2013 Phuong Nguyen (f0390946) Vrije University of Brussel (VUB) 9/30/2013 Summer Internship Report
2013
Phuong Nguyen (f0390946)
Vrije University of Brussel (VUB)
9/30/2013
Summer Internship Report
Contents 1. OVERVIEW OF THE AUTHOR AND THE INTERNSHIP................................................. 2
1.1 My information ................................................................................................................ 2
1.2 LUT – ERAMUS – IAESTE information ........................................................................ 2
2. VUB OVERVIEW .................................................................................................................. 3
2.1 Vrije University Brussels (VUB) information (http://www.vub.ac.be) ........................... 3
2.2 Supervisors ....................................................................................................................... 3
2.3 Duration of internship ...................................................................................................... 4
3. THE INTERNSHIP DESCRIPTION ...................................................................................... 4
3.1 Customer information ...................................................................................................... 4
3.2 Tasks and Responsibilities ............................................................................................... 5
3.3 Challenges ........................................................................................................................ 5
4. THE THEORY OF THE INTERNSHIP ................................................................................. 5
4.1 Flue gas cleaning .............................................................................................................. 6
4.2 Reactions kinetics ............................................................................................................. 6
4.3 Modeling .......................................................................................................................... 9
5. THE EMPICRITICAL OF THE INTERNSHIP ................................................................... 14
5.1 The experiment procedure .............................................................................................. 14
5.1.1 Experiment setup .................................................................................................... 14
5.1.2 Experiment results .................................................................................................. 16
5.2 Modeling and Athena software ...................................................................................... 17
6. INTERNSHIP EXPERIENCES .............................................................................................. 0
7. BRUSSELS EXPERIENCES .................................................................................................. 0
8. REFERENCES ........................................................................................................................ 1
1. OVERVIEW OF THE AUTHOR AND THE INTERNSHIP
1.1 My information
Surname, First name: NGUYEN Phuong
Background: bachelor’s degree of chemical engineer and master’s degree of energy
environmental engineer.
University: Lappeenranta University of Technology
Student number: f0390946
E-mail: [email protected]
1.2 LUT – ERAMUS – IAESTE information
First, IAESTE will offer traineeship program through website: www.iaeste.fi and
the deadline for application in year round is 4-15th Feb at LUT. After I checked the
traineeship program on website, I found an internship related to environmental
technology of VUB in Belgium. I wrote my cover letter, CV and prepared all documents
needed. I followed the instruction and applied online to the internship I want to get.
Secondly, Ms Niina Juuti from CIMO and IAESTE Finland contacted me after I
can match with the internship requirement. Then I took a month to wait the confirmation
from VUB and received the acceptance letter. Next, I filled mandatory forms with more
information to confirm I took the internship, including flying tickets, insurance,
emergency contacts and so on.
Thirdly, Ms Minna Niemi ([email protected]) who is responsible for IAESTE,
ERAMUS in Lappeenranta University of Technology, contacted me and provided the
necessary information. After I got her advice, I applied to ERAMUS grants and LUT
travel grants to my internship. Ms Minna helped me a lot to prepare document to get
these grants. I received the job description from my supervisor in VUB, Professor Harry
Verelst, I asked to add the internship in my degree. This internship is related to my field
and also new subject, modeling, I can learn to contribute to my knowledge.
Finally, IAESTE in Brussels contacted me to arrange the accommodation and
pick up when I arrive to Belgium. It is wonderful to get traineeship abroad to both
building my knowledge, experience and discovering another country in EU.
2. VUB OVERVIEW
2.1 Vrije University Brussels (VUB) information (http://www.vub.ac.be)
Vrije University Brussels was founded in 1834 by a Brussels lawyer, Pierre
Théodore Verhaegen. There are two parkland campuses in Brussels Capital Region: the
main campus in Etterbeek, the medical campus and the University Hospital in Jette. VUB
is the largest Dutch-speaking employer in Brussels region and has more than 11 000
students in the academic year 2010-2011. VUB has eight faculties, typically Economical,
Political and Social Sciences; Art and Philosophy; Law and Criminology; Medicine and
Pharmacy; Psychology and Educational Sciences; Sciences and Bio-Engineering
Sciences; Engineering; Physical Education and Physiotherapy. VUB has 27 bachelor
programmes, 68 Dutch and English-language master programmes and a wide range of
post-graduate and post academic programmes. There are around 134 doctoral completed
the degree and around 1200 doctoral theses in preparation with 500 of women PhD
students. VUB held 34 patents owned by VUB and 72 patents cooperated with third
parties.
2.2 Supervisors
First, I would like to thank Mr Harry for giving me the opportunity to work with
him. He explained me about the steps of modeling, from the first step to understand the
kinetic to Athena software. He is very friendly to involve me in CHIS activities and I am
familiar with the environment and people in CHIS department.
