Science of the Total Environment 599–600 (2017) 926–933

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Experimental studies on combustion of composite biomass pellets in

fluidized bed
Feihong Guo ⁎, Zhaoping Zhong ⁎
Key Laboratory of Energy Thermal Conversion and Control of the Ministry of Education, School of Energy and Environment, Southeast University, Nanjing 210096, Jiangsu, China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Explore the use of catkins as an additive This work presents studies on the combustion of Composite Biomass Pellets in fluidized bed. The use
to the CBPs fuel and assess its utilization. of catkins as an additive to the CBPs fuel is first time and proved well.
• Combustion of CBPs is a first-order reac-
tion, and it is divided into three stages.
• Combustion of CBPs is mainly focused
on the stage of volatile-release.

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

Article history: This work presents studies on the combustion of Composite Biomass Pellets (CBPS) in fluidized bed using bauxite
Received 19 November 2016 particles as the bed material. Prior to the combustion experiment, cold-flow characterization and thermogravi-
Received in revised form 1 May 2017 metric analysis are performed to investigate the effect of air velocity and combustion mechanism of CBPS. The
Accepted 3 May 2017
cold-state test shows that CBPs and bauxite particles fluidize well in the fluidized bed. However, because of
Available online xxxx
the presence of large CBPs, optimization of the fluidization velocity is rather challenging. CBPs can gather at
Editor: D. Barcelo the bottom of the fluidized bed at lower gas velocities. On the contrary, when the velocity is too high, they accu-
mulate in the upper section of the fluidized bed. The suitable fluidization velocity for the system in this study was
Keywords: found to be between 1.5–2.0 m/s. At the same time, it is found that the critical fluidization velocity and the pres-
Fluidized bed sure fluctuation of the two-component system increase with the increase of CBPs mass concentration. The ther-
Composite biomass pellets mogravimetric experiment verifies that the combustion of CBPs is a first-order reaction, and it is divided into
Characteristics three stages: (i) dehydration, (ii) release and combustion of the volatile and (iii) the coke combustion. The com-
Combustion bustion of CBPs is mainly based on the stage of volatile combustion, and its activation energy is greater than that
of char combustion. During the combustion test, CBPS are burned at a 10 kg/h feed rate, while the excess air is
varied from 25% to 100%. Temperatures of the bed and flue gas concentrations (O2, CO, SO2 and NO) are recorded.
CBPs can be burnt stably, and the temperature of dense phase is maintained at 765–780 °C. With the increase of
the air velocity, the main combustion region has a tendency to move up. While the combustion is stable, O2 and
CO2 concentrations are maintained at about 7%, and 12%, respectively. The concentration of SO2 in the flue gas
after the initial stage of combustion is nearly zero. Furthermore, NO concentration is found to be closely related

⁎ Corresponding authors.
E-mail addresses: kerry151@126.com (F. Guo), zzhong@seu.edu.cn (Z. Zhong).

