Biomass Briquettes in Malawi

Olle Faxälv Olof Nyström

Division of Energy Systems

Degree Project Department of Management and Engineering LIU-IEI-TEK-A--07/00129--SE

Minor Field Study MFS-report nr 103 ISSN 1400-3562

Abstract In Malawi 2.5 % of the forest disappears each year. The use of firewood and charcoal, deriving from forest resources, accounts for about 99 % of the household energy demand in Malawi and is a cause to the deforestation. The Government of Malawi recently launched a programme called Promotion of Alternative Energy Sources Programme (PAESP) with the aim to reduce the use of firewood and charcoal. One of the fuels included in the programme is the biomass briquette. The aim with this study is to evaluate the viability of biomass briquettes as a sustainable alternative energy source to firewood and charcoal for households in Malawi.

Research for the study was carried out during three months in Malawi. Visits were made to a number of briquette production sites to study the manufacturing methods and to collect briquette samples. The briquettes were tested using various methods and then compared with results for firewood and charcoal.

At the moment various production methods are used in Malawi, with a high difference in technical complexity and cost. Machines produced from wood using very basic mechanics can apply similar pressure as more advanced metal pressers. They also seem to be better suited than those made of metal, in terms of price and availability.

The majority of the briquette producers in Malawi use waste paper as base material. Although the paper briquettes are good, other raw materials will be needed if the production is supposed to be significantly increased.

The briquettes burn well using the most common stoves in Malawi, including the commonly used charcoal stove. While firewood is cheaper to use than other available fuels, the briquettes seem to be able to compete with the fuel costs for charcoal.

I II Sammanfattning I Malawi försvinner 2.5 % av skogen varje år. Användningen av ved och träkol, som kommer från skogstillgångarna, står för runt 99 % av hushållsenergi användningen och det orsakar avskogning. Regeringen i Malawi har nyligen introducerat ett åtgärdsprogram som heter Promotion of Alternative Energy Sources Programme (PAESP) med syftet att minska användningen av ved och träkol. Ett av bränslena som ingår i programmet är biomassabriketter. Målet med det här arbetet är att utvärdera biomassabriketten som en hållbar alternativ energikälla till ved och träkol för hushåll i Malawi.

Utvärderingen för det här arbetet gjordes under tre månader i Malawi. Flera ställen där briketter producerades besöktes för att studera tillverkningsmetoder och samla ihop olika briketter. Briketterna provades genom olika metoder och jämfördes sedan med ved och träkol.

I dagsläget används flera olika produktionsmetoder i Malawi, med stora skillnader i hur tekniskt avancerade och kostsamma de är. Maskiner producerade från trä med enkla mekaniska lösningar kan producera samma presskrafter som mer avancerade stålpressar. De verkar även lämpligare än de gjorda i stål, vad gäller pris och tillgänglighet.

Merparten av producenterna i Malawi använder återvinningspapper som utgångsmaterial. Även om pappersbriketterna är bra, så kommer andra material behöva användas om produktionen ska öka betydligt.

Briketterna brinner bra i de populäraste spisarna i Malawi, inklusive den välanvända träkolspisen. Medan ved är billigare att använda än andra tillgångliga bränslen så verkar det som att briketter kan konkurrera med bränslekostnaderna för träkol.

III IV Acknowledgements We would like to thank everybody who helped us through our work, especially:

Mats Söderström -For supervising our work

Dr. Charles Kafumba and the staff on Department of Energy Affairs in Malawi -Making it possible to carry out with the thesis and supervision

Per Lindskog and Kenneth Nyasulu -Mediate contacts

Annie Kamanga -Helping us in the various tests

Kyle and Amy Guerrero -Great accommodation and minibus service

V VI Acronyms 3SF Three-Stone Open Fire

CCS Charcoal Ceramic Stove

CCT Controlled Cooking Test

CEEDS Centre for Energy, Environment and Development Studies

CWAG Chembe Women's Aquaculture Group

DoE Malawi Department of Energy Affairs

FWS Firewood Ceramic Stove

GoM Government of Malawi

HHT Household Test

LPG Liquefied Petroleum Gas

MK Malawi Kwacha

MASEDA Malawi Socio-Economic Database

MEP Malawi Energy Policy

MWBT Modified Water Boiling Test

NGO Non-Governmental Organization

PAESP Promotion of Alternative Energy Sources Programme

PAMET Paper Making Education Trust

ProBEC Programme for Biomass Energy Conservation

SIDA Swedish International Development Agency

WBT Water Boiling Test

WESMA Wildlife and Environmental Society of Malawi

WICO Wood Industry Corporation of Malawi

WWF World Wide Fund for Nature

Exchange rates of currencies, 2006-12-31:

1 EUR = 190 MK

1 SEK = 20 MK

VII VIII Table of contents

1 Introduction...... 1 1.1 Problem Statement ...... 1 1.2 Aim...... 3 1.3 Study Boundaries ...... 3 1.4 Method ...... 3 1.4.1 Visits at production sites ...... 4 1.4.2 Testing of briquettes ...... 4 1.5 Constraints...... 5 2 Frame of Reference...... 7 2.1 Household Energy in Malawi...... 7 2.1.1 Cooking Stoves...... 8 2.1.1.1 3-stone open fire ...... 8 2.1.1.2 Improved Ceramic Stoves...... 9 2.2 Biomass Briquettes...... 10 2.2.1 Raw materials ...... 10 2.2.2 Shapes...... 11 2.2.3 Briquette burning...... 11 2.2.3.1 Airflow...... 11 2.2.3.2 Ash Removal...... 12 2.2.3.3 Positioning in fire...... 12 2.3 Briquette Production ...... 12 2.3.1 Raw Material Collection...... 13 2.3.2 Material Processing ...... 13 2.3.3 Pressing...... 13 2.3.3.1 WU-Presser...... 14 2.3.3.2 Screw presser ...... 15 2.3.3.3 Hand pressed...... 16 2.3.3.4 Heated die screw press...... 16 2.3.4 Drying...... 17 3 The visited production sites and their briquettes ...... 19 3.1 Department of Energy ...... 19 3.2 Orphanage in Ndirande, Blantyre...... 20 3.3 PAMET, Blantyre...... 21 3.4 WESMA, Lilongwe...... 23

IX 3.5 MIRTDC, Blantyre...... 24 3.6 Nordin Family, Chitedze...... 26 3.7 CWAG, Cape Maclear ...... 27 3.8 WICO, Blantyre ...... 30 4 Tests ...... 33 4.1 Water Boiling Test ...... 33 4.1.1 Method...... 33 4.1.1.1 Modified Water Boiling Test...... 33 4.1.2 Realization...... 35 4.1.2.1 The WBT for comparing the fuels...... 36 4.1.2.2 The WBT for comparing stoves...... 37 4.1.3 Results ...... 37 4.1.4 Sources of Error...... 38 4.1.5 Analysis ...... 39 4.1.5.1 Fuels...... 39 4.1.4.2 Stoves...... 42 4.2 Controlled Cooking Test ...... 44 4.2.1 Method...... 44 4.2.2 Realization...... 44 4.2.3 Results ...... 45 4.2.3.1 Sources of Error...... 46 4.2.4 Analysis ...... 46 4.3 Household Test...... 49 4.3.1 Method...... 49 4.3.2 The Household...... 49 4.3.3 Realization...... 50 4.3.4 Results ...... 50 4.3.5 Analysis ...... 50 5 Discussion ...... 53 5.1 What makes a good briquette? ...... 53 5.2 What presser to use?...... 54 5.3 The market for briquettes ...... 55 6 Conclusion ...... 57 7 Proposal to further work ...... 59 8 References...... 61 9 Appendices...... 63

X List of Tables

Table 1: Efficiency of cooking stoves...... 8 Table 2: Fuels and stoves in MWBT ...... 37 Table 4. Results for comparing stoves from WBT ...... 38 Table 5: Energy results from the tests compared with the theoretical values...... 43 Table 7: Average values of the results gained from the CCT...... 47

XI

1 Introduction

”By the year 2020, Malawi, as a God-fearing nation, will be secure, democratically mature, environmentally sustainable, self-reliant with equal opportunities for active participation by all, having social services, vibrant cultural and religious values and a technologically driven middle income country”.

This vision was announced by the former president of Malawi, Dr. Bakili Muluzi, in March 1998. It is understandable why the president find a sustainable environment to be important for Malawi, since about 90 % of the energy used in the country derive from forest resources. (MEP, 2003)

Unfortunately the handling of forests in Malawi is not sustainable today. Every year 2.5 % of the total forest in Malawi disappears according to government statistics (MASEDA, 2002). The reasons are various, but the extensive use of forest as resource for providing firewood and charcoal, is one of them.

Malawi is one of the poorest countries in the world, with 65 % of the population living on less than US$1 a day (MEP, 2003). The industrial sector in Malawi is small and the country's energy is almost exclusively used in households. Hence a situation where the forest is disappearing mainly affects the ability for households to meet their energy needs, which is most commonly spent on cooking.

The deforestation also causes increased amounts of silt in rivers creating seasonal dry ups, trouble for hydro power generation, and occasional flash floods, which threaten lives and infrastructure in the riverbanks. Furthermore it leads to sedimentation in lakes which threatens the biodiversity of fishes (MEP, 2003).

1.1 Problem Statement

The problem in Malawi is a misuse of the existing energy resources. If new more efficient methods to use the energy could be found the energy situation could be sustainable and the deforestation problem could be ameliorated.

For cooking, the major part of the population use firewood on an open fire, with a low efficiency.

1 The main reason why firewood is not more efficiently used is that the fuel can be collected for free by most Malawians. The charcoal production today is another example of how the energy is used in an inefficient way. Most of the charcoal in Malawi is produced with traditional charcoal carbonization technologies with a proven efficiency of about 10 %. The production of charcoal is an easy income source for producers, there is no need for investments or an economical capital to start producing charcoal. For the households charcoal is also a relatively convenient fuel since there is no need for advanced stoves or equipment to use it.

The Government of Malawi (GoM) has tried to take control over the charcoal production but never really succeeded. Laws and legislations have been introduced in order to reduce the environmental impact that is related to the charcoal production. Hence a licence is needed to produce charcoal legally. At present there is no producer in the country that has this licence. One of the major problems with these laws and regulations is that there are no alternative income sources to offer the illegal charcoal producers. At the moment it is also difficult for the households to find alternative energy sources on the market that can compete with the availability of charcoal.

The lack of capital among most households in Malawi makes it difficult to move from using either firewood or charcoal, to more advanced energy sources where even small initial investments must be made, for buying for example more advanced stoves or burners. Hence the substitute to these fuels needs to require a minimal capital investment, be as cheap and accessible as charcoal and firewood are, and at the same time be environmentally sustainable.

In order to fight the deforestation and reduce the dependence of the forest as a resource for energy, the Department of Energy (DoE) in Malawi launched a project called ”Promotion of Alternative Energy Sources Programme” (PAESP) in the year 2006. The aim of this project is to encourage the use of energy sources other than firewood and charcoal.

One of several alternative energy sources considered in this project is the biomass briquette (see illustration 1). Other alternatives include the biogas, Liquefied Petroleum Gas (LPG), ethanol and gel-fuel technologies. It is though evident that none of these latter alternatives can compete with the biomass briquette, in terms of the low capital investment that is required to use it.

2

Illustration 1: A typical biomass briquette made of compressed paper and sawdust.

An evaluation study about this fuel was carried out in Malawi the year 2000 (CEEDS, Biomass Briquette Extension, Production and Marketing), and then proved it to be a cheap alternative to both firewood and charcoal. Since some of the conditions for the biomass briquettes may have changed during the past six years, a new evaluation is needed to update on the information and thus validate whether this situation still obtains. That leads us to the aim of this thesis.

1.2 Aim

The aim of this study is to evaluate the viability of biomass briquettes as a sustainable alternative energy source to firewood and charcoal for households in Malawi.

1.3 Study Boundaries

The briquette evaluation will be made in terms of physical and chemical characteristics (like material content, size, weight, energy content), costs for the fuel and usability in household cooking stoves. The feasibility of the production method for each briquette type will also be evaluated. The briquettes will be compared with the characteristics of firewood and charcoal. This report does not include an evaluation of the social obstacles related to the use of biomass briquettes in households. It neither includes an elaborate market research or how the supply could meet the demand for biomass briquettes in Malawi.

1.4 Method

To carry out this study a journey was made to do field studies in Malawi for 3 months, partly

3 financed by the Minor Field Study scholarship, given by the Swedish International Development Agency (SIDA). The study has been made with support by the DoE in Malawi.

1.4.1 Visits at production sites

To get knowledge about the current status of the briquette production in Malawi visits to various manufacturers was done. The selection of places to visit was mainly done with help from the Department of Energy. During those visits an official from the DoE was always present. Visits made without the assistance of DoE was made at WESMA in Lilongwe and the Nordin family in the village Chitedze. At each site persons responsible for the production were interviewed about the briquette making activities and the used facilities were then seen. From each producer that was visited briquettes were collected to be examined further. At some sites the production was not running anymore, for various reasons, but every site had samples to give us anyway, although some of them might have been old.

1.4.2 Testing of briquettes

In order to make a statement about the briquettes as an energy source several tests were carried out. The various tests were made over a period of about two months and the necessary equipment was borrowed from ProBEC, based in Mulanje.

The performance of the briquette types was compared among themselves and with firewood and charcoal. All the tests were made in a firewood cooking stove (FWS) except the charcoal where a charcoal stove (CCS) was used. The FWS was recommended for burning the briquettes by CEED (2000). The test made at this stage was a modified water boiling test (MWBT).

Performance comparisons for the different stoves were also carried out. The stoves tested were a 3- stone fire, a FWS and a CCS. The fuels used for comparison in these tests were paper briquettes from PAMET, softwood and charcoal. The same MWBT as in the fuel testing was applied for the stove testing.

To make the comparison between briquettes, charcoal and firewood more complete, a Controlled Cooking Test (CCT) was made for each fuel. A meal consisting of nsima and vegetables was prepared by a local woman on a FWS and a CCS. The briquettes used in this phase were the ones made by the DoE.

4 Furthermore, one household was asked to use briquettes for cooking, in order to evaluate how user friendly they are and if there are any practical obstacles connected to the use of briquettes in the household. The testing lasted a week and the briquettes used during this time were produced by WESMA.

1.5 Constraints

The limited amount of available briquette samples from the producers affected the amount of tests that could be made. To assure the results of the tests (it is considered that) every test should be repeated at least three times. This is something that was not always possible, due to the lack of available fuel. The majority of the briquettes were collected in the Blantyre area, which is a few hours drive from Lilongwe, where the tests was made. The weather might have affected the results of the tests. Wind and temperature were not constant throughout all the tests, although a lot of effort was made to make the tests in sheltered environment to minimize the chances for the wind to influence the heat transfer from stove to pot. The tests were carried out during the months of November and December. This year the rainy season started approximately in early December, which brought cooler temperatures and moister air compared to the previous month. Most of the WBT was made in November before the rain, but all the CCT were made in December at the same time each day.

5 6 2 Frame of Reference

2.1 Household Energy in Malawi

As seen in the second pie chart in Diagram 1, the household sector accounts for over 80 % of the energy demand in Malawi. 99 % of the energy used by this sector is coming from biomass energy sources (the remaining 1 % comes mainly from electricity and paraffin). Approximately 4 % of the households have access to electricity.

Diagram 1: Energy consumption in Malawi displayed by source and end-user. (World Bank,1996)

About 85 % of the Malawi population, consisting of 12 million inhabitants, live in rural areas. The urban population is mostly found in the four biggest cities; Blantyre and Lilongwe (country capital), both having about 500'000 inhabitants each, and the smaller Zomba and Mzuzu, with populations of around 100'000 each. A map of Malawi with the biggest cities is shown in illustration 2.

In rural households almost exclusively firewood is used, when cooking. The firewood is normally taken from trees on the farmland or in the nearby forests, for free.

In urban areas most of the firewood users collect the wood from nearby trees for free, while about half of them sometimes buy firewood. The firewood is normally sold in street markets. Illustration 2: Map of Malawi, showing the locations of the four biggest cities: Lilongwe, Blantyre, Zomba and Mzuzu.(CIA,2006) 7 For the urban households that have to buy their fuel, an alternative to firewood is charcoal. Charcoal is made in earth kilns in rural areas and is sold in the cities, where the demand for fuel is high. The charcoal is a more convenient fuel in a way, since it has got a higher heat value (J/kg) than firewood and less smoke emissions, once it has started glowing. On the other hand, the price of charcoal is normally higher than the price of firewood (counted in money spent per meal). Therefore the fuel is mostly used by middle- to upper-income households. (World Bank, 2005)

The charcoal is almost exclusively made from hardwood, taken from indigenous forest. The GoM has tried to promote softwood charcoal, made from planted pine trees, but since the quality of this charcoal is not as good as the former alternative, households still buys the hardwood-based charcoal. The kilns where the charcoal is made are inefficient. To produce one ton of charcoal 7 tonnes of firewood is needed. (GoM, 2003)

2.1.1 Cooking Stoves

There are several cooking stoves that can be used for biomass energy in Malawi. The most simple, and affordable one is the 3-stone open fire. Other alternatives are the ceramic stoves. Those are more expensive to buy but they cook with better efficiency. The efficiencies of the three most common stoves are shown in Table1.

End use device Efficiency

Firewood Cooking devices 3-Stone Fire 10 %-14 % Improved Fire Wood Ceramic stove 20 % Charcoal Cooking Devices Improved Charcoal Ceramic stove 35 %

Table 1: Efficiency of cooking stoves. (MEP, 2003)

2.1.1.1 3-stone open fire

In African countries the 3-stone open fire (3SF) is widely used for cooking with firewood (see illustration 3). In Malawi 91 % of the urban firewood-using households prepare their meals on this open fire, according to the Malawi Energy Policy (GoM, 2003).

Illustration 3: 3 stone open fire

8 The stove consists of three stones placed on the ground, forming the edges of something like a triangle, which holds the pot a decimetre or two above the ground. Underneath the pot firewood is pushed in from different angles. The “stove” is inefficient, since much heat is lost to the surrounding environment. The advantage of this cooking device is that it is free and easily assembled (you just need three stones).

2.1.1.2 Improved Ceramic Stoves

One way of decreasing the use of biomass energy for cooking is to change the end use device. If more efficient stoves are used, less fuel is needed to cook the same amount of fuel. Therefore the DoE has been recommending ceramic stoves to the public. These stoves are made from recycled metal, and are lined with about 3 cm of clay on the inside. The clay insulates the combustion chamber against the environment (although some of the heat is absorbed by the material) and hence more heat is transferred to the pot. The stoves are sometimes given the epithet “improved” because they are a better version of the older and simpler metal stoves (without the clay lining).

The charcoal and firewood have different burning characteristics and the stoves need to have a suitable design for each fuel. While firewood transfers most of the heat by convection through the flames, charcoal transfers heat by radiation. When charcoal is used for cooking the fuel should be placed close to the pot to get the most efficient transfer of heat radiation. On the other hand, the firewood needs some space above the fuel for the hot flames. Hence the pot needs to be placed at a higher level when firewood is used, comparing to when charcoal is used. The designs of the ceramic stoves are displayed in illustration 4.

Illustration 4: Firewood ceramic stove and charcoal ceramic stove.

9 The firewood stove also has a shape that demands an opening (with or without hatch) where the firewood sticks can be pushed into the fire. (Interview with Andi Michel)

The charcoal ceramic stove (CCS) is the most common stove to use in Malawi and other African countries when burning charcoal, while few households use the FWS when burning firewood. Although the FWS is not close to being used as much as the 3SF, the stove is interesting in this study, since the Department of Energy is promoting this stove for firewood users.

2.2 Biomass Briquettes

The biomass briquette is a fuel, consisting of biomass, such as agricultural waste or waste paper, bound together and compressed into small pieces (approximately 5 to 15 cm). The fuel complies with the energy needs for poor households in developing countries, where firewood and charcoal is normally used. Briquettes are very seldom used in Malawi today. There are few producers in the country and the fuel is hard to find for consumers. If nothing else is stated, the information about biomass briquettes in this chapter is based upon facts found in the book Fuel Briquettes: Theory and Applications from around the World (2003), written by Richard Stanley for the Legacy Foundation.

Richard Stanley has a lot of experience in briquette production and he has got great knowledge about briquettes. Richard Stanley has also tried to introduce biomass briquettes as a household fuel in Malawi a few years ago.

2.2.1 Raw materials

A lot of different materials can be used for briquette making, for example agricultural residues like ground nut shells, straw, tree leaves, grass, rice and maize husks and banana leaves. It is also possible to use already processed materials such as paper, saw dust and charcoal fines. Although some materials burn better than others, the selection of raw material is usually most dependent on what is easily available in the surrounding areas of where the briquettes are made. Of course a briquette can consist of a blend between many different raw materials.

The inflammability is not the only thing that matters when the raw material is being selected. Another important characteristic is its ability to bond together, when compressed. For these reasons fibre-rich materials are good. When these materials are soaked in water and partly decomposed, the

10 fibres in the material are able to create strong bonds.

The calorific value of a basic paper/sawdust briquette will be around 15 MJ/kg. This value will of course differ depending on the selection of raw materials. It can be compared to firewood that is around 16 MJ/kg (dependent on moisture content) and Charcoal around 30 MJ/kg (CEEDS). These values should not be confused with the energy gained from the briquette when burned in different stoves.

2.2.2 Shapes

The size of the briquette has an influence on in which stove that it can be used, since it must be able to fit into the combustion chamber. The most common type of briquette is the so called doughnut shaped one. It has got a cylindrical shape with a hole in the middle. If burned properly, the central hole increases the combustion efficiency of the briquette, states the Legacy Foundation. In order to make these briquettes a presser is needed. The diameter of the briquettes is affected by which pressing equipment that is used, but usually they measure between 10 and 15 cm.

In Low Input Food and Nutrition Security: growing and eating more using less (World Food Programme, 2005) the author Stacia Nordin describes a way to make paper-based briquettes without the use of pressing equipment or tools. Paper that has been soaked for about half a day is squeezed by hand in shapes of balls or similar. The balls are then left to dry in 1-3 days, before they are ready for use.

In the section called ”Briquette Production” further below, the briquette pressing is described in greater detail.

2.2.3 Briquette burning

Below the most important theories about briquette burning is described. Richard Stanley states that there are three important factors that affect the fire when briquettes are used. Those are air flow, ash removal and positioning of the briquettes in the stove.

2.2.3.1 Airflow

Good airflow is critical for the burning of briquettes, as for other fuels. The optimal fire is reached

11 when the airflow comes from underneath the fuel. Insufficient air flow will result in a smoky fire, since the released volatile gases will not be completely combusted.

2.2.3.2 Ash Removal

The briquettes produce more ash than both firewood and charcoal. This can cause a problem in some stoves, where the air holes can get clogged, which affects the airflow. Legacy Foundation claims that the air holes in the bottom of a stove needs to have at least 1.5 inches (~37 mm) diameter, to be suitable for briquettes. The charcoal and firewood ceramic stoves do not comply with this rule since their air holes only have a diameter of about 0.75 inches. When using briquettes in a 3-stone open fire, extra tendering may be needed to remove ashes from the fire under the pot.

