Fuel 180 (2016) 324331

Contents lists available at ScienceDirect

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

Full Length Article

Combustion tests of grape marc in a multi-fuel domestic boiler

C. Schnnenbeck, G. Trouv , M. Valente, P. Garra, J.F. Brilhac
Universit de Haute-Alsace, Laboratoire Gestion des Risques et Environnement (EA2334), 3bis rue Alfred Werner, 68093 Mulhouse, France

h i g h l i g h t s

Issues for energy recovery from grape marc.

Combustion tests in domestic boiler.

Determination of combustion efficiency.

Use of blends with conventional biomasses for optimization of the process.

Measurements of gaseous and particulate pollutants.

a r t i c l e i n f o a b s t r a c t

Article history: New sources of renewable energy have to be found in order to cope with possible fossil fuel scarcity.
Received 25 August 2015 Among the renewable energy mix, biomass burning has an almost neutral CO2 balance, especially when
Received in revised form 5 April 2016 local biomass sourcing is possible. Among the diverse biomass sources, agricultural by-products such as
Accepted 6 April 2016
rice husk, corn waste, palm waste, sugarcane, date palm, olive mill wastewater. . . can be seen as renew-
Available online 17 April 2016
able biomass fuels. The present study considers the combustion process of grape marc in a domestic boi-
ler. The optimization of the heat recovery process was mainly carried out using grape marc blended with
Keywords:
wood pellets and/or miscanthus. The combustion process was assessed considering the molar carbon
Grape marc residue
Energy recovery
ratio ([CO]/[CO] + [CO2]) and the total heat losses. These parameters have been chosen as they represent
Biofuel combustion respectively the quality of the combustion in terms of CO conversion and the total efficiency of this com-
Domestic boiler bustion process. Gas and particulate matter emissions were estimated. It has been observed that the
Gaseous emissions addition of grape marc with mass ratios in the range 2533 wt% has no significant effect on the combus-
Particle emissions tion efficiency (molar carbon ratio lower than 1.3%) and on the Total Suspended Particles emissions (TSP
Total Suspended Particles ranged from 20 to 40 mg Nm3 at 10% O2) for both mixtures. An efficient combustion of the blend mis-
Electrical Low Pressure Impactor (ELPI) canthus/grape marc at weight proportion 50/50 could also be achieved. The homogenization of mixtures
of wood pellets and grape marc at 50/50 wt% was difficult to realize and led to relative high CO emissions.
Is has also been observed that Nitrogen Oxides (NOx) emissions depend on the nitrogen content of the
blend but remaining acceptable (less than 260 mg/Nm3 at 10% of O2). The feasibility of grape marc burn-
ing at domestic scale was therefore proved.
2016 Elsevier Ltd. All rights reserved.

1. Introduction efaction of fossil resources. Agricultural waste are valuable surro-

Producing energy through biomass combustion is a sustainable
process when considering the carbon dioxide balance and the rar-
Abbreviations: DIN, Deutsches Institut fr Normung; ELPI, Electrical Low
Pressure Impactor; HHV, higher heating value (kJ kg1); LHV, lower heating value
(kJ kg1); PM, particulate matter; Qa, energy loss due to the sensible heat of flue
gases (kJ kg1); Qb, energy loss due to incomplete conversion of carbon monoxide
(kJ kg1); TSP, Total Suspended Particles; DOP, di-octyl phtalate.
Corresponding author.
E-mail addresses: cornelius.schonnenbeck@uha.fr (C. Schnnenbeck), gwenaelle.
trouve@uha.fr (G. Trouv), maximilien.valente@gmail.com (M. Valente), patxi.
garra@uha.fr (P. Garra), jean-francois.brilhac@uha.fr (J.F. Brilhac).
gate fuels for energy production as they are renewable, available
in local circuits and abundant [1]. Therefore, the potential of agri-
cultural by-products as biomass fuels leads to the elaboration of
optimization strategies which have to be developed in order to
recover the maximal energy and to limit pollutant (gaseous and
particle) emissions.
The Alsace Region in France is well-known for wine production.
The vineyards approximately cover 15 600 ha in Alsace which rep-
resent 1.9% of the total region area and produce 1.2 million of hL of
wine per year, among other products. A large quantity of grape
marc is thus available in this area. From the factors given by
Caceres et al. [2], it is possible to estimate the corresponding dry

