High Heating Rate Devolatilization Kinetics of Pulverized Biomass Fuels Johansen 2018

Subscriber access provided by Kaohsiung Medical University
Combustion
High Heating Rate Devolatilization Kinetics of Pulverized Biomass Fuels
Joakim M. Johansen, Peter Arendt Jensen, Peter Glarborg,
Dr. Nikolai DeMartini, Paul Ek, and Reginald E. Mitchell
Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.8b03100 • Publication Date (Web): 20 Nov 2018
Downloaded from http://pubs.acs.org on November 21, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,

Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.
Page 1 of 27 Energy & Fuels
1
2
3
4
5
6
7
8
High Heating Rate Devolatilization Kinetics of

9
10
11
Pulverized Biomass Fuels

12
13
14
15
16
17
18 Joakim M. Johansen,† Peter A. Jensen,† Peter Glarborg,∗,† Nikolai De Martini,‡
19
20 Paul Ek,‡ and Reginald E. Mitchell¶
21
22
23 †Department of Chemical and Biochemical Engineering, Technical University of Denmark,
24
25 Denmark
26
27 ‡Aabo Akademi University, Process Chemistry Centre, 20500 Turku, Finland
28
29 ¶Stanford University, Mechanical Engineering Department, 94305 California, USA
30
31
32 E-mail: pgl@kt.dtu.dk
33
34
35 Abstract
36
37 Devolatilization kinetics for the biomass fuels miscanthus, leached miscanthus, and
38
39 KCl doped pinewood were determined at high heating rates (∼ 105 K · s−1 ), high peak
40
41 temperatures (14051667 K), and short residence times (< 70 ms). The particle tem-
42
43 perature and residence time distribution were obtained from CFD simulations. The
44
45 measured devolatilization rates, formulated in terms of single rst order reactions, were
46
47
signicantly faster than data reported in literature. This dierence was attributed
48 partly to the fast heating rate/high temperature conditions of the present study, and
49
50 partly to a more accurate estimate of the particle temperature. The current results
51
52 indicate that neither the biomass type nor the alkali content of the biomass has a
53
54 signicant impact on the devolatilization rate under the investigated conditions. The
55
56
development in the particle morphology was studied by electron microscopy as each
57 fuel underwent partial to full conversion. The char yields ranged from 0.02 (leached
58
59
60 1
ACS Paragon Plus Environment

Energy & Fuels Page 2 of 27
1
2
3
4
5
6 miscanthus) to 0.11 (KCl doped pinewood), indicating that even during fast heating
7
8 the biomass alkali content promotes char formation.
9
10
11
12
13
Introduction
14
15 In pulverized-fuel combustion of biomass, the particles experience very fast heating rates
16
17 and high temperatures. The heating rate has been shown to be important for both the
18
19 devolatilization rate and the char yield in combustion of solid fuels. Most reported studies of
20
21 biomass devolatilization under conditions representative for pulverized fuel combustion have
22
23 114
employed drop tube furnaces or entrained ow reactors. These studies have provided
24
25 important data on devolatilization rates, as well as char yields and morphology. However,
26
27 important issues are still in discussion.
28
29 Despite the interest in biomass pyrolysis under high heating rate, high temperature con-
30
31 ditions, few reported kinetic studies have been reported for these conditions. Due partly
32
33 to the uncertainty in the diculties of estimating the particle temperature, measured rate
34 10
35 constants show a signicant scatter. Recently, Johansen et al. developed a methodology to
36
obtain more reliable values. High heating rate experiments were conducted in an entrained
37
38
ow reactor with small size fuel particles. Particles were extracted at known positions along
39
40
the reactor, thus acquiring partially converted fuel samples. The temperature and residence
41
42
time history were obtained numerically by means of computational uid dynamic (CFD) sim-
43
44 ulations. The results obtained for pinewood (abbreviated as pine) indicated faster kinetics
45
46 than found in other studies and lead to predicted residence times required for full conversion
47
48 one order of magnitude lower than kinetics derived from thermogravimetric analysis (TGA).
49
50 Such a dierence has signicant implications for ignition and ame stability in pulverized
51
52 10
fuel ring. The limitation of the work of Johansen et al. was that it only covered one
53
54 biomass; an extension of the work to a broader fuel range would be useful.
55
56 In addition to the inuence on the devolatilization kinetic parameters, the heating rate
57
58
59
60 2

1
2
3
4
5
6 is known to inuence the biomass char yield.
4,11,12,15
Low char yields of 16 wt. % (d.a.f )
7
8 are reported for fast pyrolysis ( 104 105 K/s) conditions while values as high as 1517 wt. %
9
4
10 (d.a.f ) have been obtained at low heating rates ( 1020 K/min). Dall'Ora et al. showed
11
12 how fast heating rate experiments resulted in melted chars while the brous structure of the
13
14 parent fuel was retained if exposed to low heating rates.
15
16 Also the content of mineral matter in the fuel is known to inuence the devolatilization
17
18 process. Alkali metals have been found to act as catalysts for the decomposition process,
19
1619
20 promoting char formation. The char yield is thus aected by leaching or doping pre-
21 20
22 treatment of the fuel. Trubetskaya et al. showed a strong correlation between potassium
23
24 content and char yield at heating rates of 103 K/s, but data at even higher heating rates are
25 4,11
scarce.
26
27
The objective of this work is to provide kinetic data for the devolatilization of a wider
28
29
range of biomass fuels at pulverized fuel combustion conditions, exposing biomass particles to
30
31 high heating rates. The methodology of Johansen et al.,
10
described above, is followed again.
32
33 The present work involves both major categories of biomass fuels relevant for heat and power
34
35 production, i.e., herbaceous biomass and woody biomass. Annual crops are represented by
36
37 miscanthus (a typical energy crop) in a high (parent fuel) and low (leached) ash state, while
38
39 10
woody fuels are represented by pine, likewise in a high (doped) and low (parent fuel )
40
41 ash state. In addition to deriving pyrolysis kinetics, the char yield and morphology are
42
43 characterized and the impact of the alkali metal content of the biomass devolatilization is
44
45 investigated.
46
47
48
49
50
Fuel Characterization
51
52 Two biomass fuels, pine and miscanthus, have been used in this study. Pine is a high
53
54 quality, low ash fuel, while miscanthus represents an annual crop biomass with a high ash
55
56 content. To investigate systematically the impact of the potassium content, experiments
57
58
59
60 3

