Relationship between pressure fluctuations and generation of organic pollutants with different particle size distributions in a fluidized bed incinerator

Article abstract of DOI:10.1016/j.chemosphere.2004.04.046

The hydrodynamic behaviors of fluidization perhaps significantly influence the uniformity of fluidization in fluidized bed incinerator. Good uniformity of fluidization expressed the air across uniformly through the bed and the particles being distributed well in the fluid stream. The aggregates, flocs and channels of particles do not happen during fluidization. The Good uniformity will maintain high heat and mass distribution to improve reaction efficiency. These parameters include the height of static bed, gas velocity, mixing and distribution of bed particle, which have rarely been studied in previous investigations. Consequently, this study examines how the hydrodynamic parameters affect the generation of organic pollutants (BTEXs and PAHs) during incineration. The statistical and power spectral analysis of the measured pressure fluctuation during incineration are used to elucidate the relationship between behaviors of fluidization and generation of pollutants during incineration. Experimental results show the organic concentration does not increase with uniformity of fluidization decreasing. The reason may be the explosion of the gas and the consequent thermal shock destroy the coalescent bubbles to form small bubbles again and enhance the efficiency of transfer of oxygen to increase combustion efficiency. Additionally, the mean amplitude and fluidized index of pressure fluctuation similarly vary with the concentration of organic pollutants. These two indices can be used to assess the efficiency of combustion. The four particle size distributions could be divided into two groups by statistical analysis. The Gaussian and narrow distributions belong to one group and the binary and flat the other. The organic concentration of the Gaussian and narrow distributions are lower than that of the other distributions. Consequently, the bed materials should maintain narrow or Gaussian distributions to maintain a good combustion efficiency during incineration.

Full text of DOI:10.1016/j.chemosphere.2004.04.046

Chemosphere 56 (2004) 911–922
www.elsevier.com/locate/chemosphere
Relationship between pressure fluctuations and generation
of organic pollutants with different particle size
distributions in a fluidized bed incinerator
*
Chiou-Liang Lin, Ming-Yen Wey , Han-Tsung Cheng
Department of Environmental Engineering, National Chung-Hsing University, 250 Kuo-Kuang Road, Taichung 402, Taiwan, ROC
Received 1 October 2003; received in revised form 24 March 2004; accepted 22 April 2004
Abstract
The hydrodynamic behaviors of fluidization perhaps significantly influence the uniformity of fluidization in fluidized
bed incinerator. Good uniformity of fluidization expressed the air across uniformly through the bed and the particles
being distributed well in the fluid stream. The aggregates, flocs and channels of particles do not happen during flu-
idization. The Good uniformity will maintain high heat and mass distribution to improve reaction efficiency. These
parameters include the height of static bed, gas velocity, mixing and distribution of bed particle, which have rarely been
studied in previous investigations. Consequently, this study examines how the hydrodynamic parameters affect the
generation of organic pollutants (BTEXs and PAHs) during incineration. The statistical and power spectral analysis of
the measured pressure fluctuation during incineration are used to elucidate the relationship between behaviors of
fluidization and generation of pollutants during incineration.
Experimental results show the organic concentration does not increase with uniformity of fluidization decreasing.
The reason may be the explosion of the gas and the consequent thermal shock destroy the coalescent bubbles to form
small bubbles again and enhance the efficiency of transfer of oxygen to increase combustion efficiency. Additionally, the
mean amplitude and fluidized index of pressure fluctuation similarly vary with the concentration of organic pollutants.
These two indices can be used to assess the efficiency of combustion. The four particle size distributions could be divided
into two groups by statistical analysis. The Gaussian and narrow distributions belong to one group and the binary and
flat the other. The organic concentration of the Gaussian and narrow distributions are lower than that of the other
distributions. Consequently, the bed materials should maintain narrow or Gaussian distributions to maintain a good
combustion efficiency during incineration.
Ó 2004 Elsevier Ltd. All rights reserved.
Keywords: Incineration; PAHs; BTEXs; Pollutant; Pressure fluctuation
1
. Introduction
technology is considered to be effective for disposing of
waste. This approach has many advantages, such as, the
reduction of waste and the recovery of energy, and the
method has become a promising method for treating
waste. The fluidized bed incinerator has been exten-
sively employed to dispose waste due to its advantages
that include good solid mixing, high heat transfer and
large contact surface area, among others (Nam et al.,
2002).
The development of incineration technology has been
the focus of global research for several years; such
*
Corresponding author. Tel.: +886-4-22852455; fax: +886-4-
22862587.
E-mail address: mywey@dragon.nchu.edu.tw (M.-Y. Wey).
0
045-6535/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2004.04.046

