Impregnation of olive mill wastewater on dry biomasses: Impact on chemical properties and combustion performances

Energy 78 (2014) 479e489

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Impregnation of olive mill wastewater on dry biomasses: Impact on chemical properties and combustion performances Nesrine Kraiem a, b, c, Mejdi Jeguirim a, *, Lionel Limousy a, Marzouk Lajili b, Sophie Dorge c, Laure Michelin a, Rachid Said b Institut de Sciences des Mat
eriaux de Mulhouse, 15 rue Jean Starcky, 68057 Mulhouse, France UR Etude des Milieux Ionis
es et R
eactifs, IPEIM, Avenue Ibn El Jazzar, Monastir 5019, Tunisia c Laboratoire Gestion des Risques, Environnement 3 bis, rue Alfred Werner, 68093 Mulhouse, France a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 April 2014 Received in revised form 4 October 2014 Accepted 12 October 2014 Available online 6 November 2014

Mediterranean countries generate large amounts of olive oil byproducts mainly OMWW (olive mill wastewater) and EOSW (exhausted olive solid waste). Although solid residues have various valorization strategies, there is no economically viable solution for the OMWW disposal. This study aims to recover the OMWW organic contents through solid biofuels production. Hence sawdust and EOSW were used for the OMWW impregnation. The potential of the obtained samples, namely: IS (impregnated sawdust) and IEOSW (impregnated exhausted olive solid waste) were evaluated. Therefore, the physicochemical characterizations and thermogravimetric analyses of the samples were first performed. Secondly, the samples densification into pellets and their combustion in a domestic combustor were carried out. Combustion efficiencies, gaseous and PM (particulate matter) emissions as well as ash contents were evaluated. The analysis finding shows that addition of OMWW leads to an increase of energy content through the heating values increase. An increase of the impregnated samples reactivity was observed and assigned to the potassium catalytic effect. Combustion performances show that the OMWW addition has not a negative effect on their firing quality. Moreover, a beneficial effect on the pollutant emissions is observed with IEOSW pellets. The developed strategy constitutes a promising issue for the OMWW disposal and recovery. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Olive mill wastewater Impregnation Characterization Energy recovery Combustion tests Gaseous and particulate emissions

1. Introduction In a context of sustainable development, the reduction of energy costs and the use of renewable energies potential based on processes that maximize energy efficiency and protect the environment are key challenges for industrial companies. With the depletion of fossil fuel resources, countries must move towards renewable energies for the development of their economy. In this framework, agro-industrial companies are prompted to reduce and/or to valorize the generated wastes from their activities. In this way, olive oil extraction industries, representing a significant economic and social activity in the Mediterranean countries, engendered two by-products: a solid residue and an aqueous effluent

* Corresponding author. Tel.: þ33 3 89608661. E-mail address: [email protected] (M. Jeguirim). http://dx.doi.org/10.1016/j.energy.2014.10.035 0360-5442/© 2014 Elsevier Ltd. All rights reserved.

namely, OMSW (olive mill solid waste) and OMWW (olive mill wastewater), respectively [1]. The treatment of OMWW represents a serious ecological problem due to its high degree of organic pollution with a COD/ BOD ratio (chemical oxygen demand/biological oxygen demand) evaluated to be between 2.5 and 5, its pH slightly acid [2] and its high content of recalcitrant compounds such as lignin and tannin [3]. Furthermore, OMWW contains phenolic compounds and long-chain fatty acid responsible for the phytotoxic and antibacterial effect. Tunisia like the Mediterranean countries produces large quantities of olive by-products estimated at 2009 by the Tunisian national agency for waste management, namely
chets) to be annuANGED (Agence Nationale de Gestion des De ally: 600,000e1,200,000 tons of OMWW, 435,000e800,000 tons of OMSW and 120,000e170,000 tons of sludge [4]. Hence, Tunisia needs to identify an environmentally and economically viable solution for the generated waste disposal. Several OMWW disposal scenarios and methods have been studied in the literature based generally on biological or chemical-physical

480

N. Kraiem et al. / Energy 78 (2014) 479e489

treatments, membrane filtration and evaporation. Few attempts have been tested to recover the OMWW energetic potential, but there is no economically issue for OMWW due to high moisture content [1,5]. The common disposal ways were soil spreading and evaporation ponds. Although, the latter technique is the most popular, the residual oil layer floating in the pond surface prevents water evaporation. The treatment of the solid residue (OMSW) has been well investigated in literature [1,6]. It has been traditionally used as animal feed but nowadays it undergoes a second oil extraction by chemical processes in seed-oil factories in order to extract its residual oil, which leads to generation of EOMSW (exhausted olive mill solid waste). The thermal conversion is the main recovery process for EOMSW because of its interesting energy content with low heating value around 18 MJ kg1 [7]. Separate treatments of OMWW and EOMSW have been studied by several authors [5e9], but nowadays fully combined treatment is required. Such strategy could develop a green and low cost energy in order to serve mills and seed-oil factories through small sized plants [2]. Addition of OMWW to suitable proportions of EOMSW following by biomass combustion is a promising solution which can avoid the factories to pay disposal costs [2]. Recently, Chouchene and Jeguirim established a combined method for the treatment of OMWW [10e12]. This method consisted on the impregnation of the OMWW on low cost adsorbents such as sawdust or OMSW. The impregnated samples were therefore thermally oxidized in laboratory furnace. The authors showed that the addition of OMWW to both adsorbents had not negative effect on gaseous emissions [11,12]. However, the CO (carbon monoxide) and VOC (volatile organic compounds) emissions from all the tested samples were higher due to the absence of secondary air injection in their laboratory reactor [11]. Although the obtained promising results, the validation of the OMWW treatment strategy required the investigation of the impregnated samples behavior during combustion tests in a domestic boiler. In recent times, several researchers have examined the combustion of agro-industrial and agriculture residues in domestic boilers [13e17]. These biofuels were firstly densified and pelletized in order to obtain adapted fuels for the different boilers. Also, pelletization increases the biomass energy density and decreases the moisture content leading to an increase of combustion efficiency as well as a reduction of smoke during combustion [18]. Several researchers have produced pellets by blending the agroindustrial residues with wood residues. This step allows obtaining an agropellets with good quality that could be used directly in wood domestic boilers. In fact, the combustion performances as well as the gaseous emissions obtained during combustion tests could reach the European standards. This investigation aims to validate the strategy of the OMWW treatment through its impregnation on different biomasses and the direct combustion of the impregnated samples for small-scale heat generation. Hence, the impregnation of OMWW was performed on sawdust and EOMSW for 5/1 mass fraction ratio. EOMSW is preferred to OMSW due its availability in seed-oil factories as well as its higher adsorption efficiency since the residual oil was removed. In order to reach this research purpose, firstly, characterization of the impregnated samples was performed using various analytical techniques as well through thermogravimetric analysis. Secondly, pellets from the different samples were produced and characterized according to the French and European standards. Finally, combustion tests for the different pellets were performed in a residential pellets boiler to compare their combustion efficiencies as well as gas and particle emissions.

