Institutional Barrel Stoves in Northern Uganda

Theory vs. Reality

February-March 2007

“Freda”
Introduction
Twelve institutional barrel stoves were built in Gulu, Uganda and donated for use
in schools, hospitals and feeding centers. The stoves were similar in design but used
several different materials for the combustion chambers. These stoves will be monitored
for durability of components and possible design changes. The stoves varied slightly in
details and used a variety of different sized pots. Controlled Cooking Tests were done
using one of the stoves to compare results from actual cooking practices with Water
Boiling Tests done in the laboratory.
Costs of parts and labor for fabricating the stoves were recorded and compared
with estimates for manufacturing the same stove in the US and in India.

Performance
The institutional barrel stove was compared with the open fire using a modified
version of the UCB (University of California Berkeley) Controlled Cooking Test (CCT).
Identical amounts of yellow beans were prepared three times using each stove. The same
pot was used in both stoves. All food was cooked by Freda Amono who was familiar
with local cooking practices and helped choose the dish to be prepared. Freda chose all
ingredients and fuel for use in the tests. Firewood was obtained from a nearby IDP

1
(Internally Displaced Persons) camp and was weighed before and after each test. Once
Freda determined the quantity of each ingredient to be used, it was weighed and the same
quantity of each ingredient was used in all tests. The time to cook and fuel used was
recorded for each test.
These Controlled Cooking Tests differed from published UCB protocol by not
adjusting the wood used for the quantity of char remaining at the completion of cooking.
This change was made to reflect the local cooking practice of letting the fire burn out and
cleaning out all ashes and char before a new fire is begun. Char is not saved or utilized
for other purposes.

Buying firewood in Unyama IDP camp

Institutional barrel stove Three stone fire

Controlled Cooking Test (yellow beans)

2
Results of CCT comparing two stoves

Stove type/model: Stove 1 Three Stone Fire

Stove type/model: Stove 2 Institutional Barrel Stove
Location Aid Africa Office, Gulu, Uganda
Wood species Average Hardwood

Test Test St
1. CCT results: Stove 1 units 1 Test 2 3 Mean Dev
Total weight of food cooked g 14421 13512 16216 14716 1376.0
Weight of char remaining g 1271 1018 1177 1155 127.9
Equivalent dry wood consumed g 17740 13366 14990 15365 2211.0
Specific fuel consumption g/kg 1230 989 924 1048 161.1
Total cooking time min 256 205 201 221 30.7

Test Test St
2. CCT results: Stove 2 units 1 Test 2 3 Mean Dev
Total weight of food cooked g 12773 12799 12664 12745 71.6
Weight of char remaining g 399 319 341 353 41.3
Equivalent dry wood consumed g 6269 5217 4943 5476 700.0
Specific fuel consumption g/kg 491 408 390 430 53.7
Total cooking time min 116 140 132 129 12.2

Stove 2
Comparison of Stove 1 and Stove 2 /Stove1 T-test Sig @ 95% ?
Specific fuel consumption g/kg 41% 6.31 YES
Total cooking time min 59% 4.79 YES

3
 Laboratory WBT (Water Boiling Tests)
Performed at Aprovecho Research Center Laboratory in Creswell,
Oregon

Performance
Institutional Barrel Stove vs. Open Fire
Fuel Use and Emissions per Liter of water
( boiled and simmered 45 Min)
Open
Institutional Fire Inst/OF
Fuel to Cook grams/Liter 61.2 223.6 27%
Carbon Monoxide to Cook grams/Liter 1.4 11.1 12%
Particulate Matter to Cook mg/Liter 36.2 472.6 8%

Discussion of Performance

Lab tests indicated that the institutional barrel stove could save 73% of the fuel
used by the open fire. The Controlled Cooking Test indicated only a 59% savings. Why
the difference?
Lab tests (the WBT) are used as a design tool to determine what the stove itself is
capable of when carefully operated under controlled laboratory conditions. It is not
predictive of actual fuel use. The Field test (the CCT) is used to get an estimate of how
the stove might actually perform when used by real people preparing real food and using
locally available fuels. True measurement of fuel savings requires a third test (the
Kitchen Performance Test) which measures actual fuel use in many households over a
long time period. The KPT was not performed in this study.

Three stone fires in schools cooking beans

The WBT in the lab incorporates a high power phase (bringing water to boil) and
a low power phase (simmering water for 45 minutes at just under boiling temperature) to
approximate a common cooking procedure (like cooking beans or rice).