912
C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
Nomenclature
F
flatness defined in Eq. (4)
dominant frequency of pressure fluctuations
x
mean value of random signals in pressure
fluctuations (Pa)
random variables
f
p
ꢀ1
(
s
)
xðtÞ
s
/xx
x
f ðxÞ
probability density function of x
duration of sampling (s)
time shift variable in Eqs. (1) and (2) (s)
auto-correlation function of xðtÞ
angular frequency
T
S
skewness defined in Eq. (3)
power spectral density function
random signal values of pressure fluctua-
tions (Pa)
S
x
xx
d
p
average deviation of amplitude of pressure
fluctuations [–]
Some important parameters influence the operation
bubble size to reduce the mixing of bed materials. The
size of particles of the bed materials also affects the
efficiency of reaction and the hydrodynamic behaviors
of fluidization. Chiang et al. (1992) pointed out that the
organics with low molecular weight was generated easily
when the sand bed was stirred unevenly, generating lo-
cally low temperatures (400–500 °C). High temperatures
(700–900 °C) will promote the generation of high-ring
species, such as those with five or six rings. In a lack of
oxygen, the degradation of polymer tends toward a
pyrolysis reaction and the emission of volatile organics
(Zhang, 1998). Therefore, the uniformity of fluidization
directly affects the generation of pollutants. Besides or-
ganic pollutants, the carbon ash is also emitted from flue
gas during incomplete combustion. These particles
provide surfaces for capturing volatilized metals by
heterogeneous deposition and adsorbing organics to
provide active sites to transfer low-ring species to high-
ring (Wey et al., 1996; Durlak et al., 1998; Wey et al.,
1998). The generation of fly ash not only reduces the
efficiency of combustion but also impacts the generation
of organics. Consequently, the effect of these parameters
on the uniformity of fluidization and pollutant genera-
tion must be considered simultaneously.
of the fluidized bed, such as the height of the static bed,
operating velocity of the gas and the particle size dis-
tribution (PSD). Attrition and combustion increase the
number of particles and alter PSD, which may influence
the performance of the reactor, in terms of fluidizing
characteristics and stability of operation (Bemrose and
Bridgwater, 1987). PSD is governed by several para-
meters including the minimum fluidization velocity, the
terminal velocity, the elutriation rate and reaction effi-
ciency. Changes in PSD could affect hydrodynamic
behavior (Ray and Jiang, 1987; Pell et al., 1990). Arena
et al. (1983) and Chirone et al. (1985) demonstrated that
particle sizes were changed not only by fluid-dynamics
but also by thermal shock and combustion. The narrow
PSD reduces defluidization and segregation, but a wide
PSD enhances conversion and fluidity (Gauthier et al.,
1999). Consequently, PSD is an important parameter of
the operation of the fluidized bed reactor.
Many investigators have analyzed pressure fluctua-
tions in the chamber of the gas-fluidized bed to assess
the uniformity of fluidization. The pressure fluctuations
were related to the effects of the formation, coalescence
and escape of bubbles on bed height (He et al., 1997;
Van Der Schaaf et al., 1998). Several important
parameters, such as operating temperature, static bed
height, operating gas velocity and the characteristics of
the bed material, affect the pressure fluctuations and
thereby change the uniformity of fluidization (Lee and
Kim, 1988; Sun and Grace, 1992). Morse and Ballou
According to previous research, the behaviors of
fluidization affect the mixing and reaction efficiency. The
generation of air pollutants is influenced by operating
conditions.However,therelationshipsbetweenthehydro-
dynamic behaviors of fluidization and the generation of
air pollutants have rarely been studied in previous
investigations. Therefore, this investigation focuses on
further evaluating the effects of hydrodynamic behaviors
of fluidization at high temperatures on the generation of
pollutants during incineration. Pressure fluctuations will
be measured and statistics and powder spectra used to
analyze the uniformity of fluidization. The pollutants
are sampled during incineration. Then, the effect of the
uniformity of fluidization on the generation of pollu-
tants during incineration is analyzed. Experimental
studies are performed at over 800 °C on silica sand with
four different PSDs (narrow powder, binary mixture,
(
1951) firstly used pressure fluctuation to evaluate the
uniformity of fluidization. They used a small amplitude
and a high frequency of fluctuation improve fluidization,
and Kai and Furusaki (1985) obtained the same result.
However, several parameters, like operating temper-
ature, amount of excess air, mixing of bed materials and
operating gas velocity, affect not only the hydrodynamic
behaviors of fluidization but also the generation of
pollutants during incineration. In a fluidized bed incin-
erator, the amount of oxygen increases with the amount
of excess air, reducing residence time and increasing