http://dx.doi.org/10.1016/j.scitotenv.2017.05.026
0048-9697/© 2017 Elsevier B.V. All rights reserved.
 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933
to O2: the NO reaches its peak value after initial stage and later decreases with the continued depletion of O2. To-
wards the end of combustion, NO increases with the increase of O2.
Edited by: D. Barcelo
1. Introduction
Because of the energy shortage and environmental problems, many
researchers are investigating alternative to fossil fuels, where in partic-
ular biomass combustion has becomes more attractive (Panoutsou et al.,
2009; Zeng et al., 2016). Fluidized-bed combustion technology has been
proven as one of the most effective technologies for energy conversion
of different kinds of biomass.(Kaynak et al., 2005; Chyang et al., 2008;
Youssef et al., 2009). Biomass pellets type fuel has high burning efficien-
cy, environmentally friendly, with convenient storage and transporta-
tion, and can be used in the field of heating, power generation and so
on.(Fournel et al., 2015; Nunes et al., 2014; Forbes et al., 2014; Saidur
et al., 2011). In this study we focus on China's particle fuel, with a variety
of raw materials and relatively wide range of physical and chemical
characterizations.(Tian, 2006). The Composite Biomass Pellets (CBPs)
used in this experiment is mainly composed of wood chips, rice straw
and Chinese parasol catkins, produced in Nanjing with different supply
seasons. There are numerous Chinese parasol trees planted in Nanjing,
where every March, under the high temperature and wind condition,
flying catkins seriously affects the lives of the residents (Fig. 1). In this
study, we are exploring the use of catkins as an additive to the CBPs
fuel and assess its utilization.
CBPs are relatively large and can affect the mixing and flow charac-
teristics when blended with the bed materials during fluidization. The
flow characteristics of biomass particles have been studied by a number
of researchers (Abdullah et al., 2003; Clarke et al., 2005; Cui and Grace,
2007; Rao and Reddy, 2010; Si and Guo, 2008; Shao et al., 2013), how-
ever these studies are often based on relatively smaller biomass parti-
cles with regular shape, such as sawdust, rice husk and straw etc.
Furthermore, their physical properties and geometric dimensions are
very different from that of CBPs and the conclusions are not suitable
for the CBPs in this study, hence the fluidization characteristics of CBPs
need to be studied.
Fig. 1. Chinese parasol trees in Nanjing.
927
© 2017 Elsevier B.V. All rights reserved.
The combustion mechanism of biomass pellets has been studied by
numerous researchers. Li et al. (2012) investigated the reaction mecha-
nism of four typical types of Chinese biomass by TG-DTG and the results
showed the combustion of biomass could be divided into three phases:
(i) water evaporation process, (ii) volatile component combustion pro-
cess, and (iii) the fixed carbon combustion process. According to the
combustion characteristics of biomass, Demirbas (2004) found that
the burning rate of the biomass was much higher than that of coal.
Qian et al. (2011) conducted experiments on the NO emission of bio-
mass. It was noted that NO emission decreased with the excess air,
and increased with the bed temperature. Compared with char-N,
volatile-N is the more dominant reactant source for NO emission.
Ninduangdee and Kuprianov (2014) determined the best combustion
and emission performance of the conical fluidized-bed combustor was
achievable when burning palm kernel shell with a mean particle size
of about 5 mm at an excess air of 40–50%. Khan et al. (2009) pointed
out that the high percentages of alkali (potassium) and chlorine, togeth-
er with high ash content, were the major sources of agglomeration
problems. According the study of Duan et al. (2013) CO emission in-
creased with the in-bed stoichiometric oxygen ratio, but decreased
with excess oxygen ratio.
Despite numerous research devoted to the biomass study, the mech-
anisms and kinetics of the reactions involved in the combustion of the
type of CBPs proposed in this study is not understood as it has not yet
been fully investigated.
2. PROPERTIES OF CBPs
CBPs are made from wood, rice straw and catkins, with real density
of about 1100 kg/m3. High density CBPs are easy to store and can be eas-
ily mixed with the bed materials. As compared with rice husk and straw,
the shape of CBPs is cylindrical and the particle size is larger (5–7 mm)
as shown in Fig. 2. Calorific value of CBPs is 17–19 MJ/kg, less than
24.3 MJ/kg of the coal, but more than 13–15 MJ/kg of the straw. Howev-
er, the biomass may contain alkaline that could be a cause of bed
Fig. 2. Photograph of CBPs and three raw materials.
928 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933
agglomeration (Ma et al., 2016; Gatternig and Karl, 2015; Yu et al.,
2014). For better fluidization, bed materials such as quartz sand or
bauxite particles can be mixed with CBPs. In this paper, bauxite particles
with diameter in the range of 0.5–0.8 mm used as the bed material. The
actual and bulk density of bauxite is about 2760 kg/m3 and 1180 kg/m3,
respectively. The proximate and ultimate analyses of CBPs are reported
in Table 1. Generally, CBPs have a high content of carbon and oxygen,
with a very high level of volatile content.
3. Flow characteristics of CBPs
3.1. Cold-state experimental apparatus
A cold-state test facility is used to study the mixing properties of
CBPs and bauxite particles as sketched in Fig. 3. It mainly consists of a
set of air supply devices, a fluidized bed and pressure sensors. The fluid-
ization reactor is made of Plexiglas and has an overall height of 1.6 m
with a cross section of 120 mm ∗ 40 mm. In order to study the flow of
CBPs in horizontal and vertical direction, first of all, the bed material is
divided into three layers by two horizontal plates, and then each layer