2.2.3.3 Positioning in fire

The positioning of the briquettes in the stove influences the burning characteristics. The briquettes can either be burnt just like they are, or they can be broken into smaller pieces. The former method is considered to make the combustion more long-lived, but less intense than the latter, although the ignition of the briquette is made more difficult.

If the briquettes are burned without tearing them apart, doughnut shaped briquettes should be placed in an upright position (i.e. having the inner hole facing upwards). This helps the air to pass through the central hole in the cylinder, which makes the combustion much more efficient, states the Legacy Foundation. The reason for the higher efficiency is because the radiant energy from the burning material is facing inwards, and not out from the fire. This characteristic makes the briquettes suitable in stoves with low efficiency, since the impact of the heat losses are greater in those cases. The hole also creates a draft through the central hole, similar to that of a chimney, which gives a clear path for good air-flow from underneath the briquette.

2.3 Briquette Production

The making of briquettes is divided in four main steps, as described in the manual Fuel Briquettes: a Users Manual (Legacy foundation, 2003).

12 2.3.1 Raw Material Collection

As said before, a lot of different ingredients can be used for briquette making. Burnable, fibre-rich material that is both available nearby and that can be taken free of charge is preferably selected. The manual labour required for the collection of material will then be the only related cost for getting hold of raw material.

2.3.2 Material Processing

To make briquettes the raw material should be pressed together, but before this, the material has to be prepared. The preparation is necessary to release and distribute the fibres in the material. This makes the materials more susceptible to bond when compressed in the presser.

The organic matter, like agricultural residues, first needs to be chopped or pounded into smaller pieces in dry condition. Then it should be left to partially decompose in order to loosen up the structure of the material. How long time the decomposition takes varies and depends on the material and the climate. After the different materials have been decomposed properly they should be soaked in water and blended. This makes the fibres to randomly distribute in the sludgy matter that is created.

If the briquettes are to be made out of waste paper the preparing process is different, and much easier. The paper must be soaked in water for about half a day, or more, and then it should be shredded and pounded into small pieces. When this is done the material is ready to be pressed into briquettes.

The pounding of raw material, whenever it is necessary during the preparations, are the most laborious and time consuming phase in the production chain. The pounding is usually made using large mortars and pestles (about ~1.5m tall).

2.3.3 Pressing

For the material to be de-watered and to bond, it is necessary to submit it to pressure. The method of pressing will affect the final shape and burning characteristics of the briquette. A higher density gives the briquette a higher heat value (J/kg), and makes the briquette burn more slowly.

The most common type of the briquettes is the cylindrical shaped, often with a centre hole

13 (doughnut shape). To press this type of briquette it is necessary to use a cylindrical mould, most commonly a perforated tube of PVC, placed in upright position. In the centre a metal piston can be placed which enables the making of a hollow shaped doughnut briquette. The tube is then filled with raw material. The raw material in the cylinder is then compressed by descending a disc or a solid cylinder that just fits in the PVC tube.. Water, blended in the raw material, leaves the tube through the perforated holes during the compressing phase. An example of this equipment is shown in illustration 5.

When compressing the briquette the compression of raw materials requires a non linear force to distance1. There are Illustration 5: Mould kit for making different ways of applying the force for pressing cylindrical doughnut briquettes shaped briquettes. Two common technologies are explained below.

2.3.3.1 WU-Presser

The WU-presser was developed by the Washington University more than ten years ago. It is constructed from either metal or wooden parts. The wooden version has been seen in Malawi at least since 1997 (illustration 6).

Illustration 6: A wooden WU-press in Malawi.

1 When the disc first starts forcing the raw material to compress downwards, the first centimetre travelled by the disc needs a lower amount of work (since the raw material in the tube contains a lot of air and water that is easy to squeeze), if compared to the work that is necessary during the last centimetre of the pressing movement.

14 There are a few reasons why the wooden version grew popular. Wood is a cheaper material and more available than metal in Malawi and because of the lack of financial means in the country the wooden press had the economical advantage. Another reason was the high availability of skilled manpower for producing in wood. (CEED, 2000)

Illustration 7: How to use the WU-presser.(Legacy Foundation, 2003)

The WU presser is pressing the briquette in three steps described in illustration 7. Each step will press with increasing pressure. This takes advantage of the non linear force to distance property of briquette pressing described earlier.

2.3.3.2 Screw presser

The screw pressers also make briquettes in upright cylinders. The raw material is compressed by lowering a metal disc which is moved vertically by a screw that is turned by hand. The disc moves approximately 1cm/rev, with a constant exchange ratio of the force. The screwing technology is powerful and becomes handy in the final compression stage where it is able to contribute fully with its advantages of a good exchange ratio of forces.

The screw press is most commonly made in metal. This makes it sturdy but often quite expensive. Richard Stanley, who is promoting the WU-press press through Legacy Foundation, claims that a press using screwing parts is not suitable for an environment where briquettes are made, since there is too much of granular and wet material around that may cause damages on the screw threads.

15 2.3.3.3 Hand pressed

As Stacia Nordin claims in “Low Input Food and Nutrition Security: growing and eating more using less” (World Food Programme, 2005) the briquettes can be pressed by hand, using waste paper as raw material. This method does not require any use of pressing equipment or tools, which makes it cheap and available to everybody. The method is though explained only for making paper-based briquettes, and not for using agricultural residues. Illustration 8 is the instruction of how to make these briquettes from the book.

2.3.3.4 Heated die screw press

The heated die screw press is an industrialized machine for producing briquettes. It uses the natures own binder, lignin. When heating up the biomass to 300°C the lignin melts and Illustration 8: Instruction paper on how to when cooled down again it stiffens and the briquette will get make hand pressed paper briquettes. (Art by Kristof Nordin,2005) the desired shape.

In illustration 9 a principal sketch of a heated die briquetting machine is shown. In the funnel (1) the biomass is gathered. It drops down on the screw (3), which is driven by an electrical engine (4). The screw presses the biomass into the die (2). Along the die there are grooves to prevent the biomass to rotate with the screw. The die is electrically heated and heats up the biomass to 300°C so that the lignin melts. The briquette is extruded (5) and chopped off in desired length. The briquette will be hollow and have Illustration 9:Simple sketch of a heated die scew press a pyrolyzed surface from the heating.

Some machines heat up the biomass before it goes into the screw. This decreases the wear on the screw and die. To a small extent it saves the energy needed for rotating the screw.

16 Besides the cost of the investment the machine also has a cost for the electricity consumed. Another cost is the screw that gets worn and has to bee replaced frequently.

2.3.4 Drying

After the briquettes are made they have to be left to dry, usually between 3 and 8 days. The number of days depends on the weather conditions, during the dry season it is a lot quicker.

17 18 3 The visited production sites and their briquettes

In the following chapter all the visited briquette producer are described, based on observations and interviews at the production sites. The briquettes produced at each site are described with a picture and some basic physical facts. The pressure stated is from brief calculations found in appendix 4.

3.1 Department of Energy

Raw materials: Paper, sawdust Press: Stanlink Shape: Doughnut Outer diameter: 100 mm Inner diameter: 26 mm Height: 48 mm Weight: 117 g Density: 0.36 kg/dm3 Pressing pressure: 1.7 MPa Comments: Made by staff at the DoE for marketing and demonstration purpose.

Department of Energy has made briquettes for exhibition and marketing purposes. Those are made using the wooden WU-presser. DoE has several such pressers at the BARREM office in the outskirts of Lilongwe. The briquettes made by the DoE were stored at the ProBEC office in Mulanje. The briquettes consist of waste paper and sawdust.

During a visit to the BARREM office some briquettes were made together with staff from the DoE. The purpose of the visit was to get an idea of the briquette making process. Briquettes were made from paper and different additives.

19 3.2 Orphanage in Ndirande, Blantyre

Raw materials: Paper, sawdust Press: Stanlink (without iron piston) Shape: Cylindrical Outer diameter: 100 mm Height: 70 mm Weight: 163 g Density: 0.30 kg/dm3 Price: Only produced for own use Comments: Slightly irregular shape in between the briquettes

In the township of Ndirande in Blantyre disabled people make briquettes at an orphanage. The briquettes are made for own use only, and are not sold to the public. A wooden WU-presser is used. The piston has been missing, so the briquettes can not be made in a doughnut shape as intended, instead they are solid. The machine is only used once a week. This is enough for covering the demand of briquettes that they have. Raw materials used are paper waste and sawdust that is transported to the site for free. Initially the briquette production at this site started with help form the Nkhomano Development Centre. There is apparently no communication between the orphanage and Nkhomano anymore, the reason is unclear.

20 3.3 PAMET, Blantyre

Raw materials: Paper, sludge from Unilever (residue oil from food processing) Press: Screw press Shape: Doughnut Outer diameter: 150 mm Inner diameter: 56 mm Height: 50 mm Weight: 295 g Density: 0.39 kg/dm3 Pressing pressure: 1.8 Mpa Price: 5 MK Comments: Bad smell from the sludge

The non-profit organization Paper Making Education Trust (PAMET), based in central Blantyre, developed in the 90's a press based on the screw technology. The press is made from metal parts that are bolted together. The machine costs 28'000 MK (of which the mould contribute the costs 8'000 MK). The press uses one mould, where two briquettes can be made at the same time, using a divider plate. See illustration 10 of the presser in use.

The briquettes at PAMET are made of waste paper, coming mainly from the Blantyre Print and Packaging (BPP). Sometimes PAMET has to drive and fetch the waste. The waste is being used both to produce recycled paper and to produce briquettes. According to PAMET there is an increasing demand for waste paper in Blantyre, from different stakeholders. Some paper is even being exported on trucks to Zimbabwe and Mozambique, for industrial recycling.

Also some small amount of sludge (containing oil) is Illustration 10: Screwpresser in use at PAMET added to the waste paper, to increase the combustion

21 performance. This sludge is delivered free of charge from Unilever in Blantyre, where the sludge comes out as a by-product form this food industry. The sludge was kept in a bucket and did not seem to be very pleasant to work with, since it is greasy and containing a mix of unidentified ingredients.

The organization has one person employed for producing briquettes. This person produces about 100 briquettes per day. The demand is increasing so Moses Binali, executive director for PAMET, is considering employing another person for extending the briquette production. Normally the waste paper is collected about once a week, and stored at PAMET. In the afternoon the employee soaks the wastepaper that he is going to use for briquette production the day after. Next morning he pounds the waste paper and then he uses the presser to produce briquettes from the pounded material.

The briquettes are sold in the township of Chilomoni in Blantyre. PAMET delivers the briquettes with a vehicle to the salesmen in this township. Some briquettes are also sold at the PAMET office in Blantyre. Briquettes sold by PAMET to the end consumer are sold for 5 MK/each, whereas they are sold to the salesmen in Chilomoni for 3 MK/each. At the moment 3 persons are selling briquettes for them. PAMET is not producing briquettes for making profit, but for marketing the product, as an alternative energy source to charcoal and firewood. Sometimes PAMET hosts marketing events of the briquettes in townships around Blantyre. They demonstrate how easy it is to cook beans, talk about how clean the briquettes are and try to explain the whole situation of the deforestation and how it is caused by firewood. The event lasts for a day.

It should take no more than four PAMET briquettes to make nsima for a normal household, says Moses Binali.

PAMET has trained a number of women in briquette manufacturing and a few years ago there was production running by women groups (with support from NGO's and church groups) in different parts of Blantyre, but today Moses Binali does not know of anybody who is still making briquettes, apart from PAMET itself. According to him, the reason is the increasing difficulty of collecting waste paper, especially in the rural areas. If there is no raw material to be found, the production can not continue. PAMET sold a few pressers to the women groups which were trained by them. The presser that is sold is a smaller version of the one used by PAMET.

Moses also claims that the waste paper in Blantyre is not by any means enough for producing

22 briquettes for the whole Blantyre region.

PAMET is an organization financed by NGO's. They do try to make profit so that it would be economically sustainable if the NGO's would retract their funds. The incomes are generated from selling recycled paper and briquettes.

3.4 WESMA, Lilongwe

Raw materials: Paper Press: Screw press Shape: Doughnut shape Outer diameter: 100 mm Height: 70 mm Weight: 163 g Density: 0.3 kg/dm3 Pressing pressure: 1.1 MPa Price: 20 MK for a bundle of four briquettes Comments: Irregular shape in between the briquettes

A few years ago Wild Life and Environmental Society of Malawi (WESMA) started a briquette production project at the National Sanctuary in Lilongwe. Staff from WESMA made a visit to PAMET in Blantyre to learn about briquette making. WESMA also bought a presser (the same that PAMET uses today) through PAMET, for the production.2 Raw material for the production at WESMA was office paper, given by offices in City Centre. The paper was dumped free of charge by companies or NGO's, at the National Sanctuary, where it was stored in two containers. The briquettes were sold in street markets in Lilongwe, and at the WESMA office in the Natural Sanctuary. Around “one year” ago, the responsibility of the production was given over to two persons employed by WESMA. The people made briquettes, but there started to be a problem with the

2 Except for the presser from PAMET, that was bought for a few thousand MK, WESMA was also given a presser as a donation from some person. This machine is a screw press, and looks really durable, but has not been used at any time.

23 market. There were more briquettes produced, than what could be sold. So the persons were then given also the responsibility to market the briquettes that they produced. For some reason this was not very easy, so after a few months the two employed briquette makers resigned. Instead of shutting down the production, WESMA gave a group of ten women the opportunity to use their facilities for producing briquettes. The women made briquettes for a few months, until the production stopped about one month ago. The reason given for the stop was that they could not find any market for the briquettes. Today there is no production at WESMA in the Natural Sanctuary. All the facilities are there, but nobody is using them. There is a storage room filled with hundreds of paper briquettes that were made before. The containers for raw materials are still there, with quite a lot of paper in them. The presser that was bought through PAMET is partially broken. The machine can still be used, although the joint between the screw and the pressing disc is broken. (George Bokosi, 2006)

3.5 MIRTDC, Blantyre

Raw materials: Paper, sludge from Unilever (residue oil from food processing) Press: MIRTDC screw press Shape: Doughnut shape Outer diameter: 150 mm Inner diameter: 60 mm Height: 53 mm Weight: 330 g Density: 0.42 kg/dm3 Pressing pressure: 1.6 MPa Comments: Very hard packed, Bad smell from the sludge.

The Malawi Industrial Research and Technology Development Centre (MIRTDC), in Blantyre, together with some other organizations, developed a new press which was ready for the market in 1999. This project started in 1998, hoping to be able to construct a presser that could make 20 briquettes at the time, instead of only one, as did the older pressers. It was then discovered that the force applied for pressing such amount of briquettes in one move was too big.

24 The final version of the MIRTDC press that came out on the market can make 12 briquettes at the same time, using 6 cylinders, each producing 2 briquettes. See illustration 9 of the press. The force on the cylinders is applied by lowering a metal disc that squeezes the raw materials in all the 6 cylinders at the same time, instead of just working on one cylinder which is most common. The metal disc is moved up and down by screwing. The crank handle is tube shaped, which makes it possible to locate a rod (like a strong broom stick or similar) to get more leverage, when needed. The machine consists of welded metal parts, and is made in Blantyre, after design drawings from MIRTDC. MIRTDC sells the machine for 40'000 Illustration 11: The MIRTDC press MK. MIRTDC says that the only buyers of this machine have demonstrated at their shop in Blantyre. been NGO's. MIRTDC hopes that the presser could be developed to be constructed using different, less expensive, materials.

Since MIRTDC did not present any record of the buyers of this product, it was not possible to find any users of the machine. MIRTDC though offered to test the exhibition machine in the MIRTDC shop. Raw materials brought from PAMET, consisting of paper and sludge, were used when trying to produce briquettes using the MIRTDC machine. With manual force the screw was turned, pressing the briquettes. Unfortunately, the machine did not support the forces applied, so the construction broke in one of the weldings. It is worth mentioning that a 2 meter lever was used to turn the crank handle with two persons operating it at the same time. A person from the MIRTDC claimed that the machine that broke was not constructed correctly, that it did not fulfil the specification of requirements.

25 3.6 Nordin Family, Chitedze

Raw materials: Paper Press: By hand Shape: Spherical Outer diameter: 70 mm Height: 48 mm Weight: 93 g Density: 0.52 kg/dm3 Comments: Irregular shape

Kristof and Stacia Nordin lives in the village Chitedze, 20 km outside of Lilongwe. They are working under a program called Never Ending Food. At their home in Chitedze and in their work they are promoting permaculture. Permaculture is a type of sustainable agriculture on the natures own conditions. It enables the farmers to harvest food all year around and cultivate their land without ruining fertility of the soil. In their work they have created the “The Low Input Food and Nutrition Security manual”, that is all about a more sustainable living. It contains information about anything between agriculture and what food contains the necessary nutrients. One part of this manual is about making paper briquettes from office waste paper. In their instructions they soak the paper overnight and then, without pounding or shredding, pressed by hand into balls the next day. They are then left to dry for 1-3 days. The briquettes are used for cooking during power failures or for heating on a Charcoal Ceramic Stove. Similar briquettes have been made by ourselves (the authors) at home using both office paper and newspaper.

26 3.7 CWAG, Cape Maclear

Raw materials: Leaves Press: WWF hand press Shape: Doughnut shape Outer diameter: 144 mm Inner diameter: 50 mm Height: 50 mm Weight: 186 g Density: 0.36 kg/dm3 Price: 2.50 MK Comments: Very fragile, falls apart very easily

Raw materials: Corn stalks Press: WWF Hand press Shape: doughnut shape Outer diameter: 144 mm Inner diameter: 50 mm Height: 50 mm Weight: 133 g Density: 0.36 kg/dm3 Price: 2.50 MK Comments: -

The production at Panda Garden in Cape Maclear is done by Chembe Women's Aquaculture Group (CWAG), under the support of the following stakeholders:

● HEED – Malawi

● WWF – Finland

● Rotary Limbe Club

● Rotary District

● Department of National Parks

● Department of Fisheries

27 Close to the national park is the WWF headquarter situated. Here is also where the WWF started up the briquette project year 2003. The organization then developed the equipment, which can be made by people in Monkey Bay. The presser consists of metal parts and produces doughnut shaped briquettes. Instead of using a perforated PVC tube for the mould, a non-perforated metal cylinder is used. The raw material is compressed using a metal plate, a smaller cylinder and a handle that is pushed down with hand power (see illustration 12). The metal plate is slightly smaller than the bigger metal cylinder, which enables the water in the raw material to escape on top of the plate. Illustration 12:The WWF designed briquette press used in Cape The headquarter does still produce briquettes today, but the Maclear. briquettes are only made to supply the staff at the WWF. The site has since 2003 served as an educational centre for people who want to learn how to make briquettes. In Cape Maclear they today have 56 women working with briquette production, in different sites in the village. Since the equipment is so light you can conveniently keep it in your household. Today the centre has taught people from a number of villages in the Mangochi area and even one from Blantyre. At the moment briquettes are produced in four different villages in the Mangochi district, using the methods that were taught at the Panda Garden. The 56 women in Cape Maclear, who are working with briquette production, are divided in 5 groups. Each group, containing about 10 people, work together with the production. They then share the profit between themselves. A woman can make between 100 to 150 briquettes a day. The briquettes are sold at the price 2.50 MK, which means that a woman working in production can make a profit of about 250 MK per day.

The briquettes are either produced only by corn stalks or by leaves. Sometimes they add grass or paper. Paper burns well but is expensive. The materials are collected from the surrounding areas, free of charge. They are mixed with water in a basin for about 3 days. Then they are left to decompose for 2 to 3 weeks, depending on the material. The corn stalks takes about 1 week more time to decompose compared to the leaves. The decomposition makes the pounding easier, and makes the briquettes easier to compress, in the presser. When the material has decomposed, it will be pounded. This takes a lot of time and work. After this, the briquettes are made with the pressing procedure. After the pressing, the briquettes are left to dry in the sun or the shade (during the rainy season). The drying takes about 4 to 5 days, but in the rainy season it takes about 10 days.

28 In Cape Maclear briquettes are used just as much as firewood, claims Lois Chembe. This is due to the deforestation problems that this area faces. The last twenty years a lot of forest has disappeared, and the people living in Cape Maclear has to go far to collect firewood. According to Lois Chembe it is not unusual that women spend their whole day on collecting firewood, since the transport is done by feet and the wood is collected far up in the mountains. Before people used to pay the entry fee for the adjacent national park (10 MK) and then collect as much firewood as possible and then return to the village. This is illegal and the park guards now watches more carefully over its visitors.

Since it is so laborious to get firewood the prices for the fuel is high in the village. Normally you will have to pay about the double price per useful energy unit for firewood than for briquettes, according to Lois Chembe. The briquettes are sold for 2.50 MK a piece. Usually you will have to use about two or three briquettes to make nsima for a household, depending on the size of the family. The briquette is more poplar during rainy season, since the firewood in the forest is wet then. One household completely depending on briquettes as fuel for meeting their energy demand will spend between 10 and 20 briquettes per day, or even more if the family is big.

One problem that the women in Cape Maclear face is that there is not enough market for the briquettes in the village. They could easily increase production, but there is no demand for this today. Then they must be transported to other villages, and this is not convenient.

People that are interested in briquette production can visit the site in Cape Maclear to learn how to make the briquettes, free of charge. If they are interested to start up a new production site, they can then apply for getting subsidies, for paying the equipment needed.

29 3.8 WICO, Blantyre

Raw materials: Sawdust Press: Die-heated screw- press briquetting machine Shape: Cylindrical with an inner hole Outer diameter: 58 mm Inner diameter: 20 mm Height: 400 mm Weight: 931 g Density: 1.16 kg/dm3 Price: 40 MK Comments: Hard and heavy with a burnt surface. Inside the hole there was dirt/ash on our samples.

The Wood Industry Corporation of Malawi (WICO) has several sawmills in Malawi. One of them is situated in Dedza. The facility has a briquetting machine that produces briquettes form sawdust. The machine has been running periodically from 1984. The last time it was running was about 6 months ago. WICO bought the presser for a subsidized price from Japan. From the beginning it produced 500 kg briquettes per day. This number has now been reduced to 300 kg. The total amount of sawdust produced every day is about 5 tonnes in Dedza. WICO also has saw mills in other parts of Malawi, e.g. in Zomba, where the mountain of sawdust is even bigger, according to Aman Kunje, working at WICO.

The machine that has been used is driven by electricity, and produces a continuous briquette, which is then manually broken into parts, each with the weight about one kilogram. The sawdust is hold together thanks to a process where the briquette is heated on the outside, which makes a strong shell.

The reasons why the machine is not in use today are various. One is that there is some part of the equipment missing. Another is that the increasing demand of timber has made the manager

30 concentrating only on the main activity: timber production. One other reason is that the market for briquettes at the moment is not reliable. Aman Kunje thinks that it still is difficult to compete with the charcoal prices. He though thinks that if a more effective briquette machine is introduced to produce briquettes from all the WICO sawdust, it could be more profitable, if there is proof of a market to rely on.

WICO's briquettes consist of 100 % sawdust from their own production sites. WICO has tried to blend the sawdust with cassava and maize husk to get better briquettes, but it was not worth the effort. The briquette will last for a month if it is kept in normal humidity, after that it will fall apart. This is one of the major problems with the briquette, that it is not possible to store it for a longer period of time. It is possible to keep it longer if it is stored in dry conditions. The consumers then refer to charcoal that can be stored for a long time without problems.