C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
913
flat and Gaussian powder). The effects of the height of
static bed and the velocity of the operating gas are
studied.
similar Gaussian distribution powder all with nearly the
same mean diameter (d
p
¼ 719 lm). These types of PSD
were prepared according to the method suggested by
Gauthier et al. (1999). Table 1 presents the components
of the four types of PSDs. The experimental procedure
started by preheating the sand bed to the desired tem-
perature (800 °C), after which air was passed through
the preheated chamber to increase the temperature. The
minimum fluidization velocity was determined from the
2
. Experimental
2
.1. Experimental procedures
D
p
0
-versus-U diagram. The detail method of measured
Fig. 1 depicts the apparatus used in the experiment,
minimum fluidization velocity refers to Lin et al. (2002).
According to the required gas flow of fluidization, the
different masses of the artificial wastes were prepared
and the excess gas was 40%. Table 2 lists the weights of
artificial wastes (polypropylene, PP) under various
conditions and operating conditions. PP was fed into the
chamber and the pressure fluctuations were simulta-
neously recorded. When the experiment was complete
and the temperature reduced, all bed materials in the
combustion chamber and the fly ash in the cyclone and
the bag filter were collected and then the process was
repeated under other conditions.
including a bubbling fluidized bed incinerator. The
incinerator included a preheated chamber (length 500
mm), a main chamber (height 1100 mm) with an inner
diameter of 100 mm, and an expanding section (height
1
000 mm) with an inner diameter of 250 mm. The
reactor was made of 3 mm thick stainless steel (AISI
10), and was equipped with a stainless steel porous
3
plate with a 15% open area through which to distribute
gas. The temperature was controlled by a PID controller
and three thermocouples were used to measure the
temperature; the temperature was fed back to the con-
troller. The apparatus also included a data acquisition
system (ADVANTECHPC-LabCard PCL-711S and
VisiDAQ Professional Version 3.1) to record the tem-
perature and fluctuating pressure signals in the fluidized
bed. Two cyclones were connected to a bag filter to
collect the fly ash from the combustion chamber.
2.2. Measuring and analyzing pressure fluctuations
Two pressure detectors were used to measure the
difference between the pressure of the sand bed and that
of the freeboard. One end of the tube was covered with a
200 mesh screen to prevent the particles from entering
the tube, and the other end was connected to different
pressure transmitter (Huba Control 692). These pressure
Silica sand was used as the bed material in the
experiment, and had an almost constant density for all
ꢀ
3
sizes (q ¼ 2600 kg m ). Four types of PSD were
p
studied: a narrow powder, a binary mixture, a flat and
10
11
1
3
11
9
1
5
1
2
A
9
8
7
2
3
6
4
Fig. 1. The bubble fluidized bed incinerator. (1) PID controller, (2) blower, (3) flow meter, (4) preheater chamber, (5) pressure
transducer, (6) U manometer, (7) electric resistance, (8) sand bed, (9) thermocouple, (10) feeder, (11) cyclone, (12) bag filter, (13)
induced fan.

914
C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
Table 1
The components of four types of PSD
Type of
PSD
Sieves (lm)
Average
diameter (lm)
The percentage Total weight of different static bed height (g)
of total weight
%)
1
00 mm
200 mm
300 mm
(
Narrow
590–840
350–500
715.0
100
1120
1120
2240
2240
3360
3360
Gaussian
425.0
600.5
770.5
920.0
1095.0
8
25
35
23
9
5
7
00–701
01–840
8
40–1000
000–1190
1
Binary
Flat
840–1000
00–590
920.0
545.0
59
41
1120
1120
2240
2240
3360
3360
5
350–500
425.0
600.5
17
17
5
7
00–701
01–840
770.5
920.0
1095.0
19
23
8
40–1000
000–1190
1
24
dsv ¼ 719 lm
Table 2
The operating conditions for each experiment
ꢀ
1
ꢀ1
Feed rate of PP (g 15 s )
Run
Type of PSD
Static bed height (mm) (H/D)
Influent air (l min ) (Umf
)
1
2
3
4
5
6
7
8
9
Binary
Flat
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
200 (2)
100 (1)
100 (1)
100 (1)
100 (1)
300 (3)
300 (3)
300 (3)
300 (3)
120 (1.3)
113 (1.3)
109 (1.3)
117 (1.3)
138 (1.5)
131 (1.5)
126 (1.5)
135 (1.5)
156 (1.7)
148 (1.7)
143 (1.7)
153 (1.7)
138 (1.5)
131 (1.5)
126 (1.5)
135 (1.5)
138 (1.5)
131 (1.5)
126 (1.5)
135 (1.5)
1.55
1.45
1.40
1.53
1.83
1.70
1.65
1.78
2.08
1.95
1.88
2.03
1.83
1.70
1.65
1.78
1.83
1.70
1.65
1.78
Gaussian
Narrow
Binary
Flat
Gaussian
Narrow
Binary
Flat
1
1
1
1
1
1
1
1
1
1
2
0
1
2
3
4
5
6
7
8
9
0
Gaussian
Narrow
Binary
Flat
Gaussian
Narrow
Binary
Flat
Gaussian
Narrow
PP ¼ polypropylene.
probes, which were 9 mm in diameter and 250 mm in
length, were made of stainless steel and inserted into the
middle of the sand bed. The range of measurement by
flatness and dominant frequency, by statistical and
power spectral analysis.
the different pressure transmitter was 0–1000 mmH
2
O.
2.3. Sampling and analyzing organics
The pressure signals were digitized and recorded by a
data acquisition system. The sampling interval was 10
ms and 8192 points were recorded for each experiment.
These data were analyzed for mean amplitude, skewness,
During incineration, small amount pollutant may
emit during incomplete combustion. Some of these
compounds, volatile organic compounds (VOC) have