2. Materials and methods 2.1. Samples preparation OMWW and EOMSW used in this study were collected from olive mill (three-phase centrifugal olive mill) at the seed oil factory Zouila located in Mahdia, Tunisia. Sawdust was provided from sawmill located in Sayada, Tunisia. During impregnation tests, 20 kg of EOMSW or sawdust with 10% of moisture (in wet basis, wb) were slowly added to 100 kg of OMWW with 89% of moisture (wb) in a specific barrel. The impregnated samples were mixed regularly. In this specific investigation, in order to reduce the initial water content in the mixture which was initially 76% (wb) and therefore to accelerate the drying process, the barrel was heated, from the underside, using hot ashes provided by a combustor in the seed oil factory. The upper face of the barrel was exposed to ambient air. Currently, solar drying of the impregnated samples is examined. At the end of drying, different samples were taken from 10 places for each mixture at time intervals of 2 h. Therefore, a homogeneous IEOSW (impregnated exhaust olive solid waste) and IS (impregnated sawdust) were obtained with less than 15% of moisture. This level 15% was chosen in order to avoid the fermentation of the different preparations. Samples of ISW and EOMSW are analyzed in the following in their raw state. During the analysis, all samples are placed in an isothermal container to maintain their properties and moisture content.

2.2. Samples characterization 2.2.1. Proximate, ultimate analyses and energetic contents The optimization of an operating plant design for biomass combustion involves the knowledge of the fuel composition as well as its related energy properties. Hence, elemental compositions of the different prepared samples were performed by CHONS-NA 2100 protein CE instrument analyzer (Carbon, Hydrogen, Oxygen, Nitrogen, Sulfur). Proximate analysis was obtained using different techniques. Therefore, the water content (M) is measured by weighting samples before and after drying at 105  C (about 1 g, wb) in an oven until obtaining a constant mass according to the EN 14774-1 standard. The percentage of moisture content is calculated from the average of three measures per sample. Bulk density is carried out according the EN 15103 standard. The ash content is evaluated according to two norms in order to evaluate differences between them. Ash value is obtained after sample combustion (about 1 g, wb) in a muffle furnace during 3 h at two temperatures 550  C and 815  C according to EN 14775 and DIN 51719 standards respectively. The VM (volatile matter) content was obtained using the TG procedure analysis. This method consists of heating the samples under nitrogen flow at a heating rate of 10  C min1 from 20  C to 110  C and maintaining at this temperature during 10 min to remove the moisture. Temperature is then increased at 20  C min1 to 900  C and kept for 10 min to obtain the weight loss corresponding to the volatile matter content. The fixed carbon content is obtained by difference. The energy contents of the different samples were obtained using a calorimetric bomb IKA-C200 by determining the HHV (high heating values). HHV were obtained after combustion of a sample (about 0.6 g, wb) under a pure oxygen atmosphere at 35 bars according to EN 14961-1 specification. The LHV (low heating values) were calculated using the relationship between HHV and LHV given by:

LHV ¼ HHV  hgð9H þ MÞ=100

(1)

N. Kraiem et al. / Energy 78 (2014) 479e489

where: H and M are the hydrogen percentage and the moisture percentage (mass basis) of the received tested fuel respectively. Here, hg is the latent heat of steam in the same units as HHV and LHV. The energy density is obtained from the multiplication of the low heating value by the bulk density. 2.2.2. Thermogravimetric analysis TGA (thermogravimetric analyses) were carried out using a METTLER TOLEDO thermobalance (Mettler-Toledo SAS, Viroflay, France). Experiments were performed for a mass sample about 10 mg under inert and oxidative atmospheres with a gas flow rate of 12 NL h1 at a heating rate of 5  C min1 from room temperature to 900  C. 2.3. Pellets production The densification of the different impregnated sample, sawdust and EOSW (exhausted olive solid waste) was carried out using a pelletizer KAHL 15/75 type (Amandus Kahl GmbH & Co, Reinbek, Germany) containing a die diameter of 6 mm and a length of 30 mm. The specifications of the used pelletizer are: Die diameter (mm): 175, Diameter/length of roller (mm): 130/29, Number of rollers: 2, Control motor (kW/min1): 3, Roller speed (m/s): 0.5e0.8. The capacity of the pelletizer depend on its properties (frequency, temperature, ...) and sample characteristics (composition, moisture content...) and controlled manually. Generally, it is about 2e3 kg/h. Diameters and lengths of pellets produced are measured with a numeric caliper. Unit density is calculated from pellets mass and volume. The pellet energy density is obtained from the multiplication of the low heating value by the unit density.

from flue gases through hot water heat exchanger (32 kW). The heat exchanger and flue gas temperatures were recorded as well as the hydraulic circuit water flow. 2.4.2. Measurement of gaseous and particles emissions During combustion tests, a portable analyzer (TESTO 350XL/ TESTO454) measures the concentrations of the following gases (on a dry basis): CO2, CO, O2, VOC (volatile organic compounds), NO (nitrogen oxide), NO2 (nitrogen dioxide). Gaseous measurements were realized according to the NF EN 304 standard. In this present investigation, particles generated during the experimental tests were collected into 12 size fractions from 29 nm to 10 mm using an ELPI (electrical low pressure impactor) analyzer. This apparatus was described in details in previous investigations [19e21]. The obtained results are presented in number and mass concentrations as function of size distribution. 2.4.3. Performance measurements The performance measurements are evaluated through the determination of the combustion and boiler efficiencies. The combustion efficiency is calculated using the indirect method recommended by the NF EN 12809 standard. This method required the determination of thermal (qa kJ kg1), and unburned gaseous (qb kJ kg1) and unburned carbon (in the ash) (qr kJ/kg) heat losses). In the following, the used formulas for calculation are presented.

 
Cpmd  ðC  Cr Þ qa ¼ 100  Tg  Ta  0:536*ðCO þ COr Þ   ð9H þ MÞ þ CpmH2 O  1:244  100

2.4. Experimental test setup 2.4.1. Combustion equipment A commercially residential pellets boiler (Pellematic PES12 e € PVB 2000) supplied by Okofen (Barberaz, France) was used to perform combustion tests. This boiler was already tested in previous investigations [13,14]. A schematic diagram of the experimental setup is shown in Fig. 1. This boiler has a nominal power output of 12 kW and was equipped with recycling combustion to improve the combustion efficiency. The boiler is placed on a balance supplied by Sartorius to determine the fuel consumption. Combustion gases were extracted with a fan maintained constant during combustion test to have a constant draught. Heat of combustion is collected

481

(2)

qb ¼

12644  CO  ðC  Cr Þ 0:536  ðCO þ CO2 Þ  100

(3)

qr ¼

33500  Cr 100

(4)

where: Tg and Ta are the temperatures of the exhaust gases and primary air, respectively (K). Cpmd and CpmH2O are the dry flue gases and water vapor specific heats (kJ K1 m3). CO and CO2 are concentration in the dry flue gases (% of volume). Cr is the unburned carbon content passing through ash (% of mass). C is the carbon content in the pellet (% of mass). H is the hydrogen content in the pellet (% of mass). M is the moisture content in the pellet fuel (% of mass). The combustion efficiency (q %) is determined:

 q ¼ 100 

qa þ qb þ qr LHV

  100

(5)

The BE % (boiler efficiency) was calculated using the direct method according to the NF EN 303-5 standard from the heat power gain of water in the exchanger (PN kW) and the heat power content in the fuel (PC kW) ratio as described by Eqs. (6)e(8).