http://dx.doi.org/10.1016/j.fuel.2016.04.034
0016-2361/ 2016 Elsevier Ltd. All rights reserved.
 C. Schnnenbeck et al. / Fuel 180 (2016) 324331
matter produced in Alsace from grape marc as 30 000 tons per -
year. As a matter of comparison, the grape marc residue was calcu-
lated in Italy at 500 000 tons of dry matter per year, leading to an
interesting re-use of this material for energy recovery. In San Juan
province in Argentina, 82 000 tons of grape pomace are produced
each year which shows the potential of this waste [2]. Different
uses of grape marc have already been referenced such as activated
carbon transformation [3], cellulose and hemicellulose extraction
[4] and grape marc is known for its antioxidant and biological
properties [5]. The wine residue is also interesting for energy
recovery. For example, Celma et al. assessed the use of grape marc
in Extremadura (Spain) and calculated relevant economical specific
costs and on site availabilities for this surrogate fuel [6].
Toscano et al. examined in [7] the physicalchemical composi-
tion of grape marc residues in Italy and especially the carbon con-
tent. They returned values close to that of wood chips, beech chips
or wheat straw.
The thermal behavior of grape marc has previously been stud-
ied in the GRE lab through a thermogravimetric analysis under
nitrogen or under air and their kinetic data have been obtained [8].
Other authors as Miranda et al. mentioned that grape biomass
blended and diluted with other biomasses is available as a possible
way to limit pollutant emissions at a thermogravimetric scale [9].
Air pollution comes from many different sources and the use of
domestic boilers has already been pointed out as a major contrib-
utor for particle mater (PM) emissions [10,11]. This is an issue for
biomass combustion. Favez et al. proved that 20 10% of the ambi-
ent PM2.5 mass can be attributed to biomass burning during win-
ter in Paris [12]. Similarly, Ward et al. highlighted that more than
82 wt% of PM2.5 in Libby (Montana, U.S.A.) is due to biomass burn-
ing [13]. Consequently, the assessment of combustion quality,
especially at the domestic boiler scale, has to be performed in order
to limit pollutant emissions.
Burning wine residues in a domestic biomass boiler can be con-
sidered as an alternative pathway for grape marc reuse. Winemak-
ers could thus use part of this waste produced during wine
production as an on-site biofuel.
The present investigation is part of the project OUI Biomasse
(INTERREG IV Innovations for Sustainable Biomass Utilization in
the Upper Rhine Region), which aimed at developing the use of bio-
mass as a renewable energy source and/or raw material in the
Upper Rhine Region. This project brought together scientists from
Germany, France and Switzerland studying all aspects of the bio-
mass value chain to come up with alternative development scenar-
ios, analyse their potential impact in terms of sustainability criteria
and to draft guidelines for the sustainable use of biomass. Within
this framework, the present study focuses on grape marc combus-
tion for a low environmental impact. The objective is to analyse the
energy recovery of grape marc in a domestic scale boiler focusing
on the gas and particulate matter emission factors.
2. Materials and methods
2.1. Sample preparation and characterization
The wine residue tested in this study was collected directly
after the pressing step and delivered by a winemaker located in
Orschwir (Alsace, France). The residue was produced from Gewrz-
traminer grape. The grape marc may contain stalks, skins and
seeds. As they were collected before the fermentation step, the
by-products were then full of natural sugar. The sugar concentra-
tion within the corresponding juice was equal to 213 g/L.
Two different biomasses were used for the combustion experi-
ments involving multi-fuels: miscanthus giganteus produced in
Ammertzwiller (Alsace, France) and DIN CERTCO wood pellets,
325
purchased from SOFAG (Arc sous Cicon, France). In order to get ref-
erence values for the thermal performances of the boiler and for
the subsequent emissions, conventional biomasses usually used
in energy recovery such as beech chips were also burnt.
The physical and chemical characterizations of the different
samples were performed. The ultimate analysis of the different
samples was carried out by the Institut des Sciences Analytiques
(Centre National de la Recherche Scientifique, Villeurbanne,
France) in order to determine carbon, hydrogen, oxygen, nitrogen
and sulfur weight fractions, according to the standard EN ISO
16948 [14]. The samples were characterized following the stan-
dards EN 14774 for the moisture [15], EN 14775 for ash content
and BS EN 14918 for lower heating value (LHV) [16,17]. A calori-
metric bomb (C 200 from IKA) was used in order to determine
the higher heating value (HHV). The lower heating values (LHV)
were calculated based on free ash and water contents.
In the current study, grape marc and different blends were
burnt in a multi-fuel boiler and their thermal and environmental
performances were compared to that of DIN CERTCO wood pellets,
miscanthus and beech wood chips burnt in preliminary experi-
ments. Because the moisture content allowed in the boiler must
be less than 3035%, raw grape marc samples (denoted as RG) have
first to be dried in order to reduce their residual humidity from 67%
to 27%, thus obtaining a fuel (dry grape denoted as DG) compatible
with the dedicated boiler. Conventional fuels are denoted as BC for
Beech wood chips, M for Miscanthus and WP for wood pellets. All
the samples tested in the boiler and their acronyms are gathered in
Table 1.
Blends of miscanthus/grape marc and DIN wood pellets/grape
marc in weight proportion 75/25, 67/33 and 50/50 were tested in
the boiler. These proportions allow producing a blended fuel com-
patible with the boiler and presenting a maximal humidity of
about 35% without drying. Two types of blend mixtures were pre-
pared in loose contact or tight contact. The loose contact blends
were obtained by a simple mechanical mixing of the two materials.
They were prepared in a container (200 L) by hand mixing with a
stick during 30 min. The loose contact blends were quite heteroge-
neous due to the presence of grape stalks and of agglomerates of
humid grape marc. Due to this heterogeneity, it was not possible
to stabilize the combustion of such blended samples and the
results obtained will not be presented hereafter. The tight blended
samples were prepared grinding together the grape marc and the
co-combustible in order to obtain a homogenized blended mixture.
The materials were fed into a cutting mill type RETSCH SM 300
with a grid of 20 mm. The grinding process mainly reduced the size
of the stalks and split the agglomerates. The size distribution of the
other components of the fuel did not considerably change but the
mixing was clearly better. These tight contact mixtures are
denoted in Table 1 as WP/G 67/33, WP/G 50/50, M/G 75/25, M/G
67/33 and M/G 50/50.
2.2. Combustion tests
Combustion tests were performed in a multi-fuel boiler
(HKRST-FSK) supplied by REKA (Aars, Denmark) and equipped for
combustion analyses. A schematic description of the boiler is given
in Fig. 1. The boiler nominal thermal performance ranges from 30
to 40 kW depending on the fuel. The boiler basis is a moving step
grate, which allows using almost every biomass fuel as the grate
motion prevents slag formation. The boiler and the fitted hopper
are located on a balance. Hence the mass losses (miscanthus/wood
pellets/wood chips/grape marc) occurring during the combustion
process are recorded. This allows calculating the consumption rate
of the biomass. Water circulates in the double wall of the combus-
tion chamber and in the fire tube heat exchanger. The primary
combustion chamber is equipped with a fireproof lining on the
326 C. Schnnenbeck et al. / Fuel 180 (2016) 324331

Table 1
Fuel characteristics.