C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
915
generated considerable interest because of their harmful
effects. The VOC plays an important role in the forma-
tion of ozone and photochemical oxidants (Lee et al.,
According to Gulyurtlu et al. (2003), there has a
strong correlation between the combustion efficiency
and the PAHs emission. The generation of carbon
monoxide and PAHs had the same tendency. In order to
understand the effect of operation parameters on gen-
eration of pollutants, the BTEXs and PAHs were select
to represent the volatile and semivolatile organic com-
pounds and analyze the relationship between behaviors
of fluidization and generation of pollutants during
incineration.
2
002). However, these compounds bas been reported by
some stack gas and ash residues (Wheatley and Sadhra,
004). The effect of operation parameters was rarely
2
studied on generation of VOCs. The Polycyclic aromatic
hydrocarbons (PAHs) are semivolatile organic com-
pounds those includes two benzene-ring to six benzene-
ring compounds. Single ring compounds are abbreviated
as BTEXs, those contain benzene, toluene, ethyl-ben-
zene and m-, p-, o-xylene. Table 3 lists the classifications
and abbreviations of organic compounds (BTEXs and
PAHs). Amount of the VOC species, BTEXs were the
major compounds (Arion et al., 2001; Lee et al., 2002).
Additionally, several of PAHcompounds are known
carcinogens and mutagens that impair human health.
Due to these compounds are harmful human, 16 PAHs
are restricted by the US Environmental Protection
Agency (EPA).
Flue gas, containing organics, was sampled by iso-
kinetic sampling. US EPA Modified Method 5 was used
for sampling and the sampling apparatus were as de-
picted in Fig. 2. The sampling apparatus included a
stainless sampling probe, a glassy filter holder, impin-
gers, a flow meter and a vacuum pump. The flue gas
flowed through a glassy fiber to sample the fly ash. Then,
it was passed through a cooling tube to cool the flue gas
and the organic compounds were adsorbed by XAD-4
resin. The glassy fiber, which contained fly ash, and
XAD-4 resin were extracted for 20 h by the Soxhlet
extraction process and condensed to 1 ml using a Kud-
erno Danish evaporative concentrator. These two sam-
ples reflected on organic pollutants in solid and gas
phases, respectively. Finally, these samples were stored
at 4 °C. Then, these samples were analyzed using a gas
chromatograph/flame ionization detector. The recovery
efficiency was analyzed to identify the precision and
the accuracy of the analytical procedure. The recov-
ery efficiency of BTEXs and PAHs was approximately
90%.
Table 3
Classifications and abbreviations of organic compounds
(BTEXs and PAHs) in this study
BTEXs
Abbreviation
Single benzene-ring compounds
Benzene
Toluene
Ethyl-benzene
m-Xylene
p-Xylene
o-Xylene
PAHs
Two benzene-ring compound
Naphthalene
3
Nap
2
1
Three benzene-ring compounds
Acenaphthylene
Acenaphthene
Flue gas
AcPy
AcP
Flu
Fluorine
Phenanthrene
Anthracene
Pha
Ant
4
Four benzene-ring compounds
Fluoranthene
Pyrene
FluA
Pyr
8
Benzo(a)anthrancene
Chrysene
B(a)A
Chr
Five benzene-ring compounds
Benzo(b)fluoranthene
Benzo(k)fluoranthene
7
B(b)F
B(k)F
5
6
Six benzene-ring compounds
Benzo(a)pyrene
B(a)P
DbA
Fig. 2. Sampling train for organics. (1) Sampling probe, (2)
fiber glass, (3) thermometer, (4) cooling tube and XAD-4
adsorption tube, (5) 100 ml distilled water, (6) silica gel, (7) flow
meter, (8) vacuum pump.
Dibenzo(a,h)anthracene
Benzo(g,h,i)perylene
Indeno(1,2,3-cd)pyrene
B(g,h,i)P
InP