Contact information: Professor Harry Verelst
Faculty: Sciences
Department: Chemical Engineering
E-mail: [email protected]
Secondly, I would like to thank Mrs Lina for helping me to understand the
experiment and explaining my questions carefully. She always encouraged me in case I
got the mistakes on experiment. I learnt from her sharing and can handle the experiments
by myself.
Conatct information: Dr. Lina Ma
Faculty: Sciences
Department: Chemical Engineering
E-mail: [email protected]
Finally, I would like to thank Ms Wendy De Vleeschouwer who helped me with
administration during the time of the internship and involved me in team building activity
of CHIS department.
2.3 Duration of internship
In 3 months, from 1st of July, 2013 to 26th of September, 2013
3. THE INTERNSHIP DESCRIPTION
3.1 Customer information
Keppel Seghers (http://www.keppelseghers.com/en/) is the environmental
technology branch of Keppel Corporation Limited with activities to provide solutions on
waste management, water solutions and cleaning metal parts. In waste management
solutions, Keppel Seghers deliver projects related to waste-to-energy plants, combustion
systems, energy recovery systems, or flue gas cleaning systems. Keppel Seghers works
under the parent company KIE (Keppel Integrated Engineering Group) with offices in 12
countries and more than 3000 staff members worldwide. (Seghers n.d.) (Steven 2012-
2013)
The dry scrubbing process in a flue gas driven Venturi reactor followed by a
Sonair filter to absorb acid gases hydrogen chloride (HCl) and sulfur dioxide (SO2) by
injecting hydrated lime (Ca(OH)2) is one of the flue gas cleaning system of waste
incinerations of Keppel Seghers waste-to-energy plant in Great Manchester. Keppel
Seghers has asked VUB to understand the kinetics and mass transfer of lime reactions
with acid gases of the flue gas cleaning system in a lab scale packed lime bed reactor.
There are different parameters such as flow rate, acid gases concentration, temperature,
moisture, lime quantity, and so on, to collect as raw data and make mathematical models
based on the unreacted core model. (Steven 2012-2013)
3.2 Tasks and Responsibilities
My first task is testing lime conversion ability to mixture of acid gases HCl and
SO2 by changing parameters affecting the experiment results, typically lime quantity
from 0.1g to 0.25g, also (lime) pellet particle size from 100-160μm to 250-400μm, or
temperature from 130o to 170oC, or moisture from 12% to 14%, HCl concentration from
300 ppm to 1600 ppm, and SO2 concentration from 100 ppm to 400 ppm. I continued the
experiments based on the project work handled by Mr Harry and Ms Lina.
The second task is Athena software learning and computing mathematical models
based on the unreacted shrinking core model developed by Octave Levenspiel. I started to
fit the model with the testing results and found out the diffusion rates, which based on a
modeling of a Master student, Sneiders Steven.
3.3 Challenges - To experiment, I need to understand the project situation during more than a year and
what are results the customer’s expectation. One problem to the system is corrosion and
blockage of connection piping due to high moisture of mixture of acid gases. I learnt how
to find the problem and to start up the system after repairing by the maintenance team.
- To modeling, I didn’t learn in my degree also Athena software. I have only learnt
simulation on Matlab in one course at LUT. Modeling on Athena is a new knowledge I
gained from the internship.
4. THE THEORY OF THE INTERNSHIP
4.1 Flue gas cleaning
Waste-to-energy technology has been developed to waste quantity reduction by
burning solid waste in furnace together energy recovery for electricity generation and
steam. Depending on waste composition and operating parameters of furnace and energy
recovery boiler, flue gas is exhausted harmful substances such as HCl, SO2 into the air.
The necessity of flue gas cleaning is for waste incineration plants to comply with the
strict emission standards. There are different gases cleaning technique, including dry,
semi-wet and wet processes. Dry scrubbing system is a reliable and efficient method for
removal of acid gases. This process can maximum energy recovery from flue gases and
also does not produce waste water as wet process. Dry-scrubbing process has the simple
components, including a dry sorbent storage tank, an injection device as a Venturi
reactor, bag filter and residue hopper and storage tank. In Figure 1, a dry system is shown
for Keppel Seghers’s dry scrubbing system. (Steven 2012-2013) (Lina Ma 2013)
Figure 1: Keppel Seghers dry scrubbing system (Steven 2012-2013)
4.2 Reactions kinetics
Keppel Seghers showed the possible reactions of lime with acid gases in the flue gas
as in Figure 2. However, in the project, the model focused on the absorption of HCl and
SO2 through three most pertinent reactions in the red circle indicated in Figure 2. (Steven
2012-2013)
Figure 2: Reaction kinetic scheme of Ca(OH)2 with acid gases HCl and SO2 for model purposes
According to the results of XRD (X-ray diffraction), it is concluded that the major
products are Ca(OH)Cl, CaSO3.12H2O, CaSO4. .
