I wouldn't exactly call it easy, but doable :P
You're right though, it's more complicated than just guessing. The coldest point would probably be right around the manifold, but humidity could be higher in the crawlspace, where it matters most. Sticking a temp/humidity sensor into the crawlspace or somewhere into the floor sandwich, calculating the condensation point and using that data to turn on/off the pump and cooling wouldn't be too big of a deal; but can we assume that the temp/humidity is the same everywhere in the crawl?
I'll probably hook a sensor up and just try to see what sort of data I get. As for cooling, if we assume the water cools to 65F overnight (say via a ground loop), this should be doable...
[WARNING: lots of speculative math without much actual data to back it up follows]
Specific heat of water is 4.187 kJ/kgK, or 4.84 BTU/lbF. 33gal in my water heater is 264lbs, so if we assume we want to get to 78F room temperature with 65F water, that's roughly 16k BTUs of cooling capacity in the water.
Previous data I measured suggested a heat gain of 400BTU/hF or so, which would mean 2800BTU/h for a 7 degree difference. Let's triple that and round up for a sunny day to 8500 BTU/h. 16000-2400 = 13600 / 8500 = 1.6, so it would take only about an hour and a half for the water to heat up to 78 degrees.
Basically, that means if I could get the water down to 65 degrees overnight, I could run it through the loops for a total of 90 minutes *assuming ideal heat transfer* - so basically, assuming I could actually cool the air to 78 degrees on an 85 degree day with the floors alone. I don't expect the transfer to be nearly as good, so the time may be closer to 4 hours with a less pronounced cooling effect.
I guess the numbers make it look not entirely outside of the realm of possibility. I might be able to keep it down to 80 or so, if I can convince everyone to keep the doors and windows closed :P
|