When I went to UC Berkeley, studying psychology as an undergrad, lots of people made fun of Freud, Jung, and Adler who wrote (a lot) about divergent theories of what makes people tick. UCB liked to think of itself as a scientific institution, and facts are proven by statistical validity. Why should we believe speculation from the dear departed? I eventually agreed and I was intrigued by the big question: How do we know the truth?

The paradigm that captured my thinking was the BLACK BOX. In black box theory facts are like an elephant in a box. Scientists poke sticks into small holes in the box and one stick hits a toenail. A scientist exclaims, “Why, what is in the box is hard and slippery!” Other sticks hit the softer belly or the trunk, resulting in very different observations. Hopefully, after many experiments an accurate description of the elephant can emerge. (Although, what the elephant is thinking may well remain private.)

I was really excited by statistics!

If you’re trying to know the effectiveness of something, calculating the statistical significance can help. It gives you a measured amount of confidence in the hypothetical conclusion. At UCB we did experiments and 95% confidence was the minimum that students had to achieve to get an “A.” To achieve 95%, the sample size and the size of the effect had to be big enough.

Here’s the application to stoves

When we try to use found wood sticks from the forest in our experiments, one stick, for example, has two inches of bark on it. Another stick has three inches of bark and the two sticks make very different amounts of smoke. So, the variable of using sticks that emit different amounts of smoke makes it more difficult to know the truth: Did changing the air/fuel ratio, for example, result in the stove making less smoke? Using wood with no bark, we can achieve confidence in five to seven tests. We have to do a lot more tests when the fuel has added variables.

When trying to understand heat transfer or combustion efficiency in the lab (not what happens in the field) limiting variables has a great appeal to lazy researchers like me at ARC. So, we do not design stoves in the lab!

We realized quite a long time ago that we could only investigate heat transfer and combustion efficiency in the lab, and then with great relief go to an amazing place, work with wonderful people in a new culture, eat incredible meals, etc. in order to help a local team evolve a stove using found everything (and testing/statistics).

CQC stove with swinging door
CQC stove with swinging door
A door was added to the CQC stove to block some of the air entering the combustion chamber.

Last week we shared some thoughts about the importance of gathering detailed data and making direct observations when testing stoves. Here’s a question we recently considered.

Did a partial cover over the sticks entering the combustion chamber help reduce emissions in the CQC stove with Jet-Flame?

No, not in this preliminary test series, although lots of interesting things happened! Placing a swinging door over the fuel opening into the combustion chamber has often occurred to Rocket stove designers and it appears now and then in stoves. Dr. Winiarski liked the swinging door.  The thought is to get the excess air down by reducing the amount of air coming into the combustion chamber through the door. With sufficient but less excess air, the temperatures needed for cleaner combustion should rise.

Excess air did go down when a metal cover was near to touching the tops of the sticks in our experiments. It fell from an average of 2.91 times stoichiometric to 2.54. The average temperature in the combustion chamber did rise as a result, from 557°C to 596°C. The partial cover was doing its job. Firepower went up (4359 watts up from 4012 watts) and that might have helped with heat transfer efficiency. However, in this particular case, the thermal efficiency was unchanged (36% covered and 37% uncovered).

The positive changes in excess air and temperature also did not affect the emissions. PM2.5 (56mg/MJ-d covered and 55mg/MJ-d uncovered) and CO (2.63g/MJ-d covered and 2.72 g/MJ-d uncovered) were not changed enough to show a difference. Of course, using a Jet-Flame that is introducing lots of changes into the combustion chamber, including more excess air, probably makes this an unusual set of circumstances.

The swinging door may be great in a natural draft Rocket stove and as usual, Dr. Winiarski was right. It didn’t seem to be needed in this scenario. That’s OK with me, because the cover obscured the visual clues that help to tend a fire, and make tending more fun. I felt like I was flying a plane through the fog and was glad to land.

Kirk Harris tests his newest TLUD design under the Laboratory Emissions Monitoring System (LEMS) hood.

Having an emissions hood allows ARC to make more detailed observations. We try to do one ISO 19867 test every day, establishing an in-house culture of data driven inquiry. Now, having a hood isn’t necessary for creating amazing stoves. Larry Winiarski, Tom Reed, Paul Anderson, Kirk Harris, and many other people have spent happy hours tweaking prototypes that then worked better and better. But when you get to the point where it’s hard to see the smoke, visual inspection doesn’t work as well. In the modern world, health requirements are violated by fairly invisible amounts of PM2.5. Of course, the same thing goes for CO which is odorless and invisible.