4
 The actual procedure for cooking beans in northern Uganda uses high power only.
There is no simmering period. Beans are boiled vigorously for the entire time they are
cooking. No attempt to alter this practice was made during Freda’s cooking tests. After
the completion of the CCT tests however, Freda was very surprised to learn that the
temperature at a roiling boil (94°C) was no hotter than the temperature at simmer.
Lids were observed being used in about half the cases where beans were cooked.
Freda chose to use a lid both for the institutional stove and the three stone fire. When
lids were used, steam was observed escaping around the lid almost continuously. In the
CCT, the extra fuel required to produce this steam at prolonged high power helps explain
the difference in fuel consumption between the WBT and the CCT.
It should be noted that the institutional barrel stove had better heat transfer and
therefore boiled off more water than the open fire. Beans cooked with the institutional
stove were less “watery” than with the open fire and therefore weighed less. If the
weight of finished beans had been adjusted for the difference in water, the calculation of
specific consumption would have shown closer agreement between the WBT and the
CCT.
The institutional barrel stove cooked beans in 2 hours and 9 minutes (average) vs.
3 hours and 41 minutes for the open fire. The time required to reach a boil with the
institutional stove (16 minutes) was typically about half of that required with the open
fire (31 minutes).

5
 Cost of Manufacture vs. Cost Estimates

Costs for Institutional Stove from 200 Liter Barrel

basic model
(retail- Ceramic CC
USA) (India) (Uganda) (Uganda)
Combustion chamber actual estimated actual estimated
(imported to area) cost cost cost cost
309 steel (heat resistant) $28.62 $20.58 $28.62 $1.61
Labor (fabrication) $35.75 $3.25 $35.75 $0.00

Other Materials (obtained locally)

20 Ltr. (5 gal.) bucket $1.00 $0.50 $7.49 $7.49
insulation (vermiculite) $2.00 $1.00 $1.01 $0.00

200 Ltr Barrel (55 gal.) $6.00 $6.00 $25.36 $25.36

Skirt (from half of barrel) $3.00 $3.00 $12.68 $12.68

bolts (14 each 5/16") $3.42 $1.71 $3.34 $3.34

nuts (22 each 5/16") $1.54 $0.77 $1.15 $1.15
washers (32 each 5/16") $1.92 $0.96 $0.81 $0.81
spacer pipes (18" of 1/2") $0.37 $0.19 $0.29 $0.29

chimney brace $3.00 $0.00 $0.00

screws (24 each #8) $1.44 $0.72 $2.88 $2.88

elbow for chimney $4.54 $2.27 $0.00 $0.00

chimney (3 sections -9') $18.03 $9.02 $7.20 $7.20
Subtotal $110.63 $49.96 $126.59 $62.82

handles $2.59
Pot (60 liters) $65.90 $32.95 $69.16 $17.00

Assembly Labor (estimated 16

hrs.) $400.00 $6.50 $17.29 $17.29

Total $576.53 $89.41 $213.05 $97.12

12 combustion chambers made of three different grades of heat resistant steel

were brought from the United States for use in the stoves. 309 stainless (rated for
1050ºC) and 321 stainless (rated for 850ºC) were used for 4 combustion chambers each
fabricated by a sheet metal shop in Oregon. 310 stainless (rated for 1100ºC) was used
for four slightly larger combustion chambers with grates which were made in cooperation
with Dr.Brian Willson and his staff at Colorado State University. The cost of these metal

6
prototype combustion chambers was high (about $55 US) and should not be viewed as
representative.
Locally made insulative bricks from a nearby stove project (6 brick rocket stoves)
were not immediately available. The manufacture of 4 stoves utilizing this type of
combustion chamber had to be delayed until more bricks were fired. If this type of
combustion chamber proves durable it should be much cheaper (about $1.61 US) than
metal versions.
Another type of ceramic combustion chamber made in Kampala (UCODEA –
GTZ) was considered but not tried. These one piece units were designed for smaller
saucepans and are not suitable for large institutional pots. A larger version of this same
combustion chamber might be another inexpensive option.
All other components for the stoves were obtained locally. Metal parts in general
were found to be more expensive than anticipated.
Pots (called “saucepans” in Uganda) were difficult to obtain in quantity and cost
$70 to $110 US for 50 liter aluminum versions imported from India. Larger (100 liter)
pots and stainless steel pots of any size were unavailable. 70 liter 1.5mm thick stainless
steel pots could be fabricated in Kampala for about $145 US each.
A wide shallow “saucepan” with a capacity of 46 liters was available in local
markets for about $ 15 US. These thin aluminum pans were fitted with handles and used
in some of the institutional stoves.
Used oil drums (55 gal.) were available for $25 each (vs. $6 in Oregon). Many of
these drums were trucked in from Kenya so much of the increased cost was
transportation. Care must be taken to ensure that the used drums did not previously
contain toxic chemicals.
The institutional barrel stove was designed to incorporate a 20 liter World Food
Program cooking oil can to hold insulation around the combustion chamber. Many of
these cans are supplied to the IDP camps around Gulu. It was surprisingly difficult and
expensive to obtain good cans. We had estimated that they could be bought for less than
a dollar but we ended up paying $7 ½ dollars for each one. Most of the available cans
had large holes punched in the tops (which made them unusable for us). WFP cans are
also scarce because they are a prime source of tin for local sheet metal workers. We
found that it was less expensive to have a new “can” fabricated from other materials than
to try to buy “good” WFP cans.