They have sold the briquettes on the market for personal use. The tea industry has been interested in buying the briquettes but the demand could not be satisfied with WICO's small production capacity. The price for the briquettes has risen from 5 MK to 40 MK due to the increasing electricity prices and inflation. WICO claims that there is no possibility to invest in a new briquette press and make a profit, due to the low prices of charcoal.

31 32 4 Tests

4.1 Water Boiling Test

What is interesting concerning the energy content of a briquette is how much of the energy in the briquette that can actually be used. The useful energy is the energy transferred into the pot that is used while cooking. If the same test is performed on each briquette and on firewood and charcoal, a good comparison can be made. The test is called the Water Boiling Test and it will be used for comparing:

• The briquettes with each other • The briquettes with firewood and charcoal • The briquettes performance on the most common stoves in Malawi.

4.1.1 Method

Stove developers around the world have developed a standardized water boiling test (WBT) to help them in their work. This existing test, found in appendix 1, focus on comparing the efficiency of different stoves. The WBT consists of three phases: • High power test with cold start • High power test with hot start • Simmering test The high power test measures the time and fuel it takes for bringing a certain amount of water to boil, first by using a cold stove and then a hot stove. The simmering test measures the amount of fuel it takes to keep the water simmering for 45 minutes. It is very versatile but it takes quite a bit of time to perform each test. Because of limited time, supply of briquettes and the fact that the focus of the WBT is set on stove performance, a modified water boiling test (MWBT) was needed.

4.1.1.1 Modified Water Boiling Test

The information that is wanted from the MWBT is: • How quick can the fuel bring the water to boil? • How much energy is transferred into the water relative the amount of fuel consumed? • Other notations (smoke, ease of ignition etc.)

33 To save time and fuel the two last phases in the original WBT (high power test with hot start and simmering test) were combined and the 45 minutes for simmering was reduced to 15 minutes in the MWBT. The specific time of 15 minutes is approximately the time it takes to cook the most common food in Malawi, nsima. The time limitation makes the test closer to a true scenario. In the beginning of the MWBT, when the water is brought to boil, a lid is used on the pot, but once the water starts to boil the lid is taken off. The stove testers around the world still discuss if the test should be performed with or without the lid for various reasons, but to take the lid off reminds more of the real scenario when cooking nsima (because the need of stirring) and will therefore be used in this test. The first phase of the WBT (test on cold-started stove) is not very necessary since it tells more about the heat capacity of the stove than the characteristics of the specific fuel.

The WBT is designed to be suitable for comparing stove-tests that has been performed in various parts of the world where moisture content of the fuels may differ a lot. Hence this difference should be taken into account, so it does not affect the results in the original WBT. In the WBT the moisture content of each fuel should be measured before using them for testing. The value is then used for calculating the energy that is needed to vaporize this moist during the test. The calculated amount of energy is then subtracted from the total energy that has been used in the test. Since the testing included in this report is made to compare fuels during a limited time period (3 months) and for a limited geographical area (Malawi), the moisture content for the fuels was not considered very important. The true performance of a fuel is partly dependent on its moisture content and therefore its effect should not be subtracted from the results.

The time recorded to bring the water to boil is from the moment the fuel catches fire until the temperature of the water reaches local boiling point, and the lid is taken off. To prevent the use of too much fuel to bring the water to boil faster, an aim is set to immediately keep the water simmering and not heavily boiling after the lid is taken off.

The amount of energy that is transferred to the water will be calculated by measuring the increased water temperature and the amount of water disappeared from the pot during the test. The energy will be compared to the weight of the used fuel. This measure (J/kg) will be referred to as ‘utilized energy’ from now on. The procedure for calculating this is described in appendix 4.

After the test, the remaining fuel in the stove will be weighed to calculate how much of that fuel that was not combusted. This will be measured because of the limited experience of burning each type of briquettes. It can be hard to know exactly how much fuel to put into the stove, something

34 that the consumer will learn in a few weeks.

The briquettes will produce quite a lot of ash (CEEDS 2000), therefore a formula for calculating the proportion unused fuel and ash is developed. The formula is found in appendix 4. The ash produced from a fully combusted briquette will be measured at the end of the testing sessions, when it will be possible to leave the briquettes in the stove until they are fully combusted.

Every specific test shall be performed at least three times, unless there is a lack of fuel or other practical obstacles. Three times is the number recommended in the original WBT.

The exact procedure that is set up for the MWBT is found in Appendix 2.

4.1.2 Realization

The equipment used in the testing was a thermometer, a scale, and the stoves. This equipment is listed and described in Appendix 6. The thermometer was inserted in the middle of the pot and the water through a small hole in the lid. The testing equipment that was used is shown in illustration 13. For each test 20 g of softwood twigs were used to start the fire. To start the charcoal 30 g was needed because it is a fuel that is Illustration 13: Laboratory equipment for the MWBT harder to ignite. The amount of twigs needed was decided by trial and error. The testing took place in a backyard of a house in Area 6 and one in Area 43, Lilongwe.

35 4.1.2.1 The WBT for comparing the fuels

The following briquettes were tested. They are all collected by us and they have different characteristics in one way or another:

● Paper and sawdust briquette from DoE

● Paper briquette from orphanage in Ndirande

● Paper briquette from PAMET

● Paper briquette from WESMA

● Paper briquette from MIRTDC

● Leaf briquette from CWAG

● Corn stalk briquette from CWAG

● Paper briquette from the Nordin family in Chitedze

● Sawdust briquette from WICO

To get the reference to the woodfuel these were also tested:

● Softwood

● Hardwood

● Charcoal

Briquettes and firewood were tested on a firewood ceramic stove, but for the charcoal tests it would not be fair to use that stove, since the design is not suited for charcoal burning. Instead a charcoal ceramic stove was used, which is very similar to the firewood ceramic stove, but constructed for charcoal use.

The briquettes from CWAG were impossible to complete a whole test with. They produced too much smoke to be able to stay close to the stove for tendering. When they burned they produced a big amount of ash, which filled up the stove and clogged the air holes. A decision was made to interrupt these tests for health reasons.

The briquettes from the Nordin’s were not burning very well. They seemed slightly heavier then the other briquettes which the calculated density confirms. High moisture content was suspected as a possible reason and it was needed to control. To control the moisture content the briquettes were put into an oven at 70°C to vaporize the water. They stayed in the stove until there was no weight loss from them. The briquettes from Nordin’s turned out to have a moisture content of about 18 %, which can be compared to the moisture contents of briquettes made by DoE and WESMA, that measured less than 6 %.

36 4.1.2.2 The WBT for comparing stoves

For the comparison of stoves, the WBT was performed in the same way as for the fuel testing. The Firewood and charcoal were only tested on the stove produced for the specific fuel, while the briquettes were tested on all three stoves. The scheme over the testing is illustrated in table 2. During the end of the stove testing phase the charcoal stove broke so it had to be replaced. The new CCS was similar design but 1kg (~20 %) heavier. They were bought from different producers.

PAMET briquettes were used because it seemed like an average briquette concerning size and burning characteristics. It was also one of the briquettes where a surplus existed.

Firewood stove Charcoal stove 3 stone fire

Briquettes from PAMET X X X

Softwood X - X

Charcoal - X -

Table 2: A table of which fuels where tested on which stove during the stove testing in the MWBT

4.1.3 Results

The results are presented in the table below. This is a summary of all the results gained from the test. The full table of results is found in appendix 7. Some test has been declared invalid for various reasons. In a few tests the wind picked up and the losses from the pot were considered too big. Another reason was that there were too much or not enough water vaporized, this indicates that the water has not been simmering, respectively it has been boiling too hard. The failed tests are also documented in appendix 7 but in Italic. The reason why they are made invalid is described in the bottom of the column.

A general notation when doing the test was that the more compact briquettes did not need the same tendering. They burned for a longer time without any need to put more fuel into the stove.

37 MWBT Fuels Fuel Softwood Hardwood Charcoal DoE Ndirande Energy per weight unit (kJ/g) 5,1 3,7 7,3 3,6 3,1 Weight of used fuel (g) 253 366 194 426 456 Time until boiling (min,s) 15.27 12.55 17.42 10.45 14.15 Price per mass unit (MK/kg) 12,70 10,75 32,68 N/A N/A Price per energy unit (MK/MJ) 2,50 2,93 4,46 N/A N/A Density (kg/dm3) - - - 0,36 0,30 Number of tests 4461 4

Fuel PAMET WESMA MIRTDC Nordins WICO Energy per weight unit (kJ/g) 3,8 3,4 4,9 2,7 5,3 Weight of used fuel (g) 371 376 305 413 282 Time until boiling (min,s) 12.45 12.57 10.51 15.53 11.50 Price per mass unit (MK/kg) 16,96 23,52 N/A N/A 36,90 Price per energy unit (MK/MJ) 4,48 6,99 N/A N/A 6,94 Density (kg/dm3) 0,30 0,40 0,42 0,52 0,40 Number of tests 2342 1

Table 3:Results for comparing fuels from WBT

MWBT Stoves Stove 3SF 3SF FWS FWS CCS CCS Fuel Softwood PAMET Softwood PAMET Charcoal PAMET Energy per weight unit (kJ/g) 3,0 2,5 5,1 3,8 7,3 4,7 Weight of used fuel (g) 448 508 253 371 194 288 Time until boiling (min,s) 12.40 14.32 15.27 12.45 17.42 12.40 Price per mass unit (MK/kg) 12,70 16,96 12,70 16,96 32,68 16,96 Price per energy unit (MK/MJ) 4,25 6,86 2,50 4,48 4,46 3,58 Number of tests 2242 6 2 Table 4. Results for comparing stoves from WBT

4.1.4 Sources of Error

It was hard to simulate exactly the same conditions for each test. The most disturbing factor was the weather conditions. When the testing started it was in the end of the dry season. Warm winds with dry air are typical for the dry season. The wind made the testing impossible a few times. It increased the heat transfer to the surrounding environment (losses) and made it hard to keep the water simmering. This resulted in remarkably low energy values. At the end of the testing period the rain

38 season started with a few rains. The rains probably increased the humidity of the air. Unfortunately there was no available equipment to measure the humidity.

In the MWBT the stoves should be hot started each time. The definition of hot start was to wait for no more or less than 10 minutes between the finish of one test and the starting of another. Since the hot stove was defined by a time parameter more than a certain temperature, it is possible that there might have been some differences in starting temperatures of the stoves, depending on how much heat that was conserved from the fire in the preceding test.

The way that the briquettes were burnt in the stoves may have affected the outcome of the tests in some way. Because of the size differences, it was not possible to burn the large diameter briquettes (WESMA, PAMET, MIRTDC) in the FWS without first breaking them into smaller pieces. The size of the combustion chamber was simply too small to fit them inside.

4.1.5 Analysis

4.1.5.1 Fuels

At the production sites visited, the producers never had to pay for the raw materials. The production costs they have are the handling of the raw materials and the work for pressing. This makes the unit kJ/g most suitable for the comparison concerning energy, since the costs then mostly depends on the amount of raw material. These results are shown in Diagram 2.

MWBT Fuels - Results (Best first)

Charcoal WICO Softw ood Energy per MIRTDC weight unit PA MET (kJ/g) Hardw ood DoE WESMA Ndirande Nordins

0,0 2,0 4,0 6,0 8,0

Diagram 2: Results from the MWBT for comparing fuels.

39 Sludge? – MIRTDC and PAMET

In the briquettes from MIRTDC and PAMET the same raw material was used, they use plain paper but add the oily sludge from UniLever. The WESMA briquette has the same shape and density as MIRTDC. The big difference is the Sludge. According to the test results the sludge seems to increase the utilized energy of the briquette.

The difference between PAMET and the MIRTDC is the density, which indicates a higher level of compaction rate. The MIRTDC briquette has the higher level of energy, which indicates that the higher compaction increases utilized energy. The densities for these briquettes together with the rest of the briquettes are shown in Diagram 3.

MWBT Fuels - Results (Best first)

Nordins

MIRTDC

WICO Density (kg/dm3) WESMA

DoE

PA MET

Ndirande

0 0,1 0,2 0,3 0,4 0,5 0,6

Diagram 3: The measured densities of the briquettes used in the MWBT

The bad smoke and a smelly briquette is a comment from most of the test with the briquettes that has the sludge.

Paper or Sawdust? – WICO and WESMA

WICO and WESMA have the same density. WICO is from pure sawdust and WESMA is from paper. The WICO briquette gets a higher amount of utilized energy from these two. This indicates that the sawdust has higher calorific value.

40 Pressure and sawdust? – WESMA, Ndirande and DoE

The factors that differentiate the briquettes made by WESMA, Ndirande and DoE from each other are the density (pressure) and the sawdust content. When looking at the utilized energy for these three briquettes, the conclusions gained above are confirmed. Higher pressure and sawdust gives higher utilized energy.

In general it seems like the briquettes makes the water boil faster than both charcoal and firewood, see diagram 4.

MWBT Fuels - Results (Best first)

DoE MIRTDC WICO Time until PA MET boiling Hardw ood (min,s) WESMA Ndirande Softw ood Nordins Charcoal

00.00 07.12 14.24 21.36

Diagram 4: Times needed to bring the water to the local boiling point in the MWBT

Price?

Since most of the briquettes in the test are not sold in the open market it is hard to say anything about the price. Only the briquettes from WESMA and PAMET can be found in the market. The prices per kJ for these two are a lot higher than from softwood when consumed in the same stove. For charcoal the price is the same as for the PAMET briquette, while the price for the WESMA briquette is still higher. The results are shown in Diagram 5.

41 MWBT Fuels - Results (Best first)

WESMA

WICO Price per PA MET energy unit (MK/MJ) Charcoal

Hardw ood

Softw ood

02468

Diagram 5: Calculated prices from the MWBT results for comparing fuels.

4.1.4.2 Stoves

For briquettes (PAMET) it seams that a normal charcoal stove gives the higher energy efficiency from the three stoves. The 3-stone fire gives bad efficiency for both firewood and briquettes. Though when comparing with the firewood stove the firewood have 41 % less energy value in the 3-stone fire, while the PAMET briquette has only 35 % less energy value in the 3SF, see Diagram 6. This validates the theory of Richard Stanley that briquettes are less sensitive to the performance of the specific stove than firewood.

Energy Content MJ/kg

8 Charcoal PAMET 7 Softwood/Charcoal

6 Softwood 5

4 Softwood 3

2

1

0 3SF FWS CCS Stove

Diagram 6: Comparison of briquettes, charcoal and firewood in the three different stoves tested.

42 From the theory the calorific values and the efficiencies from the used stoves are known. With these values the theoretical values of the utilized energy can be calculated. These values together with the values gained form the tests are illustrated in table 5.

Theoretical Theoretical stove Theoretically Utilized energy calorific value efficiency utilized energy from tests

Firewood in 3SF 16 12 1,92 3 Firewood in FWS 16 20 3,2 5,1

Charcoal in CCS 30 35 10,5 7,3 Table 5: The values of the utilized energy from the tests compared with the theoretical values.

The cheapest fuel to use according to the test is the firewood in a FWS shown in diagram 7. However, in Malawi most people use CCS or 3SF when cooking. So the options that exist today without investments for the people in Malawi can be narrowed down to: • 3SF with firewood • 3SF with briquettes • CCS with charcoal • CCS with briquettes. From these four alternatives the use of briquettes in a CCS turns out to be the cheapest alternative.

Cost per Energy Unit MK/kJ 8

7 PAMET Softwood/Charcoal 6

5 Softwood Charcoal 4

3 Softwood

2

1

0 3SF FWS CCS Stove

Diagram 7: Calculated cost for of briquettes, charcoal and firewood in the three different stoves tested in the MWBT

43 4.2 Controlled Cooking Test

4.2.1 Method

Controlled cooking test (CCT) is a way to test stoves with focus on testing under conditions closer to the reality. A local cook is told to cook a typical meal on the specific stove. During the cooking everything will be measured and recorded. It is harder to get as precise results as the WBT will give. Still it is a good complement to examine how the stove performs under true conditions.

The detailed requirements for the CCT are described in a manual from 2004, written by Rob Bailis (University of California-Berkley) for the Shell Foundation (see appendix 3 for details). As for the WBT the CCT is not developed for comparing different fuels. The original CCT data sheet is made for firewood and can not be used for testing briquettes or charcoal. A modified controlled cooking test sheet has therefore been set up for this purpose. This does not change the way that the CCT is performed, only the corresponding documentation.

The results of the fuel performance in a CCT are measured mainly by using the following indicators:

● Total cooking time, from starting the fire to ready food.

● Amount of used fuel.

● Ease of cooking with fuel.

4.2.2 Realization

The person selected for the cooking tasks was Annie Kamanga. She is 33 years old, born and living in Malawi. She prepares food for her husband and four children every day, on either a 3-stone open fire or a ceramic charcoal stove. Together with her, an appropriate common meal was selected for the CCT. It was decided that the meal should be nsima and vegetables. This meal is very common in Malawi. Nsima, which consists of maize flour that is boiled into a thick porridge, is eaten every day by the majority of the population, and very often accompanied by a mix of fried vegetables. The exact amount of food that was going to be cooked during the test was decided (this specific recipe can be found in appendix 5). For making the meal only one stove was necessary at a time, since the vegetables was first prepared in one pot, and when ready, replaced by another pot where the nsima was made.

44 The tests were performed in the afternoons of four days in December 2006 just outside a house in Area 43 in Lilongwe. The stoves were to be cold started during each test, and therefore only one test per stove and day was possible. The CCT was performed both on a CCS and on a FWS. On each stove briquettes were tested, but also charcoal on the CCS and firewood on the FWS. The briquettes that were used during the CCT were the ones made by the DoE. The firewood and charcoal was bought at a market in Area 18, Lilongwe.

The briquettes had to be broken into pieces in order to fit the combustion chamber of the CCS, while it was possible to use whole briquettes in the firewood stove. Although it was possible, Mrs Kamanga did often break these as well.

For making the nsima the maize flour needs to be mixed with a certain amount of water to get the right consistency. The quantity of vaporized water depends on the firepower (although a lid was used to reduce this loss of water), hence the amount of maize flour that had to be used, sometimes was not exactly as planned.

4.2.3 Results

The results of the CCT are displayed in tables 6, showing the most important measures. Complete results are shown in Appendix 8.

Firewood Ceramic Stove Tested Fuels Time Fuel Used Fuel Cost Combusted Fuel Fuel Cost Firewood 27:00 312 g 3,96 MK 265 g 3,37 MK Firewood 29:20 375 g 4,76 MK 299 g 3,80 MK Briquettes 25:52 436 g 10,26 MK 380 g 8,94 MK Briquettes 22:35 387 g 9,10 MK 350 g 8,23 MK

Charcoal Ceramic Stove Tested Fuels Time Fuel Used Fuel Cost Combusted Fuel Fuel Cost Charcoal 49:02 523 g 17,09 MK 324 g 10,59 MK Charcoal 32:30 416 g 13,59 MK 196 g 6,39 MK Briquettes 27:30 379 g 8,91 MK 303 g 7,12 MK Table 6: Results from the CCT. Ash content for calculating the 'combusted fuel' are assumed as follows: firewood: 2 %; charcoal: 4 %; briquettes: 20 % (same as for the WBT tests).

45 The prices per weight unit for the tested briquettes, produced by the DoE, are assumed to be the same as for WESMA, since the DoE briquettes are not for sale. Prices for firewood and charcoal are based on the prices for this fuel at the market in Area 18, Lilongwe, where the fuel for the tests was bought. The firewood that was used is a softwood type, while the charcoal is made from hardwood.

When using firewood for cooking, it is possible to save much of the unburned fuel from the stove when the food is ready. This fuel can then be used next time a meal is going to be prepared. It is reusable. On the other hand, once you throw charcoal or briquettes into the fire, they can not be saved for use at another moment. This means that the ‘used fuel’ is not necessary the fuel that has been burned, heating the pot, since it includes all the fuel that has been put into the stove. This is an advantage for the firewood and may influence the results, but makes the test more authentic. However, if the input of fuel is not handled with care it can cause unfair comparisons. Therefore combusted fuel has been calculated for each test, with the corresponding fuel cost. This is presented in the above table, in italic.

4.2.3.1 Sources of Error

When the test results were analysed it was discovered that both the firewood stove and the charcoal stove was initially heavier (~50g, ~1 % of total weight) from what was measured during the WBT. After some investigation in the matter it was clear that it was the cold starts of the stoves that caused this difference. The WBT was always made on a hot stove, which apparently is quite lighter than a cold one. The reason might be that the cold stove has absorbed moist during the rains that makes it weigh more initially, before the testing has begun. Since the moist is then vaporized during the test, the stove afterwards will be lighter, and will thus affect the measuring of fuel and ash that is left in the stove. To overcome this problem the values for combusted fuel in the table above are calculated using stove weights taken from the WBT.

4.2.4 Analysis

Although few tests were done on the CCT it is still possible to see some tendencies about the fuel performance on each stove. The following analysis is made by comparing the average values from the results of the CCT, see table 7.

46 Firewood Ceramic Stove Tested Fuels Time Fuel Used Fuel Cost Combusted Fuel Fuel Cost Briquettes 24:14 412 g 9,68 MK 365 g 8,59 MK Firewood 28:10 344 g 4,36 MK 282 g 3,58 MK

Charcoal Ceramic Stove Tested Fuels Time Fuel Used Fuel Cost Combusted Fuel Fuel Cost Briquettes 27:30 379 g 8,91 MK 303 g 7,12 MK Charcoal 40:46 470 g 15,34 MK 260 g 8,49 MK Table 7: Average values of the results gained from the CCT.

Time comparison

The results from the CCT shows that the meals made on briquettes were ready faster than those made on firewood and charcoal respectively, see diagram 8. This indicates that the briquettes burned with a higher power than the other fuels. If the results are compared with the MWBT the results from the CCT are not very surprising. The fuel that made the water boil in the shortest time were the briquettes from the DoE, which are the same briquettes that were used for this CCT. The fact that Mrs Kamanga broke them in smaller parts may also have made the briquettes to burn livelier.

Time line for CCT

1:00:00

0:50:00

0:40:00 Charcoal in CCS Fire Wood i FWS 0:30:00 Briquettes in CCS Briquettes in FWS

0:20:00

0:10:00

0:00:00

e y s d dy fir a of Water rt Re Spinach les rea ta ma S ab si s, oil, onionTomato, Salt et e g N Ve

to cook Nsima, water

Start cook vegetabl t to Star Diagram 8:The charcoal (black line) needed much more time than the other fuels before starting to cook, but once starting the fuel suits just as well for cooking as the other alternatives. The meals made on briquettes are the fastest, although the difference with the firewood is small.

47 Cost comparison

According to the results of the CCT there are big cost-differences for the used fuels. When cooking on the firewood stove the cost for consumed firewood is less than half the cost for briquettes, see diagram 9. The results from the FWS go in line with the results from the MWBT, where firewood proved to be much cheaper than briquettes.