Fig. 1. Schema of the experimental combustion setup.

PN ¼ Qw $CPw $ðTout  Tin Þ

(6)

PC ¼ Qcomb $LHV

(7)

482

N. Kraiem et al. / Energy 78 (2014) 479e489

BE ¼

PN  100 PC

(8)

where: Qw is the mass flow of water in the heat exchanger (kg s1). CPw is calorific capacity of liquid water (kJ K1 kg1). Tin and Tout are the inlet and outlet temperatures of water in the exchanger (K) respectively. Qcomb is the pellets mass flow (kg s1).

Table 2 Proximate analysis and energy contents of raw samples. Sample characteristic

Sawdust

Moisturewb (%) Bulk densitywb (kg m3) LHVwb (MJ kg1) Fixed Carbon (%) Volatile matter (%) Ashdb (%)

9.8 103 16.4 14.5 75.2 0.6b 0.5c

a b c

2.4.4. Ash characterization Bottom ashes obtained during the combustion tests were characterized by XRF (X ray fluorescence) using a Magix from PHILIPS spectrophotometer apparatus. These residual ashes were previously shredded. Disks were manufactured under a mass of two tons during 2 min. 3. Results and discussion 3.1. Raw samples characteristics Elemental compositions as well as the “H/C” and “O/C” ratios of the different samples are listed in Table 1. The analysis finding shows that elemental compositions of the different samples are in the typical composition of biomass reported in literature [7,14,20,21]. The low standard deviations for the different samples confirm their homogeneity. Comparison of the different samples shows that the addition of OMWW leads to an increase to carbon and hydrogen contents. Such behavior is attributed to the high polyphenol contents in OMWW. The nitrogen content increases also by 1% wt with the addition of OMWW. Table 2 shows the proximate analysis and energy contents of raw samples. The main characteristics are within the typical range of agro-industrial residues [1,7]. Addition of OMWW leads to an increase of energy content through the LHV values. Such behavior is predictable since the impregnated samples have the highest carbon and hydrogen contents. In contrast, the OMWW addition leads to an increase of ash contents. This result is attributed to the high mineral contents in OMWW which has mainly from 9 to 13% ash content [10,11]. For impregnated samples (IEOSW, IS) ash contents value determined at 815  C are lower than those determined at 550  C. This result may be attributed to the evaporation of some minerals presented in OMWW. Obernberger concluded that the ash content of biomass fuels should be determined at 550  C [22]. Furthermore, Table 2 shows that the fixed carbon of samples are in the same range as other biomasses like grape marc (25%), olive husk (19%) and pine sawdust (17%) [13e15]. Adding OMWW has no significant trend on volatile matter and fixed carbon contents. 3.2. Thermal degradation Figs. 2 and 3 show the residual mass percentage (X, TG) and its time derivative (dX/dt, DTG) evolution obtained during pyrolysis

Table 1 Ultimate analysis of raw samples (dry basis). Sample

%C

Sawdust IS EOMSW IEOSW

51.3 57.0 51.3 54.4

%H ± ± ± ±

0.2 1.1 0.9 0.8

6.4 7.3 6.5 6.8

%N ± ± ± ±

0.3 0.1 0.2 0.1

0.2 1.1 0.9 1.9

0.1 0.1 0.6 0.1

0.5 3 0.1 0.1 1.0 0.1 0.1

IS 9.6 183 18.0 17.6 68.5 4b 4c

EOMSW ± ± ± ± ± ± ±

0.1 2 0.4 0.1 0.3 0.1 0.1

10 529a 16.9 25.5 61.5 3b 3c

± ± ± ± ± ± ±

0.4 10 0.3 0.2 0.2 0.1 0.1

IEOSW 15 551a 17.5 17.5 60.5 7b 5c

± ± ± ± ± ± ±

0.5 13 0.7 0.9 0.1 0.2 0.1

Bulk density. Ash at 550  C. Ash at 815  C.

and oxidation of the different tested samples, respectively. In literature, the thermal degradations of biomass under inert and oxidative atmosphere were discussed in details [12,24e26]. Pyrolysis curves have the typical biomass degradation shapes with three regions corresponding to the following steps: 1) biomass drying, 2) devolatilization step (named active pyrolysis) and 3) char formation (named passive pyrolysis). During the devolatilization step, the degradation profiles show different behaviors. Hence, EOSW and IEOSW samples have two DTG peaks correlative to the degradation of hemicellulose and cellulose with closer degradation rate. Such behavior is predictable since olive solid waste contains a similar percentage of hemicellulose (21.5%) and cellulose (24.3%). In contrast, IS and sawdust samples have a single DTG peak corresponding to cellulose decomposition since this latter is the main wood component (40%). From the DTG curves, it is important to note that the degradation of impregnated samples (IS, IEOSW) occurred earlier. Such behavior can be assigned to the potassium brought by OMWW. In fact, previous investigations showed that adding potassium catalyzes the thermal decomposition of biomass [27]. TG and DTG curves for the tested samples under oxidative atmosphere follow also the usual profiles of the biomass oxidation with three steps: 1) drying biomass, 2) devolatilization step and 3) oxidation of carbonaceous residues. The thermal decomposition of impregnated samples under oxidative condition started earlier as under inert conditions. Such behavior is also attributed to the potassium initially present in OMWW and impregnated on the different biomasses (see minerals composition in Table 7). The thermogravimetric characteristics of the different samples namely: temperature range degradation, mass loss X, peak temperature Tpeak, peak rate R and reactivity RM under inert and oxidative atmospheres are shown in Tables 3 and 4 Reactivity values are in the same range as those of some other biomasses. El May et al. found 0.37% s1  C1 and 0.55% s1  C1 for date stones under inert and oxidative atmospheres, respectively [25]. Munir et al. have found 0.45e0.49 % s1  C1 (N2) and 0.94e1.14 % s1  C1 (Air) for different types of sugarcane bagasse [24]. Comparison of the different samples shows that the addition of OMWW leads to the decrease of the peak temperature of the thermal degradation under inert and oxidative atmospheres. Moreover, a pronounced increase of the reactivity is observed during the addition of OMWW to sawdust. Such behavior is attributed to the catalytic effect of potassium. 3.3. Pellets properties

%O ± ± ± ±

± ± ± ± ± ± ±

41.5 31.1 37.9 30.8

± ± ± ±

0.5 1.0 1.6 0.9

H:C ratio

O:C ratio

0.124 0.128 0.127 0.125

0.809 0.546 0.739 0.566

Densification of samples by pelletizing increases energy density of the different tested fuels and therefore optimizes transportation costs and storage capacity. Table 5 summarizes the conditions used for pellets preparation. Different tests were performed to optimize the granulation process, namely frequency of pelletizer, the die

N. Kraiem et al. / Energy 78 (2014) 479e489

483

Fig. 2. TG and DTG curves of the different samples under inert atmosphere.

