Applied Energy 104 (2013) 286–296

Contents lists available at SciVerse ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/ locate/apenergy

Performance of a pellet boiler fired with agricultural fuels

Lara Carvalho a,b,⇑, Elisabeth Wopienka a, Christian Pointner a, Joakim Lundgren b, Vijay Kumar Verma a,Walter Haslinger a, Christoph Schmidl a

a Bioenergy 2020+ GmbH, Gewerbepark Haag 3, 3250 Wieselburg-Land, Austriab Division of Energy Science, Luleå University of Technology, 971 84 Luleå, Sweden

h i g h l i g h t s

" Performance evaluation of a pellet boiler operated with different agricultural fuels." Agricultural fuels could be burn in the tested boiler for a certain period of time." All the fuels (except straw and Sorghum) satisfied the European legal requirements." Boilers for burning agricultural fuels should have a flexible control system.

a r t i c l e i n f o

Article history:Received 26 June 2012Received in revised form 25 October 2012Accepted 29 October 2012Available online 17 December 2012

Keywords:Agricultural fuelsCombustionGaseous emissionsDust emissionsSlag

0306-2619/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.apenergy.2012.10.058

⇑ Corresponding author at: Division of EnergyTechnology, 971 84 Luleå, Sweden. Tel.: +43 7416 5299.

E-mail address: lara.carvalho@bioenergy2020.eu (

a b s t r a c t

The increasing demand for woody biomass increases the price of this limited resource, motivating thegrowing interest in using woody materials of lower quality as well as non-woody biomass fuels for heatproduction in Europe. The challenges in using non-woody biomass as fuels are related to the variability ofthe chemical composition and in certain fuel properties that may induce problems during combustion.The objective of this work has been to evaluate the technical and environmental performance of a15 kW pellet boiler when operated with different pelletized biomass fuels, namely straw (Triticum aes-tivum), Miscanthus (Miscanthus � giganteus), maize (Zea mays), wheat bran, vineyard pruning (from Vitisvinifera), hay, Sorghum (Sorghum bicolor) and wood (from Picea abies) with 5% rye flour. The gaseous anddust emissions as well as the boiler efficiency were investigated and compared with the legal require-ments defined in the FprEN 303-5 (final draft of the European standard 303-5). It was found that the boi-ler control should be improved to better adapt the combustion conditions to the different properties ofthe agricultural fuels. Additionally, there is a need for a frequent cleaning of the heat exchangers in boil-ers operated with agricultural fuels to avoid efficiency drops after short term operation. All the agricul-tural fuels satisfied the legal requirements defined in the FprEN 303-5, with the exception of dustemissions during combustion of straw and Sorghum. Miscanthus and vineyard pruning were the bestfuels tested showing comparable emission values to wood combustion.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

During the last 30 years, small-scale wood combustion systemshave been well developed and reached a high quality and perfor-mance level in Europe. The energy efficiency has increased, theemissions have decreased, fully automatic operation systems havebeen developed and the combustion technology has been opti-mised for woody biomass fuels [1,2]. Automatically stoked systemsare normally based on staged combustion, i.e. two main combus-tion zones are created in order to maximise the burnout rates.

ll rights reserved.

Science, Luleå University of238 44; fax: +43 7416 52238

L. Carvalho).

The primary combustion zone is located on a grate or burner platewhere drying, devolatilization and char combustion takes place.The secondary combustion zone is located above the grate in thecombustion chamber where the combustible gases are oxidised[2,3]. Furthermore, each combustion zone has its own air supply;primary and secondary air are supplied in the fuel bed and in thecombustion chamber [4]. In this way, automatic pellet boilerscan show efficiencies higher than 90% (based on the lower heatingvalue) with CO emissions below 50 mg Nm�3 at 13 vol.% O2 [3](which corresponds to approximately 46 mg MJ�1) under steadystate combustion conditions.

Presently, the market for fossil fuels is unstable and their pricesare constantly rising. Furthermore, Europe has the target of reach-ing its share of renewable energies to 20% by 2020 [5] and biomasscan play an important role. However, the increasing competition

L. Carvalho et al. / Applied Energy 104 (2013) 286–296 287

for woody biomass in the heating sector, sawmills and pulp andpaper industries are increasing the price of wood [6]. As a result,the interest for alternative biomass fuels is growing rapidly, cover-ing woody materials of low quality, energy crops and agriculturaland forest residues [7–12]. There are several benefits from expand-ing the spectrum of biomass raw materials used in small-scalecombustion systems [13]. Besides increasing the use of renewableenergies, energy crops can provide a supplemental income forfarmers and at the same time show the potential of restoring de-graded lands, preventing soil erosion. The value of the agriculturaland forest residues can be increased by using them as fuels andmore job opportunities for power and agricultural equipmentindustries can be created. However, burning non-woody biomassfuels in small-scale heating systems is a challenging option[11,14,15]. Compared to wood, non-woody biomass assortmentshave higher ash contents and a higher content of critical inorganicelements. Ash related problems are therefore expected, which af-fect the dust emissions [16] and cause problems during combus-tion due to e.g. slagging [10,17]. Slag on the grate of small-scalepellet boilers may disturb the combustion process (e.g. higher COemissions [11]) and lead to unwanted shutdowns of the boiler[18]. High concentrations of nitrogen (N), sulphur (S) and chlorine(Cl) in agricultural fuels increase the emissions of nitrogen oxides(NOx), sulphur dioxide (SO2) and hydrogen chloride (HCl), respec-tively [4,16]; they may also cause the formation of dioxins and fur-ans under certain combustion conditions, e.g. intermittentcombustion [19,20]. Nitrogen oxides can be formed during com-bustion by three mechanisms referred to as thermal, prompt andfuel-bound NOx. The thermal and prompt NOx formation paths be-come active at temperatures above 1300 �C [21] which are nor-mally not reached in small-scale combustion systems.Consequently, nitrogen oxides are assumed to be formed mainlyfrom fuel nitrogen during biomass combustion and therefore can-not be completely avoided [21,22]. Minimisation of NOx emissionsby optimising the combustion conditions is however possible byair or fuel staging [4,16]. The release of SO2 and to a lesser extentSO3, result from the oxidation of the fuel sulphur [11,16]. Theseemissions are usually not significant for wood combustion due tothe low sulphur contents of the fuel. However, at sulphur concen-trations higher than 0.2 wt.% dry basis, SO2 emissions start to berelevant due to its important role in corrosion [16]. Both NOx andSO2 have significant health effects (e.g. respiratory problems)[23] and are harmful gases to the environment contributing to acidrain. Incomplete combustion can lead to emissions of carbon mon-oxide (CO), OGC (organic gaseous carbon), PAH (polyaromatichydrocarbons), soot and tar. CO is generally considered an indica-tor of the combustion quality for the reason that it is oxidised toCO2 in the presence of oxygen and at a rate which depends onthe combustion temperature, residence time and mixing rate be-tween the combustible gaseous species from the fuel and air[3,4,16]. Biomass combustion also leads to relatively high dustemissions. Wood burning is one of the major contributors of pri-mary particle to the atmosphere during winter times over largeparts of Europe [24–28]. Small-scale combustion systems in partic-ular play an important role [26–28]. The dust emissions can consistof both carbonaceous particles and vaporised inorganic mattermainly alkaline metals, sulphur and chlorine. Under poor combus-tion operation practices such as unsatisfactory air supply occurringfor instance in old residential heating appliances [29,30], the dustemissions are high and dominated by particles of incomplete com-bustion. Typical dust emissions from modern pellet boilers oper-ated at 100% load can vary from 10 to 30 mg MJ�1, while theemissions from old residential heating appliances are between 65and 150 mg MJ�1 or even higher [29,30]. The inorganic part alwaysremains as background constituents. Efficient combustion of woodresults in mainly inorganic fine particles [30,31] typically