1
2
3
4
5
6 were conducted with the fuels both in their parent state and with the K-content adjusted
7
8 through pretreatment (doping or leaching). The fuel matrix thus involves both a low ash
9
10 and a high ash state for both biomasses (pine and mischantus). Results for the parent pine
11
12 were reported in our previous work. 10
13
14 The woody fuel used in this study is a high quality commercial wood pellet made from
15
16 Norwegian pine. The miscanthus has been provided by the Department of Agroecology at
17
18 Aarhus University, Denmark. Proximate and ultimate analyses of the fuels are presented in
19
20 Table 1. The ash phase is dierent both in quantity and composition. The miscanthus is
21
22 naturally rich in both potassium and silicon. Washing the miscanthus results in signicant
23
24 leaching of potassium and chlorine, while the silicon content remains high. The pine sample
25
26 was doped with KCl doping to attain a level of potassium similar to that of the parent
27
28
miscanthus.
29
30
Table 1: Proximate and ultimate analysis of the investigated fuels and mineral states.
31 Volatile fractions have been determined according to standard DS/EN 15148:2009.
32
33 Unit Pine Pine Miscanthus Miscanthus
34 Parent KCl doped Parent Leached
35 Ash wt. % (dry) 0.2 1.2 3.4 0.8
36 Volatile 82.4 77.6 77.8 85.7
37
38
HHV MJ/kg (dry) 20.58 20.79 19.28 19.64
39
LHV 19.23 19.45 17.97 18.31
40 C wt. % (dry) 51.5 51.7 48.5 49.9
41 H 6.2 6.2 6.0 6.1
42 N <0.20 <0.2 0.45 0.31
43 S 0.007 0.007 0.065 0.027
44
45
Cl 0.016 0.47 0.30 0.007
46 Al mg/kg (dry) 32 20 270 <10
47 Ca 950 430 1100 870
48 Fe 26 36 270 180
49 K 480 5700 7900 330
50 Mg 140 40 540 330
51
52
Na 95 33 340 < 10
53 P 52 < 20 740 200
54 Si <200 <200 6200 2100
55 Ti <5.0 <5.0 13 <5.0
56
57
58
59
60 4

1
2
3
4
5
6 The fuels were crushed and subsequently milled in a rotating cutter mill equipped with
7
8 laboratory accuracy screens, separating the fuels in well dened size fractions: 6390μm
9
10 and 106125 μm. The resulting dusts were analyzed by two-dimensional image analysis
11
12 using a Netzsch CAMSIZER XT R . The characteristic length describing a sphere with a
13
14 similar V /A-ratio can be approximated by the Martin minimum diameter, as discussed in
15
16 the work of Trubetskaya. 21 A Rosin-Rammler distribution is used to describe the particle
17
18 size distribution, n
19 d
20 Yd = exp − ¯ (1)
d
21
22
23 where d¯ and n are the Rosin-Rammler mean diameter and the distribution spread. The
24
25
Rosin-Rammler parameters that dene the particle size range are summarized in Table 2.
26
27 Table 2: Summary of the Rosin-Rammler parameter used to describe the included fuels.
28
29 Size fraction 63 − 90 μm 106 − 125 μm
30 Rosin RAmmler parameters d/μm n d¯/μm n
¯
31 Pine 53.6 2.41 78.8 2.45
32 KCl doped pine 45.9 1.92 63.2 2.00
33
34
Miscanthus 52.7 2.13 90.8 2.92
35
Leached miscanthus 58.9 2.56 86.0 2.98
36
37
38
39
40 Treatment of the Mineral Content
41
42 Prior to milling, the miscanthus was leached in a 15 L plastic bucket half lled with uncut
43
44 fuel (volume based without compression) and topped o with deionized water. A propeller
45
46 was inserted and vigorous stirring was applied. The water was changed every 24 hours until
47
48 a total of 72 hours were reached. The water was drained and the fuel was washed with ultra
49
50 pure water on a sieve. Afterwards the fuel was dried at 333 K until reaching a moisture level
51
52 below 2 wt. %.
53
54 In the process of doping the pine, the sample was cut to powder using a cutter mill and a
55
56 4 mm screen (producing a ne powder). The sample was soaked briey in a 73.8 mM solution
57
58 of KCl in ultra pure water. Excess solution was drained by means of a büchner funnel, a low
59
60 5