is respectively split into three equal parts as depicted in Fig. 3. By this
way, the binary mixtures in the whole fluidized bed are divided into
nine grids, and the CBPs content in each section is measured. The air
supply system contains roots blower, flow meter and wind chamber.
Gas is provided by the air blower and controlled by flow meter. Pressure
signals are monitoring by pressure sensors at the height of 200 mm
above the air distributor. In the experiment, frequency for each mea-
surement is 100 Hz and more than 1024 data are collected every time.
3.2. Experimental method
Due to the large size of CBPs, the air mainly flows along the gap be-
tween particles, and the CBPs are basically not fluidized. However, the
fluidization behavior of CBPs could be improved when they are mixing
with the bed material. Therefore, the mixing characteristics of CBPs
and bauxite particles are further studied.
The specific operational procedures are as follows: the binary mix-
tures of CBPs and bauxite particles are added into the fluidized bed
with the initial bed height of 200 mm. The pressure measuring device
is opened and the air is injected into the cold-state facility. As the air ve-
locity is increased, the bubbles gradually became larger and led to the
rise of CBPs and bauxite particles, fluidizing the bed. After 2–3 min,
when the fluidization was stable, the air supply was stopped. The mix-
tures was divided into nine sections (1–9) by the horizontal plates
and vertical separators, and, separated by a sieve and weighted. The
mass fraction of CBPs in nine parts was then obtained to study its mixing
behavior. In order to verify the accuracy of experimental data and re-
duce errors, parallel experiments were done for three times and the
error rate was less than 10%.
3.3. Results and discussion
The influence of fluidization velocity on the mixing behavior of CBPs
has been studied in this work. Fig. 4 shows that when the flow velocity
is 0.8 m/s, the blending of CBPs and bauxite particles is not satisfactory.
Table 1
Properties of CBPs.
Proximate analysis (air dry %) Ultimate analysis (air dry %)
Moisture 4.2 Carbon 44.17
Ash 10.1 Hydrogen 5.86
Volatile 72.51 Sulfur 0.03
Fixed carbon 13.19 Nitrogen 0.14
Oxygen 40.16
Note: measured in Southeast University.
At this velocity, the mass concentration of CBPs at the bottom is greater
than the upper layers. This is most probably due to the low turbulence
level at lower speeds, where larger CBPs are less affected by the air
flow and their rise are hindered. With the increase of air velocity, the
turbulence intensity increases, and the mass fraction of CBPs becomes
more uniform along the bed height. As can be seen from the chart, as
the air velocity is increased the mass fraction of CBPs at the bottom of
the bed (first floor) is gradually reduced while that of the upper part
(third floor) is gradually increased, leading to an overall distribution
balance. At the air velocity of 2.0 m/s, the binary mixture is almost even-
ly distributed in the fluidized bed. Further increase of velocity to 3.0 m/s
will lead to a decrease of the distribution uniformity of CBPs. This indi-
cates that when the air velocity reaches a certain value, most of the
CBPs are blown to the upper section of the fluidized bed. However,
when the flow of air is stopped, the bauxite density is heavy and will
fall faster than CBPs, thus CBPs tends to be concentrated in the upper
section of the fluidized bed. The similar phenomenon (segregate with
smaller denser particles at the bottom, whereas the segregation is in
the opposite direction at higher superficial velocities) has been well
known for liquid-fluidized beds.(Escudié et al., 2006)
The critical fluidization velocity umf is an important parameter to be
measured for the fluidization of different binary mixtures. In this work it
was obtained by recording the corresponding air velocity (u) and pres-
sure drop (△P) while gradually reducing the air velocity until the fluid-
ized bed slowly restored to a fixed bed. As shown in Fig. 5, at low air flow
velocities, the material is in a fixed bed state and the bed pressure drop
is proportional to the air velocity. After further increase of air velocity,
the bed begins to expand and produce bubbles, entering into the fluid-
ized phase. After a certain velocity, the pressure drop of the bed layer is
no longer changed. It can be seen from the figure, the critical flow speed
of bauxite alone is 0.6 m/s. The critical fluidization velocity of the mix-
ture with 10% and 20% of CBPs is 0.65 m/s and 0.71 m/s, respectively.
Presence of more CBPs in the mixture will lead to an increase in the crit-
ical fluidization velocity. At the same time, the total weight of mixtures
in fluidized bed will also increase, leading to a larger corresponding
pressure drop. The pressure fluctuation is also more significant due to
the existence of higher amount of large CBPs.
4. Thermogravimetric characteristics of CBPs
4.1. Experimental method
The knowledge of the kinetics of combustion is essential for moni-
toring and improving efficiency of commercial combustion. Therefore,
thermogravimetric (TG) analysis was carried out on the CBP samples
in this work. A sample of 18.84 mg CBPs was used for the experiment
in a SETSYS-1750 CS thermogravimetric apparatus. The air flow was
20 ml/min, and the heating rate was 20 K/min. The sample was heated
from 25 °C to 1000 °C, while its weight loss was continuously recorded.
4.2. Theory
Data from TG and Differential TG (DTG) curves can be used to deter-
mine the kinetic parameters, which are obtained according to the inte-
gral method of Freeman-Carroll. This method has been successfully
used for studies on the kinetics of decomposition and combustion bio-
mass fuels.
The kinetic equation is expressed as:
da=dt ¼ Ae−E=RT f ðaÞ ð1Þ
Assuming, f(a) = (1 − a)n and n = 1, is first-order reaction. To bring
it into Formula (1):
da=dt ¼ Ae−E=RT ð1−aÞ ð2Þ
 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933 929