7
20 liter WFP can being made into a combustion chamber

Other uses for WFP cans Making chimneys

Chimney sections 2 meters long and 15cm in diameter were made by a local
tinsmith. Local craftsmen are amazingly skilled and resourceful. Cost was about $4.60
for materials and $2.60 for labor.
Elbows and braces for the chimneys were fabricated from the scrap end which
had been cut out of the 55 gallon drums. This process was labor intensive and used about
a third of the time required to build the stove. Ready made elbows were not available.

8
 Vermiculite for insulating the combustion chambers was available in Kampala at
Canmin Inc. for about a dollar per stove. Wood ash is also widely available but must be
gathered as it is becomes available. Most people discard their ash daily.

Payback time for the institutional stove

No consideration is given to health benefits or time savings from using the

institutional barrel stove.
The following assumptions are made in calculating payback time:
1. Cost of wood is the same as the selling price in the IDP camps. Large
institutions may save money by purchasing larger quantities.
2. Cost of splitting wood is not considered. The observed practice with
rocket stoves in other IDP camps is that people begin gathering wood of
the appropriate size rather than splitting larger pieces.
3. Only one meal is prepared per day. If two or more meals are prepared
per day the payback time decreases proportionally.
4. Basic model stove using ceramic combustion chamber and 46 liter
“saucepan” is used.
5. Stove costs $100 US.

Cost of wood (Unyama camp) is 5500 Ush (Uganda shillings) = $3.17 US

Weight of wood 95.39 kilos

Cost per kilo = $0.0332 per kilo

Wood used by Three Stone Fire per pot of beans 15.365 kilos
Wood used by Institutional Stove per pot of beans 5.476 kilos
Wood Savings per pot of beans 9.889 kilos
Cost savings per pot of beans $0.328 US
Payback period is 305 days

Comments and Observations

When stoves were demonstrated at schools or hospitals, it was noted immediately
that the stoves used less wood than the open fires or traditional stoves they were
replacing. People tended to overestimate the fuel savings and guessed that the stoves
saved 90% of the fuel rather than the more probable 75% to 50% savings.
People were impressed by the speed with which water was brought to a boil. The
institutional barrel stove has very good heat transfer and brings water to a vigorous boil
in less than one minute per liter of water. Some people believed that there had to be a
larger fire concealed inside the barrel to provide all the heat.
Saving time was more important to cooks than saving money on fuel.
Administrators and directors (who pay the bills) were more concerned with fuel savings.
The institutional barrel stove produces very little smoke once the combustion
chamber is hot. Several people wondered “where does the smoke go?”

9
 Cooks appreciated that the chimney removed flue gases from the cooking area
and made cooking more pleasant. There was, however, little recognition that breathing
smoke could be harmful to health. .

Demonstrating stoves at Laroo Boarding School for War Affected Children

Projected cost of the stove did not seem to discourage prospective buyers. Often
the pot would cost more than the stove when larger (100 liter) pots were considered.
Other slightly larger (165 liter) stoves from Kampala were priced at $900 US per stove.
There were many requests for a smaller version of the institutional barrel stove for use in
the home.

10
 Greater savings in fuel consumption are possible if people are willing to use
stoves at low power for prolonged cooking.

Feeding center for children in Coope IDP Camp

These stoves will be monitored every three to six months to see if they are still
being used and if parts are failing. Initial enthusiasm for the institutional barrel stove
may fade if maintenance or design problems become apparent. Initial feedback from
Gulu Regional Referral Hospital was that their stove was in regular use and has replaced
three other stoves which were being used to cook for the hospital.
Continued monitoring and periodic testing are required for any successful stove
project.

Damon Ogle / Aprovecho Research Center

April 16, 2007

11