Full Scale Test of Evapro/Radiation Cooling System

These are the results from the first full scale test of the Evapro/Radiation cooling system that I experimented a bit with last year.

The full scale version uses the North roof surface of my Solar Shed as the cooling radiator to cool a 430 gallon tank of water during the night time.   When cooling is needed during the following day, the cooled water in the tank is pumped through the radiant floor heating loops in the house for cooling.

Solar Shed space heater, now used for cooling as well.

In my case, the system involves very little extra cost or work because nearly all the components are already there for the existing solar space heating system.


This is really just a first try at a practical system.  The results are encouraging, and I plan to do further work on it when our 10 month Montana winter ends :)


The Setup

The system uses our existing Solar Shed solar space heating system, which is described in detail here...

In a nutshell, the current heating system consists of 240 sqft of solar heating collector mounted on the south face of the Solar Shed, which heats a 430 gallon hot water storage tank.  The solar heated water is pumped over to the house via an insulated underground pipe and circulated through PEX loops in the floor for radiant floor heating.  This system has been operating for about 5 years and provides a significant part of our winter heating needs.


To use this same system for cooling, the north side of the Solar Shed roof is set up as a evaporative/radiation cooling surface.  Water from the 430 gallon storage tank is pumped to the ridge of the Solar Shed roof, and allowed to trickle down the roof surface.  The water trickling down the roof is cooled by a combination of radiation to the cold night sky and evaporation of the water.  The evaporative/radiation cooling idea is explained in more detail here...


The only new components added to use the same system for cooling are:

- Added a tube along the ridge of the Solar Shed to distribute water over the shed roof.


- Added a gutter and some plumbing to collect the water at the lower edge of the roof and route it back to the tank.


- A pump to pump the water to the roof top (potentially the same pump as the solar heating collectors use)


- A timer or more sophisticated control to turn the system on in the middle of the night for a set time.

The rest of the system including storage tank, plumbing to house, floor heating(cooling) loops, and even the house heating controls were used unchanged from the solar heating system.


The pictures below show the first full scale test setup.  Given that I expected to have to make changes and refinements, I did not take a whole lot of care to make it "pretty" -- but, its a very simple system, and a few hours more work would take care of the rough edges.


Spray bar along ridge line.
Slight slope allows for even spray.

System in action spraying water over roof.
Of course, this would normally be at night.


There was just not much hot weather this summer, so not a lot of incentive to get this full scale test hooked up and running.  In the end, I did not get to installing the new stuff until late August, and the testing was just one day when the temperature got up to 96F (our only really hot day of the summer).  I plan to do more with this next year, but thought I would pass on the data for the one day we did use the system.


Part 1: Cooling the Tank Water

Using this system is a two part process: part 1 is cooling the water in the "coolth" storage tank, and part 2 is transferring the stored coolth to the house interior when needed.


The plot just below shows the only night of cooling the tank with the new system results.




Purple line is tank temperature (F)

Greenish line is ambient temperature (F)

Dark blue short dash line is water returning from roof temperature (F)

Red line (lowest) is the roof surface temperature in an area with water flow over it (F)

The timer turns the pump on at about 4:10 am, and it runs until about 6:20 am. 


In this case, the tank was cooled from about 57.5F down to about 53.5 F.    This is a pretty small temperature drop, but bear in mind:

- If the system were in regular use, the starting tank temp would have been in the 60's Frather than the 57.5F.  This would require a little more run time, but there is plenty of time available during the night.


- It looks like the system could have achieved some further cooling had it not been turned off at 6:20 am.

The roof radiator was able to cool the water about 2F below the ambient night temperature, and could probably have achieved greater cooling 1) if it had been run longer, and 2) with some further optimization of the radiator.  This is supported by the earlier tests in which a lower tank temperature is achieved.


The cooling rate achieved is:

Radiator area = 240 sf


Cooling achieved = (430 gal)(8.33 lb/gal) (57.5F - 53.5F) = 14300 BTU


The cooling rate per sqft of collector, per hour of operation was:

BTU/sqft-hr  = (430 gal)(8.33 lb/gal)(1.9 hr)(1 BTU/lb-F) / (240 Sqft) = 27.6 BTU pers hour per sqft of collector.


Where 1.9 F is the average rate of cooling of the tank water per hour of operation from the plot.


