V Conferência Nacional de Mecânica dos Fluidos, Termodinâmica e Energia
MEFTE 2014, 11–12 Setembro 2014, Porto, Portugal
© APMTAC, 2014
Evaluation of Particle Fragmentation of Raw and Torrified Biomass in
a Drop Tube Furnace
F. F. Costa, M. Costa
IDMEC, Mechanical Engineering Department, Instituto Superior Técnico, Universidade de Lisboa, Lisboa, Portugal
email: [email protected], [email protected]
ABSTRACT: Particle fragmentation is one of the most important concerns in biomass combustion as it
strongly controls the ultra fine particulate matter emissions. Particle fragmentation corresponds to the
particle collapse during combustion, resulting two or more particles of smaller size. Torrefaction consists
on exposing biomass to an inert atmosphere between 260 ºC and 300 ºC. In such conditions,
hemicellulose is partially decomposed allowing some low calorific volatiles to be released from the raw
biomass, thus enhancing the remaining biomass energetic potential. One of the torrified biomass
characteristics consists on its brittle nature, and therefore, more prone and easily breakable in milling
processes. In this context, this study aims to understand whether torrefaction promotes particle
fragmentation during biomass combustion. Pine shells, olive stones and wheat straw were torrified at 280
ºC. Subsequently, all fuels, both raw and torrified, were burnt in a drop tube furnace (DTF) at 1100 ºC.
The results reported include burnout profiles and particle matter (PM) concentrations and size
distributions measured along the DTF. Overall, the results reveal that (i) pine shells present the lowest
PM concentrations due to its lower ash content and highest burnout values; (ii) particle fragmentation
does not occur during the combustion of raw and torrified olive stones; and (iii) particle fragmentation
occurs during the combustion of raw and torrified pine shells and wheat straw, but torrefaction promotes
particle fragmentation only in the case of the straw.
KEY-WORDS: Particle fragmentation; Torrefaction; Biomass; Drop tube furnace.
1
INTRODUCTION
Biomass is one of the most favourable alternatives to coal in terms of energy production, but its
combustion still poses a number of challenges. Particle fragmentation, which corresponds to the particle
collapse during combustion, is one of the most important concerns in biomass combustion as it
contributes to the ultra fine particulate matter emissions, which are especially harmful to the human
health.
Particle fragmentation is a rather complex subject and the very few studies available in the literature
have mainly concentrated on coal combustion. Xu et al. [1] examined the fragmentation of coal particles
in a drop tube furnace (DTF) and reported that, in the final stages of the combustion process, particles
tend to break and divide in more particles leading to a decrease in the particle mean diameter. Seames [2]
used scanning electron microscope techniques to evaluate the fragmentation of coal particles and
concluded that, as coal combustion occurs at constant diameter, the event of finding a collapsed particle
may indicate particle fragmentation. It should be noted that, while coal burns at approximately constant
diameter, biomass tends to retain its original desorbed shape during combustion, thus varying particle
diameter during combustion [3]. This means that the evaluation of the particle fragmentation in biomass
combustion using mean diameters or scanning electron microscope techniques may be inconclusive.
Torrefaction is a pre-treatment that consists on exposing biomass to an inert atmosphere. Under these
conditions hemicellulose is partially decomposed and, thereby, some low calorific volatiles are released
from the raw biomass, thus enhancing the remaining biomass energetic potential. In this way, it is
possible to obtain new biomass fuels, whose properties range between biomass and coal [4-10]. One of
the torrified biomass characteristics consists on its brittle nature, and therefore, more prone and easily
breakable in milling processes [8]. Accordingly, it is important to understand whether torrefaction
promotes particle fragmentation during biomass combustion.
In the present study biomass particle fragmentation is investigated based on the evolution of the
burnout and particle matter (PM) concentrations and size distributions along a DTF for both raw and
torrified biomass fuels (pine shells, olive stones and wheat straw). All biomass fuels were torrified in a
nitrogen inert atmosphere at 280 ºC. Then, both raw and torrified biomass fuels, were burnt in the DTF at
a constant wall temperature of 1100 ºC.