dominated by alkaline metals, such as potassium sulphates, chlo-rides and carbonates [32,33]. An increase in dust emissions istherefore expected when agricultural fuels are burned due to theirhigh content of alkaline metals. The release of alkali metals is inturn influenced by other fuel elements present in agricultural fuels,most importantly chlorine, silicon and sulfur. While chlorine con-tent enhances the release of alkali metals due to the formation ofvolatile alkali metal chlorides, silicates can react with potassiumpreventing its vaporisation [34,35]. Furthermore, sulphur in thefuel may inhibit the effect of chlorine through a sulfation reaction,in which the alkali metal chloride is converted to less volatile alkalimetal sulphate [36].

The emissions and ash related problems during combustion ofagricultural residues in medium and large scale combustion plantshave been thoroughly investigated, e.g. [37–39]. However, there isstill insufficient information available regarding their suitability insmall-scale systems and whether they can be burned in accordanceto the existing threshold values. Combustion of woody biomasscauses emissions of gases and particulate matter which can seri-ously affect human health [23,40]. The introduction of new bio-mass fuels that potentially may cause higher emissions into theresidential heating sector should be first thoroughly evaluatedbased on results from combustion tests. Furthermore, it is impor-tant to investigate the capability of the existing small-scale tech-nologies in burning non-woody biomass fuels. Combustion testsprovide important information to boiler manufactures by showingthe limitations of the existing boiler technology and by identifyingimportant parameters and improvements required to adapt themfor a broader spectrum of biomass fuels.

The objectives of the present study are (i) to evaluate the tech-nical and environmental performance of a 15 kW wood pellet boi-ler when burning different agricultural biomass fuels and (ii) toinvestigate the feasibility of different agricultural fuels for residen-tial heat production. The feasibility study was done by comparingthe gaseous and particle emissions as well as the boiler efficiencywith the legal requirements defined in the FprEN 303-5 (final draftof the European Standard 303-5, Heating boilers – Part 5: Heatingboilers for solid fuels, manually and automatically stoked, nominalheat output of up to 500 kW – terminology, requirements, testingand marking) [41].

2. Material and methods

2.1. Fuel properties

The combustion tests were done with pelletized fuels fromwood (Picea abies) with 5% rye flour and seven agricultural rawmaterials: straw from wheat (Triticum aestivum), Miscanthus(Miscanthus � giganteus), maize (Zea mays), vineyard pruning (fromVitis vinifera), wheat bran, hay and Sorghum (Sorghum bicolor).Straw was harvested in the Spring of 2006. Miscanthus was har-vested between January and March of 2006. Maize was harvestedin the Spring of 2006 and all the plant (stem and cob with maizegrains) was pelletized. Wheat bran is the hard outer layer of awheat grain. It is normally a by-product of milling in the produc-tion of refined wheat flour. Hay is a mixture of grasses, legumesand/or other herbaceous plants. Its composition depends greatlyon the region where it grows. Detailed information about the Sor-ghum pellets tested is given elsewhere [42]. No informationregarding the place of growth and the time of harvesting is avail-able for the pellets from vineyard pruning, wheat bran and hay.

Chemical and thermal properties of each of the experimentalpellets were analysed according to the relevant standard methodsand the properties are presented in Table 1. The ash and moisturecontents of the pellets were analysed according to CEN/TS 14775,

Table 1Properties of the investigated pellets.

Wood with 5% rye flour Straw Miscanthus Maize Vineyard pruning Hay Wheat bran Sorghum

Diameter (mm) 6 6 6 6 6 6 8 6LHVa (MJ kg�1) dry basis 18.90 15.78 16.09 15.74 16.50 15.46 15.18 15.95Moisture (wt.%) 10.2 7.9 8.7 8.7 8.4 7.5 13.7 7.9Ash (wt.%)db 0.21 5.71 3.26 3.39 2.70 7.40 5.70 7.90

Proximate analysis (wt.%) dry basisCarbon (C) 47.1 47.30 47.30 45.9 45.71 45.94 46.40 46.90Hydrogen (H) 6.17 5.80 6.00 6.30 6.30 5.72 6.08 5.81Nitrogen (N) 0.19 0.66 0.24 0.89 0.39 1.07 3.04 1.08Sulphur (S) 0.10 0.10 0.03 0.09 0 0.13 0.24 0.11Oxygen (O) 46.3 40.43 43.17 43.43 44.90 38.54 38.54 38.20

Inorganic species of the fuels (mg kg-1 fuel) dry basis (relevant for the present study)Alkali metals (K + Na) n.d. 14130 1012 7867 n.d. 11400 274649 n.d.Alkali earth metals (Ca + Mg) n.d. 3342 2613 2003 n.d. 8806 80649 n.d.Chlorine (Cl) n.d. 1720 214 1024 n.d. 1090 16140 n.d.Silicon (Si) n.d. 10620 9829 3613 n.d. 12600 15614 n.d.Phosphorous (P) n.d. 716 318 1999 n.d. 2460 174561 n.d.Zink (Zn) n.d. 5 13 17 n.d. 26 1316 n.d.

a LHV – low heating value; n.d. – not determined.

288 L. Carvalho et al. / Applied Energy 104 (2013) 286–296

at 550 �C and DIN 51718, respectively. The lower heating value(LHV) was calculated from the HHV based on the moisture contentof each fuel. The higher heating value (HHV) was determined usingan adiabatic bomb calorimeter according to ÖNORM CEN/TS14918. The content of the major elements, carbon (C) and nitrogen(N) were determined according to prCEN/TS 15104. The total sul-phur (S) was determined using the CEN/TS 15289 method. Theoxygen (O) content was calculated by difference according to Eq.(1).