Abbreviation Fuel properties Ultimate analysis (on dry basis)

1 c
LHV (MJ kg ) Moisture (%) Ashc (%) C (%) H (%) O (%) N (%) S (%)
Wood pellets (DIN CERTCO) WP 17.7 8.0 0.3 47.1 6.1 46.6b 0.1 0.1
Wood pellets/Grape 67/33 WP/G 67/33 13.2a 27.0 0.8 48.0a 6.3a 42.8a 0.9a 0.2a
Wood pellets/Grape 50/50 WP/G 50/50 10.9a 37.5a 0.3 48.6a 6.3a 41.6a 1.3a 0.2a
Dry grape (27% humidity) DG 14.0 27.0 0.8 50.0 6.4 36.5b 2.5 0.3
Beech wood chips BC 16.5 5.0 0.4 47.6d 5.7d 44.8d 0.4d 0.0d
Miscanthus M 16.7 10.0 1.4 47.7 5.9 43.7 0.3 0.1
Misc/Grape 75/25 M/G 75/25 13.5a 24.3a 1.2 49.0a 6.0a 41.9a 0.9a 0.2a
Misc/Grape 67/33 M/G 67/33 12.5a 32.0 1.5 49.1a 6.1a 40.9a 1.0a 0.2a
Misc/Grape 50/50 M/G 50/50 10.4a 38.5a 0.9 49.4a 6.2a 40.1a 1.4a 0.2a
Raw Grape (67% humidity) RG 4.0 67.0 0.4a 50.0 6.4 36.5b 2.5 0.3

n.d.: non determined.

a
Calculated.
b
Calculated by difference.
c
On raw basis.
d
https://www.ecn.nl/phyllis2/Browse/Standard/ECN-Phyllis Beech #64.

Fig. 1. Experimental combustion set-up.

sides and the top, a blower for primary and secondary air and an
automatic ash removal system. The input power is evaluated con-
sidering the heating value of the combustible. The combustion effi-
ciency is estimated using the NF EN 303-5 (revised in 2012)
standard related to boilers with a heating power lower than
500 kW [18].
The boiler heating-up was performed using a dry combustible
(wood pellets, wood chips or miscanthus). Fuel feeding is switched
to the expected combustible once the exhaust gas temperature is
reached (70 C). Emission measurements start one hour after the
stabilization of the combustion. This delay ensures steady state
conditions and a total consumption of the initial dry combustible
injected for the boiler heating-up.
The combustion efficiency was calculated using the NF EN
12809 standard related to boilers with a heating power lower than
50 kW [19]. The environmental impact of the combustion of these
different blends was estimated using the NF EN 303-5 standard
related to boilers with a heating power lower than 500 kW and
corresponding to the fifth class. Flue gas temperatures were con-
tinuously recorded. The thermal behavior of the boiler was esti-
mated at steady-state from experiments lasting 30 min. The
primary/secondary air ratio was optimized during preliminary
combustion tests of fuels with different moisture contents (mis-
canthus, DIN CERTCO pellets and wood chips) and thereafter they
were not modified for the blends. The primary air valve was
opened at about 30% and the secondary air valve at 100%. Other
parameters of the boiler were adapted for each tested fuel in order
to obtain stable combustion conditions. The operating cycle of the
moving grate was adapted to spread the combustion over the
grates while ensuring a complete combustion on the grate. The
 C. Schnnenbeck et al. / Fuel 180 (2016) 324331 327

