[Build Log] PC Case Thermodynamic Simulation Programming Project

[JrClocker]

I'm tossing around an idea of writing a thermo-dynamic simulation to model heat sources, fan locations, fan speeds, etc. inside a PC case.

Why?

Because I need a project...and get bored if I don't have one!

Also, I don't use my engineering skills as much as I used to in my current position, so I like to find projects like this to remind myself why I went to engineering school! (And to keep my Physics and Calculus skills sharp!)

Is this something that people in this forum would be interested in? First to help me test to see if the results are qualitatively accurate (I probably won't want to spend the time with the various sub-models to make it quantitatively accurate).

What I'm hoping to get out of it is a simulation result which will help put a stop to the endless debate over what is better where for what reason. Said differently, I hope the models will be able to tell that "this setup" is better than "that setup"...but it won't be accurate to 0.1 °C (probably a few degrees).

To that end, the high level overview of the tasks is:

1.) Basic user interface to draw (crudely) the case externals and internals
----- 2D drawing (i.e. looking into the side of your case)
----- I don't want to even attempt a 3D drawing interface as the main focus is not a drawing program - hehe
2.) Pre-determined boundary conditions for the following:
----- Fan intake (will require P/Q curve data for fan...assume linear relation with RPM value
----- Fan exhaust (will require P/Q curve data for fan...assume linear relation with RPM value
----- Resistance of holes/slots in the case for various standard case holes and slots
----- Assuming a standard heat transfer profile for the front and back covers of the case...can get more complicated if required
----- Assumptions about linear vs laminar vs turbulent airflow...I'm not interested in a PhD these about modeling something that goes back and forth between these
3.) Water Cooling Heat sources
----- For radiators, start with a bulk model assuming heat is dissipated evenly...get more complicated if required
----- Assume no heat is transferred through tubing...get more complicated if required
----- Assume that a portion of the heat is not removed through the water block...get more complicated if required
4.) Air cooling Heat sink modeling
----- This will be a tough one, as it will be different for every single heat sink
----- If the heat sink has a published thermal impedance (°C/Watt) this will make things easier
----- Assume even heat distribution across the entire heat sink surface area fins...not interested in a PhD thesis on this topic
5.) "Standard" fan models
----- Require P/Q curves
----- Assume equal airflow through diameter of fan...can get more complicated if required, but not interested in a PhD these on this topic
6.) Steady state temperature plot

Extra bonus (if I can figure it out)
1.) Bulk resistance (or resistance plot) of the case internals/externals...to determine what type of fan
2.) Transient temperature plots (i.e. temperature profiles over time)
3.) Fan airflow "shaping" for various types of fans...attempt to create a simple model for the fan "dead spot" behind the rotor.



Again...I'm still forming this project in my head. If anybody wants to add additional "nice-to-haves" or "must-haves", reply to this post and I'll determine if I can do it.

Let me know if you want to help out!

Thanks for reading.

 

[InfoSponge16]

JrClocker, that sounds like a really useful tool. It also sounds incredibly difficult. But I don't know how to program or a whole lot about physics.

I actually do the same thing but only on paper. It's definitely not exact but it does really come down to doing some math to get a decent idea of what to expect.

For instance, 2X 140 intake at 72 cfm each gives you an idea of intake flow.
Depending on case design.

1 120 exhaust at 65 cfm exhaust.

But you also have to assume fan curves.

Then you have to give a value for ambient temp. Then you can average out your temps based on also assumed thermal transfer efficiency.

I have only done this with sketches of liquid cooled setups with full custom loops.

Surprisingly, the numbers are very close to what I see with peoples actual systems.

I used 75,80,85,90% efficient thermal transfer numbers for my math. I found 80% to be the closest to actual numbers.
I was using a 10 percent loss at block and another 10 percent loss at the radiators.

That's as much as I can help.

Good luck on the project if you decide to do it. It was fun for me designing some projects I don't actually have the parts for.