O½wt%� ¼ 100� ðAshþ C þ H þ N þ SÞ ð1Þ

The ash forming elements (cations) were determined by microwavedigestion. The Cl content was determined by analyzing a solution ofthe combustion gases dissolved in deionized water via IC.

Fig. 1. Schematic layout of the pell

2.2. Combustion equipment

The combustion tests were performed in a commercially avail-able pellet boiler illustrated in Fig. 1. The boiler had a nominalthermal capacity of 15 kW and was equipped with a lambda sensor(LSM 11). Primary combustion air was supplied from below thegrate and secondary air from above the fuel bed. The boiler had ahorizontal feed burner with fuel pellets supply via a horizontalfeed screw. The ash removal was done by the conveying fuel andby a horizontal movement of the grate. On the opposite side ofthe fuel feeding there was an opening where the ash residues felldown into the ash box, placed below the grate. In the opening therewas an agitator that rotated perpendicularly to the fuel and ashtransport and destroyed possible ash lumps and slag. The agitatorwas not included in the commercial version. Several combustion

et boiler (commercial version).

Fig. 2. Schematic layout of the experimental setup.

Table 2Settings used during the combustion tests.

Fuels Moving grate Airratio(–)

Fuelload(%)

Frequency of themovement (h�1)

Movementduration(s)

Wood with 5% rye flour 0.5 0.1 2 100Straw, Miscanthus, maize,

vineyard pruning andhay

36 1.5 2 100

Wheat bran 36 1.5 2 50Sorghum 36 1.5 2 75

L. Carvalho et al. / Applied Energy 104 (2013) 286–296 289

parameters could be set in the boiler control. The relevant andmodified parameters for the combustion tests were: air ratio, fuelload and grate movement. The grate movement parameters con-trolled the frequency and the period of the movements.

2.3. Experimental test setup

The combustion tests were performed at the laboratory of Bio-energy 2020+, Lower Austria (48.117�N 15.136�E, 270 m abovemean sea level). The combustion tests with each type of pelletswere planned to last a minimum of 8–12 h in steady-state condi-tions. A schematic of the experimental setup is presented in Fig. 2.

The temperatures in the combustion chamber and of the fluegas (Tfg) were continuously monitored using K-type thermocou-ples. The feed and hot water temperatures (TFW and THW) weremaintained at 55 �C (±2 �C) and 75 �C (±2 �C) respectively and weremonitored using Pt-100 thermocouples. The chimney draft (P) wascontinuously measured using a pressure transmitter PTL2-K. Thedata was acquired every second during the steady-state operationof the boiler and average values were calculated for periods of 1 h.

The settings used for the moving grate, air ratio and fuel loadare presented in Table 2.

Different settings for the grate movement parameters wereused in the combustion tests with wood and agricultural fuels, ow-ing to their large differences in ash contents. The difference in fuel

load during wheat bran and Shorgum combustion is explained inSection 3.1.1.

2.3.1. Measurement of gaseous and particulate emissionsA constant flow of flue gas (0.8 L s�1) was extracted and trans-

ported to the gas analyser (NGA 2000) through heated samplingprobes and lines. The concentrations of CO2, O2, NO2, NO, SO2

and CO in the flue gas were continuously monitored. The measure-ment principles of the gas analyser were paramagnetic for O2, non-dispersive infra-red for CO, NO, and CO2, and ultraviolet cell for SO2

and NO2 determinations. The gaseous emissions were acquiredevery second. The particulate matter was sampled isokineticallyusing a gravimetric method as described in VDI 2066/sheet 2 (Mea-surement of particulate matter in flowing gases). The total coarseand the major fine dust fraction in the flue gas were collected ina cartridge filter of quartz wool. The particulate matter was sam-pled at least three times in each experiment. The emission valueswere converted to mg MJ�1 of fuel heat input.

2.3.2. Measurements of the fuel mass flowThe boiler was placed on a scale (Mettler Toledo PTA with an

accuracy of 0.04–0.3 kg) in order to determine the amount of fuelused during each combustion test (see Fig. 2). The mass flow of fuelon the grate was calculated for each experiment by dividing the to-tal amount of burned fuel by the total time of the combustion test.The mass flow of ash on the grate was calculated based on the ashcontent of each fuel and by subtracting the amount of ash lost inthe form of dust emissions.

2.3.3. Slag tendency measurementsThe slag tendency of a biomass fuel was analysed through siev-

ing, as described in [43]. The ash from the ash box was sieved at theend of each combustion test. The ash particles retained in the sieveof 5.6 mm mesh size were considered slag. The percentage of slagin the ash was calculated for each test.

Table 3Mass flow of fuel, energy load and mass flow of ash determined for each combustion test.

Fuel Vineyard pruning Hay Straw Miscanthus Maize Sorghuma Wheat branb Wood with 5% rye flour

Fuel mass flow (kg h�1) 4.45 4.25 4.21 4.06 3.62 3.11 2.52 4.06Energy load (kW) 20.39 18.27 18.44 18.14 15.83 13.76 10.61 21.32Mass flow of ash (kg h�1) 0.119 0.311 0.232 0.132 0.119 0.240 0.140 0.007

a Operated at 75% nominal load.b Operated at 50% nominal load.

290 L. Carvalho et al. / Applied Energy 104 (2013) 286–296

2.4. CO and NOx versus lambda

The lambda probe allowed the boiler to operate at a relativelyconstant lambda value. However, small lambda variations(2.0 ± 0.3) were observed during the combustion tests. These vari-ations were most probably due to the type of lambda probe usedby the boiler (see Section 2.2) which has a high resolution signalfor lambdas around one. Continuous measurements of CO andNOx acquired during combustion could therefore be compared tothe variations in the lambda value.

2.5. Efficiency

The boiler efficiency was calculated using an indirect method bymeasuring the losses occurring in the boiler as follows.

gIndirectMethod½%� ¼ 100� ðqA þ qU þ qF þ qsÞ ð2Þ

In Eq. (2), qA (%) is the thermal loss in the flue gas, qU (%) is the lossdue to unburned fuel carbon in the flue gas, qF (%) is the loss due tounburned carbon in the solid residue (ash) and qS (%) is the radia-tion heat loss from the boiler surface. qA and qU were calculatedusing Eqs. (3) and (4) respectively.

Fig. 3. CO emissions as a funct

qA ¼ ðVfg;dcpfg þ VwcpwaterÞðTfg � TairÞ1

LHVð3Þ

qU ¼ 10�6COgenVfg;d12640

LHVð4Þ

Here, Vfg,d (m3 kg�1) is the volume of dry flue gas generated duringcombustion per kg of fuel, Vw (m3 kg�1) is the volume of moisture inthe flue gas, Tfg (�C) and Tair (�C) are the temperatures of the flue gasand ambient air respectively, cpfg (kJ (kg K)�1) and cpwater

(kJ (kg K)�1) are the specific heat capacities of the flue gas and waterrespectively and COgen (ml m�3) is the CO content in the flue gas.