Fuel costs per meal

20,00 MK Total amount of fuel put into stove Combusted fuel

15,00 MK

10,00 MK Price/meal

5,00 MK

0,00 MK Briquettes Firewood Briquettes Charcoal FWS FWS CCS CCS

Diagram 9:Fuel costs for the fuels tested in firewood stove and charcoal stove. The diagram illustrates how that the results is highly dependent on how the amount of used fuel is counted, when the fuel price is compared between charcoal and briquettes (on the CCS). The difference in the diagram shows us that there is a lot more of unburned charcoal than briquettes in the stoves, when the test is finished. The highest columns present the cost when all the fuel put in the stove is counted as used. The darker orange columns describe the cost for the amount of fuel that is actually combusted.

On the other hand, when cooking on the CCS, the used briquettes are more than 40 % cheaper than the charcoal that had to be used. One possible reason why this difference is so big is that the ‘used fuel’ is defined as the fuel that is put into the stove for each test. This does not necessarily mean that this amount of fuel is the same as the amount that is actually combusted during the cooking. The charcoal burns slower than briquettes so while most of the briquettes that is put into the stove gets combusted during the cooking of one meal, much of the charcoal that is used for the same task remain unburned when the meal is ready. The diagram above shows us the difference that these ways of counting fuel cost affect the results.

48 4.3 Household Test

4.3.1 Method

To get opinions about how the briquettes perform as a cooking fuel in a Malawian household, a household test (HHT) was prepared.

A local independent person that usually cooks on firewood and charcoal was asked to use the briquettes for a period of time. After this, an interview was held and a summary of the over all experience can be presented. The expected outcome of the test was to get opinions of difficulties related to the change in the fuel-using habits, as well as how much it costs to use briquettes compared to traditional fuels, and how well the briquettes are suited for being used in the most common traditional cooking stoves.

The person asked to use the briquettes had to be considered trustworthy so that the briquettes would be used for the right purpose (cooking) in the household and that the person had honest opinions as well as accurate results to present after the testing period.

4.3.2 The Household

To get opinions about the briquette from a local energy consumer, Mrs Annie Kamanga, the same woman who performed the CCT, was asked to try using briquettes for a period of time. The household of Mrs Kamanga consists of herself, her husband, three daughters and a son. They live together in a house in Area 18, Lilongwe. The family has no electricity more than a car battery for some small electrical equipment inside. The fuels that are used for cooking are either firewood or charcoal. She uses a 3SF and a CCS, she does not use a FWS and she has never seen it being sold where she lives.

Usually Mrs Kamanga collects her own firewood, but sometimes she does not have time, or she can not find any. When this occurs she will buy firewood or charcoal at the market. Nowadays the charcoal got more expensive and she will not buy it for cooking unless it is for making a barbeque for the chicken. When the charcoal was cheaper she sometimes bought it for cooking, but she still preferred firewood because it takes less time to cook with.

The charcoal she uses she buys at the market in Area 18, in small bags. Today the price of a bag is

49 25 MK, but a month ago it was only 20 MK. One charcoal bag is just enough for cooking one dinner for Mrs Kamanga’s family in a CCS. She buys the firewood at the market in Area 18 as well. She usually buys softwood, for 5 MK per bundle. A dinner in Mrs Kamanga’s family requires 4 bundles of firewood, burned in a open 3 stone fire. Each bundle costs 5 MK, so the total energy cost for a dinner made with firewood is 20 MK.

4.3.3 Realization

Mrs Kamanga was given 19 bundles of briquettes made at WESMA. Each bundle costs 20 MK and contains 4 briquettes of about 200g each. They were delivered to her house in the afternoon Monday the 4th of December and she was told to use the briquettes instead of firewood and charcoal, when cooking in her household, until all the briquettes was finished.

Mrs Kamanga and her family used the briquettes three times a day, for breakfast, lunch and dinner. The last briquettes were consumed by breakfast Monday the 11th of December. During the testing period Mrs Kamanga tried to use the briquettes both in the 3SF and the CCS.

For the breakfast and dinner the meals were prepared for six persons (Mrs Kamanga was chef) while the lunch only was made for four persons (oldest daughter was chef) because Mrs Kamanga and her husband were working outside home.

4.3.4 Results

After the test period was finished, Mrs Kamanga said she preferred using the CCS instead of the 3SF, when burning briquettes. The first time that she tried the briquettes on each stove, she used 6 briquettes for the 3SF and 5 for the CCS. At the end of the testing period the amount of briquettes used were 5 for the 3SF and 4 for the CCS. Between the 4th and the 11th of December about 20 meals were prepared, most of them using the briquettes. The overall experience with the use of briquettes was positive, states Mrs Kamanga. She said that she will definitely consider buying briquettes instead of firewood or charcoal if she finds it in the market, at the current price.

4.3.5 Analysis

Mrs Kamanga seemed to learn quickly how to use the briquettes. It does not seem to take a lot of experience to start using briquettes.

50

The relation between consumed firewood and consumed briquettes are different from the relations in the previous WBT and CCT tests. The reason for this is unclear. The 3SF she used during the HHT might be different from the one used in the other tests. She also cooked a significant larger amount of food in the HHT. These factors could affect the results in different ways.

However the main goal of the test was to evaluate if there would be any practical obstacles concerning the use of briquettes in the household. Mrs Kamanga did not report any practical obstacles and said she would gladly buy them instead of other wood fuel if they could be found in the market and if the price would be the same or lower. Still, the fact remains that the family of Mrs Kamanga usually collects their own firewood free of charge in the forest.

51 52 5 Discussion

5.1 What makes a good briquette?

Most of the briquettes found in Malawi were made from paper with different additives, such as sawdust and the oily sludge from Unilever. Both these additives seemed to increase the utilized energy. However it is unknown how much of these additives that the briquettes samples tested contained. So the question is how much it is possible to add in a briquette and how much it will increase the utilized energy.

The sludge made the briquettes a bit greasy, bad smelling and they produced an unpleasant smoke when combusted. If it would be possible to find cleaner waste oil and add to the briquette it might be possible to reduce these negative properties from the sludge but with remaining good results.

The briquette from WICO only contains sawdust. These briquettes can be seen as an extreme case of the briquettes with sawdust as an additive. The results were good, close to firewood, which indicates that as much sawdust as possible can be used if it is available. The limitation for the simple briquette pressing (manpowered presser) would probably be the fragility of the briquette, since a normal presser does not heat up the raw material that causes the material to stick together. The WICO briquettes were difficult to burn and make the result a bit uncertain. They might give better results if the user gets used to them.

The briquettes from CWAG are the only briquettes, except from WICO, that do not consist of paper. The results from these briquettes were not satisfying in any way more than they look good. It was not possible to burn them in the tests. The tests were performed in a firewood stove, this might be the wrong stove, but the difference cannot be that significant between different stoves. The other difference is the pressure used for producing the briquette. The pressure is a lot less than for the other briquettes in the test that are produced with a mechanical presser. The question is how these briquettes would perform when produced under high pressure. Richard Stanley claims that a WU- presser works well for making briquettes based on crop residues and similar organic materials. If this is true, it creates a good possibility for making the briquetting technology more available to rural areas, which might be something that the DoE could be interested in.

53 5.2 What presser to use?

Higher pressure increases the performance of the briquettes in all the test cases. The question is until how high pressures the performance will increase and to what level it is profitable to increase the pressure, since higher pressure in general demands a more expensive pressing device.

With the screw pressers it is possible to produce very high pressure. The problem with the screw pressers are that they are expensive and each briquette takes quite a long time to press. MIRTDC has made an improved screw presser with a large lever on the screw that makes 6 (12 with half the height) briquettes at the time. However, since most of the time for making a briquette is needed for pounding, investing in an expensive machine that presses quickly makes little sense.

The WU presser presses quickly and makes as high pressure as a screw presser, though the briquettes turns out smaller. This machine can be produced cheaper from wood found locally. One of the most common reasons for stopping the production seems to be the problem of finding spare parts, and with wooden parts the replacement is a lot easier. The WU-presser has probably about the same production-pace as the press from MIRTDC, since the disadvantage of just pressing one cylinder at a time is compensated by the time saved for the compression, which is much faster than the screwing press.

If the powerful MIRTDC press was to be used for only one briquette at the time, the briquette would be pressed with very high pressure. Since no briquettes tested have been exposed to these levels of pressure it will be hard to make any conclusions of how efficient a “high density” briquette would be.

Both the PAMET and the MIRTDC press are most likely too expensive to buy without financial help from some NGO. After having visited the production sites it became obvious that these sorts of agreements between outside investors and local producers are the easiest way to get production started. However, once the agreement has terminated, it is very likely to cause problems in the long term with declining dedication, responsibility and financial support etc. When considering this, the WU-presser is of course a better alternative, although it is doubtful if even this presser would be tempting enough for a local entrepreneur to invest time and money in. In fact none of the producers that was visited during the work with this thesis, regardless of how advanced the equipment was, had started up its production without any outside support. It seems to boil down to the issue of whether the briquettes are really demanded or not by Malawian households, even if it might be very

54 much needed when looking at the country from an energy and environmental perspective.

In the light of the financial problems for starting briquette production it is interesting to discuss the briquetting ‘technology’ that is used by the Nordin family. Although it seems like the briquettes need a few more days to dry than what is stated in the paper written by Stacia Nordin, these hand made briquettes have the advantage of not requiring any capital investment (in contrast to all other technologies presented in this report). Wherever there is waste paper, usually close to offices or public buildings, the hand pressing is a very cheap, practical and accessible way to get briquettes. The disadvantage that limits the usage of this pressing method is the fact that it requires paper as raw material, which might not be the case for screw pressers or the WU-press. However, it seems like the possibility to use other raw materials than paper in the other pressers is not being widely utilized, if looking at the range of briquette samples presented in this thesis.

5.3 The market for briquettes

Briquettes proved to burn fairly well in all the tested stoves. The small air-holes in the ceramic stoves seemed to deliver sufficient with air for the briquettes to perform well and they did not clog as Legacy Foundation stated that they would.

When looking at the analysis of the stove tests it is interesting to see that the FWS is actually the one of the tested stoves that seem to be the most economic one. Using firewood in this stove is much cheaper than the use of charcoal in the CCS. It is then surprising that this stove is hardly used by any Malawians, although it should be possible to buy this one at about the same price as a similar and very much more popular CCS. When discussing the economy of stoves, the firewood fed Rocket Stove is said to be more efficient than the stoves mentioned in this report, although it might require a higher initial investment.

Most Malawian households use a 3-stone fire for cooking and for them the firewood is the most economic fuel to use, assuming the prices and results from the tests in this report. For the households who are able to use either the 3-stone fire or the charcoal stove, as for many families in urban areas, the most economical meal is made using briquettes on the CCS, if looking at the results from our tests. It should be pointed out that the fuel costs may vary a lot between different parts of the country and even within cities.

The majority of fuel consumers in Malawi never buy fuel; they just collect firewood for free in the

55 surroundings. This habit will probably not change until the availability of firewood will be so scarce that the work for finding it is worth avoiding by buying fuel at the market instead. But in households where money is spent on buying fuel for cooking, the briquettes may be able to compete. Most of the households that have got charcoal stoves can be included in this group, since they consume charcoal that is bought at the market. As our test results state, the briquettes are a more economical alternative for people who owns charcoal stoves, most charcoal using households should be able to cut costs by changing to briquettes. This line of arguments is though made in a strict economical perspective. There might of course be social obstacles concerning the selection of fuel that makes the briquettes less attractive, but that is a discussion that goes beyond this thesis.

56 6 Conclusion

A lot of materials seem to be useful for producing briquettes. Paper is a good base material, which is easy to produce briquettes from. Blending the paper with other locally available raw materials makes it possible to produce more briquettes and in many cases make the briquettes more efficient.

A higher pressure increases the efficiency of the briquette. All mechanical pressers available in Malawi today are good enough for pressing the briquettes. The more important factor to consider is the investment cost. The WU-presser is the most suited, since it is cheap and produced from wood that is locally available. It is not necessary to invest in advanced screw pressers. Briquettes made by paper only can even be pressed by hand with a decent result.

The charcoal stove is the most suited stove for the Department of Energy to recommend for the use of briquettes. There are four main reasons for this: • Briquettes combust fairly well in the charcoal stove • The charcoal stoves are wildly spread in Malawi • To find an alternative fuel for charcoal users is a major aim for the PAESP • The price for briquettes can compete with the price for charcoal

Finally it can be stated that the briquette is a good fuel to promote for the Government of Malawi, to decrease the dependence on charcoal, but not firewood which is still easy to find for free.

57 58 7 Proposal to further work

Most of the briquettes available in Malawi today consist of paper. If a large scale production of briquettes will develop, paper will probably be the limiting factor. According to Richard Stanley there is no problem to make good briquettes from crop residues and other biological materials. A greater knowledge about these kinds of briquettes can be useful when promoting the production of briquettes. The information for this exists and there is no need for research to get it.

The stoves tested are constructed for firewood and charcoal. The efficiency and production possibilities of a specific briquette stove could be evaluated. This stove produced to a low price could possibly increase the viability of briquettes as a household fuel in Malawi.

Briquette productions have been started and can be found in a few places in Malawi today. From various reasons the selling of these briquettes does not seem to work. A well laid plan how to build up a self supporting chain of production and selling of briquettes should be done to be able to get sustainable market for briquettes.

The possibilities of starting up the briquette production at WESMA in Lilongwe should be further investigated, since all the necessary facilities are present at this site.

If there will be a market for briquette production, it would be interesting to know in what extent it’s possible to produce briquettes in the matter of availability of raw material. This would prevent that too many production sites are set up, especially concerning paper briquettes.

When making briquettes from crop waste etc. (which should be left for mulching) the nutrients are taken away from the soil. In what extent is it possible to use crop waste before the soil takes any damage?

59 60 8 References

Printed Sources

Bailis, Rob (2004) Controlled Cooking Test. University of California-Berkley (for the Household Energy and Health Programme, Shell Foundation).

CEEDS (2000) Biomass Briquette Extension, Production and Marketing.

Department of Energy Affairs (2003) National Energy Policy. Montfort Press.

Department of Energy Affairs (2006) Promotion of Alternative Energy Poject.

Legacy Foundation (2003) Fuel Briquette Press Kit – A Construction Manual.

Shell Foundation (2004) The Water Boiling Test, Version 1.5. Copy given by ProBEC.

Nordin, Stacia (2005) Low Input Food and Nutrition Security: growing and eating more using less. Malawi: World Food Programme.

Verbal Sources

Binali, Moses. PAMET. Blantyre, 2006-10-23.

Bokosi, George. WESMA. Lilongwe, 2006-11-27

Chembe, Lois. WWF. Cape Maclear, 2006-10-20.

Kamanga, Annie. CCT and HHT participant.

Kunje, Aman. WICO. Blantyre, 2006-10-25.

Michel, Andi. GTZ/ProBEC. Mulanje, 2006-10-24

Mkandawire, Robert. MIRTDC. 2006-10-23.

Stanley, Richard. Legacy Foundation.

Internet Sources

Malawi Socio-Economic Database. Internet: www.maseda.info.

61 62 9 Appendices

Appendix 1 – Water Boiling Test (WBT)

Appendix 2 – Modified Water Boiling Test (MWBT)

Appendix 3 – Controlled Cooking Test (CCT)

Appendix 4 – Equations

Appendix 5 – Recipe for Nsima and Vegetables

Appendix 6 – Equipment for Testing

Appendix 7 – Results from WBT

Appendix 8 – Results from CCT

63

Appendix 1

The Water Boiling Test (WBT) Prepared by Rob Bailis, Damon Ogle and Dean Still with input from Kirk R. Smith and Rufus Edwards Household Energy and Health Programme, Shell Foundation

Introduction This modified version of the well-known Water Boiling Test (WBT) is a simulation of the cooking process that can be performed on most stoves in use throughout the world. While the test is not intended to replace other forms of stove assessment, it is designed to be a simple method by which stoves made in different places and for different cooking applications may be compared by a standardized and replicable protocol. Fuel savings among users who have adopted improved stoves can not be predicted from the results of Water Boiling Tests alone. The Kitchen Performance Test included in this booklet is the only way to demonstrate the field performance of an improved stove. However, the WBT offers a picture of stove performance that can be used during the design process. Data obtained from a few days of testing assists designers to develop better performing stoves, which can then to be tested by cooks in their intended environment. Visser (2003) has shown that by determining thermal efficiency at high and low power, as is done in this version of the WBT, fuel use can be predicted for various cooking tasks. A more complete discussion of the uses of the WBT can be found in Appendix 2. The WBT developed for the Shell HEH program consists of three phases. 1) In the first phase, the tester begins with the stove at room temperature and uses a pre-weighed bundle of wood to boil a measured quantity of water in a standard pot. The tester then replaces the boiled water with a fresh pot of cold water to perform the second phase of the test.

2) In the second phase, water is boiled beginning with a hot stove in order to identify differences in performance between a stove when it is cold and when it is hot.

3) Lastly, the tester again boils a measured amount of water and then, using a pre-weighed bundle of wood, simmers the water at just below boiling for a measured period of time (45 minutes). The third step simulates the long cooking of legumes or pulses that is common throughout much of the world.

This combination of tests is intended to measure the stove’s performance at both high and low power outputs, which are important indicators of the stove’s ability to conserve fuel. Rather than report a single number indicating the thermal efficiency of the stove, which alone can not predict stove performance, this test is designed to yield several numerical indicators including: o time to boil o burning rate o specific fuel consumption o firepower o turn-down ratio (ratio of the stove’s high power output to its low power output)

For more information on each indicator, see Appendix 3, which defines each measure and explains how it is calculated. A direct calculation of thermal efficiency derived from the Water Boiling Test is not a good indicator of the stove’s performance because it rewards the excess production of steam. Under normal cooking conditions, excess steam production wastes energy because it represents energy that is not transferred to the food. Temperatures within the cooking pot do not rise above the boiling point of water regardless of how much steam is produced. Thus, unless steam is required for the cooking process – for example in the steaming of vegetables (FAO, 1983), excess steam production should not be used to increase indicators of stove performance.

Before starting the tests… The following five steps should be completed before beginning the actual tests. 1.) Fuel and water should be procured ahead of time. Ensure that there is an adequate supply of clean water and fuel.. If possible, try to obtain all of the wood from the same source. It should be well-dried and uniform in size. If kindling is to be used to start the fire, it should also be prepared ahead of time and included in the pre-weighed bundles of fuel.

2.) Perform at least one practice test on each type of stove in order to become familiar with the testing procedure and with the characteristics of the stove. This will also provide an indication of how much fuel is required to boil 5 liters of water. As a rough guide, procure at least 20 kg of air-dried fuel for each stove to be tested in order to ensure that there is enough fuel to complete three tests for each stove. Massive multi-pot stoves may require even more fuel.

3.) The practice period is used to determine the local boiling point of water. The local boiling point of water is the point at which the temperature no longer rises, no matter how much heat is applied. This should be determined by the following procedure:

⇒ Put 5 liters of water in the standard pot and bring it to a rolling boil. Make sure that the stove’s power output is high, and the water is fully boiling!

⇒ Using the same thermometer that will be used for testing, measure the boiling temperature when the thermometer is positioned in the center, 5 cm above the pot bottom. You may find that even at full boil, when the temperature no longer increases, it will still oscillate several tenths of a degree above and below the actual boiling point.

⇒ The tester should record the temperature over a five minute period at full boil and note the maximum and minimum temperatures observed during this period. The maximum and minimum temperatures should then be averaged and this result recorded as the “local boiling temperature” on the data and calculation form. (this need only be done once for your test location). (Please see note 2).

4.) A full Water Boiling Test requires at least 15 liters of cool water for each pot being used. If water is not available in adequate supply, water used one day may cooled and reused in the next day’s testing, but do not start any tests with water that is significantly above ambient temperature.

5.) Make sure that there is adequate space and sufficient time to conduct the tests without being disturbed. The tests should occur in a space completely protected from the wind. It will take roughly 2 hours to conduct a full test of a single stove, including fuel preparation and weighing. It will save time if testers prepare enough weighed bundles of fuel to conduct one or more days of testing ahead of time. Beginning of Testing Procedure

Initial steps: to be done once for each series of tests

Equipment used for one Water Boiling Test:

o Scale of at least 6 kg capacity and 1 gram o Heat resistant pad to protect scale accuracy

o Digital Thermometer, accurate to 1/10 of a o Small shovel/spatula to remove charcoal degree, with thermocouple probe suitable from stove for immersion in liquids

o Wood moisture meter (optional) o Tongs for handling charcoal o Timer o Dust pan for transferring charcoal o Standard pot(s) (use other pots if the o Metal tray to hold charcoal for weighing standard pot does not fit the stove)

o Wood fixture for holding TC probe in water o Heat resistant gloves (see diagram in Appendix 1)

o At least 15 liters of clean water for each o 3 bundles of air-dried fuelwood each WBT (may be reused with additions to weighing between 1 and 2 kg. More fuel compensate for spills and evaporation if will be needed if the stove being tested is water is scarce) especially inefficient. A very massive earthen stove may also require more fuel for the cold start phase of the test

1. Fill out the first page of the Data and Calculations form. This includes information about the stove, fuel and test conditions. Number each series of tests for future reference.

2. Measure each of the following parameters. These should be recorded once for each series of tests. Record the measurements on page 1 of the Data and Calculation form.

a) Air temperature.

b) Average dimensions of wood (length x width x height). This is to give a rough idea of the size of fuel used for the test. Testers should use similarly sized wood for every test in order to minimize variation due to fuel differences. Fuel size should be optimized to the stove if the stove design calls for specific sizes of fuel. Otherwise, testers should use sticks 2-5 cm in diameter (see Note 5 for a discussion of the effects of fuel wood variation on stove performance).

c) Wood moisture content (% - wet basis): to be determined 1.) By weighing a sample of wet fuel, drying the sample, and weighing it again or 2.) By using the wood moisture meter included in the testing kit. (See Note 3 and the section on variables and calculations below for full details of defining and measuring moisture content). The Data and Calculation form contains a special worksheet to record and process your measurements. See the form for a more detailed explanation.

d) Dry weight of standard supplied pot without lid. If more than one pot is used, record the dry weight of each pot. If the weights differ, be sure not to confuse the pots as the test proceeds. Do not use pot lids for this, or any other phase of the WBT (see Note 4). The standard 7-liter pot (supplied with the test equipment) should be used wherever possible (see notes). If it is not compatible with the stove, use a pot that is typically used and note its dimensions in the “comments” section of the Data and Calculations worksheet.

e) Weight of container to be used for charcoal.

f) Local boiling point of water determined by using the same digital thermometer and sensor that will be used in the testing (see Note 2).

g) If you have access to a camera (not included in standard kit), photograph the stove. If you don’t have access to a camera, use a tape measure to record the dimensions of the stove and describe it in the space provided.

3. Prepare 3 bundles of fuel wood. These should be pre-weighed: one for each of the three measurement phases of the test. The third bundle should contain about twice as much wood as the first two bundles. In every case, the fuel should be relatively uniform in size and shape: split big pieces of wood and avoid using very small pieces (except for kindling, which should also be prepared in advance if necessary). (see Note 5)

4. Once these parameters have been measured and recorded and the fuel is prepared, proceed with the test. Phase 1: High Power (Cold Start) Data recorded in the remaining phases of the test should be recorded on page two of the Data and Calculation form. 1. Prepare the timer, but do not start it until fire has started.

