close

Вход

Забыли?

вход по аккаунту

?

CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process. 14(1/2), pp. 259-276, 2006.
CFD Modeling of Biomass Combustion in
Palm Oil Waste Boiler
Hong-ming Yan* and Teck-poh Lai
Chemical Engineering, School of Engineering and Science
Curtin University of Technology, Sarawak Campus, Miri, Malaysia
A computational fluid dynamics model (using CFX) has been developed to simulate
the hydrodynamic behavior of the fluid and the combustion process in the boiler that
is utilized in palm oil mills in Malaysia. The predicted gas temperature from the
model is quite close to that measured in the boiler. The temperature of the inlet air
has played an important role in the improvement of the boiler operation and in the
control of the combustion process. The temperature profile in the combustion
chamber is not uniform and a very high temperature zone has been found. The baffles
inside the boiler have greatly enhanced the convection heat transfer. Further
refinement of the model, to include the combustion kinetics for palm fiber fuel and
modeling of the heat transfer inside the boiler, will provide more accurate predictions.
Introduction
Malaysia is a leading nation in oil palm production, being the largest producer and
exporter of palm oil in the world. The total annual exports of palm oil and oil palm
products are about RM 22 billion (AustS7.86 billion). Malaysia also generates about
9 million tonnes of empty fruit bunches (EFB), 8.0 million tonnes of fiber, and
3 million tonnes of palm shell every year as wastes [ 11. EFB is not considered for fuel
uses because of its h g h moisture content (65%). Palm oil mills generally have excess
fiber and shell, which are not used and have to be disposed off separately. There are
more than 270 palm oil mills operating in Malaysia that utilize mainly palm fiber and
partly palm shell in their boilers as fuel [ 1,2].
In most palm oil mills, the boiler used is the model “Vickers Babcock TW16” and
its design is based on 1970s technology [3]. Its primary use is to generate steam for
electricity generation from a steam turbine and also the sterilization of palm fruits in a
sterilizer. The boiler operates at 21 bar gauge, and produces 21 tonnes of steam per
hour at saturation temperature. The combustion system is the air swept spout-firing
system and is a semi-fluidized system. In normal operation, palm fiber and palm shell
fed into the boiler are 6.5 tonnes and 2.25 tonnes per hour respectively. Percentage of
excess air used is about 57%. Feed water (about 99°C at 1.3 bar) is pumped into the
steam drum by an electric pump, flows down via the downcomer pipes, and circulates
via the risers back to the steam drum. During this process, steam is generated via heat
~~
* Author for correspondence @an.hong.ming@curtin.edu.my).
259
Hong-ming Yan and Teck-poh Lai
exchanger inside the boiler. Before leaving the steam drum and hence the boiler, the
generated steam at saturated temperature goes through the steam separator where
liquid water is separated mechanically via baffles and perforated plates. Then, it goes
to a non-condensing turbine for power generation, hence there is no condensate return
to the boiler. Meanwhile, the flue gas passes through a cyclone before being released
to atmosphere. No air preheater or steam superheater are installed in the process.
In general, the boiler is divided into two parts; namely, combustion chamber and
convection chamber. In the combustion chamber, there are three feed stages located in
the front side. At the bottom part of the feed stages, there are six secondary air holes.
Primary air is fed from the bottom of the combustion chamber. There are also six
secondary air holes on the rear side of the combustion chamber. These secondary air
holes will help not only to complete the combustion process, but also to create
turbulence flow inside the chamber. There are tubes which contain water mounted on
the wall of the combustion chamber. There is a baffle located between the combustion
chamber and convection chamber. In the convection chamber, there are two baffles
located on the upstream and downstream of the chamber. The heat transfer mainly
occurs in this region by means of convection heat transfer. Tubes are connected from
the top to the bottom of the water drum through the baffles.
The present utilization of palm wastes (palm fiber and shell) as fuel in the boiler is
creating serious emission problem in the industry. Kawser and Farid [ 11 showed that
boilers, using palm waste as fuel at palm oil mills, appear to produce high levels of
dust emissions of up to 11.6 g NmJ, compared to the allowable limit of 0.4 g Nm-3
thus causing serious environmental and health problems (black smoke). In addition,
further studies showed that numerous problems have been encountered in the
operation of the boilers, including salt formation in the tube interior and water carryover in the generated steam, and asWslag formation on the tube exterior [4]. Thus,
improvement and optimization of the boiler operation has become a viable tool to
overcome these problems. The aim is to present a simulation study for boilers using a
computational fluid dynamics approach.
