Recuperate some of the heat from a group of showers

[John]

I am working on a project whose objective is to recuperate some of the heat from a group of showers .

The design is very simple if there is a space of a couple of feet below de shower floor . I believe we can get back about 50% of the heat normally going down the drain. But , there is no space below the shower's floor.

Installing an electric pump defeats the purpose .

What about installing a pump coupled to a turbine propelled by the water used by the showers to raise the gray water by about 2 or 3 feet in order to feed the unit that, by gravity recuperates some of the heat? I know that will reduce the water pressure but will it work and be acceptable by the users?

The showers can have an escape drain that will bypass the recovery system if the gray water raises above a certain level, an electric pump that will operate and bring the water level to zero if the system fails or a mechanical system that revert everything to a traditional configuration.

Comments please

John

[Himpg]

The simplest solution seems to be installing a sump in the gray water system that is about 0.5 meters deeper than than the existing waste lines. DO NOT forget to instal a sewer gas trap in the original exit line so that it can continue to provide system fail safe. Then drop in in a standard float activated sump pump to deliver gay water to your heat exchange(s). This configuration should be easy to service and uses standard hardware.

[nicksanspam]

Maybe a 55 gallon plastic drum or a steel drum with a plastic liner and a lid with a locking ring, with a way to backwash it to remove crud once in a while, eg a garden hose connection.

Can we recover 90% of the heat from one shower?

We could estimate efficiency with section 7.3 of the SRCC OG300 standard, which specifies measuring solar water heater performance with 6 hourly

3 gpm 10.7 gallon hot water draws totaling 64.3 gallons per day.

If a well-enclosed shower drain is 105 F and the cold supply is 50 F, heating 64.3 gallons (536 pounds) takes 536(105-50) = 29.5K Btu.

A counterflow exchanger with equal heat capacity flow rates Cmin = Cmax = 536/6h = 89 Btu/h-F and effectiveness E = 0.9 = NTU/(NTU+1) needs NTU = 9 = AU/Cmin. With wall conductivity U = 30 Btu/h-F-ft^2, heat exchange area A = 27 ft^2, eg 100' of 1" black PE plastic water pipe inside 100' of 1.5" pipe, coiled around the the drum, surrounded by insulation.

We could pump fresh and graywater at 64.3/6/60 = 0.18 gpm with a dual metering pump and store the preheated cold water in the lower half of an electric water heater, with the lower heating element disconnected.

Nick

[nicksanspam]

Better yet, use a float switch and a single metering pump like Grainger's $204 2P305 to move 64.3 gpd at C = 22 Btu/h-F with NTU = 30x27/22 = 36 and E = 0.97. Preheated water might thermosyphon into the lower part of a water heater from the fresh water heat exchanger outlet 3' above the floor into a T 3' above the floor, connecting the upper and lower parts of the tank, with a common cold water connection at floor level.

Nick

[marson]

This existing technology would be much simpler:

[nicksanspam]

About the same cost, with 60 vs 97% efficiency. With no graywater storage, it won't work with a bathtub. With no pump, it needs a few feet of drop between the shower drain and the sewer connection.

Nick

[nicksanspam]

We could use a float switch and a metering pump like the $185 Mec-O-Matic D75 to move 64.3 gpd at C = 22 Btu/h-F (0.044 gpm) with NTU = 30x27/22 = 36 and E = 0.97. Preheated water might thermosyphon into the lower part of a water heater from the fresh water heat exchanger outlet 3' above the floor into a T 3' above the floor, connecting the upper and lower parts of the tank, with a common cold water connection at floor level, with the lower tank heating element disconnected.

This looks promising, with Poiseuille's equation for thermosyphoning:

20 PI=4*ATN(1) 30 TC=(50-32)/1.8'cold temp (C) 40 TW=(75-32)/1.8'warm temp (C) 50 RSLOPE=(998.2-988)/(20-50)'density slope 60 RHOC=998.2+(TC-20)*RSLOPE'cold density (kg/m^3) 70 RHOW=998.2+(TW-20)*RSLOPE'warm density (kg/m^3) 80 H=3*.3048'height diff (meters) 90 DP=(RHOC-RHOW)*H*9.8'pressure difference (N/m^2) 100 VSLOPE=(.001002-.000547)/(20-50)'viscosity slope 110 VC=.001002+(TC-20)*VSLOPE'cold viscosity (Ns/m^2) 120 VW=.001002+(TW-20)*VSLOPE'warm viscosity (Ns/m^2) 130 RC=24/2/12*.3048'cold pipe radius (meters) 140 RW=1.049/2/12*.3048'warm pipe radius (meters) 150 LC=3*.3048'cold pipe length (meters) 160 LW=100*.3048'warm pipe length (meters) 180 NE=6'number of elbows 190 LC=LC+50*RC'equivalent cold pipe length (meters) 200 RESC=8*VC*LC/(PI*RC^4)'cold flow resistance 210 RESW=8*VW*LW/(PI*RW^4)'warm flow resistance 220 V=DP/(RESC+RESW)'flow (m^3/s) 230 GPM=V*15850!'flow (gpm) 240 PRINT GPM

.2886552 gpm

Nick

[daestrom]

Well, that's 60% measured, actual performance versus 97% theoretical, never-been-done performance.

True

Also true. But then, how much would the pump use moving 64.3 gpd ?

daestrom

[nicksanspam]

The $185 Mec-O-Matic (love that name :-) D75 uses about 120Vx0.72A = 86 watts moving 75 gpd, so it might use 74 moving 64.3. A different kind of pump could use less power, but I think this one can run dry, and positive displacement seems good for avoiding crud complications (it comes with a little strainer) and getting the cooled graywater up into a high sewer connection.

It has a 10 minute cycle time with a variable on-time controlled by a knob for 1%-100% of 75 gpd. A simple approach would be to run it continuously, or maybe use some sort of float switch, or a 70 F thermostat on the side of the drum, or a scheme to match the pump flow rate to the actual daily flow.

Nick

[nicksanspam]

Oops. It uses about 86 W moving 97 gpd, so it might use 57 moving 64.3.

The factory folk say it can run dry and the piece of tubing can last

10 years or more, if not attacked by a few specific chemicals.

The duty-cycle knob controls a pot...

Nick

[daestrom]

Are you saying it uses 86 W * 24 hours =>2064 watt-hours to move those 96 gpd? So with a float switch it would run only 64.3/97*24= ~16 hours a day and use just 1368 watt-hours a day?

Seems a bit much (about $0.20 worth of electricity a day).

daestrom

[Mark]

But what is the temperature of the greywater when it leaves the building without a heat exchanger? Even with no fancy heat exchanger, much of the heat of the shower drain water is given off by the drain pipes in to the building environment. Also much of the inital heat is given up in the shower itself heating the walls and evaporating the water into vapor. I'm going to guess that the economics of this shows that there is no payback. Mark

[nicksanspam]

2. >So with a float switch it would run only 64.3/97*24= ~16 hours a day

Yes, but maybe it's better to turn down the flow rate and run it "continuously" (with a 10 min cycle time and a 6 min on-time) which would still use 1368 Wh/day but make the heat exchanger more efficient.

Yes... 29.5K Btu of electric water heating is 8.6 kWh/day.

Maybe we should find a more energy-efficient pump, or centrifugally pump the graywater up to a bucket above the sewer outlet and let it drip back down through the outer tubing, then back up the outer tubing to the outlet.

Nick

[nicksanspam]

Not in most houses, I ween. But in some, the house heats the cold water pipe significantly before it enters the water heater...

Nick

[nicksanspam]

Or let the shower drain fill an upper drum, with a disk filter in the bottom, and slowly drip greywater into tubing surrounding the lower tank and let it flow up and into the sewer connection with no pumps, or put a bilge pump and a float switch in the lower drum, if the sewer outlet is more than 3' above the floor. A garden hose connection to backflush the outer tubing and upper drum would be useful.

Attwood's V625 ($19.99 from BoatUS) moves 650 gph with 12VDC at 1A, ie

63 gallons with 1.2 Wh. It might be PV powered. Boat US item 168228 is a $20.99 float switch.

Nick

[nicksanspam]

Fiberglass window screen squeezed in a slot in a tight-fitting 1" PE pipe ring? It might snap into the lower expanded "rolling hoop" of a steel drum with a removable lid and a plastic film liner.

If we adjust the drip with a valve so 100 gpd flows with 3' of head,

60 might make (60/100)^2x3 = 1.08'...

