5eq5öK>(5>

r^y "T"^l T STOCKHOLM [SEI-EED--38—_ VH I ENVIRONMENT __ lj I J JL INSTITUTE S b"X" Et,D~ 3? International Institute for Environmental Technology and Management .

im-» Efficiency and Emissions of Charcoal the Improved Mbaula Cookstove

^Kaoma and G.BvKasali building ancWndustriai Minerals Research Unit National Council for Scientific Research, Zambia

REPUBLIC OF ZAMBIA Ministry of Energy and Water Development Department of Energy

Energy, Environment and Development Series - No. 38

Published by the Stockholm Environment Institute ISBN: 91 88116 94 8 in collaboration with SIDA 1994 IKTBiWWBi 8f m 0QB8MT IS UNtllfflCD MSN SAIE5 PROSISITtB Published in collaboration with

rmunwii »Wilbur

Efficiency and Emissions of Charcoal Use in the Improved Mbaula Cookstove

J. Kaoma and G.6. Kasali Building and Industrial Minerals Research Unit National Council for Scientific Research, Zambia

Anders Ellegård Stockholm Environment Institute

I^TV\Nf B

jBIMBifflflM 8F ii-15 DÖDÖVEKT IS UNL!MIT)ED FBRE1SH SALES PRPftlBITEB OCL Stockholm Environment Institute Box 2142 S-l 03 14 Stockholm Sweden Tel 446 8 723 0260 Fax 446 8 723 034S

Desk-top Publishing. A. Ellegärd Stockholm Environment Institute

© Copyright 1994 by the Stockholm Environment Institute No part of this report may be reproduced in any form by photostat, microfilm, or any other means, without written permission from the publisher.

ISBN: 91 88116 94 8 Contents

1 INTRODUCTION 1

2 EXPERIMENTAL 1

2.1 Test site, methods and instrumentation 1 2.1.1 Carbon monoxide 2 2.1.2 Carbon dioxide 2 2.1.3 Sulphur dioxide 2 2.1.4 Nitric oxide 2 2.1.5 Nitrogen dioxide 2 2.1.6 Aerosol 2 2.1.7 Respirable suspended particulates 3 2.1.8 Temperature 3 2.1.9 Calibration 3 2.1.10 Data recording 3 2.2 Fuel and Stove 3 2.3 Ignition of charcoal 4 2.4 Data analysis 5 2.4.1 Burn rate 5 2.4.2 Thermal efficiency 5 2.4.3 Air exchange rate 5 2.4.4 Measured pollutant concentration 6 2.4.5 Equilibrium pollutant concentration 6 2.4.6 Emission factor 6 2.4.7 Emission rate 7

3 RESULTS AND DISCUSSION 7 3.1 Energy conversion 7 3.2 Pollutant emissions 10 3.2.1 Carbon monoxide 10 3.2.2 Sulphur dioxide 11 3.2.3 Nitrogen oxides 12 3.2.4 Aerosol and respirable suspended particulates .... 12 3.3 Emission Dynamics 13 3.4 Improved versus Traditional Mbaula 15

4 CONCLUSIONS 17

5 ACKNOWLEDGEMENT 18

6 REFERENCES 18

APPENDIX 21

i Tables

Table 1 Proximate analysis of miombo charcoal 4 Table 2 Mean values of thermal combustion parameters for charcoal in improved mbaula 8 Table 3 Mean carbon monoxide concentration, emission factor and emission rate for charcoal in improved mbaula .... 11 Table 4 Mean sulphur dioxide concentration, emission factor and emission rate for charcoal in improved mbaula 12 Table 5 Mean nitrogen oxides emission factor and emission rate for charcoal in improved mbaula 12 Table 6 Mean aerosol and RSP concentrations, emission factors and emission rates for charcoal in improved mbaula ... 13 Table 7 Mean thermal and emission values for traditional mbaula and improved mbaula 16 Table 8 Effect of air inlet aperture on the thermal and emission characteristics of the improved mbaula 17 Table A.l Thermal combustion characteristics of charcoal in improved mbaula 22 Table A.2 Air exchange rates and mean pollutant concentrations generated during combustion of charcoal in improved mbaula 23 Table A.3 Pollutant emission factors for the combustion of charcoal in improved mbaula 23 Table A.4 Pollutant emission rates from charcoal combustion in improved mbaula 24

Figures

Figure 1 Design of the improved mbaula 4 Figure 2 Temperature variations during the combustion of charcoal in the improved mbaula 9 Figure 3 Profile of room concentration of CO for a typical burn cycle 14 Figure 4 Traditional and improved mbaula 15

Equations

(1) Burn rate 5 (2) Thermal efficiency 5 (3) Air exchange rate 6 (4) Measured pollutant concentration 6 (5) Equilibrium pollutant concentration 6 (6) Emission factor 6 (7) Emission rate 7

ii ABSTRACT

An improved chamber method was used to evaluate the thermal performance and emission characteristics of charcoal in an unvented cookstove known as the Improved Mbaula. Emission factors and rates for pollutants, bum rate and stove efficiency were determined. The pollutants that were continuously monitored were carbon monoxide (CO), sulphur dioxide (S02), nitric oxide (NO) nitrogen dioxide (N02), and respirable suspended particulates (RSP). Con­ centrations of CO, nitrogen oxides and RSP in the test chamber (a simulated kitchen) reached levels in excess of guidelines recommended in industrialized countries. Concentrations of S02 did not exceed known levels. If the test chamber actually is a good simulation of a common kitchen, the levels reached warrant concern for the health of people exposed, mostly women and children. Levels of pollution in actual kitchens will be assessed in a later study. The adjustable opening of the stove proved effective in regulating the burn rate. At half air input, burn rate decreased by about 40%, while emissions increased by about 60% compared to operation at full air input. Emissions of CO were 340 g/kg charcoal at full air input, which was taken to be the normal mode of operation. The average thermal efficiency (PHU) of the improved mbaula was 25% compared to 29% for the traditional charcoal stove.

iii Kaoma, Kasali & Ellegård

1 INTRODUCTION Zambia, like most developing countries, has the majority of its population dependent on biomass fuels for its energy needs. It was estimated that in 1990, fuelwood and charcoal accounted for 76% of Zambia's total primary energy supply (1). All this biomass fuel is harvested from native woodlands. Zambia has a relatively high population growth of 3.2% and hence measures are required to enable woodlands to sustain the ever-increasing demand for wood fuels. One way to achieve this is to decrease demand by reducing fuel consumption through the use of more efficient cookstoves. With the above in mind, the University of Zambia developed an improved charcoal stove known as the improved mbaula which was intended to reduce charcoal consumption in households. This stove has been widely disseminated in five major towns of Zambia with about 600 artisans and 2,000 consumers being trained to make and use it respectively (2). On the other hand, the efficiency of the improved mbaula has not been independently scientifically ascertained and the relationship between this efficiency and fuel emissions has also not been established. The aim of this study was, therefore, to evaluate the efficiency and emission levels of burning charcoal in the improved mbaula.