Fig. 3. Schematic diagram of the cold-state fluidized bed. 1-roots blower 2-header 3-rotameter 4-air distribution chamber 5-horizontal plate 6-fluidized bed reactor 7-pressure
measurement points 8-pressure sensor 9-data collector 10-laptop.

Under the condition of linear heating, the heating rate of w = dT/dt is − RE ; x ¼ T1, Formula (4) can be simplified as: y = a + bx. By taking a se-
brought into Formula (2): ries of points on the TG curve, the corresponding y, x value can be calcu-
lated and a straight line can be made by linear regression method, from
A −E=RT
da=dT ¼ e ð1−aÞ ð3Þ which the slope, b, and the intercept, a, can be obtained. At last, the ac-
w
tivation energy E and frequency factor A also can also be calculated.
Eq. (3) can be transformed by the integration as follow:
     4.3. Results and discussion
− ln ð1−α Þ AR 2RT E
ln 2
¼ ln 1− − ð4Þ
T wE E RT
In Fig. 6, CBPs combustion is divided into three periods; the first
stage is dehydration from 60 °C to 190 °C; the second stage is
The conversion coefficient a is defined as a = (m0 − m)/(m0 − mF),
190–410 °C for volatile combustion phase in CBPs; the third stage is
where m0 represents the initial mass of the sample; m represents the
410 °C to 590 °C mainly for coke combustion. During the three stages,
sample mass in the experimental process; mF represents the mass of
the TG curve shows three weight loss steps, and the DTG curve repre-
the combustion residue; t represents the reaction time; w represents
sents three peaks of weight loss rate: water loss, volatile release and
the heating rate; A is the pre-exponential factor; E is the apparent acti-
combustion and fixed carbon combustion. The peak of volatile release
vation energy; R is the gas constant (8.314 J/(mol·K); T is the reaction
and combustion is much greater than that of fixed carbon combustion.
temperature.
It is mainly because the mass fraction of volatile is larger than the
For most reactions, 2RT/E is much less than 1, and a ¼ ln ½− ln ð1−αÞ

T2 fixed carbon, so the release and burning rate in second stage is higher.
can be regarded as constant. Then if y ¼ ln ½− ln ð1−αÞ
T2
; a ¼ AR
lnðwE Þ; b ¼ The volatilization of CBPs occurs at about 200 °C, and its combustion
process at the early stages is mainly based on the volatile combustion.