1
2
3
4
5
6 powered vacuum pump, and a bre glass lter with a 75 μm cut-o. The wet cake was left
7
8 to dry at 333 K until the moisture content had dropped below 2 wt. %.
9
10
Methodology
11
12
13
14
15
The current work follows the methodology developed earlier by the authors. 10 High tem-
16
17 perature conditions are established by feeding the solid fuel into a laminar ow reactor
18
19 (LFR) 10,22,23 heated by a CH4 /H2 understoichiometric pilot ame. This conguration facili-
20
21 tated adiabatic ame temperatures of 14051667 K and particle heating rates on the order of
22
23 105 K/s. The partially or fully reacted particles are collected by a water cooled and helium
24
25 quenched probe at particle residence times between 4 and 200 ms.
26
27 The low ash fraction and the high volatility of some of the inorganic species, such as
28
29 alkali metals and halides, make it dicult to use ash as a tracer. To circumvent bias from
30
31 the release of volatile ash components, we again use calcium as a tracer. 10 The degree of
32
33 conversion is calculated by closing a Ca mass balance on the collected char:
34
35
36 wCa − wCa,0
χ= (2)
37 wCa (1 − wash,0 )
38
39
40 where wCa and wCa,0 are the weight percent of calcium in the char sample and in the parent
41
42 fuel, respectively, while wash,0 is the weight percent of ash in the parent fuel.
43
44 The particle residence times and temperature histories are calculated by multiphase re-
45
46 acting ow CFD simulations. All simulations are carried out using ANSYS Fluent 15.0.7.
47
48 The CFD model describes the reactor from the fuel inlet and 0.25 m downstream in an
49
50 axisymmetric representation of the domain. The pilot ame is assumed to be in thermody-
51
52 namic equilibrium and thus modeled as exhaust gases entering the domain at the adiabatic
53
54
ame temperature. Any reaction between fuel particles and species in the ame exhaust is
55
neglected. Internal gradients in the particles are neglected due to their small size. The eect
56
57 of the change in heat capacity during devolatilization is accounted for, 10 while the impact
58
59
60 6

1
2
3
4
5
6 on the heat transfer of gasses emitted has been disregarded. The gaseous ux away from the
7
8 particle is detrimental to convective heat transfer, but the eect on the particle temperature
9
10 is comparatively small. 24,25 The detailed model specication including inlet and boundary
11
12 conditions is described in the original work of Johansen et al. 10
13
14 The simulations of the reacting particles are coupled to the experiments by iteratively
15
16 updating the devolatilization kinetics applied in the CFD simulations, until the calculated
17
18 mass loss agrees with the measurements. The devolatilization is described by a single rst
19
20 order reaction (SFOR) with an Arrhenius type rate expression:
21
22
23 dV (t) −Ea
= A exp (V ∗ − V (t)) (3)
24 dt RT (t)
25
26
27 where V is the released volatile matter at time t and V ∗ is the high heating rate, high
28
29 temperature volatile content, assumed to be constant in this study. The rate expression is
30
31 tted using the pre-exponential factor, A, and the activation energy, Ea , evaluated at the
32
33 particle temperature, T . The universal gas constant is denoted R. The devolatilization is
34
35 assumed to be isothermic. A more accurate description of the devolatilization behavior may
36
37
be achieved by using a parallel reaction scheme, 26 but this is outside the scope of the present
38
work.
39
40
To facilitate comparison of the devolatilization reactivity of the pine with literature data,
41
42 a low heating rate TGA analysis was carried out. Fuel samples of 2-6 mg, milled and sieved
43
44 to 63-90 Âµm, were analyzed using a Netzsch STA 449 F1. The samples were dried out at
45
46 378 K in a pure nitrogen atmosphere and then heated to 1273 K at heating rates of 2 K/min.
47
48
49
50
51 Results and discussion
52
53
54
This section presents the results of the current study for the doped pine and the parent and
55
leached miscanthus. The data are compared to those of the parent pine. 10
56
57
58
59
60 7

1
2
3
4
5
6
7
Development in Char Morphology
8
9 The development in char morphology as the fuel particles undergo partial and eventually
10
11 full devolatilization is shown in Fig. 1 for an adiabatic ame temperature of 1566 K. The
12
13 conversion proles are calculated based on the Ca-balance of the partially reacted particles.
14
15 All fuel samples are needle-shaped with high initial aspect ratios in their virgin state. As
16
17
devolatilization progresses, the fuel particles gradually loose their structure and transform
18
19
into spherical geometries, seemingly through a melted phase. Only the leached miscanthus
20
21
tends to retain some of its needle shape structure throughout the devolatilization. Conceiv-
22
23
ably this is due to its high content of silicon; previous work 12 indicates that silicon-rich fuels
24 have a larger tendency to retain the particle morphology during pyrolysis.
25
26
Devolatilization Kinetics
27
28
29
30
31 The devolatilization kinetics are estimated based on the assumption of a rst order single
32
33
step mechanism. This way, the kinetic results are presented in a form that facilitates im-
34
35
plementation into CFD solvers. The Arrhenius parameters, derived from the experimental
36
37
data by running the iterative algorithm, are listed in Table 3 for the four fuels.
38
39
Table 3: Overview of the resulting apparent devolatilization kinetics. The data for pine were
40 drawn from Johansen et al. 10
41
42 Pine Miscanthus KCl doped pine Leached miscanthus
43 A ×10 s 3 −1
18.9 20.5 14.8 13.7
44 Ea J · mol −1
21305 17891 14350 13511
45
46
47
48
Figure 2 compares the experimental data for the 1566 K (adiabatic) ame with the sim-
49
50
ulation results, using the Arrhenius parameters listed in Table 3. Standard deviations for
51 the pine experiments were reported by Johansen et al. 10 The deviations for the present ex-
52
53 periments are of the same magnitude, but they are not shown here to avoid clutter. The
54
55 uncertainty in the data, determined by the reliability of both the experimental methods (in-
56
57 cluding using Ca as a tracer) and the particle temperature estimation by CFD, is signicant
58
59
60 8

1
2
3
4
5
6
7
1566 K.
8
9
10
Figure 1: Development in particle morphology as devolatilization progresses at an adiabatic ame temperature of
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 9