grate velocity was then reduced for fuels with a lower combustion The results presented hereafter thus correspond to the tight con-
velocity. The fuel feeding velocity was adapted to the energy con- tact blends.
tent of the fuel and then to its apparent density and LHV. According to the NF EN 12809 standard, the efficiency of the
combustion was calculated from the energy losses adopting the
indirect method.
2.3. Gaseous emissions
Table 2 gives the calculated energy losses. Qa indicates the
energy loss due to the sensible heat of flue gases (kJ kg1). It repre-
Emissions were measured in the chimney according to EN-304/
sents about 6% of the heating value for dry fuels and up to 9.5% for
A1 standard [20]. Flue gas temperatures were continuously
wet fuels (at 38% humidity). The presence of grape marc in the
recorded. O2, CO, CO2 and NOx were analyzed respectively through
mixture tends to increase Qa due to the higher level of moisture
paramagnetic (O2) and specific infra-red (CO, CO2 and NOx) ana-
content. This observation prevails for both miscanthus and wood
lyzers (ROSEMOUNT NGA 2000). According to EN 303-5 (NF EN
pellet. Qb is the energy loss due to the incomplete carbon monoxide
303-5, 2012) standard, the concentrations expressed in mg Nm3
oxidation (kJ kg1). It is lower than 1% of LHV when the complete
were referred to 10% of O2 in the exhaust line [18].
combustion of fuel is achieved which is the case for each sample
except for the dry marc DG and for the mixture WP/G 50/50. Qb
2.4. Particle matter analysis increases up to 8.5% due to the incomplete combustion of dry
grape marc DG with 27% of humidity. The ([CO]/[CO] + [CO2]) ratio
An Electrical Low Pressure Impactor (ELPI) manufactured by exhibits values smaller than 1% except for the samples DG, WP/G
DEKATI Ltd. (Tampere, Finland) was used to collect particles from 50/50 and MG/G 50/50. Even if the moisture content of the sample
7 nm to 10 lm into 12 size fractions. Within the ELPI, the particles strongly acts on the combustion efficiency, one can observe that
are first electrically charged according to their Stokes diameter and there is no high correlation between the moisture content and
then impacted on different stages according to their inertia related the ([CO]/[CO] + [CO2]) ratio. For instance, this ratio is higher for
to their aerodynamic diameter. A multistage electrometer counts DG than for WP/G 67/33 and M/G 67/33, while the moisture con-
the charged aerosol particles. The induced current is then simulta- tents are equal or higher for the two blended samples in compar-
neously measured for the twelve impactor stages and it is directly ison with DG. This emphasizes the impact of the composition of
converted by the electrometer in particle numbers and concentra- the blends. The combustion of the dry biomass (WP or M) may
tions through mathematical algorithms [21]. The number of parti- locally dry the grape marc in the blend. The contact area between
cles depends both on the induced current measured on each stage the M and G is more pronounced than that occurring between WP
and on the particle density in the exhaust gas here taken as and G and this may explain the more efficient combustion obtained
1 g cm3 [21]. with the M/G mixtures in comparison with the WP/G mixtures.
The total mass fraction of Total Suspended Particles (TSP) was Fig. 2 depicts the overall energy losses due to sensible heat Qa of
measured by gravimetry according to DIN CERTCO certification flue gas and due to the incomplete combustion of carbon monoxide
rules [22]. The sampling system was a heated filter holder at Qb as fraction of the lower heating value. In addition, the CO/(CO
70 C allowing the collection of TSP. It contains a two PALL glass + CO2) ratio obtained for each experiment is represented by a disk
fibers filter of A/E type, without binder, whose aerosol retention whose area is proportional to this ratio. When the ratio is small
at 0.3 lm for DOP is 99.98%. (miscanthus sample) i.e. a combustion under 10% humidity and
The combustion tests and further analyses were repeated (Qa + Qb)/LHV equal to 7.2%, the section of the disk is small. For
between two and six times, depending of the nature of the bio- 27% of humidity and (Qa + Qb)/LHV equal to 10.9% corresponding
masses and blends. to the combustion of grape marc, the section of the disk is larger.
The higher the humidity content, the higher is the energy loss. Con-
currently the energy loss increases when the fraction CO/(CO
3. Results and discussion
+ CO2) increases. For the low humidity fuels (less than 25% humid-
ity), corresponding to wood pellets, miscanthus, beech wood chips
3.1. Fuels properties and chemical characterization
and blends containing less than 33% of grape marc, the energy
losses are lower than 8.5%. Considering the regulation, this is an
Table 1 gives fuel properties and chemical characterization for
acceptable value as the NF EN 12809 standard related to boilers
three conventional biomasses (beech wood, wood pellets DIN
with a heating power lower than 50 kW requires an efficiency of
CERTCO and Miscanthus), for the grape marc and for the blends
at least 85%. In contrast, the energy loss increases when the humid-
prepared as described in the experimental section. The chemical
ity content of the fuel is higher (more than 25%) or when the burn-
characterizations and the fuel properties of the three biomasses
ing conditions are deteriorated i.e. a ratio CO/(CO2 + CO) higher
are in the same order of magnitude than those previously
than 1%. This is the case for the dried grape marc (27% humidity)
described in literature [2325]. The raw grape marc typically pre-
and for the blends with 50% of grape marc (about 38% humidity).
sents an elemental analysis close to those found for agricultural
residues, with a very high humidity content (67%) and a very low