Fig. 4. The effect of air velocity on CBPs mass fraction. Fig. 5. The relationship of pressure drop with air velocity.
930 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933
Fig. 6. TG-DTG curve of CBPs combustion.
The combustion process can be described by the combustion charac-
teristic parameters. As shown in Fig. 6, a vertical line is drawn from
point A and intersected the TG curve at point B at which a tangent line
can be drawn on the TG curve, and intersected at the point C from
which the temperature of ignition, Ti, can be obtained. During the com-
bustion process, the maximum weight loss rate is at the maximum
burning rate (dm/dt)rmax, which is determined by the maximum value
on the DTG curve, corresponding to the temperature of Trmax. (dm/dt)-
r
mean is the average burning rate. At DTG value of 0, the burnout temper-
ature Th can be achieved. The ignition characteristic index, Zi(Zi = (dm/
dt)rmax / T2i ) is an indication of how easy or difficult the ignition is. The
greater the value, the easier it is to ignite. The comprehensive combus-
tion characteristic index, SN(SN = (dm/dt)rmax ∗ (dm/dt)rmean / Th / T2i ),
reflects the performance of material ignition and burnout. The greater
the value, the better combustion characteristics of the fuel. For higher
volatile content, the ignition characteristic parameters and combustion
characteristic index are larger. A comparison of all the above parameters
are made with that of coal which has higher fixed carbon in Table 2. As
can be seen the T rmax, Ti and Th of CBPs are 233.2 °C, 212.7 °C and 225 °C,
respectively and lower than that of coal. However, the Zi and SN for CBPs
are much larger than that of the coal. These data indicate that the com-
bustion of CBPs is mainly focused on the volatile combustion, while for
the coal is on the fixed carbon combustion in the later stage. Overall, the
ignition and combustion characteristics of CBPs are better than coal, i.e.
CBPs more easily ignite and burnout. This is mainly because of the vola-
tile component of CBPs being more than that of coal, while its fixed car-
bon is much lower than that of coal.
Table 3 shows the kinetic parameters of CBPs at different combus-
tion stages. The correlation coefficient r was found to be close to 1,
therefore the first-order combustion reaction is a reasonable assump-
tion. The activation energy (E) of CBPs in the volatile combustion
phase is greater than that of the fixed carbon combustion, which is con-
sistent with the literature (Kirtania and Bhattacharya, 2015; de
Caprariis et al., 2015; Soria-Verdugo et al., 2016; Hu et al., 2016; Cai
et al., 2014). However, the activation energy of fixed carbon for CBPs
is less than that of coal (80–90 kJ/mol).
Table 2
Combustibility parameters of CBPs.
Ti/°C (dm/dt)rmax/(%/min) Trmax/°C Zi/%2/(°C·min) Th/°C SN/×10−8
CBPs 320 26.7 367 6.71 600 366.2
Coal 532.7 7.38 604.2 0.17 825 5.86
Note: measured and calculated in Southeast University.
Table 3
Combustion kinetic parameters of CBPs.
Temperature range/°C Equation of liner fit r A/min−1 E/(kJ/mol)
280–410 Y = −2477 × −5.721 0.9945 243.43 20.59
480–570 Y = −1987 × −7.783 0.9659 24.84 16.52
5. Combustion characteristics of CBPs
5.1. Experimental set up
The experiments are carried out in the combustion apparatus which
is illustrated schematically in Fig. 7. The system consists of a fluidized
bed reactor, heating system, gas supply system, dust collectors and an
induced draft fan. The fluidized bed reactor which is made of 316 stain-
less steel has a diameter of 100 mm and a height of 4.5 m, so that the
residence time of CBPS in fluidized bed can be maintained above 2 s,
when the fluidization velocity is 2 m/s. The combustion chamber of die-
sel engines provides additional heat by sending the high temperature
gas produced by combustion is directly into the outer jacket of the flu-
idized bed, and the thermal exchanger, preheating the initial fluidized
air. In order to prevent the possible deformation of spiral feeding rod
causing by high temperature, a water cooling device is also designed.
The operating temperature of the whole combustor could be in the
range of 300–1100 °C. Fluidization air at the bottom enters the reactor
through a 10 mm diameter tube and via 0.1 mm pores on the conical
surface of the distributor. The fluidized bed located at a height of
1.5 m is used for the secondary air. The temperature distribution along
the fluidized bed is measured by type-K thermocouples.
The aim of these measurements is to examine if the fluidized bed can
reach steady-state for CBPs combustion by observing the variation of
bed temperature. Besides, burning of CBPs with minerals involved is
likely to cause the formation of fouling and slagging. Therefore, XRF
analysis on the fly ash is carried out to investigate the matter. The exper-
imental conditions are as follows: bauxite particles with 0.5–0.8 mm
were used as the bed material, while CBPs feeding rate was about
10 kg/h on average. Bauxite particles are firstly fed into the fluidized
bed, and with the static height of 300 mm. When the temperature
reached an appropriate level (700 °C), CBPs were added. By adjusting
the feeding speed of the screw rod, the temperature of the system is
maintained in a certain range (740–780 °C). When the fluidized state
became stable, that is, the temperature change is relatively small, the
temperature profile along the bed height was recorded. The flue gas
generated by the combustion of CBPs is discharged into the chimney
through the cyclone separators and the fabric filter. The collected fly
ash was used for XRF analysis the main pollutants were monitored.
5.2. Results and discussion
The total amount of air required for the complete combustion of 1 kg
CBPs is calculated around 4.0 m3. Consequently, for the feeding rate of
10 kg/h the excess coefficient of air was chosen to be 1.25–2.0, hence
the total volume of air was varied from 50 m3/h to 80 m3/h. The exper-
imental conditions are summarized in Table 4 and the effect of fluidiza-
tion velocity on the bed temperature profiles is shown in Fig. 8.
As shown in Fig. 8, CBPs can burn stability and the temperature of
dense phase is maintained at 765–780 °C.·Because of the heat radiation,
the temperature of dilute phase decreases with the increase of bed
height. With the increase of air velocity from 1.25 to 2.25 m/s, both
the temperature of dense phase and dilute phase increased. At the
same time, the temperature difference between dense phase and dilute
phase was reduced from 35 °C to 10 °C. This shows that with the in-
crease of air velocity, CBPs get more sufficient combustion, releasing
more heat. At the same time the bed materials and CBPs rise further in
the column, and also the combustion region is also moved up, leading
to a decrease in the temperature difference. At the air velocity of
 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933 931

Fig. 7. Schematic diagram of the combustion experimental set-up. 1-blower 2-header 3-diesel tank 4-diesel combustion chamber 5-screw feeder 6-hooper 7-water cooler 8-fluidized bed
reactor 9-heating jacket 10-heat exchanger 11-fluidized gas inlet12-heating gas inlet 13-primary cyclone separator 14-secondary cyclone separator 15-ash bucket 16-fabric filter 17-
induced draft fan 18-chimney.

