Designing a Heat Exchanger for the other 90%

Our family friends, Kenneth, Diane and young Andrew Carnegie Miller (scions of the industrialist/philanthropist Andrew Carnegie) graciously sent Sybille and me the book "Design for the Other 90%", with the inscription "For two caretakers, as you are, this might inspire and inform your work."

It contains a foreward by Barbara J. Bloemink and articles such as "World Designs to End Poverty" (Cynthia E. Smith), "Design for the other Ninety Percent" (Dr. Paul R. Polak), "Fuel from the Fields" (Amy B. Smith), "Design to Kickstart Incomes" (Martin J. Fisher), "One Laptop per Child" (Interview with Nicholas Negroponte and Yves Behar), "Reliable Renewable Rural Energy" (H. Harish Hande, Ph.D.), "Rolling Water" (Pieter Hendrikse), "Informal Community Solar Kitchens" (Sergio Palleroni), "Life Line" (Cheryl Heller), "Leapfrog, Design Strategies for Global Innovation" (Sheila Kennedy), "Katrina Furniture Project" (Sergio Palleroni), "Lessons from the Marginalized: A Manifesto for a Truly Public Architecture" (John Peterson), "Hearing for All" (Modesta Nyirenda-Zabula), and "Pot-in-Pot Cooler" (Mohammed Bah-Abba).

The book is as inspirational as it is uplifting, making the reader cry out "brilliant -- wish I had thought of that" with every page turn.

Now the rub: Being such a reader, long influenced by other such great books (in particular "Eco-Pioneers: Practical Visionaries Solving Today's Environmental Problems", Buckminster Fuller's "Critical Path" and "Design Outlaws on the Ecological Frontier" among others) I have long since graduated from "wish I had thought of that" to "how do I implement the new designs I AM thinking of?"

They say that necessity is the mother of invention, and I have thus deliberately gotten used to placing myself in places and circumstances where deprivation would force me to think of ways to improve my (often self-imposed) "poverty condition" (and that of those around me), so the design ideas that are scrawled in my notebooks are not few in number. What are very few in number are my successes in translating those ideas into practical realities.

Jeffrey L. Pressman & Aaron Wildavsky wrote a widely recommended book in the planning literature called "Implementation: How Great Expectations in Washington are Dashed in Oakland; Or, Why It's Amazing that Federal Programs Work at All" and its lessons were not lost on urban planning students like me who chose first to spend a decade in the ghettoes of Los Angeles and chase that bitter drink with another decade out into the field in "developing countries" such as Indonesia, Guatemala and Egypt. Implementation is the bear, whether it is making real our designs in policy or technology.

Our latest project at Solar CITIES is an example of how challenging implementation can be on the ground when you come up with a "design for the other 90%". First of all, you must realize that very few in that 90% will be able to do much to help you outright, even when the innovation you are considering is ultimately "for their benefit". People are generally risk-averse, and the poor have learned the bitter lesson that it is safer to stick with "tradition", even when it costs more in the long run in terms of discomfort, and labor and money lost. So when you are trying out your new idea in the field, don't expect it to win converts and supporters for a long while.

The case in point here: a do-it-yourself stove top heat exchanger for the other 90% in our Cairo survey area who don't have access to insolated roof space (i.e. their roofs don't get adequate sunshine, or they aren't allowed to modify their roof in any way, or the roof can't support any weight).

In conducting our survey of the community's of Darb El Ahmar (Medieval Cairo slum area) and Muqattam/Manshiyat Nasser (informal community built by the Zabaleen garbage recyclers) we have learned that many homes not only have no solar access, so we cannot provide them with Solar CITIES hand-made solar hot water collectors, as a promised outcome of our US AID small infrastructure grant, but the dire condition of the houses (typical for "incremental housing" situations) prohibits the families from making use of standard gas and electric hot water appliances, even if and when they can afford such consumer goods.

When we actually got into the houses to see HOW people were heating their bath water, we found that most were heating 10 to 15 liter cans of water on the stove (when they had a stove) or on a small gas burner on the ground.

Water heating, the survey revealed, is done by one of the women in the household (a mother or elder daughter) and takes approximately 30 minutes per person. Thus water heating alone can take a woman with 6 children (an average family size here) up to 3 hours of her day.

Clearly something must be done to lighten this burden, particularly if gender equity and societal transformation through education are considered important and realistic goals!

Our answer to the problem: designing a stove-top heat exchanger for the other 90% who must rely on gas burners for their bathing water.