The loss due to unburned carbon in the ash (qF) was determinedfor wood pellets and other agricultural fuels, e.g. Sorghum andwere below 0.02%. The influence of this loss on the efficiencywas considered irrelevant and therefore qF was assumed zero inthe efficiency calculations (Eq. (2)).

The radiation heat loss from the boiler surface (qS) was assumedto be 1% of the total heat output, as stated in the boiler user’smanual.

ion of the excess air ratio.

L. Carvalho et al. / Applied Energy 104 (2013) 286–296 291

3. Results and discussions

The results of the combustion tests are presented and discussedwith regard to boiler technology aspects, emissions and efficiency.

3.1. Boiler technology aspects

Three important aspects of the boiler technology, (1) fuel feed-ing, (2) air ratio and (3) ash management, are presented anddiscussed.

3.1.1. Fuel feedingThe combustion tests were performed at full load, i.e. at 100%

fuel feeding, with the exception of the tests with wheat bran andSorghum pellets, performed at 50% and 75% load respectively, asshown in Table 2. The reason for the reduced loads was that bothwheat bran and Sorghum burned out more slowly than the otherbiomass fuels. As a result, a stable fire bed could not be formed be-tween two consecutive movements of the grate and unforeseenshut downs of the boiler were continuously occurring. The fuelload of wheat bran and Sorghum were therefore reduced until acontinuous operation for the stipulated amount of time waspossible.

Considering only the fuels operated at 100% nominal load andwith equivalent operating settings of the fuel screw, the mass flowof fuel varied considerably when different fuels were used (Ta-ble 3). Fuel mass flows ranging from 3.62 to 4.45 kg h�1 were ob-served, corresponding to a variation of approximately 20%. As aresult, the energy input to the boiler also varied considerably andproportionally to the mass flow of fuel. The energy load rangedfrom 15.83 kW with maize to 20.39 kW with vineyard pruning.The variability in both mass and energy flows can be explainedby differences in the bulk density. Pelletized agricultural and for-estry biomass commonly show bulk densities ranging from 480to 700 kg m�3 [11,44–48]. Other factors that may influence the fuelfeeding are pellet size [46], amount of fines [47] and the surface ofthe pellets [47]. The mass flow of ash varied between 0.12 and0.31 kg h�1, due to different fuel mass flows and different ash con-tents of the experimental fuels (ash contents ranging from 3.3 to7.9 wt.% dry basis in Miscanthus and Sorghum respectively).

3.1.2. Excess air ratioEach combustion unit has a characteristic CO versus air ratio

curve when operated with a certain biomass fuel. Such curves givethe air ratio that minimises emissions and maximises theefficiency. Furthermore, to ensure a complete combustion, a goodutilisation of the combustion chamber (with even gas flow distri-bution) as well as uniform temperature distribution (avoiding localtemperature peaks) should be achieved by optimising the combus-tion chamber and nozzle geometries [40]. Therefore, the CO

Table 4Boiler efficiency during combustion with and without prior cleaning of the heat exchangerthe heat exchanger.

Fuels Heat exchanger not cleaned between tests

Accumulated operational time of the boiler (

Miscanthus 39Straw 55Miscanthus with additivea 63Maize 85Straw with additivea 110

a The combustion tests with pellets containing additives were outside the scope of thheat exchangers. The efficiency results were therefore important to show the built of de

b The flue gas temperature increased with 5 �C in approximately 5 h.c The flue gas temperature increased with 10 �C in approximately 7 h.

characteristics can also be used to assess the suitability of thecombustion chamber dimensions in burning a specific fuel. Fig. 3illustrates the CO versus excess air ratio curves for each of theinvestigated biomass fuels for the applied combustion unit.

Most of the biomass fuels show a clear and smooth CO charac-teristic curve. As expected, each side of the curves show a sharpand a gradual increase of the CO emissions at low and high excessair ratios respectively.

In the present boiler, the minimum CO emissions are obtainedat excess air ratios of 1.6–1.8 for Miscanthus and maize, 1.8–2.0for vineyard pruning, 1.9–2.1 for hay and Sorghum and 2.0–2.2for wheat bran. The CO emissions were consistently high for strawpellets relative to the other fuels. Therefore, the most favourableexcess air ratio interval to minimise CO emissions was not clear.Nevertheless, the lowest CO emissions for straw pellets were ob-tained at excess air ratios of 2.0–2.1. The high variability of theCO emissions exhibited by straw and to a smaller extent by wheatbran and maize, indicates a difficulty of the boiler in achieving sta-ble combustion conditions with these fuels.

Several reasons can explain the differences in the excess air ra-tios to minimise CO emissions found for each fuel, (i) the combus-tion chamber was designed for woody fuels and therefore, theirsize and shape were not optimal for all the fuels; (ii) the residencetime on the grate was not optimal for all the pellets; (iii) slaggingand ash lumping on the grate; (iv) the boiler control does not fullyadapt the burning conditions to the different biomass fuels. Theboiler was operated at lambda values averaging 2.0 however, theplots of Fig. 3 showed that some of the fuels required lower orhigher air ratios in order to minimise CO emissions.

3.1.3. Managing high ash contents3.1.3.1. Ash and slag. Slag is sintered ash formed on or near thegrate and which strength (i.e. degree of sintering) is mostly af-fected by the fuels composition [17]. Ash caking and slag wasformed in different amounts during the combustion tests. The per-centage of slag present in the ash of straw, maize, Sorghum andwheat bran ranged from 20% to 45% while the ash from Miscan-thus, vineyard pruning and hay exhibited values of slag below1%. Further discussion on the different slagging behaviours of thefuels is given in [43]. The boiler was able to effectively removethe ash lumps and slag from the grate due to the moving grateand the agitator (not included in the commercial version), allowingthe boiler to operate for the stipulated time period of 8–12 h.

3.1.3.2. Ash depositions on the heat exchanger surfaces. Combustion-generated fly ash particles are responsible for ash deposition on theheat transfer surfaces which will lead to reduced heat transfer[38,39]. In the present study, several consecutive combustion testswith different fuels were performed without cleaning the heatexchanger between the tests. During these test runs, a progressive

s and total number of accumulated hours the boiler was operated without cleaning of

Heat exchanger cleaned between test

h) Efficiency (%) Efficiency (%)

89.2 90.588.4b 90.889.6 91.383.7c 89.079.9 90.2

e present work. However, they were one of the fuels tested without cleaning of theposits.