2.83 m/s, the temperature in fluidized bed reduced. At this time, some of can easily generate stable inorganic sulfate with alkali metal elements
the bed materials and CBPs are blown out because of high velocity, and at the high temperature as shown below.
the combustion efficiency is reduced.
The analysis of the flue gas is carried out on the gas stream exiting SO2 ðgÞ þ 1=2O2 ðgÞ þ H2 OðgÞ þ 2KClðgÞ→K2 SO4 ðgÞ þ 2HClðgÞ ð5Þ
from fabric filter, where specifically O2, CO, CO2, SO2 and NO concentra-
tions are measured simultaneously with MRU600 gas analyzer for emis- The maximum temperature of biomass combustion is reported to be
sion monitoring and combustion analysis (made in Germany). less than 1300 °C,(Bai et al., 2013; Zhao et al., 2014) hence NO is usually
According to the variations of O2 and CO2 concentrations, as shown in accounted for more than 90% of the total produced NOx.(Wang et al.,
Fig. 9, the combustion process can be divided into: (i) start of combus- 2015) Therefore in this paper we focus on the analysis of NO emission.
tion, (ii) stable combustion and (iii) the end of combustion. Tests with Oxidation of fuel nitrogen, called fuel NO.(Yamamoto et al., 2012) As
continuous feed of biomass are carried out and it is also found that the
discharge of pollutants in steady state is similar to that of in Fig. 9
from 600 s to 800 s. At the beginning of combustion, the O2 concentra-
tion begins to decrease, and the CO2 concentration increases. After
reaching point B, the O2 concentration is maintained at about 7%, and
that of CO2 is 12%, which refers to the stage of stable combustion. At
the start of combustion, the emission concentration of SO2 is
50–60 ppm, caused by incomplete combustion. In the stable phase,
the SO2 of CBPs combustion is very low. The level of pollutant emission
is far lower than that of air pollutant emission standards. It should be
noted that the sulfur content of CBPs is low and in addition, the SO2

Table 4
Operating parameters of CBPs combustion.

Case no. Q (m3/h) Excess air ratio Fluidized velocity (m/s)

1 50 1.25 1.25
2 60 1.50 1.75
3 70 1.75 2.25
4 80 2.0 2.83
Fig. 8. The effect of gas velocities on the bed temperature profiles.
932 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933

The calculated ratio of B/A for CBPs in this work is 2.12 and higher
than that of coal (B/A = 0.4–0.7) (Yamamoto et al., 2012). Therefore,
the heated surface of CBPs combustion can be easily deposited with
ash, and it is necessary to consider a blowing mechanism on the
heated surface. However, as bauxite is used as the bed material and
its main material is Al2O 3 , the slagging rate can be reduced
significantly.

6. Conclusions

CBPs are composite product of three kinds of biomass, which can po-
tentially solve the problem of raw material supply cycle. Meanwhile, it
will make full use of the Chinese parasol catkins in Nanjing area.
Experimental results suggest the CBPs and bauxite material can mix
and flow together, and the critical fluidization velocity of binary mixture
is about 0.6–0.7 m/s. Optimum fluidization velocity for better mixing of
Fig. 9. The variation of main gas concentrations in the combustion process.
the system was found to be 1.5–2 m/s.
The combustion of CBPs is found to follow a first-order reaction
model, and it can be divided into three stages: (i) dehydration
the fuel type N of CBPs is volatile at high temperature, and the NO is gen- (60–190 °C), (ii) the release and combustion of the volatile (190–410
erated under oxidization conditions. At point A, NO concentration °C) and (iii) the coke combustion (410–590 °C). CBPs can ignite and
reaches the peak of 120 ppm, where the concentrations of O2 and CO2 burnout and the combustion is mainly based on volatile combustion,
are equal. With the continuous depletion of O2, generating more CO where its activation energy is greater than that of char combustion.
and other reducing gases, NO can be reduced to N2 and causes a de- CBPs can burn steadily and in this work the temperature of dense
crease of NO concentration to 40–70 ppm. With the later increase of phase is maintained at 765–780 °C. With the increase of the air velocity,
O2 concentration at the end of combustion NO also increases, shown the temperature in fluidized bed rises, and the temperature difference
in region C in Fig. 9. between dense phase and dilute phase is decreased from 35 °C to 10
The ash fouling rate of CBPs is closely related to its chemical compo- °C, and shows the main combustion region has a tendency to move
sition. The analysis of ash sample from cyclone dust collector and up. O2 concentration is maintained at about 7%, and CO2 is 12%, while
depicted is presented in Fig. 10. The combustion ash is mainly composed the emission concentration of SO2 is nearly zero. NO concentration
of CaO, K2O and SiO2. CaO accounts for 35% of the ash contents. Its higher first rises to the peak value of 120 ppm, but with the consumption of
component content lowers the ash melting temperature. SiO2 and K2O O2, NO decreases to 40–70 ppm, and finally increases again with the in-
account for 16% and 13% of the ash content, respectively and could be crease of O2. The calculated ratio of alkali to acid is 2.12, suggesting the
responsible for agglomeration and heating surface fouling. The reason heated surface of CBPs combustion can be easily deposited, nevertheless
is that alkali metal oxide reacts with SiO2 to produce low temperature bauxite particles which contain most of the bed materials can help with
co-melt, which leads to serious sintering phenomenon in the fluidized the slagging reduction.
bed. In order to characterize the fouling tendency and the fouling rate
of the heated surface during CBPs combustion, the alkali to acid (B/A) Abbreviations
ratio is introduced,(Sun et al., 2010) which is defined as: umf Minimum fluidization velocity
u Superficial air velocity
△P Bed pressure drop
B=A ¼ wðCaOÞ þ wðK2 OÞ þ wðNa2 OÞ þ wðFe2 O3 Þ
þ wðMgOÞ=wðSiO2 Þ þ wðAl2 O3 Þ ð6Þ TG Thermogravimetry
DTG Derivative thermogravimetry
a Conversion coefficient
m0 Initial weight of the sample (mg)
m Actual weight of the sample (mg)
mF Final weight of the sample (mg)
T Absolute temperature (K)
n Reaction order
A Frequency factor (min−1)
E Activation energy (KJ/min)
r Correlation coefficient
XRF X-ray fluorescence

