EnvironmentAsia

Available
online
at www.tshe.org/EA
Available
online
at www.tshe.org/EA
EnvironmentAsia2 8(2)
(2015)
9-17
EnvironmentAsia
(2009)
50-54
DOI10.14456/ea.2015.18

The international journal published by the Thai Society of Higher Education Institutes on Environment

Genotoxicity Assessment of Mercuric Chloride in the Marine Fish Therapon jaruba

A Multi-Objective Unit Commitment Model for Setting Carbon Tax to
Nagarajan
Nagarani,
Velmurugan Case
Janaki Devi
Reduce
CO2Arumugam
Emission:Kuppusamy
ThailandsKumaraguru,
Electricity Generation

and Chandrasekaran Archana Devi

Nuchjarin Intalar, Apirath Phusittrakool, Chawalit Jeenanunta, Aussadavut Dumrongsiri and Pisal Yenradee
Center for Marine and Coastal Studies, School of Energy, Environment and Natural Resources,
Madurai
Kamaraj
University,
Madurai-625021,
Sirindhorn
International
Institute
of Technology,
ThammasatIndia
University,
P.O. Box 22, Thammasat Rangsit Post Office, Pathumthani 12120, Thailand

Abstract
Abstract
aimtax
of the
present
study was toinstrument
standardize
to assess
the predictive
of the
the carbon
cytogenetic
analysis
The
Carbon
policy
is a cost-effective
forand
emission
reduction.
However,value
setting
tax is one
of the
by
Micronucleus
(MN)
testmakers
in fish erythrocytes
biomarker
environmental
contamination.
Micronucleus
challenging
task for
policy
as it will leadastoa higher
pricefor
of marine
emission-intensive
sources
especially the
utility price.
frequency
baseline
in erythrocytes
was evaluated
in and
a common
chemical
wasare
determined
In a large-scale
power
generation system,
minimizing
the genotoxic
operationalpotential
cost andofthe
environmental
impact
conflicting
in
fish experimentally
exposed
inthe
aquarium
undersolution.
controlled
conditions.
Fish (Therapon
jaruba)
were exposed
96
objectives
and it is difficult
to find
compromise
This
paper proposes
a methodology
of finding
a feasiblefor
carbon
hrs
heavylevel
metal
(mercuric
chloride).
Chromosomal
damageWe
was
determined
as micronuclei
in
tax to
ratea single
on strategic
using
the operational
unit
commitment model.
present
a multi-objective
mixedfrequency
integer linear
fish
erythrocytes.
Significant
increase
in MN frequency
of fish
exposed
mercuric
programming
model
to solve the
unit commitment
problemwas
andobserved
considerintheerythrocytes
environmental
impacts.
The to
methodology
chloride.
Concentration
ppmtax
induced
thethe
highest
MN
frequencyoperating
(2.95 micronucleated
cells compared
of analyzing
of the effectofof0.25
carbon
rates on
power
generation,
cost, and COcells/1000
2 emission is also provided.
to
MNcell/1000
cells inbetween
controltotal
animals).
Thecost
study
micronucleus
test,
as
an
index
of cumulative
The1 trade-off
relationship
operating
andrevealed
total COthat
emission
is
presented
in
the
Pareto-optimal
curve to
2
exposure,
be a sensitive
model
to evaluateongenotoxic
in fish
controlled
conditions.
analyze theappears
feasibletocarbon
tax rate that
is influencing
electricitycompounds
operating cost.
Theunder
significant
outcome
of this paper
is a modeling framework for the policy makers to determine the possible carbon tax that can be imposed on the electricity
Keywords:
generation. genotoxicity; mercuric chloride; micronucleus
Keywords: CO2 emission reduction; multi-objective unit commitment; carbon tax policy; pareto-optimal curve; energy policy

1. Introduction
1. Introduction
In India, about 200 tons of mercury and its
compounds
introduced
into thedioxide
environment
Global are
warming
and carbon
(CO 2 )
annually
as
effluents
from
industries
(Saffi,
1981).
emissions reduction have become important issues. The
Mercuric
chloride
has CO
beenpolluter
used ininagriculture
as a
power sector
is a major
many countries.
2
fungicide,
in
medicine
as
a
topical
antiseptic
and
Under the pressure of global warming, it is crucial
for
disinfectant,
and
in
chemistry
as
an
intermediate
in
government to impose the effective policy to promote
the
of other mercury
compounds.
The
emission reduction.
Thailands
CO2 emissions
CO2production
contamination
of
aquatic
ecosystems
by
heavy
levels are continuing to rise in accordance with the
metals
andvolume
pesticides
has gained
increasing
attention
increased
of national
energy
consumption.
in
decades.
exposure totoreduce
and
recent
There are
several Chronic
policies introduced
accumulation
of
these
chemicals
in
aquatic
biota
CO2 emissions. The most widely proposed is a carbon
can
result because
in tissueit burdens
that produce
adverse
tax policy
is a cost-effective
instrument
for
effects
not
only
in
the
directly
exposed
organisms,
emission reduction (Baranzini et al., 2000; Nagurney
but
also2006).
in human
beings.
et al.,
Carbon
tax is the tax paid by polluters
Fish
provides
suitablefossil
model
for and
monitoring
who emit CO2 froma burning
fuels
releasing
aquatic
genotoxicity
and
wastewater
quality
CO2 into the atmosphere. In fact, CO2 emissions are
because
of its ability
to in
metabolize
xenobiotics
and
already implicitly
taxed
every country
even in those
accumulated
pollutants.
A
micronucleus
assay
developing countries that the Kyoto protocol hashas
not
been
used
successfully
in 2000).
severalHowever,
species (De
Flora,
targeted
(Baranzini
et al.,
setting
the
et
al.,
1993,
Al-Sabti
and
Metcalfe,
1995).
The
carbon tax is a challenging task for policy makers as it
micronucleus
(MN)
test
has been developed
will lead to higher
price of
emission-intensive
sources,
together
with
DNA-unwinding
assays
as
especially utility prices. The different carbon
tax rate
perspective
methods
for
mass
monitoring
of
imposed in each country depends on the regulation and
clastogenicity
several factors.and genotoxicity in fish and mussels
(Dailianis
al., 2003).have formulated mathematical

Severaletresearchers
The
MN
have
beensuch
successfully
used as
models and tests
applied
tools
as the HERMES
a measure of genotoxic stress in fish, under both

laboratory and field conditions. In 2006 Soumendra

et
al., made an attempt
to detectand
genetic
biomarkers
macroeconomic
model (Cosmo
Hyland,
2013) and
in
fish species,
Labeo
bata and
Oreochromis
thetwo
dynamic
Computable
General
Equilibrium
(CGE)
mossambica,
by MN etand
binucleate
(BN)
model (Wachirarangsrikul
al., 2013;
Thepkhun
et al.,
erythrocytes
in
the
gill
and
kidney
erythrocytes
2013) to explore the impact of the taxation of CO2 on
exposed
to thermal
plant
discharge
national economies
thatpower
undertake
carbon
reductionatin
Titagarh
Thermal
Power
Plant,
Kolkata,
the future. Some of the studies estimate theIndia.
optimal or
The
present
study
was
conducted
to
determine
appropriate carbon tax to impose on the electricity cost
the
of the
thestudies
heavy metal
compound
(Luacute
et al.,genotoxicity
2010). Most of
divided
the carbon
HgCl
in
static
systems.
Mercuric
chloride
toxic,
2
tax into
three scenarios which are a baselineis (carbon
solvable
in
water
hence
it
can
penetrate
the
aquatic
tax is zero), the average carbon tax rate, and high tax
animals.
studies
withCO
native
fish species
rate. TheMutagenic
results show
that the
2 emission can be
represent
an reduced
important
effort in the
determining
the
significantly
by imposing
fairly moderate
potential
effects
of
toxic
agents.
This
study
was
tax rate (Wachirarangsrikul et al., 2013) and highest
carried
evaluate
use of the
micronucleus
carbon out
tax to
(Cosmo
andthe
Hyland,
2013).
The carbon
test
(MN) for
the the
estimation
of aquatic
pollutionof
tax policy
shows
effectiveness
and potential
using
marine
edible
fish
under
lab
conditions.
CO emission reduction and the positive impact on the
2

economic and environmental improvement in long term.

2.
methodscarbon tax rate shows the
Materials
Although,and
the proposed
significant reduction of CO2 emission, it cannot suggest
2.1.
Sample
Collection
a policy
maker
how much it costs to reduce the CO2
emission to a certain amount based on the current
The fish
species
selected for
theexisting
present study
carbon
tax rate
in Thailand
and
power
was
collected
from
Pudhumadam
coast
of Gulf
of
generation capacity. It is important for the
strategic
Mannar,
Southeast
Coast
of
India.
Therapon
policy maker to work in a cost effective way to reduce
jarbua
belongswhile
to the
order Perciformes
of the
CO2 emissions
maintaining
minimum operating
family
Theraponidae.
The
fish
species,
Therapon
cost. In spite of this, the power sector intends to improve
jarbua
(6-6.3 efficiency
cm in length
4-4.25
g in weight)
the operating
andand
reduce
overall
operation
was selected for the detection of genotoxic effect

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

costs as much as possible while satisfying all demands.