2 EXPERIMENTAL

2.1 Test site, methods and instrumentation An improved chamber method which allows simultaneous evaluation of emissions and thermal attributes of chimneyless stoves, was used for the tests (3). This method utilizes the water boiling test to determine the overall stove efficiency or percentage of heat utilisation (PHU), while fuel emissions are quantified using the air exchange rate determined for the particular test. The design of the test-room and the experimental methods employed in this study were described in more detail in a previous report (4). Only important aspects are summarised here. The test chamber simulates a low-income house kitchen. The total internal volume of the test-room was 13.7 m3 and stratifica­ tion of pollutants in the room was eliminated with the use of two fans. The fans were placed in such a way that they did not interfere with the combus­ tion. The air exchange rate was assessed using the decay of concentration of a pollutant (usually CO) when emissions had ceased. For each experiment, the air exchange rate was assessed immediately after the monitored burn cycle.

1 Improved mbaula

Continuously measuring instruments were used to monitor carbon monoxide (CO), sulphur dioxide (S02), carbon dioxide (C02), nitric oxide (NO), nitrogen dioxide (NOz) and aerosol (smoke). The gas monitors were calibrated using commercially available standards. Respirable suspended particulates (RSP) were collected on a filter. Instruments for CO, NO, C02, aerosol and temperature were connected to a data logger which was interfaced with a computer for data storage.

2.1.1 Carbon monoxide For real-time monitoring of carbon monoxide a Drager Minipac T3 portable electrochemical CO monitor was used. Monitoring range was 0 - 2000 ppm. Data from the Minipac was recorded through the computer system.

2.1.2 Carbon dioxide Real-time monitoring of carbon dioxide was done with a Riken 411 A infrared monitor. This is a stationary instrument but can be operated with batteries. Measuring range 0-5%. Data was collected through the logger system.

2.1.3 Sulphur dioxide Sulphur dioxide concentrations were measured with an Industrial Scientific S0261 portable electrochemical monitor with digital display. Measuring range 0-200 ppm. This instrument was manually read.

2.1.4 Nitric oxide Nitric oxide was monitored with a transportable electrochemical NO monitor, Ecolyzer 2000. Measuring range 0-100 ppm. Data was collected through the logger system.

2.1.5 Nitrogen dioxide Nitrogen dioxide concentrations were measured with an Industrial Scientific N0264 electrochemical portable nitrogen dioxide monitor with digital display. Measuring range 0-1000 ppm. This instrument was manually read.

2.1.6 Aerosol Aerosol, a summary measure of particulate matter and liquid droplets, was monitored using a GCA Corp Miniram PDM-3 portable Aerosol monitor of nephelometric type. Monitoring range 0.01 - 100 mg/m3. Maximum particle size 10 jum aerodynamic diameter (calibrated for Arizona road dust). Readings were recorded manually.

2 Kaoma, Kasali & Ellegård

2.1.7 Respirable suspended particulates Respirable suspended particulates were collected on a filter using a Gil-Air portable dust sampling pump model 17 G9. The flow rate through the pump was adjusted to 1.9 litres/min. 8.0 /im pore size filters were used (Millipore Corp) for the RSP collection. A cyclone (SKC Inc.) was attached to the filter holder to remove particles larger than 7.2 /xm. Filters were desiccated before and after sampling and weighed on an A&D electronic balance model ER- 182A.

2.1.8 Temperature Temperatures were measured with thermocouple wire type K connected to the computer through the data logger.

2.1.9 Calibration Gas monitors were calibrated throughout the study using commercially available standards. Pump flow for RSP monitoring was calibrated using a high-precision rotameter (Contram-Essholm).

2.1.10 Data recording

Some of the continuously monitoring instruments (CO, COz, aerosol, temperatures) were connected to a 16-channel data logger model AAC-2, which was interfaced with a portable Toshiba computer for data storage. In addition to the automatic logging of data, all instruments were manually read and recorded in an experiment log. In the case of S02 and N02 this was the only mode of dat? recording.

2.2 Fuel and Stove Trees for charcoal production usually come from Miombo woodlands which are dominated by three genera: Brachystegia, Julbemardia and Isoberlinia. The charcoal used in this study was purchased from a local market in Lusaka. The chemical properties of charcoal were analyzed by the Industrial Minerals Research Laboratory of the National Council for Scientific Research. The results are shown in Table 1. The improved mbaula was obtained from the Department of Energy, Ministry of Energy and Water Development, which has been acting as the agency for stove dissemination. The improved mbaula is a portable cylindrical metal stove and meant for single pot cooking. An air vent with a sliding door to control air flow and power output is located in the lower portion of the stove body. The air enters through the vent passing the ash chamber to the fuel bed via a multi-holed metal grate which is placed in the middle of the stove. Combustion gases leave the stove through a circular exhaust area

3 Improved mbaula Table 1 Proximate analysis of miombo charcoal

Ash, % 6.1 Volatile matter, % 25.9 Fixed carbon, % 65.8 Moisture, % 2.19 Heat value, MJ/kg 29.0

located at the top of the stove. The top is also provided with pot rests. The stove design features are shown in Figure 1.

