ISSN: 0379-5462

IPPTA: Quarterly Journal of Indian Pulp and Paper Technical Association

EnergyPAPERS
TECHNICAL Analysis Vol. 33, E3, 2021, p. 48-51

of Multiple Effect Abstract: This article represents a brief energy analysis of Multiple Effect
Evaporator (MEE), used to enhance the solid concentration of the black liquor obtained
Evaporator in from the paper industry. MEE accounts a high share of the energy consumption from
the total industry’s energy. Nowadays reduction of energy utilization in this MEE unit

an Indian Pulp &

is a prime concern which may be acquired by integration of different energy reduction
schemes. The energy integrated MEE model may capable of increasing the industrial
self-sustainable, energy efficient and economic prosperity simultaneously. To achieve
Paper Industry the maximized energy efficiency a newly introduced metaheuristic approach named
Flow Direction Algorithm (FDA) has been applied within the constraint environment.
The obtained result shows a 75.6% enhancement in the steam economy than that of
the real time plant data available from the literatures.
Introduction:
A modern Pulp & Paper industry is self-sufficient in energy that all the steam, heat and electricity
Smitarani Pati
Department of Instrumentation and demands can be satisfied within the plant itself by using the energy of the Weak Black Liquor (WBL).
Control Engineering For which WBL (a biomass based byproduct obtained as waste) needs to be pretreated through
Dr. B. R. Ambedkar National Institute of
Technology Jalandhar, Punjab Multiple Effect Evaporator (MEE) to achieve higher concentration value before being used as fuel. MEE
is one of the most energy intensive sub-unit of Pulp & Paper industry and is capable of consuming
more than 35% of the total industry’s energy [1]. Such huge amount of energy consumption
necessitates incorporation of various Energy Reduction Schemes (ERSs) and operational changes in
order to make it more energy efficient and economical. Some of the previously reported ERSs are Feed
preheaters [2,3], Flash tanks [4,5], Thermo-Vapor Compressor [6,7], Mechanical Vapor Compressor
Om Prakash Verma [8], and their hybrids [9–12] and the operational changes includes steam- and feed- split operations
Department of Instrumentation and [6,7], declines
[13,14]. Integration of these ERSs Mechanicalthe Vapor
rate ofCompressor [8], andthat
energy utilization their hybrids
cuts [9–12]
operating andupthe operatio
cost
Control Engineering
to certain extent and increasesincludes steam-efficiency
the energy and feed-which
split operations [13,14]. Integration
leads to improvement in the of these ERSs declines
industry’s
Dr. B. R. Ambedkar National Institute of
Technology Jalandhar, Punjab socio-economic perspectives.energy utilization that cuts operating cost up to certain extent and increases the energ
which leads to improvement in the industry’s socio-economic perspectives.
The energy efficiency analysisThe of the system
energy is rendered
efficiency analysisthrough the different
of the system mathematical
is rendered through themodels
different mathemat
[10,15,16]. Fig. 1 shows the[10,15,16].
classification of the mathematical models employed
Fig. 1 shows the classification of the mathematical to themodels
different
employed to t
energy integrated MEEs in order to numerically
energy integrated MEEs analyze the system.
in order As theanalyze
to numerically real time systemsAsare
the system. the real time
generally nonlinear in practice;generally
hence, the energyinmodels
nonlinear developed
practice; hence, thefor the MEE
energy are developed
models also nonlinear
for theinMEE are als
nature. The energy efficiency ofinthe MEEThe
nature. is measured in termsofofthe
energy efficiency Steam
MEEConsumption
is measured in(SC) and
terms of Steam
Steam Consumpti
Deepak Sahu Economy (SE) and both exhibitSteam
an inverse
Economyrelationship
(SE) andwith
botheach other.
exhibit The setrelationship
an inverse of equations developed
with each other. The set o
Department of Chemical
Engineering, Dr. B. R. Ambedkar for the MEE is obtained throughdeveloped
1st principle of thermodynamics
for the MEE is obtained and the optimal
through SE value
1 st principle is achieved and the
of thermodynamics
National Institute of Technology
Jalandhar, Punjab
through formulation of a maximization
value is problem
achievedpresent
throughinformulation
the constraintof environment.
a maximization problem present in the
environment.
As per the above discussed literatures, this study
has been accounted with the solution of the
developed nonlinear mathematical models for
an energy integrated MEE configuration. These
mathematical models are disciple into a single
Surendra Pratap Singh objective optimization problem to maximize
PAPRI of JK Organization Jaykaypur, the SE and solved in search of the optimized
rayagada, Odisha
unknown process parameters (Li, Ti, and V0). A
newly developed metaheuristic based optimization
approach named Flow Direction Algorithm (FDA)
has been hired to solve the optimization problem
to obtain the best possible values of process
parameters. Furthermore, the employment of
various ERSs which includes steam, and liquor
feed split operation, Thermo-Vapor Compressor
Nishant Nilekar
Corps of Electronics and Mechanical (TVC) and feed pre-heater have been taken place
Engineers (EME), Indian Army
Ambala, India
in order to enhance the energy efficiency of the
system. The SE may be further enhanced by
integrating the MEE system with flash tanks that
extracts heat from the waste steam as a heat
Fig. 1 Solution
Fig.approaches adoptedadopted
1 Solution approaches for solving the the
for solving MEE models
source for accelerating the Kraft recovery process.
As per the above discussed literatures, this study has been accounted with the solu
MEE models

