Contents lists available at ScienceDirect

Fuel

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

Full Length Article

Adaptation of hard coal with high sinterability for solid fuel boilers inresidential heating systems

Janusz A. Lasek⁎, Katarzyna Matuszek, Piotr Hrycko, Małgorzata PiechaczekInstitute for Chemical Processing of Coal, Zamkowa 1, 41-803 Zabrze, Poland

G R A P H I C A L A B S T R A C T

A R T I C L E I N F O

Keywords:Coal combustionCoal sinterabilityFuel flexibility

A B S T R A C T

Solid fuels boilers with low emission and modern construction, which are used for residential heating, requirecoal fuel with low sinterability (RI < 15). This paper presents a process for adapting coal with high sinterabilityfor use in small boilers in residential heating systems. This solution allows for the combustion of coals with an RIof 50 without causing negative effects (in terms of combustion stability) in typical burners used in heatingdevices. The proposed solution involves the delivery of fuel particles coated with an oil film (waste oil andcanola oil) to the combustion chamber. Tests were performed using a commercial-scale boiler with a 17–23 kWoutput, equipped with a retort burner and a dosing device, which applied oil to the coal in the boiler. Tests wereperformed during the combustion of hard coals characterized by different sinterability, with and without variousstreams of oils. Emission issues are also analysed. Impregnation of the coal by canola oil caused slightly lowerNOx emissions, but higher CO and dust emissions. Analyses of the basic quality parameters (i.e., coal structureafter thermal treatment) show that changes in the coal properties also occurred. Based on these studies, a me-chanism that explains the beneficial effect of oil presence on high sinterable coal (HSC) combustion is proposed.

1. Introduction

The application of small, domestic-scale boilers and stoves(< 50 kW) as a heat source for daily activities (i.e., house and waterheating, cooking etc.) is widespread all over the world. Depending onregions, social issues and fuel resources, different fuel types are used,such as gaseous fuels (mainly natural gas) [1], coal [2–4], fossil andbioliquid fuels [5], and solid biomass (e.g., wood, agricultural productsand residues) [6–9]. The use of coal in domestic applications is

problematical in most EU countries. However, coal-fired boilers andstoves are still used in domestic applications. For example, in the CzechRepublic in 2011, out of a total of 3.6 million households, 9.2% wereheated by coal and 7.8% were heated by wood. Moreover, the problemof air pollution is enhanced due to the illegal combustion of municipalsolid waste in domestic furnaces [10]. In Poland, as of 2015, the shareof hard coal in total energy consumption in households was 32.11%[11]. In China, residential coal stoves are commonly used for cookingand heating, especially in the winter [2]. In some regions (e.g., Jingxian

https://doi.org/10.1016/j.fuel.2017.11.020Received 26 April 2017; Received in revised form 28 July 2017; Accepted 7 November 2017

⁎ Corresponding author.E-mail address: jlasek@ichpw.pl (J.A. Lasek).

Fuel 215 (2018) 239–248

Available online 21 November 20170016-2361/ © 2017 Elsevier Ltd. All rights reserved.

T

and Xinmi), the total share of coal per capita in rural household energyconsumption is more than 25% [12]. The use of coal for residentialheating around the world was analysed by Kerimray et al. [13], whonoticed that certain counties have higher per capita residential coalconsumption (in kilograms of oil equivalent per capita (kg OE/cap),2014): i.e., Poland (165 kg OE/cap), Kazakhstan (157 kg OE/cap),Mongolia (104 kg OE/cap), Ireland (87 kg OE/cap), South Africa(69 kg OE/cap), the Czech Republic (49 kg OE/cap), China (36 kg OE/cap), the Republic of Korea (14 kg OE/cap) and Hungary (10 kg OE/cap). Ryfa et al. [14] reported that about 203 million households inEurope are mostly equipped with a single heating system (boilers). Theresidential heating sector in Poland consumes approximately 12 milliontons of coal annually. Due to the price competitiveness of this fuel,compared to other fossil fuels or biomass used in this sector, the si-tuation will not change in the near future. It is estimated that ap-proximately 80% of the heating infrastructure is outdated, negativelyaffecting air quality and requiring frequent cyclic operation and su-pervision. Boilers with modern, low-emission construction, which meetthe standard PN-EN 303-5: 2012, require the use of fuel with low sin-terability (Roga index, RI < 15). Automatic coal-fired boilers (ABs), incontrast to traditional manually operated boilers (TBs), are recognizedas environmentally friendly with regard to domestic applications. Thetotal emission of gaseous pollutants (e.g., CO, VOCs) from these boilersis several times lower than that for TBs [3,15]. However, NOx emissionsare approximately 40% higher (or more) from ABs than from TBs [3].Small-scale commercial boilers (automatic or manually operated) werepreviously investigated with respect to efficiency, particulate matter(PM) and gaseous pollutant emissions (including NOx), and otherparameters [3,14–23]. Due to the increase in demand for energy andenvironmentally friendly combustion systems, different variations offuel blends have been investigated. The change in heat source (fuelswitch) in a short period of time is very difficult to manage. Thus, sometransition solutions to coal firing and other energy sources have beensuggested. The combustion performance of a mixture of different liquidfuels [5], a mixture of coal and process gas [3], a mixture of coal orlignite and biomass [4,24], and a mixture of different solid biomass[6,7,25,26] has been investigated in terms of efficiency and PM emis-sions. Xu et al. [4] observed that pelletized biomass and coal blends cansignificantly decrease emissions. Hrycko et al. [3] proposed the appli-cation of gas from the gasification process as a reburning agent.Moreover, they obtained high fuel flexibility because of the reburningfuel heat input of 63%. Buczynski et al. [18,19] and Ryfa et al. [14]performed experimental and mathematical modelling of small-scalecountercurrent fixed-bed automatic boilers. The computations includedthe combustion of coal and heat transfer from the flue gas into heatedwater. Thus, a computational fluid dynamics-based model of the wholeboiler was taken into account. Rabacal et al. [20] found that the type offuel significantly affected boiler emission characteristics (i.e., CO, NOx

and hydrocarbons). The fuel-NO mechanism is the main source of NOx

emissions in this type of boiler. A computational fluid dynamics ana-lysis of a small-scale commercial biomass pellet boiler was also pre-sented by Porteiro et al. [27]. The fuel type and its blends have alsobeen investigated in order to obtain an efficient multifuel boiler. Forexample, González-González et al. [5] presented a study involving thecombustion of different mixtures of biodiesel/gasoil under differentoperating conditions using a 26.7 kW domestic heating boiler. Schön-nenbeck et al. [6] analysed the energy recovery of grape marc in a30–40 kW multifuel domestic-scale boiler. Different blends of grapemarc and other biomasses (i.e., wood pellets and Miscanthus) weretested in terms of combustion performance and emission issues. Leys-sens et al. [7] presented a study on the combustion performance ofpellets made from sawdust and reused cardboard.