Notes

The authors declare no competing financial interest.

Acknowledgements

This research is supported by the National Natural Science Fund Pro-

grams of China (U1361115 & 51276040), Scientific Research Founda-
tion of Graduate School of Southeast University (YBJJ1644) and China
Scholarship Council Project. Thanks are due to Ali Hassanpour for assis-
Fig. 10. Analysis results of ash sample from CBPs combustion. tance with language editing and valuable discussion.
 F. Guo, Z. Zhong / Science of the Total Environment 599–600 (2017) 926–933
References
Abdullah, M.Z., Husain, Z., Pong, S., 2003. Analysis of cold flow fluidization test results for
various biomass fuels. Biomass Bioenergy 24, 487–494 (PII S0961-9534(02)00150-
26).
Bai, J., Yu, C., Li, L., Wu, P., Luo, Z., Ni, M., 2013. Experimental study on the NO and N2O for-
mation characteristics during biomass combustion. Energy Fuel 27 (1), 515–522.
Cai, J., Wu, W., Liu, R., 2014. An overview of distributed activation energy model and its
application in the pyrolysis of lignocellulosic biomass. Renew. Sust. Energ. Rev. 36,
236–246.
Chyang, C., Qian, F., Lin, Y., Yang, S., 2008. NO and N2O emission characteristics from a
pilot scale vortexing fluidized bed combustor firing different fuels. Energy Fuel 22
(2), 1004–1011.
Clarke, K.L., Pugsley, T., Hill, G.A., 2005. Fluidization of moist sawdust in binary particle
systems in a gas-solid fluidized bed. Chem. Eng. Sci. 60 (24), 6909–6918.
Cui, H., Grace, J.R., 2007. Fluidization of biomass particles: a review of experimental mul-
tiphase flow aspects. Chem. Eng. Sci. 62 (1–2), 45–55.
de Caprariis, B., Santarelli, M.L., Scarsella, M., Herce, C., Verdone, N., De Filippis, P., 2015.
Kinetic analysis of biomass pyrolysis using a double distributed activation energy
model. J. Therm. Anal. Calorim. 121 (3), 1403–1410.
Demirbas, A., 2004. Combustion characteristics of different biomass fuels. Prog. Energy
Combust. Sci. 30 (2), 219–230.
Duan, F., Chyang, C., Lin, C., Tso, J., 2013. Experimental study on rice husk combustion in a
vortexing fluidized-bed with flue gas recirculation (FGR). Bioresour. Technol. 134,
204–211.
Escudié, R., Epstein, N., Grace, J.R., Bi, H.T., 2006. Layer inversion phenomenon in binary-
solid liquid-fluidized beds: prediction of the inversion velocity. Chem. Eng. Sci. 61,
6667–6690.
Forbes, E.G.A., Easson, D.L., Lyons, G.A., McRoberts, W.C., 2014. Physico-chemical charac-
teristics of eight different biomass fuels and comparison of combustion and emission
results in a small scale multi-fuel boiler. Energy Convers. Manag. 87, 1162–1169.
Fournel, S., Marcos, B., Godbout, S., Heitz, M., 2015. Predicting gaseous emissions from
small-scale combustion of agricultural biomass fuels. Bioresour. Technol. 179,
165–172.
Gatternig, B., Karl, J., 2015. Investigations on the mechanisms of ash-induced agglomera-
tion in fluidized-bed combustion of biomass. Energy Fuel 29 (2), 931–941.
Hu, M., Chen, Z., Wang, S., Guo, D., Ma, C., Zhou, Y., Chen, J., Laghari, M., Fazal, S., Xiao, B.,
Zhang, B., Ma, S., 2016. Thermogravimetric kinetics of lignocellulosic biomass slow
pyrolysis using distributed activation energy model, Fraser-Suzuki deconvolution,
and iso-conversional method. Energy Convers. Manag. 118, 1–11.
Kaynak, B., Topal, H., Atimtay, A.T., 2005. Peach and apricot stone combustion in a bub-
bling fluidized bed. Fuel Process. Technol. 86 (11), 1175–1193.
Khan, A.A., de Jong, W., Jansens, P.J., Spliethoff, H., 2009. Biomass combustion in fluidized
bed boilers: potential problems and remedies. Fuel Process. Technol. 90 (1), 21–50.