1
2
3
4
5
6 at lower fuel conversions, while the error is small at high levels of conversion. However,
7
8 the error has only a limited impact on the derived Arrhenius parameters due to the strong
9
10 gradients. The applied kinetics appear to capture the rapid conversion well for all four fuels.
11
12
ͳǤͲͲ
13
14
15 ͲǤ͹ͷ
16
ȏǦȐ

17
18 ͲǤͷͲ

19
20
ͲǤʹͷ
21

22

23 ͲǤͲͲ
24 Ͳ ͷ ͳͲ ͳͷ ʹͲ ʹͷ
25 ȋ ȌȏȐ
26
27 Figure 2: Comparison of experimental data and modeling predictions for the degree of
28
devolatilization for the four fuels for the condition with an adiabatic ame temperature of
29
30 1566 K. The results for parent pine were drawn from Johansen et al. 10
31
32
33 An Arrhenius plot comparing the rate constants for the four fuels is shown in Fig. 3.
34
Based on earlier modeling work, 27 where kinetic parameters derived from experiments using
35
36
high and low temperature equipment were compared, the rate constants derived in the
37
38 current work are applicable at particle temperatures above 800 K. The data shown in the
39
40 Arrhenius plot indicate that the rate constant for the low ash content pine ( 0.2 wt. %) is
41
42 smaller than those of the higher ash content fuels; leached miscanthus ( 0.8 wt. %), KCl
43
44 doped pine (1.2 wt. %), and parent miscanthus ( 3.4 wt. %). However, we believe that the
45
46 dierences are within the statistical uncertainty in the parameter evaluation. The present
47
48 results indicate that neither the biomass type nor the alkali content of the biomass has a
49
50 signicant impact on the devolatilization rate under the investigated conditions.
51
52 In Fig. 4, the present rate constants are compared with selected data from literature. The
53
54 literature data were chosen to represent common methods for determining devolatilization
55
56 kinetics. Kinetic studies are commonly conducted in a TGA at heating rates of the order
57
58 of kelvins per minute. 28 This approach oers accurate temperature and mass loss curves.
59
60 10

1
2
3
4
5
6 9
7 Pine
8 Miscanthus
KCl doped pine
9
Leached miscanthus
10
11 8
12
13 ln(k) [ln(s-1)]
14
15
16 7
17
18
19
20
21 6
22 0.5 1.0 1.5
23 1/T (1000/K)
24
25 Figure 3: Comparison across the dierent fuel types of the high temperature, high heating
26
27
rate rate constants obtained in the present work.
28
29 The drawback of the TGA approach is that the low heating rate results in kinetics that are
30
31 not representative for high heating rate processes. 28,29 This limitation is conrmed by the
32
33 present results; the TGA rate constant for pine is much lower than the fast heating rate
34
35 data, even when extrapolated to high temperature (Fig. 4).
36
37 As discussed by Solomon et al., 24 it is crucial to estimate correctly the particle tempera-
38
39 ture when deriving pyrolysis kinetics. Bonvicini et al. 30 used a high temperature drop tube
40
41 furnace and derived the kinetics under the assumption that the particles assume the reac-
42
43 tor temperature instantaneously upon introduction to the furnace. This approach neglects
44
45 the particle heating period during which devolatilization occurs at lower temperatures than
46
47 assumed. This in turn results in net lower rate constants, as seen in Fig. 4.
48
49 Similar to the present work, Simone et al. 29 used an entrained ow reactor and calculated
50
51 the temperature proles using CFD. Their work provides kinetic data most similar to the
52
53 current study. However, the temperatures in their study were limited to 800 K and below;
54
55
we believe that this is the reason for their rate constant being somewhat smaller than those
56
57
of the current study. The dierence indicates that both high heating rate and high peak
58
59
60 11

1
2
3
4
5
6 20
7 Peak temperature interval
15
8 (d)
9 10 (c)
10 Peak e interval
5 (b) (a)
11 Peak e interval
ln(k) [ln(s-1)]
Peak e interval
12 0
13
-5
14
Bonvicini et al.
15 -10
Simone et al.
16 TGA pine
-15
17 Pine (a)
18 -20 Miscanthus (b)
19 KCl doped pine (c)
-25 Leached miscanthus (d)
20
21 0.5 1.0 1.5 2.0 2.5 3.0 3.5
22 1000/T [K-1]
23
24
25
Figure 4: Comparison of the rate constants for (a) pine, (b) miscanthus, (c) KCl doped pine,
26 and (d) leached miscanthus. Current work is compared to literature data of pine experi-
27 ments with slow heating rate (TGA, current study), high heating rate without temperature
28 correction (Bonvicini et al. 30 ), and high heating rate experiments at low peak temperatures
29 (Simone et al. 29 ).
30
31
32 temperature are necessary to represent particle devolatilization at power plant conditions.
33
34 In Supplementary Material, a more comprehensive comparison of the present data with
35
36 devolatilization rate constants from literature is made. Most reported studies are still limited
37
38 to heating rates on the order of 103 K · s−1 or below. The kinetics of the current study are
39
40 signicantly faster than the majority of data reported in the literature.
41
42 Activation energies for biomass devolatilization at high heating rates from the present
43
44 work and reported in literature 10,31,32 are lower than those obtained at low heating rates. 3240
45
46 On the other hand, high activation energies have been reported for high heating rate coal
47
48
pyrolysis, 24 and used successfully for single reaction devolatilization in CFD models for com-
49
50
bustion of both coal 41 and biomass. 42 Several factors may inuence the particle temperature
51
52
estimate and thereby the calculated activation energy. In the present work we assume py-
53 rolysis to be thermo-neutral, but if it proceeds as an endothermic process, 43,44 it will aect
54
55 the particle temperature. The eect of the heat of reaction has been evaluated in several
56
57 numerical studies. 4547 However, Saddawi et al. 48 have shown that the activation energy ob-
58
59
60 12