Figure 1 Design of the improved mbaula

~ /"*> - o 0

o ; © ©I ?

o

' • - 'F ' ' L. 100 tCOMIUSTION CMA»«Cn r 7 VENTILATION CH A Mir. ft 3.SLIDING 000R 1 '**

t* 700

^•^ * HO

ALL OIHC MSONS IN «-J

2.3 Ignition of charcoal The mode of charcoal ignition has been described previously (4) and suffice it here to state that the combustion process was started outside the test-room by sprinkling 10 ml of kerosene on the charcoal which was then ignited with a match stick. Immediately the charcoal caught fire the stove was taken into the test-room to commence the test.

4 Kaoma, Kasali & Ellegård

2.4 Data analysis The equations used in the calculation of the burn rate, thermal efficiency air exchange rate and the emission factor were based on the approach used by Ahuja et al. (1987). The following equations were used.

2.4.1 Burn rate The burn rate was calculated according to Equation (1). It was corrected for the moisture content of the fuel. 100 ( W W ) F.i- r f (1) t ( 100+M ) where F = burn rate (kgh1) W, = Weight of fuel at start of test (kg) W, = Weight of fuel at end of test (kg) t = Total test duration (h) M = Moisture content of fuel (%)

2.4.2 Thermal efficiency The burn rate and the net calorific value of the fuel were used in the calculation of this parameter according to Equation (2):

xlOO (2) \ Fth where n = thermal efficiency (%) WOTWf = Initial and final weights of water in the pot (kg) a = Specific heat of water (MJkg^C1) Tu = Initial and final temperature of water (°C) L = Latent heat of vaporization of water at 100 °C (MJkg'1) F = Burn rate (kgh-1) t = Total test duration (h) H = Net calorific value for the fuel ((MJkg'')

2.4.3 Air exchange rate The air exchange rate of the test-room was determined by the assumed logarithmic decay of pollution concentration after emissions had ceased (the stove brought out of the room) (Equation (3)). This decay was approximated by plotting the natural logarithm of the pollution (here CO was chosen) concentration against time, and determining the least square fit of the ensuing line. The slope of the line gives the air exchange.

5 Improved mbaula

S -. ( ^ - ^ ' (3) ', where S = air exchange rate (h1)' C„, C, = concentration at start and instantaneous concentration t, = instantaneous time (h)

2.4.4 Measured pollutant concentration

The mean pollutant concentration (Ca) was calculated by averaging the instantaneous concentration (Q) values for continuously recording instruments. The mean RSP concentration was according to Equation (4):

C = J—2. LA (4) Qt where C, = mean pollutant concentration (mgm"3) W„, W,, = Initial and final weights of the filter catch (mg) O = Average air flow rate of the sampling pump (mV) t = sampling duration (h)

24.5 Equilibrium pollutant concentration The equilibrium (steady-state) concentration (Cc) for continuously monitoring instruments was calculated according to Equation (5): C. Ce = '—— (5) ( 1 - *"' )

where Cc = equilibrium pollutant concentration (mgm.3) C, = Instantaneous concentration (mgm"3) t, = Instantaneous time (h) S = Air exchange rate (h )

The Cc values for RSP were obtained by multiplying the Ca values by the ratio of Ce and Ca for CO.

2.4.6 Emission factor The emission factors were calculated according to Equation (6):

E = —-— (6) 1000F where E = emission factor (gkg1)

1 Actually volumes/hour, i.e. the exchanges of the full volume of the room per hour.

6 Kaoma, Kasali & Ellegård

3 Ce = Equilibrium pollutant concentration (mgm* ) S = Air exchange rate (h"1) V = volume of test-room (m3) (13.67 m3) F = Burn rate (kgh-1)

2.4.7Emission rate The emission rate was calculated according to equation (7):

JER - V(a+k) [ C(7) " C(0) g'C<,4t)7] ~ VPaCo (7)

where ER = emission rate (cm*3h'') V = Volume of test-room (cm3) C(T)= Final average indoor pollutant concentration at time T (ppm) C(0) = Indoor pollutant at time 0 (ppm) C0 = Outdoor pollutant concentration (ppm) T = Test duration (h) a = Air exchange rate (h"1) k = Net rate of removal processes other than air exchange (h'1) P = Fraction of the outdoor pollutant level that permeates the test-room (unit-less).

In the case of gases the emission rate in cm3/n was converted to g/h by using the ideal gas law.

3 RESULTS AND DISCUSSION

3.1 Energy conversion Water boiling tests were conducted with the stove air vent door fully opened. It required three water boiling test runs to complete a single burn cycle. Three pots were used in the water boiling tests, designated A, B and C to reflect the different burn phases. The pots contained 3.5, 3.0 and 3.0 litres of water respectively. The different burn phases for a burn cycle are shown in Figure 3. The mean values of the determined thermal combustion parameters are presented in Table 2. The full set of data are given Table Al of the Appendix. Charcoal in the improved mbaula burned at a mean rate of 0.23 kg/h per full burn cycle, with a range of 0.16 - 0.29 kg/h. The burn rate was always highest in the first phase (A) (0.30 - 0.49 kg/h) and lowest in the last phase (C) (0.01 - 0.16 kg/h). One burn phase included the time taken to bring the water in the pot to boiling and 15 minutes during which it was kept boiling (also shown in Figure 3). On the average, a burn phase took about 1 hour,

7 Improved mbaula

Table 2 Mean values of thermal combustion parameters for charcoal in Improved mbaula

Burn phase A B C Average Burn rate (kg/h) 0.36 0.22 0.09 0.22 Test duration (h) 0.93 0.85 1.28 1.02 Thermal efficiency (%) 19.4 27.5 26.8 24.6 CO/CO, ratio 0.11 Power output (kW) 2.90 1.77 0.72 1.80

