Journal of Cleaner Production 136 (2016) 133e149

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Eco innovation strategies for promoting cleaner cement

manufacturing
Siti Aktar Ishak a, b, Haslenda Hashim a, b, *, Tan Sie Ting a, b
a
Process Systems Engineering Centre (PROSPECT), Research Institute for Sustainable Environment, Universiti Teknologi Malaysia (UTM), 81310, UTM, Johor
Bahru, Johor, Malaysia
b
Faculty of Chemical and Energy Engineering, Universiti Teknologi Malaysia (UTM), 81310, UTM, Johor Bahru, Johor, Malaysia

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

Article history: Carbon dioxide (CO2) emissions in cement plants are generated by the decarbonation of raw materials
Received 11 January 2016 and fuel combustion in the cement kiln during the cement clinker production process and account for 8%
Received in revised form of global emissions. This paper presents a mixed integer linear programming (MILP) by considering
2 June 2016
various mitigation measures, such as co-processing of fuels, kiln improvements and carbon capture and
Accepted 3 June 2016
Available online 9 July 2016
storage (CCS) yields, that can have substantial benets. The benets include cleaner cement production
with minimum production costs, while satisfying the quality standard, carbon reduction target, and fuels
substitution rate. The developed model is applied to a case study in order to demonstrate the applica-
Keywords:
Mitigation
bility of the model. For the base case, the optimal cost for clinker production is USD 90.21/t clinker while
Cleaner production CO2 emissions generated from both calcination of raw materials and fuels combustion is 531.68 kg CO2/t
Carbon reduction clinker and 325.00 kg CO2/t clinker. It was found that the highest possible CO2 emissions reduction that
Process improvement can be achieved by a combination of co-processing, kiln improvements and CCS technology is 79%, with
Optimisation an increment cost of USD 136.46/t clinker.
2016 Published by Elsevier Ltd.

1. Introduction The decarbonation reaction of raw materials - normally lime-

Cement is an important component of concrete as it is consid-
ered to be a binder that holds concrete mixture together and gives
it strength. Despite the fact that it is an energy and emission
intensive industry, the cement industry is essential for the eco-
nomic development of a country. The manufacturing process for
the cement industry consists of 3 major steps: raw material prep-
aration, clinker production, and cement production. In raw material
preparation, quarrying is done rst, then followed by pre-
homogenization and grinding of raw materials. During clinker
production, burning of fuels to provide heat and chemical reaction
occurs in a cement kiln. A chemical reaction between pre-
homogenized raw materials and fuels' ash in the cement kiln pro-
duces clinker that is then stored in clinker silos. During cement
production, blending of clinker with grinding aids for nal
adjustment occurs, followed by storage, then shipment. Fig. 1
shows the general dry and wet manufacturing process.
* Corresponding author. Process Systems Engineering Centre (PROSPECT),
Research Institute for Sustainable Environment, Universiti Teknologi Malaysia
(UTM), 81310, UTM, Johor Bahru, Johor, Malaysia.
E-mail address: haslenda@cheme.utm.my (H. Hashim).
stone (conversion of limestone to lime) or calcium carbonates
(CaCO3) rich materials in cement kiln - contributes to about 50% of
the total carbon dioxide (CO2) emissions of a cement plant while
the combustion of fuels in the cement kiln leads to 40% of the total
CO2 emissions (Benhelal et al., 2013). According to Tsakalakis and
Stamboltzis (2008), roughly two thirds of the total electrical en-
ergy consumption for cement production are used for particle size
reduction (grinding) and about 2% of the electricity produced
globally is used during the grinding process of raw materials
(Katsioti et al., 2009). Cement industry is a signicant contributor of
greenhouse gas (GHG) emission. It was found that reducing the
emission may lead to substantial reduction of overall GHG emission
(Valderrama et al., 2012). Improving thermal efciency would
create a high potential for reducing CO2 emissions from cement
plants.
An optimisation model towards emission mitigation in cement
plants has been discussed by numerous researchers. The most
relevant study was published by Kookos et al. (2011). The author
developed a mixed integer linear programming (MILP) optimisa-
tion model to minimise cement manufacturing costs by co-
processing. Carpio et al. (2008) used mathematical modelling to

http://dx.doi.org/10.1016/j.jclepro.2016.06.022
0959-6526/ 2016 Published by Elsevier Ltd.
134 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

Nomenclature V Volumetric gas ow in Nm3/t clinker under normal

condition
a, b, g, q Linearization variables for mX
Sets
a Alkalis Abbreviations and nomenclature
fg Flue gases ASU Air separation unit
h Heavy metals BP Back propagation
j Raw materials C Carbon content in fuels
k Fossil fuels C2S Dicalcium silicate
l Non fossil fuels C3A Tricalcium aluminate
o Oxides C3S Tricalcium silicate
p Clinker phases C4AF Tetracalcium aluminoferrite
s Sulfurs CaCO3 Calcium carbonate
CaO Calcium oxide
Parameters CCS Carbon capture and storage
A Availability in kg/t clinker CL Carbon looping
Bogue Bogue value CO2 Carbon dioxide
C Unit cost in $/kg GA Genetic algorithm
CEF Carbon emission factor in kg CO2/kg GAMS General Algebraic Modelling System
CO2GHG Current CO2 emission without mitigation method in kg HDP Heuristic dynamic programming
CO2/t clinker LCA Life cycle assessment
FCI Capital investment in $ M MBM Meat bone meal
M Big M constant MEA Monoethanolamine
MB Amount of pth clinker phases in clinker product in %. MgO Magnesium oxide
mw Molecular weight in kg/kmol MILP Mixed integer linear programming
nc Effects when oxy-fuel capture is selected MLD Mixed logic dynamic
NCV Net caloric value in GJ/kg NCV Net caloric value
Effects when fuels are selected N2 Nitrogen
OC Operating cost in $ M/y O&M Operating and maintenance
St Stoichiometric for O2 required for fuel combustion in O2 Oxygen
kg O2/kg OPC Ordinary Portland cement
TED Thermal energy demand in GJ/t clinker PC Petroleum coke
TEDr Thermal energy reduction in % PS Pattern search
TSR Thermal substitution rate in % PSO Particle swarm optimisation
Carbon capture and storage efciency in % RDF Refuse derived fuel
u Mass fraction in wt% S Sulfur content in fuels
SO3 Sulfur trioxide
Binary variables SS Sewage sludge
X Technology selections TED Thermal energy demand
TSR Thermal substitution rate
Continuous variables TDF Tire derived fuel
m Mass in kg/t clinker