That’s why we offer free access to serious stove developers. Kirk Harris recently visited the lab to tune up his latest TLUD. We have a nice, warm cabin for him and daily access to one of our two emissions hoods. The cafeteria is open and Kirk brings food to prepare. If we are lucky, we hear him playing his flute. Dale Andreatta makes an almost annual visit and then likely climbs some mountain and goes somewhere to watch trains.

Steps of scientific inquiry

  • Complete background research into the topic
  • Observe the natural process
  • Make a hypothesis based on the observations and test it

We’ve recently added 4 oxygen sensors and 4 temperature probes to the data generated every second for greater depth of detail. Detailed observations help to see if one set of information had an effect on another.  For example, when temperatures rose at 3” up from the floor of the combustion chamber did CO or PM2.5 change at the same time? Was a significant effect on CO or PM2.5 seen above a certain temperature? What happened with the temperature probe data taken at 4” up from the floor? And more and more. The process of learning is data driven which makes a black box theorist happy.

In so many ways, the detailed observation of what is happening in a biomass stove is in its infancy. Establishing a culture of daily data generation at our lab (3 hours) is getting us closer to having “observed the natural process.” And, it’s Christmas fairly frequently when a statistically significant relationship is observed.

The high mass CQC stove with Jet-Flame inserted from the side.
The Jet-Flame in the CQC high mass brick Rocket stove

I ask for help when moving the CQC stove. We built it on a piece of plywood and two folks can, with care, move it around the lab but it is heavy. The sand/clay/cement bricks are dense at 1.4 grams per cubic centimeter after being baked many times in the stove. Dr. Winiarski advised that, when possible, Rocket stoves should float in water at less than 1 gram per cubic centimeter.

For a long time, people have added sawdust and other lightweight materials into earthen mixtures to try to lighten up stoves. I ended up at Shengzhou Stove Manufacturer (SSM) in China because for hundreds of years ceramicists had manufactured (and sold in Africa since 1407) durable earthen stoves that weighed around 0.7 grams per cubic centimeter. Their amazing clay floats when dug out of the ground! It is full of diatomaceous earth. The Shens own a 100 year supply of clay in two mines next to the factory.

Why go to all of this trouble to lighten stoves?

The heat from the fire is diverted into the mass of the stove body and less heat is available to cook food. It is harder to start a hot, intense fire in a high mass combustion chamber. In a natural draft stove, this can be disadvantageous. The open fire has other problems but, out of the wind, the hot gases from the flames directly contact the pot and it’s common for open fires to have higher thermal efficiencies compared to high mass stoves, including Rocket stoves. Lightening the bricks helps to address this difficulty. Heat is still diverted into the stove body, but less. Well insulated, mostly metal, Rocket stoves successfully avoid most of these losses.

Indigenous cooks, experts at using fire, often use grasses and twigs to start a hot, fast fire in a high mass stove. You need to pour the BTUs into the stove to quickly prepare food. Speed to cook is almost always the first priority when talking to cooks around the world. When the SSM Jet-Flame is added to the high mass stove, the mini blast furnace immediately starts a hot, over 1,000°C fire that delivers relatively hot gases into the channel gap around the pot created by the pot skirt. (The CQC skirt creates a 5mm channel gap that is 7cm high.) 

The Jet-Flame creates a surprising result

The thermal efficiency in the first CQC/Jet-Flame test (see below) was 33%. The 5 liters of water boiled in 12.5 minutes. After the first 12.5 minutes of heating, the over 1,000°C fire started to heat up the mass and the water boiled more quickly in 10.2 minutes at 38% thermal efficiency. Three more short, but intense, heating phases resulted in the thermal efficiency incrementally rising to 41%, 42%, and 45%. The progressively hotter gases scraping against the sides and bottom of the pot in the small channel gap were more and more successful at transferring heat through the metal walls of the pot into the water.

When thermal efficiencies are in the 40% to 45% range, the performance of the high mass stove is similar to low mass, insulated Rocket stoves. This similarity was completely unexpected at ARC.

Results of five tests of the CQC Stove with Jet-Flame.
A woman sits next to two rocket stoves.
A woman sits next to two rocket stoves.
Firewood is stored between a pair of CQC’s TLC Rocket Stoves.

C-Quest Capital recently announced a collaboration with Macquarie Group Ltd., a financial services company with A$550 billion in assets under management and 16,000 employees in 35 countries. The two firms will fund and deploy efficient cook stoves with pot skirts to one million rural households across Malawi, Zambia and Tanzania. CQC’s preferred rural stoves project standard is two stoves per household to decrease user fallback on three-stone fires.

