Night Air Thermal Mass Cooling Using Water Barrels

This is a scheme that I am experimenting with to cool my shop.  It basically blows cool night time air over a collection of 50 gallon barrels filled with water to store "coolth" in the water. 

The idea is to cool the barrels down during the night hours, and then use the stored "coolth" in the barrels for space cooling during the daytime.  This requires a climate in which the night temperatures drop down to low levels even though the daytime highs are uncomfortably warm -- this is true for much of the western US.

While I am using the scheme in my shop, it could be used to cool any interior space.  Note that we use a variation on this cooling scheme to cool our house ...  The basic question is: does it make sense to use a fan forced flow of cool night air over thermal mass to store "coolth" for later use in daytime cooling?

I would call this a work in progress, and the results so far are less than stellar, but I thought I would pass them on because 1) someone may think of a way to improve on the scheme, 2) it still seems like a promising idea to me, and 3) it helps to know things that don't work on the path to things that do work.

If you have any suggestions or thoughts on this, please use the Comments section below or email me.

While its been a while since I worked on this project, I plan to do some more testing on this setup this summer (2012), and I'd appreciate any thoughts on it before I start working on it this summer.


The Prototype Setup

This prototype setup is just intended to be a test of whether the concept has merit, so the implementation is not pretty.  If it works, it can cleaned up.

The experiment was conducted in my shop, which is a 24 by 24 ft attached garage with sheet rock walls and concrete slab floor.

The fan cool air inlet:

cooling inlet
Hole for the cool air fan inlet grill.

Inlet cover from Home Depot
cool air grill
Finished cool air inlet.

The pictures above show the hole I cut in the shop wall for the fan inlet.  The fan is mounted on the inside wall of the shop directly inside this opening.  The fan blows outside air from the inlet down the cooling tunnel shown below -- the barrels are inside the cooling tunnel.


cooling mass barrels
The original 3 barrels (one more added later)
The cooling tunnel around the barrels
This is the highly sophisticated cardboard
cooling tunnel that channels the cool
fan air over the barrels.
The fan and inlet grill are at the
right end of the tunnel in this picture and
are hidden by the tunnel.
Looking down the cooling tunnel
Looking into the cooling tunnel from the
outlet end.
The fan is in the wall at the far end
of the tunnel.

The prototype of the cooling tunnel is made from cardboard is shown above.  The idea of the cooling tunnel is that it directs the cool air flow directly over the barrels and makes for a higher velocity flow of cool air over the barrels for better heat transfer.

I guess the tunnel is better than having nothing to direct the airflow over the barrels, but its not a good design in that there are dead air places between the barrels.

Positioning the barrels as the picture shows allows them to still be used in the winter to collector direct solar gain as part of the shop solar heating system.

Note that the cooling air exits the cooling tunnel and goes into the shop where it cools some of the shop thermal mass (walls etc.) before exiting out a window across the shop that is left partly open -- so the same fan provides cooling it two ways.

Granger fan fan Cooling fan
The bigger fan.

The two fans that were tried.  The Dayton fan from Grainger is only 23 watts for 665 cfm at zero pressure drop.  The Dayton fan appeared to provide nearly as much airflow as the other fan with far less power consumption.

The hinges on the fan mount board are part of a system in work that would allow the same fan to be used for blowing cool night air over the barrels at night and then pivoted on the hinges to all the same fan to blow shop air over the barrels during the daytime to extract coolth from them. 

If the 23 watt Dayton fan were able to cool the 4 barrels by 10F over (say) 6 hours of operation, the cooling would be (1670 lbs)(10F)(1 BTU/lb-F) = 16700 BTU for an energy consumption of  (6 horus)(23 watts) = 138 watt-hrs or 470 BTU.  This would be a COP of (16700BTU/470BTU) = 42  (SEER 140).  Right now, it is a not doing this well, but it seems like the potential for very efficient cooling is there. 

The COP is potentially quite a bit better than this when you consider that the air exiting the cooling tunnel blows into the shop and provides additional cooling of the rest of the shop thermal mass (walls, floors, ceiling, and contents).  

 If one has a grid-tied PV system, the electrical energy for this cooling could be provided by adding another about 15 watts of PV capacity to the system at a cost of less than $100 these days -- this results in an infinite COP -- FREE cooling :) 

Do the Numbers Say this Cooling Scheme Might be Worthwhile?

These are some preliminary calculations done before prototyping to see if it looked like it might be worth doing. 

If you want to skip this section, the actual results to date are down the page here...

Do the 4 water barrels add a worthwhile amount of thermal mass?

The shop is 24 by 24 ft by 10 ft high.   Floor area 576 sf.