bringing 3 - 3.5 litres of water to boil as shown in Table 2. Actual boiling time2 ranged from 18 to 109 minutes. i : Combustion of charcoal in the improved mbaula resulted in a mean efficiency or percentage of heat utilised (PHU) of 25% with a range of 19- 29%. The efficiency was lower in phase A (mean 19%) than in phases B and C (28 and 27% respectively). The mean efficiency found for the stove is higher than the efficiency of 20% reported for a metal Jiko stove (5) but within the range of 20.9 to 33.9% reported for the Thai bucket stove (6). Ideally, the provision of a proper fuel/air mixture in a stove should result in a complete combustion whereby all the fuel carbon is converted to carbon dioxide. However, this does not occur in a naturally vented system like the improved mbaula where the combustion of carbon yields both CO and C02. The CO/C02 ratio is a measure of the combustion efficiency or the efficiency of the supply of oxygen (air) to the fuel in the stove. The improved mbaula-charcoal system recorded a mean ratio of 0.11 as shown in Table 2 with a range of 0.09 - 0.133. Chomcharn, et al. (7) found that a bucket stove recorded efficiencies of 24.9% in calm weather and 16.6 % in a windy situation. The burn rates observed with the improved mbaula correspond to a power output of 1.8 kW on the average (range 1.3 - 2.4 kW). This power output falls in the same range as typical electric hotplates, which have a power rating of around 1.5 kW. The power output was observed to be higher at the beginning of the burning cycle (phase A; 2.5-4.1 kW) and lower at the end (phase C; 0.1-1.3 kW). This should be expected since the heat in phase A

2 Boiling time here and in the following means the time taken to bring the water to the boiling point.

3 These values were calculated for the full burn cycle only.

8 Kaoma, Kasali & Eliegård

Figure 2 Temperature variations during the combustion of charcoal in the improved mbaula

0 30 60 90 120 150 180 210 240 270 300 Minutes emanates from combustion of both volatiles (which have a relatively higher calorific value) and of carbon. In phase C, only carbon combustion takes place, and less fuel is present. Figure 2 shows typical temperature development profiles during a burn cycle of charcoal combustion in the improved mbaula. The fuel bed temperature was measured at the centre of the combustion chamber while the stove wall temperature was recorded on the outer surface wall of the chamber. The ash chamber temperature was measured underneath the grate. Three phases could be identified in the fuelbed temperature profile. In the first phase, there was a rapid rise in temperature to a maximum value within about 45 minutes. Khummongkol (8) noted that 1 kg of charcoal in a bucket stove reached the maximum temperature after 44 min. This is the devolatili- zation stage, the duration of which is a function of the charcoal volatile matter content. During the devolatilization stage, the emission of particulates (smoke) is peaking. This phase coincided with phase A (Table Al in Appendix) in the water boiling tests. The highest burn rates, which were always observed in phase A, did not result in the shortest boiling times, thus indicating that heat transfer to the pot was adversely affected. An explanation could be that much of the heat

9 Improved mbaula generated during this stage was lost through ventilation of hot gases as well as to the metallic stove body and to the preheating of unburnt charcoal pieces. During the second phase the temperature was slowly dropping from its maximum. This was phase B in the water boiling tests when burn rates dropped to an intermediate value as char combustion ensued. At this stage the entire fuelbed was glowing red-hot and this was the phase of maximum heat transfer to the pot generating the shortest boiling times as indicated in Table Al (see Appendix). In this phase the main heat transfer mechanism is probably radiation. In the final phase, a relatively slow temperature drop was noted as the fuel progressively turned into ashes. The burn rates dropped to their lowest levels. As a result, the low fuelbed temperature led to the longest boiling times in phase C (Figure 3 and Table Al in appendix). This should be expected, since as the charcoal burns to ashes, the fuelbed volume also decreases, resulting in an increase of the distance between the charcoal and the pot bottom. That means that radiation heating of the pot start becoming insignificant, and heating by conduction - metal to metal- is more important. Figure 2 also shows the ash chamber and stove wall temperature profiles which exhibited trends similar to that of the fuelbed. All the three temperatures peaked at almost identical times suggesting a high heat conductivity for the metal used in stove fabrication. This means that heat loss by conduction should be expected to be high for the improved mbaula. On the average, the ash chamber and stove wall temperatures were relatively 69 and 63 % lower than that of the fuelbed. Khummongkol (8) measured the temperature underneath the grate of a Thai bucket stove and found it to be 70% lower than the fire temperature. He attributed this to low heat radiation from the grate surface. It could also be due to the fact that relatively cold air is vented into the chamber from the bottom. The stove wall reached a maximum temperature of around 350 °C, indicating that the improved mbaula can cause severe burns if not properly handled. The heat of the ash chamber peaked at around 200 °C. Thus, the hot ash chamber can be used as an oven to roast sweet potatoes or corn. Sweet potatoes are commonly roasted in this way in Zambia in the traditional mbaula. The high surface temperatures indicate that the improved mbaula can be used for space heating during the cold months of the year.

3.2 Pollutant emissions

3.2.1 Carbon monoxide Table 3 shows the mean values for the instantaneous concentrations, emission rates and emission factors for carbon monoxide (CO). The CO concentrations

10 Kaoma, Kasali & Eliegård

Table 3 Mean carbon monoxide concentration, emission factor and emission rate for charcoal in improved mbaula

Mean concentration 229 mg/m" Mean emission factor 343 g/kg Mean emission rate 76 g/h

like all the other pollutant values reported in this study were generated under air exchange rates which ranged from 14.2 to 29.5 volumes per hour with a mean of 23.2 h"1 (see Table A2 in Appendix). The mean CO concentration in the test room was 229 mgm*3 with a range of 150-361 mgm"3 (see Table A2 in Appendix) for the improved mbaula with charcoal. This value is by far above the 30 mgm"3 of CO which the WHO (10) in Europe has recommended as a time weighted average that must not be exceeded for one hour if the general public is to be protected. Thus, if the conditions of the test room are representative for the actual use of the improved mbaula in the homes, there is need for caution. The improved mbaula should be used only in well ventilated places. To reduce the high level of CO, improvements in stove design are necessary. One possibility is to provide secondary air to burn it off. These concentrations translate into a mean CO emission factor (E) of 343 g/kg with a range of 227-624 g/kg (Table A3 of Appendix). These values are close to the mean E value of 391 g/kg*1 and a range of 336-610 g/kg reported for Haitian wood charcoal (HWC) when used in Government of Haiti stove (GHS) (9). Table 3 also shows that the mean emission rate (ER) for charcoal use in improved mbaula was 76 g/h. The ER range was from 55 to 107 g/h"1 (Table A4 in Appendix). These figure are on the lower side of the ER range of 57-473 g/h reported for the GHS-HWC (9).