1
2
3
4
5
6 tained for high temperature, high heating rates is related to the choice of model (single step
7
8 or multi-step reaction) and that similar predictions can be obtained over a wide range of
9
10 values.
11
12
13
14 Char Yield
15
16
The volatile fraction, calculated based on the Ca mass balance, is shown in Fig. 5. The error
17
18 bars indicate the experimental and the conditional standard deviation. The high ash content
19
20 fuels, miscanthus and KCl-doped pine, produce the highest amount of char on a dry and
21
22 ash free basis. This is expected since alkali matter in the inorganic phase has been shown
23
24 to promote char formation. 11,20 The ash phase of the leached miscanthus is dominated by a
25
26 large Si content that does not inuence the organic char yield.
27
28
1,00
29
30
Volatile fraction [kg/kg a.r.]
31
0,95
32
33
34
0,90
35
36
37
0,85
38
39
40
0,80
41
pine miscanthus KCl doped Leached
42 pine miscanthus
43
44
Figure 5: Comparison of the volatile fraction (dry and ash free) of the dierent fuels averaged
45
46 over experiments and reactor temperatures ( 13781566 K).
47
48
49 The volatile content in the fuel on an as-received basis makes up a signicant amount
50
51 of the fuel on both a weight and an energy basis. When comparing the results to the
52
53 standard fuel analysis (Table 1), signicantly larger fractions of volatiles are released during
54
55 devolatilization at high temperatures. This is expected since the proximate analysis is known
56
57 to underestimate the volatile yield under pulverized fuel combustion conditions.
58
59
60 13

1
2
3
4
5
6 Figure 6 compares the char yield based on the potassium content. The results are in
7
8 good agreement with data reported by Trubetskaya et al. 11,20 who tested the inuence of the
9
10 potassium content at dierent drop tube reactor temperatures ( 12731673 K) with fuel vary-
11
12 ing in potassium content ( 8028, 000 mg · kg−1 ). The data show that presence of potassium
13
14 promotes the char yield, even at the high heating rates investigated in these studies. They
15
16 conrm the trends observed by Leth-Espensen et al. 15 based on a broader data material.
17
18 14
19
20 12
21
22 10
23
Char (wt. %)
24 8
25
26 6
27 1273 K
1523 K
28 4 1673 K
29 Pine
30 KCl doped pine
2
31 Miscanthus
32 Leached miscanthus
0
33
0 5000 10000 15000
34
K content (mg/kg)
35
36
37
Figure 6: Comparison of the char residue of the dierent fuels based on potassium con-
38 tent. Results are averaged over experiments and reactor temperatures ( 13781566 K). Solid
39 symbols represents data from the current study while open symbols denote data from Tru-
40 betskaya et al. 11,20
41
42
43
44
45 Practical Implications
46
47
48 Computational Fluid Dynamics (CFD) have become an important tool for the power plant
49
50
companies to assess the importance of burner geometry, fuel properties, and operating con-
51
52
ditions. Conversion from fossil fuels such as coal to pulverized biomass involves a range
53
54
of technical challenges, including ame stabilization and burnout. The brous structure
55 of the biomass results in a more energy-intensive and less ecient milling process, leading
56
57 to particle sizes larger than those typical for pulverized coal. The devolatilization rate of
58
59
60 14

1
2
3
4
5
6 the larger particles is controlled to a signicant degree by heat transfer limitations, but the
7
8 volatile release rate for the smaller particles, important for ignition and ame stabilization,
9
10 is largely determined by the kinetics. We recommend that the kinetics derived in the present
11
12 work are used when the devolatilization is represented by a single rst-order reaction. Pre-
13
14 liminary (unpublished) CFD calculations for full scale bio-dust red burners indicate that
15
16 replacing standard devolatilization kinetics in CFD with the current fast kinetics can have
17
18 strong implications for the predicted ame stabilization, providing improved agreement with
19
20 measurement.
21
22
Conclusions
23
24
25
26
27 The single step rst order devolatilization kinetics of a range of biomass fuels were deter-
28
29 mined in a ow reactor under high temperature, high heating rate conditions. The selected
30
31 fuels covered a wide range of both silicon and ash content, representing the macro proper-
32
33 ties of both high quality woody biomass and the more dicult annual crops. The particle
34
35 temperature and residence time distribution were obtained from CFD simulations, enabling
36
37 accurate devolatilization kinetics to be derived at conditions similar to those of pulverized
38
39 fuel combustion. The devolatilization rates determined this way are signicantly faster than
40
41 results reported in literature; the dierence is attributed to the high heating rate and peak
42
temperature in the ow reactor. The present results indicate that neither the biomass type
43
44
nor the alkali content of the biomass has a signicant impact on the devolatilization rate
45
46
under the investigated conditions. The present kinetics are recommended for pulverized fuel
47
48 conditions and other applications with heating rates of the order of 103 105 K · s−1 . However,
49
50 for use in commercial CFD solvers it has to be combined with descriptions of particle heating
51
52 and low-temperature kinetics, possibly along the lines described by Johansen et al. 27 The
53
54 volatile fraction is signicantly higher when the fuel is treated at high temperatures and
55
56 high heating rates compared to a low temperature treatment. This trend is seen for all fuels.
57
58
59
60 15