Fig. 3. TG and DTG curves of the different samples under oxidative atmosphere.

temperature and moisture contents. The selected parameters are shown in Table 5. The characteristics of the prepared pellets are shown in Table 6. The obtained results are compared with the different quality requirements for pellets fuel such as EN 14961-6 (European standard for solid biomass made from sets and mixtures), AQHP (agro quality high performance) (French requirement of Agro Quality Hight Performance) and AQI (French requirement of Industrial Agro Quality) [15,28]. Table 6 shows that IEOSW pellets could meet the French AQHP and AQI standards. However, a particular attention should be paid to the higher ash content, which may lead to slag and deposit formation in the furnace during a long operation time. The IS pellets have suitable characteristics. However, the low bulk density prevents them to reach the different French and European standards. In addition, the minerals composition of the different prepared pellets was performed using inductively coupled plasma atomic emission spectroscopy (ICP-AES). The mineral compositions of the four pellets are shown in Table 7. Table 7 shows that the addition of OMWW to dry biomasses leads to a higher increase of the K content as well as a significant increase to Na, Fe and Ca contents.

3.4. Combustion tests 3.4.1. Thermal performance The main results of combustion tests of the different pellets are summarized in Table 8. All the tested pellets have good boiler and combustion efficiencies comparing to values found in literature. The comparison between the pellets produced from raw biomass with those produced from impregnated biomasses shows that the addition of olive mill wastewater has not a negative effect on the combustion quality. The different pellets have close flow rates and excess air. The air excess of impregnated biomasses is slightly lower than reference samples, which can be attributed to their higher ash content (especially for K and Fe, see Table 3), which induces a higher reactivity of the pellets and then a lower oxygen concentration in the exhaust gas. However, the values of air excess (l) are found in the same order of magnitude as those found in similar conditions of this study, which were estimated between 1.6 and 2.3 [29] and 1.4e3.9 (NF EN 303-5). The boiler efficiencies for the tested

Table 4 Thermogravimetric characteristics under oxidative atmosphere.



Table 3 Thermogravimetric characteristics under inert atmosphere.

T ( C) Tpeak ( C) R (% s1) RM*103 (% s1  C1) Char (%) at 900  C

Sawdust

IS

EOMSW

IEOSW

130e531 332 0.097 0.29 14.9

120e570 288 0.074 0.26 21.9

137e570 212e275 0.041e0.05 0.39 28.9

128e518 206e274e312 0.044e0.051e0.02 0.46 25.0

Devolatilization T ( C) Tpeak 1 ( C) step R (% s1) Combustion T ( C) step Tpeak ( C) R (%.s1) RM*103 (% s1  C1) Ash (%) at 900  C

Sawdust

IS

EOMSW

IEOSW

134e325 284 0.16 325e440 415 0.13 0.52

118e341 241 0.17 341e500 409 0.11 0.97

134e314 219e269 0.054e0.035 314e438 375 0.076 0.58

112e331 208e260 0.042e0.042 331e458 380 0.079 0.57

0.8

4.3

3.2

5.6

484

N. Kraiem et al. / Energy 78 (2014) 479e489

Table 5 Operating pelletization conditions.

Table 7 Pellets mineral contents.

Sample Pelletizer performance Frequency of pelletizer Temperature Moisture of raw material Moisture of final pellets

Sawdust 50 Hz 70  C 16% 13%

Ultimate analysis (g/kg, wb) IS 50 Hz 60  C 15% 9%

EOMSW 50 Hz 70  C 16% 12%

IEOSW 52 Hz 60  C 13% 7%

pellets are lower than that estimated by the manufacturer (92.5%) as the boiler is designed principally for labeled wood pellets DINþ. In fact, boiler parameters are preset by the manufacturer and it was not possible to modify them in order to adjust and adapt the boiler to suit other fuels than wood pellets. Combustion efficiencies of the different pellets are higher than the minimum value specified by the European Standard EN 303-5 (76%). 3.4.2. Gaseous emissions analysis Fig. 4 shows the evolution of gaseous emissions versus time during the combustion tests of the different pellets, namely O2 (oxygen), CO2 (carbon dioxide), CO (carbon monoxide), NOx (nitrogen oxides) and VOC (volatile organic compounds). The concentrations of O2 and CO2 are expressed in %vol while CO, NOx and VOC concentrations are expressed in ppmv. The comparison of the different pellets shows that sawdust (Fig. 4a) pellets have the more stable emissions. Such a behavior is predictable since the used boiler was adapted only for wood pellets. The impregnation of OMWW on sawdust leads to the occurrence of CO and VOC emission fluctuations. Such behavior may be attributed to the accumulation of ash in the combustion plate, which may obstruct the arrival of primary air-flow and therefore instantaneous higher emissions of CO and VOC were occurred. These CO and VOC emissions could not be completely oxidized by the secondary airflow. The fluctuation of gaseous emissions is more pronounced during the EOSW and IEOSW combustion tests. Such behavior is may be attributed also to a higher ash content. Since the emissions curves show a fluctuating evolution, the mean values of the different emitted gas are calculated and presented in Table 9. Moreover, emissions are corrected at 10% and 13% of oxygen in order to compare the results obtained with the different pellets with results coming from the literature as shown in Table 9. Table 9 shows different behavior concerning the impact of OMWW impregnation on the gaseous emissions. Hence, the addition of OMWW to sawdust leads to an increase of CO, VOC and NOx emissions. In contrast, the gaseous emissions are lower for IEOSW pellets comparing to EOSW pellets. Such behavior was already observed by Chouchene et al., during oxidation tests of OMWW/ OSW blends performed in a laboratory reactor [11]. Such behavior may be attributed to the well-known catalytic effect of potassium.

