After re-reading some of Steve Baer's material on the "Double Play" solar heater and cooler, and the FSEC's work on radiation home cooling, I got to wondering if the same collector that I prototyped for the simple pool heating might also be put into service as a night time radiator to cool water. The idea being that water from a storage tank would be pumped over the radiator at night, where it is cooled by radiation and evaporation. The chilled water would stored in the tank would be used for space cooling the following day.
The cooler works through a combination of radiating heat to the cool night sky and cooling due to evaporation of a small part of the water flowing down the collector.
The results from a few overnight runs of a prototype of the cooler are shown below, along with some thoughts on how such a system might be used in practical applications.
The pictures show the prototype test setup. The radiator is 0.02 thick aluminum with pre-formed grooves that I used for water channels.
Water is pumped from the 41 gallon tank by a submersible pump in the tank to the manifold along the top of the radiator. A small stream of water from the manifold squirts out onto the collector surface at each water channel, and also half way between each water channel.
The water is collected at the bottom of the radiator by a gutter that returns the water to the tank.
The tank is a galvanized steel stock tank that is somewhat insulated.
Picture of the simple prototype: The
absorber/radiator is exactly the same one used for the
The cold water storage tank is under the insulation to the left. A submersible pump in the tank pumps water to the manifold
across the top of the radiator. The manifold has multiple holes across the top that deliver water to the top of the radiator.
The water runs down the radiator (cooling on the way), and is collected by the grey gutter along the bottom, which takes it back to the
the tank about 3F cooler than when it left.
The separate piece of aluminum toward the bottom right was used to measure how far below ambient a radiator with no water
flowing over it would get.
The radiator surface temperatures were measured with an IR thermometer and surface mount thermocouples stuck to the back side of the radiator.
The water temperatures were recorded using thermistor style sensors on an Onset Computer logger. Relative humidity and ambient temperature were recorded with a similar Onset logger in a shaded location a few feet away.
Wind velocity at the surface of the radiator was measured with a simple Dwyer vane type velocity meter.
The test plots and data below show at least the following:
- The radiator surface maintains a temperature about 10F to 13F below the ambient air temperature. This is presumably due to its losing heat by radiation to the sky. The section of radiator without water flow ran about 5F cooler yet.
- The radiator was able to cool the 0.05 gpm/sqft stream of water by about 2.5F to 3F during its travel down the radiator. Some cooling is due to evaporation and some to radiation -- not sure how much to each?
- Over the 12 hour night, the radiator cooled the water in the tank from 69F down to 49.5F, or 19.5F. This amounts to 467 BTU per sqft of radiator.
This method of cooling benefits from the low humidity, clear night skies, and rapid drop in ambient temperature that are common to this area (SW Montana), and would be less effective in areas with higher night temperatures, high humidity, high winds, and cloudier skis.
Note in the plot how rapidly the radiator/cooler turns into a solar collector water heater when the sun gets on it at 8am -- clearly showing it can do both under the right conditions.
Click on plots to enlarge
The plot below shows the night of 8/29/07 -- a mostly clear night with light winds.
It shows temperatures for ambient, collector input (tank), and collector output.
T ambient air Black
T collector inlet Green -- same as the tank temperature
T collector out Blue -- temperature of cooled water returning to tank
Relative Humidity for same period.
Additional data recorded during test:
|Radiation Cooling Test 8/29/07|
|Time||Tank Depth (in)||Wind (fpm)||Sky||T absorb F (2)||T raditor F (3)|
|(1) some forest fire smoke haze|
|(2) Temperature of absorber surface midway between water channesl -- surface mount thermocouple - deg F|
|(3) Temperature of alum sheet with no water flow -- surface mount thermocouple - deg F|
Wind speed is measured at the surface of the absorber -- typically less than 2 mph on the surface on this night.
Flow rate was 0.7 gpm, this is 0.05 gpm/sqft of collector.
Collector is about 45 inches by 46 inches -- about 14.25 sqft.
The tank is a galvanized metal stock tank, 45 inches long by 20.5 inches wide with rounded ends (this is our dogs cool off tank, and she is a little pissed a me for stealing it).
It was insulated with 2 inches of polyiso underneath and a lid of 2 inch polyiso.
The sides were insulated with 2 layers of Reflectex aluminized bubble wrap style insulation.
The tank holds 3.6 gallons per inch of depth.
Start volume = 41.85 gallons
End volume = 40.05 gallons
Water used overnight = 1.8 gallons
The pump is a submersible and sits at the bottom of the tank. It draws 16 watts (55 BTU/hr) -- an outside the tank pump that does not lose all its heat to the water might improve performance a little bit -- maybe 8% at most?