292 L. Carvalho et al. / Applied Energy 104 (2013) 286–296

increase in flue gas temperature was observed. After a total opera-tional time of 110 h, the boiler efficiency had decreased withapproximately 10%, due to the continuous build up of depositionon the heat exchanger walls that inhibited proper heat transfer be-tween the flue gas and the feed water. Table 4 shows the effect ofcleaning the heat transfer surface on the boiler efficiency duringcombustion of different fuels.

The efficiencies shown in Table 4 were calculated based on theheat input and output of the boiler (direct method), owing to theinaccuracy of the indirect method in these particular cases. Effi-ciency drops due to the accumulation of ash particles in the heatexchanger walls can be avoided by frequently cleaning the heatexchangers.

3.2. Boiler limitations and potential improvements

Three main boiler limitations were found when agricultural bio-mass fuels were burned in the wood pellet boiler. The first wasconnected to the combustion chamber; based on the results ofthe CO characteristic curves shown and explained in Section 3.1.2,the design of the combustion chamber was most probably not opti-mal for all the agricultural fuels. The second limitation was relatedto the boiler lambda control; the boiler was unable to adjust to anoptimal excess air ratio. The optimum lambda value for some ofthe fuels was either lower or higher than the one the boiler wasusing (1.9–2.0). Finally, the third boiler limitation was related tothe heat exchangers. It was observed that in consecutive combus-tion tests, the efficiency was progressively decreasing due to ashaccumulation on the heat exchangers.

Several modifications in the boiler technology are required toimprove the combustion of agricultural fuels. The boiler controlsystem should be more flexible in order to better adapt the burningconditions to the different fuels. That could be done, e.g., byimproving the control algorithm of the boiler so that the operationsettings are changed according to the fuel type. The user could in-sert key fuel properties or the system could have already a data-base with the different fuel properties and the user would onlyneed to choose the fuel type prior to combustion. The boiler wouldthen adapt the air supply and fuel feeding to the desired heat out-put. In this way, the problems related with the combustion cham-ber design would also be reduced. Finally, heat exchangers with

Fig. 4. CO emissions in units of mg MJ�1and percentage of slag particles in the ash of the

automatic cleaning systems would be recommended in the interestof keeping the high comfort and low maintenance level of the pres-ent boiler. The mechanisms for cleaning of the heat exchanger andmoving grate could be connected to the controlling unit and thecleaning intervals would be a function of the fuel type.

3.3. Emissions

The gaseous and particulate emissions are presented and com-pared with the legal requirements defined in FprEN 303-5 [41].

3.3.1. CO emissionsThe CO emissions varied significantly among the studied agri-

cultural fuels and were in general higher than when woody bio-mass was used, with the exception of Miscanthus (Fig. 4). Hay,wheat bran and straw pellets showed the highest values of 280,224 and 223 mg MJ�1 respectively, which were eleven to fourteentimes higher than the emissions from wood pellets. The CO emis-sions from maize, vineyard pruning and Sorghum were compara-tively low and just two to five times higher than the emissionsfrom wood pellets.

The difference in CO emissions between agricultural and woodyfuels can be explained by the control system of the boiler, i.e. theboiler could not adapt the combustion air to the needs of the differ-ent agricultural fuels, as verified in Section 3.1.2. For example,wheat bran and hay were burned at an excess air ratio of 2.0. How-ever, a reduction in the CO emissions is to be expected if the boilerwould be operated at a lambda value of 2.1 with both fuels (Fig. 3).Additionally, the CO emissions observed during combustion ofstraw and wheat bran could be partly explained by the slagformed, whereas 40% and 37% of the ash from straw and wheatbran respectively, consisted of slag (Fig. 4). The formation of ashlumps and slag could have caused an uneven flow of combustionair through the fuel bed which could in turn impair the char andgas-phase burnout.

The FprEN 303-5 defines CO emissions for automatic stokedboilers with a nominal heat output below 500 kW operatedwith biogenic fuels. The CO emissions legal requirements forboiler classes 3, 4 and 5 are 3000 mg m�3 at 10% O2 (approximately2000 mg MJ�1), 1000 mg m�3 at 10% O2 (approximately 670mg MJ�1) and 500 mg m�3 at 10% O2 (approximately 330 mg MJ�1)

different fuels. The error bars correspond to the SEM (standard error of the mean).

Fig. 5. NOx emissions (left axis) and nitrogen conversion rate (right axes) as a function of the nitrogen content of the biomass fuels. The error bars correspond to the SEM(standard error of the mean).

Fig. 6. NOx emissions as a function of the excess air ratio for maize, hay, Miscanthus and wheat bran.

L. Carvalho et al. / Applied Energy 104 (2013) 286–296 293

respectively. All the fuels tested can fulfil the CO emissions legalrequirements for the three boiler classes.

3.3.2. NOx emissionsFig. 5 shows the NOx emissions and the nitrogen conversion rate

as a function of the different nitrogen contents of the agriculturalfuels tested. The combustion temperatures were below 1300 �Cand therefore the NOx emissions are expected to be only due tovariations in the nitrogen bound in the fuel [4,21].

As expected, the NOx emissions show a growing trend towardshigher fuel nitrogen contents. However, the two quantities wereweakly correlated, as also observed by other authors, e.g.[7,11,49]. The weak correlation can have several explanations. Onthe one hand, high concentrations of CO or other carbon com-pounds may inhibit NOx formation [50,51]. Carbon monoxide hasbeen shown to enhance the rate of NO reduction over different car-bonaceous materials [52,53]. On the other hand, the catalytic effectof char and ash may also influence the NOx emissions [39,50]; theNO formed by the char may be re-adsorbed on the char surface andform N2 by a following reaction with NO or NO can be reduced byCO catalysed by char [22]. The minerals present in biomass mayenhance the rate of the NO char reaction, indicating a potentialof this reaction for NOx control [53]. Consequently the catalytic ef-fect of char must be taken into account in fuels with high ash con-tents [54] and in fuels with high content of catalytic active

substances like calcium, magnesium, iron and potassium[22,39,55].

The combustion of agricultural fuels resulted in higher NOx

emissions than when woody fuels were burned. In terms of legalrequirements, the FprEN 303-5 [41] does not define NOx emissionlimits. Among the country deviations stated in the same draft stan-dard, Austria is the only European country with legal NOx emis-sions requirements for boilers with a nominal heat output below150 kW operated with standardised non-woody fuels. All thetested agricultural fuels, with the exception of maize, fulfil theNOx emission legal requirements (300 mg MJ�1) according to theAustrian law.