LHV (4 MJ kg1). Energy recovery could not be suitable for grape Table 2
marc considering these characteristics. However, after the drying Combustion quality and energy losses.
process previously described in order to get a residual humidity CO Qa (% of LHV) Qb (% of LHV) Qa + Qb
COCO2  (%)
content of 27%, the grape marc samples drastically increase their (% of LHV)
LHV by a factor 3.5. WP 0.1 6.4 0.04 6.4
WP/G 67/33 1.0 7.6 0.60 8.2
WP/G 50/50 9.9 8.7 6.40 15.1
3.2. Combustion and boiler efficiency DG 8.8 5.7 5.20 10.9
BC 0.4 6 0.20 6.2
The loose contact blends were quite heterogeneous due to the M 0.1 7.1 0.06 7.2
presence of grape stalks and of agglomerates of humid grape marc. M/G 75/25 0.1 7.3 0.05 7.4
M/G 67/33 0.3 8.1 0.30 8.4
Consequently, it was not possible to stabilize the combustion of
M/G 50/50 1.3 9.5 0.90 10.4
these loose blended samples in contrast to the tight contact blends.
328 C. Schnnenbeck et al. / Fuel 180 (2016) 324331
Fig. 2. Correlations between molar carbon ration and total heat losses with the fuel humidity.
In all these cases the overall energy losses exceeded 10% and even
15% for the blend containing 50% of grape marc and 50% of wood
pellets. It seems then, that a high moisture content increases the
energy loss. Comparing the blend WP/G 67/33 with the dry grape
marc, both with moisture content of 27%, it can clearly be observed
that the humidity content is not the only parameter determining
the overall energy losses. A poor combustion quality and then a
higher ratio CO/(CO2 + CO) will as well increase the overall energy
losses. The same observation can be done comparing the two
blends at 50% of grape marc.
3.3. Pollutants emissions
3.3.1. Gaseous emissions
Exhaust emissions were recorded in order to estimate the
impact of an addition of grape marc in the blended fuels. Fig. 3
gives the average values of gas concentrations and TSP in the flue
gas at steady state. Values indicated in Fig. 3 are also reported in
Table 3 with the corresponding standard deviations. The emission
factors (EF), expressed in mg MJ1, are also given in Table 3. The
values obtained with conventional fuels such as miscanthus, beech
chips and wood pellets are presented for comparison. The limit val-
ues imposed by the standard EN 303-5 are also reported in Fig. 3.
The CO concentrations measured in the present experiments are
of the same order of magnitude than those described in the litera-
ture for conventional dry fuels such as wood pellets, miscanthus
and beech chips and close to the standard values [23,26]. Further,
the combustion process is not significantly modified by an addition
of grape marc and CO emissions measured in the exhaust are still
very low for the blend M/G 75/25. For the blends 67/33 and 50/50,
the CO concentrations measured in this exhaust increase. For M/G
67/33, the CO concentration is slightly below the limit values of EN
303-5.
It should also be noticed that the standard deviations are corre-
lated to the sample humidity. The sample moisture has a negative
impact on the running conditions of the boiler. Hence only general
tendencies can be deduced from these experiments. For the sam-
ples 50/50, the highest CO concentration in the exhaust indicates
a poorer combustion quality. As shown in Table 2, the carbon
molar ratio for the blend M/G 50/50 presents a value higher than
unity, which means that the combustion process is here not com-
plete and that CO emission dominates. Moreover, as shown in
Fig. 3, the CO concentration is higher for the wood pellets/grape
mixture than for the miscanthus/grape mixture, whatever the
mass ratio considered. The distribution of the grape marc in the
biomass could explain this phenomenon. As miscanthus is consti-
tuted of fine particles, the grape marc is mixed in a better homoge-
neous way in the miscanthus/grape mixture than in the wood
pellets/grape mixture. The shape of the natural biomasses strongly
acts on the combustion efficiency. The reactive surface in the mis-
canthus blends is higher than for pellet blends, which enhances the
combustion.
Except for the DIN CERTCO wood pellets, the mean values of
NOx emissions are higher than standard ones. This result may be
linked to the relatively high nitrogen content of the blends.
The emission factors of pollutants are well documented in the
literature concerning the biomass combustion in domestic boilers
at different scales and under different conditions. They depend
on several parameters such as the nature of the wood (soft or hard
woods), its humidity, its preparation (pellets or logs), the technol-
ogy of the combustion chamber, the output and percentage of the
biomass load [27]. In the present study, the CO concentration mea-
sured for the combustion of the samples is higher than that
imposed by the European regulation DIN CERTCO 303-5 (revised
2012). However the corresponding emission factors are of the
same order of magnitude than those previously obtained during
the combustion of pellets in domestic boilers [2729], except for
DG and WP/G 50/50. For combustion in a domestic boiler, the CO
emissions depend on the operating conditions. During an intermit-
tent combustion, the CO emission factor (EF) reached a value close
to 960 mg MJ1, whereas this value strongly decreased to
36 mg MJ1 at 100% of the load [27]. The same behavior was also
described by Kelz et al. during the combustion of wood logs in
an old domestic boiler with a maximum EF equal to
12 000 mg MJ1 of carbon monoxide [28]. With a value of
66 mg MJ1 for NOx in the exhaust (see Table 3), the combustion
of wood pellets in the REKA boiler produced a similar amount of
NOx than that previously obtained by Johansson and Ozgen
(6570 mg MJ1) using the same fuel and operating conditions
[27,28].
 C. Schnnenbeck et al. / Fuel 180 (2016) 324331 329