Power generation planning is a challenging task
because of the complications of the generation process,
transmission, and generation of electric power which
lead to a variety of issues in decision making. The
power generator uses the unit commitment (UC) model
to plan the operation of the generator unit whether to
turn the unit on or off and the amount of power to be
generated in each period. In general, the UC model
is used to minimize the operating cost which mainly
include the fuel cost. In this paper, the multi-objective
unit commitment model is proposed with two
conflicting objective functions: minimization of the
operating cost and CO2 emission. In the multi-objective
problem, an optimal solution is generally more difficult
to find. Therefore, the compromise solution can be the
best for all conflicting objectives. Many solutions are
presented in the Pareto-optimal solutions on a trade-off
curve between the fuels cost and emission costs by using
a multi-objective optimization (MO) model (Catalo
et al., 2008) and another using the decommitment
procedure of unit commitment (Yamashita et al., 2010).
These solutions showed the trade-off points on the curve
for a decision maker to select. However, the studies
did not take a carbon tax into account. The models are
tested on the simulated model which contains only 10
generating units (Yamashita et al., 2010).
From the strategic policy level view point, the
impacts of various carbon tax rates on the operational
level of unit scheduling are investigated. By considering
different rates of the carbon tax, the impacts of CO2
emission from strategic level planning that lead to the
effect on the unit schedules and power generations of
power system at the operational level are investigated.
The model is coded and solved by using a commercial

optimization package, IBM ILOG CPLEX, which is a

computationally efficient solver that can handle large
scale mixed integer programming problems.
2. Materials and Methods
2.1. Data of case study
A case study of a large scale power generation
system in Thailand is used to demonstrate the
methodology. The system comprises the combination
of the thermal, gas turbine, hydro and combined
cycle generators which are 51 power plants and 171
generators. The power plants are located in five main
regions of Thailand: central, metro, north, northeast,
and south regions. There are five main sources of fuel
types: natural gas, coal/lignite, hydro, fuel oil, and diesel
oil. The generating units generate electricity by using
different types of fuel. Each fuel type has a different
efficiency, price and CO2 emission intensity. For daily
planning, input of the models is the demand on Monday
12 October, 2011 which directly obtained from EGAT.
The time period is started from 12:00 AM of day one
until 12:00 AM of day two. It is scheduled in a half-hour
schedule with a time horizon of 24 hours (48 periods)
with the time slot of 30 minutes as shown in Fig. 1.

There are two fuel modes in each generator which
are the single and mixed fuel mode. The single fuel
mode means the generating unit is allowed to use only
one type of fuel to generate electricity. The mixed fuel
mode means the generating unit is allowed to mix two
types of fuel and the mixture is considered as a new
fuel type.

The combination of fuel types produces different
levels of CO2 intensity. Therefore, we assign different

Figure 1.
Half an
houran
demand
for one for
dayone
for day
electricity
in Thailand
from EGAT
Figure
1. Half
hour demand
for electricity
in Thailand
fromdata
EGAT data

There are two fuel modes in each generator which are the single and mixed fuel mode.
The single fuel mode means the generating unit is allowed to use only one type of fuel to
10the generating unit is allowed to mix two types
generate electricity. The mixed fuel mode means
of fuel and the mixture is considered as a new fuel type.

The model is formulated to solve the unit commitment (UC) problem to minimize tot
operating cost regardless of the CO2 emission. The proportion of the fuel usage for each type
obtained by this model.
N. Intalar
et al. / EnvironmentAsia
8(2)objective
(2015) 9-17
1) Objective
Function: The
function

is the minimization of the total operatin

cost which includes the fuel cost and the startup cost of the generating units over the plannin
values of CO2 intensity depending
the fuel to
type
Where
setother
of power
generating
(thermal,
horizon,onsubjected
the load
demandI is
and
individual
unitunits
constraints
such as the fu
intensity
for
usage of each generating unit.
The
CO
gas
turbine,
combined
cycle
and
hydro),
T
is
the
length
usage. The2objective function is formulated as follows:
each generator and fuel is shown in Table 1. For carbon
of the planning horizon,
U is the total number of the

i
t
t
t

tax imposed in Thailand, in order to estimate

carbon
generating
units,
F
is
the
set
of
fuel
types (lignite, natural

Min a
c
FU
P
SC
(1
f
i
i
i si
u
U
f
F
i I tT

tax in Thailand, a baseline for setting a carbon

tax
gas
and
diesel
oil).

is Social cost of carbon (SCC). The

socialI cost

The
fuel costunits
function
is calculated
the
Where
is setofof power
generating
(thermal,
gas turbine,by
combined
cycle an
i
carbon (SCC) is used by United
StatesT Environmental
of cost
of the
fueltotal
typenumber
f used in
i (c f),
hydro),
is the length of themultiplication
planning horizon,
U is
of unit
the generating
units, F
Protection Agency and other
federal
agencies
to the
amount
used
in unit i at time t (FUit), and
the set
of fuel
types (lignite,
natural
gas of
andfuel
diesel
oil).
t
The fuel cost
is calculated
estimate the economic damages associated
withfunction
a the production
of by
unitthe
i atmultiplication
time t (Pi ). of cost of fuel type f used
i
t
small increase in CO2 emissions.
exactof fuel
used
Theinstartup
cost
(SC
is the
cost charged when
unit iHowever,
( c f ), the the
amount
unit i at
time
t (i)FU
i ), and the production of unit i at tim
value of SCC is not defined. There
are
more
than
200
any
generating
unit
is
started
to
generate
the electricity.
t
t ( Pfrom
i ). 15 baht/tCO2 to
estimated values of SCC starting
If the unit i is started at time t, sit has value of 1 and 0
24,000 baht/tCO2 (Policy Brief). Since,
has (SCotherwise.
charged
generating
unitstartup
is started to genera
The
startupwhen
cost isany
calculated
by the
TheThailand
startup cost
i) is the cost
t
not practically been imposed athe
carbon
tax
on
electricity
cost
of
unit
i
multiplied
by
startup
status.
electricity. If the unit i is started at time t, si has value of 1 and 0 otherwise. The startup co
cost, therefore, this paper relies
a carbon by
taxthe
rate
on cost
of
2) unit
General
System by
andstartup
Unit Constraints:
is calculated
startup
i multiplied
status.
IPCC Emission Trading value in 2009. Based on the
System Load Balance Constraint: There are five
Intergovernmental Panel on Climate Change
(IPCC),
an and
zones
ofConstraints:
power generation in Thailand which are
2) General
System
Unit
average carbon tax for Emission Trading
Scheme
(ET)
central,
metro,
zones. in Thailan
System Load Balance Constraint: northern,
There arenortheast,
five zonesand
ofsouthern
power generation
in Europe Emission Trading Scheme
in 2009
total amount
of production
of unitzones.
i in zone
at
which (EU
are ETS)
central,
metro, The
northern,
northeast,
and southern
Thez total
amount
is 349 baht/tCO2 (Wattanakulcharat and Wongsa, 2011). time t (Pizt) plus the total
of power transmitted through
t
production of unit i in zone z at time t ( Piz ) plus the total of power transmitted throug
transmission line from demand zone i to demand zone
t
Ltij ) minus
zonet (L
i to
demand
j at of
time
t ( pumped
2.2. Problem Formulation transmission line from demand
j at time
thezone
amount
water
by the amount
ij) minus
t
time
t t (pwkp) must meet load demand
water pumped by pump k inpump
plant kpinatplant
timept at
( pw
) must meet load demand d in zone z at time
kp
In this section, two modelst are introduced and d in zone z at time t (dzt). We also consider the transmis). We also consider
transmission
line among
capacity
among
and power
the pump
( d zminimization
formulated; the operating cost
(Modelthe sion
line capacity
zones
andzones
the pump
in power in th
2.2. Problem
Formulation
A) and the utility optimizationequation.
model (Model
B).
The constraint
is expressed
as follows:
this equation.
The constraint is expressed as follows:

t
tcost t
In
thisOperating
section, Cost
two Minimization
models are Model
introduced
(2)
pizt the Loperating
pwkp
= d z , z Z , t T
2.2.1.
(Modeland
A) formulated;
ij
minimization
(Model
A)
and
the
utility
optimization
model
(Model
B).