2. Fill each pot with 5 kg (5 liters) of cold clean water. This amount of water should be determined by placing the pot on the scale and adding water until the total weight of pot and water together is 5 kg more than the weight of the pot alone. Record the weight of pot and water on the Data and Calculations Sheet.

(If the stove can not accommodate the standard pot and the pot that is used can not accommodate 5 kg of water, OR if a multi-pot stove is used with non-standard pots that can not accommodate 5 kg of water, fill each pot ~2/3 full and record the change in procedure in the comment space. Record the weight of the pot(s) with the water on the Data and Calculation Form.) 3. Using the wooden fixtures, place a thermometer in each pot so that water temperature may be measured in the center, 5 cm from the bottom. Make sure the digital thermometer is used in the primary pot. (If there are additional pots, use the extra thermometers provided in the testing kit. Record the initial water temperature in each pot and confirm that it does not vary substantially from the ambient temperature.)

4. The stove should be at room temperature. Start the fire in a reproducible manner according to local practices. Record any starting materials that are used other than the wood from the first bundle of pre-measured wood (e.g. paper or kerosene).

5. Once the fire has caught, record the starting time. Throughout the following “high power” phase of the test, control the fire with the means commonly used locally to bring the first pot rapidly to a boil without being excessively wasteful of fuel.

6. When the water in the first pot reaches the pre-determined local boiling temperature as shown by the digital thermometer, rapidly do the following: a. Record the time at which the water in the primary pot (Pot # 1) first reaches the local boiling point of water. Record this temperature for primary pot as well.

b. Remove all wood from the stove and extinguish the flames (flames can be extinguished by blowing on the ends of the sticks or placing them in a bucket of ash or sand; do not use water – it will affect the weight of the wood). Knock all loose charcoal from the ends of the wood into the container for weighing charcoal.

c. Weigh the unburned wood removed from the stove together with the remaining wood from the pre-weighed bundle. Record result on the Data and Calculation form.

d. For multi-pot stoves, measure the water temperature from each pot (the primary pot should be at the boiling point). Record the temperatures on the Data and Calculation Form.

e. Weigh each pot, with its water. Record weight on Data and Calculation form

f. Extract all remaining charcoal from the stove, place it with the charcoal that was knocked off the sticks and weigh it all. Record the weight of the charcoal + container on the Data and Calculation Form.

Summary

⇒ Make sure that you have recorded time and temperature of the boiling water in the first pot, the amount of wood remaining, the weight of Pot # 1 with the remaining water, and amount of charcoal remaining on the Data and Calculation Form. For multi-pot stoves, be sure that you have recorded the temperature that each additional pot reached when Pot # 1 first came to its full boiling temperature.

⇒ This completes the high power (cold start) phase. Continue directly without pause to the high power (hot start) portion of the test. Do not allow the stove to cool. Phase 2: High Power (Hot Start) The stove will be hot from the first test. Be careful not to burn yourself! 1. Reset the timer, but do not start it until fire has started.

2. Refill the pot with 5 kg of fresh cold water. Weigh pot (with water) and measure the initial water temperature; record both measurements on the Data and Calculations sheet. For multi-pot stoves, fill the additional pots, weigh them and record their weights.

3. Rekindle the fire using kindling and wood from the second pre-weighed bundle designated for this phase of the test.

4. Record the starting time, and bring the first pot rapidly to a boil without being excessively wasteful of fuel using wood from the second pre-weighed bundle.

5. Record the time at which the first pot reaches the local boiling point as indicated on the Data and Calculation form. Record this temperature for the first pot.

6. After reaching the boiling temperature, rapidly do the following:

a. Remove all wood from the stove and knock off any loose charcoal into the charcoal container. Weigh the wood removed from the stove, together with the unused wood from the previously weighed supply. Record result on Data and Calculation form.

b. Record the water temperature from other pots.

c. Weigh each pot, with it’s water and record the weights.

7. Extract all remaining charcoal from the stove and weigh it (including charcoal which was knocked off the sticks). Record the weight of the charcoal plus container on the Data and Calculation form.

Without pause, proceed directly with the simmering test. Phase 3: Low Power (Simmering) This portion of the test is designed to test the ability of the stove to simmer water using a minimal amount of fuel. It begins by repeating the hot-start high power test, but then continues as the tester reduces the flame and “simmers” the water for an additional 45 minutes. For multi-pot stoves, only the primary pot will be assessed for simmering performance (see the discussion of multi-pot stove-testing in Appendix 4) Start of Low Power test 1. Set up the timer, but do not start it until the water has reached the local boiling temperature and the simmering start of the test begins (in step 5 below).

2. Refill the pot with 5 kg of cold water. You do not need to record this weight on the form (the calculations only require weight at the time of first boil and 45 minutes after first boil, as explained below).

3. Rekindle the fire using kindling and wood from the weighed bundle designated for this phase of the test. Replace the pot on the stove.

4. Bring the pot rapidly to a boil without being excessively wasteful of fuel. As soon as local boiling temperature is reached, do the following steps quickly and carefully:

5. Remove the wood from the stove and weigh it along with the wood remaining in the original bundle. Record this weight as the wt. of wood at the start of the simmer phase. Return the wood to the stove. Do not disturb the charcoal remaining in the stove – we will assume that it is similar to the mass of charcoal remaining at the end of the High Power hot-start test.

6. Quickly weigh the water in the primary pot and return it directly to the stove. Record the weight of the pot with water on the Data and Calculation form (this is P1si and goes under the heading “Start: just after Pot # 1 boils”). Replace the thermometer in the pot and continue with the simmer test by reducing the firepower. Keep the water as close to 3 degrees below the established boiling point as possible.

It is acceptable if temperatures vary up and down, but; ⇒ The tester must vigilantly try to keep the simmering water as close as possible to 3 degrees C below the local boiling point. (see note 7)

⇒ The test is invalid if the temperature in the pot drops more than 6°C below the local boiling temperature.

8. For the next 45 minutes maintain the fire at a level that keeps the water temperature as close as possible to 3 degrees below the boiling point. ( see note 6) 9. After 45 minutes rapidly do the following: a. Record the finish time of the test (this should be 45 minutes). Record this and all remaining measurements on the Data and Calculation Form under the heading “Finish: 45 minutes after Pot # 1 boils”.

b. Remove all wood from the stove and knock any loose charcoal into the charcoal container. Weigh the remaining wood, including the unused wood from the pre-weighed bundle.

c. Record the final water temperature on Data and Calculation Form – it should still be roughly 3 °C below the established boiling point.

d. Weigh the pot with the remaining water. Record the weight on the Data and Calculation Form.

e. Extract all remaining charcoal from the stove and weigh it (including charcoal which was knocked off the sticks). Record the weight of pan plus charcoal. This completes the full WBT. The full test should be conducted at least three times for each stove. Analysis Input the results of this WBT into the Data and Calculation software. Output will be viewable in the “Results” worksheet. While a full discussion of statistical theory is beyond the scope of this stove-testing manual, we will rely on some basic ideas of statistical theory to decide whether or not the results of these tests can be used to make claims about the relative performance of different stove models. For a more complete discussion of this subject, see Appendix 8.

Notes on the WBT 1. Pots: The capacity, dimensions and material of the pot have a significant influence on stove performance. Two standard pots, roughly 7 liters in capacity and based on the generic specifications for pots listed in UN Emergency Relief Items, are provided with the test kit. These standard pots should be used whenever possible. The only exception would be when a different pot must be utilized as an integral part of the stove and the stove cannot function properly with the “standard” pot. In this case, record the capacity, dimensions, weight, and material of the non-standard pot. It must be recognized that use of a non-standard pot may bias the results obtained. If a multi-pot stove is used that can accommodate more than two pots, then additional “standard pots” may be procured, or use a locally available equivalent pot. All pots should be scrubbed clean both inside and out, and dried before each test.

2. Boiling point: The local boiling point of water is the point at which the temperature no longer rises, no matter how much heat is applied. This should be determined empirically by the following procedure: Put 5 liters of water in the standard pot and bring it to a boil. Using the same thermometer that will be used for testing, measure the boiling temperature when the thermocouple is positioned in the center, roughly 5 cm above the pot bottom. The tester will find that even at full boil (when new higher temperatures are no longer observed), the temperature will oscillate several tenths of a degree above and below the actual boiling point. The tester should record the temperature over a five-minute period at full boil and note the maximum and minimum temperatures observed during this period. The maximum and minimum temperatures should then be averaged and this result recorded as the “local boiling temperature” on the data and calculation form. (this need only be done once for your test location).

The local boiling temperature is influenced by several factors including altitude, minor inaccuracies in the thermometer, and weather conditions. For these reasons, the local boiling temperature cannot be assumed to be 100° C. For a given altitude h (in meters), the boiling point of water may be estimated by the following formula:

⎛ h ⎞ o Tb =⎜100 − ⎟ C ⎝ 300 ⎠ 3. Moisture content of wood: Even well-dried fuel contains 10-20% water while fresh cut wood may contain more than 50% water by mass (wet basis). Ideally, fuel used for both testing stoves and for cooking by project beneficiaries should be dried as much as local environmental conditions allow. However, dried fuel is not always available and both stove testers and household cooks must use what they can get. In order to control for variations in fuel moisture content, stove testers should measure it and account for it in their stove performance calculations. Thus, there is a space for moisture content to be input in the Data and Calculation form and software.

There are two ways of defining fuel moisture content: on a wet basis and on a dry basis. In the former, the mass of water in the fuel is reported as a percentage of the mass of wet fuel and in the latter case, it is reported as a percentage of the mass of the dry fuel. The calculations for each are shown below followed by a plot showing how both wood moisture on a wet basis and wood mass vary with wood moisture defined on a dry basis for one kg of oven-dry wood. a Unless otherwise specified, we will report wood moisture on a wet basis. The testers should always take care to specify which basis they are using. ()Mass of fuel − (Mass of fuel) MC ()% = wet dry ∗100 and wet ()Mass of fuel wet ()()Mass of fuel wet − Mass of fuel dry MCdry ()% = ∗100 ()Mass of fuel dry

MCdry It is clear than the two moisture contents are related: MCwet = . 1+ MCdry

Wood moisture and mass

80% 4.0

70% 3.5

60% 3.0

50% 2.5

40% 2.0

30% 1.5 Wood moisture MC (%wet-basis) 20% 1.0

content (w et basis) massWet of 1 kg dry wood

10% Mass of w et w ood 0.5 (kg) 0% 0.0 0% 50% 100% 150% 200% 250%

MC (% dry-basis)

Measuring moisture content can be done in two ways. The most precise way is to use the equations listed above by weighing a sample of the fuel as it is used in the tests (in its “wet” condition) and weighing it again after it has been completely dried. A small sample (200-300 g) of the fuel is drawn randomly from the stock of fuel to be used for the tests. The sample is weighed and the mass recorded. The sample is then dried in an oven and weighed again. This may be done at the testing site if an oven is available, or the wet sample may be weighed and then stored carefully and dried later, when an oven is available. Drying should be done slowly and may take as much as 24 hours. When drying, the sample should be left in the oven overnight and then removed from the oven and weighed every two hours on a sensitive scale (±1 g accuracy) until the mass no longer decreases. It is important to ensure that the oven temperature can be carefully controlled so that it doesn’t exceed ~110°C (230°F). If the wood is exposed to temperatures near 200°C (390°F), it will lose matter that is not water, causing an inaccurate measurement of moisture content.

A second way to measure wood moisture is with a wood moisture meter. This device, which is recommended for the stove performance tests, calculates fuel moisture on a dry basis by measuring the conductivity between two sharp probes that are inserted in the wood. This is more convenient than oven-drying because the measurement can be done on site as the fuel is being prepared for testing. The probes should be inserted parallel with the grain of the wood. The recommended device may be adjusted for particular species and calibrated for different ambient temperatures. Wood moisture can vary in a given piece of wood as well as among different pieces from a given bundle. When the meter is used, take three pieces of wood randomly from the bundle and measure each piece in three places. The meter reads up to 40% moisture. If the sample of wood is wetter than 40%, the meter will yield an error. If drier wood is available, wood greater than 40% moisture should not be used for testing.1 This yields nine measurements overall. The moisture of the bundle should then be reported as the average of these nine measurements. Convert this average to a wet basis using the formula (this is done automatically in the computer

MCdry spreadsheet) MCwet = . Record this average in the Data and Calculation sheet. 1+ MCdry

4. Lids: As noted in Baldwin (1987), lids should not be used for the WBT. From Baldwin:

“If a lid is used then the amount of water evaporated and escaping is somewhat dependent on the tightness of the lid’s fit to the pot, and very dependent on the firepower. If the firepower is so low that that the temperature is maintained a few degrees below boiling, effectively no water vapor will escape. If the firepower is high enough so that the water boils, the escaping steam will push the lid open and escape,” (from Chapter 5, note 2, p. 263). The water lost has different effects on each indicator of stove performance. However, since it is difficult to standardize the lid’s “tightness of fit”, even for a standardized pot, we recommend testers not use the lid at all for the WBT. This should have little impact on the high power testing phase – indicators like specific consumption and thermal efficiency are both relatively insensitive to evaporated water. However, the indicators derived from the low power test are more sensitive to the amount of water evaporated. Again, from Baldwin, “By not using a lid, evaporation rates are higher and the stove must be run at a somewhat higher power to maintain the temperature than is the case with a lid” (p. 263). 5. Fuelwood: The type and size of fuelwood can affect the outcome of the stove performance tests. In order to minimize the variation that is potentially introduced by variations in fuel characteristics VITA (1985) recommends taking the following precautions:

o Try to use only wood that has been thoroughly air-dried. Sticks 3-4 cm in diameter may take from 3-8 months to dry fully. Drying is accelerated by ensuring wood is stored in a way that allows air to circulate through it.

o Different sizes of wood have different burning characteristics. While stove users may not have the ability to optimize fuel size, testers should try to use only similar sizes of wood to minimize this source of variation.

6. Power control: Many stoves lack adequate turndown ability. The tester may find that it is impossible to maintain the desired temperature without the fire going out (especially after the initial load of charcoal in the stove has been consumed). If this is the case, the tester should use the minimum amount of wood necessary to keep the fire from dying completely. Water temperatures in this case will be higher than 3° below boiling, but the test is still valid. The tester should not attempt to reduce power by further splitting the wood into smaller diameter pieces.

7. Procedural changes: Measurements of stove performance at both high and low power output can give an indication of how a stove will behave in actual cooking conditions. As far back as 1985, a number of stove experts started to question the wisdom of relying solely on thermal efficiency calculations, and recommended that they be replaced by another standard:

1 40% moisture on a dry basis is equivalent to roughly 29% moisture on a wet basis. …some of the procedures described here differed significantly from what has been recommended in the past. The main difference is in the concept of efficiency used. These standards are based on a broader description and justification of efficiency than Percent Heat Utilized (PHU). They interpret evaporation as a measure of energy wasted, not energy used (VITA, 1985, page ix).

The revised test presented here is based on the procedures proposed by VITA (1985) and Baldwin (1987), but has incorporated minor changes described below:

• Specific Consumption is defined as the ratio of the total amount of wood used to the amount of water “cooked” (Baldwin, 1986), but was modified for multi-pot stoves to reward heat transferred to secondary cooking pots (see Appendix 4).

• It can be difficult to make a smooth transition from high-power to low power tests. Methods used in past testing procedures have suggested extinguishing and weighing wood and charcoal as well as weighing boiling hot water, and rearranging the fire and cooking pot in rapid succession, which is both risky and stressful. This revised version of the WBT follows the suggestions described in VITA Procedural Notes 3 (VITA, 1985), which allows for a more relaxed testing procedure with minimal loss in accuracy.

• During the low power simmer test, the tester is instructed to try to keep the water temperature as close to 3°C below the predetermined boiling point as possible. Different amounts of steam are produced at each degree point below boiling. For this reason, it is necessary to minimize the variation in temperature to ensure that tests are comparable.

• Hot and cold starts are incorporated in the high power phase of the test in order to account for the differential performance of stoves that are kept hot throughout the day. This is particularly important for massive stoves, whose performance may vary significantly between cold and hot starting conditions.

• Simmering occurs for 45 minutes rather than 30, (as suggested in VITA, 1985) because the large amount of charcoal some stoves create during the high power phase can skew the results if the simmering test is too short. The presence of charcoal helps to keep small amounts of wood burning. A 45-minute simmering period is long enough for the stove at low power to establish a burning equilibrium, as excess charcoal made at high power is normally consumed within 30 minutes. REFERENCES Baldwin, S. F. (1986). Biomass Stoves: Engineering Design, Development, and Dissemination. Princeton, NJ, Center for Energy and Environmental Studies: 287. Edwards, R. D., K. R. Smith, et al. (2004). "Implications of changes in household stoves and fuel use in China." Energy Policy 32(3): 395-411. FAO (1983). Wood fuel surveys. Rome, UN Food and Agriculture Organization. FAO (1993a). Chinese Fuel-Saving Stoves: A Compendium. Bangkok, Regional Wood Energy Development Program (RWEDP): 57. FAO (1993b). Indian Improved Cookstoves: A Compendium. Bangkok, Regional Wood Energy Development Program (RWEDP): 109. Kishore, V. V. N. and P. V. Ramana (2002). "Improved cookstoves in rural India: how improved are they?: A critique of the perceived benefits from the National Programme on Improved Chulhas (NPIC)." Energy 27(1): 47-63. Smith, K. R., G. Shuhua, et al. (1993). "One Hundred Million Improved Cookstoves in China: How was it done?" World Development 21(6): 941-961. VITA (1985). Testing The Efficiency Of Wood-Burning Cookstoves: Provisional International Standards. Arlington, VA, Volunteers in Technical Assistance: 76. World Bank (2002). India: Household Energy, Indoor Air Pollution, and Health. Washington DC, World Bank Energy Sector Management Assistance Program (ESMAP): 148. Zhang, J., K. R. Smith, et al. (2000). "Greenhouse Gases and Other Airborne Pollutants from Household Stoves in China: A database for emission factors." Atmospheric Environment 34: 4537-4549.

Appendix 1 Diagram showing wooden fixture holding TC probe in pot. The dimensions are not critical, but the fixture should be made so that the TC probe fits into it tightly and the fixture itself fits securely on the pot. TC probe wire (leading to TC probe digital thermometer)

Wooden probe holder

Standard 7 liter pot

~5 cm

Side view

Slots cut out to tightly accommodate pot rim.

Top view

Hole drilled in center of fixture (diameter should be just large enough to fit TC probe tightly

Bottom view

Slots are cut out to tightly accommodate pot rim.

Appendix 2

Stove Performance Testing Tests of stove performance range from lab-based water boiling and cooking tests to qualitative and quantitative surveys of stove users in the field. There are advantages and disadvantages to both types of tests. Lab-based tests are more appropriate at the early stages of stove development in order to compare various technical aspects of stove design. For example, Baldwin recommends lab-based tests for comparing and optimizing different dimensions and other design details of the stove. Lab- based tests are also more appropriate when comparing stoves that are used in different regions of the world. There is a great amount of variation in cooking practices, fuels, and household environments throughout the world’s developing regions that makes direct comparisons of actual stoves in people’s kitchen very difficult. Lab-based tests In order to accommodate the many aspects of stove performance testing that eliminate the variability in factors that may affect stove performance other than the physical characteristics of the stove itself. In order to accommodate the many aspects of stove performance testing that designers of improved cookstoves face, the protocols described in this manual include procedures for two types of lab- based tests as well as a field test. The lab tests include a modified version of VITA’s Water Boiling Test (WBT) and as well as a Controlled Cooking Test (CCT). The field test includes two qualitative surveys: the first helps project designers to assess household cooking practices prior to the introduction of the improved stove and the other provides them with follow-up data 3-6 months after the stove has been introduced to the family. The field test also includes a procedure to compare fuel consumption in households using different types of stoves. This test is critical if project designers wish to make justifiable claims about real impacts on fuel consumption resulting from the stoves that they are promoting. Such claims can not be based on lab-based tests alone While lab-based tests allow stove developers to differentiate between well-designed and poorly- designed stoves, they give little indication of how the stoves are actually used by the people who are targeted by stove projects. In order to know if stove projects are having the desired impact (whether it is fuel conservation, smoke reduction, or both), the stoves must be measured under real conditions of use.

Two major stove programs and their use of stove performance tests In addition to the tests introduced by VITA and elaborated by Baldwin, other large-scale efforts have been conducted to assess the performance of improved stoves and stove programs. Two of the most notable efforts are those that have been conducted over the past 20 years in India and China. Taken together, these programs represent the vast majority of improved stoves introduced globally: with well over 200 million stoves disseminated between them. These programs have undergone numerous changes since their inception in the early 1980s. A full review of these programs is beyond the scope of this document, but a brief review of each is given below, with attention to the stove performance monitoring methods that have been used. References are also provided for further reading.

India’s National Improved Chula Program (NPIC) This program, which has been underway for nearly two decades, had five stated objectives: o to conserve and optimize the use of fuelwood, especially in the rural and semi-urban areas

o to help alleviate deforestation

o to reduce the drudgery associated with cooking, especially on women, and the health hazards caused by smoke and heat exposure in the kitchen o to bring about improvements in household sanitation and general living conditions

o employment generation in rural areas

By 1999, NPIC activities had disseminated over 28 million stoves (Kishore and Ramana, 2002). The graph below illustrates the number of stoves disseminated during the first 15 years of project activity. Stoves installed in India's NPIC Program: 1983-1998

3.5 35 installed (millions) 3 Cumulative (millions) 30

2.5 25

2 20

1.5 15

1 10 No.stoves of annually installed 0.5 5 Cumulative no. of stoves installed

0 0

989 992 1984 1985 1986 1987 1988 1 1990 1991 1 1993 1994 1995 1996 1997 1998 Based on data from (Kishore and Ramana, 2002)

Despite the impressive numbers, NPIC is not considered an unqualified success. Firstly, the cumulative data hide the fact that most stoves have a limited lifetime – typically no more than two years - so that the total number of stoves in use in 1998 was actually a small fraction of the cumulative number shown in the graph. Several other factors also contributed to NPIC’s problems. These are summed up well in the following passage, which is taken from a recent World Bank report on India’s experience with cookstoves in the context of indoor air pollution reduction: In the early programs it was assumed that if improved stoves were presented to people, they would be quickly adopted and the intervention would lead to self- sustaining programs. This often did not happen for several reasons. One reason was that the energy efficiencies achieved in laboratories did not translate into similar efficiency gains in rural homes. Another reason lay in an obvious failure to identify the market for improved stoves; for example, some programs introduced stoves into regions where people purchased neither their traditional stoves nor fuelwood, thus having little appreciation of efficiency gains. The health benefits of the improved stoves were not well advertised. Finally, the price of an improved stove was a significant barrier to adoption, especially in areas where there was very little cash outlay for stoves or fuel (World Bank, 2002).