Fig. 3. Concentrations of CO, NOx and TSP compared to standards.

Table 3
Concentrations and emission factors of pollutants in the exhaust line.

CO (mg Nm3 ref. 10% O2) CO (mg MJ1) NOx (mg Nm3 ref. 10% O2) NOx (mg MJ1) TSP (mg Nm3 ref. 10% O2) TSP (mg MJ1)
WP 58 16 39 11 97 5 66 3 20 2 14 1
WP/G 67/33 1895 1249 1672 1102 158 18 140 16 44 2 39 2
WP/G 50/50 13 687 7318 13 708 7329 182 59 182 59 n.da n.da
DG 36 608 10 817 32 578 9626 220 71 196 63 221 2 197 2
BC 303 94 223 69 134 7 99 5 35 2 26 1
M 61 17 44 12 222 14 161 10 40 2 29 1
M/G 75/25 89 50 78 44 264 22 230 19 n.da n.da
M/G 67/33 439 594 398 538 251 47 227 43 62 7 56 6
M/G 50/50 1637 1226 1727 1293 252 43 266 45 n.da n.da
a
n.d: non determined.

In the present study, the ratio primary air/secondary air was
maintained constant during all the combustion tests, whatever
the samples and blends. This ratio should be adapted for each indi-
vidual fuel or blend in order to find the appropriate ratio minimiz-
ing the formation of incomplete gaseous compounds such as CO
and NOx. Carvalho et al. pointed out important parameters for
the optimization of combustion of agricultural fuels in pellet boil-
ers [30]. They found that the boiler control which has to be
improved according to the different properties of the agricultural
fuels is particularly the value k of the excess of air. When optimiz-
ing the combustion of vineyard prunings in a domestic pellet boiler
having a nominal output of 15 kW they proved that k has to be
taken between 1.9 and 2.3 in order to reduce the CO emissions in
the exhaust according to the EN 303-5 regulation limits. In the pre-
sent experiments with grape marc blends, the optimal value of k
ranged between 1.6 and 1.7. An increase of k potentially reduces
the emission of carbon monoxide. Moreover, an increase of k could
decrease the air temperature during the combustion process, lead-
ing to slow down the combustion rate [31].
3.3.2. Total Suspended Particles (TSP) concentrations
The total mass concentration of TSP at 10% O2 is presented in
Fig. 3. TSP is ranging from 20 to 40 mg Nm3 for most of the fuels
tested in the present study, except BC, M/G 67/33 and DG samples,
for which the values are 35, 62 and 221 mg Nm3, respectively.
These values are very close to those found by Verma et al. during
the combustion of DIN CERTCO wood pellets in a similar boiler
[32]. Regarding deviation from standard values, TSP are lower than
or equal to the stringent limit of 40 mg Nm3 at 10% O2 imposed by
the DIN CERTCO regulation and corresponding to boilers belonging
to the fifth class, except for DG and M/G 67/33 samples. For con-
ventional biomasses such as wood pellets and miscanthus but also
for blends with a proportion of grape marc equal to 33 wt%, TSP are
close to the limit values imposed by the current standard EN 303-5.
The presence of 33% of grape marc in the blends has thus no signif-
icant influence on particle emissions. However for higher content
of grape marc in the blends, the emissions of particles and of CO
will surely be higher. EF values ranging from 14 to 56 mg MJ1
are very close to those found by Nussbaumer et al. [33] during
330 C. Schnnenbeck et al. / Fuel 180 (2016) 324331
Table 4
Particle total number concentrations and relative number distributions.
PM0.1 (number %) PM0.11 (number %)
WP 82 17
WP/G 67/33 48 55
WP/G 50/50 36 63
M 66 34
M/G 67/33 60 38
M/G 50/50 31 69
the combustion of wood pellets in various modern boilers (10
50 mg MJ1), except for dry grape marc DG.
3.3.3. Fine and ultrafine particle emissions
Table 4 gives the relative particle number distributions of
PM2.5 for both raw and blended samples of wood pellets and mis-
canthus. Quite similar distributions were also obtained for various
woody fuels and biomasses as energy crops [27,3437]. The aero-
sols emitted during the combustion process of these samples
mainly consisted of fine submicron particles with a nucleation
mode dominating. The number size distributions for raw wood pel-
lets and miscanthus prove that the diameters of the particles
which are emitted are centered between 0.12 and 0.2 lm. The
combustion process in recent pellet boilers mainly generates parti-
cles centered at 0.1 lm [11].
The addition of grape marc in wood pellets modifies the number
size distributions of the emitted particles as described in Table 4.
These distributions are indeed moved toward bigger particles,
although particles with diameters in the range 0.11 lm still dom-
inate. The fraction PM2.510 is totally negligible. The particle mat-
ters (PM0.1PM2.5) recorded during the experiments show that
mainly ultrafine particles (PM0.1 and PM0.11) are emitted. PM1
indeed represents more than 99% of the total number of particles
for all samples. The fraction PM12.5 is below 0.1% of the total
number of particles. From the number size distribution presented
in Table 4, the total PM2.5 number concentration is calculated,
which ranges from 1.0  1013 particles Nm3 to 1.5  1013 parti-
cles Nm3, related to 10% of O2 in the flue gas for raw wood pellets
and miscanthus, respectively, see Table 4. Considering a relative
standard deviation close to 20%, one may conclude that the values
of the total numbers of PM2.5 for raw and miscanthus blends
(respectively 1.3  1013 and 1.8  1013 particles Nm3) are in same
order of magnitude. Hence, the total PM2.5 number concentration
is not significantly impacted by the addition of grape marc in the
