Natural Convection Heat Sink Fouling

[MikePAMechEng]

Does anyone have a good resource recommendation or general rules for estimating fouling on natural convection heat sinks?

I'm evaluating lifetime performance of an air cooled, natural convection heat sink for an electronics enclosure. It will be in an industrial area with mildly oily air and some dirt/dust. I haven't been able to find any good values for this kind of grime that will build up on the fins. I've found fouling values for other things like diesel exhaust, air, and oily air. There just doesn't seem to be anything good for "dusty grime".

Right now, we just assume a 15% penalty on a calculated thermal resistance between the heat sink temperature and ambient air to account for fouling. I don't know if this is close to accurate, or how this value was determined other than "15% feels about right".

I plan to build a prototype and get it as dirty as other surfaces in the are (visually) to see how it acts, but I want to know if there are any better ideas to help optimize the design before building a prototype.

 

[racookpe1978]

No. YOUR prototype test piece is THE most important source. No literature or even the most carefully matches research paper will be any better than its own limits on the measurement proposed. Example: I was looking for heat transfer coefficients data for a water surface losing heat to air under various wind speeds. Fopund one similar, but it measured and reported lab data for a carefully contrived oil-covered water tank 1m x 1m with air carefully being blown over by a regulated fan at near-laminar-flow speeds. Not the open water, turbulent open air winds in real life, and certainly not at the Arctic conditions I needed to simulate.

Even under so rigorously defined conditions and lab setup, the report claimed no better than +/- 15%.

However. If you have the time (to make the prototype (better yet - make 3x prototypes: Two in "typical" locations near your actual process point, and one more outside in a clean, clear air at very close to the same net temperature difference!) and the money to test, do the test!

Now, the "dirt collection" period could be pretty long. Certainly much, much longer than the few hours to measure the various Delta-T's after the test plate stabilizes in temperature. That's the advantage of the third sample in a remote spot. While the two real samples get "real world" dirt and oil and grime, the test sample can get manually sprayed with oil, the temperatures stabilize, then measured. Wipe it off, clean it with solvent, do it again with more or less oil, more or less dirt and grime, or at a different surface temperature or ambient temperature.

 

[MikePAMechEng]

Yes. I will be testing a prototype. I was hoping to find ranges of values that others had seen in practice. If it's 15-30%, then I can probably live with it. If it's 50-60%, then it might not work at all or I'll have to make other tradeoffs to make cleaning more frequently possible.

 

[hacksaw]

Common practice in refineries and moderate heat load, uses instrument air (dry and filtered) purges.

 

[chicopee]

My input is directed to racoopke1978 about heat loss between water surface and air under various air flows. Research ME handbooks about cooling ponds if you have not done so yet.