12H2O, and CaCO3. Keppel Seghers provided the
information of three most pertinent reactions as discussed above with Gibbs free energies (GFE)
as in Table 1. The reaction 1 is twofold including reaction 2 and reaction 3 with total GFE of
-118 kJ/mol. In presence of SO2, the product, CaOHCl, is generated much to occur the reaction 4
and 5 with total GFE of -141 kJ/mol. However, the reaction 2 and 6 are considered as the
reaction pathway of hydrate lime and HCl. Moreover, the reaction of lime and SO2 is preferred
by the reaction 8 and 9 while only the reaction 7 is considered as the reaction pathway of hydrate
lime and SO2. (Steven 2012-2013)
Table 1: Reaction pathways with corresponding Gibbs free energy (Steven 2012-2013)
No. Reagents Products ∆G (kJ/mol)
1 Ca(OH)2 + 2 HCl(g) CaCl2 + 2H2O -118
2 Ca(OH)2 +HCl (g) CaOHCl + H2O -79
3 CaOHCl + HCl (g) CaCl2.H2O -39
4 CaOHCl + SO2 + 1/2O2 CaSO4 + HCl -62
5 CaOHCl + SO2 + 1/2O2 + H2O CaSO4.2H2O + HCl -62
6 CaOHCl + SO2 + 1/2 H2O CaSO3.12H2O + HCl -28
7 Ca(OH)2 + SO2 (g) CaSO3.12H2O + ½ H2O -107
8 Ca(OH)2 +SO2 +1/2O2 CaSO4 + H2O -341
9 Ca(OH)2 +SO2 +1/2O2 + H2O CaSO4.2H2O -338
There are various conclusion of parameters interaction to the rate of mass transfer in both
gas-phase and liquid-phase and they are summarized as following statements:
• ‘The increasing HCl concentration increases SO2 absorption rates’ (Liban Yassin et al.
2007)
• ‘Reaction between lime and SO2 were observed to be slower than those between lime and
HCl in a single-gas system’ (Liban Yassin et al. 2007)
• When HCl concentration is high, the lime conversion is reduced due to the clogging the
external particle surface by CaCl2 formation. (Yan et al. 2003)
• The conversion efficiency of HCl is higher than the conversion efficiency of SO2. The
reason comes from CaCl2 formation and retaining water on the surface. It causes the
increasing the HCl removal efficiency. On the other hand, calcium sulfite layer creates
the unreacted lime particle on the surface. (Liu Z-S et al. 2005)
• The concentration of SO2 is decreased through reactor but can be increased again because
the product CaSO3 can react with the excess of HCl to form CaCl2 and SO2. (Liu Z-S et
al. 2005)
• The removal efficient of SO2 increase with the increasing of the product CaCl2 between
lime and HCl. (Lina Ma 2013)
• Desulfurization activity increases by lime with smaller particle diameters and higher
surface areas. (Lina Ma 2013)
• The conversion of Ca(OH)2 increases with the increasing humidity from 4% the
conversion efficiency without moisture to 55% the conversion efficiency with 11%
moisture. Therefore, the SO2 removal reaction increases with the increasing the relative
humidity. (Lina Ma 2013)
• Fly ash can support hydrated lime to capture SO2 higher than hydrated lime alone. (Lina
Ma 2013)
• O2 and CO2 are reported not affect to the SO2 removal efficient (Lina Ma 2013).
However, CO2 can cause pore blockage on the sorbent (Terence Chin et al. 2005).
• The SO2 removal efficient increases with the increasing Ca/S molar ratio but a too large
of the ratio can’t enhance the SO2 removal. (Lina Ma 2013)
• It is concluded the improvement of SO2 removal efficient at the lowest possible
temperature and higher recycling rate of particles. (S.Kaiser et al. 1999)
4.3 Modeling
The modeling the experiment results is implemented to match the data with a set of
equations described the absorption of HCl and SO2 by hydrated lime. (Paul N.Chisholm & Gary
T.Rochelle 1999)
There are three process steps to control the overall reaction rate of lime and acid gases:
(1) diffusion through the gas film, (2) diffusion through the ash layer, and chemical reaction.
However, in case no ash layer covering the unreacted core of the fixed-size particles, two process
steps are gas-film diffusion and chemical reaction to control the overall reaction rate. (Liban
Yassin et al. 2007)
For instance, the phenomena of HCl reaction with lime is observed with the following
steps: bulk gas diffusion to lime particle surface, chemical reaction of lime and HCl, formation
of product layer, and gaseous diffusion through product layer development (PLD). (Yan et al.