=
,

i
I
i
z
j
j
k

The model is formulated to solve the unit
commitment (UC) problem to minimize total operating
(2)
2.2.1. Operating Cost Minimization Model (Model
A) Unit Constraints: This constraint is the turn on/off status for each generatin
General
cost regardless of the CO2 emission. The
proportion
of
The model is formulated to solve
the
unit
commitment
problem
tothis
minimize
total horizon starts, then it is not turned o
If the
unit
i ismodel.
already(UC)
operating
when
scheduling
the fuel usage for each type isunit.
obtained
by
this

General
Unit
Constraints:
This constraint is the
operating cost regardless of the CO2 emission.
The
proportion
of
the
fuel
usage
for
each
type
isscheduling horizon starts, then it w
at
the
start
as
(3).
If
the
unit
i
is
already
on
when
this

1) Objective Function: The objective function is the turn on/off status for each generating unit. If the unit i
obtained by this model.
eithercost
be turned
off at time is1 already
or remain
operating
atthis
time
1 as (4). horizon
If the unit
i is off when th
minimization of the total operating
which includes
operating
when
scheduling
starts,
scheduling
horizon
starts,
then
it
is
not
turned
off
at
the
start
as
(5).
If
the
unit
theObjective
fuel cost and
the startup
of the function
generating
then it is not
turned
onoperating
at the start as (3). If the unit i iisis off when th
1)
Function:
The cost
objective
is units
the
minimization
of the
scheduling
horizon
starts,
then the turn
on total
variable
must be the same as operating variable
overincludes
the planning
horizon,
subjected
to the load
demand
already on when
this scheduling
horizon starts, then it
cost which
the fuel
cost and
the Ifstartup
cost
theatgenerating
(6).
the unit
i isofoff
time t and units
on at over
time the
t+1,planning
then it was turned on at time t+1 as (7). Th
and
other
individual
unit
constraints
such
as
the
fuel
will
either
be
turned
off
at
time
1
horizon, subjected to the load demand
and other
constraints such as the fuel or remain operating
constraints
areindividual
expressed unit
as follows:
usage.
The
objective
function
is
formulated
as
follows:
at
time
1 as (4). If the unit i is off when this scheduling
usage. The objective function is formulated as follows:

horizon starts, then it is not turned off at the start as (5).

scheduling horizon starts,

Min c if FU it Pi t SC i s it If the unit i is off when this(1)
u
U
f
F
i I tT

then
the
turn
on
variable
must
be the same as operating

(1)
variable
as
(6).
If
the
unit
i
is
Where I is set of power generating units (thermal, gas turbine, combined cycle and off at time t and on at
hydro), T is the length of the planning horizon, U is the total number of the generating units, F is
CO2(lignite,
emissionnatural
intensitygas
value
the set ofTable
fuel 1.
types
and diesel oil).
The fuel cost function is calculated by the multiplication of cost of fuel type f used in
i
Type of fuel used in unit i at time t ( FU t Fuel
Type
kgCO2/MWha
unit i ( c f Generator
), the amount
i ), and the production of unit i at time

The combination of natural gas, fuel oil, and diesel oil

Thermal
The combination of natural gas and fuel oil
the cost charged when any generating unit is started to generate
The startup cost (SCi) isLignite
the electricity.
If the unit i is started
at time t, sitofhas
valuegas
of and
1 and
0 otherwise.
The startup cost
Gas Turbine
The combination
natural
diesel
oil
is calculated by the startup cost of
i multiplied
by startup
status.
Theunit
combination
of natural
gas and
diesel oil
Combined Cycle
Natural gas
a
2)
General
Systemfrom
andSutham
Unit Constraints:
Sources:
adopted
(2011) and EGAT (2010)

t ( Pi ).
t

System Load Balance Constraint: There are five zones of power generation in Thailand
which are central, metro, northern, northeast, and southern zones. The total amount of
t
production of unit i in zone z at time t ( Piz ) plus the total of power transmitted through
11
t
transmission line from demand zone i to demand zone j at time t ( Lij ) minus the amount of

water pumped by pump k in plant p at time t ( pwt ) must meet load demand d in zone z at time t

622
600
914
469
469
426

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

time t+1, then it was turned on at time t+1 as (7). The

constraints are expressed as follows:

si1 =si10s=i1 =
00

0
i

{i
Ii
|IpS Ii0|S >

{I{I
IS
p|i0p0}
>
0}0}(3)
i >

111 1 1
1i i i i i

s soso1o 1 1

i
{
iIi
{II{
Ip I| p| 0}
p 0}
0} (4)
SI |

s s0s 0 0
sss==0o 0 1

(5)

1
i

si = 0

0
0 0
Si S i 0 i

i {I I S | pi > 0}

max
for plant p (qfip ). Second, the generating unit i cannot
t
use more than one type of fuel f at time t (FUi ). The
constraints are expressed as follows:
(3) (3)(3)
qtfip qmax
p Pi , f Fi (15)
fp
tT iI
max(4) (4)(4)
t q
qtfip
FU

1fp (3) p Pii, f

I Fi (16)
i

tT iI

f F
i
{
(5)(5)
i Ii
{II{
IpSi0I|0S0p|i000}
pi0 0}
0}
SI |
FUit 1 (5)
iI

(3)
(3)
IISSI|S| pp|i ipi>0}
ii
i
{{I{

I
>0}0} (6)Spinning
(4)
Reserve Reserve
Constraint:
reserve
is provided to co
tThe spinning
max

Spinning
Constraint:
The
spinning
1 1 11 1 1
0 0 0
f F
qreserve
reserve
p
P, f F
(6)

si =sios=ii =
i

{i
Ii
|IpS Ii|S=
(6)
(6)
fip q fp
oi oi
{I{I
IS
p|i p0}
=i =
0}0} occur.
shortfalls
The
total
spinning
is
greater
than
is provided to compensate
when shortfalls occur. or
Theiequal toi the
tT iI
1 11 1 1
0 00
Reserve reserve
Constraint:
The
spinning
reserve
).total
required
(Spinning
The spinning
constraint
is expressed
as follows:
oo1i0tot1i1st 1t t1 t 1t 1 i
i i
{I{I
I
ItSI
| p||ipp
i 0}
0}
(4)
(4)greater

0}
t

(7)
is
or equal
to is
theprovided to c
osit is1sioii
,T

(7)(5)
oii oi oii si si
I
i
i
I
I SIS,IS{1,
ti{1,

{1,

, T1}
,
T1}

1} occur.
(7)
(7) t than
SI,
St ,
shortfalls
The
total
spinning
reserve
is
greater
than
or
FU
1
equal
i I to the

i constraint is expressed
t
max spinning reserve
( ). The
t P ,required
max
1
1
0
q

F
tq
t(6)
p

fip
fp ( ). The
iF
1= o
0p
0
required
constraint
isfpexpressed
(15)

f
ii

I|SSpower
=generation
0}
si1ssi
Generation
Limit Constraint:
as follows:
(
pI iqmaxfipunit
oiiunit
t tasfollows:
Pti ,f T Fi
0 0Limit

{power
IThe
{{
IIpower

IThe
p| |generation
pTi
0}
i Limit
ii
i
i (5)
ti
I 0}
iLimit
SI
Generation
Constraint:
The
generation
from
the
generating
at
t)time
Generation
Generation
Constraint:
Constraint:
The
power
generation
from
from
the
the
generating
generating
itime
atip(5)
at
time
tT iunit
1
1 1
i 111i i 1
i
i ii

fromt 1the generating

unit i at time t has upper and lower
t
1
tfollows
Spinning
Constraint: The spinning reserve is
s has
upper
andand
lower
bound
The
constraint
expressed
: :i I:
has
upper
upper
and
limits.
The
constraint
is
expressed
as as
follows
follows
olower
1obound
limits.
sit limits.
The
i
Iconstraint

I S ,ist
{1,is
,FU
T as
1}
max
t
(7)
1 i 1lower
1 bound
0
0expressed
oitit
IReserve
i
FU

1
TI is(17)
i 1
bound
limits.
The constraint
isI{expressed
as
follows:
(
p

p
)

ti
(6)
(6)
si s=
=
o
o

I
I
|
p
|
p
=
=
0}
0}
shortfalls
occur.
The
total
spinning
reserve
greater
i
i
i
(16) than o
i
i i
S S i i
f F 2.2.2. Utility Optimization

f
F

i
I
Model
(Model
B)
required unit
( ).i The
constraint
is expressed as follows:
t
t min
t tt 1ttt1 max
max power generation from the generating
t tt min
t max
at model
time
t(8)
1 Limit
oitoit
sii i
iip
iThe

I
I
I SI,S
t,ti
{1,
TI,
T,It1}
,
1}
(7)(7)
oGeneration
sConstraint:
poitopi
{1,
Ii
,,t
T
(8)
oimin
oopii oip1i

i
tConstraint:
TThe
T utilityThe
(8)
ip
optimization
(Model
B) istoformulated
the mu
i p
i p
iipo
i
Spinning
Reserve
spinning
reserve
is
provided
compensate
when
Spinning
Reserve Constraint: The spinning
reserve
is using
provided
to
has upper and lower bound limits. The constraint is expressed
as(8)
follows
: Utility
2.2.2.
Optimization
Model
(Model
B)that
technique
to
find
the
compromise
solution
minimizes
both
op
2.2.2.
Utility
Optimization
Model
(Model
B)
max
t
t
shortfalls occur. The commitment
total
spinning
reserve
is
greater
than
or
equal
to
the
spinning
reserve
shortfalls occur. The total spinning
reserve
is
greater
than
or
equal
to
the
( tp~ti ~is(Model
oimodel
pB)i )(Model
formulated
B)
t Tbytheusing
~
Generation
GenerationLimit
LimitConstraint:
Constraint:
The
The
power
power
generation
generation
from
from
the
the
generating
generating
unit
unit
i
at
i
at
time
time