calculate whether the substitution levels of the primary fuels by
alternative fuels is possible. Similarly, Oyepata and Obodeh (2015)
used Particle Swarm Optimisation (PSO) to nd the whether
optimal cost for cement production is bounded by environment
constraint and cement quality. The results shows that the optimal
cost with the use of alternative fuel can be reduced by 30e70%
without compromising the product. The studies, however, only
considered co-processing in mitigating CO2 emission from cement
plants.
The integration of various methods using MILP optimisation has
been discussed by several studies. Adebiy et al. (2015) discussed the
implementation of several energy efcient technologies in the
cement plant. Ba-Shammakh et al. (2008) considered three miti-
gation options: efciency improvement, switching to less carbon
content fuel and applying a post combustion capture system.
Ogbeide (2010) also considered several energy efcient technolo-
gies, fuel switching and a post combustion capture system. The
models that considered economic and environmental factors,
however, did not consider how the selected technologies will affect
the quality and chemistry of the product and the effects of CO2
reduction on the raw materials and fuels consumption. Moya et al.
(2010) studied a cost effective combination of retrotting of rotary
kilns, energy efcient grinding technology (substitution of ball
mills to vertical roller mills), co-generation (waste heat recovery),
and a post combustion capture system to improve energy efciency
and mitigate CO2 emissions up to 2030. The study proposed an
economical optimisation model with no consideration of quality
and chemistry of the product; the environmental aspect of the
study is driven by the economic value of the savings.
As a whole, few studies have implemented systematic and
comprehensive modelling to evaluate the economic and environ-
mental impact from various mitigation methods on cement quality
and production. There is also a lack of optimisation studies that
consider the potential of an oxy-fuel combustion system as one of
the options for capture technologies. This study integrates various
CO2 mitigation in cement plants while satisfying cement quality
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149
Fig. 1. Dry and wet cement manufacturing process.
standard and production rate. This work is performed as an
extension of work proposed by Kookos et al. (2011) which includes
more mitigation measures, such as kiln system improvements and
carbon capture and storage (CCS).
2. Literature review
CO2 emission in cement industry is mainly affected by the type
of fossil fuel used, amount of alternative fuels that replaces fossil
fuels and type of kiln used (Kajaste et al., 2015). Kara et al. (2010)
study the potential use of refuse derived fuel (RDF) using ratios of
0%, 8%, 12%, and 15% substitution with petroleum coke (PC) as
primary fuel. The results shows no disadvantages in the clinker
quality, a reduction in emissions, and increased economy due to
less cost PC. Rodrguez et al. (2013) reported that usage of sewage
sludge obtained from wastewater treatment plant could reduce
about 60%e70% of fossil fuel consumption in cement kiln.
Ariyaratne et al. (2010) determine the maximum possible
replacement of coal by meat bone meal (MBM) without affecting
the cement quality, emissions and overall process by a full scale
experimental study. The results shows that MBM could replace
more than 40% of the fossil fuel. The annual CO2 emissions could
be reduced by 10% if 40% of MBM is used in the kiln over the long
term.
Gereca et al. (2015) performed a Life Cycle Assessment (LCA)
for two scenarios: 100% PC, and co processing of 20% RDF with 80%
PC. It was found that the co-processing exhibit had a lower
environmental impact. Using RDF in cement plants helps in
mitigating the environmental impacts and could also be an
alternative strategy for waste management. Zhang and Mabee
(2016) suggested that the use of alternatives fuel is effective in
reducing emissions in cement production. The research then
studied the effects of substitution ratios to investigate the change
in CO2 emission. The CO2 decreases as the substitution increases.
LCA studies, however, only focused on the environmental impact
of alternative fuels.
Other methods that are efcient in reducing the thermal de-
mand of the cement plant are kiln improvements by reducing the
heat losses. Castan~o n et al. (2015) found that sintering and cooler
inlet secondary air temperature, and inlet NOx is the signicant
parameters that could affect the cement quality. Therefore, opti-
mising these parameters will improve the reduction of CO2 emis-
sions while maintaining the clinker quality. Atmaca and Yumrutas
(2014) studied the effects of insulating the kiln with anzast layer.
They also studied the effects of installing new chrome ore free
135
refractory bricks on the kiln's efciency and heat loss. It was found
that the efciency is increased by almost 10% and 15%. Stadler et al.
(2011) proposed the use of a controller with a model based on
mixed logic and dynamic (MLD) modelling. Controlled variables in
the modelling are the temperature during the process and the
oxygen level, while the manipulated variables are: air through the
kiln, raw meal fed, fuel fed and rotary speed of the cement kiln. The
applied controller reduces the instability of the process and
manipulated variables.
CO2 capture and storage (CCS) is another promising strategy to
reduce CO2 emissions. CCS is a set of technologies that includes the
capturing and compressing of CO2 from ue gas stream, trans-
porting and storing of the captured CO2 underground (Huisingh
et al., 2015). CO2 can be captured in three ways: post-combustion,
oxy-fuel, and pre-combustion. GarcaeGusano et al. (2015) dis-
cussed the potential of post combustion application to Spanish
cement industry using LCA and concluded that the rst cement
plants equipped with CCS will emerge in 2030 with global warm-
ing, ozone and abiotic reduction potentials of 15%, 27% and 11%.
According to Wennersten et al. (2015), due to a lack of international
agreements, slowdown in the planning of large-scale CCS has
occurred and many existing projects have been cancelled.
Vatopoulos and Tzimas (2012) studied the potential of application
for post-combustion, oxy-combustion and calcium looping post
combustion (CL) to the cement industry. For the same capture ef-
ciency of 85%, CL has an advantage over post-combustion, as the
specic energy consumption for calcium looping increases by 18%
compared to post-combustion's 45%. Oxy-combustion with lower
capturing efciency (60%), however, shows potential for lowering
overall energy consumption and the negative CO2 capture penalty
of a cement plant.
The nancial feasibility of CCS technology has also been
researched by Kuramochi et al. (2012) where in short term/mid-
term period, with CO2 avoidance cost of 65 V/t clinker, post-
combustion capture using monoethanolamine (MEA) solvent is
feasible only when low cost steam is available. In the long term, a
number of capture technologies may be feasible with 25e55
V/tCO2 avoided. Damartzis et al. (2016) investigated the effect on
absorption process for various combinations of amine solvents and
congurations of post combustion capture. It was found that the
optimal conguration improves the economic aspect in the range of
15%e35% depending on the solvents used. There has been an
attempt to adapt post-combustion CCS with an advanced amine
technology in cement plant on a pilot scale by Knudsen et al. (2014)
and resulted in a stable process with low energy consumption, low
136 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

amine consumption, low degradation and low emissions.

3. Methodology

Fig. 2 illustrates the main steps involved in determining the least

costly strategy for CO2 mitigation while meeting the kiln produc-
tion quality, as well as satisfying the emission reduction target.
The modelling approach starts with the modelling of an objec-
tive function (Eq. (1)) to calculate the minimum production cost for
when mitigating measures are to be adapted (Section 3.1). The
equation is subjected to several boundaries (Section 3.2) such as:
mass of production, product quality, raw materials and fuels
availability and CO2 emission (Eq. (2) to Eq. (32)). Then the data
needed to fully perform the models is collected for the case study,
such as: chemical properties of fuels, chemical properties of raw Fig. 3. Base material ow.
materials, costs estimations for fuels and raw materials, costs es-
timations for technology retrotting and others (Section: Overall
3.1. Assumptions
mass balance in clinker). The formulations are modeled based on
ow diagrams illustrated in Fig. 3, Fig. 4 and Fig. 5.
Several assumptions are introduced to avoid complexity of the
Fig. 3 shows the basic materials ow with no CCS installation.
models proposed:
Basically, the raw materials, fuels and combustion air are fed into
the cement kiln at the same time to produce clinker and ue gas;
i. Clinker produced is 95% of total cement production (CEM I
oxygen (O2), nitrogen (N2) and carbon dioxide (CO2). The raw ma-
Portland cement with more than 95% clinker) (Cembureau,
terials and fuel's ashes together formed clinker at a high temper-
2013)
ature while ue gases formed from decarbonation of raw materials
ii. Ashes from fuels are fully incorporated into clinker.
and fuel's combustion. Fig. 4 shows the conguration if an oxy-fuel
iii. 100% O2 is used instead of O2 rich stream (e.g: ~98% O2)
capture system is selected as one of the mitigation measures, the air
producing no inert substances from gas feed.
is fed into an air separation unit (ASU) to separate O2 and N2 in air
iv. Flue gas stack consist only CO2, excess O2 from combustion
to make sure pure O2 is fed into cement kiln. The ue gases then
air and inert N2 from combustion air.
emitted from the oxy-fuel capture system are O2 and CO2. If a post
v. 4 major phases in clinker; C3S, C2S, C3A and C4AF formed by
combustion system is selected, the ue gases formed (O2, N2 and
oxides from the raw materials and fuels in order to use Bogue
CO2) are separated using absorber/stripper unit (purier) as shown
method.
in Fig. 5.

3.2. Model formulation

The mathematical formulation includes an objective function

that minimizes the overall cost, and constraints to satisfy clinker
production, emission limits, operational restrictions, and product
quality.

3.2.1. Objective function

The objective function of this study is to nd the least clinker
production cost that satises the CO2 reduction target. The objec-
tive function for this is given by:

minimize production cost; Zmin $=t clinker

X X X
FCI OC Cj mj Ck mk Cl ml
jRaw materials kFossil lNon fossil
X X
FCIl Xl OCl ml  FCIr Xr 
l2Non fossil r2Kiln
X
FCIc Xc OCc 
c2CCS
(1)

where Cj is in $/kg of jth raw materials, mj is the mass of jth raw

Fig. 2. Methodology to determine the least costly strategy and most environmental
friendly cement manufacturing process.
materials in kg/t clinker produced. There are two types of fuels
involved in cement production which are kth fossil fuels and lth
alternative/non fossil fuels. Ck and Cl in the model represents cost of
kth fossil fuels and cost of lth non fossil fuels in $/kg fuel while mk
and ml represents mass of kth fossil fuels and mass of lth non fossil
fuels in $/t clinker. FCIl, FCIr, and FCIc is the xed capital investment
for lth non fossil fuels in $/t clinker, rth kiln improvements in $/t
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149 137

Fig. 4. Oxy-fuel carbon capture conguration.

Fig. 5. Post combustion carbon capture conguration.

clinker and cth CCS in $/t clinker produced while VCl and VCc is the
variable cost for lth non fossil fuels in $/kg of fuels and cth CCS in $/t
clinker. Binary variables Xl, Xr, and Xc are introduced for selection of
fuels, kiln improvements and CCS technologies.
The costs involved in producing clinker include: xed cost, FCI
and operating and maintenance (O&M) cost, OC of existing cement
plant and the installation cost associated with technology selec-
tions, namely co-processing, kiln improvements and CCS. The
operating cost of the existing cement plant includes electricity,
wages, maintenance cost, and etc. The third part of the objective
function is the variable cost for raw materials and fossil fuels based
on their consumption during clinker production.
The installation cost for co-processing technology includes both
xed and variable costs. The xed cost include the capital invest-
ment cost, handling and storage of alternative fuels, new facilities
and etc. Operating costs for co-processing include quality control,
safety, emissions control, etc. The installation cost for kiln im-
provements includes the technology installation cost, instruments
cost, etc. The installation cost for CCS includes the xed cost and
operating cost. The xed cost includes added capital investment of
new facilities and equipment. Operating cost include added power
consumption, added materials and etc.
Since the production cost is in the function of size, the xed cost
will be multiplied with an amortization factor to convert it into an
annual basis and will be divided by the annual capacity of the plant
(t clinker/y). The operating cost (in $/y) will be divided by the
annual capacity of the plant (t clinker/y).
3.2.2. Constraints
3.2.2.1. Mass balance around cement kiln
Basic mass balance around the cement kiln is formulated based
on Fig. 3. However, as CCS mitigation methods are considered, the
mass balance is adjusted to full the schematic of CCS mitigation
methods as shown in Figs. 4 and 5.
Overall mass balance in clinker
Clinker produced from cement kiln contains compounds like
oxides, alkalis, sulfurs and heavy metals which come from raw
materials, and ash from the burning of fuels. Oxides, alkalis and
heavy metals enter the system through raw meals, and ash from
fuels. Sulfurs in clinker product enter through raw material and
fuels. The overall constraint is formulated as:
138 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