2003)
There are two idealized models for the non-catalytic reaction of particles with
surrounding fluid, especially the progressive-conversion model by the reaction at different rates
and different locations within the particle at all times, and the unreacted-core model with a layer
of ash surrounding the particle. However, when no ash forms and does not contribute any
resistance, the shrinking unreacted spherical particle is occurred as shown in Figure 3.
(Levenspiel n.d.)
Figure 3: The unreacted shrinking core model (Steven 2012-2013) The shrinking core model is still the best simple representation for the reaction of gas-
solid systems despite of limitations of the model such as a gas reacting with a very porous solid
or solid conversion by heat releasing (Levenspiel n.d.).
The shrinking core model describes the rate of absorption of an acid gas by mass- transfer
resistances, including the gas film and the pore diffusion resistances. The limiting rates in gas
absorption are assumed by first-order kinetic and product layer diffusion as the equation below:
(Paul N.Chisholm & Gary T.Rochelle 1999)
Fa = 𝐶𝑎1𝑘𝑠+ 𝛿/𝐷𝑎
With Fa: the flux of A to the sorbent in the fixed bed
Ca: the concentration of species A
ks: the first-order surface rate constant
δ: the product layer thickness
Da: the diffusion coefficient of species A through the product layer
According to previous report, the initial reaction rate is calculated by the slope of the
curve-weight change (%) versus time (min) while the initial global reaction rate RAO as the
equation below: (Rong Yan et al. 2003)
RAO = (𝑑𝑊𝑑𝑡 )
𝐴𝑜 (𝑀 𝑝𝑟𝑜𝑑𝑢𝑐𝑡−𝑀 𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡)
With Ao : the surface area of particle
M product, M reactant: the molecular weight of solid reaction product and solid reactant
Based on the unreacted core model, the mass transfer rate from the bulk gas to the surface of the
solid: (Rong Yan et al. 2003)
RA = ks CBO CAsn
RM = KMA (CAO – CAs)
With RA: the molar rate of reaction of HCl per unit surface area of lime solid B
ks: the chemical rate constant per unit surface area
CBO: the initial molar concentration of lime
CAO: the bulk flow molar concentration of HCl
CAs: the molar concentration of HCl at the surface of the lime
n: the reaction order
RM: the molar rate of HCl reaching the surface of lime per unit surface area
KMA: the mass transfer coefficient in gas film
The rate of reaction is assumed the first order and RM AM = RA AO. (Rong Yan et al.
2003)
To the equation 1 in Table 1, the lime conversion at a specific time as the equation below: (Rong Yan et al. 2003)
Lime conversion (%) = 𝑤𝑒𝑖𝑔ℎ𝑡 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒 𝑎𝑡 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑡𝑖𝑚𝑒 (𝑚𝑔)
(𝑀 𝐶𝑎(𝑂𝐻)𝐶𝑙−𝑀 𝐶𝑎(𝑂𝐻)2)𝑥 (𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡𝑀 𝐶𝑎(𝑂𝐻)2 ) x 100%
The conversion-time equation for fixed-size particles for Ca(OH)2 is shown in Table 2 . (Liban Yassin et al. 2007)
Table 2 : Conversion-Time equations to control overall reaction rate (Liban Yassin et al. 2007)
Reaction Expression
Diffusion through gas film
Diffusion through ash layer
Chemical reaction
With: a: stoichiometric coefficient of gaseous acids
b: stoichiometric coefficient of sorbent
CAg: concentration of gaseous acids in the gaseous bulk (kmol/m3)
De: effective diffusivity (m2/s)
kg: mass-transfer coefficient (m/s)
ks: first-order rate constant (m/s)
MB: molar mass of sorbent (kg/kmol)
R: particle radius (m)
t: time (s)
xB: fractional conversion of sorbent
ρ: density (kg/m3)
θ: time required for complete conversion of unreacted particle into a product (s)
(Liban Yassin et al. 2007)
The HCl-lime reaction is predicted by the grain particle model and the conversion of the
particle is estimated by the following equation: (Claus E.Weinell et al. 1992)
Ds 𝑌
1−𝑌 = 𝑣 𝑟𝑔3𝜌𝑔(𝑟𝑡−𝑟𝑐)
3 𝑊𝑟𝑐 𝑟𝑡 = F(X)
With Ds: the solid diffusion coefficient of HCl Y = CA / CAO
X: the solid conversion at a given time
rt: the unreacted grain radius rc: the unreacted core radius W: mass of unreacted lime particles (kg) v: volumetric gas flow rate (m3/s) ρg: density of the unreacted lime grain (kg/m3) In the same report, Ds increases with the increasing of temperature and is estimated from
the slope of the linear of F(X) with Y/(1-Y). (Claus E.Weinell et al. 1992)
5. THE EMPICRITICAL OF THE INTERNSHIP
5.1 The experiment procedure
5.1.1 Experiment setup
The experiment is implemented through the equipment shown in Figure 4. The
experiment starts with the sample preparation in a packed bed reactor which is made by a quartz
tube. The certain hydrate lime quantity is weighted in an inert chamber and placed in the reactor
between 2 pieces of quartz wool to prevent displacement of lime by the pressure of the flue gas.
