Heatsink and Fan Design/Simulation

[KyleGranger]

TL;DR

I am trying to simulate the cooling performance of a fan and heatsink

I would like to do a simulation comparing the cooling performance of a heatsink with various parameters such as fin length, fin thickness, fin spacing, etc. I have Ansys Icepak available for this simulation.

From what I see, I can set the inlet pressure or velocity. The fan gives me CFM so I'm not sure how to relate these two. Any help with this or any other suggestions are greatly appreciated.

 

[jrmichler]

The CFM of a fan, any fan, is dependent on the back pressure. That relationship is shown in a fan curve. The figure below is the fan curves for two different fans that I randomly picked from the Digikey catalog: https://www.digikey.com/. There are other suppliers just as good, I randomly picked Digikey for this example.

In order to find the air flow rate for a particular heatsink, you need the system curve for that heatsink. That curve starts at 0,0 (zero pressure, zero airflow), then curves parabolically upward. The actual airflow is the intersection of the system curve with the fan curve. That intersection gives you both the inlet velocity (from the total airflow), and the inlet pressure.

Digikey has over 100,000 heatsinks in their catalog, and most of them have datasheets. You should be able find something similar to your designs, and get the system curve from that. Those datasheets also have thermal resistance information that you can use to check your results.

 

[KyleGranger]

The CFM of a fan, any fan, is dependent on the back pressure. That relationship is shown in a fan curve. The figure below is the fan curves for two different fans that I randomly picked from the Digikey catalog: https://www.digikey.com/.

Thanks. I've seen those curves and figured I would just go with the CFM listed on the specs summary from Digikey for the initial simulations. I guess that wouldn't really be a fair comparison since CFM will change depending on the physical parameters of my heatsink. I think Icepak will allow me to add a curve. I'll see if I can figure that out and add a curve from a fan Digikey has in stock to get a better comparison between heatsink designs.

 

[256bits]

Summary:: I am trying to simulate the cooling performance of a fan and heatsink

I would like to do a simulation comparing the cooling performance of a heatsink with various parameters such as fin length, fin thickness, fin spacing, etc. I have Ansys Icepak available for this simulation.

From what I see, I can set the inlet pressure or velocity. The fan gives me CFM so I'm not sure how to relate these two. Any help with this or any other suggestions are greatly appreciated.

Are you looking for optimization of materials, space or of some other factor? of the fan/ heat sink combination.

Or of just investigating how the heat transfer varies by changing the heat sink parameters with a constant air flow over the fins.

 

[KyleGranger]

Are you looking for optimization of materials, space or of some other factor? of the fan/ heat sink combination.

Or of just investigating how the heat transfer varies by changing the heat sink parameters with a constant air flow over the fins.

At the end of the day, I'm trying to optimize space. For instance, I would like to see how long is long enough for the fins. At what length do I start to just waste space because the additional heat removed is marginal. For that, I would also need to optimize fin spacing. Fin width would be another parameter to optimize but I'm adding that third degree of freedom would complicate the problem significantly.

 

[russ_watters]

At the end of the day, I'm trying to optimize space. For instance, I would like to see how long is long enough for the fins. At what length do I start to just waste space because the additional heat removed is marginal. For that, I would also need to optimize fin spacing. Fin width would be another parameter to optimize but I'm adding that third degree of freedom would complicate the problem significantly.

Are you analyzing the fan as a stand-alone parameter to analyze? Increasing the airflow by increasing fan performance is generally a big help. Downside; noise.

What is the purpose of this investigation? Not the output, but why are you doing this? The answer affects the approach. It's one thing if you're trying to do a master's thesis and another if you are just trying to optimize your TEC refrigerator.

 

[KyleGranger]

Are you analyzing the fan as a stand-alone parameter to analyze? Increasing the airflow by increasing fan performance is generally a big help. Downside; noise.

What is the purpose of this investigation? Not the output, but why are you doing this? The answer affects the approach. It's one thing if you're trying to do a master's thesis and another if you are just trying to optimize your TEC refrigerator.

Thanks for the response. I'm not doing this for a masters thesis, but I would still like to be able to do some some of the same analysis that you would expect from a thesis. So I am not wanting to just throw a bigger fan in until it meets my cooling needs. I'm trying to get a better understanding myself on how the different parameters effect the cooling system efficiency, which I'm getting a lot from literature review, then take that understanding and apply it to my actual system. That will require me to model my system and run several simulations appropriately, then interpret the results.

I figured out how to import the fan curves into Ansys so that should improve my results. I know it will still take a general understanding for what's happening since that introduces more degrees of freedom. Starting off with a constant air flow may be the better place to start for a general understanding, but I will also have to understand how changing each parameter affects pressure since that affects CFM.