1
2
3
4
5
6 The char yield is inuenced to a noticeable degree by the ash content. The promoting eect
7
8 of alkali content on the char yield agrees well with literature results for high temperature
9
10 pyrolysis of biomass.
11
12
13
14
15
Acknowledgement
16
17
18
This work is part of the Power Generation from Renewable Energy (GREEN) Research
19
20
Centre funded by the Danish Council for Strategic Research who are gratefully acknowledged.
21
22
23
24 References
25
26
27 (1) Brown, A. L.; Dayton, D. C.; Daily, J. W. A study of cellulose pyrolysis chemistry and
28
29
global kinetics at high heating rates. Energy Fuels 2001, 15, 12861294.
30
31 (2) Jimenez, S.; Remacha, P.; Ballesteros, J. C.; Gimenez, A.; Ballester, J. Kinetics of
32
33 devolatilization and oxidation of a pulverized biomass in an entrained ow reactor
34
35 under realistic combustion conditions. Combust. Flame 2008, 152, 588603.
36
37
38 (3) Lu, H.; Robert, W.; Peirce, G.; Ripa, B.; Baxter, L. L. Comprehensive study of biomass
39
40 particle combustion. Energy Fuels 2008, 22, 28262839.
41
42
43 (4) Dall'Ora, M.; Jensen, P. A.; Jensen, A. D. Suspension Combustion of Wood: Inuence
44
45
of Pyrolysis Conditions on Char Yield, Morphology, and Reactivity. Energy Fuels 2008,
46
47
22, 29552962.
48
49 (5) Biagini, E.; Simone, M.; Tognotti, L. Characterization of high heating rate chars of
50
51 biomass fuels. Proc. Combust. Inst. 2009, 32, 20432050.
52
53
54 (6) Dupont, C.; Chen, L.; Cances, J.; Commandre, J.-M.; Cuoci, A.; Pierucci, S.; Ranzi, E.
55
56 Biomass pyrolysis: Kinetic modelling and experimental validation under high temper-
57
58 ature and ash heating rate conditions. J. Anal. Appl. Pyrol. 2009, 85, 260267.
59
60 16

1
2
3
4
5
6 (7) Umeki, K.; Kirtania, K.; Chen, L.; Bhattacharya, S. Fuel Particle Conversion of Pul-
7
8 verized Biomass during Pyrolysis in an Entrained Flow Reactor. Ind. Eng. Chem. Res.
9
10 2012, 51, 1397313979.
11
12
13 (8) Li, T.; Wang, L.; Ku, X.; Guell, B. M.; Lovas, T.; Shaddix, C. R. Experimental and
14
15 Modeling Study of the Eect of Torrefaction on the Rapid Devolatilization of Biomass.
16
17 Energy Fuels 2015, 29, 43284338.
18
19
(9) Zellagui, S.; Schonnenbeck, C.; Zouaoui-Mahzoul, N.; Leyssens, G.; Authier, O.;
20
21 Thunin, E.; Porcheron, L.; Brilhac, J. F. Pyrolysis of coal and woody biomass under
22
23 N2 and CO2 atmospheres using a drop tube furnace - experimental study and kinetic
24
25 modeling. Fuel Process. Technol. 2016, 148, 99109.
26
27
28 (10) Johansen, J. M.; Gadsbøll, R.; Thomsen, J.; Jensen, P. A.; Glarborg, P.; De Martini, N.;
29
30 Ek, P.; Mancini, M.; Weber, R.; Mitchell, R. E. Devolatilization of Woody Biomass at
31
32 Short Residence Times and High Heating Rates and Peak Temperatures. Appl. Energy
33
34 2016, 162, 245256.
35
36
(11) Trubetskaya, A.; Jensen, P. A.; Jensen, A. D.; Garcia Llamas, A. D.; Umeki, K.;
37
38
Glarborg, P. Eect of fast pyrolysis conditions on biomass solid residues at high tem-
39
40
peratures. Fuel Process. Technol. 2016, 143, 118129.
41
42
43 (12) Trubetskaya, A.; Arendt Jensen, P.; Degn Jensen, A.; Steibel, M.; Splietho, H.; Glar-
44
45 borg, P.; Larsen, F. H. Comparison of high temperature chars of wheat straw and rice
46
47 husk with respect to chemistry, morphology and reactivity. Biomass Bioenergy 2016,
48
49 86, 7687.
50
51
52 (13) Holmgren, P.; Wagner, D. R.; Strandberg, A.; Molinder, R.; Wiinikka, H.; Umeki, K.;
53
54 Brostrom, M. Size, shape, and density changes of biomass particles during rapid de-
55
56 volatilization. Fuel 2017, 206, 342351.
57
58
59
60 17