Model Development
(i) Transport Equations
In CFX 5 [5], the transport equations solved are the Navier-Stokes equations with
additional ancillary equations such as the Reynolds stress equation. In order to solve
for these equations, the K-E turbulence model has to be introduced, which is the
standard industry turbulence model. The following governing equations are used in
CFXS.
The continuity equation:
-aP
-+V.(pU)
at
260
=0
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
The momentum equation:
...(2)
The energy equation:
--aph'o' 'a
at
at
+V
(pUhl(,,) = V (RV T ) + S,
... (3)
where h,,, is given by:
I
h,,,,= h + -U
2
. .(4)
*
and h is a function of temperature and pressure.
(ii) Reynolds Averaged Navier Stokes (RANS) Equation
Turbulent flow occurs inside the boiler where it will enhance the heat transfer and
mixing process, therefore, the transport equations are solved with the turbulence
model, However, turbulent flow will produce a flow field, whch is chaotic, and a
highly fluctuating instantaneous velocity. In fact, a three-dimensional flow field is
generated and the flow is dependent on an unknown time function [6].Therefore, the
transport equations above are not solved directly. In order to consider the turbulent
behavior, the Reynolds Averaged Navier-Stokes (RANS) equation is introduced.
RANS is applied to solve the model because the averaged and fluctuating components
are used to solve the transport equations for turbulent flow. This has good agreement
as suggested by Reynolds [ 6 ] .The average component is then given by:
u- =
I
--/,I+&
At
Udt
...( 5 )
where At is a time scale that is large relative to turbulent fluctuation, but small relative
to the time scale for which the equations are solved. Substituting the averaged
component into the above transport equations, the Reynolds averaged equations are
then given as:
ap
v .(pU) =0
at
,.
apu f V . ( p u ~ u ) = V . ( o - p u ) + S , ~
..
f
at
261
Hong-ming Yan and Teck-poh Lai
where pu @ u is the Reynolds stress, which is the non-linear convective term in the
non-averaged equations.
aP
-+ V *(UhIo,+ p u-h - A V T ) = at
at
...(8)
and h,,, is then given by:
I
h,,,,= h + - U ' + k
2
where k is the turbulence kinetic energy given by:
As the model involves solving for multicomponent fluid flow (N2,CO, COz, O2
and CH4), the above equations are further expanded. Multicomponent fluid is a
mixture of fluid where the mixture consists of different components (species) each
with their own physical properties. When solving the equations, the average property
of the fluid, whch is mainly governed by the composition and proportion of each
component, is used. In CFXS, when calculating the property of a component which
contains more than one material, an assumption is made so that the constituent
materials can form an ideal mixture. The materials are assumed to be mixed at the
molecular level. When solving the transport equations for a multicomponent fluid, the
flow of the species is transported and difhsed through the fluid. Hence, in order to
solve for multicomponent fluid flow, the density in the above equations has an
important role. The variation of density greatly affects conservation of the mass
balance. From knowledge of temperature and pressure for the mixture, the density of
the mixture is given by:
where Y, is the mass fraction of each component, and p, is the material density of each
component. To solve for the conservation of energy for a multicomponent fluid, an
additional diffusion term is added to Equation (1 1) as given by:
262
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
The conservation of energy equation for a multicomponent fluid is then expanded to:
The above transport equations have introduced additional ancillary equations such as
the Reynolds stress equation. In order to solve these equations, a turbulence model
has to be introduced.
(iii) Turbulence Model
The K-E model [7, 81 is the standard industry turbulence model where K is the
turbulence kinetic energy and E is the turbulence eddy dissipation. In this model, the
transport equations consist of the continuity equation (1) and the momentum equation:
92
+v
at
( p U 8U )- v ( p U , V U )
..
= V p ’i- V ~ ( , u ~ ~ V iUB) ’
where B is the sum of body forces; peg is the effective viscosity; and p ’ is the
modified pressure given by:
2
and peg is given by:
Pc,,= P +- P!