The
utility
optimization
is
CO
emission
cost.
The
CO
emission
converted
to
monetary
cost
The
utility
optimization
model
is
using
mu
2
2
generating
required
(generating
). Theunits
constraint
is
expressed
follows:
).)( and
I )(fuel
Ramp-Up/Down
Constraints:
TheThe
generating
ramp-up
ramp-down
required
((
The
constraint
isi(expressed
( ) costs,
)as
and
ramp-down
ramp-down
) as follows:
Ramp-Up/Down
Ramp-Up/Down
Constraints:
The
units
units
ramp-up
ramp-up
()and
min
max
t
t Constraints:
t
rate.
The
objective
ofthethe
start-up
cost,
lower
Ramp-Up/Down
Constraints:
The
generating
units
formulated
using
multi-objective
unit commitment
as
hasupper
upperand
and
lower
bound
bound
limits.
limits.
The
The
constraint
constraint
isisexpressed
expressed
as
as
follows
follows
:technique
: consists
commitment
to find
compromise
solution
that operating
minimizescost,
bothan
o
o
p

o
p
i

I
,
t

T
(8)
i
i
i
i
i
erate
atrate
time
t.time
For
each
ramping,
we
divided
the
equation
into
two
stages
which
are
the
initial
at at
time
t.
For
t.
For
each
each
ramping,
ramping,
we
we
divided
divided
the
the
equation
equation
into
into
two
two
stages
stages
which
which
are
are
the
the
initial
initial
The
formulation
are
presented
as
follows.
CO
emission
cost.
The
CO
emission
is
converted
to
monetary
cost
by
usin
ramp-up () and ramp-down (%) rate at time t.costs.
For
each
technique
to
find
the
compromise
solution
that
2
max2 t
t
max
t andand
t and
age
andand
the
processing
stage.
Constraint
(9) (9)
and
(10)
the
initial
stages
for
ramp-up
stage
stage
and
the
processing
stage.
stage.
Constraint
Constraint
(9)
and
and
(10)
(10)
are
stages
for
(are
prate.
othe
initial
pinitial
stages

ramp-up
tramp-up
the
Tp~ifuel
Optimization
Model
(Model
B) cost.
(Utility
pfor
ooperating
)cost
costs,

2 emission
t cost,
T
The
objective
consists
of
start-up
operating
min
minprocessing
max
max
t the
t ramping,
t tdivided
t t the
iare
i the
i ) 2.2.2.
we
equation
into
two
stages
which
minimizes
both
and CO
(17) cost, a
i
i
o
o
p
p

p
p

o
o
p
p

I
,
I
t
,

T
T
(8)
(8)
mp-down,
Constraint
(11)
and
(12)
are
the
ramp-up
and
ramp-down
limitation,
ramp-down,
ramp-down,
Constraint
Constraint
(11)
(11)
and
and
(12)
(12)
are
are
the
the
ramp-up
ramp-up
and
and
ramp-down
ramp-down
limitation,
limitation,
The
utility
optimization
model
(Model
B) is formulate
irespectively.
i are
irespectively.
i respectively.
i
i
i
i
i
i

i
I

Ramp-Up/Down
Constraints:
The
generating
units
ramp-up
(
)
and
ramp-down
(
)
1) Objective
The
objective
is the
of
costs. The
formulation
are presented
as follows.
i I2Function:
the initial stage and the processing stage. Constraint
The CO
emission
is converted
tofunction
monetary
costminimization
by
spectively.
constraints
are are
expressed
as follows:
respectively.
respectively.
The
The
constraints
constraints
are
expressed
expressed
asthe
as
follows:
follows:
commitment
findobjective
the compromise
that mi
formulated
rate
at timeThe
t.(9)
For
each
we
divided
equation
into
two below:
stages
which
are technique
the
and
(10)ramping,
are the initial
stages for
ramp-up
and
rampusing
the carbon
taxinitial
rate. toThe
consists solution
of
~and
~objective
emission
cost.
The
COB)2 emission
isisconverted
to moneta
2B)
2.2.2.
Utility
Optimization
Model
1) (Model
Objective
Function:
The
function
thet minimization
of
stage and
thedown,
processing
stage. Constraint
(9)
and
(10)
are
the
initial
for
ramp-up
CO
2.2.2.
Utility
Model
(Model
respectively.
Constraint
(11)
and
(12)
are
the
fuel
costs,
start-up
cost,
(Optimization
)stages
and
) and
ramp-down
ramp-down
)t ) operating
Ramp-Up/Down
Ramp-Up/Down
Constraints:
The
The
generating
generating
units
units
ramp-up
ramp-up
(the
2
i
t ( (
t cost, and CO
1 1 10 0Constraints:
0

Min
c
FU
P

SC
s

E
P
CT

rate.
The
objective
consists
of
the
fuel
costs,
start-up
cost,
op

The
utility
optimization
model
(Model
B)
is
formulated
using
the
multi-objective
unit
f
i
i
i
i
i
i
i
formulated
below:
(9)
p

I
(9)
(9)
p
p

p
p

I
I
ramp-down,
respectively.
Constraint
(11)
and
(12)
are
the
ramp-up
and
ramp-down
limitation,
The
utility
optimization
model
B) presented
is formulated
the m
which
(Modelare
limitation,
respectively.
emission
costs.
The
formulation
as using
iFor
i and
i each
ieach
i ramping,
i iramp-down
i i we
ate
rateatattime
timet. t.ramp-up
For
ramping,
wedivided
divided
thetheequation
equationinto
intoThe
two
two
are
thethe
initial
initial
I stages
tstages
T
uUwhich
f The
F are
costs.
minimizes
technique
formulation
are
presented
as follows.
commitment
technique commitment
to findithe
compromise
solution
that
both
operating
cost and
respectively.
The
constraints
are
expressed
as
follows:

~
to
find
the
compromise
solution
that
minimizes
both
~
~

constraints
are
expressed
as
follows:
follows.
0
1
0
0
1
1
i ramp-up
t and
t
tage
stageand
andthethe
processing
Constraint(9)(9)and
and
the
initial
initialstages
stages
forfor
and
cost.
isSC
(10)
p processing
ipp
iConstraint
(10)
i(10)

Iiare
Min
cramp-up
P(10)
s it using

FUto
(10)
to
Pi tcarbon
CTcost
pi pstage.
iistage.

iare
Ithe
I 2emission
i Emonetary

The
CO
emission
is
converted
cost

f CO
i monetary
i
i by
i the
i tax
i p
ii
i CO2 emission cost.
CO
The
emission
converted
by
usi

2
2

emission
for
each
generation:
E
FC
E
CO
T ramp-down
1)
Objective
Function: The objective
function
is
i=
f
amp-down,
ramp-down,respectively.
respectively.Constraint
Constraint(11)
(11)and
and(12)
(12)arearethetheramp-up
ramp-up
and
limitation,
limitation,
f
i Iand
t
uU f F 1) Objective
Function:
The
objective
function
2ramp-down
is the m
rate. The objective consists
of
the
fuel
costs,
start-up
cost,
operating
cost,
and
CO
t 11 t 1t 10t
t t
2 emissioncost,
rate.
The
objective
consists
of
the
fuel
costs,
start-up
cost,
operating
theformulated
minimization
of(11)
the(9)
total
below:
p ipconstraints
constraints
p pi pare
expressed

i
i
espectively.
respectively.p
The
asasfollows:
follows:
(11)
(11)cost as formulated
expressed
iIi II presented
I (9)
iThe
ip
i i
iii are
i
i
below:
i intensity from genera
The formulation
CO
emission
cost
is generation:
the
CO2 emission
costs. The formulation are
as2 follows.
costs.
The
are
presented
as
follows.

emission
for
each
E
FC
Ef
CO
2
i
=
~~1 ~ ~
0
1 1t
t
(12)
frate fromt generating t
(10)
(10)

I
i and carbon
t
t
(12)
(12)
p itp1

I
P
multiplied
by
the
total
production
(
)
tax
pi 1itppitit p0p1i 0itp

I
I
i
i
SC (18)
Min function
c minimization
FU emission
from
Easi gener
Pi
(9)
cost

f
i Pi of
i s i cost
p ip i p ip i i iii i i

i
i Function:
I I
1) Objective
The
objective
isis(9)
the
the
total
The
CO
the
CO
intensity
2 emission
2 function
1)
Objective
Function:
The
objective
is
the
minimization
o
i
i I tT
u

U
f

F
t 1
t

intensity (Ei) is calculated

by fuel consumption ( FC f ) multipl
t
(11) below:
2 emission
(11)
p0 i 0 1 p1 i ~
~ i formulated below: i CO
Iformulated
P
multiplied
by
the
total
production
(
)
and
carbon
tax
rate
from
generating
i or
(10)
(10)
p ip i
pUp/Down
iTime
~i Time

i the

i
Ithe
Itheunit
Minimum
Up/Down
Constraints:
When
unit
iunit
is scheduled
to start-up
or or
Minimum
Minimum
Time
Constraints:
Constraints:
When
is
i isscheduled
scheduled
to to).
start-up
start-up
ipUp/Down
i
ofgeneration:
When
default
value
emission factor
isiobtaine
Factor
for iteach
fuel
Ef
each
Ei = FC
type (ECO
t
t 1
ft The
2 emission
for
i
t
i
t
t t by
t
t ( FC i )f multip
(12)
(12)
p

