2017 International Conference on System Science and Engineering (ICSSE)

Analytical Modelling Fast Pyrolysis of Biomass Particles in

Fluidized Bed Reactor*
Vu Duy Pham, Huy Quoc Nguyen, Yang Van Tran, Lam Nguyen-Dinh, Hung Duong Hoang

Abstract- Determining suitable biomass particle size and Although there are many ways (i.e. direct combustion,
pyrolysis time plays a crucial role in designing and gasification and pyrolysis) to convert raw biomass into end
improvement the performance of biomass fast pyrolysis in used energy, bio-based fuels offer a partial solution to
fluidized bed reactors for bio-oils production. Current problems of energy security and environment. This is
numerical modelling investigations has not dealt with the because with proper processing they can either be combined
particle sizes and fast pyrolysis time. In this paper, analytical or used as an alternative to fossil fuels [3]. The key is to
methods were applied to investigate unsteady temperature develop processes that efficiently and sustainably convert
fields and the relationship between residence time and particle biomass to biofuels and to ensure that these fuels meet
sizes for fast pyrolysis of bagasse and wood in the fluidized bed current fuel quality standards. Although there are several
reactor. Mathematical model based on well-known heat ways to convert biomass to biofuel (i.e. fermentation,
conduction equations was developed in the MATLAB to
combustion, gasification, pyrolysis and chemical conversion),
simulate unsteady temperature fields and investigate the
pyrolysis can be seen as the most common methods for
impacts of particles sizes on the residence time. The results
shown that active pyrolysis begins occurring between 230°C to
converting biomass feedstock to bio-oil [3, 4]. Additionally,
300 °C depending on the reactor temperature. Particle sizes due to its higher yield and quality of oil compared to other
and types of biomass have significant impact of the residence pyrolysis technologies, fast pyrolysis has been considered as
time of fast pyrolysis. According to the study, the maximum a very potential solution for thennochemical conversion
radius of bagasse and wood particles for fast pyrolysis is 2.25 biomass into bio-oil [5, 6] . In order to produce high bio-oil
mm and 1 mm respectively at 500°C of reactor temperature. liquid yield, fast pyrolysis process needs to meet the below
criteria [3 , 7, 8]:
Keywords- biomass particle size, fast pyrolysis, fluidized
bed, analytical modelling, bio-oil • Particle size of biomass is small « 3mm) to improve
transferring and heating rates.
I. INTRODUCTION • Temperature of reactor should be around 500 °C to
maximize the liquid yield.
Energy security and global environmental concerns,
coupled with sustainable development requirement, are • Gas residence time should be limited to 2 seconds to
opening new chances for researching and applying renewable minimize the secondary reaction.
energy sources. Along with alternative energy resources such
as solar and wind power, biomass has been also widely • Remove rapidly char product to optimum vapor
considered as a promising renewable energy source due to its cracking.
abundance and environmental benefits [1]. It is estimated that • Condensing quickly the pyrolysis gas product to
the global biomass resources is approximately 2900 EJ per obtain the liquid product (bio-oil).
year, which has been noticed as the fourth largest energy
resource after solar, wind and geothennal energy [2]. For fast pyrolysis process, a bubbling fluidized bed
Therefore, managing and utilizing effectively this great reactor has been widely used at bench-scale and pilot-scale
energy resources would playa vital role in deciding whether plants due to its benefits such as high heat transfer rate and
or not biomass can be considered as a sustainable energy easy temperature control [9-11]. Moreover, modelling fast
source. pyrolysis in the bubbling fluidized bed reactor plays a crucial
role in determining bed hydrodynamics, designing the reactor
and maximizing the bio-oil yield. Several studies investigated
*Research supported by Ministry of Education and Training, Vietnam. experimentally and numerically the impacts of physical
V. D. Pham is with the Department of Thermal Energy Engineering, the
parameters such as size, moisture, velocity and temperature
University of Danang, University of Science and Technology (OUT),
Danang city, Vietnam (e-mail: [email protected]). on the heat, mass and momentum transport [4, 12-14].
H. Q. Nguyen , is with the Department of Thermal Energy Engineering, Papadikis, et al. [10] developed a computational model using
the University of Danang, University of Science and Technology (OUT), Eulerian model included in FLUENT's code to simulate
Danang city, Vietnam (corresponding author to provide phone: +84-905- hydrodynamics and investigate heat transfer in the fluidized
623539; e-mail: [email protected]). bed of a 150 g/h lab scale reactor. By creating a model for
V. V. Tran, is with the Department of Thermal Energy Engineering, the
University of Danang, University of Science and Technology (OUT),
validation wood flash pyrolysis in the fluidized bed reactor
Danang city, Vietnam (e-mail: [email protected]) .. [16] found that to meet flash pyrolysis wood particles should
L. Nguyen-Dinh is with the Department of Chemical Engineering, the be less than 500 !Jm. Similarly, Xiong, et al. [4] numerically
University of Danang, University of Science and Technology (OUT), investigated the impacts of operating parameters on the
Danang city, Vietnam (e-mail: [email protected]). product yield of biomass. They found that the suitable
H. D. Hoang, is with Quang Binh University, Quang Binh province,
Vietnam (e-mail: [email protected])