2
MATERIALS AND METHODS
2.1 Drop tube furnace
The DTF used in this study is described in detail elsewhere [11,12]. The combustion chamber of the
DTF is an electrically heated ceramic tube with an inner diameter of 38 mm and a length of 1.3 m. The
furnace wall temperatures are continuously monitored by eight type-K thermocouples uniformly
distributed along the combustion chamber. A water-cooled injector, placed at the top of the DTF, is used
to feed the solid fuels and the air to the combustion chamber. A twin screw volumetric feeder transfers the
pulverized solid fuels to an ejection system from which the particles are gas-transported to the watercooled injector.
2.2 Experimental methods
2.2.1 Particle sampling
Particle sampling along the combustion chamber axis was performed with the aid of a 1.5 m long,
water-cooled, nitrogen-quenched stainless steel probe [12]. The probe (see Fig. 1) comprised a centrally
located 3 mm inner diameter tube, through which quenched samples were evacuated with the aid of a
pump. The quenching of the chemical reactions was achieved by direct injection of nitrogen jets through
small holes very near to the probe tip. The tube for nitrogen delivery was surrounded by two concentric
tubes for probe cooling. On leaving the probe, the solid samples were collected in a Tecora total filter
holder equipped with a 47 mm diameter quartz microfiber filter. Subsequently, the collected solid
samples were placed in an oven at approximately 105 °C to dehydrate. Complete dehydration was
ascertained by repeated drying and weighing of the sample until the measured mass became constant. The
ash content in the solid samples was finally evaluated following the procedures described in the standard
CEN/TS 14775. Particle burnout data were obtained as follows:
(1)
where
is the particle burnout,
is the ash weigh fraction in the input biomass fuel, and
is the ash
weigh fraction in the char sample.
Uncertainties in char burnout calculations based on the use of ash as a tracer are related to ash volatility
at high heating rates and temperatures and ash solubility in water. Our best estimates indicated that the
uncertainties in char burnout calculations using ash as a tracer in the present work were negligible [12].
Particulate matter (PM) concentrations and size distributions were made with the aid of a low pressure
three-stage cascade impactor (LPI, TCR Tecora). PM was sampled isokinetically from the centreline of
the combustion chamber in five axial positions of the DTF (300, 500, 700, 900 and 1100 mm from
injection position), using the water-cooled, nitrogen-quenched stainless steel probe referred to above. The
LPI used allowed collecting three particulate cut sizes during the same sampling: PM with diameters
above 10 μm (PM10), PM with diameters between 2.5 μm and 10 μm, and PM with diameters below 2.5
μm (PM2.5). In order to avoid condensation along the line connecting the probe outlet to the impactor
inlet and also inside the impactor, a heating jacket (model Winkler WOTX1187) was used during
sampling. PM was collected on quartz microfiber filters, which were dried in an oven and weighted
before each test. After each test, the filters were again dried, to eliminate moisture, and weighted to
determine the quantity of PM captured.
2.2.2 Test conditions
The solid fuels studied in this work included raw and torrified pine shells, olive stones and wheat straw.
All raw biomass fuels were sieved to 1 mm size and the particle size distribution was measured using the
Malvern 2600 Particle Size Analyzer. The biomass fuels were then torrified in an insulated stainless steel
container through which 3 L/min of nitrogen at 310 ºC passed during 1 hour. During the process the
biomass temperature varied between 280 ºC and 300 ºC. Table 1 shows the characteristics of the raw and
torrified biomass fuels, including the fuel ash composition. The table includes the properties of a United
Kingdom bituminous coal for comparison purposes. Note that particle size distribution of the torrified
biomass fuels were similar to those of the raw biomass fuels as indicated by the measurements performed
with the aid of the Malvern Analyser.
In this study, measurements were carried out for all biomass fuels for DTF wall temperatures of
1100 ºC, and inlet air at room temperature. The solid fuels feed rate was around 23 g/h and the total air
flow rate was 4 L/min. The initial particle velocity was estimated, from the inlet air flow at room
temperature, to be 0.3 m/s.