.. .( 16)
where p, is the turbulence viscosity. In the K - E turbulence model, the turbulent
viscosity is a function of K and E. Thus, p, is then given by:
K-
P, = C,P-
. .( 17)
&
where C, is a constant with a value of 0.09.
2 63
Hang-ming Yan and Teck-poh Lai
The turbulent kinetic energy and turbulence dissipation rate are then defined as:
where CEI,Cc2,6,and 0, are constants.
(iv) Combustion Model
To model the combustion process inside the boiler, an eddy dissipation model is used
for the combustion model due to its simplicity and robustness [8,91. Moreover, the
eddy dissipation model is suitable for a wide range of turbulent reacting flows. The
transport equation used in CFX 5 to model the combustion process is given by:
where the term S,is due to the chemical reaction rate involving component i.
The chemical reaction can be expressed as follows:
?=Ail
r=A.il
..(21)
where k is the elementary reactions involving component i, and Vki is the
stoichiometric coefficient for a component in the elementary reaction k. As Si is the
rate of production of component i, thus, Si can be expressed as:
where R k is the elementary reaction rate of progress for reaction k, and is calculated in
the eddy dissipation model in CFX5. The eddy dissipation model utilizes the concept
that the reactants are mixed at the molecular level and react instantly to produce the
product, where the chemical reaction is relatively fast compared to the transport
process in the flow. That is “when it is mixed, then it is burnt”. In this model, the
reaction rate is assumed proportional to the mixing time, where:
&
rate a k
and k is the turbulence kinetic energy and E is the dissipation.
264
. . .(23)
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
Model Assumption
(0 Geometry of the Model
I
Main Air
Figure I . Geometry of the model.
A detailed geometry of the model was not created due to limitations of the available
computing hardware. Thus, all tubes in the boiler have been omitted while retaining
the basic structure of the boiler (including baffles). The simplified geometry for the
model is shown in Figure 1. The boiler consists of two chambers, namely the
convection chamber and combustion chamber. The fuel (palm fiber) is fed into the
boiler through three feed ducts connected to the side of the combustion chamber. The
feed ducts are named as the first, the middle and the third, respectively, according to
the sequence from the rear to the front. The air is fed into the boiler via three sets of
inlets. The primary air inlet is located at the bottom of the combustion chamber, but
the secondary air is fed into the boiler through both the inlets above the feed ducts and
on the rear side of the combustion chamber. More detailed information concerning the
three-dimensional geometry of the boiler model is available [ 101.
(ii) Primary Air Inlet Duct Size
The primary air inlet duct size is designed with a 5 mm hole diameter, and a distance
of 20 mm between the centers of each hole. With an overall boiler size of 4600 mm x
3074 mm, the total number of primary air holes is 35,35 1. To include this number of
265
Hong-ming Yan and Teck-poh Lai
holes in the modeling is extremely difficult. Hence, the dimension of the primary air
hole is changed while the total area is retained. The diameter of the primary air hole is
then changed to 100 mm with a distance of 300 mm from each other. The air hole size
has a significant effect on the mixing of fuel and air and, therefore, this will be further
compensated for by the ease of mixing of methane gas and air in their molecular state.
The momentum of CH4 can also hrther enhance the mixing between fuel and air.
(iii) Fuel
Palm fiber and shell are wastes produced from the palm oil mills. The proximate and
ultimate analyses of the palm fiber are given in Table 1. This version of CFX-5
available cannot simulate the combustion process of solid fuels, hence methane gas
(CH,) is chosen in the simulation to infer palm fiber. The amount of CH4 used is
based on the heating value of fiber as given by:
Mass of CH4
=
...(24)
Y(MW,,)
HC".
where Hfis the heating value of fiber (Jkg); F is the feed rate of fiber (kg/hr); H C H is~
the heating value of methane ( J h o l ) ; and M W C H ~is the molecular weight of
methane (kg/kmol). However, a direct use of Equation (24) may produce some
significant discrepancies in the final result, as methane gas and fiber are not
physically identical. Therefore, some fine-tuning was implemented so that the
combustion mechanism can take into account this physical effect. The ash inside the
fiber may absorb a fraction of heat, and thus may reduce the total energy released.