978-1-5386-3422-6/17/$31.00 ©2017 IEEE 685

Authorized licensed use limited to: RMIT University Library. Downloaded on June 14,2021 at 12:04:16 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on System Science and Engineering (ICSSE)

nitrogen temperature and optimal biomass particle diameter

are 800 K and 0.9 mm, respectively. aTI =0 (4)
ar r~O
The literatures shown that maximum time for fast The (1) can be rewritten as:
pyrolysis should not be more than 2 seconds to avoid
secondary reactions and obtain the highest yield bio-oil
-aT - -k- (a
--T+2-aT]
-
2
production [3 , 7, 15, 17]. However, the numerical modelling (5)
investigations have not been discussed in detail about the
at pCp ar2 r ar
relationship between radius and reaction times. In this paper, By solving (5), the temperature field of the biomass
the authors applied analytical methods to simulate unsteady- particle in the fluidized bed reactor can be determined. By
state temperature fields of biomass particles that were using nondimensionalization method (known as Buckingham
pyrolyzed fast in hot nitrogen (L,) environment in the Pi Theorem) combined with boundary conditions, the
fluidized bed. The temperature fields and relationship solution for (5) is expressed as (6) below:
between biomass fast pyrolysis time and its suitable particle
sizes is also clarified.
T(r,t)=T - (T "
- 7;)~a
00 sin(If I..c)
R exp ( - 1f 2a
'- 2
tJ (6)
j=l r R
II. MA THEMA TICAL MODEL
00 00 I 1

If; R
In this study, biomass particle is estimated as a uniform
sphere with characteristic radius of R [mm]. It is also where, a =-
k is thermal diffusivity ; 'l/j and eJ.J
consumed that the physical parameters (e.g. specific density pCp
C p [J/kg/K] , thermal conductivity k [WIm/K] and density p
coefficience are detennined by the equation tgljl = ~ and
[kg/m3 ]) is constant during fast pyrolysis processes. This I-Bj
particle is pyrolyzed in hot nitrogen environment with
temperature (Too) and heat transfer coefficience h [W/m 2/K].
Fig. 1 shows the model of a spherical biomass used to solve
temperature field in fluidized bed. Temperature field of the
particle is a function of radius (r < R), residence time (t) and Applying penetration theory, average heat transfer
physical parameters (p, Cp, k and h). coefficience for a biomass particle in the bubbling fluidized

•
bed reactor during a certain contact time t can be given [18]:

Too k .(pC )
h= m p m (7)
Jr.t
where, k m is mixture thermal conductivity of sand and
gases; ( p Cp)m is mixture specific heat.

It is noteworthy that increasing the subscript i of the (6)

results in the significant increase in the value of If/. As a
result, the exponential function ,exp reduces (-wt ::)
quickly to zero. Therefore, calculating is performed for three
fust partial sum of infinite series with an expected under
0.5% of error:

Figure I. Spherical biomass geometry for fast pyrolysis in fluidized bed 3 sin(ljI;-.C) ( at]
T(r,t)=T -(T -~)~>. R exp _ ljI2 _ (8)
i =l r R2
Heat conduction is written for a sphere biomass particle.
00 00 I I

ljI;!i
The temperature field T(r,t) is computed by the equation
below [15]. It is noted that there is small part of biomass (munreact) at
the center of the particle that is not pyrolyzed during the fast
pyrolysis [4]. Fig. 2 shows the model of biomass pyrolysis. It
(1) is assumed that this unreacted part is totally spherical at the
center of the biomass with the radius (r ill,) and its surface
The initial conditions are: temperature (T~,). If initial radius and mass of the biomass
particle are Rand m o, residence time ~ to pyrolyze the
T(r,t=O)=T" (2)
The boundary condition at the surface of the particle is
particle from initial radius R to surface ofunreacted sphere r l:!
can be determined by equation:
.
given by:
_ 3 sin(If;(Y) [ 2atpj
-k aT
ar I =h(T-T) S 00
(3)
Tp -T -(T -To)"a
00 L...l
00
;~ 1
(Y
3
If; Y
exp - 1fJR
- 2
(9)
r =R

and at the center (r = 0) of the particle:

686

Authorized licensed use limited to: RMIT University Library. Downloaded on June 14,2021 at 12:04:16 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on System Science and Engineering (ICSSE)

with

Y =m:: act
or y =[; J (10)

By experimental investigation, Xiong, et al. [4] found that

unreacted biomass mass accounts for 20%, 14% and 10% of
initial mass of biomass at different inner reactor temperatures
of 450, 475 and 500 °C, respectively.
0.6
T
(a) Ra d iu s ( RIm)
0.2 0. 1 0. 2 0.4
Time (second)
T

2
0.8 06
Figure 2. Model of particle biomass pyrolysis (b) . 0.4

Radius (mm)
o Time (seco nd)
III. RESULTS AND DISCUSSION
Figure 3. Temperature fields ofparctile biomass pyrolysis at reactor
A. Physical parameters temperatures Too = 500°C; a) R = 0.5 mm; b) R = I mm

Assumptions have been made in order to calculate and

determine the temperature fields of different biomasses (e.g.
bagasse and wood) as well as gas residence time in the
section II. Physical parameters of wood and bagasse shown in
the Table I is used to investigate the temperature fields and
relationship between particle size and residence time.

TABLE 1. PHYSICAL PARAMETERS OF BAGASSE AND WOOD

Parameters Unit Bagasse Wood Nitro

Initial temperature, To °C 25 25
Density, p kg/m3 280 700 0.456
Specific heat, Cp J/kg/K 1125 1500 1091.6
Radius (mm) Time (second)
Thermal conductivity, k W/m/K 0.1 0.15 0.0563
(a)
Thermal diffusivity, u m2/s 3x10-7 9.5xlO-s 1.1 3x IO-4

B. Modelling temperature fields

MA TLAB has been used in order to simulate and solve

equations in the section II. Fig. 3 shows the temperature field
profiles for R = 0.5 mm and R = 1 mm of wood particles at
various reactor temperatures (Too = 450 °C and 500 0c). As
can be seen in the temperature field profiles, temperature of
surface of biomass particle increase by about 8 °C when the
temperature of the reactor increases by 50 °C. It is observed 2
that active pyrolysis begins occurring between 230 °C to 300
°C depending on the reactor temperature. This result is
consistence with results obtained by [10, 19,20]. Ra dius (mm) Time (second )

(b)
Figure 4. Temperature fields ofparctile biomass pyrolysis at reactor
temperatures Too = 450 °C; a) R = 0.5 mm; b) R = I mm

687

Authorized licensed use limited to: RMIT University Library. Downloaded on June 14,2021 at 12:04:16 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on System Science and Engineering (ICSSE)

Furthermore, to validate the achieved modelling results, unreacted surfaces is not considerable, especially at above
thermo-gravimetric analysis (TGA) was utilized to active pyrolysis points (> 300 °C).
investigate the mass loss of the wood and bagasse due to BAGASSE
decomposition. A sample of 4.62 mg of the wood and 4,3 mg 2.5 ,------,....------r-------r--------,------,.-------,
of the bagasse is heated from 30 °C d~n 700 °C in Perkin
Elmer STA6000 with the heating rate of 80 K.min·'. Results 2
obtained by TGA show that wood and bagasse start
degrading in the range of 260 °C to 300 °C for wood and 200
°C to 240 °C for bagasse (Fig. 5). These results are good ..c
3 1.5

agreement with the modelling results. ~ I

- - T = 450 °C
- Wood sample
0.5
- - - Bagasse sample "' O"' T = 475 °C
4 ~~ ......... T= 500 °C
\--..: Pyrolysis slarting
0.5 0.8 1. 1 1.4 1.7 2.3
Radius (m m)