It is worth emphasizing that the co-combustion of coal and liquidfuels in domestic-scale boilers has not been investigated before.However, co-combustion of coal and glycerol (1 wt%) in industrialpulverised fuel boiler (180MWth) was reported by Topolnicka et al.

[28]. Some liquid fuels also have the status of renewable sources ofenergy. If canola oil cannot be used in the food industry (due to in-sufficient parameters), it can be employed as a fuel. Bhimani et al. [29]noticed that canola oil has a negligible ash content and can be used inboilers as an almost foul-free fuel.

The combustion of HSC is problematic in modern ABs. The mainnegative phenomenon involves the effect of AB burner blocking. Thus,it is necessary to find a solution to this problem, as it has not beenpreviously investigated. Another novelty of this work is the extendedfuel flexibility of a solid fuel-fired domestic-scale boiler using a liquidfuel as a secondary fuel. The objective of this work is to demonstratethat the stable combustion of HSCs in a domestic automatic boiler ispossible after coal impregnation by liquid fuels. Combustion tests arealso performed, along with an analysis of combustion parameters(flame stability, boiler efficiency) and the emissions of particles andgaseous pollutants.

2. Experimental

2.1. Fuels analysis

The proximate analyses of the examined coal samples were per-formed according to the technical procedure Q/LP/06/A:2011 for thedetermination of ash content and Q/LP/05/A:2011 for the determina-tion of moisture content. The volatile matter analysis was conducted inaccordance with the technical procedure Q/LP/07/A:2011. The ulti-mate analysis was performed using a LECO TrueSpec (LECO, USA) CHNanalyser and a LECO SC 632 analyser (LECO, USA). The heating valueswere analysed using LECO AC500 (LECO, USA) apparatus. The oil ig-nition temperature was determined by Marcusson method. The liquidfuel sample was placed in an open vessel and it was heated up (rate of3 °C/min). The gaseous fuel flame was approaching the liquid fuelsample within the temperature increase of 1 °C until ignition of liquidfuel volatiles appeared. The viscosity of liquid fuels samples were de-termined using Engler viscometer.

Fuel samples were also observed using an Axio Imager M1m (CarlZeiss, Germany) optical light microscope. Samples of coal were pro-cessed before the microscopic observation. The coal samples were cut inorder to obtain 2 cm cubes, then combusted in a flame. After combus-tion, the samples were embedded in acrylic resin, which was cut toobtain cross sections and then polished. Sample cross sections wereobserved using reflected polarized light. The results of the analyses ofthe applied fuels are presented in Tables 1 and 2. HSC was used in thetests. Usually, this type of coal is applied to the coking process. How-ever, this type of coal was used to evaluate whether the proposed coal

Table 1Proximate and ultimate analysis results of low sinterable coals and HSCs.

Parameter Unit Low sinterable coal High sinterable coal

Mrt % 10.1 ± 0.3 2.4 ± 0.3

Ma % 4.3 ± 0.1 1.0 ± 0.1Aa % 4.1 ± 0.1 7.0 ± 0.1Va % 34.16 ± 0.12 30.16 ± 0.12Vdaf % 37.29 ± 0.14 32.78 ± 0.14Cat % 75.8 ± 0.4 79.6 ± 0.4

Hat % 4.56 ± 0.16 4.5 ± 0.16

Nat % 1.22 ± 0.10 1.29 ± 0.10

Oad % 9.63 ± 0.31 6.07 ± 0.31

Sat % 0.56 ± 0.03 0.65 ± 0.03SaC % 0.39 ± 0.03 0.11 ± 0.03SaA % 0.17 ± 0.03 0.54 ± 0.03HHVa kJ/kg 30345 ± 68 32700 ± 68LHVa kJ/kg 29244 ± 70 31693 ± 70LHVr kJ/kg 27324 ± 94 31210 ± 94RI – 11 ± 2 57 ± 2

a – analytic basis, r – as received basis, daf – dry ash free basis, M – moisture content, A –ash content, V – volatile matter content.

J.A. Lasek et al. Fuel 215 (2018) 239–248

240

processing is effective regarding the stable combustion of HSC. Thesecond coal (with lower sinterability) is also important from a practicalpoint of view. The application of these coals is troublesome in thecoking process (too low RI) and low-emission retort boilers (too highRI). Thus, if a method for the stable combustion of high sinterable coalis determined, it will also be useful for medium sinterable coals.

2.2. Combustion tests

Experiments were conducted using a domestic commercial-scaleretort boiler (nominal thermal power of 20 kW). The boiler was fedwith Polish hard coal (which is called “pea coal”, whose grain diameterranges from 5 to 25mm). The canola oil was originally used in foodindustry. The waste oil (spent engine oil) came from cars service center.The scheme of a boiler is shown in Fig. 1. The combustion chamber wasequipped with a horizontal flame burner known as a “retort burner”.Fuel was fed from a hopper into the burner by a horizontal screwfeeder, working by impulses.

To determine the oil-coal ratio, a preliminary test (without com-bustion, cold conditions) was conducted. Samples of oils (50 g) weretaken from oil tanks, while the coal samples were placed (together withoil) in glass chambers. The oil film was drained from the coal after 2 h.It was assumed that the increase of coal mass was equal to the maximalcapacity of the coal and oil connection. An oil feeding test was per-formed using the boiler screw without combustion (called the cold test).The control valve of the oil was set in order to obtain an oil feeding rateof 3.6 kg/h. The coal feeding rate was 18 kg/h; thus, the oil massfraction was∼17%. Combustion tests were conducted for oil fractionsin the range of 10–30%. Fig. 2 shows the coal samples before (right)and after (left) feeding/impregnation with oil. The oil was uniformlyspread on the coal surface.

The temperature of the inlet and outlet circulating water wasmeasured using a resistance thermometer (Pt-100, accuracy below0.5 °C). The composition of the exhaust gas from the boiler was

measured continuously using a mobile set of analysers manufactured bySiemens, consisting of two types of analysers: Ulramat (non-dispersiveinfrared spectroscopy) and Oxymat 61 Siemens (paramagnetic). Theflue gas was analysed for CO (0–5 vol%), CO2 (0–25 vol%) and NO(0–1000 ppm) contents. The flue gas from the boiler stack was con-tinuously sampled by a probe system with a heated ceramic filter, aheated hose and a gas conditioning system (PSS-5, M&C Products). Heatfrom the boiler was removed by cooling towers. The boiler efficiencywas determined using formulas from the standard PN-EN 303-5:2012.The fundamental equation in the procedure wasη=(1− qA− qU− qS− qB)× 100%, where qA is the loss throughsensible heat of the products of combustion (values relative to the heatinput), qU is the loss through incomplete combustion (values relative tothe heat input), qS is the loss through radiation, convection and con-duction (values relative to the heat input), and qB is the loss throughunburned fuel in ash (values relative to the heat input).