X X X X fraction of SO3 in lth non fossil fuels' ash. uSk is the mass fraction of
mo ma ms mh mclinker sth sulfurs in kth fossil fuels and uSl is the mass fraction of sth sulfurs
o2Oxides a2Alkalis s2Sulfurs h2Heavy Metals in lth non fossil fuels.
(2) The mass fraction of sulfurs expressed as SO3 is multiplied with
the mass fraction of fuel ash to get the amount of sulfurs contrib-
where mo, ma, ms and mh represents the total mass of oth oxides, uted by fuels' ashes. Factor of 80/32 is obtained from the mass
total mass of ath alkalis, total mass of sth sulfurs and total mass of hth fractions of the sulfurs in fuels. Sulfur content in fuels reacts with
heavy metals kg/t clinker. O2 to produce SO3, as shown in the chemical reaction below:
The mass of oxides in the resultant clinker are subjected to the
summation of oxides from raw materials and oxides from the ashes 3
S O2 / SO3
of combusted fuels. 2
X X X From the chemical reaction, 1 mol of S is needed to produce1
mo uo;j mj uo;k mk uo;l ml mol of SO3. So, in order to nd the amount of SO3 contributed from
j2Raw materials k2Fossil l2Non fossil
the sulfur content in fuels, the S wt. % from fuels are subjected to a
ca 2Oxides molar mass of 80 kg SO3/mol divided by 32 kg S/mol.
(3) Overall gas mass balance in clinker
To nd the amount of gas emitted from the cement plant, total
where uo,j is the mass fraction of oth oxides in jth raw materials, uo,k ue gas formed is required. Flue gases emitted from the cement
is the mass fraction of oth oxides in kth fossil fuels' ash and uo,l is the plant are O2, N2, and CO2. The volumetric owrate of ue gas
mass fraction of oth oxides in lth alternative fuels' ash. Mass fractions emitted is calculated by assuming an ideal gas law under normal
are in wt. %. conditions.
The mass of alkalis in the resultant clinker were subjected to the
X mfg
summation of alkalis from raw materials and alkalis from ashes of Vfg 22:414 cfg 2Flue gas (7)
burned fuels. mwfg
fg2Flue gases
X X X
ma ua;j mj ua;k mk ua;l ml where mfg is the mass of fgth gas in ue gas in kg per t clinker, mwfg
j2Raw materials k2Fossil l2Non fossil is the molecular weight of fgth ue gas in kg/kmol. 22.414 is the
ca 2Alkalis molar volume of an ideal gas under normal conditions, in Nm3/
(4) kmol.
The decarbonation of raw materials and combustion of fuels
where ua,j is the mass fraction of ath alkalis in jth raw materials, ua,k contribute to CO2 emissions. CO2 emissions from fuels are a com-
is the mass fraction of ath alkalis in kth fossil fuels' ash and ua,l is the bination of CO2 emissions from fossil fuels and CO2 emissions from
mass fraction of ath alkalis in lth alternative fuels' ash. Mass fractions non-fossil fuels. Mass of CO2 emitted from decarbonation (mcb) and
are in wt. %. combustion (mcm) in kg/t clinker are formulated as below:
The mass of heavy metals in the resultant clinker were subjected
to the summation of heavy metals from raw materials and ashes of mCO2 mcb mcm (8)
combusted fuels. At a high temperature in cement kiln, CO2 is formed from the
X X X decomposition of calcium carbonate (CaCO3) and magnesium car-
mh uh;j mj uh;k mk uh;l ml bonate (MgCO3) in raw materials. CO2 emissions from decarbon-
j2Raw materials k2Fossil l2Non fossil
ation of raw materials are then illustrated as below:
ch 2Heavy metals
(5) CaCO3 /CaO CO2

where uh,j is the mass fraction of hth heavy metals in jth raw ma- MgCO3 /MgO CO2
terials, uh,k is the mass fraction of hth heavy metals in kth fossil fuels'
Carbon content, C in fuels reacts with combustion air during
ash and uh,l is the mass fraction of hth heavy metals in lth alternative
their combustion to produce CO2. The CO2 emissions from fuels
fuels' ash. Mass fractions are in wt. %. Since oxides, alkalis and heavy
combustion are then illustrated as below:
metals are obtained from ashes, the mass fraction of oxides, alkalis
and heavy metals in fuels are multiplied with mass fraction of ash.
C O2 /CO2
Sulfurs expressed as sulfur trioxides (SO3) in the resultant
clinker were subjected to SO3 produced from sulfates in raw ma- From the chemical reactions, the CO2 emissions are then
terials and fuels' ashes, as well as sulfur content in fuels (due to fuel formulated as:
combustion), as written below.
44 X 44 X
X X mcb uCaOj mj uMgOj mj
ms uSO3 j mj uSO3 k mk 56 40
j2Raw materials j2Raw materials
j2Raw materials k2Fossil (9)
X X  80

uSO3 l ml uSk  mk X X
32 mcm CEFk k Xk mk CEFl l Xl ml (9a)
l2Non Fossil k2Fossil
X  
80 k2Fossil l2Non fossil
uSl  ml (6)
32
l2Non fossil where uCaOj is the mass fraction of CaO in jth raw materials and
where uSO3j is the mass fraction of SO3 in jth raw materials, uSO3k is uMgOj is the mass fraction of MgO in jth raw materials. CEFk is the
the mass fraction of SO3 in kth fossil fuels' ash and uSO3l is the mass carbon emission factor of kth fossil fuels and CEFl is the carbon
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149 139

emission factor of lth non fossil fuels. Both factor 44/56 and 44/40 in consists of 76.8% N2 and 23.2% O2, with O2 released at a controlled
Eq. (9) are the molar fractions of CO2 with CaO and MgO in raw level of 10%, complete combustion is achieved, and the possibility of
materials. From the chemical reaction, 1 mol of CaCO3 produces oxy-fuel combustion to be selected as a mitigation method results
1 mol CO2 and 1 mol of CaO. So, in order to nd the amount of CO2 in a 100% O2 feed as shown in Fig. 4; constraints indicated are as
produced from the decarbonation of CaCO3 in raw materials, the follows:
CaO wt. % from raw materials are subjected to a molar mass of 44 kg X
CO2/mol divided by 56 kg CaO/mol. Similarly, 1 mol of MgCO3 mN2 76:8%  mair  76:8%  nc  bc (15)
produces 1 mol CO2 and 1 mol of MgO. So, in order to nd the c2CCS
amount of CO2 produced from the decarbonation of MgCO3 in raw X
materials, the MgO wt. % from raw materials are subjected to molar mO2 23:2%  mair  23:2%  nc  bc
mass of 44 kg CO2/mol divided by 40 kg MgO/mol. The formulation c2CCS
0 1
uses the same principal to calculate sulfurs contributed from fuels
X X X
in Eq. (6). nc  gc  @ Stk mk Stl ml A
Greenhouse gases are contributed to by CO2 emissions as c2CCS k2Fossil l2Non fossil
biogenic and non-biogenic. When a biogenic fuel is selected, the
(16)
CO2 emitted from biogenic fuel combustion will not be evaluated in
the impact of the GHG since the CO2 generated is recycled to the  
32
sources that replace the fuel. However, some alternative fuels de- mO2 Vfg 10% (17)
cays and produces products that affects the GHG impact (Mikul ci
c 22:414
et al., 2014). To account for this, is introduced. It is important to
where mN2, and mO2 are the mass of N2 in kg/t clinker and mass of
note that Eq. (i), produces nonlinear function. Through exact
O2 in kg/t clinker. mair represents mass of air fed in kg/t clinker. Stk
linearization, the complexity and non-linearity of the model is
and Stl are the stoichiometric O2 required for complete combustion
avoided. The method of linearizing the non-linear problem is based
of kth and lth fuels. Mass of O2 emitted from cement kiln is calcu-
on an established method that has been described in Yue and You
lated as the difference between O2 in fed air with O2 used for
(2014):
complete combustion. The O2 required for complete combustion of
a. Introducing a new variable, a, as an equal to the nonlinear term:
fuels is calculated as per Green and Perry (2008).
a mX bc and gc are linearized variables where bc is linearized variable
for mairXc while gc is linearized variable for mO2inXc. Binary Xc is
b. a lie between certain boundaries: introduced for CCS technology selection. nc is a parameter intro-
duced to include the effect from selected CCS technology. mO2in is
0  a  MX (9b) the mass of pure O2 fed into cement kiln in kg/t clinker. In Eq. (15),
the second term; 76.8%ncbc is introduced to make sure that once
m  M1  X  a  m (9c) oxy-fuel CCS is selected, N2 will not be present since 100% O2 is
assumed to be fed into the cement kiln. Similarly, in Eq. (16), second
where M is any large number. term; 23.2%ncbc is introduced to eliminate the mass of air term
Inequalities in (ii) and (iii) enforce the following equivalence: while third term; ncgc is introduced to account for the pure O2 fed
to the burner. As discussed before, the linearized variables are
 If some X 0, then (ii) ensures that a 0 leaving (iii) redundant. subjected to the following constraints:
 If some X 1, then (iii) ensures a m leaving (ii) redundant.
0  bC  MXC (18)
The introduction of a resulted in changes in (i):
X X mair  M1  XC  bC  mair (19)
mcm CEFk k ak CEFl l al (10)
k2Fossil l2Non fossil 0  gC  MXC (20)