Min
c
FU
P

SC
s

E
P
CT

t
t
t
t

1
emission
intensity
(E
)
is
calculated
fuel
consumption
CO

(18)

2
i

i
i
i
)
or
minimum
down
time
(
)
before
the
utdown,
it
has
a
duration
of
minimum
up
time
(
Min
c
FU
P

SC
s

E
P
CT

(i
)

or
) Ior
minimum
down
time
()i before
) before
the
shutdown,
shutdown,
itp has
it
has
a duration
apduration
of
of
minimum
minimum
upup
time
time
(
fminimum
i Guideline
i down
i ( fNational
i Greenhouse
i
i i Gas
f i B co
time
IPCC
for
Inventories.
Model

i (11)
i the
i
i
i
i
i2006
(11)
p

i
i
I

i i
i i
i I tT
uU f F
i i
i I tT
uU f The
F
isofthe
CO
CO2 emission
cost
2 emission
fuel
function
with
constraints
(2)-(17).
The
default
value
emission
factor isintens
obtain
Factor
for
each
type
(E
itunit
isunit
startup
ort shutdown
because
the
generating
unit
cannot
be
immediately
turned
or
is isstartup
startup
or
or
shutdown
shutdown
because
because
the
the
generating
generating
unit
unit
cannot
cannot
be
be
immediately
immediately
turned
turned
on
on
or
or
f).on
~
~
t Minimum
t t1 1 Up/Down Time Constraints: When the
t

itotal
(12)
(12)
p
p

p
p

I
I
i
P
Minimum
Up/Down
Time
Constraints:
When
the
unit
i
is
scheduled
to
start-up
or
multiplied
by
the
production
(
)
and
carbon
tax
rate
2006
IPCC
Guideline
for
National
Greenhouse
Gas
Inventories.
Model
rned
off.off.
The
constraints
up/down
time
expressed
in (13)
and
(14)
asE
follows:
turned
turned
off.
The
constraints
of
minimum
up/down
up/down
time
are
are
expressed
expressed
ingeneration:
in
(13)
(13)
and
and
(14)
(14)
asFC
as
follows:
follows:
iThe
i constraints
i of minimum
i of
iminimum
emission
for aeach
Ef
COtime
(19) B frc
2are
CO
Ei = iFC f Ef
f generation:
2 emissioni =for each
unit
i tisi scheduled
to start-up or shutdown,
it has
t
t
with
constraints
i ) or3.function
minimum
down
( (2)-(17).
) before
the (Ei) is calculated by fuel
shutdown, it duration
has at duration
of minimum
up(time
(minimum
Results
and
discussion
emission
intensity
COtime
t t minimum
t of
t t
2CO
up
time
down
i), or
The
CO
emission
cost
is CO
the
intensity
from
generating (19)
unitconsumption
i (Ei) gene
s

I
,
t
{
T
>

}
2| tT
2 emission
s
s

I
I
t
,

t
{

{
T
|
t
|
>
t

>
}

}
The
emission
cost
isor
the CO
2
2 emission intensity from
i
i
i
(13)
i
i
i
i
i
i
(13)
(13)
Minimum
Up/Down
Up/Down
Time
Time
Constraints:
Constraints:
When
When
the
the
unit
unit
i
i
is
is
scheduled
scheduled
to
to
start-up
start-up
or
unit is Minimum
startup
or
shutdown
because
the
generating
unit
cannot
be
immediately
turned
on
or
t
time
(
%i) before the unit is startup or shutdown because
).
The
default
value
of
Factor
for
each
fuel
type
(E
'
t
'
'
f generating unit i (CT ). emission
=t ti=t1t=
Pidiscussion
multiplied
by time
the total3.
production
((13)
)and
and
carbon
tax rate
from
The the
it
1i 1
In in
this
section,
analysis
results
multiplied
the
total
(P
) and
carbon
taxobtained
rate fromi from
generating
Results
and
turned
off.ittThe
constraints
of
minimum
up/down
(14)
as
follows:
i of
or
) orexpressed
minimum
minimum
down
down
time
time
( (production
)2the
before
)emission
before
thethe
hutdown,
shutdown,
ithas
has
agenerating
aduration
duration
of
of
minimum
minimum
up
uptime
time
( (i)iare
the
unit
cannot
be
immediately
turned
on or
by
The
CO
cost
isthethe
CO
2006
IPCC
Guideline
for
National
Greenhouse
Gas
Inventori
2 emission
i optimization model i(Model
t t t t
Minimization
Modelby(Model
A) and the (utility
emission
intensity
(E
)
is
calculated
fuel
consumption
FC
)
multiplied
by
Emission
CO
2
i
turned
off.
The
constraints
of
minimum
up/down
time
intensity
from
generating
unit
i
(E
)
multiplied
by
the
t shutdown
emission
intensitywith
(E
ison
calculated
by iffuel consumption ( FC f ) multi
function
t t
nit
unitis isstartup
startuporor~
the
generating
unit
cannot
beimmediately
immediately
turned
turned
on
ororthe(2)-(17).
i)constraints
t~t ~t t t because
sshutdown

i
I
,

{It T
tT|{unit
>t|Tt~
}tCO
were
formulated
find
that from
satisfies
s1so1iin
1obecause
oi andthe
ias
It,
,
{|
i~
>}i }2~imodels
}be
In this
section,
the
analysis
ofcompromise
the resultssolution
obtained
theb
t to
i
generating
Iti,follows:
{tT
>These
cannot
i s
i|>
(14)
i(13)
(14)
(14)factor
i i ioi of
(13)
are
expressed
(14)
total
production
(P
)
and
carbon
tax
rate
from
generating
urned
turnedoff.
off.The
The
constraints
constraints
of
minimum
minimum
up/down
up/down
time
time
are
are
expressed
expressed
in
in
(13)
(13)
and
and
(14)
(14)
as
as
follows:
follows:
i
).
The
default
value
of
emission
is
obtained
from
Revised
Factor
for
each
fuel
type
(E
f forlevels.
't =t
~
).
The
default
value
of
emission
factor
is
obtai
Factor
each
fuel
type
(E
'
'
~
~
'
f
operational
Minimization
Model
(Model
A)
and
the
utility
optimization
model
(Model
1tt=1t1 1
t =tti=
ti i i
CO
intensity
(E
) iscontains
calculated
2006 IPCC Guideline These
for
National
Greenhouse
Gas
Inventories.
Model
objective
i). The
2 emission
iB
3.i (CT
Results
and
discussion
2006
IPCCunit
Guideline
for
National
Gas
Inventories.
Model B
t t
t t
models
were
formulated
to
findGreenhouse
i the compromise solution that satisfies b
t s s
o
o

I
,
I
t
,

{
T
{
T
|
t
|
>
t
>

}
}
(13)
by
fuel
consumption
(FC
)
multiplied
by
Emission
function with constraints
(2)-(17).
i
i
i~ti
i i t
(13)
(13)
f
function
with
constraints
(2)-(17).
Solution
of Operating
Cost, CO2 emission Cost and CO2 emiss
operational
levels.
FuelFuel
Usage
are
two
for
thefor
fuel
usage
limitation.
First,
1Constraints:
oi There
iare
are
Itwo
,constraints
ttwo
{constraints
Tconstraints
| t >3.1.
~i }Compromise
Fuel
Constraints:
There
There
for
the
the
fuel
fuel
usage
usage
limitation.
limitation.
First,
First,
tiUsage
1i Constraints:
1si
t ' =t '=Usage
Factor
for
each
fuel
type
(Efthe
). The
default of
value
(14)
In
this
section,
analysis
the ofresults obtain
t
t
t
t ' =ttof
~ti fuel
1of of
e the
total
amount
usages
f off 3.
the
unit
i unit
in
plant
p
at
time
t
(
q
)
must
not
exceed
the
the
total
total
amount
amount
fuel
fuel
usages
usages
of
f Results
of
the
the
unit
i
in
i
in
plant
plant
p
at
p
at
time
time
t
(
t
(
q
q
)
must
)
must
not
not
exceed
exceed
the
the
emission
factor
is
obtained
from
Revised
2006
IPCC
Minimization
Model
(Model
A)
and
the
utility
optimization
fip
fip
In this
weof present
theCost,
resultCOof2 emission
Model ACost
andand
B CO
using
th
and discussion
3.1.
Compromise
Solution
Operating
2 emis
3.~ Results
andfipsection,
discussion
t~t
t t
~
~
(14)
Guideline
for
National
Greenhouse
Gas
Inventories.
max
s
s