Kirtania, K., Bhattacharya, S., 2015. Coupling of a distributed activation energy model with
particle simulation for entrained flow pyrolysis of biomass. Fuel Process. Technol.
137, 131–138.
Li, P., Wang, Q., Xu, Q., Yu, W., Yue, Y., Liang, Z., Dong, X., Hu, S., 2012. Combustion reaction
mechanism of four typical Chinese biomass by TG and DTG. Asia Pac. J. Chem. Eng. 72,
S209–S215.
933
Ma, T., Fan, C., Hao, L., Li, S., Song, W., Li, W., 2016. Biomass-ash-induced agglomeration in
a fluidized bed. Part 1: experimental study on the effects of a gas atmosphere. Energy
Fuel 30 (8), 6395–6404.
Ninduangdee, P., Kuprianov, V.I., 2014. Combustion of palm kernel shell in a fluidized bed:
optimization of biomass particle size and operating conditions. Energy Convers.
Manag. 85, 800–808.
Nunes, L.J.R., Matias, J.C.O., Catalao, J.P.S., 2014. A review on torrefied biomass pellets as a
sustainable alternative to coal in power generation. Renew. Sust. Energ. Rev. 40,
153–160.
Panoutsou, C., Eleftheriadis, J., Nikolaou, A., 2009. Biomass supply in EU27 from 2010 to
2030. Energ Policy 37 (12), 5675–5686.
Qian, F.P., Chyang, C.S., Huang, K.S., Tso, J., 2011. Combustion and NO emission of high ni-
trogen content biomass in a pilot-scale vortexing fluidized bed combustor. Bioresour.
Technol. 102 (2), 1892–1898.
Rao, K.V.N.S., Reddy, G.V., 2010. Cold flow studies of rice husk, saw dust, and groundnut
shell fuels in a fluidized bed. Energy Sources Part A-Recovery Utilization and Environ-
mental Effects 32 (18), 1701–1711.
Saidur, R., Abdelaziz, E.A., Demirbas, A., Hossain, M.S., Mekhilef, S., 2011. A review on bio-
mass as a fuel for boilers. Renew. Sust. Energ. Rev. 15 (5), 2262–2289.
Shao, Y., Ren, B., Jin, B., Zhong, W., Hu, H., Chen, X., Sha, C., 2013. Experimental flow behav-
iors of irregular particles with silica sand in solid waste fluidized bed. Powder
Technol. 234, 67–75.
Si, C., Guo, Q., 2008. Fluidization characteristics of binary mixtures of biomass and quartz
sand in an acoustic fluidized bed. Ind. Eng. Chem. Res. 47 (23), 9773–9782.
Soria-Verdugo, A., Goos, E., Arrieta-Sanagustin, J., Garcia-Hernando, N., 2016. Modeling of
the pyrolysis of biomass under parabolic and exponential temperature increases
using the distributed activation energy model. Energy Convers. Manag. 118, 223–230.
Sun, Z., Shen, J., Jin, B., Wei, L., 2010. Combustion characteristics of cotton stalk in FBC. Bio-
mass Bioenergy 34 (5), 761–770.
Tian, Y.S., 2006. Agricultural biomass energy development strategy in China. China Energy
28 (9), 16–18 (in Chinese with English abstract).
Wang, Y., Wang, X., Hu, Z., Li, Y., Deng, S., Niu, B., Tan, H., 2015. NO emissions and combus-
tion efficiency during biomass co-firing and air-staging. Bioresources 10 (3),
3987–3998.
Yamamoto, T., Kuwahara, T., Nakaso, K., Yamamoto, T., 2012. Kinetic study of fuel NO for-
mation from pyrrole type nitrogen. Fuel 93, 213–220.
Youssef, M.A., Wahid, S.S., Mohamed, M.A., Askalany, A.A., 2009. Experimental study on
Egyptian biomass combustion in circulating fluidized bed. Appl. Energy 86 (12),
2644–2650.
Yu, C., Tang, Z., Zeng, L., Chen, C., Gong, B., 2014. Experimental determination of agglom-
eration tendency in fluidized bed combustion of biomass by measuring slip resis-
tance. Fuel 128, 14–20.
Zeng, T., Weller, N., Pollex, A., Lenz, V., 2016. Blended biomass pellets as fuel for small
scale combustion appliances: influence on gaseous and total particulate matter emis-
sions and applicability of fuel indices. Fuel 184, 689–700.
Zhao, K., Jensen, A.D., Glarborg, P., 2014. NO formation during oxy-fuel combustion of coal
and biomass chars. Energy Fuel 28 (7), 4684–4693.