With this information, the amount of CH4 is further reduced (corresponding to the
heat absorbed) so that the temperature rise in the boiler is reasonable with reference to
the experimental data.
Ash and residue contents of the fiber are also substituted by inert materials such as
COz or Nt when modeling the combustion process with CFX-5. Thus, the reduction in
CH4, and the addition of COz or N1,are used to compensate for the physical effects of
CHI and fiber. The fuel composition used in the simulation is CH4 20%, C02 40%,
and H 2 040%.
Tablel. Proximate and ultimate analysis ofpalm fiber.
Ultimate analysis
Hydrogen
Carbon
Surfur
Nitrogen
Oxygen (by d$)
Ash
Total
266
(wt%)
6.00
47.20
0.30
1.40
36.70
8.40
100.00
Proximate analysis (dry)
Volatile matter
Fixed carbon
Ash
Total:
Heating value (MJkg)
(Wt%)
72.80
18.80
8.40
100.00
20.35
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
Table 2. Feed conditions
Feed
Fuel
Primary air
Secondary air
Mass jlow rate (kg/hr)
2,414
48,279
20,69 1
Temperature (K)
298
298
298
Pressure (bar g)
1
1
1
(iv) Feed Conditions
The total mass flow rate of CHI at the feed inlet is 2414 k g h . The ratio of the excess
air is 1.5, the total primary air feed rate and secondary air feed rate are 48279 kglhr
and 20691 k g h , respectively. Both fuel and air are fed into the boiler at atmospheric
pressure. At the boiler outlet, the pressure approaches zero gauge pressure. The feed
conditions are given in Table 2.
(v) Meshing
CFX-5 provides automatic mesh generation, which includes unstructured triangular
surface mesh generation and tetrahedraYprismatic volume mesh generation. There are
three modes of volume meshing in CFX-5, including Advancing Front and Inflation
(AFI), Paving and Isomeshing, and Volume Mesh Import (VMI) [ 5 ] . Mesh generation
is the process by which spatial discretisation of the model is accomplished. By
default, CFX-Build meshing is based on trianguladtetrahedral element discretisation,
which is used in the current modeling work. In CFX-5, the mesh is prepared in two
stages [ 5 ] :
Interactive facilities within CFX-Build are used to generate the surface mesh
of triangular elements.
0
The volume mesh is generated from the surface mesh during the creation of
the GTM file.
The most common method for meshing is the AFI method in CFX-5. It is suitable
for most geometries. The AFI method is a region-based mesh generation application.
This method uses all the current mesh control settings to determine the appropriate
size of the mesh in a particular region. In general, the element size is determined by
the minimum length scale from all mesh controls, the local length scale from surface
mesh parameters and global length scale [ 5 ] . To determine mesh parameters correctly,
the maximum edge length for volume mesh elements is specified as its default value
of 14. Generally, the maximum edge length is used to provide a maximum
background volume mesh scale. The default is a value between 1-5%of the maximum
model dimension.
Special consideration should be given to the fuel feed region, where mesh control
is required for better meshing around that region. When creating mesh control, the
constant surface mesh spacing is determined so that it will set the mesh control edge
length to a fixed value of 0.6. The expansion factor is defined as 1.2, which
determines how fast the mesh edge length scale is expanded from the edge of the
radius of influence into the background volume mesh.
267
Hong-ming Yan and Teck-poh Lai
(vi) Solution Convergence
In CFXS-, the Equations (1)-(23) are solved by the finite volume method. Starting
everything from zero, trial and error is required in order to obtain a satisfactory result.
It is often difficult to model the reacting flow when compared to the non-reacting
flow, Typically, the time required for non-reacting flow is small and usually the
number of iterations should be less than 100. However, when the reacting flow is
modeled, the complexity not only increases with time but also is very dependent upon
the number of iterations. In order to obtain reasonable and accurate results, the root
mean square (RMS) value calculated is controlled to less than
The smaller the
value of RMS, the greater the time required to perform the calculation. However, the
physical time step has to be defined so that a robust calculation can be provided. In
this case, the time step is 2 seconds as a base value to allow better convergence of the
reaction rate. This value can be fixther reduced to achieve better resolution of the
simulation conditions.