2.3. Experimental uncertainty analysis

It is known that the relative (ΔA) uncertainty and the absolute (δA)uncertainty of a considered parameter A are related by the equationΔA= δA/A×100%. Additionally, it can be assumed that, for measuredparameters A and B, the rules for the uncertainty calculation are

± + ± ≅ + ± +A δA B δB A B δA δB( ) ( ) ( ) ( ),± − ± ≅ − ± +A δA B δB A B δA δB( ) ( ) ( ) ( ),± × ± ≅ × ± +A A B B A B A B( Δ ) ( Δ ) ( ) (Δ Δ ),± ± ≅ ± +A A B B A B A B( Δ )/( Δ ) ( / ) (Δ Δ ).For example, for the decrease in combustible matter in the ash ΔCM

(see Fig. 6 in Section 3), the uncertainty can be estimated by the fol-lowing procedure.

If the combustible constituents in the residues (combustible matter,CM) in the ash are calculated using the equation CMi= 100%-Ari,where i= coal or coal+ liquid fuel, Ar is ash content in the matter aftercombustion (i.e., in the slag of coal or coal+ liquid combustion, as areceived state), then absolute uncertainty is δCMi= δAri and relativeuncertainty is ΔCMi= δAri/CMi× 100%.

The decrease in combustible constituents in the residues is:

= − +ΔCM (CM CM )/CMcoal coal liquid coal (1)

Thus, relative uncertainty of ΔCM is ΔΔCM=(δCMcoal+ δCMcoal+liquid)/(CMcoal− CMcoal+liquid)× 100%+ ΔCMcoal = (δAr coal + δAr coal+li-

quid)/(CMcoal− CMcoal+liquid)× 100%+ δAr coal/CMcoal × 100%.For the presented procedure, the relative uncertainty of CM and

boiler efficiency did not exceed 2.8% and 3%, respectively. A similarprocedure for experimental uncertainty analysis was recently presented

Table 2Analysis results of the applied liquid fuels (as received).

Parameter Unit Canola oil Waste oil

C % 77.5 87.9H % 10.90 13.92S % 0.01 0.30LHV kJ/kg 37,200 41,368Density (at 15 °C) kg/m3 915 922Viscosity (at 20 °C) m2/s 83× 10−6 107× 10−6

Ignition temperature °C 321 192

Fig. 1. Scheme of the boiler with a system for oil application: 1) boiler, 2) heat exchanger(flue gas-water), 3) retort burner, 4) screw feeder with electric drive, 5) coal tank, 6) airfan, 7) main oil tank, 8) valve to control oil feeding and 9) medial oil tank.

Fig. 2. Coal before (right) and after (left) the feeding of oil.

J.A. Lasek et al. Fuel 215 (2018) 239–248

241

by Lasek et al. [30].

3. Results and discussion

3.1. Combustion test

The combustion of HSC (without processing) is not stable. In thepreliminary test, a problematic process scheme was observed as follows.The introduced portion of coal was preliminarily ignited; however, thefurther addition of coal caused agglomeration and fused pieces. As aconsequence (due to the sintering process), the flame was inhibited andthe temperature inside the combustion chamber decreased dramati-cally, such that further stable combustion was not possible.

One possible way to obtain stable combustion properties of HSC isby pyrolysis. During the pyrolysis process, the coal passes through theplastic stage, which means that the reactive coal macerals, such as vi-trinite, liptinite and a one-third part of semi-fusinite, start to soften andbecome a plastic mass. During the plastic stage, ordering of the carbonmolecular structure occurs, while optically isotropic coal becomes ananisotropic carbon material in polarized reflected light. After the firststep of degasification, decomposition and depolymerization, the plasticmass commences resolidation to become semi-coke. Coal grains ag-glomerate in a state of plastic mass if they are compacted and theirplastified surfaces come into contact. The combustion process of coke orsemi-coke is considered in processes including catalyst regeneration[31,32], oil recovery [33], and combustion in a porous bed reactor[34]. Pyrolysis performed as a separate process before combustion is anenergy- and cost-consuming process. Thus, it was assumed that thepyrolysis process could be conducted inside a boiler if the coal were tobe covered with liquid fuel (e.g., waste oil or vegetable oil). The co-combustion of coal and oil droplet mixtures was presented by Wilk et al.[35]. Combusting an oil film on the coal surface inhibits oxygentransport to the fuel surface, such that a coking “in situ-like” processduring combustion can be achieved. This process is discussed and ex-plained in the following section.

The assumed power of the boiler (17–22 kW) was achieved in everycase. Examples of measured parameters concerning the function of timeare presented in Figs. 3 and 4. These results were obtained duringcombustion of HSC with waste oil 77/23. The measured parameterschange periodically; however, they represent stable conditions ofcombustion. From every test under quasi-stable conditions, the averagevalues of the parameters are determined as arithmetic mean values.Moreover, the standard deviation of the analysed parameters can beestimated, which can help to evaluate the range of periodic change inmeasured parameters. The mean values and standard deviations(marked as± ) are: O2 11.0 ± 1.0 vol%, CO 605 ± 157 ppm, NO266 ± 18 ppm, SO2 230 ± 23 ppm, CO2 8.4 ± 0.9 vol%, outputpower 17.5 ± 0.8 kW. The relative change in flue gas concentration(calculated as standard deviation/mean value×100%) was no higher

than 12% in almost all cases. Only the change in CO was greater than26%. The variations in measured parameters in the case of domesticboilers have also been observed by other researchers [9,36,37]. Moreintensive variations of CO concentration, compared to value changes inother parameters (i.e., O2, CO2 and NOx), were confirmed by Liu et al.[36] and Lajili et al. [37]. The mean values of certain measured para-meters were transformed into units of mg/mn

3 (ref. to 10% O2).The variations in the values were caused by the periodic character

of process parameters, i.e., solid fuel feeding (5 s feeding and a 23 spause), as well as the spontaneous removal of ash from the retortburner. Moreover, it should be mentioned that the additional benefit ofco-combustion is the increase in bio-originated fuel fraction. In otherwords, the part of fossil fuel is replaced by biofuel. The maximal liquidfuel share (in terms of calorific value) is 36.8%. The liquid fuel “lf”share (in terms of calorific value), zElf, is calculated using the equationzElf = glf × LHVr

lf/(glf × LHVrlf + (1− glf)× LHVr

coal), where g is themass fraction of liquid fuel (lf) and coal.