Na K P Mn Fe Mg Si Ca Al

Samples Sawdust

IS

EOMSW

IEOSW

0.009 1.224 0.056 0.042 0.043 0.098 0.039 1.215 0.030

4.761 22.431 0.819 0.078 3.553 0.743 0.644 3.602 0.384

3.201 15.127 0.541 0.007 0.173 0.485 0.366 4.665 0.199

7.627 32.125 0.985 0.025 4.287 0.922 0.909 6.193 0.571

This behavior may be depended on the dry biomass since it was not observed in the case of sawdust. Fig. 5 shows a comparison of boiler power, combustion efficiency and gaseous emissions of the different tested pellets with European requirement and other biomass pellets examined in the literature. Several investigations have examined recently the combustion of various biomass pellets in low and medium boilers power [13e17,30]. The obtained values in Table 7 and Fig. 6 show that the combustion efficiency meets European standard (76%) and are in the same range as other biomasses and agropellets as Brassica and poplar [31], Miscanthus and Swithgrass [32]. The obtained CO emissions were 459, 722 and 743 mg Nm3 at 13% O2 for IEOSW, EOSW and IS pellets, respectively. These values meet European standard (3000 mg Nm3 at 13% of O2). The values calculated at 10% O2 are lower or closer than those obtained in
lez et al. found CO literature for various pellets. In fact, Gonza emissions of 1822 ppm at 13% of O2 (eq. 3340 mg Nm3 at 10% O2) for tomato pomace and 1680 ppm at 13% of O2 for cardoon (eq. 3080 mg Nm3 at 10% O2) burned in a mural boiler of a 12 kW of nominal power [29]. Limousy et al. found 353 ppm (606 mg Nm3 at 10% of O2) for pellets corresponding to a blend from spent coffee grounds and pine sawdust and 1785 ppm (3069 mg Nm3 at 10% of O2) for spent coffee grounds pellets burned in a domestic boiler (12 kW of nominal power) [14]. Garcia-Maraver et al. found 550 ppm at 10% of O2 (eq. 733 mg Nm3 at 10% O2) for Portuguese pine and 1900 ppm at 10% of O2 (eq. 2533 mg Nm3 at 10% O2) for olive pruning burned in a domestic top feed pellet-fired boiler of 22 kW of nominal power with a thermal load of 10 kW [17]. DíazRamírez et al. found 285 mg Nm3 at 10% of O2 for Brassica, 202 mg Nm3 at 10% of O2 for blend Poplar 50%-Brassica 50%, 186 mg Nm3 at 10% of O2 for Poplar and 117 mg Nm3 at 10% of O2 for DINplus [31]. Verma et al. found 4221 mg Nm3 at 10% O2 for peat pellets [15]. VOC emissions were 491, 777 and 1207 mg Nm3 at 10% of O2 for IEOSW, EOSW and IS pellets, respectively. The obtained value for IEOSW is in the same range of wood DINplus (324 mg Nm3 at 10% of O2) and spent coffee grounds pellets (539 ppm be 530 mg Nm3 at 10% of O2) [14]. In addition, NOx

Table 6 Pellets characteristics. Samples

Moisturewb (%) Diameter (mm) Length (mm) runit (kg m3) rapp (kg m3) LHVwb (MJ Kg1) Ash, 550db (%) Ash, 815db (%)

Standards

Sawdust

IS

EOMSW

IEOSW

EN 14961-6

AQHP

AQI

13 6.0 ± 0.1 20e27 1150 ± 57 601 ± 12 16.4 ± 0.3 0.6 ± 0.1 0.5 ± 0.1

9 6.0 ± 0.2 11e21 1065 ± 74 550 ± 5 18.5 ± 0.4 4 ± 0.2 4 ± 0.1

12 6.0 ± 0.1 17e20 1233 ± 62 626 ± 6 16.3 ± 0.2 3 ± 0.1 3 ± 0.1

7 6.0 ± 0.1 13e21 1245 ± 50 690 ± 14 19.8 ± 0.2 7 ± 0.4 5 ± 0.1

12 e e e 600 14.1 6 5

11 6e8 3.15e40 e 650 15.8 e 5

15 6e16 3.15e40 e 650 14.9 e 7

N. Kraiem et al. / Energy 78 (2014) 479e489

485

Table 8 Comparison between different pellets combustion parameters. Echantillon

q (kg h1)

DTe ( C)

Pu (kW)

Pf (kW)

hboiler (%)

l

[O2] (%)

Tamb ( C)

Tf ( C)

hcombustion (%)

IEOSW EOMSW IS Sawdust

2.00 2.70 2.3 2.1

19.2 21.1 19.5 16.6

9.3 10.3 9.7 8.5

11 12.3 11.8 10.2

84.9 83.7 81.8 83.4

2.07 2.40 2.35 2.61

10.9 12.3 12.1 13.0

29.8 23.1 24.6 23.0

110 124 113 120

91.5 88.4 88.2 91.0

qcomb is the fuel mass flow (kg h1), DTeau is the gradient of temperature of water in the heat recovery circuit (K), l is the air factor, [O2] is the mean oxygen proportion in fumes (%) and Tf is the flue temperature ( C).

emissions for IEOSW, IS and EOSW were 223, 255 and 384 mg Nm3 at 10% of O2, respectively. They are close to poplar (266 mg Nm3 at 10% O2) and Brazil nut shells (247 mg Nm3 at 10% O2) [16,23], higher than DINplus (176 mg Nm3 at 10% O2) [31] and sunflower husk (179 mg Nm3 at 10% O2) [32]. Garcia-Maraver et al. found 105 mg Nm3 at 10% of O2 for Portuguese pine and 340 ppm at 10% of O2 (512 mg Nm3 at 10% of O2) for olive pruning [17]. Limousy et al. found 201 ppm (377 mg Nm3 at 10% of O2) for the blend from spent coffee grounds and pine sawdust pellets and 206 ppm (407 mg Nm3 at 10% of O2) for spent coffee grounds pellets [14]. Díaz-Ramírez et al. found 672 mg Nm3 at 10% of O2 for Brassica and 418 mg Nm3 at 10% of O2 for blend Poplar 50%Brassica 50% [31].Hence, one may conclude that OMWW impregnation does not generate an excess of pollutants comparing to other biomasses. 3.4.3. Particle matter analysis During the experimental combustion tests performed with the different pellets, particles were collected in an ELPI (electrical low pressure impactor), in order to observe the impact of the biomass composition and the combustion quality upon particle concentrations and distributions. As it was expected, sawdust presents the lowest particle emissions (143 mg Nm3 at 10% O2) due to its elemental composition (0.6% wt of ash, Table 2) and also to the fact that pellet boilers are especially developed for this kind of fuel.

Results obtained with the IS pellets indicate that the impregnation of sawdust with olive mill wastewater leads to a great increase of the particle concentration in the exhaust gas. As it is presented in Table 11, the impregnation step induces an important increase of the PM (particulate matter) concentration for IS (659 mg Nm3) which could be explained by the high content of K and Ca in comparison with the pure sawdust (2.24 and 0.36 %wt in comparison with 0.12 and 0.12 %wt on wet basis respectively). This result is different than the one obtained with the EOMSW sample, where PM concentration reaches the highest value (1038 mg Nm3), while it corresponds to an intermediate value for IEOSW pellets (558 mg Nm3). While the impregnation of olive mill wastewater induces an increase of K and Ca contents, respectively 3.2 and 0.62 %wt for IEOSW pellets, and 1.5 and 0.46 %wt for the EOMSW (on wet basis), the PM concentration evolves at the opposite in comparison to the others. It seems that the presence of K and Ca at the surface of the EOMSW sample has a catalytic effect. Nevertheless, the results obtained with the non impregnated samples (sawdust and EOMSW) are in good agreement with those found in the literature with the same input power (see Table 10), while Elmay et al. obtained particle concentrations close to 100 mg Nm3 for date stones (with a ash content of 0.8 %wt on dry basis) and of 400 mg Nm3 for rachis pellets (with a ash content of 5.6 %wt) [13]. Limousy et al. studied the energetic potential of pure spent coffee grounds or blended with pine sawdust (50/50 %wt) [14]. They

(a)

(b)

(c)

(d)

Fig. 4. Gaseous emissions during combustion tests of the different pellets: (a) Sawdust, (b) IS, (c) EOSW, (d) IEOSW.

