Gasifier stoves - Science, technology and field outreach

H. S. Mukunda*, S. Dasappa+, P J Paul*, N K S Rajan*,Mahesh Yagnaraman∗∗, D. Ravi kumar++, Mukund Deogaonkar∗∗ ∗

February 28, 2010

Abstract

This paper is concerned with the developmentof a new class of single pan high efficiency,low emission stoves, named gasifier stoves thatpromise constant power that can be controlledusing any solid biomass fuel in the form of pel-lets. These stoves use battery-run fan basedair supply for gasification (primary air) andfor combustion (secondary air). Design withthe correct secondary air flow ensures near-stocihiometric combustion that allows attain-ment of peak combustion temperatures withaccompanying high water boiling efficiencies(up to 50 % for vessels of practical relevance)and very low emissions (of carbon monoxide,particulate matter and oxides of nitrogen).The use of high density agro-residue based pel-lets or coconut shell pieces ensures operationalduration of about an hour or more at powerlevels of 3 kWth (∼12 g/min). The principlesinvolved, the optimization aspects of the de-sign are outlined. The dependence of efficiencyand emissions on the design parameters are de-scribed. The field imperatives that drive thechoice of the rechargeable battery source andthe fan are brought out.

The process development, testing and inter-nal qualification tasks were undertaken by In-

∗* Department of Aerospace Engineering, + Centrefor sustainable technologies, Indian Institute of Science,Bangalore 560 012, ∗∗ First Energy Private Limited,Pune ++ General Electric, India

dian Institute of Science. Product develop-ment including improvements on the ceramicchamber, battery pack, the fan and the finalshape of the stove on the one hand and thefuel pellet production process on the other weredealt with by First energy Private Ltd (de-noted by FEPL). The process of business devel-opment was of course completely dealt with byFEPL. Close interaction at several times dur-ing this period has helped progress the projectfrom the laboratory to large scale commercialoperation.

This technology was initially developed asa business by BP, India (earlier) and subse-quently taken on by FEPL. At this time, overfour hundered thousand stoves have been soldin four states in India. Oorja-Plus is widelysold currently and Oorja-Super is yet to enterthe market. The implications of such devel-opments to the domestic cooking scenario ofIndia are briefly brought out.

Keywords: Gasifier stove, domesticstove, biomass stove

1 Introduction

Research, development and dissemination onbiomass based domestic combustion devicesotherwise termed cook stoves has a long his-tory of over 5 decades. Many of the devel-opments have occurred due to intuitive ap-proaches to examine heat transfer aspects rel-egating the combustion issues to a peripheral

1

state. A broad summary of the scientific workperformed till mid-eighties is reflected in thework of Prasad et al1. Over a hundred designsfrom all parts of the world have been docu-mented and these are discussed in a website de-voted to cook stoves (www.bioenergylists.org).Stoves are made of metal, mud, refractoriesof various qualities with and without chimney.Chimney based stoves are considered superiorto chimney-less stoves because any emission istaken out of the kitchen thus preserving theindoor air quality. However, these emissionscreate a load on the environment somethingthat has become a matter of serious concern inthe last several years. Further, even the sin-gle pan stoves without chimney have air sup-ply by free convection since electric supply wasnot to be found in most rural dwellings whichis where these stoves were largely expected tobe used. Most of the these stoves had utiliza-tion efficiencies assessed by water boilling testsbetween 10 to 20% (see for instance, Bhat-tacharya et al2,3) while one slightly complexdesign shows an efficiency of 40% at smallpower levels (see Mukunda et al4 for details).It has been known that both kerosene andLPG (liquified petroleum gas) stoves have ef-ficiencies as high as 65 and 70 %. The scien-tific question as to what limits the efficienciesin biomass stoves in comparison to keroseneand LPG stoves had remained un-examinedtill Mukunda et al4 studied this aspect ex-perimentally. They specifically showed thatthe efficiency is directly related to the oper-ation of the stove at near-stoichiometric con-ditions that lead to peak combustion tempera-ture of the product gases; it is this temperaturethat influences the heat transferred to the ves-sel within the limited bottom area available.Hence, the stove operation can be expected tobe the best when the air-to-fuel ratio is main-tained near stoichiometry, albiet on the leanside to ensure minimum CO emissions; limit-ing the emissions of NOx is taken care of natu-

rally by the lower combustion temperatures ofbiomass-air system as thermal NOx is the pre-dominant source. Limiting the velocities in thefuel residing zone will help limit the particulatecarry over.

The stove designs that have (a) free convec-tion based air ingestion and (b) biomass loadedas and when it is thought fit into the hot com-bustion zone will function with air-to-fuel ra-tios that can be very rich or very lean. Theformer condition leads to sooting and the lat-ter to smoking. Also parts of the fuel over thestove may still be cold leading to air flow outof the stove body without participating in thecombustion while other areas are deprived ofthe needed air for combustion. Coupled to allthese, the biomass may either be dry or verywet. The latter condition will lead to poorcombustion with heavy smoking. Thus im-provement of biomass based stove combustioninvolves biomass quality as well as combustionaerodynamics, a view point that has not beenadequately reflected either in the stove litera-ture or in the development of improved cookstove initiatives in the past.

Typical specifications of a domestic stovecan be derived from comparable stoves basedon kerosene or LPG that have a input powerlevel of about 2.0 kWth (∼ 3 g/min ofLPG). Accounting for efficiency differences,the biomass stove input power level can be setat about 3 kWth (12 g/min of biomass). Thestove must allow continuous operation and ifnot, must last the complete cooking durationthat has been estimated as about one hour.One would expect to have clean combustionwith minimal emissions. Minimizing emissionsof CO and other unburnt products of combus-tion calls for maximizing the combustion effi-ciency and minimizing CO2 emissions impliesmaximizing the utilization efficiency (or waterboiling efficiency on vessels of practical rele-vance). One would desire a smooth start-upand some control on the power level. The life

2

of the stove must be large, at least for two tothree years. In doing all these, the cost of thestove must be affordable.

Thw work described here is arranged as fol-lows: (a) principles of gasifier stove, (b) freeconvection vs. forced convection operation, (c)gasifier stove vs. in-situ combustion stove, (d)fuel for the stove, (e) the stove, efficiency andemissions, (f) power pack related aspects, (g)stove performance evaluation by others, and(h) comparison with other developments.

The broad implications of these develop-ments to the country and other countries thathave similar requirements or challenges are alsobrought out.