9
16
C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
time in the sand bed decreases. However, a gas velocity
3
. Results and discussion
maintained at 1.5 and 1.7 Umf yielded similar amounts
of generated organic pollutants. The concentration of
organic pollutants decreases as the bed height increases.
3.1. Distribution of BTEXs and PAHs on gas and solid
phases at various conditions
Fig. 3 present the typical results of organic pollutants
in gas and solid phases. Under all conditions, the dis-
tribution of organic pollutants in the gas and solid
phases were similar. Fig. 3 shows that the BTEXs and
the two-ring PAHs (Nap, Acpy, Acp and Flu) are most
distributed in the gas phase. The most of high-ring
PAHs (three to six rings) are distributed in solid phase.
Chiang et al. (1992) pointed out that a high temperature
of 700–900 °C promoted the generation of five and six-
ring PAHs. Griest and Tomkins (1986) and Durlak et al.
3.2. Dominant frequency and mean amplitude of pressure
fluctuations
The power frequency spectrum is measured by taking
the Fourier transform of the auto-correlation function
of a pressure fluctuation signal. The calculation func-
tions are defined as
Z
T
2
1
u ðsÞ ¼ lim
xðtÞxðt þ sÞdt
ꢀT
2
ð1Þ
xx
T !1
T
(
1998) also implied that the ability of adsorption of
Z 1
organics on fly ash increased with an increase in the
number of benzene rings in the organic components. At
the same temperature and concentration, the high-ring
organic pollutants have a high capacity to be adsorbed
on the fly ash. Wey et al. (1996, 1998) also pointed out
that the simultaneous presence of fly ash and heavy
metal provides active sites for the reaction of the PAHs.
Therefore, when the low-ring PAHs condense on the
surface of the fly ash, the active sites, which contain
carbon or heavy metals, will catalyze the formation from
low-ring species to five and six ring PAHs. Therefore,
most of these high-ring PAHs are distributed in the solid
phase.
ꢀxt
S
xxðxÞ ¼
u ðsÞe ds
xx
ꢀ
1
Z
sm
¼
2
u ðsÞ cos 2pf sds
ð2Þ
xx
ꢀ
sm
Fig. 5 presents the dominant frequency of the pressure
fluctuations. Fig. 5 shows that the dominant frequency
increases with gas velocity, at a low gas velocity. How-
ever, when the gas velocity exceeds 1.5 Umf , the domi-
nant frequency begins to decline. Kai and Furusaki
(1985) revealed that reducing the amplitude of press-
ure fluctuations and the bubble size, and increasing the
main frequency, improved the uniformity of fluidization.
According to earlier studies, the uniformity of fluidiza-
tion is best when the operating gas velocity is maintained
at 1.5 Umf . From Fig. 5, the dominant frequency of the
Fig. 4 illustrates the PAHs and BTEXs concentra-
tions in the flue gas. The amount of organic pollutants
increases with the gas velocity, because the gas residence
4
3
2
1
0000
0000
0000
0000
0
gas phase
solid phase
Organic species
Fig. 3. The distribution of concentration of BTEXs and PAHs in gas and solid phase (Gaussian distribution, 2.5 H/D and 1.5 Umf ).

C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
917
Fig. 4. BTEXs and PAHs concentration in flue gas at different gas velocity and static bed height.
Fig. 5. Dominant frequency of pressure at different gas velocity and static bed height.
four PSDs decreases as the static bed height increases,
implying a decrease in the uniformity of fluidization
because the sources of pressure fluctuations increase to
reduce dominant frequency.
as the velocity of the gas increases (Sadasivan et al.,
1980; Svoboda et al., 1983; Fan et al., 1990; Leu and
Lan, 1990). The increasing mean amplitude represents a
considerable change in the characteristics of fluidization,
perhaps reducing the uniformity of fluidization.
Fig. 6 show the mean amplitude of pressure under
various conditions. According to two-phase theory, the
increase in the mean amplitude is attributable to the
increase in bubble size, when the operating gas velocity
exceeds the minimum fluidization velocity. Meanwhile,
slugs may increase the amplitude of pressure fluctuation
From Fig. 6, the mean amplitude begins to decrease
as the static bed height increases above 2 H/D. Tradi-
tionally, the bubble has a longer residence time in a
higher bed, and may coalesce with other bubbles,
increasing the mean amplitude. However, these results

918
C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
Fig. 6. Mean amplitude of pressure at different gas velocity and static bed height.
reveal that the bubble does not coalesce with other
bubbles to form a large bubble. The plastic combustion
process may affect the behaviors of bubble. Wey and
Chang (1995) considered that the plastics crack to gen-
erate organic vapor, when the organic component
reaches a particular concentration, the volatiles explode
to cause the pressure to changing suddenly. The volatiles
explode to destroy the coalesced bubbles, and form
small bubbles again, which may decrease the mean
amplitude.
height increases. Comparing the skewness with the mean
amplitude and the dominant frequency reveals that the
volatile substances explode during plastic combustion,
destroying the periodic bubble behavior. The coales-
cence of bubbles, breakdown of bubbles and explosion
of volatile substances occur simultaneously, causing the
pressure fluctuations to change suddenly and influence
behaviors of fluidization.
Fig. 8 presents the value of flatness. As the gas
velocity increases, the flatness decreases, showing that
the fluidization is regulated. However, the flatness in-
creases with the static bed height, implying that the sig-
nal distribution of pressure from convergence to width.
This result is consistent with the skewness, implying that
the pressure fluctuations to change suddenly.
3
.3. Skewness and flatness
Skewness and flatness can be used to measure the
extent of symmetry and sharpness in the probability
density function about the mean value. When the pres-
sure fluctuations are normally distributed, the skewness
is 0 and the flatness is 3. Comparing the normal distri-
bution and variations of skewness and flatness at the
same standard deviation reveals that the skewness is
positive or negative according to the direction of
asymmetry about the mean value. The flatness is larger
or smaller than three, according to the extent of the
sharpness about the mean value (Lee and Kim, 1988).
Flatness and skewness are measured by these functions.
3
.4. Fluidized index
In order to describe whether bed behavior is good or
bad, Morse and Ballou (1951) provided a uniformity
index to evaluate numerically the uniformity of fluid-
ization. Shuster and Kisliak (1952) modified this index
for pressure fluctuations. The fluidization index is given
by
pffiffi
Z þ1
2 2dp
1
3
3
Index ¼
ꢁ 100
ð5Þ
S ¼ r
ðx ꢀ xÞ f ðxÞdx
ð3Þ
ð4Þ
pf
p
ꢀ1
Z þ1
Shuster and Kisliak (1952) also pointed out that the
uniformity of fluidization is good when the index is less
than 5. If the fluidization index exceeds 5, then the
uniformity of fluidization is poor. Fig. 9 shows that the
index decreases as the bed height increasing, improving
1
4
F ¼ r
ðx ꢀ xÞ f ðxÞdx
4
ꢀ1
Fig. 7 shows the measured skewness. The skewness in-
creases with the gas velocity, but decreases as static bed