3.2.2 Sulphur dioxide The sulphur content of the charcoal used in this study was not determined.

However, S02 emissions were consistently detected during charcoal combus­ tion. The results are presented in Table 4 below. The sulphur dioxide concentrations ranged from 0.2 to 0.9 mgm"3, (Table A2 in Appendix). This could be compared to the Canadian indoor air quality guidelines (11) for an acceptable short term exposure range (ASTER) of 1 mg/m3 for 5 minutes. The range for the emission factor was 0.3 - 1.0 g/kg (Table A3 in

Appendix). The mean E(S02) value found in this study was equal to the average E value of 0.6 g/kg found for HWC combusted in a round Haitian

11 Improved mbaula stove (9). The sulphur dioxide emission rate ranged from about 0.1 to 0.2 g/h (Table A4 in Appendix).

Table 4 Mean sulphur dioxide concentration, emission factor and emission rate for charcoal in improved mbaula

Mean concentration 0.40 mg/m* Mean emission factor 0.60 g/kg Mean emission rate 0.1 g/h

3.Z3 Nitrogen oxides The oxides of nitrogen were determined as nitrogen dioxide (N02) and nitric oxide (NO). Table 5 shows the observed emission results. N02 con­ centrations ranged from 1.2 to 5.8 mgm"3, while those for NO ranged from 0.2 3 to 1.3 mgm" (Table A2 in Appendix). The N02 emission factor ranged from 1.8 to 18.6 g/kg while the range for NO was 0.3-3.7 g/kg (Table A3 in Appendix. Emission rates had ranges of 0.4-3.4 and 0.1-0.6 g/h for N02 and NO, respectively (Table A4 in Appendix). The mean nitrogen oxide concentrations found in this study substantially exceeded the 0.48 mgm"3 (5 min) ASTER level (11).

Table 5 Mean nitrogen oxides emission factor and emission rate for charcoal in improved mbaula

NO, NO Mean concentration 2.3 0.6 mg/m' Mean emission factor 4.7 1.0 g/kg Mean emission rate 1.0 0.2 g/h

3.2.4 Aerosol and respirable suspended particulates In this study aerosol was monitored as a real-time indicator of smoke intensity in the room. However, a quantitative relationship between aerosol (smoke) and suspended particulate concentrations was not established as the aerosol was not calibrated with the gravimetric particulate sampler. Table 6 shows the aerosol and respirable suspended particulate (RSP) results obtained for charcoal use in improved mbaula. There was very little visible smoke generated during the combustion of charcoal in improved mbaula. The aerosol concentrations ranged from 0.2 to 0.4 mgm*3. The aerosol emission factor ranged from 0.03 to 0.16 g/kg while

12 Kaoma, Kasali & Ellegård

Table 6 Mean aerosol and RSP concentrations, emission factors and emission rates for charcoal in improved mbaula

Aerosol RSP Mean concentration 0.07 0.75 mg/m3 Mean emission factor 0.09 1.14 g/kg Mean emission rate 0.02 0.25 g/h

the emission rate ranged from 0.01 to 0.03 g/h (Table A2, A3 and A4 in Appendix). Combustion of charcoal in improved mbaula resulted in RSP concentrations which ranged from 0.4 to 1.3 mgm"3 (Table A2 in Appendix) while the emission factor range was from 0.5 to 3.2 g/kg (Table A3 in Appendix). The emission rate ranged from 0.1 to 0.5 g/h (Table A4 in Appendix). The RSP mean concentration values presented in Table 6 are considerably higher than the Canadian domestic indoor air quality 1-h guideline of 0.1 mgm"3 (11).

3.3 Emission Dynamics Knowledge of the instantaneous trends of emissions is essential especially for those intending to incorporate emission control techniques into stove designs. The data can also be used to educate users as to when and where stoves should be used to minimise the risk of pollutant exposures. Figure 3 shows the variation in CO concentrations with time during the combustion of charcoal in improved mbaula. The initial phase was marked by rapid rise in CO levels until the peak occurred after 30 to 50 minutes. Wachter, et al (12) reported that CO emissions from charcoal peaked at 34 minutes from ignition. Islam and Smith (9) also found that in all runs with Haitian wood charcoal CO rose up initially and then fell down towards the end. This is the same trend observed in this study. Burnet et al. (13) noted that in the initial phase of combustion, emissions can increase because volatile fuel gases are generated at a faster rate than they can be efficiently com­ busted. In the present study the highest burn rates occurred in the initial phase of the burn cycle (Table Al in Appendix).

S02 exhibited a similar trend to CO as the concentrations were higher in the initial than in the latter stages of the burn-cycle. Sanborn (14) found that high burn rates corresponded to hotter fuelbed temperatures which resulted in increased S02 emission rates. In this study the initial phase of charcoal combustion was characterised by high burn rates and high tempera­ tures (Figure 2). Smoke is usually a result of the devolatilization of the fuel

13 Improved mbaula

Figure 3 Profile of room concentration of CO for a typical burn cycle (No.8) of the improved mbaula with charcoal. A, B and C denote the different burn phases, AXC denotes the concentration peak used for assessing the air exchange rate.

800

700- CARBON MONOXIDE AXC

]j;i],.,ii.,.jjiiiiii]iiiinii:;]uii;iiii:L»ii.;iiiii:liL.i::ii.UL.u.iijiji»Jirnui!i:»i]:LLij]:iiJMlJ»tJUilliJli»lui]iiiji«j::iJiJiJn.iuiJl.lJJ.Ui«jj;u»uiu«i]ilLn:ri»Lii:;gjiiJ.Li]a].,i' 30 60 90 120 150 180 210 Minutes and this process normally marks the initial phase of combustion. The charcoal used had a low volatile matter content (Table 1) hence smoke generation did not persist for a long time in the burn cycle. In the first 20 minutes of the burn cycle nitric oxide (NO) could be detected and then later nitrogen dioxide started to be detected. There is no readily available explanation for this behaviour of nitrogen oxides, but this was probably due to the fact that the detection level of the NO monitor started from 0.1 ppm, while that for N02 began from 1 ppm. Thus apparently, it needed some time for NOz to reach levels that could be detected by the used monitor. Wendt, et al (15) showed that the presence of sulphur in the fuel can either enhance or inhibit the formation of fuel nitrogen oxides, however, it is not clear whether any such interaction even took place under the conditions of this study. The fuelbed temperatures encountered in this study (Figure 2) were well below the 1300 °C needed for NOx formation from nitrogen in the combustion air and hence the recorded oxides were of fuel origin.