The reactor is put inside a cylindrical furnace or oven (2) to have a homogenous distribution of
temperature. The flue gas is mixed in the mixing zone (1) with five gas stream (air or oxygen,
CO2, N2, HCl, SO2) and a water stream. All six gas streams are controlled by flow control valves
to get various concentrations in mixed flue gas. The mixed flue gas can be guided through either
a by-pass or the reactor in a specific condition before analyzed in mass spectrometer (3). (Steven
2012-2013)
To stabilization and correction of the experiment, N2 is purged in the system to ensure a
clean system before the next test, together to prevent the corrosion problems (Terence Chin et al.
2005). In addition, the three-way valve is turned to by-pass piping to have a clear background
and preparing concentration values of the mixed moisture flue gas (Steven 2012-2013).
After all concentrations are stable and accurate as a specific condition, the three-way
valve is turned toward the reactor to start the experiment. The concentration of acid gases, HCl
and SO2, begin to drop and the experiment is stopped when a breakthrough of the acid gases
concentration by time is observed and back near to the input values constantly. (Steven 2012-
2013)
Figure 4: The experiment equipment (upper picture) and the packed bed reactor (lower picture)
(Steven 2012-2013)
The experiment was done with the changing of following parameters during the
internship period:
• Lime quantity is fixed at 0.1g, but pellet size is from 100-160μm to 250-400μm.
• The moisture is from 12% to 14%.
• Temperature is from 130oC, 150oC to 170oC.
• HCl concentration is from 300ppm to 1600ppm and SO2 concentration is from 100ppm to
400ppm. The acid gases ratio HCl/SO2 is various from 0.75, 1,2,3,4, and 5.
• O2 and CO2 concentrations are fixed at 8% of air mixtures with N2 added to obtain 100%
at flow rate 50Nml/min.
5.1.2 Experiment results
The phenomenon of the experiment mentioned in Steven’s report is characterized by four
different areas as in Figure 5: (1) the initial rate of absorption, (2) the changed rate of absorption,
(3) the breakthrough and (4) the saturation of the sorbent. (Steven 2012-2013)
Figure 5: An example of an experiment for four different areas of reaction profile
In the simulated flue gases, HCl concentration is dropped in 50-100 min and raised
slowly to the input value which means the product layer formation and the reaction rate is
controlled by gaseous diffusion. On the other hand, SO2 concentration is also dropped in 10-20
min and the saturation of SO2 is faster to back the input value. According the changing of various
parameters, including lime quantity, relative humidity, temperature, acid gases concentration, the
graphs are drawn to observe the amount of time the lime absorption efficiency or before a
breakthrough occurs in the second area in Figure 5. (Steven 2012-2013)
There are conclusions of behaviors of different parameters to lime the amount of time for
a breakthrough: (Steven 2012-2013)
• A doubling of flow rate has a halving of the time before breakthrough, for
instance, HCl concentration from 500 ppm to 1000 ppm with the time before
breakthrough from 800 min to 400 min, or SO2 from 100 ppm to 200 ppm with
the time before breakthrough from 200 min to 100 min.
• The time before breakthrough is reduced with the smaller of lime quantity.
• Moisture is not affected on SO2 breakthrough time but the time of breakthrough
of HCl is longer at the moisture 12%. Although the moisture is excluded in the
model, it is concluded the moisture up to 12% cause the corrosion problem.
• Temperature does not influence to SO2 conversion while the time of HCl
breakthrough is longer with the lower temperature.
5.2 Modeling and Athena software
5.2.1 Modeling techniques and Athena software
Mathematical models are used for problem solving with the basic stages as Figure 6:
(Aris n.d.)
Figure 6: The basic stages of problem solving (Aris n.d.)
A model is constructed to understand the possible mechanisms for the phenomenon
conceived a purpose and based on theory viewpoint. A model can be formed with some
particular assumptions. A model can be reformulated by changing the assumptions during the
model evaluation phase. When a solution of mathematical model is different equations
formulated in modeling manipulation phase, a model operates to results in given numbers and
these numbers are compared with the experiment observations in modeling evaluation phase.