o
o

I
,
I
t
,

{
T
{
T
|
t
|
>
t
>

}
}
max
max
These
models
were
formulated
to
find
the
compromise
solution
programming.
The
comparison
is
divided
into
three
terms
which
are
the
total
op
(14)
(14)
i
i
i
i
i
i
aximum
amount
of fuel
f fuel
available
for for
plant
pplant
( qpfp(pq().fpqSecond,
the the
generating
unitunit
i unit
cannot
use useuse
maximum
maximum
amount
amount
of of
fuel
f available
f available
for
plant
). Second,
the
generating
generating
i cannot
i cannot
fp). Second,
Model
B
contains
objective
function
with
constraints
Fuelt'=
Usage
constraints
for
the
fuel
usage
limitation.
First,
operational
levels.
total
CO
emission
cost
and
the
total
CO
emission.
For
the
total
operating
cost
t '=~ti~1i 1 Constraints: There areIntwo
this
section,
we
present
the
result
of
Model
A
and
B
using
th
this
section,
the
analysis
of
the
results
obtained
from
the
Operating
Cost
2In
2
In this
section, the analysis of the results obtained from the
t
t t
t as follows:
ore
than
one
type
of
fuel
fusages
at
(Minimization
FU
The
constraints
are
expressed
more
more
than
one
one
type
of of
fuel
fuel
ftime
atf at
time
tthe
( tFU
FU
).
The
).
The
constraints
constraints
are
are
expressed
expressed
as
as
follows:
follows:
(2)-(17).
Utility
Optimization
Model
(Model
B)
has
only
0.28
percent
higher
cost
tha
programming.
The
comparison
is
divided
into
three
terms
which
are
the
total
i ().
Model
(Model
A)
and
the
utility
optimization
model
(Model
B)
is
presented.
the
totalthan
amount
oftype
fuel
ft time
of
unit
i
in
plant
p
at
time
t
(
q
)
must
not
exceed
the
i i
Minimizationfip Model (Model A) and the utility optimization model (Modeo
emission
3.1.
Compromise
Solution
Operating
CO2 operating

Fuel
Usage
Constraints:
There
are
two
constraints
Cost
Minimization
(Model
A)
which
iscompromise
considered
acceptable.
theCos
to
total
CO
emission
cost
and
the First,
total
COthe
emission.
ForCost,
the
total
cos
These
models
were
find
theModel
compromise
solution
that
satisfies
both
strategic
and
2to
2of
These
models
were
formulated
to
find
solution
thatFor
satisfies
Fuel
Fuel
Usage
UsageConstraints:
Constraints:
There
There
are
are
two
two
constraints
constraints
forfor
the
the
fuel
fuel
usage
usage
limitation.
limitation.
First,
max formulated
maximum
amount
fuelusage
f available
for
plant
p
(
q
).
Second,
the
generating
unit
i
cannot
use
fp
cost,
Model
B
is
1.23
percent
lower
than
Model
A.
Moreover,
the
total
CO
for theoffuel
limitation.
First,
the
total
amount
of
3.
Results
and
discussion
Utility Optimization
Model (Model B) has only 0.28 percent higher cost th
t
levels.
hethetotal
totalamount
amountofoffuel
fuelusages
usagesf of
f operational
ofthe
theunit
unit
i in
i inplant
plantp pattoperational
attime
timet (t q( tfip
qlevels.
) must
) mustnot
notexceed
exceed
thethe we present the result of Model A a
fip
In
this
section,
t
Model
B
is
lower
than
Model
AA)approximately
4.82 million
kilograms
fuel
usages
f
of
the
unit
i
in
plant
p
at
time
t
(q
)
must
Cost
Minimization
Model
(Model
which is considered
acceptable.
For theofto
fip
more than one type of fuel f at time t ( FU i ). The max
constraints
are
expressed
as
follows:
max
programming.
The
comparison
is
divided
into
three
terms
which
(kgCO
)
or
4,816
tons
of
carbon
dioxide
(tCO
)).
Once,
the
multi-objective
u
cost,
Model
B
is
1.23
percent
lower
than
Model
A.
Moreover,
the
total
CO
maximum
maximumamount
amount
ofoffuel
fuelfthe
available
f available
forforplant
plantp p( qof
( fpqfuel
).
).
Second,
Second,
the
the
generating
generating
unit
unit
i
cannot
i
cannot
use
use
not exceed
maximum
amount
f
available

In
this
section,
the
analysis
of
the
results
obtained
emission
Cost
and
CO
emission
3.1.
Compromise
Solution
of
Operating
Cost,
CO
2
2
2
2
Cost and CO
3.1. Compromise Solution of Operating Cost,
fp
2
2 emi
total
CO
emission
cost
and
the
total
CO
emission.
For
the
tota
2 Model
2 operating
approach
is
applied,
Model
B
is
not
only
minimize
total
cost,
but
al
Model
B
is
lower
than
A
approximately
4.82
million
kilograms
o
t t
more
morethan
thanone
onetype
typeofoffuel
fuelf at
f attime
timet (t FU
( FU
).The
The
constraints
constraints
are
areexpressed
asas
follows:
follows:
Optimization
Model
(Model
B)
hasthe
only
0.28
i ).
i In
this
section,
we
the
result
of
Model
Athe
and(tCO
B 2using
the
mixed
integer
(kgCO
)expressed
or 4,816
tons
of we
carbon
dioxide
)).ofOnce,
2present
In
thisUtility
section,
present
result
Model
Amulti-objective
and
B percent
using
Cost
Minimization
Model
(Model
A)
which
is
considered
programming. The comparison
is
divided
into
three
terms
which
are
the
total
operating
cost,
theaccep
approach
is
applied,
Model
B
is
not
only
minimize
total
operating
cost,
but a
programming. The comparison is divided into three terms which are the
total
12
cost,
Model
B
is
1.23
percent
lower
than
Model
A.
Moreove
total CO2 emission cost total
and the
total
CO
emission.
For
the
total
operating
cost
comparison,
the
2 cost and the total CO2 emission. For the total operating co
CO2 emission
Model
B is 0.28
lower
than
A
approximately
4.82 millio
Utility Optimization Model
B)
has only
percent
higher
cost
than
the Operating
Utility(Model
Optimization
Model
(Model
B)Model
has only
0.28
percent
higher
cost t

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

Table 2. Total cost comparison for each model

Model
Operating Cost Minimization Model
Utility Optimization Model

CO2 Emission Cost (THB)

136,723,966.31
135,043,207.04

from the Operating Cost Minimization Model (Model

A) and the utility optimization model (Model B) is
presented. These models were formulated to find the
compromise solution that satisfies both strategic and
operational levels.

Total Operating Cost (THB)

633,856,023.21
635,609,670.08

gas instead. The natural gas is typically expensive;

therefore, it leads to higher total cost in Model B than
Model A. Once the CO2 emission cost is involved
in the objective function, the proportion of fuel type
consumption is changed. For Model B, the fuel usage
of lignite in thermal generator is decreased by 20.15
percent and the usage of natural gas is increased by
21.05 percent.

3.1. Compromise Solution of Operating Cost, CO 2

emission Cost and CO2 emission

3.2. Sensitivity Analysis

In this section, we present the result of Model

A and B using the mixed integer programming. The
comparison is divided into three terms which are the
total operating cost, the total CO2 emission cost and
the total CO2 emission. For the total operating cost
comparison, the Utility Optimization Model (Model B)
has only 0.28 percent higher cost than the Operating
Cost Minimization Model (Model A) which is
considered acceptable. For the total CO2 emission
cost, Model B is 1.23 percent lower than Model A.
Moreover, the total CO2 emission from Model B is
lower than Model A approximately 4.82 million
kilograms of carbon dioxide (kgCO2) or 4,816 tons of
carbon dioxide (tCO2). Once, the multi-objective unit
commitment approach is applied, Model B is not only
minimize total operating cost, but also the total CO2
cost. Therefore, the main generators are in combined
cycle power plants which lead to a lower amount of
coal-fired consumption. Therefore, the CO2 emission
is significantly dropped.

The results of CO2 emission comparison are shown
in Table 3. The unit commitment optimization approach
decreases the overall emission by 2.65 percent from
Model A. The total proportion of emission of Model B
from the thermal plants is decreased by 17.44 percent.
The emission from the gas turbine and combined cycle
plants are increased by 18.81 and 8.85, respectively. The
system reduces the usage of lignite and shift to natural

In this section, the sensitivity analysis of Model

B shows the comparison of the power generation for
each generator by varying the carbon tax rates from
0.10 - 3.00 baht/kgCO2 (Fig. 2). The high carbon tax
rate directly impacts the behavior of power generation,
especially for the thermal and combined cycle plants.

There is no significant change when the carbon
tax rate is between 0.10 and 0.60 baht/kgCO2 which
indicates that the prices are too small to trigger the
change of system behavior compared with the operating
cost. However, change occurs slowly when the prices
are between 0.70 and 1.80 baht/kgCO2. The proportion
of combined cycle plant is slowly increased while the
thermal plant is slowly decreased. The obvious change
is occurred when the price is greater than 1.90 baht/
kgCO2.

The stacked chart of percent changed of power
generation for thermal, gas turbine, and combined
cycle plant is shown in Fig. 3. At the rate of 0.70 baht/
kgCO2, the generation from thermal starts to decrease
by 10 percent while the combined cycle is increased
by 7.14 percent. This shows the carbon tax has higher
effect than the operating cost in the thermal plant. On
average, from the rate of 0.20 - 1.90 baht/kgCO2, the
generation from thermal is decreased by 14.74 percent
while the combined cycle is increased by 10.02 percent.