The principle is simple and is derived from the standard gas heating appliance found in some of the homes where hot and cold water plumbing infrastructure has been provided: in a standard on-demand gas heater (the kind sold in Cairo) a copper tube snakes around a metal heat absorbing plate that surrounds several flame pipes. You ignite the flame pipes and the roaring fire quickly heats up the absorber plate. When you turn on the water, it flows through the copper tube and heats up. The output is instant hot water that lasts as long as the flame is on.

There is no storage tank, and the system is very efficient. It suffers, however from a couple of serious problems, even for people who can afford its modest (~ 500 LE, or $90) price tag (roughly a months salary for most people). One is that the diaphragm that allows the gas to flow to start the heating relies on adequate water pressure to operate. In our area, the water is frequently cut, and when it comes on, even in homes that have their own water pump, the water pressure can be very inconsistent. When the water is cut, heavy salts and other minerals in the water, which is of bad quality, crystalize and perforate the rubber diaphragm, rendering the appliance inoperable. While it is a simple matter to repair the diaphragm (I did it myself three times over the course of the year we lived in Maadi using a gas heater) and it only costs a few pounds (about 50 cents) if you know where to buy a replacement, most people DON'T know how to take apart and repair the heater and hiring a plumber (which costs an average of 60 pounds for the day, and who may charge at least 20 pounds for this procedure) is often out of the question for people at this level of poverty. Thus we have seen many homes with unused gas heaters who have gone back to boiling water on the stove.

We have witnessed the same thing with regards to electric hot water heaters: families will go to great expense to install them (necessitating finishing a bathroom with hot and cold water pipes and tiles -- not a trivial expense) and will use them for a couple of years until a combination of calcification and overheating when the water cuts out causes the heating element to burn out. At this point people will abandon using the heater and go back to using the stove, not knowing that a new electric heating element can be purchased for as little as 35 LE (7 dollars). The trick in many heaters, of course, is knowing how to install the new element. Most people simply consider the appliance broken and either throw it out or leave it sitting unplugged.

We have also found families who have unplugged their electric heaters and gone back to heating on the stove not because the electric heater didn't work, but because the electric prices proved prohibitive. It is "cheaper" to let the women in the family prepare the water in the "traditional way", and, as one respondent put it "I prefer to have my sister boil the water for my bath -- that way it is ready for me when I come to bathe; with my electric heater I not only have the cost to deal with, but it takes a half an hour to heat up, since I can't afford to leave it plugged in all the time. That means I have to think to switch it on and then wait. I would rather my sister simply calls me when it is time to bathe -- that way she can think about the heating."

Regardless of the reasons, a significant number of the households we have surveyed are forced to use gas stove heating for their hot water, and we cannot help them with solar hot water technology. So what to do?

The design of a stove top heat exchanger seemed most logical. Rather than having the mother or sister sit for hours tending to a pot of boiling water on an open flame, carrying the heavy and dangerous hot water from kitchen to bath each time it is ready (and risking scalding themselves or their family; third degree burns from bathing account for around 30 deaths and 300 hospitalizations a year according to Ain Shams Universities study in "Burns" Magazine), we proposed a copper heating coil in a cooking pot attached to a small 30 liter elevated cold water tank (basically a bucket with an output at the bottom) by a hose on one end and to a hose with a faucet on the other, attached to a hot water bucket.



The idea is that rather than placing the cooking pot on the stove and waiting for all the water in it to heat up, then emptying it each time, one could simply put the heat exchanger pot on the stove, heat it up in five minutes, then continuously run cold water through it (as in the "on-demand" gas heating appliance), running the hot water that results into the bucket at a safe bathing temperature (around 40 degrees) for transfer to the bathroom (or one could extend the hot water hose all the way to the bathroom itself).

Sounds simple.

The devil, of course, is in the details.

And here are the details from the first experiment here in Essen, Germany, where we are spending the Thanksgiving break:

Materials: (purchased from Bauhaus at Langmarckstr. 2)
81 Euros.
Breakdown:

7 meters of copper pipe (roll, 12 mm x 1mm), sold at discount 5 meter price (3.95 per meter): 19.75
2 brass winkel @ 2.48 ea:
4.96
Rohrnippel Verzinkt:
1.31
Teflonband 12x0.1mm:
1.02
2 Flexible Schlauch, 1/2" @12.26 ea:
24.52
Kugelauslaufhahn M.
9.25
Kugeldurchgangs
6.78
Getraenkefass 30 L
13.29

Basically, the faucet and on-off valve, and the two plumbers hoses, were the most expensive parts of the system. Plumbing parts always are. Our experience is that the plumbing supplies drives the costs of experimenting with new designs to prohibitive levels and discourages people from being innovative. Imagine, a small valve costing 9.25 euro!