Model modification can be minimized by the equations dimensionless. (Aris n.d.)
There are computer softwares to implement a model, such as Matlab, Athena. According
to Athena Visio Studio Technical Guide, Athena Visual Studio is software integrated modeling,
estimation, optimization for chemically reactive or non-reactive systems.
5.2.2 Mathematical model development
Problem Identification
A solution is possible?
Model building
Simulation
Pay- off
The multi-core model is used to visualize the way of HCl and SO2 move in the particle.
Due to difference in diffusivity (DHCl > DSO2), two unreacted core can be moved at different
rates, for instance, the dimensionless core radius to SO2 (r’1) is 0.9999m whereas to HCl is
0.9998m. An example of the multi-core model is explained in Figure 7. (Steven 2012-2013)
Figure 7: A multi-core model example (Steven 2012-2013) To formula the model, there are some parameters transferred in their dimensionless form
as these equations below (Steven 2012-2013).
With r’: dimensionless of core radius t’: dimensionless time Φ: dimensionless yield According to the theory of the unreacted core model, the mathematical model is built
based on two first-order equations as below (Steven 2012-2013).
With SdA, SdB: dimensionless stoichiometric coefficient
GA, GB: dimensionless time in pellet
There are some controllable parameters such as initial concentration, temperature, flow
rate and so on, and some non-controllable parameters like diffusion coefficient, density, molar
weight and so on.
The acid gases react with the particle along its path and the particles are near the inlet
absorbed more than the particles close to the outlet. Therefore, the reactor is divided into smaller
volume elements or ‘Parts’ which act as CSTRs and all equations are recalculated to address the
different concentration in the inlet and the outlet. Figure 8 show Parts in the packed bed reactor.
(Steven 2012-2013)
Figure 8: The diagram of a smaller volume element in packed bed reactor (Steven 2012-2013)
5.2.3 Model fitting to data of experiments It was concluded the fitting model to data of experiments with two observations,
including fast breakthrough and non-zero minimum concentration during absorption. The
breakthrough in the model is slower than in experiments results. Diffusion coefficient estimation
from the model to one component such as SO2 or HCl are shown in Figure 9 and 10. They are
some results of the model of single core for SO2 and HCl in various temperatures. (Steven 2012-
2013)
Figure 9: An example of fitting model for SO2 concentrations profile (Steven 2012-2013)
Figure 10: An example of fitting model for HCl concentrations profile (Steven 2012-2013) Following to Figure 11 and 12, they are shown the fitting the model results.
Figure 11: Fitting SO2 in the mixed flue gases (2 components) (Steven 2012-2013)
Figure 12: Fitting HCl in the mixed flue gases (2 components) (Steven 2012-2013)
5.2.4 Reference models and results
As a previous report, the removal of SO2 by an electrostatic spraying absorber (ESA)
with Ca(OH)2 is studied as the following reaction below and got the results of diffusion
coefficients of SO2 in gas phase and liquid phase, and Ca(OH)2 in water as indicated in Table 3.
(Binlin Dou et al. 2008)
Ca(OH)2 +SO2 + H2O = CaSO3.12H2O + 3/2 H2O
Table 3: Diffusion coefficients in the mass transfer through a gas film and a liquid film (Binlin
Dou et al. 2008)
Parameter Value Unit
Diffusion coefficient of SO2 in liquid phase 1.8 x 10-9 m2/s
Diffusion coefficient of SO2 in gas phase 1.4 x 10-5 m2/s
Diffusion coefficient of Ca(OH)2 in liquid phase 1.6 x 10-9 m2/s
According to Liban Yassin et al. (2007), a model is simulated the reactions of hydrate
lime and acid gases mixture in a Venturi reactor of an energy-from-waste combustion plant. In
the report, the diffusion coefficients are reported in range of 10-14 - 10-9 m/s for gas-solid
systems.