USAID in-field testing in Africa showed that Rocket stoves with pot skirts reduced smoke emissions by 40% due to the use of less wood while cooking. Addressing health by increasing the air exchange rate in the kitchen and home is a fundamental component of this project. This is done by strategic placement of windows and doors, and promoting half-wall kitchens or well-protected external cooking spaces. A minimum of one visit per year by trained staff to each household to help repair, maintain, and ensure good use of the Rocket stoves is also essential to elevating adoption rates in the targeted areas.

Over the next decade, this investment will deliver over 40 million high quality carbon credits with verified Sustainable Development Contributions to the Voluntary Carbon Market. It is the first leg of a three-pronged program to transform the lives of low-income communities across Sub-Saharan Africa at scale. Ken Newcombe, CEO of CQC, comments, “Our hope is to include something like 100,000 Jet-Flames, assembled by Ener-G-Africa in Lilongwe, Malawi, in the project. Field tests have indicated that the Jet-Flame dramatically reduces PM2.5 emissions and exposure to cooks and their families, further protecting health. If the deployment doesn’t get to 100,000 sold by end of next year it’s not because of the demand – it’s because we couldn’t get the working capital and distribution channels to get the product to the market. Of course, we are exploring all possibilities.”

Sam Bentson and a Winiarski designed down-feed downdraft Rocket stove with added cooling fins.

We used to have problems with the upper portion of the sticks catching fire in Dr. Winiarski’s down-feed downdraft Rocket stoves. The draft had to be strong, and the sticks at the right moisture content and size/weight to burn up completely without any smoke back drafting up into the room. The vertical metal feed tube got too hot and could catch the top part of the sticks on fire, overcoming the draft moving into the stove. (Usually something like 3 MPH.)

When Sam Bentson and Karl Walter were making a 20 watt thermoelectric generator (TEG) for the EPA SBIR supported Integrated Stove project we had the same problem, until Sam added aluminum cooling fins to the top of the vertical feed tube, as seen above. The team had previously designed and built the cooling fins unit that Sam is holding in the photo to cool the cold side of the TEG. I was amazed how well cooling fins work!

But then I remembered how small the radiators are in automobiles with huge horse powers. I saw that Sam and Karl had added a fan to their radiator to make sure it could dissipate the 1.5 kilowatts running through the hot side of the TEG. Adding fins to parts of a stove that could use more dissipation of heat brings to mind several possible applications: the heat exchanger cylinder in the photo, a chimney that has high exit temperatures, or the outside of a metal combustion chamber to preserve the metal. But I would not use fins where they can get dirty! They don’t work, for example, on the bottom of a cooking pot where soot quickly fills the space between the fins. (I didn’t think of this before we tried it.)

Manufacturing pot skirts

In 2013, C-Quest Capital (CQC) began distributing and installing the TLC Rocket Stove (TLCRS), a high-efficiency, long-life metal and brick improved cookstove, to the rural poor of Malawi. Early learning has resulting in many upgrades to the stove to improve sustained use and a long life. Over the past two years, CQC has installed the TLCRS in 450,000 Malawian households. Beginning in January 2020, Ener-G-Africa (EGA), a Malawian entity formed by CQC and Malawian entrepreneurs, began manufacturing all the metal stove parts for CQC’s sub-Saharan Africa TLCRS program and has since produced more than 300,000 sets of parts.

Interior view of EGA Stove factory in Lilongwe, Malawi
Stove Kits ready to ship at Ener-G-Africa’s factory in Lilongwe, Malawi

More recently, in February 2021, CQC placed irrevocable orders for the first 10,000 Jet-Flames from Shengzhou Stove Manufacturer in China, marking the first large scale commercial commitment to Jet-Flame distribution in the world. With CQC’s funding, EGA’s factory in Lilongwe is currently building the second solar panel assembly plant in sub-Saharan Africa and will begin manufacturing the solar panels, and eventually the batteries, needed for the Jet-Flame Kit.  CQC is hoping the superior cost and cooking amenity provided by the Jet-Flame will make serious inroads to the charcoal user market.

Through the growing partnership between CQC and EGA, the TLCRS will be installed on a two stove per household basis in three million households across eight sub-Saharan African countries in the next four years. Together, CQC and EGA are setting a new standard for cookstove projects in rural Africa. 

Manufacturing pot skirts
Welded pot supports
Parts ready for packing
Manufacturing area at Ener-G-Africa’s factory in Malawi
CQC stove set up for testing under the LEMS hood

ARC is investigating how to optimize the performance of the SSM Jet-Flame in the CQC earthen brick stove. Forty six thirty-minute ISO 19867 Water Heating Tests were completed under the LEMS hood at seven fan speeds. Two 4 cm x 4 cm douglas fir sticks were burned side by side. Five liters of water in a seven liter pot were heated, and the CQC pot skirt was used in all tests.