The plots in Fig. 6 show the effect of the lambda variation on theNOx emissions. When maize was burned at a lambda of 1.7, the NOx

emissions were reduced with about 22% to 270 mg MJ�1 withoutcompromising complete combustion (see Section 3.1.2). In thisway, maize was compliant with the Austrian legal requirementsin terms of NOx emissions. Similarly, Miscanthus operated at alambda of 1.6–1.7 would reach a 16% reduction in NOx emissionsto about 175 mg MJ�1. It can also be observed that when the boilerwas operated at a lambda of 2.1 with hay and wheat bran insteadof 2.0, which was found to be the optimal level in terms of COemissions (see Section 3.1.2), the NOx emissions would increasewith approximately 18% and 6% respectively, making wheat branincompliant with the Austrian NOx emission requirement. The

Table 5Sulphur content, SO2 emissions (±SEM) and sulphur conversion rate (±SEM) of the tested fuels. SEM stands for standard error of the mean.

Fuel Wheat bran Hay Sorghum Maize Straw Miscanthus Vineyard pruning Wood with 5% rye flour

S content in the fuel (wt.%) dry basis 0.24 0.13 0.11 0.09 0.10 0.03 0.02 0.01SO2 emissions (mg MJ�1) 220 ± 2.7 75 ± 0.6 70 ± 1.3 58 ± 0.5 44 ± 0.8 26 ± 0.7 7.2 ± 0.9 3.4 ± 0.0Sulphur conversion rate (%) 73 ± 0.9 44 ± 0.4 52 ± 1.0 51 ± 0.3 35 ± 0.7 70 ± 1.8 32 ± 1.7 37 ± 0.0

294 L. Carvalho et al. / Applied Energy 104 (2013) 286–296

reductions and increases on the NOx emissions when the air ratiowas decreased and increased respectively were probably con-nected with the conditions in the primary combustion zone. It iswell known that a considerable reduction of the NOx emissionscan be reached by applying staged combustion, i.e. creating anunderstoichiometric primary combustion zone and providing amean residence time of 0.5 s between primary and secondary com-bustion will favour the reduction of fuel-N to N2 [4].

3.3.3. SO2 emissionsAs presented in Table 5, the SO2 emissions varied as a function

of the fuel sulphur content. The lowest and the highest emissionsof SO2 were obtained during combustion of Miscanthus(26 mg MJ�1) and wheat bran (227 mg MJ�1) respectively. Corro-sion problems might be expected during combustion of wheat brandue to its high sulphur content (above 0.2 wt.% dry basis) [16]. Thesulphur conversion rate was not strongly affected by the sulphurcontent. The lowest and the highest sulphur conversion rates wereobserved during combustion of vineyard pruning (32%) and wheatbran (73%) respectively.

3.3.4. Dust emissionsFig. 7 shows the dust emissions from the boiler when operated

with the tested fuels. In general, with increasing ash content of thebiomass fuel also the particle emissions increased considerably[56]. However, no clear correlation was found between the particleemissions and the ash content of the fuels. The emissions fromburning hay, wheat bran and Sorghum pellets were lower than ex-pected, given the high ash contents of these fuels. Alkaline metalsare key fuel components responsible for the inorganic fraction ofthe dust emissions. Other relevant species are S, Cl and some heavymetals such as Zn (zinc) [16,31,36]. The high dust emissions ob-served during straw, maize and wheat bran combustion may beconnected to their high contents of alkali metals and sulphur.The high dust emissions from burning straw can also be connectedwith a high level of products of incomplete combustion given the

Fig. 7. Particle emissions in units of mg MJ�1. The error bar

fact that the CO emissions were high (220 mg MJ�1). Despite thehigh alkali metals and sulphur contents in wheat bran (275 and2.4 g kg�1 of fuel respectively), the particle emissions were rela-tively low (78 mg MJ�1). In comparison to straw, the alkali metalsand sulphur contents were approximately 20 and 2 times higherrespectively in wheat bran, while the resulting dust emissionswere 45% lower. The dust emissions while burning hay were 60%lower than when straw was burned. However, the alkali metalsand sulphur contents in the two fuels were very similar. The rela-tively low dust emissions shown by wheat bran and hay could beexplained with the relatively high content of silica. The volatilityof potassium was probably significantly reduced by the incorpora-tion into silicate structures [34,35]. The high amount of sinteredash observed during combustion of wheat bran, 40% of slag(Fig. 4) confirms the retention of potassium in the ash.

The emissions while burning agricultural fuels were in generalhigher than when wood was used. All the fuels tested, with theexception of straw and Sorghum, fulfilled the dust emissionsrequirement of the FprEN 303-5 [41] for boiler class 3 operatedwith non-woody fuels (200 mg m�3 at 10% O2 or approximately133 mg MJ�1). The dust emissions requirements for boiler classes4 and 5 are respectively 60 and 40 mg m�3 at 10% O2, correspond-ing to approximately 40 and 27 mg MJ�1. These requirements areonly applicable to woody fuels. However, vineyard pruning andMiscanthus fulfilled the dust requirements for both boiler classes.Among the deviations in the dust emission requirements of someof the European countries, it is important to stress the cases ofAustria and Germany due to their stricter requirements. In Austria,the dust emissions of boilers with automatic stoking operated withstandardised non-woody fuels should not exceed 60 mg MJ�1 andfrom January 1, 2015 this limit will be lowered to 35 mg MJ�1.According to the current Austrian law, only Miscanthus, vineyardpruning and hay could be burned in the studied boiler. The Germandeviation only includes emission limits for non-woody biomassfuels to appliances installed after 31.12.2004. From that date on,the dust emissions from boilers with a nominal heat output

s correspond to the SEM (standard error of the mean).

Fig. 8. Boiler efficiency during combustion of the investigated fuels. The error bars correspond to the SEM (standard error of the mean).

L. Carvalho et al. / Applied Energy 104 (2013) 286–296 295

between 4 kW and 30 kW and burning e.g. straw and grains,should not exceed 20 mg m�3 at 13% O2 (approximately13 mg MJ�1). Among the agricultural fuels tested, only Miscanthuscan fulfil the future emission limits of Austria and Germany in thetested boiler.

3.4. Efficiency

Fig. 8 represents the combustion efficiency of the boiler oper-ated with the tested biomass fuels. The boiler showed efficienciesclose to or higher than 90% when operated with agricultural fuels.Boiler efficiency was highest with wheat bran (94%) and Miscan-thus pellets (94%) and lowest with Sorghum pellets (89%). The boi-ler satisfied the minimum efficiency requirements defined in theFprEN 3003-5 [41] with all the fuels tested and was always abovethe requirements for boiler classes 3, 4 and 5 (74%, 82% and 88%respectively).

4. Conclusions

The technical and environmental performance of a 15 kW pelletboiler fired with different fuels were evaluated.

– Boiler technology:� The automatic feeding system and the lambda control were

unable to adjust to an optimal level for a broad variety ofnon-woody biomass fuels.