2 Principles of gasifier stove

Gasifiers are essentially devices that enableconverting solid fuel to gaseous fuel by a ther-mochemical conversion process. This processinvolves substoichiometric high temperatureoxidation and reduction reactions between thesolid fuel and an oxidant - air in the presentcase. This is arranged such that air and thegas passes through a fixed packed bed.

The gasification knowhow has a long history;it became particularly important in the Euro-pean scene during the World war II when fos-sil fuel availability was scarce. Research anddevelopment into gasification process came toa low after fossil fuel availability became nor-mal and their prices low. The area has be-come active in the last thirty years particu-larly in oil importing countries like India. Re-search relevant to the objectives of this papercan be traced to rice hull gasification systemsthat used a vertical cylinder filled with rice hulland air drawn through the packed bed from thetop. While these were first developed in Chinaas early as in 1967 (see Bhattacharya5), the sci-entific investigation was conducted by Kaupp6.In this reactor, shown on the left side of Fig-

Figure 1: The left side of the figure is schematicview of a rice hull gasifier. Note that the airflows from the open top towards the bottom.The right side is a schematic of a reverse down-draft gasifier stove where the gasification airflows from bottom to top

ure 1, one can use biomass like wood chips,and pellets apart from ricehulls for which itwas first conceived. If now the reactor is oper-ated such that air flows from the bottom andthe fuel surface at the top is lit, one wouldget combustible gases that will burn above thethe top surface with ambient air or with ad-ditional air supplied towards the top. Such aconfiguration, shown on the right side of Figure1 termed reverse downdraft gasifier constitutesthe essence of a gasifier stove. An importantconsequence of this mode of operation is thatthe gas exiting from the top of the packed bedbears a fixed ratio to the amount of air intro-duced for gasification (primary air flowing fromthe bottom). The reduction reactions follow-ing the oxidation limit the amount of fuel tobe consumed due to the endothermic natureof these reactions. The interesting feature isthat the relative amounts of fuel consumed andair introduced remain the same and increasedamount of solid is consumed when primary airflow rate increases. Thus the power of the stove

3

is proportional to the primary air flow rate.The gases coming out of the bed will be at atemperature of 800 to 1100 K and will be com-posed of CO, H2, CH4, H2O (as gas), somehigher hydrocarbons and N2. These gases areburnt to CO2 and H2O with a second streamof air which is introduced in the top region forthis purpose.

Based an early work of La-Fontaine andReed7, Reed and colleagues developed a freeconvection based gasifier stove (Reed andLarson8) and have subsequently discussed thedevelopment of forced convection based gasi-fier stove (Reed et al9,10). The biomass islargely wood chips and the forced convectiondepended on a fan; it formed the basis ofcamp stove and this is indeed how it got mar-keted. While the broad principles were clearat this time, the aspects of process optimiza-tion and development of components needed tobe addressed to deal with the affordability is-sue in emerging economies like India with fuelsdepending on agro-residues rather than woodchips. One of the key drivers for the devel-opment was the evolution of low cost storagebattery based fan; the idea of building a do-mestic stove became meaningful when the fansfor supplying air were being built in numbersfor computer applications and at this stage,low power fans could be obtained or builtat less than 5 USD per system including theelectrics. Work at the IISc laboratory on gasi-fication systems had begun in 1983 (Srinivasaand Mukunda11) and reached reasonable ma-turity by the end of nineties, and the workon the development of stoves of high efficiencybegan in 1999 with a study to determine thedependence of various parameters on the wa-ter boiling efficiency and emissions with severalagro-residues.

3 Free vs. Forced convectionstoves

Nearly all the domestic cook stoves over theworld are free convection driven. In several ofthese designs the free convective driving po-tential is so small that the smallest of ambientdisturbances can affect the air flow through thestove. Even in better of these designs witha column of hot zone (due to the combus-tion chamber) that enables stabilize the flowthrough the stove, flow disturbances caused byvarying heat release across the section of thecombustion chamber due to randomly locatedand moved-around pieces of “firewood” leadsto widely varying air-to-fuel ratios locally. Thiskind of a variation reduces the peak flame tem-perature most of the time and leads to emis-sions of CO, unburnt hydrocarbons (UHC) andparticulates significantly. Even when better ef-ficiencies are obtained, the stove operation be-comes very rich and significant sooting will re-sult. This is particularly because, faster cook-ing is understood to be obtained at a largerfuel burn rate and this is achieved by intro-ducing more fuel into the combustion space.Getting good emission performance from suchstoves is usually a difficult task. Thus labo-ratory test results and operational data fromrealistic environment can be widely different.Much of the current debate that is going onwith regard to the standards and protocols fortesting stoves arising from the fact that the labtests are different from real experiments is dueto this fundamental aspect of the free convec-tion stove; by design, the efficiency as well asemissions are dependent on the user and theways the stove-fuel combination is used. Freeconvective stoves were and even now, are con-sidered inevitable because of the i) perceptionthat such stoves provide the “choice” and flex-ibility and allow the users to continue to usewhatever non-processed raw biomass and/or

4

wood they were using, ii) use of any forceddraft that requires electricity is seen as diffi-cult and/or unavailable. The point (i) needs tobe challenged given the variable and distinctlypoorer performance on efficiencies and emis-sions that have been observed and recorded(to be brought out subsequently). As far aspoint (ii) is concerned, improvement in elec-tricity and infrastructure and technology ofsmaller fans and power sources have given thereal option to provide forced convection in thestove designs to improve performance and re-duce emissions. All these do not imply thatthe rural environment enjoys the availabilityof electricity all the time; perhaps, they haveit over a few hours a day. Yet the fact thatelectricity is available over some period can bemade use of for the use of electricity enabledstove designs. Also, it is possible that noth-ing better than free convection stove can becontemplated in regions deprived of electricitytotally. This should not mean that other re-gions that have electricity support even overthe part of day should be deprived of moderntechnology interventions.

4 In-situ combustion stove vs.gasifier stove

Another stove design that has been revealedin recent times that has had a limited dissem-ination is a in-situ combustion stove. Such astove is not different from a furnace in princi-ple. One supplies enough air for combustion di-rectly around the fuel zone so that some partsof the fuel will undergo volatilization and thevolatiles will burn with the air around, otherparts that have become char will also simul-taneously be oxidized. In some stoves (likePriyagni in India) the air flow occurs throughthe grate and over it just from around a cham-ber (of short height) to complete the combus-tion process. While direct combustion systems

are common at large power levels where fuelfeeding can be automatic and combustion man-agement is relatively easy, at small through-puts like in a domestic stove, fuel loading hasto done at frequent intervals (typically everyfew minutes apart) manually to ensure com-plete combustion with smaller packets of fueland this may be considered a burden on theuser in comparison to dealing with conven-tional stoves. The power variation with timeis dependent on this attention. If perchance,the amount loaded at one time is more thanwhat the system can carry, significant emis-sions will result. Here again it may turn outthat a trained operator in a laboratory can getgood emission performance, but users in thefield may end up with poor emission perfor-mance.