The boiler efficiency, flue gas temperature and combustible con-stituents in the residues are presented in Fig. 5. It is known that the lossof boiler efficiency is caused by three main phenomena: i.e., thermalheat losses in the flue gas, chemical heat losses in the flue gas, and heatlosses due to combustible constituents in the residues [3,17]. Thelowest thermal efficiency of a boiler (i.e., 78.3%) was obtained for thecombustion of HSC with canola oil (15 wt%). This can be explained bythe high temperature of flue gas, relatively high oxygen excess and COconcentration (see Table 3), and high combustible constituents in ash.According to the literature, the thermal efficiency of solid fuel-fireddomestic boilers can achieve values in the wide range of 85–97% [7]and 74.7–84% [14,19]. Some regulations (NF EN 12809 standard) re-lated to boilers with a heating power lower than 50 kW require an ef-ficiency of at least 85% [6]. The combustion efficiency requirement ofthe EN 303-5 standard is equal to 80% [7]. According to the PN-EN303-5:2012 standard, requirements (in terms of efficiency) depend on aboiler’s class and power. Depending on emissions and efficiency, atested boiler can be classified into the proper class. It should be notedthat a higher class means a more efficient and more environmentallyfriendly unit. For smaller domestic boilers (power Q < 100 kW) thefollowing equations (Eqs. (2)(4)) are suggested (according to the PN-EN303-5:2012 standard) in order to determine the class of a boiler:

Fifth class, Q < 100 kW:

= +η 87 logQ,%k (2)

Fourth class, Q < 100 kW:

= +η 80 2logQ,%k (3)

Third class, Q < 300 kW:

= +η 67 6logQ,%k (4)

where ηK is boiler efficiency (%) and Q is output power (kW).In this study, the aforementioned level of 85% is achieved by the

reference test (86.8%), the mixture of coal (RI= 11) with canola oil(70/30; 88.8%), and the mixture of coal (RI= 11) with canola oil (85/15; 86.2%). The highest level (in terms of efficiency), namely, the fifthclass, is achieved by the mixture of coal (RI= 11) with canola oil (70/30; 88.8%). The level of the fourth class (in terms of efficiency) isachieved by the reference test (86.8%) and the mixture of coal(RI= 11) with canola oil (85/15; 86.2%). The rest of the cases achievedthe level of the third class (in terms of efficiency).

In almost all cases, the combustible matter decreased (see Fig. 6, ascalculated from Eq. (1)) compared to the ash from the combustion ofcoal without oil addition. The loss of ignition also decreased. In the caseof canola oil (mass fraction of 30%), combustible matter significantlyincreased. The flue gas compositions from the combustion of coals withliquid fuels are presented in Table 3. The oxygen fraction at the in-flamezone was in the range of 6.7–11%, which is typical for such coal-fireddomestic boilers. Garcia-Maraver et al. [8] reported that excess air can

Fig. 3. Examples of measured O2 vol%, CO2 vol%, and estimated power during thecombustion of HSC with waste oil (77/23).

J.A. Lasek et al. Fuel 215 (2018) 239–248

242

also be higher (referring to O2 of∼18%). All emission values werecalculated for reference 10% O2 in flue gas. The values of the referencetest for HSC combustion was not presented due to unstable combustionparameters.

The emission of PM and gaseous pollutants from our study is com-pared with the other results in Table 4. The values (taken from litera-ture) refer to 10 vol% of O2 in flue gas. Biomass-fired and coal-firedsmall-scale boilers were taken into account. The overall conclusion

from this comparison is that the emission of pollutants is at a com-parable level (order of magnitude).

The emission of CO was in a wide range of 58mg/mn3 to 2427mg/

mn3. However, the typical value is in the range of several hundred mg/

mn3. This is consistent with the CO emission from other boilers de-

scribed in the literature [6–8,14,19] (see Table 4). The combustion ofsome fuel blends (low sinterable coal with canola oil, 85/15 and 30/70)are characterized by high levels of CO emissions (i.e., 2427mg/mn

3 and1270mg/mn

3), which are considerably higher than emissions from thecombustion of coal without impregnation (i.e., 788mg/mn

3). On theother hand, the combustion of coal impregnated by waste oil (90/10)emits a considerably lower amount of CO (i.e., 335mg/mn

3). Accordingto PN-EN 303-5:2012, the permissible emission of CO for automaticboilers fired by fossil fuels is 3000mg/mn

3, 1000mg/mn3 and 500mg/

mn3 for third, fourth and fifth classes, respectively. It is known that CO

emissions from small-scale boilers depend on the boiler type, burnerconstruction, thermal load and air excess ratio [8,19]. Verma et al. [38]described the T-3 rule; T-3 stands for temperature (high combustiontemperature), turbulence (to mix fuel with the necessary oxygen) andtime (sufficient residence time for the combustion products to reachcomplete combustion). These authors concluded that emissions of COfrom a combustion device could result from a low combustion tem-perature, insufficient oxygen, poor mixing of fuel with the combustionair and/or too short a residence time of the combustion gases in thecombustion zone. Garcia-Maraver et al. [8] noticed that CO emissionswere affected by the gas residence time in the combustion chamber,temperature and turbulence/mixing. NOx emissions depend on theabove-mentioned process parameters, as well as fuel properties. It isknown that a crucial parameter influencing NOx emissions during the

Fig. 4. Examples of measured CO ppm, SO2 ppm andNO ppm during combustion of HSC with waste oil(77/23).

Fig. 5. Boiler efficiency, flue gas temperature and combustible constituents in the re-sidues during the combustion of coals with liquid fuels. The maximal relative un-certainties of the parameters were 0.3% for the flue gas temperature, 0.3% for thecombustible constituents and 3% for boiler efficiency.

Table 3Flue gas composition during the test for coal combustion.