WOOD

-- - 2. 1

1.8
0L-~
I 00~--~200~--~3~OO~--~400L---~5~OO~--&ffi~--~700
Temperature("C)
.e
$1.5
Figure 5. Thennogravimetric (TG) curves for the wood and bagasse i=
1.2

c. Effects of biomass particle sizes on residence time --+- T =450 'c

-t- T =475 ' C
....... T =500 ' C
R=O.5 mm
0.5 0.6 0.7 0.8 0.9
Radius (mm)

Figure 7. Suitable biomass particle sizes and resid ence time in fluidized
bed

Fig. 7 shows the relationship between the biomass

particle sizes and residence time of fast pyrolysis at various
- Bagasse (panicle surface)
temperature of the fluidized bed reactor. Overall, the
- - - Bagasse (unreaCled surface) residence time for both wood and bagasse increase if their
- Wood (particle surface)
- - - Wood (unreacted surface) particle radius rise. As above mentioned, the residence time
0.8 1.2 1.6 of fast pyrolysis should be lower 2 seconds to maximize the
Time (s) liquid yield. Therefore, the maximum radius of biomass
R=lmm particle is obtained at 2 seconds of residence time. It can be
found in the Fig. 7 that at different temperatures of the
reactor Too = 450°C, 475 °C and 500°C, the maximum radius
of the bagasse particles is 2.18 mm, 2.22 mm and 2.25 mm,
respectively. Meanwhile, the radius of the woody biomass
particles should not be higher than 0.97 mm, 0.98 mm and 1
mm, respectively. This is because bagasse has lower density
and specific heat, so it results in the higher range of radius
- Bagassc(particlc surfocc)
- - - Bagasse (unreacled surface) particle in comparison with wood. It is interesting that the
- Wood (panicle surface)
- - - Wood (Ul1fCaCled surface)
difference in residence time of bagasse is negligible at given
1.5 2.5 reactor temperatures (Too = 450°C, 475 °C and 500°C) when
Time (s)
its particle sizes is smaller than 1.5 mm. This result are well
Figure 6. Temperature rise of wood and bagasse at particle and unreacted expected due to the low density of bagasse and small Biot
surface numbers.

Fig. 6 illustrates the temperature at particle and unreacted IV. CONCLUSION

surface of wood and bagasse at different radiuses. It can be
Analytical methods was used based on well-known heat
seen that increase in the temperature of the surfaces depends
conduction equations to investigate unsteady temperature
on the types of biomass and its sizes. Due to lower density
fields and the relationship between residence time and
and specific heat, the rise of temperature distribution inside
particles sizes for fast pyrolysis of bagasse and wood in the
the bagasse particles is more significant than wood particles.
fluidized bed reactor.
Additionally, the difference of temperature at particle and

688

Authorized licensed use limited to: RMIT University Library. Downloaded on June 14,2021 at 12:04:16 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on System Science and Engineering (ICSSE)

Temperature field profiles of woody particles at different [15] K. Papadikis, S. Gu, and A. Bridgwater, "CFD modelling of the fast
sizes and reactor temperature has been modified. It was found pyrolysis of biomass in fluidised bed reactors. Part B: Heat, momentum
and mass transport in bubbling fluidised b eds," Chemical Engineering
that active pyrolysis begins occurring between 230 °C to 300 Science, vol. 64, pp. 1036-1045, 2009.
°C depending on the reactor temperature. The results were
[16] Z. Luo, S. Wang, and K. Cen, "A model of wood fl ash pyrolysis in
validated with the TGA results and numerical modelling fluidized bed reactor," Renewable E nergy, vol. 30, pp. 377-392,2005.
results. The results also shown that increase in the
[17] A. Anca-Couce and N. Zobel, "Numerical analysis of a biomass
temperature of the particle and unreacted surfaces depends pyrolysis particle model: Solution method optimized for the coupling to
significantly on the types of biomass and its sizes. reactor models," Fuel, vol. 97, pp. 80-88, 2012.
Furthennore, the particle sizes and types of biomass have [18] K. Papadikis, S. Gu, and A. V. Bridgwater, "Computational modelling
considerable impacts on the residence time of fast pyrolysis. of the impact of particle size to the heat transfer coefficient between
According to the study, the maximum radius of bagasse and biomass particles and a fluidised bed," Fuel Processing Technology,
wood particles for fast pyrolysis is 2.25 mm and 1 mm vol. 91 , pp. 68-79, 2010.
respectively at 500°C of reactor temperature. For small sizes [19] P. R. Bonelli, E. L. Buonomo, and A. L. Cukierman, "Pyrolysis of
of light biomass (low density such as bagasse) residence time Sugarcane Bagasse and Co-pyrolysis with an Argentinean
Subbituminous Coal," Energy Sources, Part A: Recovery, Utilization,
is not considerably affected by reactor temperatures. and Environmental Effects, vol. 29, pp. 731-740, 2007/04/ 11 2007.
Overall, analytical methods can be considered as a new [20] C. Di Blasi, "Comparison of semi-global mechanisms for primary
approach to investigate the impacts of particle sizes and pyrolysis of lignocellulosic fuels ," Journal of Analytical and Applied
residence time. Pyrolysis, vol. 47, pp. 43-64 , 9// 1998.