In comparison to the above mentioned de-signs, gasifier stoves with loaded fuel providean alternative. Gasifier stove that has loadedfuel pellets/pieces can be conceived from theend user as a starting point. Starting withthe amount of cooking energy needed to cookone meal, the design of the stove has been ar-rived at in combination with amount of fuelto be loaded in such a way that the operationis relatively foolproof. The stove operation isdependent on thermo-chemistry which modu-lates the gas production process. At any fixedpower level, the gas composition and temper-ature remain steady and with appropriate sec-ondary air flow, the flame will get maintainedat its peak temperature in a near-steady mode.However, when power variation is demanded,it takes a transition time (of about a few min-utes) much unlike a gas stove since the thermalinertia in the hot fuel bed has to be overcomein the process of reaching a new steady state.This design while providing flame control tothe users, also makes it less user-dependent forachieving desired efficiency and emissions andit may be expected that field results would beclose to those obtained at the laboratory.

5

Table 1: National fuel usage in rural and urban households and energy efficiency, hh = hosehold,mhh = million household, mmt = million metric tonnes, yr = year, t = tonnes, GJ = GegaJoules

Fuel Rural Urban Fuel used Unit fuel Unit energytype mhh mhh mmt/yr t/yr/hh GJ/yr/hhFirewood 87 15 250 2.5 40.0Agro-residue 20 2 120 5.5 77.0Cowdung cake 20 2 95 4.3 55.0Coal, coke 2 2 6 1.5 27.0Kerosene 2 8 5 0.5 21.0LPG 9 25 8 0.24 5.7Others 1 2 - -Total 141 66 465a 3.2a,b 47a,b

a includes solid bio-fuels only, b national averages

5 Fuel for the stoves

Historically, biomass stoves imply “firewood”stoves. It is generally thought that firewood ofany size, shape should be acceptable; moisturein the firewood is known to be undesirable, butthere is no rigor in ensuring dry biomass usein stoves. The national statistics on fuel use incook stoves in rural and urban environments ispresented in Table 1. This table is composedfrom the data provided in Ravindranath andHall12, Shukla13, Kishore et al14 and Anon15.There differences between various studies andthe data of Table 1 can be expected to haveinaccuracies up to 15 %.

The data in the table is very reveal-ing. While wood and agro-residues are bothbiomass, the amount of agro-residues used ona per house-hold basis is nearly twice thatof wood. While it is generally understoodthat wood use itself is inefficient, the degreeof wastefulness of agro-residues is enormous, afact about which there is very little apprecia-tion all-round. If developing improving cookstoves on firewood is considered important, itis far more important to develop stoves to burnagro-residues that are light and odd shaped toobtain high efficiency and reduce the emissions.The magnitude of the use of cowdung cake

as a source for fuel is non-insignificant, butits use is about as energy-inefficient as agro-residues. However, the emissions from its useare significant and any improvement in the useof cowdung cake should address this aspect aswell. Coal is used in a wasteful way largely be-cause of ignition problem. Many of the stovesare lit in the open for the volatiles to escape(about 30 % in comparison to biomass with 70% volatiles) until coal becomes virtually cokeand its combustion becomes vigorous. Chinathat has encouraged a large production of coal-powder based beehive briquettes has serious in-door air pollution problems related to this fact(Sinton and Smith16).

LPG and kerosene are more sought-after fu-els and are in the upper region of the energyladder. Hence, they turn out to be impor-tant as reference for performance comparison.Kerosene is a fuel used to a larger extent bythe urban poor with relatively small kitchensand cannot afford the costs of LPG for cook-ing. Both these have higher water boiling ef-ficiency. Laboratory experiments have shownwater boiling efficiencies of 70 to 75 % for LPGstoves and 60 to 65 % for kerosense stoves.Kerosene use as fuel on a per-household basisappears large while the usage of LPG seems

6

not unreasonable (one 14 kg cylinder for threeweeks for a family of five).

The calorific values of biomass, kerosene andLPG are 16, 42, and 45 MJ/kg and on this ba-sis, one would have expected kerosene usageto be about 30 % higher than LPG allowingfor differences in the utilization efficiency. Per-haps, the magnitude reported on kerosene usefor cooking may be inaccurate as it is gener-ally known that significant amount of kerosenebought under public distribution scheme atsubsidized prices is sold away as either cook-ing fuel or fuel for adulteration with gasolineat higher prices.

The calorific value ratio coupled with effi-ciency differences allows speculation on howmuch of solid biomass is needed for domes-tic cooking on a national scale. The equiva-lent of 0.24 tonnes/yr/hh of LPG translates toabout 1.20 t/yr/hh [0.24 × (45/16) × (70/40)].Acheiving this implies that one would aim ata total solid biofuel use for cooking of 253mmt/year as against the current estimate of465 mmt/year. The magnitude of the task canbe understood if we note the current efficien-cies of firewood stoves as 20 %, of agro-residuebased stoves (or their use in the same firewoodstoves) at 10 % and cowdung cakes as 12 %.Enhancing the efficiency of the use of agro-residue based fuels and cowdung cakes mustoccupy the highest attention, next to which isfirewood. Cowdung can perhaps be integratedinto the strategy for better fuel making withoutany special stove design for cowdung cakes; fu-els based on agro-residues and cowdung shouldbe dealt with as a separate development task.

Lastly, it is noted from Table 1 that the aver-age energy per year per person is estimated at 5to 10 GJ in Ravindranath and Hall (1995) fromsome detailed surveys. The present estimate of8.9 GJ/person/year (the average household inrural India is 5.3 as per the recent census doc-uments) falls within the range obtained earlier.