Table 1. Characteristics of the raw and torrified biomass. Coal is included for comparison purposes.
Parameter
Proximate analysis
(wt%, as received)
Volatiles
Fixed carbon
Moisture
Ash
Ultimate analysis
(wt%, daf)
Carbon
Hydrogen
Nitrogen
Sulphur
Oxygen
High heating value (MJ/kg)
Ash analysis
(wt%, dry basis)
SiO2
Al2O3
Fe2O3
CaO
SO3
MgO
TiO2
P2O5
K2O
Na2O
Cl
Other oxides
Particle size (μm)
Under 30
Under 70
Under 100
Under 300
Under 500
Under 1000
Under 2000
SMD
3
Pine shells
Raw Torrified
Olive stones
Raw
Torrified
Wheat straw
Raw
Torrified
Coal
58.9
25.9
13.9
1.3
69.4
27.6
1.0
2.0
57.8
19.7
9.4
13.1
60.6
23.1
0.3
16.0
64.9
11,5
8,9
14,7
55.0
24.0
0.9
20.1
44.6
51.4
1.7
2.3
47.8
5.6
0.3
0.0
46.3
18.82
54.4
5.5
0.4
0.0
39.7
23.52
43.2
5.6
1.9
0.0
49.3
17.54
47.8
5.1
2.3
0.0
44.8
20.67
39.4
5.2
0.5
0.0
54.9
19.0
44.0
4.3
0.8
0.0
50.9
19.4
79.3
5.9
1.9
0.5
12.4
35.04
9.6
11.9
2.1
50.9
2.0
12.1
0.0
5.6
4.1
0.8
0.0
0.9
30.4
10.6
9.9
9.9
0.7
7.2
0.9
7.4
18.7
2.9
0.7
0.7
34.3
7.7
2.9
24.1
2.6
3.7
0.4
3.4
15.1
0.8
3.5
1.5
39.4
23.2
24.7
3.5
1.5
1.0
1.7
0.2
1.2
2.6
0.0
1.0
13.6
28.5
34.7
53.1
67.0
95.1
100
177.8
9.9
17.5
21.3
37.1
49.8
94.1
100
97.0
5.9
12.4
16.1
37.6
52.9
92.2
100
111.8
38.4
61.4
69.9
94.2
100
100
100
19.7
RESULTS AND DISCUSSION
3.1 Biomass characteristics
The biomass ash content and composition is very important for the evaluation of particle fragmentation.
As seen in Table 1, the pine shells ash content is much lower than those of the olive stones and wheat
straw. Moreover, the pine ashes are dominated by calcium (50.9 wt%), while the olives stones and wheat
straw ashes are dominated by silica (30.4% and 34.3%, respectively). It is also important to underline the
potassium content in the ashes, which are 4.1%, 18.7% and 15.1% for pine shells, olive stones and wheat
straw, respectively. Note that the unburned carbon in the ashes has a significant dependence on the
potassium content, as it enhances the ashes agglomeration [3].
Figure 1 shows the van Krevelen diagram for all biomass fuels studied, both raw and torrified, and for
the coal used for comparison purposes. The data reported presents a high correlation coefficient and it is
clear that torrified biomass fuels approach coal characteristics [13].
0.14
0.12
H/C
0.10
0.08
Raw pine shells
Torrified pine shells
Raw olive stones
Torrified olive stones
Coal
Raw wheat straw
Torrified wheat straw
0.06
0.04
0.02
0.00
0.00
0.50
1.00
1.50
O/C
Figure 1. Van Krevelen diagram for all biomass fuels studied. Coal is included for comparison purposes.
3.2 Particle burnout
Figure 2 shows the particle burnout data obtained for all biomass fuels, both raw and torrified, at
1100 ºC. The data for pine shells and olive stones were taken from Costa et al. [13]. Fig. 2 shows that the
pine shells present the highest levels of burnout, as compared with olives stones and wheat straw. The
figure also reveals that the wheat straw presents lower particle burnout values than the olive stones in the
first half of the DTF, but the olive stones present higher values than the wheat straw near the exit of the
DTF.