Eq. (10) is then subjected to these constraints:

mO2  M1  XC  gC  mO2 (21)
in in

0  ak  MXk (11)

mk  M1  Xk  ak  mk (12) 3.2.2.2. Chemical analysis of clinker

In general, the quality of a clinker blend is composed of multiple
0  al  MXl (13) solid components. Based on the clinker analysis suggested by
Bogue, 4 major phases in clinker; tricalcium silicate (C3S), dicalcium
silicate (C2S), tricalcium aluminate (C3A) and tetracalcium alumi-
ml  M1  Xl  al  ml (14) noferrite (C4AF) are formed by oxides from the raw materials and
the ashes of fuels:
For N2 and O2 balance, assuming that air fed for combustion where MBpL; in kg/t clinker, represent the lower limits for pth

X   X   X  
MBpL mclinker  Boguep;o mo;j Boguep;o mo;k Boguep;o mo;l  MBU
p mclinker (22)
j2Raw materials k2Fossil l2Non fossil
140 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

clinker phase while MBU p ; in kg/t clinker, represent the upper limits Co-processing
of pth clinker phase. mo,j, mo,k, and mo,l are the mass of oxides in jth The total number of fuels mixed must not exceed the maximum
raw materials, kth fossil fuels' ash and lth non fossil fuels' ash. The amount of fuels that can be mixed in the burner. The amount of
lower and upper limit of the constraints are the typical concen- fuels that are allowed to be mixed depends on the type of burner
tration of pth phases in clinker product. Values of Boguep,o param- used in the cement kiln. For example, according to Vaccaro (2006),
eters is tabulated in Table 6. a modern multichannel burner that can burn up to ten different
fuels was installed in a rotary cement kiln in Switzerland.
3.2.2.3. Thermal energy X X
Thermal energy demand Xk Xl  max number of fuels mixed (30)
The type of kiln used will affect the amount of heat supplied kFossil lNon fossil
from fuels. However, when kiln improvements are applied as
Xk and Xl are the binary variables for the use of kth and lth fuels.
mitigation measures, the thermal efciency increases, improving
The binary variable represents the kth or lth fuel type used and could
the thermal demand from the kiln. Heat consumed in kilns is
have a value of zero if the kth or lth fuels are not selected.
provided by both types of fuels. The thermal energy demand
constraint is then formulated as below: 
1 if fuel k is used
X X X Xk
0 otherwise
mk NCVk ml NCVl TED 1  Xr TEDr
k2Fossil l2Non fossil rKiln 
1 if fuel l is used
(23) Xl
0 otherwise
where NCVk is the net caloric value of kth fossil fuels in GJ/kg of fuels Kiln improvements
and NCVl is the net caloric value of lth non fossil fuels in GJ/kg of Since it is possible to apply all the kiln improvements, maximum
fuels. TED represents the thermal energy demand/consumption in number of improvements has been used.
GJ/t clinker. Binary Xr is included to represent the selection of rth kiln X
improvements to thermal demand where TEDr is the thermal ef- Xr  maximum selections (31)
ciency in % improved once rth kiln improvements are installed. rKiln
Thermal substitution rate
As the burner will not realistically be fully functioning on 100% Xr is the binary variable for the selection of rth kiln improvements.
alternative fuels, the use of alternative fuels will be subjected to a 
limit called thermal substitution rate (TSR) in %. The constraint can 1 if kiln improvement r is used
Xr
be written as: 0 otherwise

X   X Carbon capture and storage (CCS)

TSR
ml NCVl  TED 1  Xr TEDr (24) Eq. (32) illustrates that only one capture process can be installed
100
l2Non fossil rKiln at a time.
X
Xc  1 (32)
3.2.2.4. CO2 emission reduction target cCCS
The total carbon emissions after the implementation of CO2
Xc is the binary variable for the installation of cth CCS, where the
mitigation measures should not be more than the total CO2 emis-
binary variable could have a value of zero if the cth CCS technology
sion for the base case, and can be formulated as:
is not selected or otherwise.
X
mCO2  c qc  1  %CO2 CO2GHG (25) 
1 if CCS c is used
c2CCS Xc
0 otherwise
mCO2 is the amount of CO2 emitted obtained from Eq. (8). CO2GHG
is current CO2 emission without mitigation method (base) in kg/t
3.3. Case study
clinker. c is the percent CO2 reduction when cth CCS is imple-
mented. qc is the linearization variable; multiplication of binary
The applicability of the optimisation model is demonstrated
variable Xc with mCO2. The variable is subjected to:
using the case study of a cement plant with an annual capacity of
A 1 Mt Ordinary Portland Cement (OPC) with a lifespan of 25 years.
0a X (26) The plant consists of 5-stages pre-heater pre-calciner kiln system,
AC
Vintage 3 channel type burner with fuel to air ratio is 2 to 1. The
mCO2  M1  Xc  qc  mCO2 (27) thermal energy demand is 3.25 GJ/t clinker. Fixed capital invest-
ment and the operating cost of the existing plant is $ 284 M and $
25 M/y (IEA GHG, 2008) with a 10% interest rate. On the other
3.2.2.5. Fuels availability hand, the demand is based on the projection of cement produc-
The fuels used must not exceed the fuels availability. tion for the year 2020 estimated at 25.2 Mt (Trading Economics,
2016).
mk  Ak Xk (28) The raw materials selected for this study include limestone,
clay, sand, or iron source while the fuel options are coal or pe-
ml  Al Xl (29) troleum coke (PC) as main fuels and refuse derived fuel (RDF),
sewage sludge (SS), tire derived fuel (TDF), or meat bone meal
where Ak and Al are the availability of kth and lth fuels in kg/t clinker. (MBM) as additional/alternative fuels. Kiln improvements include
kiln refractories, implementation of process control in kiln, kiln
3.2.2.6. Technologies selection combustion system improvements, and increase preheater stages
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149 141

Table 1
Raw materials data.

Raw materials Limestone Clay Sand Iron source

Cost, C ($/kg) 0.02 0.04 0.01 0.01

Chemical Formula Molecular weight, mw Mass fraction dry material, u
SiO2 60.08 0.1 0.6 0.8 0.01
Al2O3 101.961 0.01 0.1 0.05 0.005.
Fe2O3 159.69 0.01 0.05 0.025 0.95
CaO 56.0774 0.5 0.1 0.03 0.02
MgO 40.3044 0.005 0.02 0.02 0.01
K2O 94.196 0.003 0.02 0.015 0.001
Na2O 61.97,894 0.001 0.01 0.005 0.001
SO3 80.066 0.0001 0.0005 0.0005 0.003
Others 0.371 0.0995 0.05 0.002

Source: Kookos et al. (2011) for mass fraction of dry raw materials; Willett (2011) for price of limestone; and U.S. Geological Survey (2014) for price of other raw materials.

from 5 stages to 6 stages. On the other hand, either post com-
bustion or oxy-combustion will be selected for the carbon capture
system.
Tables 1e5 shows the data used in this study. The models are
then executed using the General Algebraic Modelling System
(GAMS); a high-level modelling system for mathematical pro-
gramming and optimisation. The model required 52 variables and
72 constraints. The computer times were 0:00:00.242 s of CPU time
on Intel Core i7-3520M 2.90 GHz machine.
4. Results and discussion
Results for this study are divided into two sections, base case
study and cement manufacturing cleaner process improvement. In
section 4.1, results indicate the base case scenario where no
mitigation methods are implemented. Results obtained in section
4.1 will be used as the benchmark in section 4.2, where the highest
possible CO2 reduction target is researched by changing the value of
the CO2 reduction target. Section 4.3 will discuss the effects of the
production rate on CO2 emissions and technology selections.
Sensitivity analysis will be presented in section 4.4.
4.1. Base case scenario
Base case scenario results are obtained when there is no miti-
gation method applied. The CO2 calculated in base case is used as
current CO2 emissions in the proposed models for subsequent
sections. Below are the conditions for a base case scenario:
i. Thermal substitution rate (TSR) 0%

Table 2
Characteristics of fuels.