One can get an assessment of the cooking

energy needs by determining the amount offood cooked per day, assessing how much ofsolid biofuel is needed to achieve this. Table2 contains the data of the average monthlyconsumption of food items in India drawnfrom a recent study by Chatterjee et al17.The amount of fuel required for cooking theseitems (excepting for meat/fish/egg item) asdetermined from actual cooking experiments(Mukunda et al4) in a biomass stove with 40 %water boiling efficiency are also shown in Ta-ble 2. Using these data, the amount of fuel re-quired to cook food for one person for a day isobtained (food/p/d). The last two columns in-dicate the calories in each of the food items andthe calorie intake in the food per day. As canbe seen from the table, the total fuel per per-son per day is 136.2 g and for a house hold thistranslates to 722 g of biomass fuel/day whichimplies 0.26 t/yr/hh. This will perhaps be anabsolute lower bound of the possibility. Thedifference between this value and the one de-duced earlier - 1.2 t/yr/hh is not small andmay be related to the difference in food intakebetween LPG users and others and needs rec-onciliation. At this point, it is appropriate toconclude that a more realistic target would be1.2 t/yr/hh of biomass.

5.1 Issues with agro-residues andcowdung

Most agro-residues are characterized by sea-sonal availability and low intrinsic as well asbulk density. Typically, intrinsic density variesfrom 50 to 200 kg/m3 where as bulk densitywould be 50 to 70 % of this value. Thereis also substantial moisture (up to 50 %) atthe time the material is harvested. The veryfact that densities are low permits loss of mois-ture even in open air storage, the extent of lossdepending on the ambient conditions. Trans-portation and combustion are affected by theseaspects. Drawing from the practice of firewood

7

Table 2: Average food consumption, energy intake, fuel required for cooking and cooking energy;cpm = consumption per month, cpd = consumption per day, m = month, p: person, d: day

Food item cpm cpd fuel/food Fuel/p/d Food cals Calorieskg/p/m g/p/d g/g g/p/d cals/g Cals/d

Rice 7.2 240.0 0.23 55.2 1.5 360Wheat 4.4 146.7 0.23 33.7 3.0 440Other cereals 1.3 43.3 0.23 10.0 1.0 43Pulses 0.9 30.0 0.15 4.5 0.7 1Dairy 4.2 140.0 0.04 5.6 1.0 140Edible oil 0.5 16.7 0.06 1.0 9.0 150Meat/fish/egg 1.6 53.3 0.30 16.0 4.0 213Veg/fruit 10.2 340.0 0.03 10.2 0.4 136Sugar/spices 1.5 50.0 3.5 175Processed food 1.6 53.3 3.0 160Beverages 2.3 76.7 3.0 230Total 136.2 2048

transportation, it is understood to be econom-ical to transport about 10 tonnes of firewoodover hundreds of kilometers, shorter the dis-tance better. A truck carrying 10 tonnes offirewood whose typical density is upwards of500 kg/m3 can carry less than a few tonnesof agro-residues economically. The region tobe covered to maintain a continuous supplyof a variety of these residues becomes large,many times lerger than economics can support.Hence, densification is an important element inthe process. A question then arises - is it notgood enough to densify the material to as muchas firewood (implying about 500 kg/m3)? Theanswer to this question comes from the stovethat uses this material. Approaching high effi-ciencies of a LPG stove demands that the com-bustion volume be brought down to as low avalue as possible. This can be obtained by in-creasing the density of the fuel. Also in a stovewith a fixed storage volume for the fuel likethe one considered here, increasing the den-sity helps in reducing the overall chamber vol-ume that would have benefits of lower inertmaterial content and associated heat loss. Aquestion that next arises is the level to whichthe material must be densified. Here, pactice

in industries suggests that binderless briquet-ting achieves a density of 1000 to 1100 kg/m3.This would form an upper limit to which thematerial must be densified. Since the require-ment in stoves will turn out to be small sizedpieces, one needs to produce pellets, typicallyof 10 to 12 mm diameter and up to 40 mmlong to obtain good packing density. Produc-ing pellets of this size at large throughputs inan economical way has many challenges in theprocess as well plant throughput sizing both ofwhich have not been addressed adequately inthis country, yet. Pellet making and briquet-ting have very different process fundamentals.Briquetting process uses very high pressures, ofthe order of 1200 atms to generate heat due tofriction between the material and the die wallto raise the temperature up to 350 ◦C to enablelignin to be released. This lignin acts as a pow-erful binder. Pelleting process uses moistureor steam in extruding the material through asmall die; the temperatures achieved are nothigh, typically, around 100 to 120 ◦C. Underthese conditions, the crude protein present inthe biomass is softened and helps in the bind-ing process. The densities achieved are usuallylower than in binderless briquetting and are

8

typically about 600 to 800 kg/m3. The pres-ence of crude protein is important and henceany kind of ingredients having some amountwould aid in the process and enahancing thethroughput of the system.

Cowdung is a major cooking fuel in north In-dia, perhaps in conjunction with agro-residuesor firewood. Huge mounds of dry cowdung areset around houses to enable peal-off and usein the stoves. Cowdung itself has componentsthat could help pelleting in both the intrinsicfeed material (to the bovines) and due to whathappens in the rumen of the bovines. This isan imporatant task for the future.

Typical agro-fuels used currently are bagasseand groundnut husk as primary fuels alongwith tamarind husk, doiled ricebran, sawdust,and other seasonally available wastes as sec-ondary ingredients. The team at FEPL hashad wide experience in using fodder pelletmaking machines. The fuel pellet throughputcomes down to 50 to 60 % of the fodder pelletthroughput with increased maintenance. Af-ter much consideration, two pellet machineswere obtained from Europe and they have beenset up, one of 20 t/d Dharwad (Karnataka)and another of 50 t/d at Islampur (Maharash-tra). These installations are producing pelletsand with other procurements a total of 1600tonnes/month is being supplied into the mar-ket.

6 The stove, efficiency andemissions

The development was carried out over a periodwith different fuels, different sized combustionchambers and designs of air supply to estab-lish the performance as a function of variousparameters on efficiency. Actual emission mea-surements were undertaken in the last stage asit was clear that obtaining the highest combus-tion efficiency (and accompanying utilization

Figure 2: The mass of the fuel with time dur-ing the stove efficiency tests. The slope of thecurve gives the mass loss rate

efficiency) with relatively short visible flameheights was needed to be achieved first; it wasinferred that this would also imply reducingthe emissions. The amount of secondary airwas varied to obtain a short visible flame. Thismeant that the ambient air demanded for com-bustion was minimal. The relevant dimensionsof the stoves built for this purpose and theirperformance are presented in Table 3. The ex-periments were made with the stove and vesselwith water kept on a balance. The mass of thesystem was continuously measured along withthe temperature of the water after it was vig-orously stirred once in a while. Figure 2 showsthe plot of the mass of the fuel burnt with time.As can be noticed, there are two distinct phasesof heat release. The first part is due to flamingand the second part due to char combustion.The power output during the second phase isabout a fourth of the power in the first phase.The slope of the mass vs. time plot gives themass loss rate that is the same as burn rate;this quantity multiplied by the calorific valuewill give the power of the stove.