Results

Tier 4 ISO Voluntary Performance Targets:

  • Thermal Efficiency           40% to 49%
  • CO                                     <4.4g/MJd
  • PM2.5                               <62mg/MJd

Time to boil: The time to boil decreased with an increase in fan speed.

Thermal efficiency: The thermal efficiency stayed close to 35% in most cases and was higher at 3 and 8 volts (around 40%).

Firepower: The firepower rose to 6.8kW at 8 volts, starting at 2.6 kW at 2 volts.

Emissions of Carbon monoxide: Generally emissions decreased with increasing fan speed.

Emissions of PM2.5: 7 and 8 volts scored the best, at half of the result of 5 volts.

Combustion chamber temperatures: The mid combustion chamber temperatures rose with increases in fan speed from 382C to 730C.

Excess air:  Lambda fell as voltage increased from 4.1 to 1.9.

We recommend that the project do enough field testing to determine what settings are preferable to local cooks, remembering that higher voltages consume more power. In this way, the Jet-Flame/CQC stove can be tailored to regional cooking, keeping in mind the power output and use patterns of the CQC photovoltaic solar system.

Here’s what the flame looks like when varying the voltage:

Kabanyana Murabukirwa Domina and Jean Marie Vianney Kayonga in Rwanda
Kabanyana Murabukirwa Domina and Jean Marie Vianney Kayonga in Rwanda
Kabanyana Murabukirwa Domina and Jean Marie Vianney Kayonga in Rwanda

One of the roles of the ARC engineer is to give accurate technical information to the in-field decision makers who are directing the stove project. The folks on the ground have to make sure that cooks really like the stove, that the price is market based, that manufacturing is arranged for, etc. ARC engineers and the field team work closely together as the project evolves.

A New Project in Rwanda

In Rwanda, Kabanyana and Jean-Marie and their NGO, ENEDOM, are working with C-Quest Capital and ARC on a carbon credit supported Jet-Flame project. We met Jean-Marie through the internet and realized that he is well known in the sector. In fact, he knows many of our friends in Africa. Dr. Dan Lieberman at Global Health Labs sent Jean-Marie twenty Jet-Flames, and he showed them around to many of organizations, like the World Bank, that have large projects in the country.

Real World Use Guides Product Improvement

Moving the Jet-Flame to the side of the CQC stove
Moving the Jet-Flame to the side of the CQC stove

When we envisioned the Jet-Flame we imagined that it would be inserted into the fuel door of a Rocket stove. Mr. Shen at SSM directed the effort to manufacture the Jet-Flame and it includes a beautiful stainless steel stick support that also protects the fan. However, it only took several weeks of trails for ENEDOM to make a strong recommendation to move the Jet-Flame to the side of the CQC stove. Cooks in their homes were accidentally burning up the cord!

We gratefully thank ENEDOM for helping us make fewer mistakes. It’s another great example of trying to make sure that reality is in the product.

cover of Clean Burning Biomass Cookstoves 2nd edition
Click here to download the free pdf file (16mb)

If stoves pollute in the lab, they certainly will in the field. We estimate at least 3 times more. Commercially available biomass cookstoves that meet WHO standards are very rare. ARC continues to be committed to doing research and development to help to get the needed new stoves to market so that field studies will show success in sales, protecting health, saving wood, and making cooks happy. We believe that sharing what we learn is very important! So, we updated our “textbook” and it’s available for free here. The chapters have been updated and rewritten to try and share everything that we have learned in the lab in the last five years.

Enjoy!

Here are some highlights:

  • With clean outdoor air, doubling the air exchange rate halves the concentrations of PM and CO in the kitchen.
  • Using an EPA model of Oakridge, Oregon, the outdoor air concentration of PM2.5 would only be increased from 13.1 μg/m3 to 13.3 μg/m3 if homeowners used an ISO Tier 4 PM2.5 cooking stove.
  • A catalytic converter works well with gases (30-95% reduction of CO) but not with smoke (30-40% reduction of PM2.5) (Hukkanen et al., 2012).
  • We think that the Harris TLUD is perhaps the first “close to optimal” cookstove. It scored 0.7mg/minute PM2.5 with pellets at Lawrence Berkeley National Laboratory. It has a 3 to 1 turn down ratio. Large natural draft static mixers create thorough mixing. Decreasing primary air reduces the rate of reactions (production of wood gas) if the air/fuel mixture becomes too rich. A stationary fan blade spins the flame for longer dwell time. And cooks at ARC love to use it.
  • When carefully tested at ARC, the SSM Jet-Flame in the CQC earthen stove scored Tier 4 for thermal efficiency, CO, and PM2.5.
  • Renewably harvested biomass can be a carbon neutral energy source when burned very cleanly.

We are getting closer to practical solutions! The ones we know about are in the book.