Test No. Fuel O2 Referred for 10% O2 in flue gas

CO NOx CO2 dust TOC 16 WWA B(a)P% mg/m3

n mg/m3n % mg/m3

n mg/m3n mg/m3

n μg/m3n

1 Low sinterable coal RI= 11 8.60 788 363 9.34 56 60 0.4 32.22 Blend of low sinterable coal (85 wt%) and canola oil (15 wt%), 85/15 6.67 2427 320 9.74 302 65 1.02 36.93 Blend of low sinterable coal (70 wt%) and canola oil (30 wt%), 70/30 7.52 1270 346 9.56 235 70 1.47 57.44 Blend of low sinterable coal (90 wt%) and waste oil (10 wt%), 90/10 9.16 335 376 7.99 96 59 0.39 2.25 High sinterable coal RI= 57 Unstable values6 Blend of high sinterable coal (85 wt%) and canola oil (15 wt%), 85/15 10.80 894 584 9.29 168 130 0.56 9.27 Blend of high sinterable coal and waste oil, 77/23 10.99 831 599 9.28 125 95 0.53 5.4

J.A. Lasek et al. Fuel 215 (2018) 239–248

243

combustion of solid fuels is N-fuel content [39]. However, the impact offuel properties and process parameters on NOx emissions from do-mestic-scale boilers is still being discussed. Garcia-Maraver et al. [8]concluded that fuel-NO is the main mechanism of NO formation in thistype of boiler since the temperatures inside the combustion chamberare typically too low (< 1200 °C at the in-flame zone) for the onset ofthe NO thermal mechanism. However, they noticed inconsistent results

when NOx emissions varied for the fuels consisting of similar amount ofnitrogen (i.e., Portuguese pine (N=0.5 wt%), Spanish pine (N=0.3 wt%), cork (N=0.4 wt%), olive wood (N=0.3 wt%)). They explainedthis phenomenon with reference to the difference in process conditions.During combustion, the occurrence of local fuel-rich conditions, due toturbulence and mixing, can result in the relatively fast conversion offuel carbon into CO, which competes for oxygen, leading to a reduced

Fig. 6. Decrease in combustible matter in the ash.

Table 4The comparison of CO, NOx and PM emissions (ref. to 10 vol% O2 in flue gas).

Boiler type Ref.

Fuel type/blend CO NOx dust

mg/m3n mg/m3n mg/mn3

This study1 Low sinterable coal RI= 11 788 363 562 Blend of low sinterable coal (85 wt%) and canola oil (15 wt%), 85/15 2427 320 3023 Blend of low sinterable coal (70 wt%) and canola oil (30 wt%), 70/30 1270 346 2354 Blend of low sinterable coal (90 wt%) and waste oil (10 wt%), 90/10 335 376 965 High sinterable coal RI= 57 Unstable values6 Blend of high sinterable coal (85 wt%) and canola oil (15 wt%), 85/15 894 584 1687 Blend of high sinterable coal and waste oil 77/23 831 599 125

Multi-fuel boiler (30–40 kW), moving step grate [6]8 Wood pellets, LHVa= 17.7MJ/kg, N=0.1 wt% 58 97 209 Miscanthus, LHVa=16.7MJ/kg, N=0.3 wt% 61 222 4010 Blend of wood pellets (67 wt%) and grape (33 wt%), WP/G 67/33, LHVa=13.2MJ/kg, N=0.91 wt% 1895 158 4411 Blend of Miscanthus (67 wt%) and grape (33 wt%), M/G 67/33, LHVa 12.5MJ/kg, N=1.0 wt% 439 251 62

Dry grape, LHVa= 14.0MJ/kg, N=2.5 wt% 36608 220 221

Pellets boiler, 8–12 kW (automatic feeding) [7]12 Sawdust, LHVa=16.5MJ/kg, N < 0.1 wt% 283 144 2413 Blend of cardboard (50 wt%) and sawdust (50 wt%), LHVa= 14.5MJ/kg, N < 0.1 wt% 67 190 1214 Blend of cardboard (74 wt%) and sawdust (26 wt%), LHVa= 14.1MJ/kg, N < 0.1 wt% 484 216 1815 Cardboard, LHVa= 13.9MJ/kg, N < 0.1 wt% 158 239 26

Domestic top-feed pellet-fired boiler with a maximum thermal capacity of 22 kW (10–17 kW), with forced draught; data for 17 kW [8]16 Portuguese pine, LHV=17.1MJ/kg, N=0.5 wt% 772 96 13917 Spanish pine, LHV=16.9MJ/kg, N=0.3 wt% 1053 149 12618 Cork, LHV=19.1MJ/kg, N=0.4 wt% 1225 296 26019 Olive wood, LHV=16.0MJ/kg, N=0.3 wt% 1157 386 31620 Olive pruning, LHV=16.5MJ/kg, N=1.1 wt% 1494 616 616

Coal-fired boilers [14,19]21 domestic central heating retort boiler (25 kW) demineralised coal of commercial name ECORET, LHV=28.24MJ/kg, N=1.39wt

%)445 402 n/a [19]

22 domestic central heating retort boiler (25 kW), demineralised coal of commercial name ECORET, LHVa= 30.54MJ/kg,Na= 1.49 wt%)

378 267 n/a [14]

a On dry basis.

J.A. Lasek et al. Fuel 215 (2018) 239–248

244

availability of oxygen for the fuel-NO formation. Another possible ex-planation is that the reduction in the fuel-NO formation may result fromcatalytic effects of the char surface, due to the presence of ash, on thegaseous NO. Similar to Garcia-Maraver et al. [8], Verma et al. [38,40]suggested that NOx emissions in small domestic-scale boilers can beaffected by the catalytic effects of the ash surface on NOx formation[40] and/or the catalytic effects of the char surface on the gas phase NOreduction [38]. In our study, NOx emissions changed in the range of320–599mg/mn

3. This emissions level is related to N-fuel content (seeTable 1) and other parameters (T-3). It should be emphasized thattemperature can play a crucial role in NOx conversion during thecombustion of coal on a domestic scale. Even though some researchers(e.g., Garcia-Maraver et al. [8] and Verma et al. [40] Carvalho et al.[41]) have claimed that fuel-N plays a main role in NOx conversionduring combustion in domestic-scale boilers, it should be explained thattheir investigations were carried out using biomass as a fuel. In thesecases, the temperature inside the flame zone reached up to 1200 °C [8].Buczyński et al. [19] noted that the highest temperature (∼1360 °C)occurs in the flame during coal combustion in domestic-scale boilers,just below the deflector. It is known that the Zeldovich mechanism(thermal NO) is of the utmost important at temperatures above 1500 °C[39]. However, Tomeczek and Gradoń [42,43] claimed that, at suchtemperatures, the thermal mechanism is much faster than as describedby the Zeldovich reactions with the recommended rate constants. Theyexplained this by introducing the term “extended thermal mechanism”(see Eqs. (5)–(9), where M is the third body, e.g., N2, CO2 and others).In this “extended thermal mechanism” (including the reactions from theZeldovich mechanism: (5) and (6)), the role of nitrous oxide (N2O) wasunderscored in the formation of NOx. The role of nitrous oxide for thethermal NO build-up is crucial in temperatures below 1500 °C. Thus,due to the high temperature in the combustion chamber (i.e., higherthan 1300 °C), a high air excess coefficient or a high residence time forthe gases in the combustion chamber, these parameters can play animportant role in NOx conversion during the combustion of coal indomestic-scale boilers. This issue needs more focused investigation.