Three cooking vessels were used for deter-mination of the thermal utilization efficiency.The consideration behind this choice is that

9

Table 3: Initial experiments - stove geometry, fuels and performance, WC = Wood chips, RHB= rice husk briquette pieces, CS = coconut shell pieces, MP = Marigold waste pellets, + Thesplit-up is between flaming and char combustion times, ∗ Power over the flaming time

biomass ρbulk moisture ash biomass burn time+ Power∗ ηwb

kg/m3 % % loaded, g mins kWth %Stove dia. = 100 mm, chamber volume = 0.6 liter, Vessel = 10 liter

WC 220 10 0.8 130 14 + 5 2.3 49.3RHB + WC 500 7 18.3 250 + 50 27 + 10 1.9 49.3CS + WC 430 9 0.6 230 + 30 30 + 10 2.2 53.3MP + WC 366 12 11.3 225 + 30 18 + 7 2.5 49.1

Stove dia. = 125 mm, chamber volume = 0.9 liter, Vessel = 10 literWC 200 10 0.8 170 30 + 8 5.0 48.0MP + WC 366 14 10.0 325 + 25 40 + 11 3.5 51.5

small families may use smaller vessels andlarger families, larger vessels. It would be valu-able to determine the efficiency with vessel size.It can be expected that larger diameter ves-sels extract more heat compared to smaller ves-sels and hence designs that allow greater heatextraction from the same stove would be theappropriate choice. The cooking vessels werealuminium vessels of 10 liter volume (diame-ter of 320 mm, height of 160 mm, and 0.96 kgweight), 6 liter volume (diameter of 260 mm,height of 130 mm and 0.61 kg weight) and 2.5liter volume (diameter of 205 mm, height of105 mm and 0.34 kg weight).

The lighting up process used two ap-proaches. The first one used about 15 ml ofkerosene on the top region of the fuel bed andthe second one used about 25 g of fine woodchips over which about 10 ml of kerosene wasdoused. The kerosene soaked fuel was lit with amatch stick and the primary air was turned on.After a two minutes of vigorous combustion,the secondary air was turned on. The vesselcontaining water was kept just about a minuteafter the lighting. The procedure used in thetests was based on Indian standard specifica-tions except that the tests were done with ves-sels larger than indicated in the specifications;the vessel diameters and volumes of water to

be taken for the tests at various power levelswere set in the 1991 specifications of the Bu-reau of Indian Standards keeping in mind lowerefficiencies that were expected in the better ofthe stoves at that time. The currently achivedefficiencies are nearly double of those in earliertimes; this is the reason for the choice of largervessel sizes in the current tests.

The results of efficiency measurements canbe noted from Table 3. The efficiency valuesare close to 50 % in most cases. The resultson vessel diameter effect with the 100 mm diastove are also shown in Figure 3. It is clearthat there is significant enhancement in the ef-ficiency with vessel diameter. This implies thatgiven an option of a cooking vessel with a de-sired volume, it would be better to select avessel with as low a height-to-diameter valueas is practical. This result was also found inthe earlier study of Mukunda et al4 albiet withlower efficiencies with a stove named “Swos-thee” (for Single pan WOod SToves of HighEfficiEncy) that is similar in configuration andperformance to the currently prevalent rocketstove.

10

Figure 3: The effect of vessel diameter on ef-ficiency, the lower plot is for the stove namedSwosthee (Mukunda, et al4)

6.1 Considerations for the practicalstove

The geometric parameters of the stove relateto the diameter and the height of the combus-tion shamber. These were decided based onthe following considerations.

The power level of the stove was set at 3kWth amounting to a nominal pellet fuel con-sumption rate of 12 g/min. For a 1 hour burntime with a 40 minute normal burn, some du-ration of a 9 g/min burn and remaining of charburn it was inferred that 600 g of fuel would beadequate for a cooking cycle. At a bulk den-sity of 400, 450 and 500 kg/m3, 600 g of fueltranslates to a fuel loading volume of 1.5, 1.33and 1.2 liters. Based on a nominal pellet bulkdensity of 450 kg/m3, a fuel loading volume of1.3 liters was chosen. This implies that if fuelof higher density was loaded, one gets largerburn time. On the other hand, if wood chipsof a bulk density of 200 kg/m3 were loaded (260g), the burn time of the stove would be 25 min-utes. Thus, while the stove would operate withmany fuels in the form of small pieces, the func-tional role of cooking would be met with withfuels with bulk density more than 450 kg/m3.

In order to make a choice of the combustionchamber diameter, two consistent view pointswere synthesized. The first one is that the pri-

mary airflow had to be about 1.5 times the fuelconsumption rate for the gasification process.Hence the air flow rate would be 18 g/min. Theother parameter of importance is the super-ficial velocity through the combustion cham-ber. If this was large (say, of the order of ormore than 0.1 m/s) then particulate carry-overwould be large and this would directly affectthe partculate emissions from the stove. If thevelocity was too low (say, less than 0.04 m/s),the combustion process in the char mode couldeither be extinguished or worse, so weak thatCO emissions would be large, perhaps beyondacceptable limits. The choice of 0.05 m/s forsuperficial velocity was made after some trialstudies. This would correspond to 95 mm di-ameter. Thus the choice of combustion cham-ber diameter between 95 to 105 mm wouldmeet the objectives of power, burn time, andminimum emission of particulate matter.

A second view arises from comparison to atypical gas stove burner; this has an outer di-mension of about 100 mm. Since these stoveshave been used very efficiently widely, it is pru-dent to make this choice. Hence, the inner di-ameter was chosen as 100 mm.

Scaling the design of this stove for otherpower levels is straightforward. We first rec-ognize that at the design superficial velocity of0.05 m/s, the power level (P ) of the stove is3 kWth for a inner diameter of 100 mm. Thepower level scales as the area of the combustionchamber (Ac). Thus P = 3(Ac/78)2 kWth.One can use a circular or square combustionchamber.

If the burn time has to be increased at a fixedpower level, the combustion chamber depth isto be increased linearly.