REFERENCES
[11 N . Prakash and T. Karunanithi, "Kinetic modeling in biomass
pyrolysis-a review," Journal of Applied Sciences Research, vol. 4, pp.
1627-1636, 2008.
[21 G. Ayugi, E. Banda, and F. D ' Ujanga, "Local Thermal Insulating
Materials For Thermal Energy Storage," Rwanda Journal, vol. 23,
2011.
[3] S. Papari and K. Hawboldt, "A review on the pyrolysis of woody
biomass to bio-oil: Focus on kinetic models," Renewable and
Sustainable Energy Reviews, vol. 52, pp. 1580-1595, 2015.
[4] Q. Xiong, S. Aramideh, and S.-C. Kong, "Modeling effects of operating
conditions on biomass fast pyrolysis in bubbling fluidized bed
reactors," Energy & Fuels, vol. 27, pp. 5948-5956, 2013.
[5] A. V. Bridgwater, "Review of fast pyrolysis of biomass and product
upgrading," Biomass and bioenergy, vol. 38, pp. 68-94, 2012.
[6] B. M. Phan, L. T. Duong, V. D. Nguyen, T. B. Tran, M. H. Nguyen, L.
H. Ng uyen, et ai. , "Evaluation of the production potential of bio-oil
from Vietnamese biomass resources by fast pyrolysis," Biomass and
Bioenergy, vol. 62, pp. 74-81 , 2014.
[7] A. Bridgwater, D. Meier, and D. Radlein, "An overview of fast
pyrolysis of biomass," Organic Geochemistry, vol. 30, pp. 1479-1493,
1999.
[8] R. G. Graham, M. A. Bergougnou, and R. P. Overend, "Fast pyrolysis
of biomass," Journal of Analytical and Applied Py rolysis, vol. 6, pp.
95-135, 6// 1984.
[9] J. Saastamoinen, "Simplified model for calculation of devolatilization
in fluidi zed beds," Fuel, vol. 85, pp. 2388-2395, 2006.
[10] K. Papadikis, A. Bridgwater, and S. Gu, "CFD modelling of the fast
pyrolysis of biomass in fluidised bed reactors, Part A: Eulerian
computation of momentum transport in bubbling fluidi sed b eds,"
Chemical Engineering Science, vol. 63 , pp. 4218-4227, 2008.
[II] T. N . Trinh, P. A. Jensen, K. Dam-Johansen, N. O. Knudsen, and H. R.
S0rensen, "Influence of the pyrolysis temperature on sewage sludge
product distribution, bio-oil, and char properties," Energy & Fuels, vol.
27, pp. 1419-1427, 2013 .
[I2] C. D. Blasi, "Modelling the fast pyrolysis of cellulosic particles in
fluid-bed reactors," Chemical Engineering Science, vol. 55, pp. 5999-
6013, 12// 2000.
[13] B. Babu and A. Chaurasia, "Modeling & simulation of pyrolysis: effect
of convective heat transfer & orders of reactions," in Proceedings of
Tnternational Symposium and 55th Annual Session of flChE
(CHEMCON-2002), OU, Hyderabad, Tndia , 2002, pp. 105-106.
[14] Q. Xue, D. Dalluge, T. Heindel , R. Fox, and R. Brown, "Experimental
validation and CFD modeling study of biomass fast pyrolysis in
fluidized-bed reactors," Fuel, vol. 97, pp. 757-769, 2012.

689

Authorized licensed use limited to: RMIT University Library. Downloaded on June 14,2021 at 12:04:16 UTC from IEEE Xplore. Restrictions apply.