486

N. Kraiem et al. / Energy 78 (2014) 479e489

Table 9 Emissions values of CO2, O2 and CO, NOx and VOC reported at (a) 10% and (b) 13% of O2. Sample

CO2 (%)

O2 (%)

CO (mg Nm3)

NOx (mg Nm3)

VOC (mg Nm3)

IEOSW

9

11

EOMSW

8

12

IS

8

12

Sawdust

6

15

Wood DINplus [14]

6

12

spent coffee grounds-pine sawdust [14] Spent coffee grounds [14] Peat [15]

9

8

(a) 631 (b) 459 (a) 993 (b) 722 (a) 1022 (b) 743 (a) 346 (b) 252 (a) 263 (b) 191 (a) 606 (b) 441

(a) 223 (b) 162 (a) 384 (b) 279 (a) 255 (b) 185 (a) 116 (b) 84 (a) 84 (b) 61 (a) 377 (b) 274

(a) 491 (b) 357 (a) 777 (b) 565 (a) 1207 (b) 878 (a) 914 (b) 665 (a) 324 (b) 236 (a) 205 (b) 149

5

4

19

1

e

11

Sunflower husk [16]

e

13

Portuguese pine [17]

e

18

Olive pruning [17]

e

18

(a) 407 (b) 296 (a) 53 (b) 39 (a) 247 (b) 179 (a) 179 (b) 130 (a) 105 (b) 76 (a) 512 (b) 372 e

(a) 530 (b)385 e

Brazil nut shells [16]

(a) 3069 (b) 2232 (a) 2363 (b) 1718 (a) 290 (b) 211 (a) 315 (b) 229 (a) 733 (b) 533 (a) 2533 (b) 1842 (a) 3340 (b) 2247 (a) 3080 (b) 2240 (a) 797 (b) 580 (a) 839 (b) 610 (a) 186 (b) 135 (a) 285 (b) 207

e

e

(a) 660 (b) 480 (a) 302 (b) 220 (a) 266 (b) 193 (a) 672 (b) 489

e

Tomato pomace [29]

7

10

Cardoon [29]

7

10

Industrial wood wastes [30] Peach stones [30]

e

e

e

e

Poplar [31]

e

10

Brassica [31]

e

11

e e e e e

e e e

found that the particle concentration increases with the ash content from 426 mg Nm3 for the blend (with an ash content of 1.15 % wt) to 1472 mg Nm3 for pure spent coffee grounds pellets (with a ash content of 2 %wt). Garcia-Maraver et al. measured a particle concentration close to 100 mg Nm3 (10% O2) in the exhaust gas

Fig. 6. Concentration (mg Nm3) of the PM emissions according to their size distribution obtained from the combustion of different pellets in a domestic low power pellet boiler (12 kW).

during the combustion of pine pellets (ash content of 0.9 %wt on wet basis) at an input power of 14 kW, which corresponds also to the results we obtained [17]. They also observed a PM emission corresponding to about 500 mg Nm3 for olive pruning (ash content of 5.5 %wt on wet basis), which corresponds to EOMSW emissions.

Fig. 5. Evolution of gaseous emissions resulting from Sawdust, IS, EOSW and IEOSW combustion.

N. Kraiem et al. / Energy 78 (2014) 479e489

487

Table 10 Comparison of PM emissions obtained with different pellet fired boilers from various biomass fuels. Pellets

PM mg Nm3 (at 10% O2)

Boiler power (kW)

Reference

Sawdust IS EOMSW IEOSW Pine

143 659 1038 558 120 674 220 200e410 95e123 1472

12 12 12 12 10 11 16 12 12 12

This This This This [17] [33] [33] [13] [13] [14]

426 655 500

12 40 10

[14] [15] [17]

Date rachis Date stone Spent coffee ground (SCG) SCG/Pine (50/50) Sun flower husk Olive pruning

work work work work

Fig. 6 presents the mass repartition of PM observed with the different tested biomass according to the particle size distribution. As we can observe, the repartition of sawdust pellets mass particles is quite homogeneous for diameters from 0.12 to 3.09 mm, and for diameter from 0.32 to 3.09 for IS, EOMSW and IEOSW pellets. An interesting observation is that the difference of particles mass distribution observed from sawdust and IS is correlated with an increase of the number of particles (116  106 Particles Nm3 for sawdust and 187  106 Particles Nm3 for IS), while the opposite is observed with EOMSW and IEOSW (149  106 Particles Nm3 for EOMSW and 249  106 Particles Nm3 for IEOSW). Then, we decided to represent the size distribution of particles in order to understand this behavior (Fig. 6). The size distributions obtained with EOMSW and sawdust are in good agreement with those presented in previous works for similar biomasses [13,15,17]. As we can see in Fig. 7, ultrafine particles (0.04 mm of mean diameter) are present by a majority for the impregnated pellets (68% for IEOSW and 35% for IS), while their contribution remains low for the non-impregnated ones (0.4% for EOMSW and 11% for sawdust). It means that the increase of K and Ca contents in biomass may have various effects on particle emissions: (1) an increase of the number of particles emitted during the combustion process, which could be explained by a catalytic effect and the fractionation of big particles into smaller ones (2) a better oxidation of the remaining fixed carbon contained in large particles, which could be quickly oxidized in the presence of CO2 and H2O in the flue gas [34,35]. Then, if combustion is performed with pine pellets (sawdust), the increase of particle number as well as flying ash concentration (mass) after impregnation with olive mill is associated to very low carbon content in pine pellet ashes, while impregnation leads to larger particles which distribution is quite similar to the IEOSW (Fig. 7) distribution and also to the increase of ultrafine particles (Fig. 7). When combustion is carried out with EOMSW pellets, the bulk density increases a lot (see Table 2) in comparison with pine pellets. Then, high pollutant and particle emissions may be associated to the difficulty for oxygen to diffuse inside the pellets. This observation was also done by Limousy et al. [14] for the combustion of spent coffee ground. Finally, the presence of higher K and Ca levels in IEOSW induces an increase of the carbon oxidation rate associated to a better catalytic effect. These compounds are mainly present in the composition of ultrafine flying ash because their densities are low and also because they are easily volatilized at low temperature during the combustion process as shown by Torvela et al. [36]. They observed that ultrafine particles (<50 nm) are mainly composed by K, O, S, Cl and Zn during

Fig. 7. Size distribution of the PM emissions (%) obtained from the combustion of different pellets in a domestic low power pellet boiler (12 kW).

the combustion of wood logs using a biomass grate combustion unit with a nominal power output of 40 kW.