14 Kaoma, Kasali & Ellegård

3.4 Improved versus Traditional Mbaula The improved mbaula was developed to improve upon the performance of the traditional mbaula. Principally the promoters have emphasized the ability of improved mbaula to reduce fuel (charcoal) consumption. The design details of the traditional mbaula have been described previously (4). However, there are similarities and differences which need to be stated.

Figure 4 Traditional (left) and improved mbaula.

The traditional and improved mbaula are both charcoal stoves with a cylindrical shape and are made by tinsmiths from scrap metal. The improved mbaula has one air inlet gap with a door for air adjustment, while the traditional mbaula can have 3-5 such gaps without doors located around the ash chamber. There are no openings of any kind in the combustion chamber wall of the improved mbaula while that of the traditional one is perforated with holes all around it. The previous tests with the improved mbaula were conducted with the air inlet fully opened. Table 7 shows that there is little difference in burn rate and combustion efficiency (CO/C02 ratio) between the stoves. Unexpectedly, the results show that the average stove efficiency (PHU), was slightly lower in the improved mbaula. On the other hand, carbon monoxide emissions were considerably lower from the improved compared to the traditional mbaula.

15 Improved mbaula

Table 7 Mean thermal and emission values for traditional mbaula and improved mbaula

Traditional Improved Burn rate (kg/h) 0.21 0.22 Efficiency (%) 29 25 CO/CO, ratio 0.13 0.11 CO emission factor, (g/kg) 601 343 Power output (kW) 1.7 1.8

Probably the most important reason for the higher emissions in the traditional mbaula is that the stove is so wide that the pot rests immediately on the charcoal bed. This causes constriction of the air flow to the charcoal. In the improved mbaula the pot rests on pot rests located on the top of the stove, which results in better ventilation and increased residence time for hot gases within the combustion chamber. This raises the fuelbed temperature causing the burn rates to increase and consequently enhancing the combustion efficiency of the flue gases. Burnet, et al (13) have shown that higher burn rates which produce increased fuelbed turbulence and temperatures result in lower air pollutant emissions. Thus, operating the improved mbaula with the air inlet door fully opened will result in higher charcoal consumption but lower carbon monoxide emission levels. In the manual of Energy Saving Cookers (16) the users are instructed to close the door halfway when they want their food to cook slowly on improved mbaula. This exercise affects the combustion characteristics of the fuel and hence it was decided to set up experiments to determine the effect of this operation mode on burn rate and CO emission. The stove was operated with the air inlet door fully and halfway opened. A full charge of fuel was used for each mode of operations, but no water boiling test was performed (no pot was placed on the stove). Thus the values obtained cannot be compared to the values obtained in the previously presented water-boiling tests, but can serve for inter-comparison of the operation modes only. The results presented in Table 8 show the mean values obtained. Operating the improved mbaula with the air inlet door halfway open decreased the burn rate by 44%, while the CO emission factor and rate increased by 66 and 20%, respectively. The provision of an adjustable opening on the improved mbaula thus allows the user to save energy but it also increases pollutant exposures for the household. Most of the dishes (maize meal, beans, fish, meat and vegetables) prepared in Zambia require high stove power output for them to be

16 Kaoma, Kasali & Eliegård

Table 8 Effect of air inlet aperture on the thermal and emission characteristics of the improved mbaula

Fully open Halfway open Burn rate (kg/h) 0.36 0.25 CO emission factor (g/kg) 205 341 CO emission rate (g/h) 73 88 Power output (kW) 2.9 2.0

sufficiently cooked in ample time. There is, therefore, every likelihood that the halfway mode will be mostly used for space heating purposes, inside the house, after the cooking has been stopped. It is also necessary to educate the improve mbaula user on the adverse environmental effects of this energy saving exercise.

4 CONCLUSIONS This study has revealed certain hitherto unknown aspects concerning energy flow and emission patterns arising from the combustion of charcoal in the improved mbaula. The following has so far been established in this study:

1 The thermal performance of the improved mbaula was comparable to most other charcoal cookstoves of similar design and make. It can, therefore, be adequately used for cooking and water heating.

2 The high surface temperature of the stove stresses the importance of adequate safety measures in operation the improved mbaula. This is especially important if there are children present while cooking, or if the stove is used for space heating in the night.

3 The combustion of charcoal in improved mbaula generated pollutant emissions which persisted throughout the burn cycle until burn-out. The levels of most of these emissions exceeded acceptable levels of indoor air quality standards. Charcoal use in improved mbaula should, therefore, be conducted in well ventilated places.

4 Compared to the traditional charcoal brazier (mbaula),the improved mbaula consumed slightly more charcoal when operated with the air inlet door fully opened. Fuel savings with improved mbaula can only be achieved if the users are properly educated about the functions of

17 Improved mbaula

the stove's adjustable opening, and accept the lower burn rates which this causes.

Operating improved mbaula with the air inlet door halfway open reduced fuel consumption but increased the emission levels. The technique of reducing air flow rate through the stove to improve efficiency should be critically re-examined by stove designers, especially in the case of unvented cook stoves.

5 ACKNOWLEDGEMENT The authors are grateful to the Stockholm Environment Institute (Sweden) for funding this research study and to the National Council for Scientific Research (Zambia) for providing the physical and manpower resources.

6 REFERENCES

1. Hibajene, S.H., Chidumayo, E.N. and Ellegård, A., 1993, Summary of the Charcoal Industry Workshop. Policy and Management: Challenges for the Future, May 10-14, 1993, Siavonga, Zambia. Stockholm Environment Institute; Energy, Environment and Development Series No 24

2. Nsofwa, G., 1991, Dissemination of Information and Introduction of New Technologies in Communities - Problems and Solutions. Paper presented to the Work-shop on Energy Saving Cookers in SADCC Countries, 12-17 August, 1991, Commonwealth Youth Centre, Lusaka, Zambia.