The 41 gallons (340 lbs) cooled from 69F down to 49.5F overnight -- a drop of 19.5F degrees.
This is (340 lb)(69F - 49.5F) (1 BTU/lb-F) = 6630 BTU, or 467 BTU per sqft of collector.
The cooling rate per sqft of collector, per hour of operation was:
BTU/sqft-hr = (40.5 gal)(8.33 lb/gal)(1.43F/hr)(1 BTU/lb-F) / (14.25 Sqft) = 33.8 BTU pers hour per sqft of collector.
Where 1.43 F is the average rate of cooling of the tank water per hour of operation from the plot.
So, this is a sort of figure of merit for comparing radiator cooling efficiency for different radiator consructions.
A 200 sqft radiator would then give about 93,000 BTU of cooling -- kind of a like a 1 ton AC operating for 8 hours.
The COP is about (6630 BTU)/( (16 w)(10hr)(3.412 BTU/ wh) ) = 12.1 (around 41 SEER?)
This seems good to me, and the pump was not chosen for efficiency, so one might do even better.
Several additional night long runs were done with similar results.
This thought just struck me: Suppose you transferred all of the water
from the storage tank to a fairly shallow and fairly large basin, and just let
it sit under the sky overnight, then in the morning, transfer it back to the
storage tank. How would the cooling compare to the scheme described above?
One of these transfers could be gravity flow.
Click pictures to enlarge
|Tan and submersible pump||Top manifold with water going.
To make a full home cooling system, you also need a way to distribute the coolth that has been collected in the storage tank to the house on the following day. In my case, I am thinking of using the system that I built to provide solar heating in the winter. This system has a set of solar heating collectors that heat the water in a 500 gallon storage tank. The heated water in the storage tank is in turn pumped through the radiant floor piping when the house needs heat.
For cooling, I am thinking of using the same tank and radiant floor system to store and deliver coolth instead of heat.
The south roof/wall of the garden shed have the solar heating collectors on them. These (I believe) would not make good radiators due to the IR opaque glazing used on them. I think that the the north roof of the same garden shed could be used for the cooling radiators. If the cooling rate per sqft that the test shows can be replicated, the shed roof will provide plenty of area to cool the water in the 500 gallon tank. I am thinking about a manifold along the roof ridge line to deliver water from the tank to the roof. The water would flow down the roof slope, cooling as it goes, and then be collected by a gutter, and directed back tot the tank. I believe that the same pump that is used to pump water to the collectors could be used to water the roof radiator.
If the 500 gallon tank can be cooled by the 20F that the test indicates, then it would store about 83000 BTU of coolth. This would provide us with the same effective cooling as running a 1 ton AC for 7 hours a day.
click on pictures to enlarge
The solar shed with collectors on the
The storage tank is inside the shed.
The cooling radiators would make use of the two north facing roof slopes -- in excess of 200 sqft of radiator area.
The tank inside the garden shed under
The cooling radiators could store cooth in the same tank that the collectors store heat.
The stored coolth is piped to the house via buried pipes, and circulated through the radiant floor.
I'm starting to think that this might just work.
Total cost of adding cooling would be a few dollars worth of pipe and valves, and some kind of timer or reverse differential controller to control the pump.
I realize that radiant floors are not ideal for cooling, but in our case where cooling demands are not high, and humidity is low, I think the radiant floor should be OK.
One might even envision a system in which collector panels similar to trickle style swimming pool panel are glazed during the winter for efficient heat collection (like a Thomason trickle collector), and then the glazing panels are removed in the summer to provide efficient evapro-radiative cooling in the summer. It might seems a bit impractical to remove the glazing panels each year, but I was struck by how easy this was on the system I have with the large but lightweight twinwall polycarbonate panels. I had to remove and reinstall all 6 panels once to fix a problem, and the round trip on 240 sqft of glazing only took about an hour -- I think this could be improved if it was designed to make the changeover easy -- I think this could become a 15 minute job with good design. This would make a system that could be used equally well for both heating and cooling even in climates with quite cold winters. It would also be dirt cheap to build -- perhaps $3 per sqft for the collectors.
This is a less elegant solution than the "double play" system, but the addition of the seasonally removable glazing might allow it to work in colder climates.
Another dual use system would use the trickle style swimming pool collector to heat the swimming pool during the day, and during the same night use the same collector to chill water in a storage tank that could then be used for house cooling on the fooling day.
Any thoughts or comments on any of this would be most welcome -- Gary
9/6/2010 -- updated to include BTU/sf-hr performance of the metal radiator.