3.4.4. Ash characterization Results of the boiler dust analyses of the traditional chemical data for 17 oxides (normalized to 100%) are given in Table 11 and Fig. 8. The major 6 oxides present in biomass ash of HAR (herbaceous and agricultural residue) are respectively in descending order of magnitude K2O > SiO2 > CaO > P2O5 > MgO > Al2O3 [37]. This order can change depending on the variety of biomass. Other compounds as TiO2, SrO, CuO, Cr2O3, ZrO2, ZnO, Rb2O and other oxides like Mn are minorities. Ash biomass from the food processing industry is generally high on K2O, CaO and P2O5. Other major components other than these can be introduced as Fe2O3, which characterize contaminated biomass. Indeed, the presence of iron oxides between the major compounds in ash may be attributed to the introduction of Fe during operations that accompany food products (harvesting, processing, etc.) and their biomass (storage, transport, drying, etc.) [17]. Moreover, the presence of SrO, ZrO2 and Rb2O may be also attributed to the biomass storage and transport. In fact, these compounds are usually present in natural clays [38] and therefore left in the ash residues during combustion. Major compounds values for impregnated samples of IS and IEOMSW are close to those of reference samples except for CaO which is higher for sawdust and EOMSW because this compound characterizes wood and woody biomass. K2O of IS and IEOMSW is higher than that of woody biomass (5e15 wt %) and the same order as agricultural biomass (20e30 wt %) as shown in Table 10.

488

N. Kraiem et al. / Energy 78 (2014) 479e489

Table 11 Ash analysis (%wt). Element C K2O SiO2 CaO Fe2O3 SO3 P2O5 Na2O MgO Cl Al2O3 TiO2 SrO CuO Cr2O3 ZrO2 ZnO Rb2O MnO2 Br BaO Other oxides Sum

Sawdust 9.587 15.802 6.376 47.800 5.513 1.383 3.168 1.167 4.585 0.745 1.614 0.196 0.115 0.019 0.170 0.010 0.033 0.043 1.613 0.232 0.001 100

IS

EOMSW

IEOMSW

6.304 26.837 6.983 13.879 13.625 4.705 3.588 6.118 2.735 12.069 1.453 0.152 0.061 0.119 1.139

0.680 28.003 18.125 29.586 1.896 4.335 4.990 4.273 3.748 2.255 1.566 0.168 0.243 0.111 0.120

0.034 0.017 0.486 0.018

0.016 0.010

5.159 27.293 18.656 17.655 9.219 4.708 3.880 4.266 2.925 3.689 2.020 0.231 0.135 0.070 0.044 0.040 0.049 0.012

0.000 100

0.000 100

0.013 0.001 100

Pine sawdust [37]

Portuguese pine [17]

Grape marc [37]

Olive husks [37]

e 14.38 9.71 48.88 2.10 2.22 6.08 0.35 13.80 e 2.34 0.14 e e e e e e e e e e 100

e 11.50 20.90 26.20 21.60 0.30 4.20 2.50 4.30 0.04 6.20 e e e e e 1.00 e e e e e 100

e 36.84 9.53 28.52 1.77 6.29 8.80 0.67 4.77 e 2.63 0.18 e e e e e e e e e e 100

e 4.30 32.70 14.50 6.30 0.60 2.50 26.20 4.20 e 8.40 0.30 e e e e e e e e e e 100

results confirmed that the combined process of the treatment of olive mill wastewater presented in this work may be a promising issue for the valorization of its organic content. Hence, it will be possible to use this wet biomass as feedstock for biofuel production and thereby divide by 2 the consumption of solid biofuels such as OSW. This operation could generate as an example a benefit for Zouila company of 650.000 euros/year. Moreover, the use of OMWW for biofuels production will reduce their bad environmental impact and the different constraints for managing them in ponds. Acknowledgments

Fig. 8. Elemental composition of boiler dust from pellets used in this study.

4. Conclusion This investigation aims to recover the organic contents of olive mill wastewater through the production of solid biofuels. Hence two dry biomasses (sawdust and exhausted olive mill solid waste) were used for the impregnation of the olive mill wastewater. The IS (impregnated sawdust) and IEOSW (impregnated exhausted olive solid waste) were characterized and evaluated for biofuels production. The samples characterization shows that the addition of olive mill wastewater leads to an increase of energy content through the low heating values. Nevertheless, it leads to an increase of ash contents. Thermal degradation analysis shows a pronounced increase of the reactivity during the addition of olive mill wastewater which can be assigned to the catalytic potassium effect. The comparison between the pellets produced from raw biomasses with those produced from impregnated biomasses during combustion tests shows that the addition of olive mill wastewater has not a negative effect on the combustion quality. In particular, the gaseous emissions are lower for impregnated exhausted olive solid waste pellets comparing to exhausted olive solid waste pellets. However, an increase in particles emissions is observed which is attributed to the high content of K and Ca in OMWW. The main

This work has been partially carried-out in the frame of the CNRS-DGRS 2013 GREENPOL (EDC 25950) and the authors gratefully acknowledge CNRS-DGRS support. Authors would like to thank Zouila Company (Mahdia, Tunisia) for its technical support. Nesrine Kraiem thanks Institut Français de Tunisie (Grant
n 808475L), Agence Universitaire Francophone (Bourse mobilite ^me
es des 2013/2014) and Association Françaises des Femmes Diplo
s (Bourse Rene
e Ge
rard 2014) for their financial support. Universite References [1] Vitolo S, Petarca L, Bresci B. Treatment of olive oil industry wastes. Bioresour Technol 1999;67:129e37. [2] Caputo AC, Scacchia F, Pelagagge PM. Disposal of by-products in olive oil industry: waste-to-energy solutions. Appl Therm Eng 2003;23:197e214. _ Gündüz U, Yücel M. Comparison of physicochemical charlu E, Erog lu I, [3] Erog acteristics and photofermentative hydrogen production potential of wastewaters produced from different olive oil mills in Western-Anatolia, Turkey. Biomass Bioenergy 2009;33:706e11. [4] ANGED. Strategic options for the promotion of value of organic waste VDO Tunisia. 2009. [5] Miranda MT, Cabanillas A, Rojas S, Montero I, Ruiz A. Combined combustion of various phases of olive wastes in a conventional combustor. Fuel 2007;86: 367e72. [6] Masghouni M, Hassairi M. Energy applications of olive-oil industry byproducts: d I. The exhaust foot cake. Biomass Bioenergy 2000;18:257e62.
G, Zagrouba F. Study on the [7] Chouchene A, Jeguirim M, Khiari B, Trouve emission mechanism during devolatilization/char oxidation and direct oxidation of olive solid waste in a fixed bed reactor. J Anal Appl Pyrol 2010;87: 168e74. [8] Droussi Z, D'orazio V, Provenzano MR, Hafidi M, Ouatmane A. Study of the biodegradation and transformation of olive-mill residues during composting