N2 + O ↔NO + N (5)

N + O2 ↔NO + O (6)

N2 O + M ↔N2 + O + M (7)

N2 O + O ↔NO + NO (8)

N2 O + O ↔N2 + O2 (9)

This study found that the addition of canola oil (in the case of lowsinterable coal) slightly decreases NOx emissions. The relative differ-ence between NOx emissions (calculated as (363 – 320)/363×100%)during the combustion of non-impregnated coal (NOx=363mg/mn

3)and canola oil-impregnated coal (NOx=320mg/mn

3) is less than 12%.However, assuming that the temperature plays an important role, theaddition of more calorific fuel (i.e., impregnation by oil) could increasethe temperature on the coal surface and, as a consequence, increaseNOx emissions. In fact, a slight decrease in NOx emissions was observedin this case. Slightly lower emissions of NOx were probably caused bythe existence of fuel-rich zones. CO emissions during the combustion ofcanola oil-impregnated coal (i.e., 85/15 and 70/30) was higher than forthe combustion of non-impregnated coal (see Table 3). A higher COconcentration is typical for the reburning zone due to the reactionCHi+O→ CO+H+… [44]. This is consistent with the observationsby Hrycko et al. [3] regarding lower NOx emissions for manual feedboilers, which were characterized by higher CO emissions compared toautomatic retort boilers. The nitrogen content in the coals tested byHrycko et al. [3] ranged between 0.95–1.28 wt% and 1.00–1.46 wt%for retort and traditional boilers, respectively. Thus, the lower NOxemissions were probably caused by the creation of local reburningzones when the oil film (on the coal surface) was burned. Local re-burning occurs when higher amounts of reducing compounds are pro-duced, as well as combusted fuel. These reducing agents affect the NOx

formation mechanism. Reducing compounds can be produced, for ex-ample, during the volatile release from biomass, which is co-combustedwith a coal. Local reburning effects were investigated by Thanapal et al.[45] during the co-combustion of coal and raw/torrefied mesquiteusing a 30 kWth drop tube reactor. This phenomenon needs more spe-cific investigations to explain the observed results. Nevertheless, theimpact of the aforementioned local reburning should also be taken intoaccount.

Emissions of PM can be also affected by process parameters and fuelproperties. Garcia-Maraver et al. [8] noticed that the type of fuel(biomass pellet of pine, cork, olive wood and olive pruning) has a moresignificant impact on PM emissions that boiler thermal load. However,

Fig. 7. Observation of a sample cross section after combustion in a flame. The numbers represent the different zones of fuel conversion: 1) unreacted coal, 2) beginning of coaltransformation, 3) appearance of incipient anisotropic effects, 4) char optical texture, 5) zone of char matrix degradation (oxidation).

J.A. Lasek et al. Fuel 215 (2018) 239–248

245

they concluded that, in some specific cases (e.g., agro-pellets), PMemissions were more sensitive to the boiler operating conditions.Moreover, they reported that values for total PM emissions varied from5 to 800mg/m3

n (values corrected to 10% of O2 in the flue gas), de-pending on the burner type and boiler configurations. Our study foundthat the combustion of impregnated coals (by liquid fuels) emits morePM (96–302mg/m3

n) than the reference base (i.e., 56 mg/m3n). None-

theless, this is still a lower value compared with those reported inaforementioned literature. According to PN-EN 303-5:2012, the per-missible emission of dust is 150mg/mn

3, 60mg/mn3 and 40mg/mn

3 forthird-, fourth- and fifth-class boilers, respectively.

3.2. Proposed simplified mechanism for coal adaptation

The laboratory-scale investigations of an adaptation mechanismwere performed via the combustion of coal samples (2 cm cubes) in aburner flame, while observations of their cross sections used a lightmicroscope and polarized light. An example of an analysed sample ispresented in Fig. 7. It should be explained that the presented zones weredivided according to the observed changes in coal structure (using alight microscope and polarized light) along its cross section. Zone 1 wasthe furthest away field from the flame front. In this zone, no significantchanges were observed; thus, it may be referred to as the unreactedcore. In Zone 2, the initial changes in coal transformation can be ob-served, i.e., the appearance of voids (pores), which appear as blackfields on a photo. It is known that, during the devolatilization stage, thereleasing volatile matter can become trapped in the plasticized coalparticles and form bubbles. The volatile matter diffuses directly out ofthe particle’s surface, as well as into bubbles. This causes the growth ofbubbles and particle swelling [46]. Thus, in Zone 2, an initial stage ofdevolatilization can be observed. Zone 3 is characterized by the ap-pearance of incipient anisotropic effects; thus, this zone represents thefurther transformation of coal into a char-like structure. Zone 4 presents

Table 5Microscopic observation of coal samples after combustion in a flame using polarized reflected light and a Lambda plate (magnification ×200).

Residence time in a flame Coal without oil impregnation Coal impregnated by oil

3min

6min

9min

Fig. 8. Simplified mechanism of HSC adaptation to stable combustion in automatic boi-lers.

J.A. Lasek et al. Fuel 215 (2018) 239–248

246

the appearance of a char optical texture with a visible pore structure.Some bigger spherical voids and smaller slender voids can be seen in aphoto. The observed differences between the voids’ size are consistentwith the mechanism of bubble growth during the swelling of cokingcoal. During the initial stage, proto-bubbles are created. Next, thebubbles grow and create bigger fields by coalescence [47]. Thus, theexistence of voids in the char structure is the result of volatile mattergeneration and transportation during coal pyrolysis [48]. Zone 5 is thefield that has direct contact with the flame front. Consequently, it is thezone of char oxidation.

The cubic samples of coal (fresh and after combustion) were em-bedded in acrylic resin, then cut across the plane of the temperaturegradient. Samples were polished using diamond and silica suspensions.The changes in the optical character of the samples, according to thetemperature gradient vector, were observed.

The results of these observations are presented in Table 5. Thephotos represent a cross section of the zone closest to the flame. Thesamples impregnated with oil had a more anisotropic structure, whichcan be detected by the presence of colourful and variously shaped do-mains in the reflected polarized light (right side of Table 5). A mosaicanisotropy of the optical texture is typical for samples after pyrolysis.Pyrolysis rearranges the carbon molecular structure; thus, larger ani-sotropic domains can be observed in the case of impregnated samples[49–51]. The following simplified mechanism of coal adaptation isproposed. The impregnation of coal by oil creates an oil film on thesurface of coal grains. The combustion of the oil film in the high-tem-perature zone of a boiler causes oxygen deficiency near the coal surface.High temperature and lack of oxygen are the necessary conditions forthe pyrolysis process, which occurs on the surface of the coal grains.Resolidation of a plastic mass occurs faster because of the higher tem-perature of the combusting oil, while grains are not able to agglomerateby fusing as fast as in the case of the combustion of coal without oil. Asthe pyrolysis continues, the coal loses its sinterability. The highertemperature caused by the oil combustion process on the surface of acoal grain resulted in more beneficial conditions for the ordering of thecarbon molecular structure. As a consequence, after the combustion ofthe oil film, the stable combustion of processed coal is possible. Thissimplified mechanism is presented in Fig. 8. The main stages of coalprocessing and combustion are presented.