The air for combustion of the gases is pro-vided above the top of the bed. The amountof air (secondary air) that has to be providedhere is the difference between the stoichiomet-ric combustion air for the biomass and theair supplied for gasification. For biomass, the

11

stoichiometric combustion air depends on theCHNO analysis of the fuel. For the range offuels considered here, the stoichiometric air-to-fuel ratio is about 6.0 (the fuel is allowed 10 %ash with correponding reduction in the air-to-fuel ratio). For the current design, for 12 g/minof burn rate, the primary air is 18 g/min andthe secondary air will therefore be 54 g/min.This is supplied through a large number ofholes of small diameter. In the current design,18 holes of 6.5 mm dia (an area of 597 mm2)are provided. This leads to an inward air veloc-ity of 1.8 m/s through the holes. This critical-ity of this air flow is more towards determiningthe emission rather than efficiency. For, if thisair flow is inadequate, some part of the air fromthe ambient atmosphere is drawn in for com-bustion; but the oxidation of carbon monoxidein the product stream is affected, for it is veryslow to combust. Hence, the provision of aslightly larger secondary air flow will not affectthe performance. Also, the fuel pellets madefrom a wide variety of agro-residues cannot beexpected to have the same CHNO composition.As such, some variations in the stoichiometricair requirement can be expected; thus, if by de-sign a slight excess air is introduced, it wouldaccount for these variations in limiting the COemissions. Before release of the stove for de-velopments involving engineering and produc-tion, emission measurements carried out at fuelconsumption rates of 12 and 9 g/min showedthat the CO emissions were 1 and 1.3 g/MJwhere as particulate emissions were 10 and 6mg/MJ for the two power levels. Figure 4shows the variation of the volumetric ratio ofCO-to-CO2 with time of operation. The firstphase - flaming mode shows a low CO/CO2

(less than 0.01). The transition to char mode ofoperation increases the emission of CO signifi-cantly. The fact that char combustion in stoveshas significant CO emissions is well known instove literature (see for recent data, Smith etal18). There are also test-to-test variations in

Figure 4: The variation of the volumetric ratioof CO to CO2 during a stove operation. Theoperation till 30 mins is in flaming mode; charmode operation begins after a short transition

the CO emissions. These are largely becauseof the differences in the packing of the pelletsin the bed. The overall CO:CO2 was found as0.01 (volumetric) even though char burn alonecreates far more CO. These meet the require-ments of the Indian standards and hence areconsidered acceptable.

6.2 The power pack

The first choice was to integrate the powerpack based on commercially available elements.The first version has 6 V, 4 ampere-hour lead-acid battery that has a life of least 2 years(about 2000 cooking cycles). A single fullcharge permits 10 cooking cycles and wouldneed about 4 hours of charging before fresh use.In order to eliminate the use of lead acid bat-tery and also reduce the demand of fan power,after much in-house research and developmentat FEPL, a new version that allowed local man-ufacture of the fan and the power pack wasdeveloped. This is based on a 1.2 V Ni-MHbattery with 1.2 ampere-hours that has alsolasted more than 2 years. A single full chargecan last about 5 to 6 cooking cycles and the

12

Figure 5: The Oorja stove (a) notice the powerpack towards the left bottom section of the leftplate (b) the ceramic combustion chamber withgrate at the bottom and secondary air holestowards the top

charging time is an hour and a half. The fanitself was simply rebuilt with special bearingsto the same specifications. Now the team isevolving the Ni-Mh battery pack with feed-back of consumers to take the same batteryat 2.3 amp-hours (from 1.2 amp-hours). Thebattery charger is supported by circuitry that(i) boosts the voltage, (ii) starts warning theusers with a beep when the voltage drops to 0.7V giving consumers ample time to finish thecooking and charging it; the user will neverhave an issue of the battery draining and/orfan stopping since they will be alerted well be-fore.

6.3 The stove in the market

Figure 5 shows the photograph of the stovein two orientations. The inner wall is madeof ceramic composition that removes the lim-itations of the material limited life issues inthe combustion chamber. The bottom grate ismade of cast iron that ensures long life. Theprimary air comes through the grate and thesecondary air issues out of holes seen at thetop. The design ensures that the outer walltemperature does not exceed 60 ◦C complyingwith Indian standard requirement even thoughthere is a warning on the outside for the usernot to touch the body. The combustion pro-

Figure 6: The combustion process in the Oorjastove

cess inside the stove is seen in Figure 6. Thecup-like flames are those formed around the airjets issuing from the wall. Except for the ini-tial lighting process during which the flamesare yet to acquire the character of the com-bustion of a gasified fuel, all subsequent flamebehavior is similar to what is seen in Figure 6.

These stoves were first built with metal ver-sion. The metal was enamelled to extend thelife. Extensive combustion tests in the fac-tory showed that the life of this version wasabout 12 months. This was considered inad-equate. The next version included an innercast iron sheathing. This version gained wideacceptance and is expected to have a life ofmore than two years. This was also consid-ered inadequate primarily because it was con-cluded that the life could be affected by roughuse and this might leave the life question unre-solved. After much research and developmentby FEPL in coordination with private ceramicindustries, ceramic combustion chamber wasdeveloped (the stove using the ceramic com-bustion chamber is termed Oorja-plus). Thishas provided a quantum jump in the quality ofthe product. Over the last two years, at thistime of writing, about 420,000 stoves (includ-

13

ing the advanced metal and ceramic versions)have been sold in the states of Maharashtra,Karnataka and Tamilnadu. Based on the feedback from users, another version of the stovewith a square ceramic combustion chamber hasbeen developed including an ash removal tray(termed Oorja-super). This stove has still tosee large scale commercialization. The Oorja-plus stove has been subject to laboratory test-ing nationally and internationally.

While the stove configuration at the time oftechnology transfer to BP, (India) had speci-fied the secondary air hole area of 550 to 600mm2, there were no measurements on the in-fluence of a choice of this area on efficiencyand emissions even though the qualitative in-fluences could be deduced. The results of thestudy on the variations in the area of sec-ondary air introduction are summarized in theTable 4 (more detailed results can be seen athttp://cgpl.iisc.ernet.in). The ash content ofthe fuel is on the outer boundary (it is ex-pected to be less than 10 %). The power levelof the stove is 2.7 to 3 kWth over the flam-ing duration shown in the third column of Ta-ble 4. The water boiling efficiency is indepen-dent of the secondary area in this range. Theamount of CO2/unit fuel and NO in mg/MJbasis are roughly constant and SO2 seems in-significant as expected for biomass but CO de-creases with increase in the secondary hole areasignificantly. This is vividly clear in Figure 7.