Sheet rock mass aprox   (55 lb/ft^3)(960sf + 576sf)(0.5 inch/12inch/ft) = 3520 lbs,  water equiv = (3520 )(0.24) = 850 lb water equivalent

Half of Slab to 2 inches deep (576 sf*0.5)(2/12)(150 lb/ft^3) = 7200 lbs,  water equiv = (7200)(0.18) = 1300 lbs water eqivalent
(half of the floor is covered by carpet and slab under carpet is not counted.  The 0.18 is the specific heat of concrete. )

So, the shop wall, floor and ceiling thermal mass is about 1300 + 850 = 2150 lbs water equivalent of thermal mass. 

The 4 barrels of water are (50 gal)(8.33 lb/gal)(4 barrels) = 1670 lbs of water equivalent thermal mass

So, the thermal mass of the 4 barrels is of the same order as the thermal mass of the shop itself -- this seems like a worthwhile addition to me.  Especially when you consider that it might be possible to get the water barrels cooled down closer to the lowest night air temperature because you can flow the night air past them at good velocity for effective heat transfer.

How much cooling might we get?

Night air even on hot days typically gets into the 50'sF at night.  So, if we start blowing the cool night air over the barrels at (say) 10 pm, and stop just before sunrise (which is typically the low for the night), then if the night low temp is 50F, we might hope to get the barrel water down to 55F (?).  

If the barrels start the morning at 55 F, and we circulate air over them during the day to cool the shop with the idea that we would like to keep the shop to about 80F or a bit more, then the useful stored coolth is (75F - 55F)(1670lb)(1BTU/lb-F) = 33,400 BTU of available cooling.  This seems quite worthwhile compared to the estimated cooling load (just below).

(The preliminary results below indicate that getting 20F of cooling overnight is probably not in the cards, but 10F seems very doable and still provides plenty of cooling.)

How much cooling do we need on a hot day?

A very rough calculation of the heat gain to the shop  per degree F of temperature difference from outside to inside (the UA):

Garage door (18 ft*8 ft)/R4 = 36 BTU/hr-F

Outside walls (26 ft * 10ft)/R15 = 17 BTU/hr-F

Window and door (40 sf)/R3 = 13 BTU/hr-F

1 ACH infiltration (5760 ft^3)(0.07 lb/ft^3)(0.24 BTU/lb-F) = 97 BTU/hr-F

So, a total UA of about 163 BTU/hr-F     -- this means that if its 1F hotter outside than inside, the shop will gain 163 BTU over 1 hour.

 We get days once in a while with highs in the 90'sF, and its nice to keep the shop to 80F or less.  Just to get an idea how much cooling is needed to do this, lets say there are  the equivalent of 8 hours with an average outside temperature that is 10F over the desired inside temperature.  This gives a heat gain to the shop of (160 BTU/hr-F)(10F)( 8 hrs) = 12800 BTU.   So, it says the calculated heat loss is well under the coolth that might be stored in the barrels and that there should be enough stored coolth to help keep the shop from getting too hot.

A 2nd estimate of shop heat loss based on how fast the shop thermal mass warms up gives a UA that is more like 200 BTU/hr-F -- pretty close.

Is the heat transfer from cool air to barrel water good enough for the scheme to work?

This may be the key question -- is the heat transfer coefficient between the moving cool air and the barrel water sufficient with the available barrel surface are to get a useful amount of heat extracted from the barrels?

This Typical Overall Heat Transfer Coefficients in Heat Exchangers  gives heat transfer coefficients for various heat transfer arrangements.  Of interest:

Tubes - Liquid inside and gas outside      3 to 15 BTU/ft^2-F-hr

Air-cooled heat exchangers - Cooling of water      100 to 130 BTU/ft^2-F-hr

The Engineering Toolbox - Typical Overall Heat-Transfer Coefficients gives somewhat lower coefficients:

Liquid free convection - Forced Convection Gas   2 to 10 BTU/ft^2-F-hr

So, this is not too helpful in that the range of possible coefficients is so large.

If one uses (say) 20 BTU/ft^2-F as the heat transfer coefficient, and a barrel surface area of 68 sf for all four barrels, and the temperature difference between the barrel and the air is 10F, then the heat transfer is:  (68 ft^2)(10F)(20 BTU/ft^2-F-hr) = 13600 BTU/hr.   This would appear to be plenty, but its unclear to me if we are getting anywhere near the 20 BTU/ft^2-F-hr heat transfer coefficient.

 Results So Far

I only had time to work in a couple of nights testing, so, the results are incomplete, but still interesting.

The plot above is for the night of 8/12

Air tunnel inlet temp -- green line
Air tunnel outlet temp -- red line
Water barrel temp -- purple line
Shop temp  -- black line

In this case, the fan moving air over the barrels starts just after midnite and runs until about 7:15 am -- it was supposed to start earlier, but the timer did not work.