The microscopic pictures presented in Table 5 show evident dif-ferences between the optical texture of the impregnated and non-im-pregnated samples. The samples that were not impregnated with oilshowed lower levels of molecular ordering. The optical texture of thesesamples was characterized by a lower intensity of anisotropy and asmaller anisotropic domain. The texture was dominated by isotropic,incipient anisotropic and fine mosaic textures. For the impregnatedsample, larger anisotropic domains and a higher intensity of anisotropycould be observed.

4. Conclusions

An adaptation method for stable combustion (in the retort boiler) ofHSC was proposed. The adaptation of this type of coal for combustion ina boiler was achieved by the oil impregnation of coal. The oil film wascombusted on the coal surface, creating the conditions for the pyrolysisprocess, which occurs on the surface of coal grains. As the pyrolysiscontinues, the coal loses its sinterability, such that the grains are notable to agglomerate during the subsequent stage of combustion.Moreover, the stable combustion of impregnated coal samples wastested and confirmed using a commercial-scale boiler (20 kW). Theadditional benefit of co-combustion is the increase in bio-originatedfuel fraction. This means that the part played by fossil fuel is replacedby biofuel. Nevertheless, higher emissions of CO (in the case of thecombustion of impregnated low sinterable coal by canola oil: 85/15 and70/30) and dust (all impregnated coal cases) were reported. Resolvingthis issue should still be pursued. The emission of NOx was slightly

decreased (in the case of the combustion of impregnated low sinterablecoal by canola oil: 85/15 and 70/30) and slightly increased (in the caseof the combustion of impregnated low sinterable coal by waste oil: 90/10), although the change was insignificant compared to the combustionof low sinterable coal without impregnation. The emission of NOx

during the combustion of impregnated HSC (comparable to fuel-N) wassignificantly higher than for the combustion of low sinterable coal,highlighting the significant role of temperature and other parameters(T-3 rule) in NOx conversion during the combustion of coal in domestic-scale boilers.

Acknowledgements

The results presented in this paper were obtained during the re-search projects entitled “Badania nad opracowaniem kompozycji nis-koemisyjnych paliw stałych dla kotłów retortowych“ (IChPW no.11.11.018) and “Utrzymanie potencjału badawczego Centrum BadańTechnologicznych” (IChPW no. 11.17.011), financed by the Ministry ofScience and Higher Education, Republic of Poland. We would like tothank Ms. Roksana Muzyka for her help during description ofExperimental section.

References

[1] De Paepe M, T'Joen C, Huisseune H, Van Belleghem M, Kessen V. Comparison ofdifferent testing methods for gas fired domestic boiler efficiency determination.Appl Therm Eng 2013;50:275–81.

[2] Shao L, Hou C, Geng C, Liu J, Hu Y, Wang J, et al. The oxidative potential of PM10from coal, briquettes and wood charcoal burnt in an experimental domestic stove.Atmos Environ 2016;127:372–81.

[3] Hrycko P, Lasek J, Matuszek K. Biomass gasification and Polish coal-fired boilers forprocess of reburning in small boilers. J Cent South Univ 2013;20:1623–30.

[4] Xu Y, Wang Y, Chen Y, Tian C, Feng Y, Li J, et al. Characterization of fine andcarbonaceous particles emissions from pelletized biomass-coal blends combustion:implications on residential crop residue utilization in China. Atmos Environ2016;141:312–9.

[5] González-González JF, Alkassir A, San José J, González J, Gómez-Landero A. Studyof combustion process of biodiesel/gasoil mixture in a domestic heating boiler of26.7 kW. Biomass Bioenergy 2014;60:178–88.

[6] Schönnenbeck C, Trouvé G, Valente M, Garra P, Brilhac JF. Combustion tests ofgrape marc in a multi-fuel domestic boiler. Fuel 2016;180:324–31.

[7] Leyssens G, Trouvé G, Caplain I, Schönnenbeck C, Cazier F. Energetic performancesand environmental impact of the combustion of cardboard/sawdust in a domesticboiler. Fuel 2014;122:21–7.

[8] Garcia-Maraver A, Zamorano M, Fernandes U, Rabaçal M, Costa M. Relationshipbetween fuel quality and gaseous and particulate matter emissions in a domesticpellet-fired boiler. Fuel 2014;119:141–52.

[9] Limousy L, Jeguirim M, Dutournié P, Kraiem N, Lajili M, Said R. Gaseous productsand particulate matter emissions of biomass residential boiler fired with spentcoffee grounds pellets. Fuel 2013;107:323–9.

[10] Horak J, Kubonova L, Krpec K, Hopan F, Kubesa P, Motyka O, et al. PAH emissionsfrom old and new types of domestic hot water boilers. Environ Pollut2017;225:31–9.

[11] Energy consumption in households in 2015. Warsaw: Central Statistical Office(Główny Urząd Statystyczny); 2017.

[12] Xiaohua W, Kunquan L, Hua L, Di B, Jingru L. Research on China’s rural householdenergy consumption – household investigation of typical counties in 8 economiczones. Renewable Sustainable Energy Rev 2017;68:28–32.

[13] Kerimray A, Rojas-Solórzano L, Amouei Torkmahalleh M, Hopke PK, GallachóirBPÓ. Coal use for residential heating: patterns, health implications and lessonslearned. Energy Sustainable Dev 2017;40:19–30.

[14] Ryfa A, Buczynski R, Chabinski M, Szlek A, Bialecki RA. Decoupled numerical si-mulation of a solid fuel fired retort boiler. Appl Therm Eng 2014;73:794–804.

[15] Schmidl C, Luisser M, Padouvas E, Lasselsberger L, Rzaca M, Ramirez-Santa Cruz C,et al. Particulate and gaseous emissions from manually and automatically firedsmall scale combustion systems. Atmos Environ 2011;45:7443–54.

[16] Dias J, Costa M, Azevedo JLT. Test of a small domestic boiler using different pellets.Biomass Bioenergy 2004;27:531–9.

[17] 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;35:80–9.