6.4 Efficiency and emissions

Two studies on the stove performed by outsidegroups on this stove are Bryden and Taylor19

as well as Karabi Datta20. They have usedthe procedure developed by the university ofBerkeley discussed in detail by Smith et al18.Table 5 shows the results. They both reportsimilar efficiencies – 64 % for nominal powerand 65 % for low power. These appearedlarge to the present authors and when the de-

Figure 7: CO, NO and SO2 Emission indiceswith secondary air hole area

tails of the measurements were examined, onequantity that seemed inappropriate was theheat of combustion; Bryden and Taylor19 useda low value for the lower heating value re-sulting in unjustified enhancement of the effi-ciency. Fortunately, Bryden and Taylor reportthe amount of fuel required to heat 1 liter ofwater. This value (45 g/liter) seems to matchwith the results from this laboratory if we takenote of the different vessel sizes and the influ-ence of the vessel size on the efficiency. The re-sults of emissions reported by Karabi Datta20

seem to relate to indoor air pollution ratherthan from the stove directly. Emissions of COand particlulate matter (at 10 and 2.5 µm di-ameter) seem to match roughly.

6.4.1 Comparison with other studies

There is a body of a number of studies onstoves in respect of efficiencies and emissions.Most early studies are on free convection basedstoves made of metal, mud, and ceramics withsingle, two and three pots. Smith et al18 haveconducted an exhaustive and careful study ofthe emissions of a variety of stoves in India.There are a number of studies by Kirk Smithand colleagues on the greenhouse gas emis-sions from domestic stoves in several countries

14

Table 4: Emission and efficiency performance with secondary air entry area variation, fuel used= 650 g, As = secondary hole area, tb = Power during flaming time, WBE = Water boilingefficiency with 10 liter vessel, emissions in g and g/MJ of fuel energy

As Ash tb WBE CO CO2 NO SO2 CO NO SO2

mm2 % min % g g g g g/MJ mg/MJ g/MJ320 11.5 40 52.2 59.1 1039 0.027 1.18 6.5 2.9 0.13320 11.0 47 53.2 62.3 1008 0.037 2.13 6.8 4.1 0.23320 10.5 44 53.6 954 0.030 1.10 5.8 3.2 0.12382 12.1 40 52.0 32.5 1029 0.035 0.53 3.6 3.9 0.06510 12.0 38 51.7 9.6 879 0.037 - 1.1 4.1 -510 11.3 41 52.6 10.6 815 0.037 - 1.2 4.0 -

Table 5: Comparison of the results on efficiency and emissions on Oorja, WBE = Water boilingefficiency, BT08 = Bryden and Taylor, 2008, KD08 = Karabi Datta, 2008, n = nominal, nr =not reported

Item Present BT08 KD08Power, kWth (n) 2.7 2.3 nrWBE, % 51 64 64Fuel to boil, g/liter 36∗ 45+ nrLow power, kWth 2.0 1.7 nrLow power WBE, % - 65 65CO, n, g/MJ 1 1.2 nrPM10 nominal, mg/MJ 8 1.8 nrPM2.5 nominal - 1.2 nr

∗ 10 liter vessel, + 5 liter vessel

15

Figure 8: CO emissions (g/MJ) vs. water boil-ing efficiency from many studies

(see the publication list of Prof. Kirk Smith).Bhattacharya et al2,3 have presented the re-sults of similar stoves from south east Asia andIndia. Still21 has compiled the measurementsof efficiency of and emissions from about 20stoves, only six of which are relevant here (sev-eral stoves are with chimney). These stovescontain the data of fan based stoves as well.A class of stoves termed TLUD (implying toplit up-draft) around the development of Reedand co-workers8,9,10 has been popularized byAnderson (see Anderson et al22). Recently, acombustion stove with small continuous feedand supply of air from a fan has been developedand is in marketing trials by Phillips. The re-sults are set out in a summary form in Table6 and in Figure 8 with data from the abovesources.

The wide range of efficiencies and emissionsin free convective based designs is not un-expected since there is no possibility of con-trolling the emissions due to free-convectivemode of operation. The key problem of free-convection based stove is that while a certainarrangement of fuel sticks on the grate andtending will provide reasonably good efficiency

and low emissions, it is never clear what tend-ing will provide good results. At least soot-ing can be observed and controlled. How-ever, gaseous emissions cannot be observedand hence no observable physical control strat-egy can be devised. A well controlled labora-tory test may provide good performance anda whole range of field test data may indicatebad-to-average results. Rigorous protocols fortesting are not of any great use since theywill not represent an average user. What isamply clear from the plot is that fan basedstoves that promise near stoichiometric oper-ating conditions for combustion perform in afar superior way both with regard to efficiencyand emissions. A further optimization broughtabout in the Oorja design relates to the choiceof a high density fuel in the form of pellets.This feature is emphasized by characterizingthe Oorja as a fuel optimized forced convection(FOFC) stove. The choice of high density forthe fuel pellets helps reduce the volume of thecombustion chamber, provides guidance as tothe amount that would normally be requiredfor cooking by needing to fill a fixed amountand reduces opportunities to obtain an infe-rior performance by not having to demand pe-riodic loading or tending. Piece-by-piece load-ing is resorted only to extend the cooking byanother ten to fifteen minutes when requiredrather than a basic need to do it at undesirablyshort periods as in Phillips stove operation.

Attaining high combustion efficiency ap-pears to be a prerequisite for a “new gener-ation” stove to not only meet the requirementsof cooking, but also meet the obligations oflow green house gas emissions, a fact clearlybrought out by Kirk Smith in most of his writ-ings on indoor air pollution (see for instance,Smith23). Keeping away from fan based de-signs by invoking the lack of eletricity is contin-uously getting weakened with larger emphasison rural electrification; availability of electric-ity even for a small period during the day or

16

Table 6: Efficiency and emissions of various classes of stovesNature of stoves WBE CO PM

% g/MJ mg/MJFree convective based designsmud, ceramic, metal 15 - 35 1.5 - 15 30 to 1000Fan based stoves 35 - 45 0.8 - 1.2 2 to 20Optimized gasifier fan stove 40 to 50 0.8 - 1.0 2 to 9

night is adequate to charge the batteries usedfor cooking.

7 Concluding Remarks

This paper is concerned with the developmenta new cooking solution that is based on “gasi-fier” stove that uses an engineered “solid fuel”based on agricultural residues in the form ofhigh density pellets. Several aspects of thestove design and the fuel are brought out. Thevery high utilization efficiency as well as lowemissions are a consequence of the generationof near-constant throughput of gaseous fueldue to gasification and a correct air-to-fuel ra-tio used for combustion of the gases. The roleof secondary air in strongly controlling the COemissions is emphasized. An important aspectbrought out is that while free convection basedstoves may be appropriate in totally unelectri-fied areas, forced convection stoves may be theonly way to the future in improving the qual-ity of the environment around the stove, apartfrom higher efficiency; the minimum demandon “tending” is met with by the current de-sign.