Parameters Coal PC RDF SS TDF MBM

Net caloric value, NCV (GJ/kg) 0.03 0.033 0.026 0.016 0.032 0.017
Oxygen required, St (kg O2/kg) 2.32 2.67 2.16 1.59 2.34 1.45
CO2 emission factor, CEF (kg CO2/kg) 2.75 3.30 2.20 1.58 2.56 1.54
Environment effects factor, 4 0 0 0.5 0.5 0 1
Cost ($/kg) 0.11 0.10 0.05 0.04 0.03 0.02
Availability, A (t/y) e e 30,000 30,000 20,000 20,000
Mass fraction of dry material
C 0.75 0.9 0.6 0.43 0.7 0.42
H 0.05 0.03 0.1 0.09 0.07 0.06
O 0.08 0.01 0.25 0.272 0.1 0.153
S 0.003 0.04 0.01 0.002 0.015 0.004
N 0.0001 0.01 0.001 0.018 0.005 0.075
Ash 0.1 0.01 0.1 0.2 0.1 0.3
Mass fraction of ash
SiO2 0.525 0.4 0.4 0.4 0.22 0.005
Al2O3 0.3 0.1 0.25 0.15 0.1 e
Fe2O3 0.1 0.07 0.02 0.05 0.015 e
CaO 0.03 0.01 0.2 0.2 0.11 0.2
MgO 0.01 0.03 0.025 0.025 0.015 e
K2O 0.015 0.005 0.01 0.01 0.01 0.005
Na2O 0.005 0.005 0.01 0.01 0.01 e
SO3 0.015 0.02 0.02 0.01 0.15 e
Cl 0.001 0.001 e e 0.004 e
NiO e 0.15 e e e e
V2O5 e 0.2 e e e e
ZnO 0.0002 0.00005 0.000085 0.0007 0.35 e
Cd 0.00001 0.000001 0.000001 0.000005 0.000005 e
Pb 0.0002 0.00001 0.00005 0.003 0.00005 e
Tl 0.000004 0.00008 e e 0.0000001 e
As 0.000002 0.000005 0.000005 0.00002 0.0000001 e
Hg e e 0.000002 0.000015 e e

PC Petroleum coke.
RDF Refuse derived fuel.
SS Sewage sludge.
TDF Tire derived fuel.
MBM Meat bone meal.
Source: Kookos et al. (2011).
142 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

Table 3 Table 5
Kiln improvements data. Bogue analysis data.

Measures FCI ($/t Thermal improvements, Phases, p Oxides, o

clinker) TEDr (%)
SiO2 Al2O3 Fe2O3 CaO Mass limit, MB (%)
Improved kiln refractories 0.25 19
Lower Upper
Improved process control in kiln 1.70 6
Kiln combustion system 1.00 15 C3S 7.6 6.72 1.43 4.07 50 60
improvements C2S 8.6 5.07 1.08 3.07 10 35
Increase preheater stage 0.88 3 C3A 0 2.65 1.69 0 1 15
C4AF 0 0 3.043 0 1 15
Source: ECRA (2009) for increase preheater data; and Worrell et al. (2013) for other
kiln improvements data. Source: Kookos et al. (2011).

ii. CO2 reduction target 0% selected includes: co-processing, kiln refractories improvements
iii. Xr, and Xc 0% and improvements in the combustion system. Once the reduction
target is set to 20%, oxy-fuel combustion is also chosen. With the
Table 6 shows the results for the base case scenario. The optimal ability to mitigate up to 61% of CO2 emissions, oxy-fuel combustion
cost for clinker production for this particular scenario is $ 90.21/t is selected until a 70% CO2 reduction target is set, which is where
clinker. The total CO2 emissions generated from both calcination of post-combustion is chosen. Please note that CO2 produced is
raw materials and fuels combustion is 856.68 kg CO2/t clinker dened as the amount of CO2 produced from decarbonation
where 531.68 kg CO2/t clinker is from calcination of raw materials (calcination) of raw materials and from the combustion of fuels,
and 325.00 kg CO2/t clinker from fuels combustion. Majority of raw while CO2 released is dened as the amount of CO2 released to the
materials used came from limestone with 1315.49 kg/t clinker, atmosphere after capture technology is implemented.
followed by sand with 161.48 kg/t clinker and iron source with Fig. 6 shows the effects of CO2 reduction target on optimal
10.89 kg/t clinker. The thermal supplied to cement kiln came fully production cost and total CO2 emissions released to the atmo-
from petroleum coke with 98.49 kg/t clinker. Flue gases for this case sphere. 10% CO2 reduction can be achieved with an optimal cost
are comprised of CO2 emissions, O2 excess and inert N2 with from $ 90.21/t clinker to $ 87.17/t clinker. This is because, with kiln
856.68 kg/t clinker, 310.13 kg/t clinker and 1897.11 kg/t clinker. refractories improvements and improvement in the combustion
Results obtained in this section will be used as the basis for section system, the thermal demand from kiln is improved, lowering the
4.2, where CO2 emitted in the base case scenario is used as the amount of fuels needed for combustion and with the low cost for
current CO2 emissions value. implementation of these methods, the optimal production cost is
reduced. As oxy-fuel combustion is chosen at 20% reduction target,
production cost increases from $ 87.17/t clinker to $ 94.05/t clinker
4.2. Highest possible CO2 reduction
due capital and installation cost of oxy-fuel combustion. The
optimal cost remains constant until 60% reduction target. This can
This section discusses the highest potential of CO2 emission
be explain by the efciency of oxy-fuel combustion in mitigating up
reduction that can be achieved by the implementation of various
to 61% of CO2 emissions fullling the constraint set by the reduction
mitigation strategies. The mathematical model can be used to
target. This is also reected on the CO2 emission.
predict the highest CO2 reduction by changing the value for CO2
Fig. 7 shows the effects of CO2 reduction target on fuels usage
reduction target %. The current CO2 emission is the same as the base
and CO2 emitted from fuels. It is observed that fuels usage de-
case scenario and the TSR value throughout the optimisation is 30%.
creases from 98.49 kg/t clinker of petroleum coke to 44.97 kg/t
The optimisation is separated into two: rst series of optimi-
clinker of petroleum coke and 19.88 kg/t clinker of tire derived fuels
sation and second series of optimisation. The rst series of opti-
when CO2 reduction target is implemented as kiln improvements
misation is done to determine the range of the highest CO2
methods are chosen. The decrease remains constant until 70% of
reduction target. This can be achieved by changing the amount of
CO2 reduction target 10% until an infeasible value is obtained.
Once an infeasible value is achieved, the maximum CO2 reduction
target falls between the last and second last reduction target. The Table 6
Base case scenario.
second series of optimisation is done for a more accurate value of
the maximum CO2 reduction target for this study and is acquired by Minimized cost ($/t clinker) 90.21
changing the reduction target with a smaller value of 1.0% be- CO2 released (kg CO2/t clinker)
Total CO2 856.68
tween the range obtained in the rst series of optimisation. Table 7
CO2 released from fuels combustion 325.00
and Table 8 summarises the results from the optimisation pro- CO2 released from decarbonation of raw materials 531.68
cesses. Figs. 6 to 9 shows the results in graphical form. Base case Raw materials usage (kg/t clinker)
results are included for comparison purposes. Limestone 1315.49
As can be seen in Table 7, the maximum CO2 reduction that can Clay 0
Sand 161.48
be achieved lies within the 70%e80% range. The range will be the Fe Source 10.89
basis for second series of optimisation. Among the technologies Fuels usage (kg/t clinker)
Coal 0
Petroleum coke 98.49
Table 4 Flue gas (kg/t clinker)
CCS data. Volumetric ue gas (Nm3/t clinker) 2172.26
Mass of ue gas 3063.92
FCI ($ M) OC ($ M/y) % CO2 mitigated nc
Mass of air 2470.19
Post combustion 205.60 22.90 0.74 0 Mass of O2 310.13
Oxy-fuel combustion 50.10 4.00 0.61 1 Mass of CO2 856.68
Mass of N2 1897.11
Source: IEA GHG (2008)
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149 143

Table 7
Results for rst series of optimisation.

