Single pump drainback scheme for modest pump power and modest pump head in a tall house

-- with the heat storage tank in the basement.

David M. Delaney
Started March 11, 2010
Latest update, March 18, 2010

Keywords: solar water heating, solar hot water, solar water heater, freeze protection, drain back, drainback, drainback tank,  drain back tank ,  pump power, pump head, tall house, basement heat store, basement tank, heat exchanger

In the urban core of Canadian cities, most of the good locations for solar thermal collectors are on rooftops.  Retro-fitting solar thermal and space heating means that collectors must often be on the roofs of tall houses or small apartment buildings, and that the large water tanks required for heat storage must be in the basements.  Drainback syatems are the simplest four season solar water heater systems. Although they use plain water in their collector loops they provide reliable freeze protection if the water in the collector drains away into a drainback tank quickly enough when the circulating pump stops. 

The  conceptual drainback scheme presented here is unconventional. It is intended for tall houses having solar collectors on their roofs and a heat store in the basement.  The heat storage meduim is water.  The design makes it possible to avoid boiling in the collectors, even in very tall syatems.  Its requirements for pump power and pump head are modest, making it feasible to power it with an inexpensive PV system.   It is intended to work well and quietly in systems that are taller than usual for drainback systems..

WARNING. I have not built a system like the one described here, nor have I tested its principles experimentally. It is unconventional in the placement of its drainback tank in the riser rather than in the descender.  It has an unconventional air passage between the descender and the drainback tank.  The passage must pass air in both directions in order to drain and fill the collector.  If the air passage does not move air into the collector quickly enough to prevent freezing when the pump stops, the system will be unworkable.

System Diagram

Drainback  scheme to reduce pump power and pump head

The working fluid circuit is a closed loop (the collector loop) comprising, proceeding anti clockwise, the collector, the descender, the heat exchanger, the riser, the header / drainback tank, and the pump. The water and air within the collector loop are at atmospheric pressure approximately near the middle of the system, but will pressurize and depressurize slightly as the system  heats and cools.

Once the pump has filled the collector and the top of the descender, the work required from it should be no more than that required to overcome frictional losses around the loop, because there is no further net lifting.
?????????????????????????????/  Fix this.  May always have to lift from top of DBT.   MAy not be able to ever guarantee that the upper part of  the descender will fill completely with water. This would mean that the pump will always have to lift from the top of the DBT to the top of the collector ------------------------------  Even a pump of modest power should be able to deliver adequate flow. The pump head required at startup is no more than needed to lift water from the drainback tank to the top of the collector.   The position of the air insertion / removal point in the descender pipe should prevent the pump having to drive a slug of air around the circuit at startup, which would expose it to the necessity of providing sufficient initial head to lift water all the way from the storage tank below.

Details of Operation

State 1.  The pump has been turned off for some time.  The collector is cold and filled with air. The drainback tank is almost full of water. The level  of the water in the descender is at the height of the surface of the water in the header / drainback tank.   The surface of the water level in the descender is several inches below the lowest part of the air slug diverter. 

State 2.  The collector has become hot, causing the control system to turn on the pump.  The pump is pushing water up the upper part of the riser into the collector.  The water level has not yet reached the top of the collector.  The level of the water in the drainback tank is dropping. The level of the water in the descender is also dropping.  The path through the piping anti clockwise from upper interior of  the collector that is not yet full of water, over the top of the collector, down the upper part of the descender, through the air slug diverter, into the top of the header / drainback tank, is free of water. Air is forced anti clockwise around this path from the collector into the header / drainback tank as the water level rises in the collector.  

State 3.  The collector is full to the very top and water has started to fall over the top and down into the upper part of the descender.   The falling water,  pushed over the top and down by water moving upward in the collector under the impulsion of the pump, pushes a slug of air down the descender.    Some water may slide down around this air slug, and temporarily  allow the level of the  water in the descender to block the air slug diverter, but whenever the level of water in the lower part of the descender is high enough to block the air slug diverter, the column of water in the lower part of the descender must be moving downward, since it outweighs  the column  of water in the riser up to the surface of  the water in the header / drainback tank, that surface being lower than the lowest point of the air slug diverter, and since the pressure of air in the header / drainback tank cannot build up because of the emptying action of the pump.  No sizable slug of air can be propelled  below the air slug diverter into the lower part of the descender, because it will clear  the air slug diverter of water and allow rapid bleeding of the air of the air slug across to the header / drainback tank.  Eventually, the descender will be clear of sizable air bubbles.

State 4.  The pump has been running for some time. The collector is hot and is heating the water passing through it. All of the air in the system is resident in the upper part of the header / drainback tank.  The pump no longer has to expend energy to lift water, but merely to overcome forces of friction to keep water moving around the system. A very small amount of water will be dribbling into the header / drainback tank from the constricted air channel across to the descender.  The rate of entry of water from the descender to the air channel is not sufficient to fill the air channel as it empty into the drainback tank.  The bulk of the water moves anti clockwise in a closed cycle around the system, being heated in the collector and being cooled in the heat exchanger while heating the water in the heat store.