1
2
3
4
5
6 (14) Niemela, N. P.; Tolvanen, H.; Saarinen, T.; Leppanen, A.; Joronen, T. CFD based
7
8 reactivity parameter determination for biomass particles of multiple size ranges in high
9
10 heating rate devolatilization. Energy 2017, 128, 676687.
11
12
13 (15) Leth-Espensen, A.; Glarborg, P.; Jensen, P. Predicting Biomass Char Yield from High
14
15 Heating Rate Devolatilization using Chemometrics. Energy Fuels 2018, 32, 95729580.
16
17
(16) Raveendran, K.; Ganesh, A.; Khilart, K. C. Inuence of Mineral Matter on Biomass
18
19
20
Pyrolysis Characteristics. Fuel 1995, 74, 18121822.
21
22 (17) Nik-Azar, M.; Hajaligol, M. M. R.; Sohrabi, M.; Dabir, B. Mineral matter eects in
23
24 rapid pyrolysis of beech wood. Fuel Process. Technol. 1997, 51, 7.
25
26
27 (18) Jensen, A.; Dam-Johansen, K.; Wo, M. A.; Serio, M. A. TG-FTIR Study of the Inuence
28
29 of Potassium Chloride on Wheat Straw Pyrolysis. Energy Fuels 1998, 12, 929938.
30
31
(19) Jones, J.; Darvell, L.; Bridgeman, T.; Pourkashanian, M.; Williams, A. An investigation
32
33
34
of the thermal and catalytic behaviour of potassium in biomass combustion. Proc.
35
36
Combust. Inst. 2007, 31, 19551963.
37
38 (20) Trubetskaya, A.; Jensen, P. A.; Jensen, A. D.; Steibel, M.; Splietho, H.; Glarborg, P.
39
40 Inuence of fast pryolysis conditions on yield and structural transformation of biomass
41
42 chars. Fuel Process. Technol. 2015, 140, 205214.
43
44
45 (21) Trubetskaya, A. Single Biomass Particle Combustino and Fuel Characterization. PhD
46
47 thesis, Technical University of Denmark, 2015.
48
49
(22) Campbell, P. A.; Mitchell, R. E.; Ma, L. Characterization of Coal Char and Biomass
50
51
52
Char Reactivities to Oxygen. Proc. Combust. Inst. 2002, 29, 519526.
53
54 (23) Ma, L.; Mitchell, R. Modeling Char Oxidation Behavior under Zone II Burning Condi-
55
56 tions at Elevated Pressures. Combust. Flame 2009, 156, 3750.
57
58
59
60 18

1
2
3
4
5
6 (24) Solomon, P. R.; Serio, M. A.; Carangelo, R. M.; Markham, J. R. Very Rapid Coal
7
8 Pyrolysis. Fuel 1986, 65, 182194.
9
10
11 (25) Solomon, P. R.; Serio, M. A.; Suuberg, E. M. Coal Pyrolysis - Experiments, Kinetic
12
13 Rates and Mechanisms. Prog. Energy Combust. Sci. 1992, 18, 133220.
14
15
(26) Hayhurst, A. N. The kinetics of the pyrolysis or devolatilisation of sewage sludge and
16
17
18
other solid fuels. Combust. Flame 2013, 160, 138144.
19
20 (27) Johansen, J. M.; Jensen, P. A.; Glarborg, P.; Mancini, M.; Weber, R.; Mitchell, R. E.
21
22 Extension of Apparent Devolatilization Kinetics from Thermally Thin to Thermally
23
24 Thick Particles in Zero Dimensions for Woody Biomass. Energy 2016, 95, 279290.
25
26
27 (28) Johansen, J. M.; Aho, M.; Paakkinen, K.; Taipale, R.; Egsgaard, H.; Jakobsen, J. G.;
28
29 Frandsen, F. J.; Glarborg, P. Release of K, Cl, and S during combustion and co-
30
31 combustion with wood of high-chlorine biomass in bench and pilot scale fuel beds.
32
33 Proc. Combust. Inst. 2013, 34, 23632372.
34
35
(29) Simone, M.; Biagini, E.; Galletti, C.; Tognotti, L. Evaluation of Global Biomass De-
36
37 volatilization Kinetics in a Drop Tube Reactor with CFD Aided Experiments. Fuel
38
39 2009, 88, 18181827.
40
41
42 (30) Bonvicini, G.; Coraggio, G.; Faleni, M. Biofuels Devolatilization and Char Combustion
43
44 Characterization with the IPFR ; International Flame Research Foundation, 2013.
45
46
47 (31) Hajaligol, M. R.; Howard, J. B.; Longwell, J. P.; Peters, W. A. Product composition
48
49 and kinetics for rapid pyrolysis of cellulose. Ind. Eng. Chem. Process Des. Dev. 1982,
50
51 21, 457465.
52
53 (32) Antal, M. J. J.; Várhegyi, G. Cellulose Pyrolysis Kinetics: The Current State of Knowl-
54
55 edge. Ind. Eng. Chem. Res. 1995, 34, 703717.
56
57
58
59
60 19

1
2
3
4
5
6 (33) Akita, K.; Kase, M. Determination of Kinetic parameters for pyrolysis of cellulose
7
8 and cellulose treated with ammonium phosphate by dierential thermal analysis and
9
10 thermal gravimetric analysis. J. Polym. Sci. Part A-1 Polym. Chem. 1967, 5, 833848.
11
12
13 (34) Min, K. Vapor-phase thermal analysis of pyrolysis products from cellulosic materials.
14
15 Combust. Flame 1977, 30, 285294.
16
17
(35) Fairbridge, C.; Ross, R. A.; Sood, S. P. A Kinetic and Surface Study of the Thermal De-
18
19
composition of Cellulose Powder in Inert and Oxidizipg Atmospheres. J. Appl. Polym.
20
21 Sci. 1978, 22, 497510.
22
23
24 (36) Antal, M. J. J.; Friedmand, H. L.; Rogers, F. E. Kinetics of Cellulose Pyrolysis in
25
26 Nitrogen and Steam. Combust. Sci. Technol. 1980, 21, 141152.
27
28
29 (37) Bilbao, R.; Arauzo, J.; Millera, A. Kinetics of thermal decomposition of cellulose. Part
30
31 I. Inuence of experimental conditions. Thermochim. Acta 1987, 120, 121131.
32
33
(38) Bilbao, R.; Arauzo, J.; Millera, A. Kinetics of thermal decomposition of cellulose. Part
34
35
II. Temperature dierences between gas and solid at high heating rates. Thermochim.
36
37 Acta 1987, 120, 133141.
38
39
40 (39) Várhegyi, G.; Antal, M. J. J. Kinetics of the thermal decomposition of cellulose, hemi-
41
42 cellulose, and sugar cane bagasse. Energy Fuels 1989, 3, 329335.
43
44
45 (40) Grønli, M.; Antal, M. J.; Várhegyi, G. A Round-Robin Study of Cellulose Pyrolysis
46
47 Kinetics. Ind. Eng. Chem. Res. 1999, 38, 22382244.
48
49
(41) Jones, J. M.; Patterson, P. M.; Pourkashanian, M.; Williams, A.; Arenillas, A.; Ru-
50
51
biera, F.; Pis, J. J. Modelling NOx formation in coal particle combustion at high
52
53 temperature: an investigation of the devolatilisation kinetic factors. Fuel 1999, 78,
54
55 11711179.
56
57
58
59
60 20