Table 3. Emission comparison for each power plant type

Plant Type
Thermal
Gas Turbine
Combined Cycle
Total

Emission (kgCO2)
Model A
201,056,448.00
29,027,026.93
139,715,145.47
369,798,620.39

Model B
171,202,150.53
35,750,944.71
153,286,247.06
360,239,342.31

13

%
Change
- 17.44
18.81
8.85
- 2.65

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

Figure 2.
Sensitivity
analysisanalysis
of the power
for eachfor
plant
type
to the
Figure
2. Sensitivity
of thegeneration
power generation
each
plant
typecarbon
to the tax
carbon tax

The significant change obvious occurs when the called the marginal operating cost and the marginal
price is greater than 2.00 baht/kgCO2. On average, CO2 emission, respectively. The obvious trade-off
from the rate of 2.00 - 3.00 baht/kgCO2, the generation between the operating cost and the CO2 emission
from thermal is decreased by 57.58 percent while the occurs at the price of 0.70 baht/kgCO2. The CO2
combined cycle is increased by 33.93 percent. This emission is decreased by 5.13 percent while the
operating cost is increased only 1.66 percent. For the
leads to higher CO2 emission reduction.
Fig. 4 shows the percentage of change of the based case, the imposition of carbon tax at 0.349 baht/
operating cost and the CO2 emission when the carbon kgCO2 reduces approximately 1.96 percent while the
total
operating
increases
by only 0.04 percent.
tax rateFigure
is increased
by 0.10
baht/kgCO
are for
2. There
2. Sensitivity
analysis
of the power
generation
each
plant typecost
to the
carbon tax

Figure 3. The percent change of power generation for each plant type from the carbon tax rate of 0.10
baht/kgCO2

Fig. 4 shows the percentage of change of the operating cost and the CO2 emission when
the carbon tax rate is increased by 0.10 baht/kgCO2. There are called the marginal operating cost
and the marginal CO2 emission, respectively. The obvious trade-off between the operating cost
and the CO2 emission occurs at the price of 0.70 baht/kgCO2. The CO2 emission is decreased by
5.13 percent while the operating cost is increased only 1.66 percent. For the based case, the
imposition of carbon tax at 0.349 baht/kgCO2 reduces approximately 1.96 percent while the
Figure 2. Sensitivity analysis of the power generation for each plant type to the carbon tax
total operating
cost increases by only 0.04 percent.
Figure 3.
The percent
change of
powerofgeneration
for eachfor
plant
type
from
thefrom
carbon
rate tax
of 0.10
2
Figure
3. The percent
change
power generation
each
plant
type
the tax
carbon
rate baht/kgCO
of 0.10
baht/kgCO2

Fig. 4 shows the percentage of change of the operating cost and the CO2 emission when
the carbon tax rate is increased by 0.10 baht/kgCO
14 2. There are called the marginal operating cost
and the marginal CO2 emission, respectively. The obvious trade-off between the operating cost
and the CO2 emission occurs at the price of 0.70 baht/kgCO2. The CO2 emission is decreased by

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

CO2

Figure 2. Sensitivity analysis of the power generation for each plant type to the carbon tax
Figure 4. The increase in percent of total operating cost and reduction of the emission
Figure 4. The increase in percent of total operating cost and reduction of the emission

3.3. Pareto-Optimal
Analysis
of Utility
Optimization
Model
which
is 3 percent. From 2.40 to 2.50 baht/kgCO2 and
3.3. Pareto-Optimal
Analysis
of Utility
Optimization
the marginal reduction of CO2 emission is equal to the
Model
In this section we present the Pareto-Optimal
analysis
by of
using
the efficient
frontier
marginal
increment
operating
cost which
is 5 percent.
of the
utility
optimizationis model.
It
betweenwe
thepresent
total operating
cost and the COThe
2 emission
suggestion
from
this relationship
the approapriate
In graph
this section
the Pareto-Optimal
is
illustrated
as
a
guideline
for
a
strategic
planner
to
determine
the
compromise
solution
for
the
analysis by using the efficient frontier graph between carbon tax range that can decreased large proportion
generation
based
considerations.
Besides,iswe
suggest0.60
the to
methodology
between
0.70 baht/kgCO2.
of CO2 emission
the totalpower
operating
cost and
theon
COenvironmental
2 emission of the
to
set
an
appropriate
carbon
tax
that
maximizes
the
reduction
of
the
CO
emission
while
the tax at
2
carbon
utility optimization model. It is illustrated as a guideline In the utility optimization model, the
operating
cost
increases
in
the
lowest
proportion.
for a strategic planner to determine the compromise any rate can be set to test the effect of the operating
The trade-off graph of Model B is obtained by plotting the point representing the total
solution for the power generation based on environmental cost and the CO2 emission instantly. Therefore, the
operating cost and total CO emission for each carbon tax rate as shown in Fig. 5. This graph
considerations. Besides, we suggest2 the methodology computational time of the utility optimization model
shows the different behavior of the increment of the total operating cost over the decrement of
to set an appropriate carbon tax that maximizes the is fast and continent.
the total CO2 emission as the carbon tax rate increased. The marginal CO2 emission is the
reduction
of thein
COthe
while thewhen
operating
cost tax has an increment by unit. The marginal
2 emission
change
CO2 emission
the carbon
4. Conclusions
increases
in
the
lowest
proportion.
operating cost is the change in the operating cost when
the carbon tax has an increment by unit.

The trade-off
graph
of
Model
B
is
obtained
by
When the carbon tax is changed from 0.60 to 0.70 baht/kgCO2, the marginal CO2
This paper
modeling
framework to
plottingemission
the point
representing
the total
operating
is decreased
by 6.30
percent
while the operating
cost ispresented
increaseda only
1.72 percent.
determine
the
possible
carbon
tax
rates
cost andThe
total
CO
emission
for
each
carbon
tax
rate
2 change of the operating cost and the CO2 emission is 1:4. From 1.90 to that
rate of
2.00can be
imposed
on
the
electricity
operating
cost
for
as shown
in
Fig.
5.
This
graph
shows
the
different
baht/kgCO2, the marginal reduction of CO2 emission is equal to the marginal increment the
of policy
makers.
It was significant
to know reduction
the best achievable
behavior
of the increment
total operating
costto 2.50
operating
cost whichofisthe
3 percent.
From 2.40
baht/kgCO2
and the marginal
of
level
of
the
total
operating
cost
and
the
CO
emission
over theCO
decrement
of
the
total
CO
emission
as
the
emission
is
equal
to
the
marginal
increment
of
operating
cost
which
is
5
percent.
The
2
2
2
based
on
the
capacity
of
existing
power
generators.
carbon suggestion
tax rateFigure
increased.
The
marginal
CO
emission
from
this
relationship
is
the
approapriate
carbon
tax
range
that
can
decreased
large
2
3. The percent change of power
generation for each plant type from the carbon tax rate of 0.10
Optimization
proposed in this
proportion
emission
is the
between
0.70Utility
baht/kgCO
is the change
in
theof
COCO2
when
carbon0.60
tax toThe
utilitywas
optimization
baht/kgCO
2. In themodel
22emission
model, theby
carbon
tax at
any rateoperating
can be set
the effect
of theaoperating
cost and the
CO2 for a
paper
to generate
good compromise
solution
has an increment
unit. The
marginal
costto test
Fig. 4Therefore,
shows
thethe
percentage
operating
cost and theproblem
CO
emission
instantly.
timeofofthethe
utility
optimization
model
is
fast when
2 emission
multi-objective
unit commitment
in large-scale
is the change
in the
operating
cost when
thecomputational
carbon of
taxchange
the carbon
called the The
marginal
operating
and continent.
2. There areplanning.
power generation
operating
cost cost
was
has an increment
by unit.tax rate is increased by 0.10 baht/kgCO
emission,
respectively.
The
obvious
trade-off
between
the
operating
cost
and
the
marginal
CO
2

When the carbon tax is changed from 0.60 to 0.70 increase in small proportion compared with large
andmarginal
the CO2 emission
occurs
the price ofamount
0.70 baht/kgCO
. The CO2itemission
decreased
by
of CO2 2emission
reduced.isThe
trade-off
baht/kgCO2, the
CO2 emission
is at
decreased
5.13
percent
while
the
operating
cost
is
increased
only
1.66
percent.
For
the
based
case,
the
by 6.30 percent while the operating cost is increased relationship between the operating cost and CO 2
imposition
ofofcarbon
baht/kgCO
approximately
percent
while the
2 reduces
emission
and the
marginal cost1.96
of CO
only 1.72 percent.
The rate
changetax
of at
the0.349
operating
2 reduction were
totalemission
operatingiscost
by only
0.04 percent.
illustrated to use as a guideline for a planner to make
cost and the CO
1:4.increases
From 1.90
to 2.00
2
baht/kgCO2, the marginal reduction of CO2 emission an optimal decision for the unit commitment dispatch
is equal to the marginal increment of operating cost with the acceptable emission allowance.

