I came across this interesting hot climate design. At night, water is pumped from a storage tank and sprayed on the roof where radiation, convection and evaporation cool it down. It then flows back to the tank (like a rain water collection system). During the day, a water source heat pump uses the tank water instead of a ground loop. Do the numbers work - maybe in hot dry climates.

You can get a lot of evaporative cooling (day or night) when the dew points are in the 40s F or lower.

The emissivity of water is pretty high, but not dramatically more than most roofing materials, so while a wet roof has modestly better radiational cooling than a dry roof, the evaporative cooling would probably dominate.

I don't see how using this scheme at night makes better sense than during the day.

> I don't see how using this scheme at night makes better sense than during the day.

It's another twist on diurnal storage. A tank/cistern full of 70F water is a better source for cooling than an 80F ground loop or 95F air. Try getting 70F water running off the roof during the day (100F+ is more likely). On the other hand, with the right circumstances (calm, clear night, etc) you can get ice off a roof even when the air temperature stays well above freezing.

So if I were to hook up some kind of water distribution system across the surface of my roof, where the water flow would be such that none makes it to the gutters (evaporates), would the attic temperature be cooled enough where the HVAC kWh reduction would more than offset the cost of water supply?

I've often wondered about this, but don't know how to put a pencil to the numbers.

My residence is in Dallas where the summers can be blistering (currently upwards of 106 degree highs). It's single story, 3400 sf (A/C) and thus roof sf is a lot more. Roof shingles are typical asphalt type. Attic insulation is typical R38 fiberglass fill on the floor, and nothing else. Geothermal units (2) are in the unconditioned attic, along with all ducts (steel pipe wrapped with R8 insulation.

Dew points this week are in the mid-60s. Last week they were in the upper 60s to low 70s. Cooling Degree Days annually are about 3400.

Would greatly appreciate comment from those much smarter than me. Is it time to put a soaker hose up along the roof's ridge, with a slow trickle such that no water makes it to the gutters, or would I be wasting my money (water cost)?

Many thanks!

Best regards, Bill

The heat of vaporization is about 970 BTUs per pound of water.  There are 8.34 lbs of water per gallon(or 62.4 lbs per cubic foot), so you soak up about 8090 BTUs of heat per gallon (or 60528 BTUs per cubic foot).

But it's all happening on the outside of the house, not the inside.  The model for how much cooling that delivers to the conditioned space is not universal or simple.  The solar reflective index of the roofing materials and average temperature reduction on the interior roof deck you get a  known flow rate might get you closer, but I doubt you'd hit better than order-of-magnitude on the actual cooling efficiency even with some in-situ measuring & monitoring. 

Evaporative cooling (swamp coolers) is usually cost effective relative to air-source AC when the evaporation is extracting the heat directly from the conditioned space, but isn't very useful in places as humid as Dallas, since it's only lowering the temperature converting the sensible heat to latent-heat, and you already have a substantial latent load you're trying to remove with the heat pumps.

I would try it. Put a sprinkler up there and measure the change in attic temperature. Then you have some numbers that are easier to work with.

Latent vs sensible loads vary greatly, with one factor being how well the house is sealed (a perfectly sealed house would approach zero latent load, independent of climate).

Im not sure I agree with the statement about zero latent loads. Houses have occupants. Respiration, showers,indoor plants etc. all contribute to latent loads.

That's why "approach zero" and not zero, although the hours when nobody is home could come very close. A couple of people would add around .03 tons latent (still very close to zero), not accounting for fresh air needs.

Some of this is actually relevant in cases like unmanned data centers.

What is being described here is a crude form of wet cooling tower. Heat can be rejected at temperatures approaching we bulb temperature via use of a wet cooling tower instead of a dry coil.

Wet bulb temperatures are nearly always well below dry bulb temperatures, but often 5-10 degrees higher than dewpoint. Wet towers consume some water and require lots of maintenance, so are not common in small systems, but savings of 30% are attractive to operators of large systems in office buildings or campuses.