Results and Discussion
(0 Temperature Profile
Figure 2 shows that after the fuel enters the boiler, the mixing and preheating occurs
first at the bottom section. The combustion process then progresses instantly when the
mixing process occurs and the gas mixture temperature is increased. This
phenomenon can be seen from the temperature profile in Figure 2, that is, the region 1
(approx. 850°C) around the region 2 (approx. 25°C). At the interface between the
regions 1 and 2 is where the molecular level of each reactant species interacts with
each other and where the combustion process occurs. Note that the region 3 (over
1500°C) is the highest temperature zone.
Other regions at different temperatures are due to the cooling process of the feed
air and feed fuel, which are not preheated. A significant difference between the
temperature profiles at the first and third feed ducts is shown in Figures 2a and 2c,
where they should be the same. This is probably due to the inappropriate location of
the boiler feed ducts, or because of the flame behavior where the shape and direction
are unpredictable.
A large region 1 is seen in Figure 2b in the area around the second feed duct
because primary feed air is concentrated at the center of the boiler. A relatively high
temperature region occurs at the rear side of the combustion chamber near the
secondary process air inlet. It is proposed that this effect is caused by the high
momentum of the hel. Methane gas then reaches the rear side of the combustion
chamber, mixes with air, and combusts without further cooling by the incoming air.
However, thls does not apply to biomass where it will be further retarded by the
gravitational force before it reaches the rear of the combustion chamber. Combustion
gases produced at high temperatures have a lower density compared to the air flowing
into the boiler. Due to the variation in density, buoyancy forces result in gases
flowing upward [8]. Hot gases that are flowing upward in the boiler will eventually
create an induction flow and entrain the surrounding fluids.
From these results, the temperature is found to be uniform at the convection
chamber (see region 4, approx. 1100 - 1300°C) when compared with the combustion
268
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
Figure 2 a Firstfeed duct temperature profile.
Figure t b . Second feed duct tenipernture projle.
[Note: A reprint of this paper with coloured regions shown on Figures 2, 3 and 4 is available
directly froin rlie author, HMY, by etnail as a PDFfile.1
269
Hong-ming Yan and Teck-poh Lai
Figure 2c. Thirdfeed duct temperature profie.
chamber. This is significant, as it ensures uniform heat transfer in the convection
chamber. The combustion gas temperature has also dropped from a high temperature
(near flame temperature, the region 1) to a temperature of around 120OOC (region 4).
This is due to heat exchange between the hot gas and cold feed air and methane gas as
previously discussed. After this point, the temperature of the inlet air will have an
important influence on the temperature profile in the boiler. Thus, the amount of inlet
air is an essential parameter for boiler operation and control of the combustion
process. The predicted combustion gas temperature is quite closed to the design
temperature (12OOOC).
The uniform temperature profile in the convection chamber is because heat
transfer between hot gases and the heat exchanger inside the convection chamber is
not considered in the model, as mentioned previously.
(lo Velocity Profle
The velocity profile of the combustion gases in the boiler is shown in Figure 3a. The
first secondary air, which is fed into the boiler just above the feed duct, does not
contribute significantly to the mixing process. However, the other secondary air fed
into the combustor from the opposite side to the feed duct at the lower position has
made direct contact with the fuel flowing down from the feed duct, and both are
diverted upward by the primary air. Therefore, the presence of the secondary air at the
2 70
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
lower position does not assist in completion of the combustion process, but rather
assists in the direct mixing process together with the primary air. Figure 3b shows the
gas streamlines, it can be seen that swirl flows are generated inside the boiler, and
circulation is taking place at the bottom of the combustion chamber and convection
chamber as well as between the baffles.
The vortex (see Figures 3b and 3c) created inside the convection chamber of the
boiler is due to the turbulent flow caused by the presence of baffles. Intuitively the
circulation generated at the bottom of combustion chamber will improve the mixing
of he1 and air. In addition, the flow is entrained inside the combustion chamber rather
than causing vigorous mixing [8]. As discussed, the buoyancy generated by the hot
combustion gas causes the gas to flow upward to the convection chamber. The gas
flows from the combustion chamber to the next chamber by passing through the
baffles, and the presence of baffles has enhanced turbulent flow of the hot gases. It is
interesting to note that whenever the fluid flows through the edge of the baffles, a
circulation region is generated which enhances the heat transfer process. Circulation
is caused by the low-pressure generated by the turbulence flow.