Materials and methods developed nonlinear mathematical models for an energy integrated MEE configura
mathematical models are disciple into a single objective optimization problem to maxi
Kapil Kumar Goyal The preferred MEE system is and
a seven stage
solved evaporator
in search of thesystem installed
optimized in an process
unknown Indian pulp and paper
parameters (Li, Ti, and V0
Department of Industrial and mill used to concentrate the WBL. Black liquor
developed is the waste
metaheuristic basedproduct obtained
optimization when the
approach KraftFlow
named recovery
Direction Algori
Production Engineering, Dr. B. R.
Ambedkar National Institute of process proceeds towards thehaspaper
been making
hired toby extracting
solve the lignin,problem
the optimization celluloseto and other
obtain the cellulose
best possible values
Technology Jalandhar, Punjab fibers from the pulpwood. It is normally Furthermore,
parameters. concentratedthe about 60 – 85%
employment by using
of various MEE
ERSs systems
which to steam, and
includes
produce energy and recover the cooking chemicals. Hence, the pulping industry has been using it as
one of the energy sources since 1930s by extracting the water and utilizing it as a heat source to the
evaporator system.

48
 order to enhance the energy efficiency of the system. The SE may be further enhanced by
In similar way for the 4 and 5 stages of the MEE the energy equations are deduced as in Eqs.
integrating the MEE system with flash tanks that extracts heat from the waste steam as a heat
source for accelerating the Kraft recovery process.
(7) - (8).
Materials and methods ( Li  Li 1 )(i  hci )  Li ( H i  hi )  Li 1 (hi 1  H i )  0 (7)
Energy Analysis of Multiple Effect Evaporator in an Indian Pulp & Paper Industry TECHNICAL PAPERS
The preferred MEE system is a seven stage evaporator system installed in an Indian pulp and paper
U i Ai (Ti1  Ti )  (Li  Li 1 )(i  hci )  0
mill used to concentrate the WBL. Black liquor is the waste product obtained when the Kraft
recovery process proceeds towards the paper making by extracting the lignin, cellulose and other
(8)
the 4th and th
cellulose fibers from the pulpwood. It is normally concentrated about 60 – 85% by using MEE
The schematic
systems to producediagram
energy and of the MEE
recover system
the cooking is shown
chemicals. Hence, in theFig. 2. industry
pulping This diagram
has where, iwhere,
denotesi the stage the
denotes number
stage of number of 5thestage. Further,
4th and employment
5th stage. of feed split
Further,
shows theit backward feedsources
flow sequence
since 1930s byalong with various ERSsitsuchas a as employment th of feed
th split operation to the 6th and 7th stages with a split
operation to the 6 and 7 stages with a split fraction of (1-k) and k respectively. The energy
been using as one of the energy extracting the water and utilizing
heat source to the evaporator system.
steam split, feed split, and feed preheater. The first two stages of this MEE fraction of (1-k) and k respectively. The energy balance equation get modified
The schematic diagram of the MEE system is shown in Fig. 2. This diagram shows the backward
system
feed floware fed along
sequence with with
thevarious
live steam