3. Ahuja, D.R; Joshi, V; Smith, K.R. and Venkataraman.C, 1987, Thermal Performance and Emission Characteristics of Unvented Biomass-burning Cookstoves: A proposed standard method for evaluation. Biomass 12, 247-270.

4. Kaoma, J. and Kasali, G., 1994, Efficiency and Emission Characteristics of two Zambian Cookstoves Using Charcoal and Coal Briquettes. Stockholm Environment Institute; Energy, Environment and Develop­ ment Series No. 36

18 Kaoma, Kasali & Eliegård

5. Joseph, S; Shanahan, Y. and Young, P., 1982, The comparative performances of Kenyan Charcoal Stoves. Intermediate Technology Development Group, London.

6. Osuwan, S. and Boonyakiat, K., 1982, Study of variables affecting charcoal stove efficiency, Journal of Chemical Engineering, Food and Fuel Technology, 5-59.

7. Chomcharn, A; Jonjitirat, P; Ultarapong, P. and Andsrisom, B., 1984, Improved Biomass Cooking Stove for Household Use: A report submitted to the National Energy Administration, Ministry of Science, Technology and Energy, p33.

8. Kummongkol, P., 1990, Standard Method of Testing Cooking Stoves Performances. Final report submitted to the AEAN Subcommittee on Non-conventional Energy Research and ASEAN-Australia Energy Cooperation Programme. King Mongkut's Institute of Technology, Thonburi, Bangkok, Thailand.

9. Islam, N.R. and Smith, K.R., 1989, Combustion Tests on Lignite Briquettes. Paper presented at the Third Coal Research Conference (October), Wellington, New Zealand.

10. WHO, 1987, Air Quality Guidelines for Europe, WHO Regional Publications, European series, No. 23, ISBN 92-890-1114-9.

11. Canadian Domestic Indoor Air Quality Guidelines, Federal Provincial Advisory Committee on Environment and Occupational Health. Environmental Health Directorate, Health Protection Branch, April, 1987, Canada.

12. Wachter, A.E; de Priest, J.C. and Ahmed, N., 1992, Comparative study of Pakistan Coal Briquettes and Traditional Pakistan Domestic Fuels. Report prepared by Oak Ridge National Laboratory for the USDOE DE-AC05-84OP21400; ORNL/TM-12087

13. Burnet, G.P., Edmisten, G.N., Tlegs, E.P., Houck, E.J. and Yoder, A.R., 1986, Particulate, carbon monoxide, and acid emission factors for residential wood burning stoves. Journal of Air Pollution Control Association, 36, 1012-1018.

19 Improved mbaula

14. Sanborn, C.R., 1983, Evaluation of emissions from the residential coal- fired space heaters; Proc. 76th Annual Meeting of the Air Pollution Control Association, Atlanta, Georgia, Vol. 83, pl4

15. Wendt, J.O.L., Corley, T.L. and Marcomb, J.T., 1977, Interactions between sulphur oxides and nitrogen oxides in combustion processes. Proceedings of the 2nd stationary source combustion symposium, Vol. IV: Fundamental combustion research, EPA-600/7-77-073d, NTIS PB 274- 029/AS, July 1977, pp 101-137.

16. Manual of Energy Saving Cookers (1991). Zambia Alliance of Women, Lusaka, Zambia.

20 APPENDIX

21 Table A.1 Thermal combustion characteristics of charcoal in improved mbaula

BURN RATE THERMAL EFFICIENCY TEST DURATION COM­ BUSTION kg charcoal / hour % heat utilization (PHU) hours EFFICI ENCY BURN BURN PHASE BURN BURN PHASE BURN BURN PHASE BURN CYCLE CYCLE CYCLE CYCLE (CO/CO, A B C AVG A B C AVG A B C TOTAL ratio)

1 0.38 0.23 0.11 0.22 21 26 30 26 0.85 0.77 1.33 2.95 0.12 2 0.33 0.19 0.04 0.19 16 30 39 28 1.18 0.83 1.12 3.13 0.13 3 0.33 0.15 0.07 0.16 18 31 19 23 1.05 1.00 2.07 4.12 0.13 4 0.30 0.17 0.01 0.19 18 20 19 19 1.18 1.40 0.55 3.13 0.13 5 0.49 0.34 0.11 0.29 23 25 39 29 0.72 0.57 0.98 2.27 0.11 6 0.36 0.26 0.12 0.24 21 26 30 26 0.83 0.68 1.05 2.56 0.11 7 0.33 0.28 0.08 0.19 20 23 28 24 0.90 0.72 1.83 3.45 0.09 8 0.37 0.21 0.13 0.21 21 27 21 23 0.82 0.70 1.58 3.10 0.10 9 0.37 0.21 0.11 0.22 15 27 18 20 0.95 0.70 1.30 2.95 0.12 10 0.38 0.20 0.16 0.25 21 40 25 29 0.85 0.62 0.95 2.42 0.09

MEAN 0.36 0.22 0.09 0.21 19 28 27 25 0.93 0.80 1.28 3.01 0.113 S.D. 0.05 0.05 0.04 0.04 2 5 8 3 0.15 0.23 0.43 0.51 0.015

Notice that the tost duration of each burn phaae generally la 0.25 h longer than the time It taket to bring the water to the boiling point

22 Table A.2 Air exchange rates and mean pollutant concentrations generated during combus­ tion of charcoal in improved mbaula

Burn Air ex­ Mean pollutant concentration (mg/m1) cycle change rate CO SO, NO, NO Aerosol RSP (h')

1 14.2 259 0.56 2.58 0.44 0.05 0.59 2 15.0 275 0.87 1.67 0.38 0.14 0.56 3 21.0 176 0.44 5.83 0.38 0.02 0.42 4 29.3 229 0.32 1.35 0.04 1.26 5 20.0 361 0.34 1.85 0.84 0.11 1.17 6 29.5 195 0.30 0.33 0.07 0.67 7 24.6 150 0.31 1.25 0.07 0.70 8 22.8 244 0.23 1.32 0.07 0.57 9 28.7 178 0.31 1.34 0.06 0.89 10 26.4 218 0.32 0.18 0.06 0.72