While each of these problems presents significant obstacles for stove project designers, the problem of linking lab-based efficiency to actual fuel consumption in rural homes is of greatest concern to the ideas presented in this document. We will discuss this issue in much more detail below. NPIC reports claimed that each stove reduced fuelwood consumption by 30-40% relative to a traditional chulha, which represents roughly 700 kg of fuelwood family per year (FAO, 1993b). The World Bank reports more modest savings of 19-23% relative to traditional stoves. Kishore and Ramana also report smaller improvements. They cite one study that found savings of about 35 kg fuelwood per year and another that actually found a net increase in fuel consumption. The results of the latter study are shown in Table 1 below. These data are the result of fuel consumption surveys that followed VITA’s kitchen performance test protocol (VITA, 1985; Kishore and Ramana, 2002). The results show that in many cases, NPIC stoves consumed more fuel than the traditional chulas that they were meant to replace. Kishore and Ramana’s review of the NPIC stove performance relied on field tests, but the NPIC program itself based its assessments of stove performance primarily on “thermal efficiency” tests. Thermal efficiency was defined as “the ratio of heat actually utilized to the heat theoretically produced by complete combustion of a given quantity of fuel,” (FAO, 1993b, p. 96). In order to qualify for inclusion in NPIC, improved chulhas were required to have a minimum efficiency of 20% for fixed mud stoves and 25% for portable metal stoves. The lab-based tests that were used differ substantially from the WBT designed by VITA and presented below with slight modifications. See (FAO, 1993b, Annex 2) for a full description of the procedure. Table 1: Comparison of fuel consumption in Improved and Traditional Chulhas in 3 Indian States Improved chulha Traditional chulha

No. of Wood consumption No. of Wood consumption Percent savings of improved households kg/day/stovea households kg/day/stovea chulha compared to traditional chulha

Tamil Nadu Manachai 14 5.99 8 5.34 -12 Muthupattai 10 7.27 10 4.77 -62 Rajasthan Motuka 11 7.2 12 5.91 -22 West Bengal Golti 14 9.34 14 11.65 19.8 Iswarigacha 15 7.34 13 7.83 6.3 Average -14 From (Kishore and Ramana, 2002) a Assuming six persons per family, one chulha per family and a calorific value of 17.6 MJ/kg for wood.

NPIC continued until 2002, when the Indian government devolved authority to the state level so that individual states in India are now responsible for implementing improved stove programs. Additionally, in recent years, many non-governmental agencies have become involved in stove development and dissemination, both in partnership with and independent of, state-run projects.

The Chinese National Improved Stove Program (CNISP) China has undertaken the most extensive improved stove program in the world. Between its inception in 1983 and 1998, roughly 185 million stoves were disseminated in China (Edwards, Smith et al., 2004). Like India’s NPIC, CNISP was originally intended to conserve biomass fuels and reduce the time and effort household members had to devote to meeting their energy needs.2 Because of its sheer scale and the extent of social and economic changes China has undergone during the time that the program has been in place, the impact that CNISP has had is difficult to assess independent of other changes in rural China. Biomass fuel consumption in some rural areas has decreased, but this may be attributed to fuel switching rather than to gains in energy efficiency. Many rural families have begun using coal, electricity, or other fuels in addition to biofuels, which would likely result in decreased biofuel consumption regardless of the type of stove in use.

2 One expert has pointed out that in addition to the goals of biofuel conservation and reduced workload on the rural household, China’s stove program was also a state-driven effort to modernize biofuel consumption in order to reduce and/or delay growth in demand for fossil fuels among China’s massive rural population (Smith, Shuhua et al., 1993). As Smith indicated over 10 years ago, information about quantitative aspects of China’s improved stoves is not widely available (Smith, Shuhua et al., 1993). Data about the performance of stoves disseminated through CNISP are not widely published. Zhang and colleagues tested emissions and efficiencies of 28 fuel-stove combinations in China. They report average lab-based efficiencies derived from a 3 repetitions of a modified version of VITA’s WBT (Zhang, Smith et al., 2000). A sample of their results are reproduced below in the graph below showing four pairs of traditional and improved Chinese stoves each using a different solid fuel: coal, wood, maize stalks and wheat stalks. Three popular fossil fuel options are also included for comparison. Notice the efficiency determined by the modified WBT is higher in improved stoves. However, the biomass stoves tested are far less efficient than liquid and gaseous fossil fuels. Another interesting result concerns health impact of stoves; the biofuels used in improved stoves appear to result in higher emissions of particulate matter (measured as Total Suspended Particulates or TSP). All of these stoves tested have chimneys, so the increase in TSP is not necessarily a cause for concern as long as the chimneys are functioning well, however it does indicate that higher overall efficiency in these improved stoves has likely been achieved at the epense of combustion efficiency. Chinese Stoves: Efficiency (WBT) and TSP Emissions

50 10.0

45 Stove efficiency 9.0 TSP emissions 40 8.0

35 7.0

30 6.0

25 5.0

20 4.0 Efficiency (%) Efficiency

15 3.0 emissionsTSP (g/MJ delivered)

10 2.0

5 1.0

0 0.0 Coal LPG Kerosene Traditional Improved Traditional Improved Traditional Improved Traditional Improved Gas Coal honeycomb Wood Maize stalk Wheat stalk Based on data from (Zhang, Smith et al., 2000)

However, it is difficult to predict fuel consumption by rural families under real cooking conditions solely through the results of energy efficiency determined by WBTs. Unfortunately, few reports of actual fuel consumption by improved stoves in China have been published outside of China. The authors would welcome any information about field performance of improved stoves in China. Appendix 3

An explanation of the calculations used in the WBT The WBT consists of three phases: a high-power phase with a cold start, a high power phase with a hot start, and a low power (simmer) phase. Each phase involves a series of measurements and calculations. The calculations for the one-pot test are described below. For stoves that accommodate more than one pot, the calculations will be adjusted to account for each pot. These adjustments are explained below. Variables that are constant throughout each phase of the test HH Gross calorific value (dry wood) (MJ/kg) V

LHV Net calorific value (dry wood) (MJ/kg)

m Wood moisture content (% - wet basis)

ceff Effective calorific value (accounting for moisture content of wood)

P Dry weight of empty Pot (grams)

k Weight of empty container for char (grams)

Tb Local boiling point of water (deg C)

Explanations of Variables HHV – Higher heating value (also called gross calorific value). This is the theoretical maximum amount of energy that can be extracted from the combustion of the moisture-free fuel if it is completely combusted and the combustion products are cooled to room temperature such that the water produced by the reaction of the fuel-bound hydrogen is condensed to the liquid phase. LHV – Lower heating value (also called net heating value). This is the theoretical maximum amount of energy that can be extracted from the combustion of the moisture-free fuel if it is completely combusted and the combustion products are cooled to room temperature but the water produced by the reaction of the fuel-bound hydrogen remains in the gas phase. For woodfuels, LHV typically differs from HHV by 1.32 MJ/kg.3 m – This is the % wood moisture content on a wet basis, defined by the following formula: ()mass of wet fuel - (mass of dry fuel) m = ∗ 100 mass of wet fuel This can be determined gravimetrically (by weighing a sample of wet fuel, drying the sample, and weighing it again) or through the use of a wood moisture meter (see description of test procedure). If the Delmhorst J-2000 moisture meter is used in this test to measure wood moisture content, be aware that it provides moisture content on a dry basis. In order to use ‘m’ in the following analysis, the output of the instrument must be converted to moisture content on a wet basis. Dry basis must be converted to wet basis using the following equation:

3 Dry wood typically consists of 6% hydrogen by mass. Thus, one kg of dry wood contains 60 g of hydrogen, which reacts to form 540 g of H2O. The difference in enthalpy between the liquid and gaseous phases of 540 g of water at room temperature is roughly 1.32 MJ, thus, for a typical sample of moisture-free wood, HHV and LHV differ by 1.32 MJ. In Baldwin (1986), the difference between HHV and LHV is given as 1.39 MJ/kg, but this applies to water vapor at 100 ºC, which is not typically how LHV is defined (Baldwin, 1986, p. 55). MCdry MCwet = 1+ MCdry ceff – This is the effective calorific value of the fuel, with takes account of the energy required to heat and evaporate the moisture present. This is calculated in the following way: LHV - m ∗ (80 ∗ 4.186 + 2260) c = eff mass of wet fuel where 80ºC represents the typical change from ambient temperatures to the boiling point of water, 4.186 kJ/(kg•ºC) is the specific heat capacity of water, and 2260 kJ/kg is the energy required to evaporate one kilogram of water. The graph below shows ceff as a function of wood moisture content (wet basis) assuming an HHV of 20,000 kJ/kg (LHV of 18,680 kJ/kg), which is a typical value for hardwoods. Note that at 50% moisture, which is not uncommon for freshly cut (green) wood in moist climates, the effective energy content of the fuel is reduced by more than half.

Effective calorific value

20,000 18,000 16,000 14,000 12,000 10,000 8,000

kJ/kg-wet fuel 6,000 4,000 2,000 - 0% 20% 40% 60% 80% Wood moisture (% wet basis)

P – This is the weight of the empty pot. For multi-pot stoves, this is followed by an index number 1 – 4. k – This is simply the weight of the charcoal container that will be used to hold the char when it is removed from the stove and weighed.

Tb – This is the local boiling point of water, which should be determined empirically in order to account for variations as a result of altitude. 1. High power test (cold start)

Variables that are directly measured Variables that are calculated

fci Weight of fuel before test (grams) fcm Wood consumed, moist (grams)

Pci Weight of Pot with water before test (grams) Δcc Change in char during test phase (grams)

Tci Water temperature before test (ºC) fcd Equivalent dry wood consumed (grams)

tci Time at start of test (min) wcv Water vaporized (grams)

fcf Weight of wood after test (grams) wcr Water remaining at end of test (grams)

cc Weight of charcoal and container after test Δtc Duration of phase (min) (grams)

Pcf Weight of Pot with water after test (grams) hc Thermal efficiency

Tcf Water temperature after test (ºC) rcb Burning rate (grams/min)

tcf Time at end of test (min) SCc Specific fuel consumption (grams wood/grams water)

T SC h Temp-corrected specific consumption (grams wood/grams water)

FPc Firepower (W)

Explanations of Calculations fcm - Wood consumed (moist): This is the mass of wood that was used to bring the water to a boil found by taking the difference of the pre-weighed bundle of wood and the wood remaining at the end of the test phase:

fcm = fcf – fci

Δcc - Net change in char during test phase: This is the mass of char created during the test found by removing the char from the stove at the end of the test phase. Because it is very hot, the char will be placed in an empty pre-weighed container of mass k (to be supplied by testers) and weighing the char with the container, then subtracting the two masses.

Δcc = cc – k fcd - Equivalent dry wood consumed: This is a calculation that adjusts the amount of wood that was burned in order to account for two factors: (1) the energy that was needed to remove the moisture in the wood and (2) the amount of char remaining unburned. The calculation is done in the following way:

fcd = fcm ∗ (1− (1.12 ∗ m)) − 1.5 ∗ Δc c The factor of 1− ()1.12 ∗ m adjusts the mass of wood burned by the amount of wood required to heat and evaporate m ∗ fcm grams of water. It takes roughly 2260 kJ to evaporate a kilogram of water, which is roughly 12% of the calorific value of dry wood. Thus if wood consists of m% moisture, the mass of wood that can effectively heat the pot of water is reduced by roughly 1− ()1.12 ∗ m because the water must be boiled away (see (Baldwin, 1986) for further discussion).

The factor of 1.5 ∗ Δc c accounts for the wood converted into unburned char. Char has roughly 150% the calorific content of wood, thus the amount of wood heating the pot of water should be adjusted by 1.5 ∗ Δc c to account for the remaining char. Note, in the simmer phase it is possible that there will be a net loss in the amount of char before and after the test, in which case Δc is negative and the equivalent dry wood increases rather than decreases. wcv - Water vaporized: This is a measure of the amount of water lost through evaporation during the test. It is calculated by simple subtraction of initial weight of pot and water minus final weight of pot and water.

w cv = Pci − Pcf wcr - Water remaining at end of test: This is a measure of the amount of water heated to boiling. It is calculated by simple subtraction of final weight of pot and water minus the weight of the pot.

w cr = Pcf − P

Δtc – Duration of phase: This is simply the time taken to perform the test. It is a simple clock difference:

Δtc = tcf – tci hc - Thermal efficiency: This is a ratio of the work done by heating and evaporating water to the energy consumed by burning wood. It is calculated in the following way.

4.186 ∗ (Pci − P)∗ (Tcf − Tci ) + 2260 ∗ (w cv ) hc = fcd ∗ LHV In this calculation, the work done by heating water is determined by adding two quantities: (1) the product of the mass of water in the pot, (Pci – P), the specific heat of water (4.186 J/gºC), and the change in water temperature (Tcf – Tci) and (2) the product of the amount of water evaporated from the pot and the latent heat of evaporation of water (2260 J/g). The denominator (bottom of the ratio) is determined by taking the product of the dry-wood equivalent consumed during this phase of the test and the LHV. rcb - Burning rate: This is a measure of the rate of wood consumption while bringing water to a boil. It is calculated by dividing the equivalent dry wood consumed by the time of the test.

fcd rcb = tci − t cf SCc - Specific fuel consumption: Specific consumption can be defined for any number of cooking tasks and should be considered “the fuelwood required to produce a unit output” whether the output is boiled water, cooked beans, or loaves of bread. In the case of the cold-start high-power WBT, it is a measure of the amount of wood required to produce one liter (or kilo) of boiling water starting with cold stove. It is calculated in this way:

fcd SCc = Pcf − P T SC c – Temperature corrected specific fuel consumption: This corrects specific consumption to account for differences in initial water temperatures. This facilitates comparison of stoves tested on different days or in different environmental conditions. The correction is a simple factor that “normalizes” the temperature change observed in test conditions to a “standard” temperature change of 75 ºC (from 25 to 100). It is calculated in the following way.

T fcd 75 SC c = ∗ Pcf − P Tcf − Tci FPc – Firepower: This is a ratio of the wood energy consumed by the stove per unit time. It tells the average power output of the stove (in Watts) during the high-power test.

fcd ∗LHV FPc = 60 ∗ ()tci − tcf Note, by using fcd in this calculation, we have accounted for both the remaining char and the wood moisture content. High power test (hot start) In this test, measurements and calculations are identical to the cold start test. Simply substitute the subscript ‘h’ for the subscript ‘c’ in each variable as in the table below. Variables that are directly measured

fhi Weight of fuel before test (grams)

Phi Weight of Pot with water before test (grams)

Thi Water temperature before test (ºC)

thi Time at start of test (min)

fhf Weight of wood after test (grams)

ch Weight of charcoal and container after test (grams)

Phf Weight of Pot with water after test (grams)

Thf Water temperature after test (ºC)

thf Time at end of test (min)

Variables that are calculated

fhm Wood consumed, moist (grams) fhm = fhf – fhi

Net change in char during test phase Δc Δc = c – k h (grams) h h

fhd Equivalent dry wood consumed (grams) fhd = fhm ∗ (1− (1.12 ∗ m)) − 1.5 ∗ Δch

whv Water vaporized (grams) w hv = Phi − Phf

whr Water remaining at end of test (grams) w hr = Phf − P

Δth Duration of phase (min) Δth = thf – thi

4.186 ∗ (Phi − P)∗ (Thf − Thi )()+ 2260 ∗ w hv hh Thermal efficiency hh = fhd ∗ LHV

fhd rhb Burning rate (grams/min) rhb = thi − thf

Specific fuel consumption (grams fhd SCh SCh = wood/grams water) Phf − P

T Temp-corrected specific consumption T fhd 75 SC h SC h = ∗ (grams wood/grams water) Phf − P Thf − Thi

fhd ∗LHV FPh Firepower (W) FPh = 60 ∗ ()thi − thf

Low power (simmering) test In this test, the initial measurements are the same as in the high power tests, however the goal of this test is to maintain water at a high temperature with minimal power output from the stove. Since the goal differs, the interpretations of the calculations also differ from those of the high power phases. In addition, one important assumption is made using data from the hot start high power test and one additional calculation is performed that does not appear in the high power tests. These are both explained below. The assumption made in this test is based on the amount of char present when the water first boils. The low power phase starts by repeating the high power hot start test, however when the water comes to a boil, it is quickly weighed without disturbing the char and then the fire is tended to maintain the water within a few degrees of boiling for 45 minutes There will be char remaining in the stove from the wood that was used to bring the water to a boil. Removing that char from the stove, weighing it and relighting it disturbs the fire and may result in the water temperature dropping too far below boiling. Thus, the recommended procedure is to assume that the char present at the start of the simmer phase is the same as the char that was measured after the high power hot start test (Δch). While this is not entirely accurate, the error introduced by this assumption should be minimal – especially if the tester(s) followed an identical procedure in bringing the water to a boil.

Variables that are directly measured

fsi Weight of unused fuel when the water first boils (grams)

Psi Weight of Pot with water when the water first boils (grams)

Tsi Water temperature at boiling (Tsi = Tb) (ºC)

tsi Time at start of simmer phase test (min)

fsf Weight of unburned wood remaining after test (grams)

cs Weight of charcoal and container after test (grams)

Psf Weight of Pot with water after test (grams)

Tsf Water temperature at end of test (ºC)

tsf Time at end of test (min)

Variables that are calculated

fsm Wood consumed, moist (grams) fsm = fsf – fsi

Net change in char during test phase Δc Δc = c – k – Δc s (grams) s s h

fsd Equivalent dry wood consumed (grams) fsd = fsm ∗ (1− (1.12 ∗ m)) − 1.5 ∗ Δc s

wsv Water vaporized (grams) wsv = Psi − Psf

wsr Water remaining at end of test (grams) wsr = Psf − P

Δts Duration of phase (min) Δts = tsf – tsi

4.186 ∗ (Psi − P)∗ (Tsf − Tsi )()+ 2260 ∗ w sv hs Thermal efficiency hs = fsd ∗ LHV

fsd rsb Burning rate (grams/min) rsb = t si − t sf

Specific fuel consumption (grams fsd SCs SCs = wood/grams water) Psf − P

fsd ∗LHV FPs Firepower (W) FPs = 60 ∗ ()t si − t sf

FP TDR Turn-down ratio TDR = h FPs

There is no temp-corrected specific consumption in the simmer phase because the test starts at Tb and the change in temperature should be limited to a few degrees. It is important to remember that the goal of this part of the test is to maintain the water at a temperature just under boiling, and one should interpret the results accordingly. Whereas the specific consumption in the high power tests (SCc and SCh) indicated the mass of fuel required to produce one liter (or kilogram) of boiling water, the specific consumption in the simmer phase (SCs) indicates the mass of wood required to maintain each liter (or kilo) of water three degrees below boiling temperature. These are not directly comparable, but rather tell two different measures of stove performance. The same is true for other indicators, like burning rate and firepower. It is also important to acknowledge that over-reliance on thermal efficiency can lead to misleading results, particularly in the simmer phase. Because thermal efficiency accounts for sensible heat as well as evaporative losses, it rewards for the generation of steam. In most cooking conditions, excess steam production does not decrease cooking time, as the temperature in the pot is fixed at the boiling point. Thus, producing excess steam, while it does reflect wood energy transferred to the cooking pot, is not necessarily a good indicator of stove performance. As we state elsewhere, we wish to de-emphasize the role that thermal efficiency plays in discussions of stove performance and stress other, more informative indicators such as the burning rate and specific consumption at high and low power, and the turn-down ratio, which indicates the degree to which power output from the stove can be controlled by the user. Appendix 4

WBT Modifications for multi-pot stoves Some stoves are designed to cook with more than one pot. If this is the case, the tester should use the number of pots that the stove can accommodate (the Data and Calculation Form has space for up to four pots). The testing procedure will remain the same except for the additional measurements of weight and temperature. In addition, the calculations will be modified slightly. To test stoves that accommodate multiple cooking pots, the data forms have been modified (see the file Template_for_multipot_WBT.xls). The modifications allow for weight and temperature of up to four pots to be recorded. The calculations are also modified to take these additional measurements into account. The modifications are explained below.

High-power tests: In order to closely mirror the single pot test and ensure that the task can be completed in a reasonable amount of time, the high power tests are stopped when the primary pot (the pot closest to the source of heat) comes to a boil. The indicators of stove performance account for the water heated in the additional pots. To do so they are modified in the following way. Calculations that are modified to account for multiple pots in the high power tests*

fcm Wood consumed, moist (grams) Same as for single-pot stove Net change in char during test phase Δc Same as for single-pot stove c (grams) Equivalent dry wood consumed f Same as for single-pot stove cd (grams)

4 wcv Water vaporized (grams) w cv = ∑()Pjci − Pjcf j=1

4 “Boiled” water remaining at end of ⎛ ⎛ Tj − Tj ⎞⎞ w w = ⎜()Pj − Pj ∗ ⎜ cf ci ⎟⎟ cr test (grams) cr ∑⎜ cf ⎜ ⎟⎟ j=1 ⎝ ⎝ Tb − Tjci ⎠⎠

Δtc Duration of phase (min) Same as for single-pot stove ⎡ 4 ⎤ ⎢4.186 ∗ ∑()()Pjci − Pj ∗ Tjcf − Tjci ⎥ + 2260 ∗ ()w cv hc Thermal efficiency ⎣⎢ j=1 ⎦⎥ hc = fcd ∗ LHV

rcb Burning rate (grams/min) Same as for single-pot stove f SC = cd Specific fuel consumption (grams c SC 4 ⎡ ⎛ Tj − Tj ⎞⎤ c wood/grams water) ⎜ cf ci ⎟ ∑ ⎢()Pjcf − Pj ∗ ⎜ ⎟⎥ j=1 ⎣⎢ ⎝ Tb − Tjci ⎠⎦⎥ Temp-corrected specific T T 75 SC c consumption (grams wood/grams SC c = SCc ∗ water) Tcf − Tci

FPc Firepower (W) Same as for single-pot stove * These calculations use the subscript-c for the cold-start test, however the modified hot-start calculations are identical. In each case, j is an index of each pot (1-4) ⎛ Tj − Tj ⎞ ⎜ cf ci ⎟ The factor ⎜ ⎟ is used to “discount” the water heated in additional pots that does not come to a full boil. ⎝ Tb − Tjci ⎠ For example, when calculating specific consumption, which, in this test, measures the amount of wood required to boil a unit amount of water, we want to give credit for the water heated in other pots, although it was not boiled. Since the energy (Q) required to bring water to a boil is a roughly linear function of the temperature change ()Q ∝ ΔT we discount the water that was not boiled by a factor that varies between zero and one, reflecting the fraction of sensible heat absorbed by the water relative to the heat required to boil it.

Low-power test: In the low power test it is more difficult to incorporate the output from additional cooking pots. For this reason, multi-pot stoves may appear to be at a disadvantage in this part of the test, which assesses the ability of the stove to maintain a pot of water just below the boiling temperature. In lowering the power delivered to the primary cooking pot, the stove will probably not be able to deliver much heat to secondary pots. Fluctuations in temperature in the other pots will greatly complicate the assessment, thus they will be ignored. The Stove Performance Test used in assessing improved stoves in China adopts a similar procedure (FAO, 1993a). Of course, we acknowledge the strengths of well-designed multi-pot stoves, which lies in the stoves’ ability to provide high power to the primary cooking pot, while simultaneously providing low power to an additional pot (or pots). However, this test is designed to only bring the water in the primary pot to boiling temperatures and the stove performance indicators calculated from the results of the simmer test will only rely on the measurements taken from the primary pot. While this may not capture all of the strengths of the multi-pot stove, those strengths should be captured in the results of the high power test, as well as in the controlled cooking test and kitchen performance (field) tests, which also must be conducted to fully assess stove performance.