� The design of the combustion chamber is not optimal for allthe biomass fuels (see Section 3.1.2).

� Adequate ash management devices were important toenable the boiler to manage fuels with high ash contents.

� Ash accumulation on the heat exchanger walls rapidlyreduced the boiler efficiency.

– It is recommended that boilers for burning agricultural fuels haveflexible control systems, automatic cleaning of the heat exchang-ers and should be adapted to manage high ash content fuels.

– The emissions from burning agricultural fuels were in generalhigher than when woody biomass was used. Nevertheless, allthe agricultural fuels satisfied the legal requirements definedin the FprEN 303-5 with the exception of straw and Sorghumregarding dust emissions.

– Only Miscanthus, vineyard pruning and hay can fulfil Austrianlegal requirements of NOx and dust emissions (stated in thecountry deviations of the FprEN 303-5 [41]).

– The boiler showed efficiencies close or higher than 90% with allthe agricultural fuels and met the minimum permissible effi-ciency requirement defined FprEN 303-5 for boiler classes 3–5.

Acknowledgments

The results presented have been derived in projects carried outwithin the Kplus-Programme of the Austrian Federal Government.Financial support was received from the Austrian Research Promo-tion Agency, the Styrian Regional Government, the Styrian Econ-omy Promotion Agency and the Lower Austrian RegionalGovernment. The authors are gratefully acknowledged for theirfunding. The collaborations with Hargassner and FJ-BLT Wiesel-burg are also highly acknowledged.

References

[1] Obernberger I, Thek G. The current state of Austrian pellet boiler technology.In: Proc. of the 1st world conference on pellets, Stockholm, Sweden; 2002. p.45–8.

[2] Fiedler F. The state of the art of small-scale pellet-based heating systems andrelevant regulations in Sweden, Austria and Germany. Renew Sustain EnergyRev 2004;8:201–21.

[3] Musil-Schläffer B, McCarry A, Schmidl C, Haslinger W. European wood-heatingtechnology survey: an overview of combustion principles and the energy andemissions performance characteristics of commercially available systems inAustria, Germany, Denmark, Norway and Sweden. Final report prepared for theNew York State Energy Research and Development Authority (NYSERDA);2010.

[4] Nussbaumer T. Combustion and co-combustion of biomass: fundamentals,technologies, and primary measures for emission reduction. Energy Fuels2003;17:1510–21.

[5] European Parliament, Council. Directive 2009/28/EC of the EuropeanParliament and of the Council of 23 April 2009 on the promotion of the useof energy from renewable sources and amending and subsequently repealingDirectives 2001/77/EC and subsequently repealing Directives 2001/77/EC and2003/30/EC (Text with EEA relevance). Off J Eur Union 2009;L140/16.

[6] Uslo A, Gomez NC, Belda MS. Demand for lignocellulosic biomass in Europe.Elobio Biofuels policies for dynamic markets; 2010. <http://www.elobio.eu/fileadmin/elobio/user/docs/Additional_D_to_WP3-Linocellulosic_biomass.pdf> [accessed 25.05.12].

[7] Dias J, Costa M, Azevedo JLT. Test of a small domestic boiler using differentpellets. Biomass Bioenergy 2004;27(6):531–9.

[8] Kristensen EF, Kristensen JK. Development and test of small-scale batch-firedstraw boilers in Denmark. Biomass Bioenergy 2004;26(6):561–9.

[9] Gilbe C, Öhman M, Lindström E, Lindström E, Boström D, Backman R, et al.Slagging characteristics during residential combustion of biomass pellets.Energy Fuels 2008;22(5):3536–43.

[10] Lindström E, Larsson SH, Boström D, Öhman M. Slagging characteristics duringcombustion of woody biomass pellets made from a range of different forestryassortments. Energy Fuels 2010;24(6):3456–61.

296 L. Carvalho et al. / Applied Energy 104 (2013) 286–296

[11] Verma VK, Bram S, Gauthier G, De Ruyck J. Evaluation of the performance of amulti-fuel domestic boiler with respect to the existing European standard andquality labels: Part 1. Biomass Bioenergy 2011;35(1):80–9.

[12] Fernández RG, et al. Study of main combustion characteristics for biomassfuels in boilers. Fuel Process Technol; in press. http://dx.doi.org/10.1016/j.fuproc.2011.12.032.

[13] Saidur R, Abdelaziz EA, Dermibas A, Hossain MS, Mekhilef S. A review onbiomass as a fuel for boilers. Renew Sustain Energy Rev 2011;15(5):2262–89.

[14] Carvalho L, Wopienka E, Lundgren J. Challenges in small-scale combustion ofagricultural biomass fuels. Int J Energy Clean Environ 2008;9:127–42.

[15] Wopienka E, Friedl G, Haslinger W. Reality check for agricultural biofuels. In:Proc of the world sustainable energy days (WSED), Wels; 2009.

[16] Obernberger I, Brunner T, Bärnthaler G. Chemical properties of solid biofuels-significance and impact. Biomass Bioenergy 2006;30(11):973–82.

[17] Öhman M, Boman C, Hedman H, Nordin A, Boström D. Slagging tendencies ofwood pellet ash during combustion in residential pellet burners. BiomassBioenergy 2004;27(6):585–96.

[18] Öhman M, Boström D, Nordin A, Hedman H. Effect of kaolin and limestoneaddition on slag formation during combustion of wood fuels. Energy Fuels2004;18:1370–6.

[19] Hübner C, Boos R, Prey T. In-field measurements of PCDD/F emissions fromdomestic heating appliances for solid fuels. Chemosphere 2005;58(3):367–72.

[20] Hedman B, Näslund M, Marklund S. Emission of PCDD/F, PCB, and HCB fromcombustion of firewood and pellets in residential stoves and boilers. EnvironSci Technol 2006;40(16):4968–75.

[21] Stubenberger G, Scharler R, Zahirovic S, Obernberger I. Experimentalinvestigation of nitrogen species release from different solid biomass fuelsas a basis for release models. Fuel 2008;87(6):793–806.

[22] Glarborg P, Jensen AD, Johnsson JE. Fuel nitrogen conversion in solid fuel firedsystems. Prog Energy Combust Sci 2003;29(2):89–113.

[23] World Health Organization (WHO) Europe WHO Air quality guidelines forparticulate matter, ozone, nitrogen dioxide and sulphur dioxide, global update,summary of risk assessment; 2005. <http://whqlibdoc.who.int/hq/2006/WHO_SDE_PHE_OEH_06.02_eng.pdf> [accessed 26.07.11].

[24] Szidat S, Ruff M, Perron N, Wacker L, Synal HA, Hallquist M, et al. Fossil andnon-fossil sources of organic carbon (OC) and elemental carbon (EC) inGöteborg, Sweden. Atmos Chem Phys 2009;9:1521–35.