Automatically, with a $145 Taco 561 3/4" 3-way heat motor zone valve (or a solenoid valve scrounged from a dead washing machine and a $3 hose bibb vacuum breaker) that closes the tubing outlet and opens a pressurized water connection behind it for 10 minutes when the upper drum is full too often, eg when an upper-drum float switch has been open more than 90% of the time? Every 1-3 months?

To detect a leak in the inner tube (which has a 50-year guarantee), we might close the drip valve once a year and check for greywater outflow. And drain the outer tube automatically with a hose bibb vacuum breaker if water pressure ever fails. In light of the 2006 ICC plumbing code, we might prove that it's extremely unlikely that someone would drink more than 1 quart of essentially toxic bathwater...

Nick

[nicksanspam]

Automatically, with... a solenoid valve from a dead washing machine and a $3 Hose Bibb Vacuum Breaker (HBVB) that closes the tubing outlet and opens a pressurized water connection behind it for 10 minutes when the upper drum is full too often.

An HBVB ("HeeBeeVeeBee") is like a relay with ports P, B, and A:

. . . . . . . . B . A (for "Atmosphere") . ^ . . P . . . . . . . .

When pressurized water is not present at port P, a spring check valve prevents flow at P and allows flow between A and B. When pressurized water is present at P (during backflush), it can flow from P to B, but flow between A and B is prevented.

The greywater outlet might look like this, viewed in a fixed font:

fpt---- --------mpt | $3 Dawn 100D saddle T hose / | ..................... clamp / | 1.5"

-------------------------------------- / | cap

-------------------------------------- / | . ^ 1" PE pipe . bbb--- | . | . hose | . -- 1.049" ID . clamp |

----------------------------------------------------- |

----------------------------------------------------- | fpt-- . . bbb----------rmpt \\ mht . . | --- . graywater HBVB . ^ . outlet . P . . . . . . . . . B . A ==> backflush HBVB . ^ . pressure . P . release ----------- mht | | backflush| solenoid | We might drill a hole in the 1.5" cap -------->| valve | for the 3/4" male adapter, then add | | a rubber washer, then screw on ----------- a garden hose adapter. | ^ | | | | The Dawn "Kwik-Seal" irrigation saddle T | backflush | clamps on the outside of a PE pipe and | water | allows using smaller pipe sizes, vs | | an insert T.

Nick

[nicksanspam]

Oops. That probably won't work, since the 1" male adapter ID is probably smaller than the 3/4" pipe OD, so it seems better and simpler to wrap the 3/4" pipe with electrical tape and a hose-clamp, like this:

Nick

[nicksanspam]

A garden hose connection might backflush the outer tubing and upper drum automatically, with a solenoid valve from a dead washing machine and a $3 hose bibb vacuum breaker that closes the tubing outlet and opens the hose connection behind it for 10 minutes when the upper drum is full too often, eg when an upper-drum float switch has been open a long time.

If the basement is 70 F and the switch heats a 1-quart water bottle with

12V 82 and 91 ohm 1 W resistors glued to the bottom, inside a 6"x6"x1' R20 box with conductance G = 2.5ft^2/R20 = 0.125 Btu/h-F, the final temp will be 161 F. With the upper drum always full and RC = 2/0.125 = 16 hours, the bottle will reach 140 F in -16ln((140-161)/(70-161)) = 23 hours.

An $8 Grainger 2E248 snap disk thermostat on the other side of a 6"x6" R5 wall between the thermostat and the bottle could open the backflush solenoid when it reaches 140 F and close it when it the thermostat reaches

120 F. A 33K 1/2 W hysteresis resistor could keep the thermostat closed with 3.412x120^2/33K = 1.49 Btu/h of heat until the bottle temp drops to 120-1.49R5/0.25ft^2 = 90 F.

A rubber stopper with 2 tubes in the bottleneck might move P pounds of 60 F backflush water into the 140 F water, so the bottle temp reachs 90 F when (60P+(140(2-P)/2 = 90, ie P = 0.8. If this takes 10 minutes, 0.8/(10x8.33) = 0.0096 gpm would flow through the tubes.

Nick

[daestrom]

Why not just a 1" external pipe cap with a hole for the 3/4 pipe to extend out from? ISTM that it would be easier to seal around the 3/4 pipe in the 1 inch cap than to build up with tape. With the 3/4 pipe extending out a short way, you can then do all sorts of things to attach the garden-hose bib.

daestrom P.S. Sorry, my ASCII-art skills are no where up to yours, you'll have to try and picture what I mean :-)