Appendix 8

Aspects of statistics to think about when conducting the WBT At least three tests should be performed on each stove. If two models of stove are being compared, the testers should pay attention to the statistical significance of the results of the series of tests. For example, if testers want to compare an indicator of stove performance like specific fuel consumption, it is not possible to say conclusively that one stove is better than another with 100% surety. They can only declare one stove better than another with a certain level of confidence. This level depends on several factors, including the difference in the average specific consumption of each stove, the variability of the test results, and the number of tests that were performed. While a full discussion of statistical theory is beyond the scope of this stove-testing manual, we will rely on some basic ideas of statistical theory to decide whether or not the results of these tests can be used to make claims about the relative performance of different stove models. For example, Table 2 shows data from a series of cold-start water boiling tests conducted at the Aprovecho Institute on two different single-pot woodstoves. Each stove was tested three times. From the data, it is clear that the Stove-2 performs much better than Stove-1 in most indicators of stove performance. Notice however, that some indicators of stove performance, namely burning rate and firepower, show no difference between stoves. This indicates the importance of considering a multiple indicators when defining stove performance. Table 2: Results of three high-power cold start Water Boiling Tests on two different stoves Stove-1 Stove-2 Statistics

% difference Significant units Mean SD CoV Mean SD CoV between Stove-1 T-test with 95% and Stove-2 confidence?

4 13 Wood consumed g 837 34 468 60 -44% 7.55 YES % % 7 10 Time to boil 5 liters of water min 36 3 20 2 -44% 6.89 YES % % 4 14 Thermal efficiency -- 0.19 0.01 0.28 0.04 49% -3.30 YES % % 3 18 Rate of wood consumption g/min 23 1 24 4 1% -0.04 NO % % 5 12 Specific fuel consumption g/liter 155 8 91 11 -41% 6.77 YES % % 3 18 Firepower kW 6.6 0.2 6.6 1.2 1% -0.04 NO % % SD = Standard deviation; CoV = Coefficient of variation (CoV = SD ÷ mean)

Table 3, on the other hand, shows the impact of greater variability on the statistical confidence. The table shows the specific consumption derived from two pairs of stove comparisons based on three trials each. In both the higher and lower variability cases, the stoves have identical average specific consumptions, favoring the Stove-2 by 23% (104 compared to 134 g wood per liter of water boiled). However, in the lower variability case the coefficient of variation (CoV)4 is 6% and 9% for Stove-1 and Stove-2 respectively, while in the higher variability case the CoV is higher (9% and 13% respectively). In the lower variability case, the difference in the two stoves is statistically significant with 95% confidence, while in the higher variability case, it is not. Thus, even though the specific fuel consumption of Stove-2 appears to be better than Stove-1 by over 20% we can not say with 95% confidence that Stove-2 is better based on the data with higher variability. In order to

4 The coefficient of variation (CoV) is the ratio of the standard deviation to the mean for a given set of data. rectify the situation, we either need to lower our standards of confidence, or conduct additional tests. If we do the lower our standards, we would find that we can say the observed difference between Stove-1 and Stove-2 is significant with 90% confidence. Alternatively, if we want to maintain the standard of 95% confidence, we can try conducting more tests. For example, if we perform additional tests and the standard deviation in the test results does not change from that shown in the higher variability case of Table 3, then 5 tests of each stove will be sufficient to declare that the observed difference of 23% between Stove-1 and Stove-2 is significant with 95% confidence. Table 3: Hypothetical test results showing effect of data variability on statistical confidence based on three tests of each stove

Stove-1 Stove-2 Statistics

Specific Consumption units % difference Significant with Mean SD CoV Mean SD CoV between Stove-1 T-test 95% confidence? and Stove-2 Lower variability g/liter 134 8 6% 104 9 9% -23% 3.4 YES 13 Higher variability g/liter 134 12 9% 104 13 % -23% 2.4 NO

Appendix 2

Modified Water Boiling Test (MWBT) The following testing protocol is a modified version of the Water Boiling Test (Version 1.5) that was presented by the Shell Foundation the year 2004. The original test was made for comparing stoves, and not for comparing fuels and hence the test had to be modified for the current requirements.

For each test session Procedure 1. Make a fire to heat up the stove from firewood. 2. Put on a pot with water, measure the daily boiling point. 3. Weigh the pot. 4. Weigh the stove. 5. Measure the outside temperature. 6. Do the test according to ‘Every test’. 7. Let the fuel turn into ashes and weigh the stove, without spilling any ash.

Results ● Daily boiling point ● Weight of pot being used ● Weight of empty stove ● Outside temperature ● Ash content of the last briquettes tested.

For each test

Procedure (Underlined tasks demands documentation)

1. Measure 2000 g of water 2. Measure temperature of cold water 3. Weigh a pile of fuel 4. Weigh 20 g of firewood twigs 5. Start the fire from firewood twigs and fuel ~10 minutes after the last fire was put out 6. Note the time when it catches fire. 7. Put on the pot of water with a lid 8. Note the time when the temperature of boiling is reached 9. Take of the lid and keep the water simmering for 15 min 10. After 15 minutes take of the pot 11. Weigh the pot with water 12. Weigh the stove 13. Weigh the remaining fuel that is not used 14. Note if there was lots of smoke (1-5) 15. Note other notations

Results

● Cold water temperature ● Initial weight of fuel ● Time when fire starts ● Time when the water reaches the boiling point ● Weight of pot with water after 15 min boiling ● Weight of stove with remaining fuel ● Weight of remaining non-used fuel ● How much smoke was produced ● Other, like ease of ignition etc.

What do we get from the testing?

● Energy transferred from fuel (J/g) to the water in the pot.

● How quick the burning fuel can make the water boiling.

● Other notations like smoke, ease of ignition etc.

Appendix 3

Controlled Cooking Test (CCT) Prepared by Rob Bailis for the Household Energy and Health Programme, Shell Foundation

(Not currently included in Shell HEH Stove Performance Protocols) The controlled cooking test (CCT) is designed to assess the performance of the improved stove relative to the common or traditional stoves that the improved model is meant to replace. Stoves are compared as they perform a standard cooking task that is closer to the actual cooking that local people do every day. However, the tests are designed in a way that minimizes the influence of other factors and allows for the test conditions to be reproduced.

Equipment The equipment required to conduct a series of CCTs is similar to the equipment required to perform the WBT. In addition, a sufficient quantity of food will be needed to conduct all of the tests. This is discussed in more detail below. • Fuel: A homogeneous mix of air-dried fuel wood should be procured. Sufficient wood for all of the CCTs should be obtained ahead of time. Use local input to determine the quantity of fuel required to cook a “standard meal” on a traditional stove. Assume that each stove will be tested at least 3 times and allow for some margin of error. For example, if local people report that a standard meal requires ~2.5 kg of fuel wood and three stoves are to be tested, then the full range of tests will require

kg tests 2.5 x 3 stoves x 3 x 2 . meal stove The final factor of two is included to allow for aborted tests and other contingencies. This is roughly 45 kg of wood. As in the WBT, the fuel may be divided into pre-weighed bundles to save time during testing. • Food and water: Testers should be sure they have sufficient food and water for the entire range of tests. Like fuel, the food should be homogenous so that variability in food does not bias the results of the test.

• Cooking pot(s): if possible, use the standard pots supplied with the testing kits. If the standard pots do not fit one or more of the stoves being tested, use the most appropriate pots and be sure to record the specifications in the Data and Calculation form. If possible, the same type (size, shape, and material) of pots should be used to test each stove. However, unlike the WBT, lids should be used if local cooks commonly use them. • Scale: Supplied with testing kit: (at least 6 kg capacity and 1 gram accuracy): (see note in WBT section). • Heat resistant pad to protect scale when weighing hot charcoal. • Wood moisture meter • Timer. • Thermometer (this is only for recording ambient temperature – food temperatures are not recorded in this CCT). • Small shovel/spatula to remove charcoal from stove for weighing. • Dust pan for transferring charcoal. • Metal tray to hold charcoal for weighing. • Heat resistant gloves. CCT testing procedure The CCT described here is meant primarily to compare the performance of an improved stove to a traditional stove in a standardized cooking task. The procedure that follows should be applied to type of stove commonly in use in the community as well as the model or models of stove being promoted. Three repetitions of the CCT for each stove that is being compared are recommended. 1. The first step in conducting the CCT is to consult with people in the location where the stove or stoves are going to be introduced in order to choose an appropriate cooking task. This should be done well ahead of time, to ensure that sufficient food can be obtained to conduct all of the necessary tests.

• If the stove is designed for home use, then the task should be a typical meal consisting of foods that are regularly eaten in the community. It may include one or more dishes, though foods requiring complicated preparations should be avoided in the interest of time. In addition to the type of food, the testers and community participants must also decide on the precise quantity of food that is best representative of a typical family’s meal. This is critical to ensure that tests are uniform. If local measures are used, the testers should convert this into standard measurements and record these on the Data and Calculations form. The Box below shows an example of the food used for a CCT in West Africa (from Baldwin, 1987).

• If the stove is designed for specialized applications, for example making tortillas or chapati, then the cooking task requires less input and testers must simply decide on the exact amount of food on which to base the test.

• Once a cooking task has been decided on, Example of food used in a CCT (adapted ensure that sufficient food is available to from Baldwin, 1987, p. 94) conduct the tests. Dish Ingredient Quantity (g) 2. After deciding on a cooking task, the procedure Porridge water 4000 should be described in as much detail as possible Millet flour 1000 and recorded in a way that both stove users and testers can understand and follow. This is important Sauce: oil 100 to ensure that the cooking task is performed meat 450 identically on each stove. If possible, include an tomatoes 300 objective measure of when the meal is “done”. In other words, it is preferable to define the end of the water 2500 cooking task by an observable factor like “the skins onions 70 come off the beans” rather than a subjective spices 50 measure like “the sauce tastes right” (VITA, 1985, CCT Procedural note 2).

After sufficient ingredients and fuel have been obtained and the steps of the cooking task are written up and well understood by all participants, the actual testing can begin. The cooking itself should be done by a local person who is familiar with both the meal that is being cooked and the operation of the stove to be tested. If the stove is a new design that differs significantly from traditional cooking practices, some training will probably be required before conducting the actual tests.5 When comparing stoves with the CCT, if more than one cook is used, each cook should test each stove the same number of times, in order to remove the cook as a potential source of bias in the tests. In addition, to ensure that the testers have control over the testing environment, the tests should be conducted in a controllable setting such as a lab or workshop rather than in a private home.

5 Of course, if a great deal of training is required in order for a local user to “master” the use of the stove, then the stove-testers should probably reconsider that particular stove design. 3. Record local conditions as instructed on the Data and Calculation form.

4. Weigh the predetermined ingredients and do all of the preparations (washing, peeling, cutting, etc) as described by the cooking directions recorded in step 2 above. To save time, for non- perishable food, the preparation can be done in bulk, so that food for all of the tests is prepared at once.

5. Start with a pre-weighed bundle of fuel that is roughly double the amount that local people consider necessary to complete the cooking task. Record the weight in the appropriate place on the Data and Calculation form.

6. Starting with a cool stove, allow the cook(s) to light the fire in a way that reflects local practices. Start the timer and record the time on the Data and Calculation form.

7. While the cook performs the cooking task, record any relevant observations and comments that the cook makes (for example, difficulties that they encounter, excessive heat, smoke, instability of the stove or pot, etc).

8. When the task is finished, record the time in the Data and Calculation form (see the comments on determining when the task is complete in step 2 above).

9. Remove the pot(s) of food from the stove and weigh each pot with its food on the balance. Record the weight in grams on the Data and Calculation form.

10. Remove the unburned wood from the fire and extinguish it. Knock the charcoal from the ends of the unburned wood. Weigh the unburned wood from the stove with the remaining wood from the original bundle. Place all of the charcoal in the designated tray and weigh this too. Record both measurements on the Data and Calculation form.

11. The test is now complete – you may now enjoy the food that was cooked or proceed by testing the next stove – each stove should be tested at least 3 times.

Note: this procedure only requires the use of one standardized cooking task. However, stove testers are encouraged to develop a CCT for several different cooking tasks – particularly if the communities where the stove is being promoted cook meals that are equally popular, but differ significantly in their specific cooking requirements (for example, one task that involves slow boiling and another task that involves frying).

Analysis After each test, transfer data from the Data and Calculation forms into the software. Once three tests for each stove are complete, the software provides a value of specific consumption and total cooking for each individual test as well as an average of three tests for each stove. Once CCTs for two stoves are completed, the software will compare the results and test for statistical significance. In addition, any qualitative observations made during each test should be noted. Each data form contains space for qualitative observations to be recorded and summarized on the “Results” page.

Analysis of the CCT The calculations produced by the Data and Calculation form are somewhat more straightforward than the calculations for the WBT. They are explained in Appendix 5. Appendix 5

Analysis of the CCT

Variables As in the WBT, there are a number of variables that are directly measured. These include environmental variables and physical test parameters. The environmental variables may vary slightly from one test to another, but should be nearly constant. The physical test parameters should be constant for all tests. Environmental variables: Wind conditions Air temperature

Physical test parameters: Variable Label

Avg dimensions of wood (centimeters) --

Wood moisture content (% - wet basis) m

Empty weight of Pot # 1 (grams) P1

Empty weight of Pot # 2 (grams) P2

Empty weight of Pot # 3 (grams) P3

Empty weight of Pot # 4 (grams) P4

Weight of container for char (grams) k

Local boiling point of water (°C) Tb

Measurements and Calculations Upon finishing the test, a number of measurements are taken. These include: Initial weight of fuelwood (wet basis) (grams) fi Final weight of fuelwood (wet basis) (grams) ff Weight of charcoal with container (grams) cc The weight of each pot with cooked food (grams) Pjf (j is an index for the cooking pot ranging from 1–4 depending on the number of pots used for cooking) Start and finish times of cooking (minutes) ti and tf

These measurements are then used to calculate the following indicators of stove performance: Total weight of food cooked (Wf) – this is the final weight of all food cooked; it is simply calculated by subtracting the weight of the empty pots from the pots and food after the cooking task is complete: 4 Wf = ∑ ()Pjf − Pj where j is an index for each pot (up to four). j=1

Weight of char remaining (Δcc) – the mass of charcoal from within the stove, including the char removed from the ends of the unburned fuel that is extinguished just at the end of the cooking task. This is found by simple subtraction: Δcc = cc – k

Equivalent dry wood consumed (fd) – This is defined as for the WBT, adjusting for the amount of wood that was burned in order to account for two factors: (1) the wood that must be burned in order to vaporize moisture in the wood and (2) the amount of char remaining unburned after the cooking task is complete. The calculation is done in the following way:

fd = ()ff − fi ∗ (1− (1.12 ∗ m)) − 1.5 ∗ Δc c Specific fuel consumption (SC) – This is the principal indicator of stove performance for the CCT. It tells the tester the quantity of fuel required to cook a given amount of food for the “standard cooking task”. It is calculated as a simple ratio of fuel to food: f SC = d ∗1000 Wf Notice this is reported in grams of fuel per kilogram food cooked, whereas Wf is reported in grams. Thus a factor of 1000 is included in the calculation. Total cooking time (Δt) – This is also an important indicator of stove performance in the CCT. Depending on local conditions and individual preferences, stove users may value this indicator more or less than the fuel consumption indicator. This is calculated as a simple clock difference: Δt = tf - ti

Appendix 4

Equations

Equation for calculating remaining fuel during combustion

Variables f0 = weight of used fuel af = weight of ash at full combustion of f0 dx = weight of fuel and ash at time x fx = weight of fuel at time x ax = weight of ash at time x k = weight percentage of ash from fuel at complete combustion fx is unknown and want to be expressed in the known variables f0,dx,k

Obvious relationships a f = f 0 ⋅ k

a f = f x ⋅ k + ax

d = ax + f x

How to get the equation

⎧a f = f 0 ⋅ k ⎫ f x ⋅ k + ax = a f ⎨ ⎬ ⎩ax = d − f x ⎭

f x ⋅ k + d − f x = f 0 ⋅ k

f x ⋅ (k −1) = f 0 ⋅ k − d

d − f ⋅ k f = 0 x 1− k

Equations for the energy transferred to the water

ΔQ = k ⋅ Δm

ΔQ = c p ⋅ m ⋅ ΔT

Where c = 4,179kJ /(kg ⋅ K) p k = 2,257MJ / g Pressing force from the WU-presser

Note that these calculations are just for getting a brief idea of how much pressure the presser can produce. Below a simple sketch of the forces that acts on the presser is described. The distances are straight from the paper. Since it’s just relation ships between forces the real distances are not needed. The place in the presser for the final pressing is approximated according to the picture.

Fa - The force produced by the person using the presser, using their body weight ~50kg~500N

Fc - The force acting on the briquette

⎧Fa ⋅50 =F b⋅11⎫ ⎨ ⎬ ⇒ Fc ≈ 11,7kN (1170kgf ) ⎩Fb ⋅ 62 =F c⋅12⎭

Distributed on the briquette which are is

A = π ⋅ R 2 = π ⋅ (482 −132 ) ≈ 6,7 ⋅10−3 m 2 (67cm 2 )

The pressure will then be

F 6 2 P = c ≈ 1,7 ⋅10 Pa (17kgf / cm ) A Pressing force from the screw presser Note that these calculations are just for getting a brief idea of how much pressure the presser can produce. The measures the calculations are based on are not the exact measurements. The force produced by the screw presser will be:

2 ⋅π ⋅ F ⋅ r F = η a s F – The force on each side of the lever r – The distance from the screw axis to the force applied on the lever s – The ratio on the screw (distance/rev) η – Efficiency factor

The ratio on the presser at MIRTDC was calculated to 6,81mm/rev and the efficiency is set to 0,4. These values will be used for all pressers.

F P = a A

Presser at PAMET and F = 15kgf ≈ 150N r = 50cm = 0,5m A = 152cm2 = 0,0152 m2

Fa ≈ 27670N (2767kgf )

P ≈ 1,8⋅106 Pa (18kgf / cm2 )

Presser at WESMA F = 15kgf ≈ 150N r = 30cm = 0,3m A = 157cm2 = 0,0157m2

Fa ≈ 16600N (1660kg)

P ≈ 1,1⋅106 Pa (11kg / cm2 )

Presser at MIRTDC F = 40kg r = 100cm A =6x148cm2

Fa ≈ 147600N (14760kgf ) P ≈ 1,6⋅106 Pa (16kg / cm2 ) ( P ≈ 9,6⋅106 Pa if only one briquette at the time would be presses)

Appendix 5

Recipe for Nsima and Vegetables The exact amount of ingredients used in each CCT is presented below. The recipe is for about 3 persons and the amounts presented below were measured when Mrs. Annie Kamanga performed the first CCT test. This recipe was then used for all of the CCT’s in order to make the comparison of results as fair as possible.

Vegetables Oil: 10g Onions: 47g Tomatoes: 95g Salt: 3g Spinach: 116g Water: 35g Procedure: The vegetables are fried together in oil. Salt and water is added. It will cook for some minutes under the lid before it is finished.

Nsima Water: 1163g Maize flour: 323g Procedure: Water is heated. Maize flour is added when the water starts to boil. Then it should boil for a couple of more minutes, while additional flour is added. The nsima is ready when it looks like a thick porridge.

Appendix 6

Equipment for Testing Scale Model: Adam CDW-30 Maximum weight: 30kg Accuracy: 1g Manufacturer: Adam Equipment Co. Ltd, United Kingdom.

Thermometer Model: Fluke 54 II Accuracy: 0.1 °C Manufacturer: Fluke Corporation, Everett WA, USA.

Cooking Equipment Firewood Ceramic Stove The stove was borrowed from ProBEC in Mulanje. Outer diameter of stove: 245mm Inner diameter of stove: 165mm

Charcoal Ceramic Stove The stove was bought at the market by the river in area 2, Lilongwe. Outer diameter of stove: 260mm Inner diameter of stove: 200mm

3 Stone Open Fire The 3 stone open fire was made by three bricks.

Pots The pots used in the tests were bought at the market by the river in area 2, Lilongwe. Diameter: 210mm Weight: 170g, 287 g

Fuels The softwood, hardwood and charcoal used in the tests was bought at the street market in Area 18, Lilongwe.

Appendix 7

MWBT Fuel testing page 1/2 123456789101112131415161718192021 Producer of the briquettes in test Softwood 1Softwood 2Softwood 3Softwood 4Hardwood Hardwood Hardwood Hardwood Charcoal 1 Charcoal 2 Charcoal 3 Charcoal 4 Charcoal 5 Charcoal 6 Charcoal 7 Charcoal 8 NRC a 1 NRC a 2 NRC b 1 NRC b 2 PAMET 1 Raw materials in briquette Softwood Softwood Softwood Softwood Hardwood Hardwood Hardwood Hardwood Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal maize huskmaize huskpaper paper paper paper paper leaves leaves sludge Date of testing 06-11-13 06-11-21 06-11-21 06-11-21 06-11-21 06-11-21 06-11-21 06-11-21 06-11-22 06-11-23 06-11-23 06-11-23 06-11-23 06-11-23 06-11-24 06-12-05 06-11-13 06-11-13 06-11-15 06-11-16 06-11-15

Data specific for each test session Outside temperature ºC 28.6 26.8 25.6 25.6 28 23.7 23.7 23.7 24.6 27.6 26.6 26.6 25.9 24.6 24.5 24.5 28.6 30.2 27.5 26.3 26.7 Local Boiling Point ºC 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 Weight of pot g 169 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 169 169 169 168 168 Weight of lid g 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 Weight of stove g 5557 5540 5544 5541 5539 5538 5538 5534 4308 4287 4286 4286 4283 4274 4271 5161 5558 5556 5554 5554 5571

Weight of stove with ash 5599 Ash content (%) 22222222444444442020202020

Data specific for each test inicial weight of twigs g 66 20 20 20 20 20 20 20 30 30 30 30 30 30 30 30 20 20 20 20 20 Inicial weight of fuel g 590 797 897 689 1251 689 1060 989 5656 4878 4500 4124 3731 3289 2801 4590 940 466 875 780 881 Cold water temperature ºC 28 25.3 27 28.5 26.2 28.2 27.8 27.2 28.3 27.4 29.6 28.6 28.4 27.8 28 25.9 27.7 27.5 27.6 25.5 26.5 Time when fire start 10.20.00 09.36.10 20.49.40 21.26.50 10.18.05 22.02.40 22.43.10 23.23.30 20.55.00 20.38.30 21.22.20 22.12.50 22.56.00 23.40.10 00.31.50 15.19.40 11.02.20 11.44.00 13.28.15 09.50.20 12.13.00 Time when water boils 10.31.00 09.50.35 21.01.05 21.51.50 10.31.05 22.17.30 22.54.20 23.36.10 21.27.00 20.55.20 21.47.00 22.27.40 23.13.20 00.05.45 00.48.12 15.35.50 11.14.55 11.56.00 13.43.15 10.08.30 12.26.50

Water simmers for 15 min until:2 10.46.00 10.05.35 21.16.05 22.06.50 10.46.05 22.32.30 23.09.20 23.51.10 21.42.00 21.10.20 22.02.00 22.42.40 23.28.20 00.20.45 01.03.12 15.50.50 11.29.55 12.11.00 13.58.15 10.23.30 12.41.50 Weight of pot with water g 1880 1819 1911 1835 1868 1890 1792 1808 1984 1810 1876 1804 1748 1972 1782 1769 1823 1821 1871 1884 1848 Weight of stove with remaining fuel and ash g 5576 5557 5555 5560 5713 5588 5580 5568 4485 4496 4497 4496 4526 4585 4461 5380 5649 5677 5733 5728 5703 Weight of remaining non used fuel g 350 507 662 397 734 352 611 558 5365 4500 4124 3731 3289 2801 2418 4164 466 0 408 218 475

Rate smoke of fuel (0-5) 33333333 24341

Used three fire sticks, white smoke, lot's of fire, but irritating some going for eyes. out of hatch, Harder to windy. smoky Used three maybe more handle the when new Wind, fire Only used fire sticks, smoke than briquettes in fuel is applied going out of two sticks in lot's of fire, SW, sprakar, Not sure New stove, fire. More simmering at in stove. stove caused fire, a bit too some going bad “start” of Too litle fuel about initial last one efficient, relatively low otherwise Other remarks by wind less. out of hatch fire in stove stove weight broke cause door temp. windy close to dark smoke can be shut. none.