State 5. The collector has become cold enough to cause the control system to stop the centrifugal pump. The pump has just stopped. A centrifugal pump is quite open to reverse flow when it is not running, so the  pressure in the drainback tank rises immediately as the weight of the column of water in the upper riser is no longer supported by the action of the pump.  This rising pressure forces air through the air channel toward the descender.    The channel has a large diameter (except for its constriction), and is sloped gently upward toward the descender. The constriction is sufficiently small so that the flow rate of water entering from the descender is never enough to fill the channel. Water stops entering the channel immediately because the pressure at the descender end of the air channel is lower than the pressure at the tank end. The pressure at the tank end is the pressure at the surface of the water level in the tank, which is several inches lower than the descender end, and is therefore at a higher pressure. 
Air flows immediately over the top of the remaining water in the channel as that remaining water continues to drain toward the header / drainback tank. Air   forces it way through the constriction into the descender, and rises to the top of the collector, allowing the water in the collector to fall. The design of the air slug diverter keeps the constriction free of water on both sides to allow a free flow of air through the constriction. The collector and upper portions of the riser and descender empty quickly of water and are refilled with air.

State 6.   Same as state 1.

The Air Slug Diverter

The air slug diverter allows an air slug to cause water to drain away from the constriction

Outstanding issues:

(1) The water left standing in the descender pipe when the pump stops first cools off, then is later injected into the heat exchanger in the storage tank when the pump starts up again.  

It is desirable to keep the diameter of the lower part of the descender pipe narrow to minimize the amount of cool water injected at pump startup, (this was also the reason for  the unconventional location of the drainback tank in the riser.)  There is no thermal penalty for a larger pipe diameter in  parts of the collector loop other than the lower part of the descender, the part of the descender below the drainback tank and above the storage tank. In particular, the portion of the descender above the air slug diverter can have a 3/4 inch diameter or larger to facilitate drainback by allowing air and water to pass freely in opposite directions during drainback.

In opposition to the need to keep the thermal mass of the water standing in the lower portion of the descender small by keeping the diameter of the lower part of the descender small, it is desirable to keep the diameter of all pipes of the collector loop large enough to keep the loop friction small so that a modest pump head will drive the collector loop water all around the circuit -- total length approximately 400 ft, assuming 300  ft of pipe in the heat exchanger.  Half inch (12.7 mm) diameter pipe for the lower part of the descender should be narrow enough to keep the volume of the standing cool water in the descender  "small". (A 1/2 inch descender would inject about 150 Btu of coolth into the storage tank when the pump starts after having been off long enough for the descender pipe to cool to the interior temperature of the house, assuming a 130F tank temperature and a 70F house temperature. It would be worthwhile to insulate the descender well to reduce the thermal penalty of short off times.) 

Working on the descender pipe size issue issue:

There might be flow rate problems with a 1/2" descender depending on the size of the collector.   Here's a table of the flow rate required for different collection powers in gallons per minute.. The flow rate FR is in gallons per minute, dT is the temperature rise through the solar collector  in Fahrenheit, power P in Btu/hr,  1 gallon = 0.134 ft^3, 1 gallon water  = 0.134 ft^3 * 62.4 lbm/ft^3 =  8.36 lbm water, 1 gallon/minute=1 gpm = 8.36 lbm/min

P <Btu/hr> = FR <gal/min> * 8.36 lbm/gal * 60 min/hr  * dT < F> * 1 Btu/lbm.F

FR = P / (8.36 * 60 * dT) = P/(501.6 * dT)

The following table has the form  dT  = >  (P1, FR1), (P2, FR2), (P3, FR3) ...

10 F => ( 1000 Btu/hr, 0.199 gpm), (5000 Btu/hr, 0.999 gpm),(10,000 Btu/hr,1.99 gpm),(20,000 Btu/hr,3.99 gpm),(40,000 Btu/hr,7.98 gpm)

20F => (1000 Btu/hr,0.0997 gpm), (5000 Btu/hr,0.498 gpm),(10,000 Btu/hr,0.997 gpm),(20,000 Btu/hr,1.99 gpm),(40,000 Btu/hr,3.98 gpm)

40F => (1000 Btu/ hr,0.0498 gpm),(5000 Btu/hr,0.249 gpm),(10,000 Btu/hr,0.498 gpm),(20,000 Btu/hr,0.996 gpm),(40,000 Btu/hr,1.99 gpm)

60F => (1000 Btu/hr,0.0332 gpm),(5000 Btu/hr,0.166 gpm),(10,000 Btu/hr,0.332 gpm),(20,000 Btu/hr,0.664 gpm),(40,000 Btu/hr,1.328 gpm)

80F => (1000 Btu/hr,0.0249 gpm),(5000 Btu/hr,0.125 gpm),(10,000 Btu/hr,0.250 gpm),(20,000 Btu/hr,0.50 gpm),(40,000 Btu/hr,1.00 gpm)

Note:  1 Btu/hr = 0.293 watt,    1000 Btu/hr = 293 watt,    40,000 Btu/hr =  11,711 watt

2) Will the small available head (~ 1 to 2  ft water, decreasing progressively to zero) drive air to the left  through the air slug diverter fast enough to fill the collector with air promptly?  The head available to drive the air flow to the left into the collector when the loop is emptying is provided by the difference in elevation of the air slug diverter and the surface of the water in the DBT.  This difference of elvation is limited to a foot or two (?) by the need to ensure that when the loop is full and the pump is running, the pressure in the descender at the level of the air slug diverter is greater than the pressure in the drainback tank to ensure there is no air flow into the descender from the drainback tank.

Friday March 17, 2010, To be continued....