The reaction of HCl with Ca(OH)2 is investigated with the specific condition as 10%
humidity, 250 ppm or less HCl in nitrogen and at 100oC. The shrinking core model is used to
describe experimental data and the rate limiting is controlled by product layer diffusion or
chemical reaction, also a first-order reaction is assumed the HCl concentration throughout the
reaction progress. The solid-state diffusion coefficient of chlorine is around 1.78 x 10-14 m2/s and
the gas diffusion coefficient of HCl in air is 2.55 x 10-5 m2/s. (Michael Koch et al. 2005)
Following results of an article, the SO2 diffusion coefficient in gas phase is estimated
around 1.4 x 10-5 m2/s. (Fabrizio Scala & Michele D'Ascenzo 2002)
Moreover, the simulation of the gaseous phase with a cascade of Continuous Stirred Tank
Reactors (CSTRs) is studied for the kinetic modeling of a gas-solid reaction. The diffusion
coefficient in the solid product layer is obtained in the range 10-13 to 10-14 m2/s, and depends on
both temperature and humidity of the gas. In Table 4, it shows the diffusion coefficient in
different temperature of the experiment between HCl and Ca(OH)2 with activation energy 26.8
kJ/mol. In addition, the diffusion coefficient is found in linear relationship with relative
humidity. (Ana M.Fonseca et al. 2003)
Table 4: Diffusion coefficient of HCl and lime reaction in different temperature
Temperature (oC) Diffusion coefficient x 1012 (m2/s)
50 2.5 – 3.7
70 0.91 – 2.4
90 0.23 -1.4
110 0.77
130 0.55
Model Acid gas Changes Sources Concentration Lime quantity Flow rate Temperature Moisture Da/Db
1 component
SO2
Temp
From Reference 1
100ppm 0.1g 50 Nml/min 110oC 0% 1.7E-12 100ppm 0.1g 50 Nml/min 130oC 0% 1.7E-12 100ppm 0.1g 50 Nml/min 150oC 0% 2E-12
From Reference 2
100ppm 0.1g 50 Nml/min 130oC 12% 5E-12 100ppm 0.1g 50 Nml/min 150oC 12% 5.8E-12 100ppm 0.1g 50 Nml/min 170oC 12% 7.2E-12 100ppm 0.1g 50 Nml/min 190oC 12% 7.4E-12
Results
100ppm 0.1g 50 Nml/min 110oC 12% 9.5E-13 100ppm 0.1g 50 Nml/min 130oC 12% 9E-13 100ppm 0.1g 50 Nml/min 150oC 12% 1.15E-12 525ppm 0.1g 50 Nml/min 170oC 12% 9.8E-13 525ppm 0.1g 50 Nml/min 190oC 12% 5.58E-12
Lime quantity
From Reference 1
100ppm 0.1g 50 Nml/min 150oC 0% 1.6E-12 100ppm 0.25g 50 Nml/min 150oC 0% 1.2E-12 100ppm 0.5g 50 Nml/min 150oC 0% 1.6E-12
From Reference 2
100ppm 0.1g 50 Nml/min 150oC 0% 1.2E-12 100ppm 0.5g 50 Nml/min 150oC 0% 1.74E-12
Results 100ppm 0.1g 50 Nml/min 150oC 0% 3.5E-12 100ppm 0.25g 50 Nml/min 150oC 0% NA 100ppm 0.5g 50 Nml/min 150oC 0% 1.75E-12
Moisture
From Reference 2
100ppm 0.1g 50 Nml/min 150oC 0% 3.5E-12 100ppm 0.1g 50 Nml/min 150oC 4% 4.8E-12 100ppm 0.1g 50 Nml/min 150oC 6% 2.45E-12
Results
100ppm 0.1g 50 Nml/min 150oC 0% 3.5E-12 100ppm 0.1g 50 Nml/min 150oC 4% 4.8E-12 100ppm 0.1g 50 Nml/min 150oC 6% 2.5E-12 100ppm 0.1g 50 Nml/min 150oC 12% 3.4E-12 100ppm 0.1g 50 Nml/min 150oC 14% NA
HCl Temp
From Reference 1
500ppm 0.1g 50 Nml/min 130oC 0% 2.5E-11 500ppm 0.1g 50 Nml/min 150oC 0% 3.4E-11
Results 500ppm 0.1g 50 Nml/min 110oC 0% 1.97E-11 500ppm 0.1g 50 Nml/min 130oC 0% 1.89E-11 500ppm 0.1g 50 Nml/min 150oC 0% 2.8E-11
Lime From 500ppm 0.1g 50 Nml/min 150oC 0% 3.4E-11
Based on the table above, the results of fitting the model with the experiments data is shown for the key points are concluded as the
following statements:
- To SO2: diffusion coefficient does not change by the change of temperature, moisture or lime quantity, but to moisture of acid gas,
Da seems lower in comparison with dry acid gas in the same temperature.