Procedure:

We started by going to the Zoomarkt pet store and buying fine sand (1 Euro for a 2kg packet in the bird section) which we pouredwith a funnel into the copper tube after unravelling it to its full 7 meter length. Once it was filled with sand, we tightly coiled the copper tube, fitting it inside a cooking pot with a 25 cm diameter so that as much as possible was on the bottom of the pot and the rest spiralled up the sides. Then we removed the coil from the pot and took it outside and dumped the sand out. We then replaced the coil in the pot.

I connected the outlet of the 30 L plastic tank with a 3/4" to 1/2" nipple to the plumbers hose and connected that to the valve. to this I connected the brass winkel which I placed on one end of the copper tube sticking out of the pot.

On the other end of the copper coil, also sticking out of the pot, I placed the other brass winkel (angle or corner) to which was connected the other plumbers hose. On the other end of that was the faucet. This was placed over a 10 liter plastic bucket.

We placed the 30 liter plastic tank (sold here for beer making) on top of another, inverted, bucket next to the stove so that its outlet was above the pot with the copper coil. After experimenting with simply heating the pot with the coil inside, filled with water, and achieving lukewarm results, we finally filled both the coil and the pot with water, and let the entire pot heat up to 40 degrees reasoning that the best way to transfer heat to the entire coil was through outside water as a medium. The water in the pot would theoretically stabilize the temperature inside the coil which otherwise would simply radiate any heat it gained from the metal of the pot out into the air.

A log of the experiment is as follows:

9:17 PM, November 21, 2007:
Begin heating pot with copper coil on small electric stove burner. Pot contains 2 cms of water to cover the bottom coils. Water in is at 13 degrees C.
After 25 minutes the water in the pot has reached 93 degrees.
I open the faucet and let water flow to the ground bucket. It starts coming out at 84 degrees, but the temperature drops quickly as cold water from the tank flows through the coil. Within 3 minutes of flow the water coming out of the exchanger is at 23 degrees. By 4 minutes it is 22 degrees.

After 8 minutes the 10 liter bucket is full and I switch off faucet. The mixed temperature of the 10 liters in the bucket is 30 degrees.

At 9:50 we decide to switch to the larger electric stove top burner, figuring that it has higher heat output.

We open the faucet and start filling another 10 liter bucket. Inlet temperature of water into the coil is still 13 C.

9:52 The temperature of the water leaving the coil is now 28 C.
9:54 Brigitte suggests I fill the whole pot with the lukewarm water in the first bucket, now at 29 C, to see if that helps transfer heat. I do so.

Outlet temperature drops to 26 C and stabilizes there.

9:56 I close the faucet to let the pot heat up.
10:00 The water is 37 C in the pot
10:02 Water is 40 C in the pot
10:03 Water is 45 C in the pot
10:05 Water has reached 50 C in the pot START EXPERIMENT ANEW

10:07 Water outlet to bucket starts flowing at 53 C and quickly drops to 40 C.
10:08 Outlet 36 C
10:09 Outlet 34 C
10:14 29 C
The 10 liter bucket is full, and after mixing, is at 35 C

10:15 New bucket, Water in pot is now at 35-36 C, water inlet at 13 C, outlet at 26 C
10:21 full bucket, 10 liters mixed at 28 C

10:22 New bucket. Outlet starts at 30 C, drops to 26 C in 30 seconds, bounces up to 27 C.
10: 23 outlet temperature goes up to 29 C!
10:25 outlet temp 32 C
10:26 outlet temp 31 C

10:35 New bucket. Pot temperature is 33-34 C; inlet temp is 13 C, outlet starts at 29 C.
10:36 outlet temp 27-28 C
10:37 outlet temp 29 C
10:38 outlet temp 31 C
10:41, I restrict the flow rate and outlet temp naturally rises, to 35 C.
10:48 outlet temp 31 C, but of course it is taking longer to fill bucket.

10:49 New bucket. Outlet temp 36 C; I return to orignal flow rate by opening the faucet.
10:50 31 C
10:51 I decide to pour the 10 liters of 30 C water I collected into the remaining 20 liters of cold water (13 C) in the tank -- previously I had been refilling from the faucet at 13 C, but this was wasting water, as I was pouring the outlet water buckets each time into the garden. One thing to be remembered in implementation of designs into reality is that you are dealing with real, not theoretical, masses, volumes and materials. Water is expensive.
10:51 continued: The mix gives me a new input temperature -- water in tank is now 18 C. Temp at outlet now 32 C.
10:55 outlet temp 33 C -- a difference of ~ 13 degrees from input temp of 18 C, consistent with what we observed when input temp was 13 C and we were getting output temp of ~ 26C. It may be that the thermal gain from this electric stove burner with this system is at best 13 degrees over input at this flow rate.
11:00 outlet temp has climbed to 38 C, perhaps because entire pot is hotter...