miscanthus blends. This result is coherent with those obtained
by Miranda et al. through a thermogravimetric analysis [9].
4. Conclusion
A short local circuit for the use of waste to energy recovery
ensures social and economic benefits. In such a context, the use
of grape marc could be relevant in regions producing wine. The
moisture content of raw grape marc (between 65% and 80%) has
to be reduced before the combustion process as it leads to high
concentrations of carbon monoxide, of volatile organic compounds
and of particles in the exhaust line. However, the present study
demonstrates that the combustion of grape marc blended up to a
level of 33 wt% with conventional biomasses miscanthus or wood
pellets respects the required efficiency of the NF EN 12809 stan-
dard. Emission factors are not so high and should be lowered with
an adaptation of the primary/secondary air ratio for each individ-
ual fuel or blend. From the two conventional biomasses tested in
this study, the miscanthus is the best candidate to be blended with
wine residue and particularly grape marc. Its reduced particle siz-
ing and high surface area lead a good contact between miscanthus
PM12.5 (number %) PM2.5 (p Nm3 at 10% O2)
<0.1 1.0  1013 0.5  1013
<0.1 1.0  1013 0.5  1013
<0.1 1.3  1013 0.5  1013
<0.1 1.5  1013 0.3  1013
<0.1 1.8  1013 0.5  1013
<0.1 1.3  1013 0.5  1013
sawdust particles and grape marc. Using blended fuels, the emis-
sion of gas and of particulate matter are more or less similar to
those observed for other biomass resources or by-products. Emis-
sions of fine and ultrafine particles in the exhaust line are not
affected by the addition in wood pellets and miscanthus of grape
marc in weight proportions up to 50 wt%. Emissions of NOx are
limited (<260 mg Nm3 at 10% of O2) which may be attributed to
the initial nitrogen content of the grape marc.
Further studies focusing on the emissions factors of grape marc
burning regarding air excess, optimized fuel appearance, or per-
haps a pelleting process could lead to an optimization of the com-
bustion process and therefore to a new source of energy in the
Alsace region but also throughout the world in vineyard areas.
Acknowledgements
This work was performed within the framework of the OUI Bio-
masse INTERREG Program Innovations for sustainable biomass uti-
lization in the Upper Rhine Region (www.oui-biomasse.info). The
authors wish to thank the European Community (Fonds Europens
de DEveloppement Rgional, France and Germany) and the Trina-
tional Metropolitan Region for their financial support.
The authors also wish to thank the wine maker Domaine
Franois Schmitt (Orschwir, France) for supplying grape marc used
in the present investigation and for helpful discussions about the
wine making process.
References
[1] Saidur RE, Abdelaziz A, Demirbas A, Hossain MS, Mekhilef S. A review on
biomass as a fuel for boilers. Renew Sustain Energy Rev 2011;15:226289.
http://dx.doi.org/10.1016/j.rser.2011.02.015.
[2] Caceres CX, Caceres RE, Hein D, Molinac MG, Piaa JM. Biogas production from
grape pomace: thermodynamic model of the process and dynamic model of
the power generation system. Int J Hydrogen Energy 2012;37:101117.
[3] Deiana AC, Sardella MF, Silva H, Amaya A, Tancredi N. Use of grape stalk, a
waste of the viticulture industry, to obtain activated carbon. J Hazard Mater
2009;172:139. http://dx.doi.org/10.1016/j.jhazmat.2009.06.095.
[4] Spigno G, Pizzorno T, De Faveri DM. Cellulose and hemicelluloses recovery
from grape stalks. Bioresour Technol 2008;2008(99):432937. http://dx.doi.
org/10.1016/j.biortech.2007.08.044.
[5] Torres JL, Varela B, Garca MT, Carilla J, Matito C, Centelles JJ, et al. Valorization
of grape (Vitis vinifera) byproducts. Antioxidant and biological properties of
polyphenolic fractions differing in procyanidin composition and flavonol
content. J Agric Food Chem 2002;50:754855. http://dx.doi.org/10.1021/
jf025868.
[6] Celma AR, Rojas S, Lpez-Rodrguez F. Waste-to-energy possibilities for
industrial olive and grape by-products in Extremadura. Biomass Bioenergy
2007;31:52234. http://dx.doi.org/10.1016/j.biombioe.2006.08.007.
[7] Toscano G, Riva G, Duca D, Pedretti EF, Corinaldesi F, Rossini G. Analysis of the
characteristics of the residues of the wine production chain finalized to their
industrial and energy recovery. Biomass Bioenergy 2013;55:2607. http://dx.
doi.org/10.1016/j.biombioe.2013.02.015.
[8] Valente M, Brillard A, Schnnenbeck C, Brilhac JF. Investigation of grape marc
combustion using thermogravimetric analysis. Kinetic modeling using an
extended independent parallel reaction (EIPR). Fuel Process Technol
2015;131:297303. http://dx.doi.org/10.1016/j.fuproc.2014.10.034.
[9] Miranda T, Romn S, Montero I, Nogales-Delgado S, Arranz JI, Rojas CV,
Gonzlez JF. Study of the emissions and kinetic parameters during combustion
of grape pomace. dilution as an effective way to reduce pollution. In: Fuel
Process Technol. 4th international congress of energy and environment
engineering and management, vol. 103. p. 1605. http://dx.doi.org/10.1016/
j.fuproc.2011.10.002.
 C. Schnnenbeck et al. / Fuel 180 (2016) 324331
[10] Sippula O, Hokkinen J, Puustinen H, Pasi YP, Jokiniemi J. Comparison of particle
emissions from small heavy fuel oil and wood-fired boilers. Atmos Environ
2009;43:485564. http://dx.doi.org/10.1016/j.atmosenv.2009.07.022.
[11] Kaivosoja T, Jalava PI, Lamberg H, Virn A, Tapanainen M, Torvela T, et al.