[25] Caseiro A, Bauer H, Schmidl C, Pio CA, Puxbaum H. Wood burning impact onPM10 in three Austrian regions. Atmos Environ 2009;43(13):2186–95.

[26] Glasius M, Ketzel M, Wåhlin P, Jensen B, Mønster J, Berkowicz R, et al. Impactof wood combustion on particle levels in a residential area in Denmark. AtmosEnviron 2006;40(37):7115–24.

[27] Yttri KE, Dye C, Braathen O-A, Simpson D, Steinnes E. Carbonaceous aerosols inNorwegian urban areas. Atmos Chem Phys 2009;9:2007–20.

[28] Bari MA, Baumbach G, Kuch B, Scheffknecht G. Temporal variation and impactof wood smoke pollution on a residential area in southern germany. AtmosEnviron 2010;44(31):3823–32.

[29] Schmidl C, Luisser M, Padouvas E, Lasselsberger L, Rzaca M, Cruz CR-S, et al.Particle and gaseous emissions from manually and automatically fired smallscale combustion systems. Atmos Environ 2011;45(39):7443–54.

[30] Bäfver LS, Leckner B, Tullin C, Berntsen M. Particle emissions from pellet stovesand modern and old-type wood stoves. Biomass Bioenergy 2011;35(8):3648–55.

[31] Wiinikka H, Gebart R, Boman C, Boström D, Nordin A, Öhman M. High-temperature aerosol formation in wood pellets flames: spatially resolvedmeasurements. Combust Flame 2006;147(4):278–93.

[32] Valmari T, Kauppinen EI, Kurkela J, Jokiniemi JK, Sfiris G, Revitzer H. Fly ashformation and deposition during fluidized bed combustion of willow. J AerosolSci 1998;29(4):445–59.

[33] Sippula O, Hytönen K, Tissari J, Raunemaa T, Jokiniemi J. Effect of wood fuel onthe emissions from a top-feed pellet stove. Energy Fuels 2007;21(10):1151–60.

[34] Jensen PA, Frandsen FJ, Dam-Johansen K, Sander B. Experimental investigationof the transformation and release to gas phase of potassium and chlorineduring straw pyrolysis. Energy Fuels 2000;14(6):1280–5.

[35] Knudsen JN, Jensen PA, Dam-Johanen K. Transformation and release to the gasphase of Cl, K and S during combustion of annual biomass. Energy Fuels2004;18(5):1385–99.

[36] Sippula O, Lind T, Jokiniemi J. Effects of chlorine and sulphur on particleformation in wood combustion performed in a laboratory scale reactor. Fuel2008;87(12):2425–36.

[37] Obernberger I. Decentralized biomass combustion: state of the art and futuredevelopment. Biomass Bioenergy 1998;14(1):33–56.

[38] Baxter LL, Miles TR, Miles Jr TR, Jenkins BM, Milne T, Dayton D, et al. Thebehaviour of inorganic material in biomass-fired power boilers: field andlaboratory experiences. Fuel Process Technol 1998;54(1–3):47–78.

[39] Werther J, Saenger M, Hartge E-U, Ogada T, Siagi Z. Combustion of agriculturalresidues. Prog Energy Combust Sci 2000;26:1–27.

[40] Van Loo S, Koppejan J, editors. Handbook of biomass combustion and co-firing.IEA Bioenergy Task 32; 2002. ISBN:9036517737.

[41] Final draft of the European Standard FprEN 303-5. Heating boilers – Part 5:heating boilers for solid fuels, manually and automatically stoked, nominalheat output of up to 500 kW – terminology, requirements, testing andmarking; February 2012.

[42] Picco D, Vecchiet A, Santacroce FC, Pointner C, Wopienka E. Combustion testsof pellet made of different herbaceous energy crops in a small scale pelletboiler. In: Proc of the 19th European biomass conference and exhibition,Berlin, Germany; 2011.

[43] Wopienka E. Properties and melting behaviour of ashes from small-scalebiomass combustion. PhD dissertation (unpublished results). Institute ofChemical Eng. TU Wien.

[44] Obernberger I, Thek G. Physical characterisation and chemical composition ofdensified biomass fuels with regard to their combustion behaviour. BiomassBioenergy 2004;27(6):653–69.

[45] Zamorano M, Popov V, Rodríguez ML, García-Maraver A. A comparative studyof quality properties of pelletized agricultural and forestry lopping residues.Renew Energy 2011;36(11):3133–40.

[46] Ohnmacht R. Untersuchung der Fördercharakteristik von Schneckenförderernfür eine Pelletabwurffeuerung. Diplomarbeit: Institute of Chemical Eng. TUWien; 2009.

[47] Dai J, Grace JR. Biomass granular screw feeding: an experimental investigation.Biomass Bioenergy 2011;35(2):942–55.

[48] Theerarattananoon K, Xu F, Wilson J, Ballard R, Mckinney L, Staggenborg S,et al. Physical properties of pellets made from Sorghum stalk, corn stover,wheat straw and big bluestem. Ind Crops Prod 2011;33(2):325–32.

[49] Dahl J, Obernberger I. Evaluation of the combustion characteristics ofperennial energy crops (Arundo donax, Cynara caadunculus, Miscanthus �giganteus and Panicum virgatum). In: Proc of the 2nd world conference andexhibition on biomass for energy, industry and climate protection, Rome, Italy;2004.

[50] Garijo EG, Jensen AD, Glarborg P. Kinetic study of NO reduction over biomasschar under dynamic conditions. Energy Fuels 2003;17(6):1429–36.

[51] Staiger B, Unterberger S, Berger R, Hein KRG. Development of an air stagingtechnology to reduce NOx emissions in grate fired boilers. Energy2005;30(8):1429–38.

[52] Aarna I, Suuberg E. The role of carbon monoxide in the NO–carbon reaction.Energy Fuels 1999;13(5):1145–53.

[53] Sørensen CO, Johnsson JE, Jensen A. Reduction of NO over wheat straw char.Energy Fuels 2001;15(6):1359–68.

[54] Verma VK, Bram S, Delattin F, Laha P, Vandendael I, Hubin A, et al. Agro-pelletsfor domestic heating boilers: standard laboratory and real life performance.Appl Energy 2011;90(1):17–23.

[55] Zevenhoven R, Hupa M. The reactivity of chars from coal, peat and woodtowards NO, with and without CO. Fuel 1998;77(11):1169–76.

[56] Brunner T, Bärnthler G, Obernberger I. Fine particulate emissions from state-of-the-art small-scale Austrian pellet furnaces – characterization, formationand possibilities of reduction. In: Proc of 2nd world conference on pellets, 30May–1 June, 2006, Svebio, Jönköping, Sweden; 2006. p. 87–91.