The barrels cool down by about 2.7F over this time -- about 0.4 F per hour.  The barrel cooling would have been greater if 1) the fan had started at the right time, and 2) the barrel water temp was higher at the start of the period (as it would normally be if the barrels had been used for cooling the day before), and 3) the heat transfer from air stream to barrels was further optimized as discussed elsewhere.

But, even with the relatively small 2.7F, the cooling is (50 gal)(4 barrels)(8.33 lb/gal)(2.7F)(1 BTU/lb-F) = 4500 BTU. 

 It think that a reasonable target for an optimized system would be more like a 10F barrel cooling, which would be worth about 17,000 BTU -- this would exceed my normal daytime cooling requirement.

It is clear that more effective heat transfer from the air stream to the barrels would be highly desirable -- any ideas?

The air temperature rise from tunnel inlet to outlet is a pretty impressive (to me) 10F.

Note that the air exiting the tunnel is well below the shop temperature, and provides additional cooling to the shop.  If the flow rate is about 250 cfm (see below), then the additional cooling is of the order of (250 ft^3/min)(0.07 lb/ft^3)(60 min/hr)(70F - 64F)(0.24 BTU/lb-F) = 1500 BTU/hr, or about 10,000 BTU over the 7 hours.  Hopefully most of this goes into cooling the shop thermal mass down.

Flow Rate Through the Tunnel

I did a rough velocity survey of the airflow around the barrels by measuring the air velocity with a Kestrel anemometer at three points in space between the barrels and the tunnel walls on the north and on the south sides with these results.  The north and south channels are each 42 inches high --there is no airflow above or below the barrels.

South Air Channel Width
South Channel Velocity
North Channel Width
North Channel Velocity
8 70 2.5 160
8 70 2.5 120
8 70 2.5 90

If you combine these to estimate the flow rate, it comes out to a flow area of 3 sf with a flow rate of 253 cfm. 

This is less than half the fan name plate flow rate.  The lower flow than the fan name plate is a combination of lower flow rate due to pressure drop through the tunnel, and to the leakage out of the the tunnel (which is pretty leaky at the entrance and along its edges).

The tunnel design is not to satisfying in that it exposes less than half of the barrel surface area to the high velocity tunnel flow -- any idea how to do a better tunnel?

One advantage of this kind of system is that one could hold the coolth stored in the barrels for use later in the day.  That is, let the thermal mass of the living space itself (which has been cooled down by the exhaust air from the barrel tunnel) take of the cooling in the first part of the day, then as the space gets warmer than desired, use forced circulation over the barrels to provide more cooling later in the day.

I did log performance for a couple other nights, but the results are about the same as for the plot above. 

Improvement Possibilities?

So, its not working as well as hoped in the current configuration -- but it does look promising (to me) with some further optimization:


Other Related Schemes

Irrigation Water Scheme

Early in the day, before cooling is needed, fill the barrels with cold water from the well -- so, the barrels start the day full of 55F water.  During the day, air flowing past the barrels cools the room -- this could be natural convection or fan forced as required to extract the cooling.   At some point after the heat of the day, use the water in the barrels for lawn or garden watering.  

If the barrels start the day at 55F, and end the day at 75F, then the cooling available for this per barrel would be (50 gal)(8.3 lb/gal)(75F - 55F) (1BTU/lb-F) = 8300 BTU per barrel  -- a useful amount.   Depending on your irrigation and cooling needs, more than one barrel could be used.

While we are trying various strategies to decrease our irrigation water use, it is still more than enough during the hottest time of the summer to provide water for this cooling scheme. 


Same Water Mass for Heating and Cooling

In the south wall of the house, include a setup similar to what I have for my south shop wall.  That is, an outer layer of glazing (perhaps twinwall or triple wall polycarbonate), a movable layer of insulation (the rollup garage door in my case, and then a "wall" of thermal mass water containers. 

In the winter, on sunny days, the movable insulation is move out of the way to allow sun to shine on the water containers.  This provides some light, some heat direct to the room, and some heat stored in the water containers.  At night, the moveable insulation is moved to insulate the room from the outside, and the water containers provide heat to the room via natural convection.

In the summer, use the scheme outlined above to circulate cool night air over the water containers to store coolth for the following hot day.  If the south wall had a suitable overhang to block direct solar radiation from the high summer sun, the movable insulation could be moved partly or completely aside when solar daylighting is desired during the daytime.

The same thermal mass that store heat in the winter stores coolth in the summer. 



This just shows the type of climate that this cooling scheme might work well in.

The graph below from Weather Spark  shows the summer temperatures and clearly shows the pattern of large drops in night time temperature -- this allows cooling methods that can make use of the cold night air to work even though daily highs can be uncomfortably warm.  The advantage of this is that the only electrical power required for this method is to run a fan, using very little power.  This temperature pattern of hot days and cold nights is common in many parts of the US west.





blog comments powered by Disqus