According to Fig. 2c) for the raw wheat straw the char combustion stars at x = 900 mm, while for
torrified wheat straw the char combustion starts at x = 700 mm. This is due to the volatiles release during
torrefaction, which leads to an earlier onset of the char combustion process for the torrified biomass.
Similar observations are valid for the olive stones, which are fully discussed in reference [13].
3.3 Particle matter concentrations
Figure 3 shows the PM concentrations for all biomass fuels studied, both raw and torrified. PM
concentrations are divided in particles sizes above 10 μm (PM10), between 10 μm and 2.5 μm (PM2.5 –
PM10), and below 2.5 μm (PM2.5). For pine shells the total PM concentrations are lower than those for
the olive stones and wheat straw due to the lower pine shells ash content (cf. Table 1).
Figures 3a and 3b reveal that for the raw pine shells PM under 10 μm at x = 900 mm represent 10.9% of
the total PM and at x = 1100 mm correspond to 24.6%, while for the torrified pine shells PM under 10 μm
at x = 900 mm represent 17.9% and at x = 1100 mm account for 32.1%. However, the particle burnout
data show little increase from x = 900 mm to x = 1100 mm; specifically from 98.4% to 98.9% for the raw
pine shells and from 97.8% to 99.4% for the torrified pine shells at x = 900 mm and 1100 mm. This
indicates that, near the exit of the DTF, particles strongly reduce size despite burnout almost remaining
constant. In this case it is clear that particle fragmentation occurred, leading to a reduction in particle size
and a higher concentration of smaller particles.
Figures 3c and 3d show that for the raw olive stones PM under 10 μm at x = 900 mm represent 6.5% of
the total PM and at x = 1100 mm correspond to 6.9%, while for torrified olive stones PM under 10 μm at
x = 900 mm represent 4.8% and at x = 1100 mm account for 5.6%. Particle burnout data show an increase
between these two axial positions from 90.8% to 93.8% for the raw olive stones and from 82.1% to
86.1% for the torrified olive stones. In this case, near the exit of the DTF, the PM concentrations and
distributions remain unchanged. It can therefore be concluded that for olive stones, both raw and torrified,
particle fragmentation does not occur. It is interesting to note that beyond x = 700 mm the PM size
distribution and concentration remain constant, which suggest the combustion of particles of olive stones
take place at a constant diameter.
Figure 3e and 3f show that for raw wheat straw PM under 10 μm at x = 900 mm represent 8.4% of the
total PM and at x = 1100 mm account for 26.4%, while for torrified wheat straw PM under 10 μm at x =
900 mm represent 14% and at x = 1100 mm correspond to 34.5%. Burnout data show that between x =
900 mm and x = 1100 mm particle burnout increases from 84.3% to 87.3% for raw wheat straw and from
76.4% to 80.8% for torrified wheat straw This indicates that, near the exit of the DTF, particles strongly
reduce size despite particle burnout increasing only slightly. In this case, it is clear that particle
Particle burnout (%)
100
90
80
70
60
50
40
30
20
10
0
Particle burnout (%)
100
90
80
70
60
50
40
30
20
10
0
Particle burnout (%)
fragmentation occurred, leading to a reduction in particle size and higher concentration of smaller
particles. For wheat straw particle fragmentation occurred more significantly than for pine shells.
Furthermore, in the case of the torrified wheat straw, particle fragmentation is more severe as it generates
more particles of smaller size than in the case of the raw wheat straw.
100
90
80
70
60
50
40
30
20
10
0
a)
Raw pine shells
Torrified pine shells
0
200
400
600
800
1000
1200
b)
Raw olive stones
Torrified olive stones
0
200
400
600
800
1000
1200
c)
Raw wheat straw
Torrified wheat straw
0
200
400
600
800
1000
1200
Axial distance (mm)
Figure 2. Particle burnout profiles along the axis of the DTF. a) Pine shells, b) olive stones, d) wheat
straw.