C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
919
Fig. 7. The skewness of pressure at different gas velocity and static bed height.
Fig. 8. The flatness of pressure at different gas velocity and static bed height.
the behavior of fluidization. According to the dominant
frequency and the skewness, increasing bed height in-
creases the sources of pressure and decreases the period
of bubble, which changes are bad for fluidization.
However, these changes of fluidization do not decrease
the efficiency of combustion. Among the results of sta-
tistical analysis and power spectral analysis, the mean
amplitude and fluidized index vary in the same way with
the concentration of organic pollutants. Therefore, these
two indices can be used to assess the efficiency of com-
bustion.
3.5. Comparison of organic pollutants by statistical and
power spectral analysis under various conditions
3.5.1. Static bed height (H/D)
Fig. 4 plots the concentration of organic pollutants
for different static bed heights. Comparing the statistical
analysis and power spectral analysis with organic con-
centration shows that the mean amplitude (Fig. 6) and
fluidized index (Fig. 9) vary similarly with the organic
concentration, implying that the generation of organic
pollutants is strongly related to the mean amplitude and

920
C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
Fig. 9. Fluidized index of pressure at different gas velocity and static bed height.
the fluidized index. According to the preceding discus-
sion, the sources of pressure fluctuation increases and
the fluidization passes through bad periods, causing the
uniformity of fluidization to decrease as the bed height
increases. However, the organic concentration does not
increase with bed height, which result is not as expected.
Other investigators (Rasul et al., 1999, 2000; Rasul and
Rudolph, 2000) stated that when a system of binary
particles with various sizes and densities is fluidized, the
behaviors of different particle tend to segregation. The
large particle primarily feeds at top layer, while the other
particles stay at the bottom layer of the bed. As the gas
velocity is increased, the gradual inversion of layers is
observed. Larger particles that originally stay on the top
will become segregated at the bottom of bed. Conse-
quently, the plastic may fall gradually to the bottom
during combustion. The organic pollutants, which are
cracked from the plastic surface have a long residence
time during which they contact bed materials. The
inversion of layers will promote combustion efficiency
and thus reduce the concentration of pollutants.
plastic begins to combust. The volatiles explode, causing
the pressure in the combustion chamber to change
suddenly, destroying the coalescent bubbles to form
small bubbles again, possibly increasing the exchange of
oxygen. Although the gas explosion causes the pressure
to change and reduce the period of bubble. These
behaviors will decrease the efficiency of reaction in the
operating of fluidized bed. However, in fluidized incin-
eration the efficiency of conversion of oxygen between
bubbles and wastes rises, increasing the combustion
efficiency and reducing the concentration of organic
pollutants.
3.5.2. Gas velocity (U/Umf
)
According to the two-phase theory of fluidized bed,
bubble size increases as gas velocity increases, when the
operating gas velocity exceeds the minimum fluidization
velocity. Meanwhile, the residence time of gas and
oxygen conversion efficiency is decreased. Accordingly,
Fig. 4 plots the increase of the concentration of organic
pollutants with gas velocity. The mean amplitude (Fig.
6) and fluidized index (Fig. 9) have similar tendency.
These results indicate that the uniformity of fluidization
decreases as gas velocity increases.
According to the dominant frequency, the skewness
and the flatness, the signal of pressure fluctuations
gradually distributed extensively as bed height increases.
Traditionally, a rising bubble has a longer residence time
in a higher bed and may coalesce with other bubbles
while rising. At a low gas velocity, the oxygen in a large
bubble is not exchanged with the outside. The oxygen
concentration declines to reduce the efficiency of com-
bustion. However, Wey and Chang (1995) found that in
the combustion of plastics, the surface of the plastics
firstly swales and cracks to generate organic vapor.
When the organic component reaches a particular con-
centration, the volatile substances explode and the
Figs. 7 and 8 plot skewness and flatness, the skewness
increases with gas velocity, while the flatness decreases.
These findings show that the pressure fluctuations have
high periods and distribute from width to convergence.
The effect of plastic combustion on the exchange of
oxygen between bubbles and the emulsion phase be-
comes weaker as the gas velocity increases. Conse-
quently, the residence time associated with gas and
oxygen transfer, decreases, increasing the concentration
of organic pollutants.