This compares with 33.8 BTU/sf-hr achieved by the 14 sf aluminum radiator in the previous test...
So, to a first cut, it looks like the asphalt shingle roof performs about 82% of the aluminum sheet radiator.

Some additional runs would be needed to confirm this.

One practical design observation for this system is that rain water that falls on the shed roof ends up in the storage tank.  This is a good news/bad news thing.  The good news is that the collected water would help offset the evaporation of water as it flows over the roof.  The bad news is that you could easily overflow the tank in a good rainstorm.  So, for a practical design, the tank has to have an overflow outlet that will run excess water off into the weeds, or you will end up with a mess. 


I suppose that an enterprising person could use this scheme to combine a large rainwater catchment system with a house cooling system.


The only control I used to turn the water on to the roof spray bar was a simple 24 hour timer set to come on for a couple hours before dawn.   For around here that might work fine with just an manual override to keep the system from turning on if the tank is already cold enough.  Fancier controls that run the system long enough to cool to a lower set point might also be used.


Part 2: Cooling the house

The 2nd part of the system is making use of the coolth stored in the tank to meet the cooling needs of the house during the day.


We were only able to test this on the 96 F day mentioned above -- these are the results from this one day test.  Since we ventilated the night before and morning of the predicted 96F day, no cooling was needed until early afternoon.



  Radiant floor loops Main floor     Outside  
Time Floor In(F) Floor Out(F) Tfloor1 Tfloor2 Tfloor3 TroomAir Tambient  
1:49 PM 51 F 61 F 76 F 77 F 74 F 76 F 90 F System turned on
2:10 PM 52 60 74.6 76 74.1 74 92  
2:30 PM 52 60 73 73.4 73 75.5 92  
4:50 PM 56 60 74.6 73.7 76.6 76 93  
6:00 PM 58 62 74.6 74.4 77 77 91  
8:00 PM 60 63 77.1 75 79.1 76 79 system turned off



- Floor In is the temperature of the water from the tank as it enters the radiant floor loops.


- Floor Out is the temperature of the water leaving the radiant floor loops


- Tfloor1 through Tfloor3 are the temps of 3 places on the main floor where floor loops are installed -- measured with an IR temperature gun.


- TroomAir  is the room temperature of the main floor living area measured 4 ft above the floor.


- Tambient is the outside ambient temperature.

I did not record relative humidity, but this is a dry area, and typically has low humidity.


So, during the time the system was on, the tank water was warmed from 51F up to 60F.  This corresponds to:

Coolth Delivered = (430gal)(8.33 lb/gal)(60F - 51F)(1 BTU/lb-F) = 32,237 BTU  of cooling.


This is roughly equivalent to a half ton AC running 6 hours.

If you look at the room air temps, the system was able to keep comfortable temps in the house throughout the afternoon and evening until the outside air cooled enough to not need any more cooling.  Maximum room temp at 6pm was 77F.


The floor temps (in areas cooled by the radiant floor loops) were in the 73F up to 77 F area



- As the tank water warms, it is delivering less coolth to the  room.  When the tank is at 51F, the temp rise over the floor loops is 10F, but when the tank is up to 60F, the temp rise is reduced to 3F, so about 1/3rd the cooling by the end of the day.

I guess that a couple message are that 1) getting the tank temp as low as possible is good, and 2) a larger tank may be necessary for higher cooling loads than we have.

On the other hand, this was easily the hottest day of the year here, and the system met our full cooling need.


- The system does not lower humidity at all, so areas with both high temps and high humidity would need some additional means to reduce humidity.


Another interesting thing was that the same controls that control the radiant floor heating in the winter for the solar heating system also work with the cooling system.    There are two thermostats that are hooked in series -- one that senses storage tank temperature and the other room temperature.  In the winter they the tank thermostat is in "cooling mode", and turns on when the tank is hot enough to provide heat.  The room thermostat is in "heating mode", and turns on when the room is cool enough to need heat. 

For the cooling system, the tank thermostat is switched to "heating mode", and the room thermostat is switched to "cooling mode", and it works with no further changes except to reset the temperature set points -- almost like it was planned that way.



A very rough go at energy efficiency of this cooling system:

I use a Taco 008 pump to deliver flow to the winter solar heating collectors.   This pump uses 90 watts.


Since it would be desirable to use the same pump summer and winter, and its about the right size for the summer circulation to the roof radiator, I will use its power consumption for the power use during cooling.   With this arrangement, resetting a couple valves to change the plumbing around would be all that is involved in a winter to summer switch over.