Comparison of emissions and toxicological properties of fine particles from
wood and oil boilers in small (2025 kW) and medium (510 MW) scale.
Atmos Environ 2013;77:193201. http://dx.doi.org/10.1016/j.
atmosenv.2013.05.014.
[12] Favez O, Cachier H, Sciare J, Sarda-Estve R, Martinon L. Evidence for a
significant contribution of wood burning aerosols to PM2.5 during the winter
season in Paris, France. Atmos Environ 2009;43:36404. http://dx.doi.org/
10.1016/j.atmosenv.2009.04.035.
[13] Ward TJ, Lynn RR, et Lange T. The 2003/2004 Libby, Montana PM2.5 source
apportionment research study. Aerosol Sci Technol 2006;40:16677. http://dx.
doi.org/10.1080/0278682050049453.
[14] ISO 16948. Method for the determination of total carbon, hydrogen and
nitrogen contents in solid biofuels; 2015.
[15] EN 14774-2. Solid biofuels determination of moisture content oven dry
method Part 3: Moisture in general analysis sample; 2010.
[16] FprEN 14775. Solid biofuels determination of ash content; 2009.
[17] BS EN 14918. Solid biofuels. Determination of calorific value; 2009.
[18] NF EN 303-5. Heating boilers Part 5: Heating boilers for solid fuels, manually
and automatically stoked, nominal heat output of up to 500 kW
terminology, requirements, testing and marking. NF EN 303-5; 2012.
[19] NF EN 12809/A1. Residential independent boilers fired by solid fuel nominal
heat output up to 50 kW requirements and test methods; 2005.
[20] NF EN 304/A1. Heating boilers test code for heating boilers for atomizing oil
burners. NF EN 304/A1; 1998.
[21] Marjamki M, Ristimki J, Virtanen A, Moisio M, Luoma R, Keskinen J. Testing
porous metal as a collection substrate in ELPI. J Aerosol Sci, Eur Aerosol Conf
2000;31(Suppl. 1):767. http://dx.doi.org/10.1016/S0021-8502(00)90083-1.
[22] DIN CERTCO, DIN EN 13229. Certification scheme fireplace insert appliances
and cassettes fired by solid fuels with low-pollution combustion. Berlin; 2008.
[23] Verma VK, Bram S, Gauthier G, De Ruyck J. Evaluation of the performance of a
multi-fuel domestic boiler with respect to the existing European standard and
quality labels: Part-1. Biomass Bioenergy 2011;35:809. http://dx.doi.org/
10.1016/j.biombioe.2010.08.028.
[24] Leyssens G, Trouv G, Caplain I, Schnnenbeck C, Cazier F. Energetic
performances and environmental impact of the combustion of
cardboard/sawdust in a domestic boiler. Fuel 2014;122:217. http://dx.doi.
org/10.1016/j.fuel.2014.01.034.
[25] Dorge S, Jeguirim M, Trouv G. Thermal degradation of Miscanthus pellets:
kinetics and aerosols characterization. Waste Biomass Valoriz 2011;2:14955.
http://dx.doi.org/10.1007/s12649-010-9060-4.
331
[26] Verma VK, Bram S, Gauthier G, De Ruyck J. Performance of a domestic pellet
boiler as a function of operational loads: Part-2. Biomass Bioenergy
2011;35:2729. http://dx.doi.org/10.1016/j.biombioe.2010.08.043.
[27] Johansson LS, Leckner B, Gustavsson L, Cooper D, Tullin C, Potter A. Emission
characteristics of modern and old-type residential boilers fired with wood logs
and wood pellets. Atmos Environ 2004;38:418395. http://dx.doi.org/
10.1016/j.atmosenv.2004.04.020.
[28] Kelz J, Brunner T, Obernberger I, Jalava P, Hirvonen MR. PM emissions from old
and modern biomass combustion systems and their health effects. In:
Proceedings of the 18th European biomass conference, 37 May 2010, Lyon,
France.
[29] Ozgen S, Caserini S, Galante S, Guigliano M, Angelino E, Marongiu A, et al.
Emission factors from scale appliances burning wood and pellets. Atmos
Environ 2014;94:14453. http://dx.doi.org/10.1016/j.atmosenv.2014.05.032.
[30] Carvalho L, Wopienka E, Pointner C, Lundgren J, Verma VK, Haslinger W, et al.
Performance of a pellet boiler fired with agricultural fuels. Appl Energy
2013;104:28696. http://dx.doi.org/10.1016/j.apenergy.2012.10.058.
[31] Porteiro J, Patio D, Collazo J, Granada E, Moran J, Miguez JL. Experimental
analysis of the ignition front propagation of several biomass fuels in a fixed-
bed combustor. Fuel 2010;89:2635. http://dx.doi.org/10.1016/
j.fuel.2009.01.024.
[32] Verma VK, Bram S, Vandendael I, Laha P, Hubin A, De Ruyck J. Residential pellet
boilers in Belgium: standard laboratory and real life performance with respect
to European standard and quality labels. Appl Energy 2011;88:262834.
http://dx.doi.org/10.1016/j.apenergy.2011.02.004.
[33] Nussbaumer T. Overview on technologies for biomasscombustion and
emission levels of particulate matter. Expert Group on Techno-Economic
Issues (EGTEI) under the Convention on Long-Range Transboundary Air
Pollution (CLRTAP). Zurich, Swiss; 2010.
[34] Zuberbhler U, Baumbach G, Straub D, Struschka M. Fine particle emissions
from wood combustion in different firing systems. In: Proceedings of the 6th
international conference on technologies and combustion for a clean
environment. Oporto, Portugal; 2001. p. 4.
[35] Jeguirim M, Dorge S, Trouv G. Thermogravimetric analysis and emission
characteristics of two energy crops in air atmosphere: Arundo donax and
Miscanthus giganthus. Bioresour Technol 2010;101:78893. http://dx.doi.org/
10.1016/j.biortech.2009.05.063.
[36] Tissari J, Hytnen K, Lyyrnen J, Jokiniemi J. A novel field measurement
method for determining fine particle and gas emissions from residential wood
combustion. Atmos Environ 2007;41:833044. http://dx.doi.org/10.1016/j.
atmosenv.2007.06.018.
[37] Tissari J. Fine particle emissions from residential wood combustion [Ph.D.
report thesis]. Finland: Department of Environmental Science, Kuopio
University; 2008.