C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
921
3
.5.3. Particle size distribution (PSD)
According to statistical analysis and power spectral
again. The gas explosion possibly increases the efficiency
of conversion of oxygen between bubbles and wastes,
increasing the combustion efficiency and reducing the
concentration of organic pollutants.
analysis, four PSDs exhibit similar tendencies of pres-
sure distribution. However, for mean amplitude, flatness
and fluidized index reveal that the four PSDs can be
divided into two groups. The Gaussian and narrow
distributions belong to one group and the binary and
flat distributions belong to the other. The mean ampli-
tudes and fluidized indices of the Gaussian and narrow
distributions are smaller than those of the binary and
flat distributions, implying that the uniformity of fluid-
ization of the Gaussian and narrow distributions is
higher than that of the binary and flat distributions. The
concentrations of organic pollutants generated with
Gaussian and narrow distributions are lower than that
generated with the other distributions.
Among the results of statistical analysis and power
spectral analysis, the mean amplitude and fluidized
index similarly vary with the concentration of organic
pollutants. These two indices can be used to assess the
efficiency of combustion. The mean amplitude and the
fluidized index show that the four PSDs can be divided
into two groups. The Gaussian and narrow distributions
belong to one group and the binary and flat distribu-
tions belong to the other. The uniformity of fluidization
with Gaussian and narrow distributions is better than
that of the binary and flat distributions. The concen-
tration of organic pollutants was generated using
Gaussian and narrow distributions are also lower than
that generating by other distributions. Consequently, the
bed materials should be maintained with narrow or
Gaussian distributions to maintain a high combustion
efficiency during incineration.
Gauthier et al. (1999) indicated that the fluidization
behavior of the binary and flat distributions tends to
segregation. Therefore, the plastics do not mix com-
pletely with the sand bed and appear segregated,
reducing the efficiency of combustion. The narrow
distribution reduced defluidization and segregation,
enhancing the stability of operation. The efficiency of
combustion of the narrow distribution is better than that
of other distributions. For a Gaussian distribution, the
extreme-sized particles represent a small proportion of
all particles, so the behavior of the Gaussian distribution
was similar to that of the narrow distribution. Conse-
quently, the fluidization with the narrow and Gaussian
distributions is uniform, yielding greater mass and heat
transfer and a related decrease in organic pollutants.
The bed materials should maintain narrow or Gaussian
distributions to keep high combustion efficiency during
incineration.
Acknowledgements
The authors would like to thank the National Science
Council of the Republic of China, Taiwan for financially
supporting this research under Contract No. NSC 91-
2
211-E-005-012.
References
Arena, U., D’Amore, M., Massimilla, K., 1983. Carbon
attrition during the fluidized combustion of a coal. AIChE
J. 29, 40–49.
Arion, A., Baronnet, F., Lartiges, S., Birat, J.P., 2001.
Characterization of emissions during the heating of tyre
contaminated scrap. Chemosphere 42, 853–859.
Bemrose, C.R., Bridgwater, J., 1987. A review of attrition test
methods. Powder Technol. 2, 97–126.
4
. Conclusion
This work elucidates the effect of hydrodynamic
parameters of fluidization on the generation of organic
pollutants (BTEXs and PAHs) during incineration. The
relevant hydrodynamic parameters include PSD of bed
materials, the static bed height and the gas velocity. To
measure the pressure fluctuations during incineration,
statistical analysis and power spectral analysis are used
to assess the relationship between uniformity of fluid-
ization and the generation of pollutants in a fluidized
bed incinerator.
Chiang, P.C., You, J.H., Chang, S.C., Wei, Y.H., 1992.
Identification of toxic PAHcompounds in emitted partic-
ulates from incineration of urban solid wastes. J. Hazard.
Mater. 31, 29–37.
Chirone, R., D’Amore, M., Massimilla, M., Mazza, A., 1985.
Char attrition during the batch fluidized bed combustion of
a coal. AIChE J. 31, 812–820.
Durlak, S.K., Biswas, P., Shi, J., Bernhard, M.J., 1998.
Characterization of polycyclic aromatic hydrocarbon par-
ticulate and gaseous emissions from polystyrene combus-
tion. Environ. Sci. Technol. 32, 2301–2307.
In general, the reaction efficiency will reduce with the
uniformity of fluidization declining in the fluidized bed
reactor. However, the concentration of organic pollu-
tants does not increase with uniformity of fluidization
decreasing, which result is not as expected. It may be
that the explosion of the volatiles causes the pressure in
the combustion chamber to change suddenly and de-
stroy the coalescent bubbles to form small bubbles
Fan, Z., Chen, G.T., Chen, B.C., Yuan, H., 1990. Analysis of
pressure fluctuations in
Technol. 62, 139–145.
a 2-D fluidized bed. Powder
Gauthier, D., Zerguerras, S., Flamant, G., 1999. Influence of
the particle size distribution of powders on the velocities of
minimum and complete fluidization. Chem. Eng. J. 74, 181–
196.