1
2
3
4
5
6 (42) Ma, L.; Jones, J. M.; Pourkashanian, M.; Williams, A. Modelling the combustion of
7
8 pulverized biomass in an industrial combustion test furnace. Fuel 2007, 86, 19591965.
9
10
11 (43) Milosavljevic, I.; Oja, V.; Suuberg, E. M. Thermal Eects in Cellulose Pyrolysis: Re-
12
13 lationship to Char Formation Processes. Ind. Eng. Chem. Res. 1996, 35, 653662.
14
15
(44) Rath, J.; Wolnger, M. G.; Steiner, G.; Krammer, G.; Barontini, F.; Cozzani, V. Heat
16
17
18
of Wood Pyrolysis. Fuel 2003, 82, 8191.
19
20 (45) Simmons, G. M.; Gentry, M. Particle size limitations due to heat transfer in determining
21
22 pyrolysis kinetics of biomass. J. Anal. Appl. Pyrolysis 1986, 10, 117127.
23
24
25 (46) Kothari, V.; Antal Jr., M. J. Numerical studies of the ash pyrolysis of cellulose. Fuel
26
27 1985, 64, 14871494.
28
29
30 (47) Narayan, R.; Antal, M. J. J. Thermal Lag, Fusion, and the Compensation Eect during
31
32
Biomass Pyrolysis. Ind. Eng. Chem. Res. 1996, 35, 17111721.
33
34 (48) Saddawi, A.; Jones, J. M.; Williams, A.; Wojtowicz, M. A. Kinetics of the Thermal
35
36 Decomposition of Biomass. Energy Fuels 2010, 24, 12741282.
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 21

1
Wood
2
3
4
5
6
7
8
9
10
11
WoodKCl
12
13
14
15
16
17
18
19
20
21
22
hus
23
24
ant
25
26
c
27
s
Mi
28
29
30
31
32
d
33
he
34
ac
35
e
husl
36
37
38
ant
39
40
sc
41
Mi
42
43 ACS Paragon Plus Environment
44
45
46
1.00
1
2
3
4
5
6
7
8
0.75 Pine
Conversion [‐]
9
10 Pine
11
12
13 Miscanthus
14
0.50
15
16
Miscanthus
17
18
19
Doped pine
20
21
0.25 Doped pine
22
23
25
26
28
29 0.00
30
31 0 5 10 15 20 25
33
34 Residence time (CFD) [ms]
35
36
37
38
39
40
41
9
Energy & PFuels
i n e Page 24 of 27
M is c a n t h u s
K C l d o p e d p in e
L e a c h e d m is c a n t h u s
18
2
ln ( k ) [ln ( s -1 ) ]
3
4
5
67
7
8
106
11 0 . 5 1 .0 1 .5
1 / T (1 0 0 0 / K )
2 0
Page
1 5
25P of
e a k 27Energy
t e m p e r a &t Fuels
u r e in t e r v a l
(d )
1 0 (c )
P e a k e in t e r v a l
1 5 (b ) (a ) P e a k e in t e r v a l
P e a k e in t e r v a l
l n ( k ) [ l n ( s -1 ) ]
2 0
3 -5
4 B o n v ic in i e t a l.
5-1 0 S im o n e e t a l.
T G A p in e
6-1 5
P in e ( a )
7-2 0 M is c a n t h u s ( b )
8-2 5 K C l d o p e d p in e ( c )
ACS Paragon L e a c h e d m Plusi s c a n t Environment
h u s (d )
9
10 0 . 5 1 . 0 1 . 5 2 .0 2 .5 3 .0 3 .5
11 1 0 0 0 / T [K ] -1
12
13
14
15
1
2
3
4 1,00
5
6
Volatile fraction [kg/kg a.r.]
7
8 0,95
9
10
11 0,90
12
13
14 0,85
15
16
17
18 0,80
19 pine miscanthus KCl doped Leached
20 pine miscanthus
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
1 4
1 2
11 0
2
8
C h a r (w t. % )
3
4
6
5 1 2 7 3 K
1 5 2 3 K
64 1 6 7 3 K
7 P in e
82 K C l d o p e d p in e
M is c a n t h u s
9 ACS Paragon Plus LEnvironment
e a c h e d m is c a n t h u s
100
11 0 5 0 0 0 1 0 0 0 0 1 5 0 0 0
K c o n te n t (m g / k g )

High Heating Rate Devolatilization Kinetics of Pulverized Biomass Fuels Johansen 2018

Enviado por

Dados do documento

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

High Heating Rate Devolatilization Kinetics of Pulverized Biomass Fuels Johansen 2018

Enviado por

Direitos autorais:

Formatos disponíveis

Subscriber access provided by Kaohsiung Medical University

is published by the American Chemical Society. 1155 Sixteenth Street N.W.,

High Heating Rate Devolatilization Kinetics of

Pulverized Biomass Fuels

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

ACS Paragon Plus Environment

Você também pode gostar