Results Time taken to boil 2 l of water (mm:ss) 11.00 14.25 11.25 25.00 13.00 14.50 11.10 12.40 32.00 16.50 24.40 14.50 17.20 ######## 16.22 16.10 12.35 12.00 15.00 18.10 13.50 Weight of used fuel g 225.5102 278.5714 228.5714 278.5714 350 292.8571 415.3061 405.102 118.75 176.0417 171.875 190.625 207.2917 184.375 201.0417 215.625 478.75 431.25 360 485 342.5 Weight of fuel put into stove g 240 290 235 292 517 337 449 431 291 378 376 393 442 488 383 426 474 466 467 562 406 Number of used briquettes 4.081415 3.676471

Weight of water that has been vaporized g 289 351 259 335 302 280 378 362 186 360 294 366 422 198 388 401 346 348 298 284 320 Heat transfered from fuel to water 1 kJ 1229 1391 1170 1329 1273 1207 1432 1400 994 1394 1227 1398 1526 1025 1452 1499 1360 1366 1253 1239 1311 - For bringing up to boil 577 599 585 573 592 575 578 583 574 582 563 572 573 578 577 594 579 581 580 598 589 - For simmering 15min 652 792 585 756 682 632 853 817 420 813 664 826 952 447 876 905 781 785 673 641 722 Heat transfer per g used fuel 1 kJ/g 5.4 5.0 5.1 4.8 3.6 4.1 3.4 3.5 8.4 7.9 7.1 7.3 7.4 5.6 7.2 7.0 2.8 3.2 3.5 2.6 3.8

123456789101112131415161718192021 Not included in avreage XX X

Significant Reason Wind Too less fuel diffrent value MWBT Fuel testing page 2/2 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Producer of the briquettes in test PAMET 2 MIRTDC 1 MIRTDC 2 MIRTDC 3 MIRTDC 4 Nordins 1 Nordins 2 Nordins 3 Cape MacleCape MacleGTZ 1 WICO 1 WICO 2 WICO 3 OrphanageOrphanageOrphanageOrphanageWESMA 1 WESMA 2 WESMA 3 Raw materials in briquette paper paper paper Paper Paper paper Papper Papper maize huskmaize huskPaper Sawdust Sawdust Sawdust office pape office pape office pape office pape Paper Paper Paper sludge sludge sludge Sludge Sludge Sawdust sawdust sawdust sawdust sawdust Date of testing 06-11-16 06-11-15 06-11-16 06-11-19 06-11-22 06-11-15 06-12-19 06-12-19 06-11-15 06-11-16 06-11-19 06-11-19 06-11-22 06-11-22 06-11-16 06-11-16 06-11-16 06-11-22 06-12-05 06-12-05 06-12-19

Data specific for each test session Outside temperature ºC 23 26.7 23.4 33.3 21.6 27.5 31.4 31.4 28.3 28.2 33.3 33.3 24.7 23.6 27.2 22.1 21.7 22.8 29.7 29.6 32.5 Local Boiling Point ºC 97 97 97 98 97 97 97 97 97 97 98 98 97 97 97 97 97 97 97 97 97 Weight of pot g 174 168 170 170 170 169 169 169 168 168 170 170 170 170 168 168 168 170 170 170 170 Weight of lid g 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 112 Weight of stove g 5547 5558 5542 5535 5531 5554 5573 5553 5554 5545 5535 5536 5673 5554 5554 5554 5528 5542 5542 5551

Weight of stove with ash 5610 5604 5589 5726 5628 5610 Ash content (%) 20 20 20 20 20 20 20 20 20 20 20 2 2 2 20 20 20 20 20 20 20

Data specific for each test inicial weight of twigs g 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Inicial weight of fuel g 3070 962 1205 877 571 468 828 446 687 920 2346 2564 2177 1105 735 826 1590 808 843 447 2290 Cold water temperature ºC 24.8 27 25.4 28.9 27.3 27.6 25.6 26.8 27.3 25.9 29 29 27.4 27.6 25.7 24.8 24.6 27.7 25.8 26.2 27 Time when fire start 16.00.10 12.52.50 16.40.05 14.55.25 00.00.50 14.08.30 10.09.30 10.52.15 15.01.55 11.11.55 13.43.45 14.19.20 21.50.30 22.05.45 10.32.30 14.30.50 15.16.35 22.52.20 08.39.45 09.17.20 12.23.05 Time when water boils 16.11.50 13.03.00 16.51.02 15.06.15 00.12.15 14.32.30 10.24.52 11.08.40 11.25.05 13.54.30 14.31.10 22.27.30 10.46.20 14.48.25 15.31.40 23.02.50 08.51.50 09.31.10 12.40.45

Water simmers for 15 min until:2 16.26.50 13.18.00 17.06.02 15.21.15 00.27.15 14.47.30 10.39.52 11.23.40 11.40.05 14.09.30 14.46.10 00.15.00 22.42.30 11.01.20 15.03.25 15.46.40 23.17.50 09.06.50 09.46.10 12.55.45 Weight of pot with water g 1779 1743 1621 1830 1873 1896 1974 1748 1762 1694 1848 1810 1855 1799 1845 1867 1936 Weight of stove with remaining fuel and ash g 5689 5642 5606 5596 5595 5604 5670 5652 5641 5673 5693 5702 5754 5675 5660 5663 5634 Weight of remaining non used fuel g 2608 623 885 572 285 446 19 147 1898 2182 546 278 248 978 361 447 52 1857

Rate smoke of fuel (0-5) 23344 33 323 2333

Stoped boiling for a while, open hatch, might Extremely A little bit of be wrong smoky and Faild to start wind, cold Black amount of constant up good, then weaher, too dark smoke. smelling water, wet tendering I put in too many smelly. smoke, dark smoke, briquettes(offi was much fuel. briquettes strange probably didn't boil last ce necessary. A They contain cause bad little smoke. sparkling from sludge, minute cause briquettes), lot of ash. dark smoke, ash in the combustion? smells a little sounds,big almost no of stoped Simmered 1 Bad briquette test failed, middle, that low temp Put in lots of Other remarks bad. sludge ”thing” tenderingputting in failed!!! fire Open hatch min too long. for firewood fire went out has to be simmering fuel inside. dark smoke fuel. burned out. fire went out stove. dark smoke dark smoke emptied windy

Results Time taken to boil 2 l of water (mm:ss) 11.40 10.10 10.57 10.50 11.25 24.00 15.22 16.25 ######## 13.10 10.45 11.50 ######## 21.45 13.50 17.35 15.05 10.30 12.05 13.50 17.40 Weight of used fuel g 400 318.75 320 305 277.5 7527.5 438.75 387.5 7801.25 966.25 426.25 281.6327 7870.408 570.4082 397.5 537.5 515 375 347.5 342.5 437.5 Weight of fuel put into stove g 462 339 320 305 286 468 382 427 687 773 448 382 2177 559 457 578 612 447 396 395 433 Number of used briquettes

Weight of water that has been vaporized g 395 425 549 340 297 2169 273 195 2168 2168 422 408 2170 476 320 358 313 371 325 303 234 Heat transfered from fuel to water 1 kJ 1495 1544 1838 1345 1253 5475 1213 1027 5476 5487 1529 1498 5479 1654 1318 1411 1312 1417 1329 1276 1113 - For bringing up to boil 603 585 598 578 583 580 597 587 583 594 577 577 582 580 596 603 605 579 595 592 585 - For simmering 15min 892 959 1239 767 670 4895 616 440 4893 4893 952 921 4898 1074 722 808 706 837 734 684 528 Heat transfer per g used fuel 1 kJ/g 3.7 4.8 5.7 4.4 4.5 X 2.8 2.6 X X 3.6 5.3 X 2.9 3.3 2.6 2.5 3.8 3.8 3.7 2.5

22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 41 Not included in avreage ? X X X X X ?

Stoped Could not Could not Too bad Reason burning keep fire alive keep fire alive Fire went out combustion MWBT 3SF 3SF 3SF 3SF FWS FWS FWS FWS FWS FWS CCS1 CCS1 CCS1 CCS1 CCS1 CCS1 CCS1 CC2 CCS2 CCS2 Stove testing page 1 12345678910111213141516171819 20 Producer of the briquettes in test Softwood Softwood PAMET PAMET Softwood 1Softwood 2Softwood 3Softwood 4 PAMET 1 PAMET 2 Charcoal 1 Charcoal 2 Charcoal 3 Charcoal 4 Charcoal 5 Charcoal 6 Charcoal 7 Charcoal 8 PAMET PAMET Raw materials in briquette Softwood Softwood Paper Paper Softwood Softwood Softwood Softwood paper paper Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Paper Paper Sludge Sludge sludge sludge Sludge Sludge Date of testing 06-12-03 06-12-03 06-12-03 06-12-03 06-11-13 06-11-21 06-11-21 06-11-21 06-11-15 06-11-16 06-11-22 06-11-23 06-11-23 06-11-23 06-11-23 06-11-23 06-11-24 06-12-05 06-12-05 06-12-05

Data specific for each test session Outside temperature ºC 26.3 28.9 26.3 29 28.6 26.8 25.6 25.6 26.7 23 24.6 27.6 26.6 26.6 25.9 24.6 24.5 24.5 23.8 24.4 Local Boiling Point ºC 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 97 Weight of pot g 170 170 170 170 169 170 170 170 168 174 170 170 170 170 170 170 170 170 171 171 Weight of lid g 111 111 111 111 112 112 112 112 112 112 112 112 112 112 112 112 112 112 111 111 Weight of stove g 00005557 5540 5544 5541 5571 5547 4308 4287 4286 4286 4283 4274 4271 5161 5156 5164

Weight of stove with ash 75 Ash content (%) 2 2 20 20 22222020444444442020

Data specific for each test inicial weight of twigs g 20 20 20 20 66 20 20 20 20 20 30 30 30 30 30 30 30 30 20 20 Inicial weight of fuel g 2024 1848 2403 1830 590 797 897 689 881 3070 5656 4878 4500 4124 3731 3289 2801 4590 1308 973 Cold water temperature ºC 25.8 26.6 25.8 27.8 28 25.3 27 28.5 26.5 24.8 28.3 27.4 29.6 28.6 28.4 27.8 28 25.9 26.2 26 Time when fire start 10.02.30 14.10.30 10.42.30 12.48.40 10.20.00 09.36.10 20.49.40 21.26.50 12.13.00 16.00.10 20.55.00 20.38.30 21.22.20 22.12.50 22.56.00 23.40.10 00.31.50 15.19.40 13.59.20 14.36.35 Time when water boils 10.16.00 14.22.20 10.56.00 13.04.15 10.31.00 09.50.35 21.01.05 21.51.50 12.26.50 16.11.50 21.27.00 20.55.20 21.47.00 22.27.40 23.13.20 00.05.45 00.48.12 15.35.50 14.10.50 14.50.25

Water simmers for 15 min until:2 10.31.00 14.37.20 11.11.00 13.19.15 10.46.00 10.05.35 21.16.05 22.06.50 12.41.50 16.26.50 21.42.00 21.10.20 22.02.00 22.42.40 23.28.20 00.20.45 01.03.12 15.50.50 14.25.50 15.05.25 Weight of pot with water g 1799 1885 1851 1896 1880 1819 1911 1835 1848 1779 1984 1810 1876 1804 1748 1972 1782 1769 1792 1867 Weight of stove with remaining fuel and ash g 73 37 150 131 5576 5557 5555 5560 5703 5689 4485 4496 4497 4496 4526 4585 4461 5380 5251 5254 Weight of remaining non used fuel g 1459 1425 1831 1309 350 507 662 397 475 2608 5365 4500 4124 3731 3289 2801 2418 4164 973 663

Rate smoke of fuel (0-5) 12

windy. smoky when new fuel is applied in Stopped the stove. little smoke. Not sure New stove, test 90 Ash content otherwise smells a little Too litle fuel about initial last one Stopped 30s Other remarks seconds too availible close to bad. in stove stove weight broke too late late Dark smoke none.

Results Time taken to boil 2 l of water (mm:ss) 13.30 11.50 13.30 15.35 11.00 14.25 11.25 25.00 13.50 11.40 32.00 16.50 24.40 14.50 17.20 ######## 16.22 16.10 11.30 13.50 Weight of used fuel g 502.0408 393.8776 527.5 487.5 225.5102 278.5714 228.5714 278.5714 342.5 400 118.75 176.0417 171.875 190.625 207.2917 184.375 201.0417 215.625 300 275 Weight of fuel put into stove g 565 423 572 521 240 290 235 292 406 462 291 378 376 393 442 488 383 426 335 310 Number of used briquettes

Weight of water that has been vaporized g 371 285 319 274 289 351 259 335 320 395 186 360 294 366 422 198 388 401 379 304 Heat transfered from fuel to water 1 kJ 1432 1232 1315 1197 1229 1391 1170 1329 1311 1495 994 1394 1227 1398 1526 1025 1452 1499 1447 1280 - For bringing up to boil 595 588 595 578 577 599 585 573 589 603 574 582 563 572 573 578 577 594 592 593 - For simmering 15min 837 643 720 618 652 792 585 756 722 892 420 813 664 826 952 447 876 905 855 686 Heat transfer per g used fuel 1 kJ/g 2.9 3.1 2.5 2.5 5.4 5.0 5.1 4.8 3.8 3.7 8.4 7.9 7.1 7.3 7.4 5.6 7.2 7.0 4.8 4.7

1234567891011121314151617181920 Not included in avreage XX

Significant diffrent value, not very Too lite fuel much water Reason in stove vapourized

Appendix 8

Date CCT with Firewood in Firewood Stove 06-12-01 Ingredients Weight Time

FCS Stove 5564 Fuel 1983 Air temp. Water temp.

Start of fire 16.23.30 12:00:00 AM

Start to cook 16.26.10 12:02:40 AM

Oil 10 16.26.10 12:02:40 AM Onions 47 16.26.10 12:02:40 AM Tomato 95 16.27.45 12:04:15 AM Salt 3 16.27.45 12:04:15 AM Spinach 116 16.30.10 12:06:40 AM Water 35 16.32.30 12:09:00 AM Summa 306 Ready 209 16.34.00 00:10:30

Start 16.36.00 12:12:30 AM Water 1163 Nsima 323

Ready 1277 16.50.30 00:27:00

Stove 5594 Used fuel Fuel 1671 312

Notes 30g twigs + 10g paper to light fire

Date CCT with Firewood in Firewood Stove 06-12-12 Ingredients Weight Time

FCS Stove 5607 Fuel 2245 Air temp. Water temp.

Start of fire 17.12.15 12:00:00 AM

Start to cook 17.15.25 12:03:10 AM

Oil 10 17.15.25 12:03:10 AM Onions 47 17.15.50 12:03:35 AM Tomato 95 17.17.30 12:05:15 AM Salt 3 17.18.50 Spinach 116 17.21.50 12:09:35 AM Water 35 17.23.30 12:11:15 AM Summa 306 Ready 372 17.26.00 00:13:45

Start 17.26.15 Water 1163 17.26.15 12:14:00 AM Nsima 323 17.29.00 12:16:45 AM

Ready 1548 17.41.35 00:29:20

Stove 5624 Used fuel Fuel 1870 375

Notes It is raining in the beginning, but we are under roof. The fuel burns slow, so Annie puts in more after pouring in salt in the pot. Missed to take the exact time when the maize flour was poured in first time. The Nsima water was boiling very much and Annie left 19g of maize flour, that was not used. 30g twigs + plastic bag was used to light fire

Date CCT with Charcoal in Charcoal Stove 06-12-02 Ingredients Amounts Time

CCS Stove 3755 Fuel 6837 Air temp. 22,2 Water temp. 26,7

Start of fire 19.02.20 12:00:00 AM

Start to cook 19.14.38 12:12:18 AM

Oil 10 19.14.38 12:12:18 AM Onions 47 19.15.10 12:12:50 AM Tomato 95 19.16.29 12:14:09 AM Salt 3 Spinach 116 19.19.40 12:17:20 AM Water 35 19.21.25 12:19:05 AM Summa 306 Ready 190 19.24.19 00:21:59

Start Water 1163 19.24.25 12:22:05 AM Nsima 323 19.46.30 12:44:10 AM

Ready 1358 19.51.22 00:49:02

Stove 3967 Used fuel Fuel 6314 523

Notes Rain in the air, put on more charcoal 19.33.40. maybe not enough initial fuel was used. 30g twigs, 10g paper, plastic bag used to light fire.

Date CCT with Charcoal in Charcoal Stove 06-12-14 Ingredients Weight Time

Stove 5222 Fuel 3373 Air temp. Water temp.

Start of fire 17.02.20 12:00:00 AM

Start to cook 17.12.15 12:09:55 AM

Oil 10 17.12.15 12:09:55 AM Onions 47 17.12.50 12:10:30 AM Tomato 95 17.15.00 12:12:40 AM Salt 3 17.16.25 Spinach 116 17.17.00 12:14:40 AM Water 35 17.17.55 12:15:35 AM Summa 306 Ready 386 17.21.00 00:18:40

Start Water 1163 17.21.20 12:19:00 AM Nsima 323 17.23.00 12:20:40 AM

Ready 1542 17.34.50 00:32:30

Stove 5401 Used fuel Fuel 2957 416

Notes

Date CCT with Biomass Briquettes in Firewood Stove 06-12-02 Ingredients Amounts Time

FCS Stove 5583 Fuel 1429 Air temp. 22,2 Water temp. 26,7

Start of fire 18.16.29 12:00:00 AM

Start to cook 18.20.09 12:03:40 AM

Oil 10 18.20.09 12:03:40 AM Onions 47 18.20.43 12:04:14 AM Tomato 95 18.21.50 12:05:21 AM Salt 3 18.23.15 12:06:46 AM Spinach 116 18.23.36 12:07:07 AM Water 35 18.25.28 12:08:59 AM Summa 306 Ready 198 18.27.30 00:11:01

Start Water 1163 18.27.50 12:11:21 AM Nsima 323 18.31.50 12:15:21 AM

Ready 1294 18.42.21 00:25:52

Stove 5674 Used fuel Fuel 993 436

Notes Big fire, open hatch, briquettes in pieces, rain in the air, more briquettes 18.27.50, 30g twigs, 10g paper

Date CCT with Biomass Briquettes in Firewood Stove 06-12-14 Ingredients Weight Time

Stove 5595 Fuel 611 Air temp. 24 Water temp.

Start of fire 16.24.10 12:00:00 AM

Start to cook 16.27.37 12:03:27 AM

Oil 10 16.27.37 12:03:27 AM Onions 47 16.28.10 12:04:00 AM Tomato 95 16.28.35 12:04:25 AM Salt 3 16.29.25 Spinach 116 16.29.46 12:05:36 AM Water 35 16.31.00 12:06:50 AM Summa 306 Ready 199 16.33.40 00:09:30

Start Water 1163 16.34.00 12:09:50 AM Nsima 323 16.38.28 12:14:18 AM

Ready 1598 16.46.45 00:22:35

Stove 5649 Used fuel Fuel 224 387

Notes

Date CCT with Biomass Briquettes in Charcoal Stove 06-12-12 Ingredients Weight Time

CCS Stove 5254 Fuel 998 Air temp. 23,4 Water temp. 26

Start of fire 16.18.30 00.00.00

Start to cook 16.21.25 00.02.55

Oil 10 16.21.25 00.02.55 Onions 47 16.22.00 00.03.30 Tomato 95 16.23.50 00.05.20 Salt 3 16.24.40 Spinach 116 16.25.00 00.06.30 Water 35 16.27.15 00.08.45 Summa 306 Ready 367 16.29.50 00.11.20

Start Water 1163 16.30.10 00.11.40 Nsima 323 16.35.40 00.17.10

Ready 1346 16.46.00 00.27.30

Stove 5298 Used fuel Fuel 619 379

Notes 18 g additional maize flour was used to get the right consistency 30g twigs+paper was used to light fire

Framläggningsdatum Institution och avdelning 2007-03-29 Institutionen för ekonomisk och industriell utveckling. Publiceringsdatum (elektronisk version)

Språk Rapporttyp ISBN:

Svenska Licentiatavhandling ISRN: LIU-IEI-TEK-A--07/00129--SE X Annat (ange nedan) X Examensarbete

C-uppsats Engelska D-uppsats Serietitel Övrig rapport

Serienummer/ISSN

URL för elektronisk version

Titel: Biomass Briquettes in Malawi

Författare: Olle Faxälv, Olof Nyström

Sammanfattning

In Malawi 2.5 % of the forest disappears each year. The use of firewood and charcoal, deriving from forest resources, accounts for about 99 % of the household energy demand in Malawi and is a cause to the deforestation. The Government of Malawi recently launched a programme called Promotion of Alternative Energy Sources Programme (PAESP) with the aim to reduce the use of firewood and charcoal. One of the fuels included in the programme is the biomass briquette. The aim with this study is to evaluate the viability of biomass briquettes as a sustainable alternative energy source to firewood and charcoal for households in Malawi.

Research for the study was carried out during three months in Malawi. Visits were made to a number of briquette production sites to study the manufacturing methods and to collect briquette samples. The briquettes were tested using various methods and then compared with results for firewood and charcoal.

At the moment various production methods are used in Malawi, with a high difference in technical complexity and cost. Machines produced from wood using very basic mechanics can apply similar pressure as more advanced metal pressers. They also seem to be better suited than those made of metal, in terms of price and availability.

The majority of the briquette producers in Malawi use waste paper as base material. Although the paper briquettes are good, other raw materials will be needed if the production is supposed to be significantly increased.

The briquettes burn well using the most common stoves in Malawi, including the commonly used charcoal stove. While firewood is cheaper to use than other available fuels, the briquettes seem to be able to compete with the fuel costs for charcoal.

Nyckelord Biomass briquettes, Household energy, Briquette press, MFS, Alternative energy sources