922
C.-L. Lin et al. / Chemosphere 56 (2004) 911–922
Griest, W.H., Tomkins, B.A., 1986. Influence of carbonaceous
particles on the interaction of coal combustion stack ash
with organic matter. Environ. Sci. Technol. 20, 291–295.
Gulyurtlu, I., Karunaratne, D.G.G.P., Cabrita, I., 2003. The
study of the effect of operating parameters on the PAH
formation during the combustion of coconut shell in a
fluidized bed. Fuel 82, 215–223.
Rasul, M.G., Rudolph, V., Wang, F.Y., 2000. Particles
separation using fluidization techniques. Int. J. Miner.
Process. 60, 163–179.
Ray, Y.C., Jiang, T.S., 1987. Particle attrition phenomena in a
fluidized bed. Powder Technol. 49, 193–206.
Sadasivan, N., Barreteau, D., Laguerie, C., 1980. Studies on
frequency and magnitude of fluctuations of pressure drop in
gas–solid fluidized beds. Powder Technol. 26, 67–74.
Shuster, W.W., Kisliak, P., 1952. The measurement of fluid-
ization quality. Chem. Eng. Prog. 48, 455–458.
Sun, G., Grace, J.R., 1992. Effect of particle size distribution in
different fluidization regimes. AIChE J. 38, 716–722.
Svoboda, K., Cermak, J., Hartman, M., Drahos, J., Selucky,
K., 1983. Pressure fluctuations in gas-fluidized beds at
elevated temperatures. Ind. Eng. Chem. Process Des. Dev.
22, 514–520.
He, Z., Zhang, W., He, K., Chen, B., 1997. Modeling pressure
fluctuations via correlation structure in a gas–solids fluid-
ized bed. AIChE J. 43, 1914–1920.
Kai, T., Furusaki, S., 1985. Behavior of fluidized beds of small
particles at elevated temperatures. J. Chem. Eng. Jpn. 18,
113–118.
Lee, G.S., Kim, S.D., 1988. Pressure fluctuations in turbulent
fluidized beds. J. Chem. Eng. Jpn. 21, 515–521.
Lee, S.C., Chiu, M.Y., Ho, K.F., Zou, S.C., Wang, X., 2002.
Volatile organic compounds (VOCs) in urban atmosphere
of Hong Kong. Chemosphere 48, 375–382.
Van Der Schaaf, J., Schouten, J.C., Van Den Bleek, C.M.,
1998. Origin, propagation and attenuation of pressure
waves in gas–solid fluidized bed. Powder Technol. 95,
220–233.
Leu, L.P., Lan, C.W., 1990. Measurement of pressure fluctu-
ations in two-dimensional gas–solid fluidized bed at elevated
temperatures. J. Chem. Eng. Jpn. 23, 555–562.
Wey, M.Y., Chang, C.L., 1995. Kinetic study of polymer
incineration. Polym. Degrad. Stab. 48, 25–33.
Lin, C.L., Wey, M.Y., You, S.D., 2002. The effect of particle
size distribution on minimum fluidization velocity at high
temperature. Powder Technol. 126, 297–301.
Wey, M.Y., Chao, C.Y., Yu, L.J., 1996. The influence of heavy
metal on PAHformation during incineration. Toxicol.
Environ. Chem. 56, 35–45.
Morse, R.D., Ballou, C.O., 1951. The uniformity of fluidization
its measurement and use. Chem. Eng. Prog. 47, 199–204.
Nam, W., Kim, J., Han, G., 2002. Photocatalytic oxidation of
methyl orange in a three-phase fluidized bed reactor.
Chemosphere 47, 1019–1024.
Wey, M.Y., Yan, J.T., Wei, M.C., 1998. The major species of
heavy metal aerosol resulted from water cooling system and
spray dryer system. J. Air Waste Manage. Assoc. 48, 1069–
1076.
Pell, M., Williams, J.C., Allen, T. (Eds.), 1990. Handbook of
Powder Technology. Elsevier, Amsterdam.
Wheatley, A.D., Sadhra, S., 2004. Polycyclic aromatic hydro-
carbons in solid residues from waste incineration. Chemo-
sphere 55, 743–749.
Rasul, M.G., Rudolph, V., 2000. Fluidized bed combustion of
australian bagasse. Fuel 79, 123–130.
Zhang, X.J., 1998. The effect of the oxygen fraction on volatile
organic-compound emissions from waste combustion. J.
Environ. Sci. Health A 33, 147–163.
Rasul, M.G., Rudolph, V., Carsky, M., 1999. Segregation poten-
tial in binary gas fluidized beds. Powder Technol. 103, 175–181.