Mean 23.2 229 0.40 2.26 0.56 0.07 0.75 S.D. 5.61 61.0 0.19 1.64 0.40 0.03 0.27

Table A.3 Pollutant emission factors for the combustion of charcoal in improved mbaula

Emission factor (g emission / kg charcoal burned) Burn cycle CO SO, NO, NO Aerosol RSP

1 227 0.49 2.26 0.38 0.04 0.48 2 329 1.04 3.74 0.45 0.16 0.62 3 297 0.75 18.6 0.64 0.03 0.66 4 625 0.86 3.68 0.10 3.17 5 342 0.32 1.76 0.79 0.09 1.02 6 350 0.53 0.59 0.12 1.10 7 238 0.48 1.99 0.10 1.03 8 351 0.33 1.91 0.09 0.76 9 328 0.58 2.48 0.10 1.52 10 344 0.51 0.29 0.09 1.04

Mean 343 0.59 4.68 0.97 0.09 1.14 S.D 109 0.23 6.17 1.20 0.04 0.77

23 Table A.4 Pollutant emission rates from charcoal combustion in improved mbaula

Emission factor (g emission / kg charcoal burned) Bum cycle CO SO, NO, NO Aerosol RSP

1 55 0.12 0.54 0.09 0.01 0.12 2 62 0.19 0.70 0.08 0.03 0.11 3 55 0.14 3.44 0.12 0.01 0.12 4 100 0.14 0.59 0.02 0.51 5 107 0.10 0.55 0.25 0.03 0.32 6 86 0.13 0.14 0.03 0.27 7 55 0.11 0.46 0.02 0.24 8 83 0.08 0.45 0.02 0.18 9 76 0.13 1.99 0.02 1.03 0 85 0.13 0.07 0.02 0.26

Mean 76 0.13 0.96 0.19 0.02 0.25 S.D. 19 0.03 1.10 0.18 0.007 0.12

24 SEI Board

Chairman: Karin Söder Sweden

Prof Umberto Colombo Italy Elizabeth Dowdeswell Canada Dr Peggy Dulany USA Prof Gordon T. Goodman UK Mrs Fatima Ahmed Ibrahim Sudan Prof Thomas B. Johansson Sweden Ms Kerstin Niblaeus Sweden SirShridath Ramphal Guyana Dr Katsuo Seiki Japan Dr Alvaro Umafia Costa Rica Dr Alexi V. Yablokov Russia

STOCKHOLM ENVIRONMENT INSTITUTE Director: Prof Michael J. Chadwick Director SEI-Boston: Dr Paul Raskin Vice Director: Dr Lars Kristoferson Director SEI-Tallinn: Dr Tönis Kaasik Head of Administration: Ann-Charlotte Bradley Information Manager: Dr Arno Rosemarin Librarian: Krister Svärd

Stockholm Environment Institute SEI at York Address: Lilla Nygatan 1, Box 2142 Address: University of York, S-103 14 Stockholm, Heslington Sweden York YOl 5DD, UK Telephone: Int +46 8 723 0260 Telephone: Int+44 904 43 2897 Telex: 19580 SEI S Telex: 57933 YORKUL G Telefax: Int +46 8 723 0348 Telefax: Int +44 904 43 2898

SEI-Boston SEI-Tallinn Address: 11 Arlington Street Address: Box 160 Boston, MA 02116-3411 EE-0090 Tallinn USA Estonia Telephone: Int +1 617 266 8090 Telephone: Int + 372 2 601 844 Telex: 279926 ESRG BSN UR Telefax: Int + 372 2 440 982 Telefax: Int +1 617 266 8303 Stockholm Environment Institute The Stockholm Environment Institute (SEI) was established by the Swedish Parliament in 1989 as an independent foundation for the purpose of carry ing out global environment and development research. The Institute is governed by an international Board whose members are drawn from developing and industrial­ ized countries worldwide. Central to the Institute's work have been activities surrounding the Rio UNCED conference, and previous to this, the Brandt and Palme Commissions and the work of the World Commission for Environment and Development. Apart from its working linkages with the relevant specialised agencies of the UN system, a particular feature of SEP s work programme is the role it has played in the development and application of Agenda 21, the action plan for the next century. A major aim of SEI's work is to bring together scientific research and policy development. The Institute applies scientific and technical analyses in environmental and development issues of regional and global importance. The impacts of different policies are assessed, providing insights into strategy options for socially responsible environmental management and economic and social development. The results of the research are made available through publications, the organisation of and participation in conferences, seminars and university courses, and also through the development of software packages for use in the exploration of scientific problems. SEI has also developed a specialised library which functions as a central catalyst in the short-term and long-term work of the Institute.

Research Programme A multidisciplinary rolling programme of research activities has been designed around the following main themes, which are being executed via internationally collaborative activities with similar institutions and agencies worldwide:

Q Environmental Resources, including energy efficiency and global trends, energy, environment and development, and world water resources;

• Environmental Technology, including clean production and low waste, energy technology, envi­ ronmental technology transfer, and agricultural biotechnology;

• Environmental Impacts, including environmentally sound management of low-grade fuels, climate change and sustainable development, and coordinated abatement strategies for acid depositions;

• Environmental Policy and Management, including urban environmental problems, sustainable environments and common property management; and

Q POLESTAR, a comprehensive modelling and scenario-based activity, investigating the dynamics of a world with 10 billion people by the middle of the next century.

SEI's Network SEI has chosen a global network approach rather than a more traditional institutional set-up. The work programme is carried out by a worldwide network of about 60 full- and part-time and affiliated staff and consultants, who are linked with the SEI Head Office in Stockholm or to the SEI Offices in Boston (USA), York (UK) and Tallinn (Estonia). SEI has developed a large mailing register to communicate to key members of society in government, industry, university, NGOs and the media around the world. r~*\ •* 1 "I" STOCKHOLM ^S |H I ENVIRONMENT C_y 1 J JL INSTITUTE International Institute tor Environmental Technology and Management Postal address: Box 2142, S-103 14 Stockholm, Sweden Telephone: +46-8-723 0260 Telex: 19580 seis Telefax: +46-8-723 0348 E-mail: [email protected]