15

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

emission
and
the
cost
the
Utility
Optimization
Model.
5.
Efficient
frontier
of
CO
Figure
5.analysis
Efficient
of2the
CO
emission
and
the type
operating
cost
using
theto
Utility
Optimization
FigureFigure
2. Sensitivity
Figure
2. Sensitivity
of frontier
thethe
power
analysis
generation
of2 the
for
power
eachoperating
generation
plant
for
to using
the
each
carbon
plant
tax
type
the carbon
tax
Model.

The
carbon tax rate that can be imposed Catalo JPS, Mariano SJPS, Mendes VMF, Ferreira LAFM.
4. possible
Conclusions
Short-term scheduling of thermal units: emission
to reduce CO2 emissions in long term is between 0.60

constraints
andthe
trade-off
curves.
European
Transactions
. In order
to seta the
carbonframework
tax
to 0.70 baht/kgCO
This 2paper
presented
modeling
to
determine
possible
carbon
tax rates

on
Electrical
Power
2008;
18(1):
1-14.
on electricity
it is needed
involvement
of cost for the policy makers. It was significant to
that cancost,
be imposed
on thethe
electricity
operating
Cosmo VD, Hyland M. Carbon tax scenarios and their effects
government
andachievable
related parties.
knowoffice
the best
levelThe
of appropriate
the total operating cost and the CO2 emission based on the

on the Irish energy sector. Energy Policy 2013; 59: 404existing
generators.
The Utility Optimization model was proposed in this
carboncapacity
tax rate of
setting
willpower
lead to
CO2 emissions
14.
paper
generate
good
solution
a multi-objective
commitment
problemofinEnergy.
reduction
in atolong
term.a In
the compromise
beginning phase,
the for Energy
Policy and unit
Planning
Office Ministry
large-scale
power
generation
planning.
The
operating
cost
was
increase
in
small
lower carbon tax rate can be imposed on electricity
Energy Statistics of Thailand 2012.proportion
Report.
compared
with
large
amount
of
CO
emission
it
reduced.
The
trade-off
relationship
between
the Power
2
cost to observe the impact on different perspective
Electricity
Generating
Authority
of Thailand
(EGAT).
operating
cost
and
CO
emission
and
the
marginal
cost
of
CO
reduction
were
illustrated
to
use
2 0.349 baht/kgCO2 as
2
by using the carbon tax rate of

Development
Plan 2010 (PDP2010). EGAT, Bangkok,
as a It
guideline
a planner
make
an optimal
for the2010.
unit commitment dispatch with

Thailand.
a baseline.
might beforapplied
for atofew
years.
Once decision
acceptable emission allowance.
IPCC Guidelines for National Greenhouse Gas Inventories.
the COthe
2 emissions can be decreased by this policy, the
The
possible
carbon
tax
rate
that
can
be
imposed
to reduce
CO2 2.1,
emissions
Chapter
2, Equation
Energy;in
2: long
2.11. term
2006.
government sector can increase the carbon tax rate and
is
between
0.60
to
0.70
baht/kgCO
.
In
order
to
set
the
carbon
tax
on
electricity
cost,
it
is
needed
IPCC
Guidelines
for
National
Greenhouse
Gas
Inventories.
2
reduce more emissions. The options that can
be used
parties.
Chapter
2, Table 2.2carbon
Defaulttaxemission
factors for
the
involvement
of
government
office
and
related
The
appropriate
rate
setting
to limit the CO2 emissions from electricity generation

stationary
combustion
in thethe
energy
industries,
will lead to CO2 emissions reduction in a long term.
In the beginning
phase,
lower
carbonEnergy;
include increasing the usage of renewable energy,
2: 2.16.
2006. on different perspective by
tax rate can be imposed on electricity cost to observe
the impact
using fuels with lower CO2 emission per kWh produced Lu C, Tong Q, Liu, X. The impacts of carbon tax and
using the carbon tax rate of 0.349 baht/kgCO as a baseline. It might be applied for a few years.
and/or increasing the efficiency of the electricity2
complementary
policies
on Chinese
economy.
Once the CO2 emissions can be decreased by this policy,
the government
sector
can increase
the Energy
production.
In
future
research,
the
model
should

Policy
2010;
38(11):
7278-85.
carbon tax rate and reduce more emissions. The options that can be used to limit the CO2
consider
the sharefrom
of renewable
and formulate
Nagurneythe
A, Liu
Z, Woolley
T. Optimal
endogenous
emissions
electricityenergy
generation
include increasing
usage
of renewable
energy,
using carbon

taxes
for
electric
power
supply
chains
with
the equation
that
also
involved
renewable
energy.
fuels with lower CO emission per kWh produced and/or increasing the efficiency ofpower
the plants.

2 generation
Figure 3. The percent
Figure
change
3. The
of power
percent
change offor
power
eachgeneration
plant type
from
each
theplant
carbon
type
tax
from
rate
the
of 0.10
carbon
tax rate2006;
of 0.10
for
Mathematical
and
Computer
Modelling
44: 899electricity
production.
In future research, the model should consider the share of renewable
baht/kgCO
baht/kgCO
2
2

916.
Acknowledgments
energy and formulate the equation that also involved renewable energy.
Social Cost of Carbon Fact Sheet. United States EnvironFig. 4 shows the Fig.
percentage
4 shows
ofthe
change
percentage
of the operating
of change
cost
ofmental
the
andoperating
the
CO2 cost
emission
and the
when
CO2 emission
when
Protection
Agency.
November
2013. 1-4.

TheAcknowledgments
authors gratefully acknowledge the support of
the carbon
tax
rate
the
is
carbon
increased
tax
by
rate
0.10
is
increased
baht/kgCO
by
0.10
baht/kgCO
.
There
are
called
.
the
There
marginal
are
called
operating
the
marginal
cost
operating
cost
2
Grant 51-2115-023-JOB NO.544-SIIT from EGAT
and Sutham2 P. Investigation on Carbon Intensity of Cement
and the
and
CO2the
marginal
CO2 emission,
emission,
respectively.
Therespectively.
obvious
trade-off
obvious
between
trade-off
the
operating
between
costthe
operating
cost
The
Iron
Industry
and Energy
Industry
(Fossil
fuel
the marginal
BCP Research
Grant
fromgratefully
Bangchak
Petroleum
Public
The
authors
acknowledge
the support
ofIndustry,
Grant 51-2115-023-JOB
NO.544-SIIT
from
and the
CO
and
the
CO
emission
occurs
at
emission
the
price
occurs
of
0.70
at
baht/kgCO
the
price
of
.
0.70
The
baht/kgCO
CO
emission
.
The
is
decreased
CO
emission
by
is
decreased
by
2
2
2
2
2
2

power
plants)
in
Thailand.
The
Joint
Graduate
School
of
Company
Limited
EGAT
and (BCP).
the BCP Research Grant from Bangchak Petroleum Public Company Limited (BCP).
5.13 percent while
5.13
thepercent
operating
while
cost
theis operating
increased cost
onlyis1.66
increased
only
For
1.66
theEnvironment.
percent.
based case,
For the
the
basedbycase,
the
percent.
Energy
and
Funded
Thailand
imposition
of carbon
imposition
tax at of
0.349
carbon
baht/kgCO
tax at 0.349
baht/kgCO
approximately
1.96
approximately
percent
while
1.96theOrganization
percent while(TGO).
the
2 reduces
Greenhouse
Gas
Management
2 reduces
References
total operating cost
total
increases
operating
by cost
onlyincreases
0.04 percent.
by only 0.04 percent.
2011.
Baranzini A, Goldemberg J, Speck S. A future for carbon

taxes. Ecological Economics 2000; 32(3): 395-412.

16

N. Intalar et al. / EnvironmentAsia 8(2) (2015) 9-17

Thepkhun P, Limmeechokchai B, Fujimori S, Masui T,

Shrestha RM. Thailands low-carbon scenario 2050:

The AIM/CGE analyses of CO2 mitigation measures.

Energy Policy 2013; 62: 561-72.
Wachirarangsrikul S, Sorapipatana C, Puttanapong N,

Chontanawat J. Impact of carbon tax levy on electricity

tariff in Thailand using computable general equilibrium

model. Journal of Energy Technologies and Policy 2013;

3(11): 220-28.
Wattanakulcharat A, Wongsa K. Policy brief: imposition of

carbon tax policy to reduce carbon dioxide emissions

and impact on Thailands economy. Thai Universities

for Healthy Public Policies (TUHPP). 2011. 1-4.
Yamashita D, Niimura T, Yokoyama R, Marmiroli M.

Pareto-optimal solutions for trade-off analysis of CO2

vs. cost based on DP unit commitment. Proceedings of

2010 International Conference on Power System

Technology, Hangzhou, China. October 24-28, 2010.
1-6.
Received 12 December 2014
Accepted 19 February 2015
Correspondence to
Assistant Professor Dr. Chawalit Jeenanunta
Head of the School of Management Technology,
Sirindhorn International Institute of Technology,
Thammasat University, 12120
Thailand
Tel: +66-2501-3505 ext.6000
E-mail: chawalit@siit.tu.ac.th

17