Facing some work in a hot attic of a rental, I set sprinklers on the roof - not much help. An area outdoorsy restaurant "crab shack" runs water on its tin roofs during hot weather...it seems to help. My guess is that is 72*F well water, because they run a lot of it.

I have toyed with the idea of connecting a water source heat pump to a swimming pool, and running a fountain or other water feature whenever pool temperature exceeds a set limit. Obviously there would have to be an intermediate heat exchanger or careful management of water chemistry

This isn't answering Bill's question about daytime roof use, but I'll guess that someone who already has a good ground loop that is < 90F peak wouldn't find ~70F from a night-time roof radiator worth it (maybe 25% savings). On the other hand, if the loop is hitting 100F+ then maybe and if the system is shutting down due to over heating, very probably.

I have no idea if it makes sense to cool down the ground loop instead of storing water in a tank. The former is certainly easier. One can automate use - running on dry, clear, cooler nights yields lower temps.

It is not necessary to have the system work with evaporation. You can use a radiator system installed under the roofing material to dump heat at night. I have been testing a product for the last 2 months to see what it can do. According to a study commissioned by the State of New Mexico in 2007 can get rid of 75% of your AC bill even in the hottest areas of New Mexico and Arizona.

Do you have a reference to the study?

Jon, I just put up a quick page on my web site. I do not have permission from the author yet so I may have to take it down. But it is there for now. http://www.sol-ice.com/solar-products/night-sky-radiation-cooling/ Forgive how sloppy the page is I just wanted to get you some information from what we are working on to try and make this a reality for the average american family.

It shows how radiative cooling drops the temperature to below ambient. But it also seems to show that radiative and nighttime evaporative combined would make it even more effective. Would be interesting to know more about how well this would work in other climates.

A test that showed 50% savings and a very rapid payback.

http://www.ornl.gov/sci/eere/international/Website/Night%20Sky%20Radiant%20Cooling.doc

jonr

That is a cool study. In the testing I did last month on our closed loop system I had my sprinklers hit the system and the U-Factor went from 2.4 to about 6. Check this link on my web site to see the chart from that night. http://www.sol-ice.com/night-sky-radiation-coolingbuilding-integrated-solar-thermal-bist/

This relates to a topic I am studying.
I have about 500 square feet of solar panels
on my roof. I saw some experiments showing
the panels to lose efficiency at high temperatures.

I was thinking a simple circuit could allow
a small amount of water to trickle across the
solar panels in the daytime.
(The circuit would need to ensure that
the panels are not so hot already that
they would crack)

The above article offers an idea
what to do with this circulation equipment
at night!

There are all kinds of factors and I
don't know when I will work on this,
but perhaps for the 2012 season...

Posted By DennisDuce on 02 Sep 2012 10:27 PM
jonr

That is a cool study. In the testing I did last month on our closed loop system I had my sprinklers hit the system and the U-Factor went from 2.4 to about 6. Check this link on my web site to see the chart from that night. http://www.sol-ice.com/night-sky-radiation-coolingbuilding-integrated-solar-thermal-bist/

Cool, that is in my neck of the woods, Prescott Arizona.

The numbers seem to work for using NSRC with a water source heat pump. If you are only interested in cooling, it might be a viable alternative to a geothermal ground loop. A large water tank is the primary cost.

Posted By jonr on 02 Sep 2012 08:48 PM
A test that showed 50% savings and a very rapid payback.

http://www.ornl.gov/sci/eere/international/Website/Night%20Sky%20Radiant%20Cooling.doc

I don't think residential would see as much or as quick payback. The study showed the application on a commercial building that had cooling bills of over $5,000 a year. I'm in southern Arizona and my residential cooling bill or A/C is around $1,300 a year (May/June - September/October).

If the unit is around $10k installed, and you cut my A/C bills by 50%, It would take 15 years to pay off the unit. This not taking into account any breakdowns. Unless the residential unit would cost half of the commercial unit or $5k, then it would take 7.5 years to pay it off. I am guessing that the residential unit costs half, someone would have to verify this.

I would like to see some studies on residential homes and what the costs were.