N. Kraiem et al. / Energy 78 (2014) 479e489

[9] [10]

[11]

[12]

[13] [14]

[15]

[16] [17]

[18] [19]

[20]

[21]

[22]

[23]

using FTIR spectroscopy and differential scanning calorimetry. J Hazard Mater 2009;164:1281e5.
lez JF, Martínez G, Roma
n S. Catalytic pyrolysis of exhausted Encinar JM, Gonza olive oil waste. J Anal Appl Pyrol 2009;85:197e203.
G, Favre-Reguillon A, Le Buzit G. Combined Chouchene A, Jeguirim M, Trouve process for the treatment of olive oil mill wastewater: absorption on sawdust and combustion of the impregnated sawdust. Bioresour Technol 2010;101: 6962e71.
G, Le Buzit G, Khiari B, Chouchene A, Jeguirim M, Favre-Reguillon A, Trouve et al. Energetic valorisation of olive mill wastewater impregnated on low cost absorbent: sawdust versus olive solid waste. Energy 2012;39:74e81. ne A, Re
guillon AF, Trouve
G, Le Buzit G. A new valorJeguirim M, Chouche isation strategy of olive mill wastewater: impregnation on sawdust and combustion. Resour Conserv Recycl 2012;59:4e8.
G, Jeguirim M, Said R. Energy recovery of date palm residues Elmay Y, Trouve in a domestic pellet boiler. Fuel Process Technol 2013;112:12e8.
P, Kraiem N, Lajili M, Said R. Gaseous Limousy L, Jeguirim M, Dutournie products and particulate matter emissions of biomass residential boiler fired with spent coffee grounds pellets. Fuel 2013;107:323e9. Verma VK, Bram S, Delattin F, Laha P, Vandendael I, Hubin A, et al. Agro-pellets for domestic heating boilers: standard laboratory and real life performance. Appl Energy 2012;90:17e23. Cardozo E, Erlich C, Alejo L, Fransson TH. Combustion of agricultural residues: an experimental study for small-scale applications. Fuel 2014;115:778e87. Garcia-Maraver A, Zamorano M, Fernandes U, Rabaçal M, Costa M. Relationship between fuel quality and gaseous and particulate matter emissions in a domestic pellet-fired boiler. Fuel 2014;119:141e52. Lehtikangas P. Quality properties of pelletised sawdust, logging residues and bark. Biomass Bioenergy 2001;20:351e60.
G. Thermogravimetric analysis and emission Jeguirim M, Dorge S, Trouve characteristics of two energy crops in air atmosphere: Arundo donax and Miscanthus giganthus. Bioresour Technol 2010;101:788e93.
G. Thermal degradation of Miscanthus pellets: Dorge S, Jeguirim M, Trouve kinetics and aerosols characterization. Waste Biomass Valorization 2011;2: 149e55. El May Y. Measurement of gaseous and particulate pollutants during combustion of date Palm wastes for Energy recovery. Aerosol Air Qual Res 2012;12:814e25. Obernberger I, Thek G. Physical characterisation and chemical composition of densified biomass fuels with regard to their combustion behaviour. Biomass Bioenergy 2004;27:653e69. Ait Baddi G, Hafidi M, Gilard V, Revel J-C. Characterization of humic acids produced during composting of olive mill wastes: elemental and spectroscopic analyses (FTIR and 13 C-NMR). Agronomie 2003;23:661e6.

489

[24] Munir S, Daood SS, Nimmo W, Cunliffe AM, Gibbs BM. Thermal analysis and devolatilization kinetics of cotton stalk, sugar cane bagasse and shea meal under nitrogen and air atmospheres. Bioresour Technol 2009;100:1413e8.
G, Said R. Study on the thermal [25] El may Y, Jeguirim M, Dorge S, Trouve behavior of different date palm residues: characterization and devolatilization kinetics under inert and oxidative atmospheres. Energy 2012;44:702e9.
G. Pyrolysis characteristics and kinetics of Arundo donax [26] Jeguirim M, Trouve using thermogravimetric analysis. Bioresour Technol 2009;100:4026e31. [27] Jensen Anker, Dam-Johansen Kim, Wojtowicz Marek A, Serio Michael A. TGFTIR study of the Influence of potassium Chloride on Wheat Straw pyrolysis. Energy Fuels 1998;12:929e38.
enne, Norme française NF EN 14961-1/;: Bio[28] AFNOR. Norme europe
cifications des combustibles, Partie 1/: combustibles solides, Classes et spe
ne
rales. 2010. Exigences ge
lez JF, Gonz

lez J, Sabio E, Gan ~a
n J, et al. [29] Gonza alez-Garcıa CM, Ramiro A, Gonza Combustion optimisation of biomass residue pellets for domestic heating with a mural boiler. Biomass Bioenergy 2004;27:145e54. [30] Rabaçal M, Fernandes U, Costa M. Combustion and emission characteristics of a domestic boiler fired with pellets of pine, industrial wood wastes and peach stones. Renew Energy 2013;51:220e6. [31] Díaz-Ramírez M, Sebasti
an F, Royo J, Rezeau A. Influencing factors on NOX emission level during grate conversion of three pelletized energy crops. Appl Energy 2014;115:360e73. [32] Association d'Initiatives Locales pour l'Energie et l'Environnement. Les
sultats des tests de combustion en laboratoire et sites pilotes, Colloque final re du projet Life Green Pellets 23 septembre e Rennes n.d. [33] Fernandes U, Costa M. Particle emissions from a domestic pellets-fired boiler. Fuel Process Technol 2012;103:51e6. [34] Zhang Y, Gong X, Zhang B, Liu W, Xu M. Potassium catalytic hydrogen production in sorption enhanced gasification of biomass with steam. Int J Hydrogen Energy 2014;39:4234e43. [35] Mitsuoka K, Hayashi S, Amano H, Kayahara K, Sasaoaka E, Uddin MA. Gasification of woody biomass char with CO2: the catalytic effects of K and Ca species on char gasification reactivity. Fuel Process Technol 2011;92:26e31.
n A, et al. Effect of [36] Torvela T, Tissari J, Sippula O, Kaivosoja T, Leskinen J, Vire wood combustion conditions on the morphology of freshly emitted fine particles. Atmos Environ 2014;87:65e76. [37] Vassilev SV, Baxter D, Andersen LK, Vassileva CG. An overview of the chemical composition of biomass. Fuel 2010;89:913e33. [38] Williams LB, Metge DW, Eberl DD, Harvey RW, Turner AG, Prapaipong P, Poret-Peterson AT. What ́makes a natural clay antibacterial. Environ Sci Technol 2011;45:3768e73.