CO2 reduction target (%) Base 10 20 30 40 50 60 70 80

Minimized cost ($/t clinker) 90.21 87.17 94.05 96.69 96.66 96.66 96.66 132.77 Infeasible
CO2 emission (kg CO2/t clinker)
Total CO2 produced 856.68 732.26 685.35 732.26 732.26 732.26 732.26 732.26 Infeasible
CO2 released from fuels combustion 325.00 199.28 160.46 199.28 199.28 199.28 199.28 199.28 Infeasible
CO2 released from decarbonation of raw materials 531.68 532.98 524.88 532.98 532.98 532.98 532.98 532.98 Infeasible
CO2 captured 0 0.00 0.00 446.68 446.68 446.68 446.68 541.87 Infeasible
CO2 released to atmosphere 856.68 732.26 685.35 285.58 285.58 285.58 285.58 190.39 Infeasible
Raw materials usage (kg/t clinker)
Limestone 1315.49 1318.73 1273.27 1318.73 1318.73 1318.73 1318.73 1318.73 Infeasible
Clay 0 0.00 155.47 0.00 0.00 0.00 0.00 0.00 Infeasible
Sand 161.48 161.44 31.28 161.44 161.44 161.44 161.44 161.44 Infeasible
Fe Source 10.89 11.16 22.72 11.16 11.16 11.16 11.16 11.16 Infeasible
Fuels usage (kg/t clinker)
Coal 0 0.00 42.47 0.00 0.00 0.00 0.00 0.00 Infeasible
Petroleum coke 98.49 44.97 0.00 44.97 44.97 44.97 44.97 44.97 Infeasible
Refused derived fuels 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Infeasible
Sewage sludge 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Infeasible
Tire derived fuels 0 19.88 17.06 19.88 19.88 19.88 19.88 19.88 Infeasible
Meat bone meal 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Infeasible
Cement kiln upgrade selection
Improved refractories Infeasible
Process control Infeasible
Improved kiln combustion system Infeasible
Increase preheater stage Infeasible
CCS technology selection
Oxy-fuel combustion Infeasible
Post combustion Infeasible

CO2 reduction target. This is because the kiln thermal energy de-
mand remains constant due to the same kiln improvements
methods chosen, as can be seen in Table 7. The constant fuel usage
reects the amount of CO2 emitted from fuels where they remain
constant until the 70% reduction target.
Fig. 8 shows the effects of the CO2 reduction target on raw
materials usage and CO2 emissions from decarbonation of raw
materials. In general, the majority of raw materials used is lime-
stone, followed by sand and iron source. The amount of raw
materials used increase as the CO2 reduction target increase from
0% to 10%. This is due to the decrease in fuel supply lowering the
amount of fuel ash to be incorporated in the clinker thus increasing
the raw materials feed to compensate for the amount of composi-
tion reduced from the absence of the fuel ash. The increase of raw
materials usage, however, contributes to higher CO2 emissions from
raw materials calcination (from 531.68 kg/t clinker to 532.98 kg/t
clinker). For 20% CO2 reduction, co processing of raw material are
used, 1273.27 kg/t clinker limestone, 155.47 kg/t clinker clay,

Table 8
Results for the second series optimisation to determine the highest achievable CO2 reduction target.

CO2 reduction target (%) 71 77 78 79

Minimised cost ($/t clinker) 132.77 132.77 133.40 136.46

CO2 released (kg CO2/t clinker)
Total CO2 produced 732.26 732.26 723.02 691.94
CO2 released from fuels combustion 199.28 199.28 189.88 160.46
CO2 released from decarbonation of raw materials 532.98 532.98 533.14 531.47
CO2 captured 541.87 541.87 535.04 512.03
CO2 released to atmosphere 190.39 190.39 187.99 179.90
Raw materials usage (kg/t clinker)
Limestone 1318.73 1318.73 1319.12 1309.09
Clay 0.00 0.00 0.00 37.24
Sand 161.44 161.44 161.51 128.19
Fe Source 11.16 11.16 11.15 15.03
Fuels usage (kg/t clinker)
Coal 0.00 0.00 0.00 42.47
Petroleum coke 44.97 44.97 42.85 0.00
Refused derived fuels 0.00 0.00 0.00 0.00
Sewage sludge 0.00 0.00 0.00 0.00
Tire derived fuels 19.88 19.88 18.94 17.06
Meat bone meal 0.00 0.00 0.00 0.00
Cement kiln upgrade selection
Improved refractories
Process control
Improved kiln combustion system
Increase preheater stage
CCS technology selection
Oxy-fuel combustion
Post combustion
144 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

140.00 900.00

800.00
120.00

CO2 emissions (kg CO2/t clinker)

Production cost ($/t clinker)
700.00
100.00
600.00

80.00 500.00

60.00 400.00

300.00
40.00
200.00

20.00
100.00

0.00 0.00
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00
CO2 emission 856.68 732.26 685.35 285.58 285.58 285.58 285.58 190.39
Cost 90.21 87.17 94.05 96.69 96.66 96.66 96.66 132.77

CO2 reduction target (%)

Fig. 6. Effects of CO2 reduction target on optimal production cost and total CO2 emissions.

31.28 kg/t clinker sand and 22.72 kg/t clinker iron. This co pro-
cessing lead to lower CO2 emissions from 532.98 kg/t clinker to
524.88 kg/t clinker. Amount of raw materials usage and CO2
emissions from calcination goes back to their original values once a
reduction target is set at 30% with oxy-fuel combustion, co-
processing and improvements in combustion and refractories sys-
tem selected. This is because CCS technology did not affect the
amount of materials usage in cement kiln as stated earlier. The
values remain constant until 70% CO2 reduction target.
Fig. 9 shows the effects of a CO2 reduction target on the total CO2
produced, captured and released. Note that the total CO2 produced
is different from the amount of CO2 released. As illustrated in Fig. 9,
it can be observed that the CO2 produced decreases from 732.26 kg/
t clinker in base case to 685.35 kg/t clinker at 20% reduction target
and with no CO2 captured. This is because the implementation of
kiln improvements is selected to full the reduction target at a
lower production cost as compared to the capture system. In order
to achieve 30% CO2 reduction, a combination of kiln improvements
with co-processing and oxy-fuel combustion system are recom-
mended. 30% CO2 reduction target can be achieved through the
implementation of oxy-fuel combustion capture system, together
with kiln improvements; generating the amount of CO2 captured
and released of 446.68 kg/t clinker, and 285.58 kg/t clinker will
remain constant until 60% reduction target. On the other hand,
post-combustion technology is selected to achieve 70% CO2
reduction. This will result in the amount of CO2 captured and

120.00 350.00
CO2 emission (kg CO2/t clinker)
Fuels usage (kg/t clinker)

100.00 300.00

250.00
80.00
200.00
60.00
150.00
40.00
100.00

20.00 50.00

0.00 0.00
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00
Coal 0.00 0.00 42.47 0.00 0.00 0.00 0.00 0.00
PC 98.49 44.97 0.00 44.97 44.97 44.97 44.97 44.97
RDF 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
SS 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
TDF 0.00 19.88 17.06 19.88 19.88 19.88 19.88 19.88
MBM 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
CO2 emission from fuel 325.00 199.28 160.46 199.28 199.28 199.28 199.28 199.28

CO2 reduction target (%)

Fig. 7. Effects of CO2 reduction target on fuels usage and CO2 emissions from fuels combustion.
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149 145

1,400.00 534.00

Raw materials usage (kg/t clinker)

CO2 emissions (kg CO2/t clinker)

1,200.00 532.00

1,000.00 530.00

800.00 528.00

600.00 526.00

400.00 524.00

200.00 522.00

- 520.00
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00
Limestone 1,315.49 1,318.73 1,273.27 1,318.73 1,318.73 1,318.73 1,318.73 1,318.73
Clay 0.00 0.00 155.47 0.00 0.00 0.00 0.00 0.00
Sand 161.48 161.44 31.28 161.44 161.44 161.44 161.44 161.44
Fe Source 10.89 11.16 22.72 11.16 11.16 11.16 11.16 11.16
CO2 emission from decarbonation 531.68 532.98 524.88 532.98 532.98 532.98 532.98 532.98

CO2 reduction target (%)

Fig. 8. Effects of CO2 reduction target on raw materials usage and CO2 emissions from decarbonation of raw materials.

released changing to 541.87 kg/t clinker and 190.39 kg/t clinker,
reecting the higher capture efciency of post-combustion tech-
nology compared to oxy-fuel combustion.
Table 8 shows the result for the second series optimisation. Note
that constant results are not included in table (between 71%e77%
CO2 reduction target). Technology selections in the remaining kiln
improvements play an important role in nding the highest CO2
reduction potential, since co-processing, improved kiln refractories
and combustion system, and post-combustion technology had
already fullled the environmental constraint up to 77% of CO2
reduction target. Optimal cost increases at 78% ($ 133.40/t clinker)
and 79% ($ 136.46/t clinker) as remaining kiln improvement
methods are chosen.
At a 78% reduction target, increases in pre-heater stage help in
reducing thermal demand in the cement kiln from 44.97 kg/t
clinker of petroleum coke and 19.88 kg/t clinker of tire derived fuel
mixture, to 42.85 kg/t clinker of petroleum coke and 18.95 kg/t
clinker of tire derived fuel mixture; simultaneously reducing total
CO2 produced and CO2 produced from fuel combustion from
732.26 kg/t clinker and 199.28 kg/t clinker to 723.02 kg/t clinker
and 189.88 kg/t clinker. As discussed before, reduced fuel contrib-
utes to higher raw materials usage and CO2 emissions from raw
materials calcination, as can be seen in Table 8. Lower CO2 produced
reduced the amount of CO2 captured and released: from 541.87 kg/t
clinker and 190.39 kg/t clinker to 535.04 kg/t clinker and 187.99 kg/
t clinker.
At a 79% reduction target, the additional process control in the
kiln system reduces the thermal demand from a mixture of

900.00 600.00

800.00
500.00
CO2 (kg CO2/t clinker)

700.00

600.00 400.00

500.00
300.00
400.00

300.00 200.00

200.00
100.00
100.00

0.00 0.00
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00
CO2 emission from combustion 325.00 199.28 160.46 199.28 199.28 199.28 199.28 199.28
CO2 emission from decarbonation 531.68 532.98 524.88 532.98 532.98 532.98 532.98 532.98
CO2 released 856.68 732.26 685.35 285.58 285.58 285.58 285.58 190.39
CO2 captured 0.00 0.00 0.00 446.68 446.68 446.68 446.68 541.87
CO2 produced 856.68 732.26 685.35 732.26 732.26 732.26 732.26 732.26

CO2 reduction target (%)

Fig. 9. Effects of CO2 reduction target on total CO2 produced, captured and released.
146 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149

Table 9
Selection of CO2 mitigation strategies for increase of clinker production rate.

