to: aperri

Hi Tony,

 

Thanks for the comprehensive details, I will try to answer everything as best I can although we would really need to see the CFZ to comment on the model - feel free to share a dropbox link or something via pm.

 

Using the black box approach is proably the best starting point, you can always add detail in later should you need to. Just ensure that they all have a good uniform mesh across them (5 elements inlet to outlet)

 

Yes, you can provide non-static functions. Just edit the material, save it into your own material database rather than using the default and start to change the material properties.

I am not sure you could have the thermal resistance change with flow rate but you could certainly have pressure and flow rate related within a resistance. Within a resistance you can have the conductivity change in relation to temperature or pressure.

 

For the radiator I would suggest starting with a resistance with a negative volumetric heat load applied to it.

 

Regarding the reservior, if it just introduces a pressure drop - just use a resistance.

 

Do you have natural convection in the external air domain or is air forced past the assembly to help cool it? This is a point I am not clear on presently and it would impact some of the answers above regarding heat extraction.

 

Kind regards,

Jon

 



Jon Wilde
Please use plain text.
New Member
aperri
Posts: 2
Registered: ‎11-03-2013

Re: Closed Loop Cooling system

11-05-2013 07:06 AM in reply to: wildej

Thank you Jon,

 

I apologize for my delayed reply.

 

Thank you for each comment concerning the adjustment of the resistance material to better match the materials I have via the manufacturer. I never thought using a resistance material for the pump in order to get the pressure/flow relation. (I guess the pump could be considered a anti-resistor in that regard? ) .

 

I also wanted to thank you about the detail considering the number of elements for the inlet/outlet. As a newbie, I would have never thought of that.

 

I will take your advice concerning the beginning attempts for the radiator. If I normalize the volume (or just take it into account), perhaps I could achieve the same effect aft the Thermal Resistance vs Flowrate curves I get from the data detailing the radiators performance.

(Our radiator: http://www.swiftech.com/MCRX20-XP-RADIATOR-SERIES-1.aspx#tab1)

 

The system (once it works, I am baby stepping it), it will be subject to a moving external flow and the radiatior will have 120 mm fans in order to aide its cooling. Once the model is expanded to that level, the radiator is my next worry spot as I know that the convection cooling properties depend on the surface area (something I don't have directly...).

 

I hope you don't mind if I keep this thread open until the simulation is complete incase the radiator solution needs a bit more.

 

I will post a share link later in the day as I must create a copy in my dropbox. The delay is due to my absymal internet connection  (speedtest.net rated it an F+ ... Plus! :smileyhappy: at least it has something positive going for it...  ) Due to this connection, the uploading of files, even small, becomes a project of its own.

 

When my simulation is complete I will post what I had to do inorder to get it going (sharing the information with all fellow CFD starters)

 

Thanks again for your reply and Ill keep this thread up to date with my efforts,

Tony 

 

Please use plain text.
Product Support
wildej
Posts: 726
Registered: ‎08-25-2011

Re: Closed Loop Cooling system

11-05-2013 07:32 AM in reply to: aperri

Hi Tony,

 

Sounds like you are taking the right approach, even we try to take baby steps where we can.

 

I have some additional comments to make now, as I think I have caused some confusion.

The pump - try using one of our axial fan models if it is in-line. Here you can prescribe a flow/pressure curve and this will cause the fluid to move - you cannot do this with a resistance, they only provide a pressure drop. I presume it is only these that cause the fluid to move?

 

Regarding the full scale rad model - do you have a CAD model at all or do you have the manufacturers performance data? If you only had the CAD and this is just a suggestion but you could always run just a small section to characterise it's performance and then use a resistive region in the full model again.

 

Thanks for keeping the thread up to date, it is users like you who are/will be a huge help to others just starting out on their CFD journey.

 

I am happy to help you as you progress - as will other forum members I am sure.

(that is a particularly spectacular speedtest rating :smileyhappy: )

 

Thanks,

Jon

 

 

 



Jon Wilde
Please use plain text.