Lets target cooling the 430 gallons (3580 lbs) from about 61F down to 49F for a total cooling of (12F)(3580 lb)(1 BTU/lb-F) = 43,000 BTU.
43K BTU is about equivalent to a half ton AC run for  7 hours -- fine for my house even on very hot days.


Based on the test above that achieved 28 BTU per sqft of radiator per hour of cooling, and using the same 240 sf of north roof radiator area, the run time to cool the tank the full 12 F would be:

43,000 BTU = (28 BTU/sf-hr)(240 sf) (Trun hrs)


Trun = (43000 BTU)/((240sf)(28 BTU/sf-hr) ) = 6.4 hours of run time for 43K of cooling.

Pump energy consumed is then (90 watts)(6.4 hr) = 576 watt-hr, or 1965 BTU


COP = (energy out) / (energy in) = (43,000 BTU) / (1,965 BTU) = 21.9   (or a rough SEER of 75)


This does not include the energy required to distribute the heat to the house on the following day -- this might add another 300 or so watt-hours and cut the COP correspondingly.  I'm not sure exactly how the COP/SEER is figured on conventional AC units -- I do know that for heat pumps and furnaces, the energy for distribution of heat/coolth is not normally included in the calculation.


So, this seems like pretty impressive efficiency?  Or, am I missing something?


Bottom Line, and Next Summer Tests

The water tank nighttime cooling rates achieved by the 240 sqft asphalt roof cooling radiator are somewhat less than the earlier test of the small metal radiator, but still probably acceptable.   For next summer I plan to:

- Do some more tests of the small metal radiator to confirm its higher efficiency, and get a better idea how it performs under different conditions.
And, determine how much below ambient it is capable of cooling.


- Do some more tests on the 240 sf asphalt shingle radiator on the north solar shed roof.  Determine how much additional cooling longer run times will achieve, and how fart below ambient it can cool compared to the metal collector.


- Decide on whether the metal radiator performance is enough better to be worth doing a larger metal radiator on the Solar Shed roof (probably just the lower part of the roof).


- Hook up the plumbing such that the same pump that is used for winter circulation of water to the collectors can be used as summer circulation to the radiator.

Any ideas would be appreciated.

Reports on any similar systems would also be appreciated.


The house cooling part of the test that used the existing radiant floor system to distribute coolth to the house seemed to work fine, and I'll probably just use it in the same way as we did for the first test.


I may also just have a go at using well water for cooling with the water first used for cooling, and then irrigation.



For those interested in more reading on innovative cooling systems by Steve Baer and the FSEC:

Double Play -- An Experimental Solar Heating and Cooling System
Steve Baer

How to get articles from Home Power ...

and the Zomeworks site


Roof Integrated Solar Absorbers: The Measured
Performance of “Invisible” Solar Collectors
Preprint (700K pdf)

Roof Integrated Solar Absorbers -- slides

Roof Integrated Solar Absorbers
Home Power magazine article, issue 102

This innovative system uses unglazed collectors that are integrated with the roof structure and a unique thermal storage system to provide solar thermal heating in the winter and radiation cooling in the summer. Amazing!

The pdf report provides detailed performance measurements from the FSEC & NREL.

Theoretical Evaluation of the NightCool Nocturnal Radiation Cooling Concept
Danny S. Parker
Florida Solar Energy Center (FSEC)

New report on the NightCool performance in test buildings:
Experimental Evaluation of the NightCool Nocturnal Radiation Cooling Concept: Performance Assessment in Scale Test Buildings,
Danny S. Parker, John R. Sherwin
Florida Solar Energy Center (FSEC)

Daily performance of the test buildings,


This is a simple, building integrated, cooling scheme that uses nighttime  radiation cooling from a metal roof to cool air in the attic space.  Attic air is then circulated into the living area to provide cooling.
A simulation model is used to predict performance in various climates -- the scheme works very well in dry areas, and fairly well even in very difficult moist-warm climates (e.g. Florida).
One of the attractive features of this system is that the roof is very conventional -- no massive ponds or moving insulation.
There is the potential in some climates to use the same metal roofing/attic system for space heating in colder weather.
The new test report shows the results for two 10' by 16' structures that were built to compare NightCool performance to conventional AC in Florida.  Very interesting and promising results.
Tests of a a simple dehumidification system using  desiccants are now underway.




Gary September 6, 2010