- To HCl: diffusion coefficient does not change by the change of temperature, but it is higher with higher lime quantity
- To the mixture of SO2 and HCl: in a dry mixture, diffusion coefficient of SO2 is from 1 to 3E-12 m/s2 same as in a moisture mixture
while diffusion coefficient of HCl is around form 1E-10 to 4E-11 m/s2
quantity Reference 1 500ppm 0.15g 50 Nml/min 150oC 0% 7E-11 500ppm 0.25g 50 Nml/min 150oC 0% 1E-10
Results
500ppm 0.1g 50 Nml/min 150oC 0% 2.8E-11 500ppm 0.15g 50 Nml/min 150oC 0% 5.5E-11 500ppm 0.25g 50 Nml/min 150oC 0% 1.04E-10 500ppm 0.5g 50 Nml/min 150oC 0% 2E-10
2 components
SO2 NA From Reference 1
100ppm 0.1g 50 Nml/min 150oC 0% 3.4E-12
HCl NA 500ppm 0.1g 50 Nml/min 150oC 0% 1.9E-10 SO2 NA
Results
100ppm 0.1g 50 Nml/min 150oC 12% 3.5E-12 HCl NA 500ppm 0.1g 50 Nml/min 150oC 12% 4.35E-11 SO2 NA 100ppm 0.1g 50 Nml/min 150oC 12% 2.5E-12 HCl NA 500ppm 0.1g 50 Nml/min 150oC 12% 2.5E-10 SO2 NA 100ppm 0.1g 50 Nml/min 150oC 0% 2.82E-12 HCl NA 500ppm 0.1g 50 Nml/min 150oC 0% 1.12E-10 SO2 NA 100ppm 0.1g 50 Nml/min 150oC 0% 1.02E-12 HCl NA 500ppm 0.1g 50 Nml/min 150oC 0% 1.88E-11
6. INTERNSHIP EXPERIENCES
The internship gave me a chance to work on flue gas cleaning system as in my degree,
also learn new knowledge related to modeling which can help me in my career. I have an
opportunity and experience to work with Mr Harry, a professor of chemical department, who has
much experience in modeling and also absorption technology application in environment
treatment system. In addition, Ms Lina shared not only her experience of the project but her life
in Belgium also. I met other professors and PhD students in chemical department, and they are
friendly and helped me whenever I needed. Moreover, I knew and made friends with many
students who are trainees from IAESTE in other countries.
7. BRUSSELS EXPERIENCES
I have a wonderful time in Belgium. Beside the time of working in lab, I spent my time to
visit famous places in Brussels with my new friends, especially I joined many events for summer
time, especially National Holiday, music festival, cosmic festival. It is a good feeling and a sweet
memory when I enjoy beer, moules, frites, and chocolate in Brussels.
8. REFERENCES
Ana M.Fonseca et al. 2003, 'A new approach to the kinetic modeling of the reaction of gaseous
HCl with solid lime at low temperatures', Chemical Engineering Science, pp. 3499-3506.
Aris, R, Mathematical Modeling Techniques.
Binlin Dou et al. 2008, 'Flue gas desulfurization with an Electrostatic Spraying Absorber',
Energy & Fuels, pp. 1041-1045.
Claus E.Weinell et al. 1992, 'Hydrogen Chloride reaction with lime and limestone: kinetics and
sorption capacity', pp. 164-171.
Fabrizio Scala & Michele D'Ascenzo 2002, 'Absorption with instantaneous reaction in a droplet
with sparingly soluble fines', AIChE.
Levenspiel, O, Chemical Reaction Engineering (second edition), Oregon State University.
Liban Yassin et al. 2007, 'Study of the Process Design and Flue Gas Treatment of an Industrial-
Scale Energy-from-Waste Combustion Plant', pp. 2648-2656.
Lina Ma 2013, 'Review of using Ca-based sorbent to remove acidic gases generated from
municipal solid waste incineration'.
Liu Z-S et al. 2005, 'Advanced experimental analysis of the reaction of Ca(OH)2 with HCl and
SO2 during the spray dry scrubbing process', pp. 5-11.
Michael Koch et al. 2005, 'Reaction mechanism of a single calcium hydroxide particle with
humidified HCl', Chemical Engineering Science, pp. 5819-5829.
Paul N.Chisholm & Gary T.Rochelle 1999, 'Dry Absorption of HCl and SO2 with hydrated lime
from humidified flue gas', pp. 4068-4080.
Rong Yan et al. 2003, 'Kinetic study of hydrated lime reaction with HCl', Environmental Science
Technology, pp. 2556-2562.
S.Kaiser et al. 1999, 'Modeling a dry-scrubbing flue gas cleaning process', Chemical Engineering
and Processing, pp. 425-432.
Seghers, K, , <http://www.keppelseghers.com/>.
Steven, S 2012-2013, 'Master thesis: Characterization of Ca(OH)2 based systems for the dry
removal of HCl and SÒ from flue gases', Chemical Engineering, Vrije University Brussels.
Terence Chin et al. 2005, 'Hydrated Lime Reaction with HCl under simulated flue gas
conditions', pp. 3742-3748.
Yan et al. 2003, 'Kinetic study of hydrated lime reaction with HCl', pp. 2556-2562.