11:01 New bucket. I pour the previous bucket (which also averaged 30 C) into the tank (which was at 18 C) and make it 22 C. The temp in the pot is now 40 C. The outlet temp is 36 C (again reflecting the idea of a roughly 13 degree difference)

11:08 outlet 37 C
11:10 outlet 38 C

11:12 new bucket
I pour the bucket of water into the tank, raising its temperature to 26 C.
Outlet temp is 37 C
11: 13 outlet temp 38 C
11:15 pot temp is 42-43 C, tank temp is 26 C and outlet temp is 39 C, again confirming idea that heat gain at this flow rate is roughly 13 C.
11:17 outlet temp 40 C
11:21 pot temp is 44 C, outlet temp is 40 C, tank temp is 26.
We now see that to get a 40 degree bath from our system on an electric stove burner the input temp needs to be 26 degrees.

11:22 New bucket. After mixing, tank is at 29 C. Pot is 44 C. Outlet is only 37 C, but flow rate is faster for some reason...
11:26 outlet 38 C
11:29 outlet 39 C

11:30 New Bucket outlet hovers between 38 and 39 C, stabilizes at 39 C
11:35 outlet temp 39 C... Is the water in the pot now taking the heat we want to transfer to the copper coils?

11:36 new bucket, tank temp is 32 C, pot temp is 43 C, outlet is 39 C.
11: 38 outlet temp is still only 39 C, and not rising, even though tank temp is 32 C. If it followed the curve from previous results it should have heated to at least 45 C. How can that be? Are we at a point where heat is now being tranferred to the air from the water in the open pot? Helmut suggests that we should keep the pot sealed in the next phase of the experiment.

11:40 New bucket
Tank temp 32 C, pot temp is 44 C, outlet temp 39 C

11:46 new bucket, we cool off tank, making it 16 C. Outlet temp now 29 C.
11:50 we make tank 18 C, outlet temp now 28 C

11:52 New bucket; tank is 16 C, pot is at 33 C, outlet temp 26 C, jumping between 26 and 27 C.
Now we are looking at 10 degree heating difference. Has the flow rate really increased? Why and how?
11:55 Tank is 17 C, outlet temp 26-27 C. I restrict the flow rate with the faucet and temp rises to 29 C. I open it up again.

12:00 AM, new bucket. Added warm water to tank.
Tank: 22 C, pot 34 C, outlet 30 C

12:05 new bucket
Tank 25 C, outlet temp 30 C
12:10 outlet temp 32 C, clearly the water is not taking on as much heat, and we notice that the electric range burner is throbbing from bright orange to deep red. Is something wrong with the burner after leaving it on for over 2 and a half hours? We terminate the experiment. The 10 liter bucket is now full with 30 C water.

CONCLUSIONS:

Much more experimentation needs to be done, which means more set up, more time, more mess, more trying the patience of Sybille's family and tying up the kitchen. Since our beneficiaries use gas burners, the results from the electric range oven may not be appropriate. We must test on a gas oven, and we must test at different flow rates. It was taking roughly 6 minutes to fill a 10 liter bucket, which is 1.6 liters per minute flow rate. Even if we could find a way to get the heat gain up, so that we were filling a 10 liter bucket every five or six minutes with 40 degree water, would that justify the effort and expense?

We need to calculate how long it takes to heat 10 liters to 40 degrees on the stove in a pot by itself. It may turn out to take roughly the same amount of time, in which case the only advantage of the heat exchanger would be if the output was piped directly to the bathroom, so that no carrying of hot water was necessary. This is assuming 1.6 liters per minute is an adequate flow rate for showering/bathing.

The cost of materials alone for this project was as much as purchasing a gas heater or electric heater in Cairo; we must see what the cost for materials is when purchased in Cairo (that will be the next step).

Perhaps a different design would work better -- is there a way to put the copper coil inside a hollow can (like a Nido Milk can) and place that directly on the gas burner? Are we losing heat by using a pot? By filling the pot with water? What is the best way to transfer the heat to the water in the copper pipes?

All of these are questions to be resolved through further experimentation. But that, of course, takes further investments in time and money and effort -- all things that the poor can ill afford, to say nothing of a graduate student long overdue on writing his thesis!