Forecast production (25.2 Mt cement 23.94 Mt clinker; 95% clinker); CO2 reduction target 30%; TSR 30%

Production increase (%) Base 0 20 40 60 80 100 120 140 160 180 200 220 240
Total cost ($M) 2159 2422 2906 3391 3875 4360 4845 5471 7692 8331 8973 9614 10,331 Inf
CO2 emission (Mt) 20.51 6.84 8.20 9.57 10.94 12.31 13.67 14.26 10.94 11.85 12.76 13.67 14.35 Inf
Fuels selection
Fossil fuel PC PC PC PC PC PC PC Coal PC PC PC PC Coal Inf
Non fossil fuel TDF TDF TDF TDF TDF TDF TDF TDF TDF TDF TDF TDF Inf
Kiln improvements
Refractories Inf
Control system Inf
Combustion system Inf
Increase preheater Inf
Carbon capture and storage
Oxy-fuel Inf
Post combustion Inf

25 12,000

10,000
20
CO2 emission (Mt CO2)

8,000

Total cost ($M)

15

6,000

10
4,000

5
2,000

0 -
Base 0 20 40 60 80 100 120 140 160 180 200 220
Production increase (%)

Fig. 10. CO2 emissions for increased of clinker production rate.

petroleum coke and tire derived fuel to the lower caloric value of higher cost raw materials (clay) and also from the substitution of a
coal and tire derived fuel mixture. The higher production cost is due lower cost petroleum coke to a higher cost coal. This substitution,
to high newly selected technology installation cost, addition of however, helped in reducing CO2 emission from fuel as coal possess

30% TSR 30% TED 30% NCV Coal 30% NCV PC 30% NCV TDF

-3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5

Changes in production cost (%)

Fig. 11. Sensitivity analysis for production cost.