The vector profile shown in Figure 3c allows a better analysis of the
hydrodynamic behavior inside the boiler. It shows that there are four circulation flows
generated in the combustion chamber. Based on this flow behavior, once CH4 has
entered the combustion chamber, then mixing starts at the lower right comer. The
combustion process begins after this mixing, and the flame then moves upward by the
buoyancy force of the hot gases. In fact, the flow pattern in the combustion chamber
is not a major issue as radiation heat transfer is the dominate heat transfer mechanism.
In the convection chamber, the circulation generated at the lower right comer does
not contribute significantly to the heat transfer and the air accumulating in this region
may act as a cooling medium. In the case of any entrained flow for the biomasses,
accumulation of the solid particles may also occur and the heat transfer will be
retarded by conduction rather than convection. This may subsequently affect the heat
transfer performance in the convection chamber.
(iii) CH,Mass Fraction Projile
The mass fraction profiles of methane at different feed ducts are shown in Figures 4a
and 4b. Overall, the simulated combustion of methane gas in the boiler has not
reached completion where the mass fraction of methane gas remaining is about 20%.
It appears that the overall performance of the boiler is poor. However, it should be
noted that the design of this boiler is not intended for gas combustion but for solid
combustion. The incomplete combustion process of methane gas is due to an
insufficient amount of oxygen supplied, which is based on the real air flow rates to
the boiler as shown in Table 2 . The high fraction of methane gas at the top comer of
the combustion chamber is caused by the circulation flow where methane gas is
accumulated. This occurs in the first and second feed ducts as shown in Figures 4a
and 4b respectively. The primary air at the sidewall causes a large region 1 at the
bottom of the feed duct. The region 2 between the regions 1 and 3 is where the fuel
and the process air are mixed and the combustion process starts. The region 3 is
slowly transformed into the region 4 as more methane gas is combusted.
271
Hong-ming Yan and Teck-poh Lai
Figure 3a. Boiler velocity profile.
Figure 3b. Boiler streamline plot.
2 72
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
Figure 3c. Boiler vector plot.
Figure 4a. CH,mass fraction profile at first feed duct.
2 73
Hong-ming Yan and Teck-poh Lai
Figure 46. CH4mass fraction profile at second feed duct.
.
7-
Figure 5. CH4 inass fraction versus vertical distance (m) in the combustion chamber.
2 74
CFD Modeling of Biomass Combustion in Palm Oil Waste Boiler
There are considerable similarities between Figures 4a and 4b, the second feed
duct is not restricted by any space confinement and the region 1 is smaller. Once
again, the secondary air above the feed duct has aided the combustion process as
methane gas is flowing upward. In Figure 4b, as there is no circulation, the
momentum of methane gas whch exceeds the momentum of primary air has created
the region 3 at the surface of the primary process air holes. The primary air comes out
of the holes, and the mixing process commences instantly.
To further analyze the combustion process of methane gas in the boiler, Figure 5
shows the methane mass fraction from the bottom of the combustion chamber to the
top, at a position close to the secondary feed stage. Figure 5 shows that at the bottom
of the combustion chamber (distance is negative about 1.2 m), the mass fraction of
CH4 is zero since there is pure primary process air. As the vertical distance increases,
the CH4 mass fraction also increases because pure methane gas is feeding into the
chamber. The CH4 mass fraction will never become 1.0 as some CH4 is mixed with
the surrounding process air and combusted before reaching the bottom of the boiler.
The mass fraction then decreases as methane gas is burned, and eventually the
combustion process stops due to:
Insufficient molecular mixing between methane gas and process air;
Insufficient ignition energy to reach the ignition temperature for methane gas.
Conclusions
A computational fluid dynamics model (using CFX-5) has been developed to simulate
the hydrodynamic flow fields inside the boiler in palm oil mills by including the
combustion reaction. The predicted combustion gas temperature from the model is
quite close to the design operating temperature of the boiler. The temperature of the
inlet air has plays an important role in the operational improvement of the boiler and
assists in controlling the combustion process. There is a considerable difference
between the temperature profiles at the first and third feed ducts of the boiler.