There appears to be rising demand for thisclass of stoves in the “rural” user groups thatdo not lack the fuel as well as the “urban”users who have difficulty in sourcing any kindof fuel in an affordable manner. This demandis traced to a clear feeling of the user in owningsomething that operates in a manner close towhat LPG stove does. The journey from con-cept and laboratory studies to production and

commercial out-reach to more than a thousandusers a day hoping to move to a million userssoon has been possible due primarily to thepresent authors being together through manyproblem solving sessions over the last severalyears.

A view point that emerges from the presenteffort is this: the crucial aspect of produc-ing adequate amounts of dense pellet fuel andmaking them available in an affordable mannerforms the primary limiting feature in resolvingthe cooking fuel problem of our country (aswell as other similar countries). Another wayof expressing this point is that any progress inmain-streaming solid bio-fuels will alleviate thecooking problem of a large segment of the In-dian population. The role of the current effortsin making positive contributions to the climatechange problem and related current serious de-bate has been outside the scope of this study.

8 Acknowledgements

It is a pleasure to record the support receivedfrom Ms Manikankana Sharma who performedmost of the early experiments and from Mr.Srinath who performed all the tests on manyvariants of Oorja stove, both on efficiency andemissions.

References

[1] Prasad K. K., Sangen, E and Visser P.,Wood burning cook stoves, Advances in

17

Heat transfer, v 17, 1985

[2] Bhattacharya, S. C., Albina, D. O., andKhaing, A. M., Effects of selected pa-rameters on performance and emissionof biomass-fired cookstoves, Biomass andBioenergy, 2002, 23, 387 - 395.

[3] Bhattacharya, S. C., Albina, D. O., andSalam A. P., Emission factors of wood andcharcoal fired stoves, Biomass and Bioen-ergy, 2002, 23, 453 - 469.

[4] Mukunda, H. S., Shrinivasa, U., and Das-appa, S., Portable single-pan wood stovesfor high efficiency, Sadhana, 1988, 13, 237- 270,

[5] Bhattacharya, S. C., Biomass gasificationin Asia, chapter 7, Handbook biomassgasification Ed: Knoef, Biotechnologygroup, BV, 2005, 162 - 176.

[6] Kaupp, A. Gasification of Rice Hulls.GTZ, Eschborn, Germany, 1984; also seeKaupp, A. The gasification characteristicsof rice hulls for the generation of electric-ity and shaft power on a small (5 - 30 hp)scale, 1st Int. Prod. Gas Conf., Colombo,Sri Lanka, 1982.

[7] La Fontaine, H. and Reed, T. B.,An Inverted Downdraft Wood-Gas Stoveand Charcoal Producer, in Energy fromBiomass and Wastes XV, D. Klass, Ed.,Washington, D. C., 1993.

[8] Reed, T. B. and Larson, R., A wood-GasStove for Developing Countries, in De-velopments in Thermochemical BiomassConversion, Ed. A. V. Bridgwater, BlackieAcademic Press, 1996.

[9] Reed, T. B. and Walt, R., The “Turbo”Wood-Gas Stove, in Biomass: Proceed-ings of the 4th Conference of the Americas

in Oakland, Ca, Ed R. P. Overend and E.Chornet, Pergamon Press, 1999.

[10] Reed, T. B., Anselmo, E., and Kircher,K., Testing and modeling the wood-gasturbostove, Progresss in ThermochemicalBiomass Conversion Conference, Sept. 17-22, Tyrol, Austria, 2000.

[11] Shrinivasa, U. and Mukunda, H. S., WoodGas Generators for Small Power Require-ments, Current Science, 1983.

[12] Ravindranath N.H., and Hall D.O.,Biomass Energy and Environment- A De-veloping Country Perspective from India,Oxford University Press, Oxford, 1995.

[13] Shukla, P.R., Biomass Energy in India,Conference Proceedings - Biomass En-ergy: Key Issues and Priority Needs, In-ternational Energy Agency, Paris, France,1997.

[14] Kishore, V.V.N., Preety M. B., andGupta, P., Biomass energy technologiesfor rural infrastructure and village power- opportunities and challenges in the con-text of global climate change concerns,Energy Policy, 2004, 32, pp. 801 - 810.

[15] Anonymous, TERI energy data directoryand yearbook (TEDDY), 2004/05.

[16] Sinton, J. E. and Smith, K. R., In-door air pollution database for China,WHO/ENG/95.8, UNEP, Geneva, 1995;also see FAO document RWEDP Report32, GCP/RAS/154/NET, Proceedings ofthe regional expert consultation, 1997.

[17] Chatterjee, S., Rae, A. and Ray, R.,Food Consumption and Calorie Intakein Contemporary India, Departmentof Applied and International Eco-nomics, Discussion Paper No. 07.05,Massey university, 2007; also see URL:

18

http://www.fao.org/docrep/006/Y5022E/y5022e04.htmand http://www.weightlossforall.com/food-calories-list.htm

[18] Smith, K., Uma, R., Kishore, V. V. N.,Lata, K., Joshi, V., Zhang, J., Rasmussen,R. A., Khalil, M. A. K., Greenhouse gasesfrom small-scale combustion devices in de-veloping countries, Phase IIa, Householdstoves in India, USEPA, EPA-600-00-052,2000.

[19] Bryden, M and Taylor, P., Emission testresults for BP-Oorja stove, April, 2008.

[20] Karabi Datta, A comparison of stoveemission levels in rural households usingvarious models of improved stoves for atleast an year and the constraints influenc-ing the utilization of improved stovea ina rural community in Maharashtra state,Better air quality 2008, Air quality andclimate change: scaling up win-win solu-tions for Asia, November, 2008, Bangkok,Thailand

[21] Still, D., MacCarty, N., Ogle, D., Bond,T. and Bryden, M., Comparison of stoves,Aprevecho research center, Shell founda-tion, Partnership for clean indoor air,2006.

[22] Anderson, P. S., Reed, T. B., and Wever,P. W., Micro-gasification: What it is andwhy it works, Boiling point, 2007, 53.

[23] Smith, K. R., Health, energy and green-house gas impact of biomass combustionin household stoves, Energy for sustain-able development, 1994, 1, 23 – 29.

19