 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149
30% TSR 30% TED 30% NCV Coal
-14 -12 -10 -8 -6 -4 -2
Changes in CO2 emission (%)
Fig. 12. Sensitivity analysis for CO2 emission.
lower carbon content as compared to petroleum coal (from
189.88 kg/t clinker to 160.46 kg/t clinker). Lower fuels usage is
compensated by additional raw materials. From Table 8 it can be
seen that clay is one of the raw materials used as opposed to the
usual limestone, sand and iron source combination. The raw ma-
terials combination however produces lower CO2 emission albeit at
a higher cost (contributed by clay) because of the lower CaO and
MgO in the raw materials. With a lower total CO2 produced, the
amount of CO2 captured and released is also reduced to 512.03 kg/t
clinker and 179.90 kg/t clinker.
4.3. Clinker production rate on CO2 emission and technology
selections
This section discusses the implication of technology selection
and CO2 emissions trends if the cement production is increased by
20%. The results are summarised in Table 9. The base case results
are included for comparison's sake.
Table 9 shows the selection of mitigation technologies as the
production is increased by 20%. For the base case, there is no
reduction target nor are mitigation methods applied. Once the
ability to select mitigation methods - 30% TSR and 30% reduction
target - are applied at the base production value, refractories
improvement and combustion system improvement are selected.
Oxy-fuel combustion is selected as the capture system and fuels
consumed are PC and TDF. Technologies selected remains the
same until a 100% of production rate increase; at a 100% pro-
duction rate increase, all of the kiln improvements are selected
while for co-processing, lower carbon coal is used instead of PC.
Technology selections change at 120% where post combustion
capture is chosen instead of oxy-fuel capture. Kiln improvements
selected are improvement in refractories and combustion system.
Fuels consumed are PC and TDF. Selections remain the same until
200% production increase; at 220%, fuels consumed changes from
PC to coal while the rest of the mitigation methods selected are
the same. The results reached are infeasible once a 240% pro-
duction increase is applied; this indicates that the allowable
production rate reached its maximum rate within the 220%e
240% range.
Fig. 10 shows the CO2 emissions from increased clinker pro-
duction rate. As expected, CO2 emission shown in increases
trend as the production increases. For the base case production,
about 20.51 Mt CO2 emitted from 23.94 Mt clinker productions.
If the clinker production remain as the base case with the
147
30% NCV PC 30% NCV TDF
0 2 4 6 8 10 12 14
mitigation option taking place, the CO2 emission reduces from
20.51 Mt CO2 to 6.84 Mt CO2. It was found that CO2 emission
increases up to 14.26 Mt CO2 for a 120% increase in production
rate. However, the emission decreases to 10.94 Mt CO2 for a
140% increase in production rate. This occurred due to the
changes of capture technology selection, where post combustion
a higher capture efciency is selected rather than the oxy-fuel
capture system. The CO2 emission then increases as the pro-
duction rate increases afterwards until a 220% production rate
increase. The optimal production cost for increased production
rate increases as the production rate is increased. From $
2159 M at base case to $ 10,331 M at 220% production rate
increased.
4.4. Sensitivity analysis
Sensitivity analysis was performed to identify parameters that
affect the cost of clinker production and CO2 emission. For this
analysis, three parameters are evaluated: TSR, TED and NCV. TSR is
assumed to be changed by 30%, TED by 30% and NCV by 30%.
Fig. 11 shows that TED has a signicant impact on production
cost. In this study, TED represents the amount of fuels usage from
various fuel options such as coal, petroleum coal, refused derived
fuels, sewage sludge and tire derived fuel. Unique operating char-
acteristic of fuel could affect the energy and heat generated to be
used for cement manufacturing process. Therefore, to achieve 30%
increased of TED, PC and TDF are more favourable fuel option in
order to satisfy the energy and heat demand requirement. PC is the
most expensive fuel and lead to an incremental rise of production
cost by more than 3%.
Fig. 12 shows the effects of TSR, TED and NCV on CO2 emission.
As illustrated, TED has signicant impact on CO2 emission. This is
due to the selection of PC to achieve increased of TED since PC has
the highest CO2 emission factor as compared to other fuels option
with 3.30 kg CO2/kg fuel. As a result, increased of TED by 30%
caused an increase of CO2 emission up to 11%. It is therefore vital to
improve the TED to reduce the cost of clinker production cost and
emissions.
5. Conclusion
A mixed integer model has been successfully developed
throughout this work to determine single and multiple mitigation
measures for cleaner cement production. Various mitigation
148 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149
measures considered in this work include co processing, kiln pro-
cess improvement and CCS. The effect of CO2 emission reductions
and the operational cost for different technology selections while
satisfying cement quality standard is quantied, so it is an advan-
tage for designing and operating cleaner cement production
manufacturing.
The highest possible CO2 emissions reduction that can be ach-
ieved by combining all the discussed measures is 79%. The optimal
cost at 79% reduction is $ 136.46/t clinker and the CO2 emitted to
the atmosphere of 179.90 kg/t clinker. An increase in the optimal
cost is expected as the additional costs of technology installation
are taken into consideration. The raw materials usage increases
where 1309.32 kg/t clinker limestone, 37.24 kg/t clinker clay,
128.19 kg/t clinker sand and 15.03 kg/t clinker iron source are used.
The use is increased in order to compensate the decreased fuels
usage (decrease in clinker composition contributed by fuel ash).
The fuels consumption is 42.47 kg/t clinker coal and 17.06 kg/t
clinker TDF.
The production of 23.94 Mt clinker expected in 2020 requires
strategies to imminently reduce the increase in CO2 emissions. The
models are then used to discover the combination of mitigation
strategies that must be implemented in order to reduce the emis-
sion and maximum production. Using 30% as a CO2 reduction
target, maximum production is achieved at 220% increased pro-
duction with the implementation of co-processing, two kiln im-
provements measures and a post combustion capture system as the
capture technology chosen, with a total cost of $ 10,331 M.
The implementation of post combustion capture can be affected
by the presence of other pollutants, especially sulfur dioxide (SO2)
and nitrogen oxides (NOx) due to the reaction of MEA solvent with
the acidic compounds to form salt. SO2 and NOx emissions from the
cement plant are also subjected to the emission standard limit in
Environmental Quality (Clean Air) Regulations, indicating the need
to mitigate multiple pollutants from the cement plant. Clinker
production is the main source of SO2 and NOx emission (Wang et al.,
2016). The issue of multiple pollutants will be addressed in a future
publication.
Acknowledgement
Special thanks go to the Universiti Teknologi Malaysia (UTM)
and the Ministry of Higher Education of Malaysia. This research is
funded by GUP Research Grant (Q.J130000.2544.10H67).
References
Adebiy, K.A., Akinyemi, O.O., Akintan, A.L., Ajayeoba, A.O., 2015. Modelling the
hazard control in a wet process plant e a case of cement plant. Braz. J. Oper.
Prod. Manag. 12, 338e348.
Ariyaratne, W.K.H., Melaaen, M.C., Eine, K., Tokheim, L.A., 2010. Meat and bone meal
as a renewable energy source in cement kilns: investigation of optimum
feeding rate. In: International Conference on Renewable Energies and Power
Quality (ICREPQ'11), 13th e 15th April, 2010, Spain, PL: 609.
Atmaca, A., Yumrutas, R., 2014. Analysis of the parameters affecting energy con-
sumption of a rotary kiln in cement industry. Appl. Therm. Eng. 66, 435e444.
BaeShammakh, M., Caruso, H., Elkamel, A., Croiset, E., Douglas, P.L., 2008. Analysis
and optimization of carbon dioxide emission mitigation options in the cement
industry. Am. J. Environ. Sci. Vol. 4 (5), 482e490.
Benhelal, E., Zahedi, G., Shamsaei, E., Bahadori, A., 2013. Global strategies and po-
tentials to curb CO2 emissions in cement industry. J. Clean. Prod. 51, 142e161.
Carpio, R.C., Junior, F.S., Coelho, L.S., Silva, R.J., 2008. Alternative fuels mixture in
cement industry kiln employing particle swarm optimization algorithm. J. Braz.
Soc. Mech. Sci. Eng. 4, 335e340.
Castan~o
n, A.M., GarcaeGranda, S., Guerrero, A., Lorenzo, M.P., Angulo, S., 2015.
Energy and environmental savings via optimisation of the production process at
a Spanish cement factory. J. Clean. Prod. 98, 47e52.
Cembureau, 2013. Cements for a Lowecarbon Europe: a Review of the Diverse
Solutions Applied by the European Cement Industry through Clinker Substi-
tution to Reducing the Carbon Footprint of Cement and Concrete in Europe. The
European Cement Association, Brussels, Belgium.
Damartzis, T., Papadopoulos, A.I., Seferlis, P., 2016. Process owsheet design opti-
mization for various amine-based solvents in post-combustion CO2 capture
plants. J. Clean. Prod. 111, 204e216.
European Cement Research Agency (ECRA), 2009. Development of State of the Art
Techniques in Cement Manufacturing: Trying to Look Ahead. Report from ECRA
and Reviewed by ECRA and Cement Sustainability Initiative (CSI). European
Cement Research Agency (ECRA)/Cement Sustainability Initiative (CSI), Dus-
seldorf, Germany.
GarcaeGusano, D., Garran, D., Herrera, I., Cabal, H., Lecho n, Y., 2015. Life Cycle
Assessment of applying CO2 postecombustion capture to the Spanish cement
production. J. Clean. Prod. 104, 328e338.
Green, D.W., Perry, R.H., 2008. Perry's Chemical Engineers' Handbook, eighth ed.
McGraw Hill, New York, USA.
Gereca, L.P., Torres, N., JuarezeLopez, C.R., 2015. The coeprocessing of municipal
waste in a cement kiln in Mexico. A lifeecycle assessment approach. J. Clean.
Prod. 107, 741e748.
Huisingh, D., Zhang, Z., Moore, J.C., Qiao, Q., Li, Q., 2015. Recent advances in carbon
emissions reduction: policies, technologies, monitoring, assessment and
modeling. J. Clean. Prod. 103, 1e12.
IEA Greenhouse Gas R&D Programme (IEA GHG), 2008. CO2 Capture in the Cement
Industry. Mott MacDonald Ltd, Brighton, UK.
Kajaste, R., Hurme, M., 2015. Cement industry greenhouse gas emissions e man-
agement options and abatement cost. J. Clean. Prod. 112, 4041e4052.
Kara, M., Gunay, E., Tabak, Y., Durgut, U., Yildiz, S., Enc, V., 2010. Development of
refuse derived fuel for cement factories in Turkey. Combust. Sci. Technol. 183
(3), 203e219.
Katsioti, M., Tsakiridis, P.E., Giannatos, P., Tsibouki, Z., Marinos, J., 2009. Character-
ization of various cement grinding aids and their impact on grindability and
cement performance. Constr. Build. Mater. 23 (5), 1954e1959.
Knudsen, J.N., Bade, O.M., Askestad, I., Gorset, O., Mejdell, T., 2014. Pilot plant
demonstration of CO2 capture from cement plant with advanced amine tech-
nology. Energy Procedia 63, 6464e6475.
Kookos, I.K., Pontikes, Y., Angelopoulos, G.N., Lyberatos, G., 2011. Classical and
alternative fuel mix optimization in cement production using mathematical
programming. Fuel 90 (3), 1277e1284.
Kuramochi, T., Ramrez, A., Turkenburg, W., Faaij, A., 2012. Comparative assessment
of CO2 capture technologies for carboneintensive industrial processes. Prog.
Energy Combust. Sci. 38, 87e112.
Mikulci
c, H., von Berg, E., Vujanovi
c, M., Duic, N., 2014. Numerical study of coering
pulverized coal and biomass inside a cement calciner. Waste Manag. Res. 32 (7),
661e669.
Moya, J.A., Pardo, N., Mercier, A., 2010. Energy Efciency and CO2 Emissions e
Prospective Scenarios for the Cement Industry. Ofce for Ofcial Publications of
the European Communities, Luxembourg.
Ogbeide, S.O., 2010. Developing an optimization model for CO2 reduction in cement
production process. J. Eng. Sci. Technol. Rev. 3 (1), 85e88.
Oyepata, J.S., Obodeh, O., 2015. Cement production optimization modeling: a case
study BUA plant. J. Eng. Technol. 7 (4), 53e58.
Rodrguez, N.H., MartnezeRamrez, S., BlancoeVarela, M.T., Donatello, S.,
Guillem, M., Puig, J., Fos, C., Larrotcha, E., Flores, J., 2013. The effect of using
thermally dried sewage sludge as an alternative fuel on Portland cement clinker
production. J. Clean. Prod. 52, 94e102.
Stadler, K.S., Poland, J., Gallestey, E., 2011. Model predictive control of a rotary
cement kiln. Control Eng. Pract. 19, 1e9.
Trading Economics, 2016. Malaysia Cement Production Forecast 2016 e 2020.
Accessed: February 5th, 2016. www.tradingeconomics.com/malaysia/cement-
production/forecast.
Tsakalakis, K.G., Stamboltzis, G.A., 2008. Correlation of the Blaine value and the size
of the cement particle size distribution. ZKG Int. 61 (3), 60e68.
U.S. Geological Survey (USGS), 2014. Cement Statistics and Information, and Other
Commodities. US Geological Survey. Internet: www.minerals.usgs.gov/
minerals/pubs/commodity/cement/mcs-2014-cemen.pdf. Accessed: February
5th, 2016.
Vaccaro, M., 2006. Burning alternative fuels in rotary cement kilns. In: IEEE Cement
Industry Technical Conference, 2006. 9 e 14 April, 2006, Phoenix, AZ, 127 e 136.
http://dx.doi.org/10.1109/CITCON.2006.1635711.
Valderrama, C., Granados, R., Cortina, J.L., Gasol, C.M., Guillem, M., Josa, A., 2012.
Implementation of best available techniques in cement manufacturing: a life-
cycle assessment study. J. Clean. Prod. 25, 60e67.
Vatopoulos, K., Tzimas, E., 2012. Assessment of CO2 capture technologies in cement
manufacturing process. J. Clean. Prod. 32, 251e261.
Wang, W., Jiang, D., Chen, D., Chen, Z., Zhou, W., Zhu, B., 2016. A Material Flow
Analysis (MFA) e based potential analysis of ecoeefciency indicators of
China's cement and cement-based materials industry. J. Clean. Prod. 112,
787e796.
Wennersten, R., Sun, Q., Li, H., 2015. The future potential for Carbon Capture and
Storage in climate change mitigation e an overview from perspectives of
technology, economy and risk. J. Clean. Prod. 103, 724e736.
Willett, J.C., 2011. 2010 Minerals Yearbook: STONE, CRUSHED. U.S. Department of
the Interior. U.S. Geological Survey, USA.
Worrell, E., Kermeli, K., Galitsky, C., 2013. Energy efciency improvement and cost
saving opportunities for cement making. In: An ENERGY STAR Guide for En-
ergy and Plant Managers. Ernest Orlando Lawrence Berkeley National Labora-
tory, California, USA.
Yue, D., You, F., 2014. Reformulation-linearization method for global optimization of
 S.A. Ishak et al. / Journal of Cleaner Production 136 (2016) 133e149 149

mixed integer linear fractional programming problems with application on 949 e 954(2014). Elsevier B.V, Amsterdam, Netherlands.
sustainable batch scheduling. In: Klemes, J.J., Varbanov, P.S., Liew, P.Y. (Eds.), Zhang, L., Mabee, W.E., 2016. Comparative study on the life-cycle greenhouse gas
Computer Aided Chemical Engineering 24 European Symposium on Computer emissions of the utilization of potential low carbon fuels for the cement in-
Aided Process Engineering, Budapest, Hungary, June 15 e 18, 2014, Volume (33), dustry. J. Clean. Prod. 122, 102e112.