The temperature profile in the combustion chamber is not uniform, there is htgh
radiation intensity and radiation is the dominant heat transfer mechanism. However,
the temperature profile is uniform in the convection chamber and there is stable heat
transfer. The presence of baffles inside the boiler results in the higher velocity of the
combustion gases in the convection chamber. The hydrodynamic behavior in the
convection chamber may not be satisfactory because tortuous flows, which would
enhance heat transfer, are not formed. The CH4 mass fraction profiles showed that the
combustion process could not go to completion and at the boiler exit there is still 20%
CH4 remaining. Hence, improvements are required to enhance the combustion
process, e.g. the boiler should be modeled using solid fuel and including water tubes.
Acknowledgments
The authors acknowledge Mr Wong Lin Siong, mill manager, at Bintulu Lumber
Development Palm Oil Mill, Sarawak, Malaysia, for technical information related to
the boiler. Curtin University of Technology (Sarawak) research grant scheme is also
acknowledged.
2 75
Hong-ming Yan and Teck-poh Lai
Nomenclature
k-E turbulence model constant (1 -44)
k-E turbulence model constant (1.92)
k-E turbulence model constant (0.09)
Specific heat capacity at constant pressure (Jkg K)
Reynolds stress model constant (1.45)
Reynolds stress model constant (1.9)
Specific static (thermodynamic) enthalpy (Jkg)
Specific total enthalpy (Jkg)
Turbulence kinetic energy per unit mass (Jkg)
Static thermodynamic pressure (Pa)
Modified pressure (Pa)
Reynolds number
Energy source (Jkg)
Momentum source (kg d s )
Static temperature (K)
Velocity of vector Uxyz ( d s )
Fluctuating velocity component in turbulent flow ( d s )
Molecular weight, ideal gas fluid mixture (kg/mol)
Molecular diffbsion coefficient of component
Mass fraction of component i in the fluid
Turbulence dissipation rate (m2/s)
Molecular (dynamic) viscosity (Pas)
Turbulence viscosity (Pa.s)
Effective viscosity (Pa.s)
Density (kg/m3)
Prandtl number
k-E turbulence model constant (1 .O)
k-E turbulence model constant (1.3)
References
Kawser, J.M., and Farid Nash, A. 2000. Oil Palm Shell as a Source of Phenol, J. Oil Palm Research,
12(1), 86-94.
2. Ministry of Primary Industries Malaysia, hflpr//www.kpu.gov.my/commo~j~/commo~i~~.
htm
3. Vickers Hoskins (M) Sdn Bhd, 2001. htfpt//winv.vickershoskins.com.ni~//Hi./ifnr/
4. hi, T.P., and Yan, H.M. 2003. Internal Industrial Survey Report on Palm Oil Mills, Chemical
Engineering, Curtin University, Sarawak Campus, Malaysia.
5 . CFX-5. CFX International, AEA Technology, Harwell Laboratory, Didcot, Oxfordshire, UK.
6. Baukal, C.E. 2000. The John Zink Combustion Handbook, CRC Press, Florida, USA.
7. Turns, S.R. 2000. An Introduction to Combustion - Concepts and Applications, 2nd Edn, McGrawHill, Singapore.
8. Bonnan, G.L., and Ragland, K. W.I998 Combustion Engineering, McGraw-Hill, Singapore.
9. Fletcher, D.F., Haynes, B.S. Christo, F.C., and Joseph, S.D. 1999. A CFD-based Combustion Model
of an Entrained Flow Biomass Gasifier, Applied Mathematical Modeling 24, Australia, pp.165-182.
10. Goh, Y.B., and Yan, H.M. 2003. Creation of Three-Dimensional Geometry Model for the Steam
Boiler in Palm Oil Mills, Internal Research Report, Chemical Engineering, Curtin University Sarawak
Campus, Malaysia.
1.
2 76
Документ
Категория
Без категории
Просмотров
0
Размер файла
2 245 Кб
Теги
palm, cfd, waste, biomasa, boilers, modeling, oil, combustion
1/--страниц
Пожаловаться на содержимое документа