4
CONCLUSIONS
The main conclusions from this work can be summarized as follows: (i) Pine shells present the lowest PM
concentrations due to its lower ash content and highest burnout values; (ii) particle fragmentation does not
occur during the combustion of raw and torrified olive stones; (iii) particle fragmentation occurs during
the combustion of raw and torrified pine shells and wheat straw, but torrefaction promotes particle
fragmentation only in the case of the straw.
1400
PM (mg/Nm3)
1200
1400
a)
1200
1000
1000
800
800
600
600
400
400
200
200
0
500 mm
700 mm
900 mm
1100 mm
PM (mg/Nm3)
c)
1200
1000
800
800
600
600
400
400
200
200
300 mm
500 mm
700 mm
900 mm
1100 mm
1200
500 mm
700 mm
900 mm
1100 mm
500 mm
700 mm
900 mm
1100 mm
500 mm
700 mm
900 mm
1100 mm
d)
0
300 mm
1400
1400
PM (mg/Nm3)
300 mm
1400
1000
0
PM<2.5
2.5≤PM<10
PM≥10
0
300 mm
1400
1200
b)
e)
1200
1000
1000
800
800
600
600
400
400
200
200
0
300 mm
500 mm
700 mm
900 mm
1100 mm
f)
0
300 mm
Figure 3. Size-classified PM concentrations along the DTF. a) Pine shells, b) torrified pine shells, c) olive
stones, d) torrified olive stones, e) wheat straw, f) torrified wheat straw.
ACKNOWLEDGMENTS
This work was supported by Fundação para a Ciência e a Tecnologia (FCT), through IDMEC, under
LAETA Pest-OE/EME/LA0022 and under PTDC/EME-MFE/116832/2010.
REFERENCES
[1] M Xu, D Yu, H Yao, X Liu, Y Qiao (2011). Coal combustion-generated aerosols: Formation and properties.
Proceedings of the Combustion Institute 33, 1681–1697.
[2] WS Seames (2003). An initial study of the fine fragmentation fly ash particle mode generated during pulverized
coal combustion. Fuel Processing Technology. 81, 109– 125.
[3] A Williams, JM Jones, L Ma, M Pourkashanian (2012). Pollutants from the combustion of solid biomass fuels.
Progress in Energy and Combustion Science 38, 113-137.
[4] JJ Chew, V Doshi (2011). Recent advances in biomass pretreatment – Torrefaction fundamentals and technology.
Renewable and Sustainable Energy Reviews. 15, 4212-4222.
[5] K Svoboda, M Porohely, M Hartman, J Martinec (2009). Pretreatment and feeding of biomass for pressurized
entrained flow gasification. Fuel Processing Technology. 90, 629-635.
[6] WH Chen, WY Cheng, KM Lu, YP Huang (2011). An evaluation on improvement of pulverized biomass property
for solid fuel through torrefaction. Applied Energy 88, 3636-3644.
[7] V Repellin, A Govin, M Rolland, R Guyonnet (2010). Energy requirement for fine grinding of torrefied wood.
Biomass and Bioenergy 34, 923-930.
[8] M Phanphanich, S Mani (2010). Impact of torrefaction on the grindability and fuel characteristics of forest
biomass. Bioresource Technology. 102, 1246-1253.
[9] MJ Prins, KJ Ptasinski, F Janssen (2006). Torrefaction of wood Part 2. Analysis of products. Journal of Analytical
and. Applied. Pyrolysis 77, 35-40.
[10] B Moghtaderi, C Meesri, TF Wall (2004). Pyrolytic characteristics of blended coal and woody biomass. Fuel 83,
745-750.
[11] G Wang, R Zander, M Costa (2014). Oxy-fuel combustion characteristics of pulverized-coal in a drop tube
furnace. Fuel 115, 452-460.
[12] G Wang, RB Silva, JLT Azevedo, S Martins-Dias, M Costa (2014). Evaluation of the combustion behaviour and
ash characteristics of biomass waste derived fuels, pine and coal in a drop tube furnace. Fuel 117, 809-824.
[13] FF Costa, G Wang, M Costa (2015). Combustion kinetics and particle fragmentation of raw and torrified pine
shells and olive stones in a drop tube furnace. Proceedings of the Combustion Institute, in press.