[18] Buczyński R, Weber R, Szlek A, Nosek R. Time-dependent combustion of solid fuelsin a fixed-bed: measurements and mathematical modeling. Energy Fuels2012;26:4767–74.

[19] Buczyński R, Weber R, Szlęk A. Innovative design solutions for small-scale domesticboilers: combustion improvements using a CFD-based mathematical model. JEnergy Inst 2015;88:53–63.

[20] Rabaçal M, Fernandes U, Costa M. Combustion and emission characteristics of a

J.A. Lasek et al. Fuel 215 (2018) 239–248

247

domestic boiler fired with pellets of pine, industrial wood wastes and peach stones.Renewable Energy 2013;51:220–6.

[21] Šyc M, Horák J, Hopan F, Krpec K, Tomšej T, Ocelka T, et al. Effect of fuels anddomestic heating appliance types on emission factors of selected organic pollutants.Environ Sci Technol 2011;45:9427–34.

[22] Fernandes U, Costa M. Formation of fine particulate matter in a domestic pellet-fired boiler. Energy Fuels 2013;27:1081–92.

[23] Vega-Nieva DJ, Ortiz Torres L, Míguez JL, Morán JC. Measuring and predicting theslagging of woody and herbaceous Mediterranean biomass fuels on a domesticpellet boiler. Energy Fuels 2016. just accepted.

[24] Tsuchiya Y, Yoshida T. Pelletization of brown coal and rice bran in Indonesia:characteristics of the mixture pellets including safety during transportation. FuelProcess Technol 2017;156:68–71.

[25] Zeng T, Weller N, Pollex A, Lenz V. Blended biomass pellets as fuel for small scalecombustion appliances: influence on gaseous and total particulate matter emissionsand applicability of fuel indices. Fuel 2016;184:689–700.

[26] Nunes LJR, Matias JCO, Catalão JPS. Mixed biomass pellets for thermal energyproduction: a review of combustion models. Appl Energy 2014;127:135–40.

[27] Porteiro J, Collazo J, Patiño D, Granada E, Moran Gonzalez JC, Míguez JL.Numerical modeling of a biomass pellet domestic boiler. Energy Fuels2009;23:1067–75.

[28] Topolnicka T, Zuwała J, Hrycko P, Muzyka R. Crude glycerol as a bioliquid forrenewable energy generation in power sector [In Polish: Gliceryna techniczna jakonośnik energii odnawialnej dla energetyki zawodowej]. Przemysł Chemiczny2013;92:111–9.

[29] Bhimani S, Alvarado JL, Annamalai K, Marsh C. Emission characteristics of me-thanol-in-canola oil emulsions in a combustion chamber. Fuel 2013;113:97–106.

[30] Lasek JA, Kopczyński M, Janusz M, Iluk A, Zuwała J. Combustion properties oftorrefied biomass obtained from flue gas-enhanced reactor. Energy2017;119:362–8.

[31] Zagoruiko AN, Belyi AS, Smolikov MD, Noskov AS. Unsteady-state kinetic simula-tion of naphtha reforming and coke combustion processes in the fixed and movingcatalyst beds. Catal Today 2014;220–222:168–77.

[32] Kelling R, Kolios G, Tellaeche C, Wegerle U, Zahn VM, Seidel-Morgenstern A.Development of a control concept for catalyst regeneration by coke combustion.Chem Eng Sci 2012;83:138–48.

[33] Bazelatto Zanoni MA, Massard H, Ferreira Martins M. Formulating and optimizing acombustion pathways for oil shale and its semi-coke. Combust Flame2012;159:3224–34.

[34] Fadaei H, Sennoune M, Salvador S, Lapene A, Debenest G. Modelling of non-con-solidated oil shale semi-coke forward combustion: influence of carbon and calcium

carbonate contents. Fuel 2012;95:197–205.[35] Wilk R, Zieliǹski Z, Kòlarz W. Determination of overall burning rate coefficient of

coal-oil droplet mixtures. Fuel 1990;69:110–6.[36] Liu H, Chaney J, Li J, Sun C. Control of NOx emissions of a domestic/small-scale

biomass pellet boiler by air staging. Fuel 2013;103:792–8.[37] Lajili M, Jeguirim M, Kraiem N, Limousy L. Performance of a household boiler fed

with agropellets blended from olive mill solid waste and pine sawdust. Fuel2015;153:431–6.

[38] Verma VK, Bram S, Gauthier G, De Ruyck J. Performance of a domestic pellet boileras a function of operational loads: Part-2. Biomass Bioenergy 2011;35:272–9.

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

[40] Verma VK, Bram S, Delattin F, Laha P, Vandendael I, Hubin A, et al. Agro-pellets fordomestic heating boilers: standard laboratory and real life performance. ApplEnergy 2012;90:17–23.

[41] Carvalho L, Wopienka E, Pointner C, Lundgren J, Verma VK, Haslinger W, et al.Performance of a pellet boiler fired with agricultural fuels. Appl Energy2013;104:286–96.

[42] Tomeczek J, GradoŃ B. The role of nitrous oxide in the mechanism of thermal nitricoxide formation within flame temperature range. Combust Sci Technol1997;125:159–80.

[43] Tomeczek J, Gradoń B. The role of N2O and NNH in the formation of NO via HCN inhydrocarbon flames. Combust Flame 2003;133:311–22.

[44] Smoot LD, Hill SC, Xu H. NOx control through reburning1. Prog Energy Combust Sci1998;24:385–408.

[45] Thanapal SS, Annamalai K, Ansley RJ, Ranjan D. Co-firing carbon dioxide-torrefiedwoody biomass with coal on emission characteristics. Biomass Convers Biorefin2016;6:91–104.

[46] Yu J, Lucas JA, Wall TF. Formation of the structure of chars during devolatilizationof pulverized coal and its thermoproperties: a review. Prog Energy Combust Sci2007;33:135–70.

[47] Jenkins DR, Mahoney MR. A mechanistic model for the softening of coking coal andits use for predicting the dilatation of blends. Fuel 2015;153:585–94.

[48] Yu J, Lucas J, Wall T, Liu G, Sheng C. Modeling the development of char structureduring the rapid heating of pulverized coal. Combust Flame 2004;136:519–32.

[49] Patrick JW, Reynolds MJ, Shaw FH. Development of optical anisotropy in vitrainsduring carbonization. Fuel 1973;52:198–204.

[50] Patrick JW, Walker A. The relation between vitrain properties and the developmentof optical anisotropy during carbonization. Fuel 1991;70:465–70.

[51] Patrick JW, Reynolds MJ, Walker A. Coal carbonization and coke anisotropy: somecomments. Fuel 1983;62:131–2.

J.A. Lasek et al. Fuel 215 (2018) 239–248

248