ERSs such byassplitting
steam split,itfeed
with
split,aandfraction of y and
feed preheater. balance and
equation get modified
rendered and(9)rendered
as in Eqs. - (12). as in Eqs. (9) - (12).
Thelast
the first two
twostages of thisare
effects MEEfedsystem
withare the
fed with the live
black steamby
liquor by splitting
splitting it with a fraction
it with a feed
of y and the last two effects are fed with the black liquor by splitting it with a feed spit fraction of
spit fraction
k. The of k. The
feed pre-heater feed
is used pre-heater
to heat is used
the black liquor to heatit tothetheblack
before sending evaporatorliquor before
system, (L6  L5 )(5  hc5 )  L6 (H 6  h6 )  L7 (h7  H 6 )  (1 k)L f (h f  H 6 )  (1 k )L f CpT  0 (9) (9)
sending
however thisit topre-heater
the evaporator system,
takes energy from however
the steam thissixth
obtained from pre-heater
effect with atakes energy
steam split
fraction of m and the total steam coming out from the seventh effect. U 6 A6 (T5  T6 )  ( L6  L5 )(5  hc 5 )  (1  k ) L f CpT  0 (10) (10)
m( L7  (1  k )L f  L6 )(6  hc6 )  L7 ( H 7  h7 )  kL f (h f  H 7 )  kL f CpT  0 (11) (11)
U 7 A7 (T6  T7 )  m( L7  (1  k ) L f  L6 )(6  hc 6 )  kL f CpT  0 (12) (12)
These fourteen nonlinear equations of the MEE treated as equality constraints with fourteen
These fourteen nonlinear equations of the MEE treated as equality constraints
number with
of unknown
fourteendecision
numbervariables
of unknownsuch decision
as fresh steam supplied
variables such(Vas 0), fresh
outputsteam
temperature of
each stage (Ti (i=1,2,…,6)), and output liquor flow rates (Li (i=1,2,…,7)). Eqs. (13) - (14)
supplied (V0), output temperature of each stage (Ti (i=1,2,…,6)), andillustrates
output liquor flow rates (Li (i=1,2,…,7)).
the side constraints of Ti and Li for the mentioned stages of the proposed Eqs. (13) - (14) illustrates the MEE
side constraints of Ti and Li for the mentioned stages of the proposed
continuous processMEE with feasiblea continuous
to maintain bound ofprocess [0:3] with
(kg/s) of fresh
feasible bound of steam (V0), require
[0:3] (kg/s)
the
thethe continuous
side constraintsprocess.
of fresh
ofboundary TheLboundary
Ti Tand
steam (V0),
for values of Tstages
required to acceleratei and Lof
the i that
continuous subjected
process. Theto comp
side constraints of i and
iL
values of Tithe
i for and mentioned
the mentioned
Li that subjected tostages
computeof the
thethe proposed results MEE
proposed
optimal MEE to
results
continuous are mentioned
process withare in Table
feasible
mentioned 1.
inbound
Table 1. of of
[0:3] (kg/s) of fresh steam (V to),maintain
required
continuous process
the side with feasible
constraints of Ti and bound
Li for the [0:3]
mentioned (kg/s)
stages ofofthefresh
proposed steamMEE 0(V 0), require
a
thethe continuous process.
continuous
continuoustheprocess. The
process
The boundary
with feasible
boundary values
bound of
,3...7)values of
[0:3] T and
(kg/s)
of of
(T8T)i and L
fresh that
steam
iLthat
subjected
(V ), required to
to compute
accelerate
i , (i subjected
Lsubjected 1,2...7) to compu
i i
Ti  T i 1 , ( iThe 2boundary T 7  52 Li 1  0
(13)
continuous process. values of Ti and Li that to compute the optimal
results
Fig. 2 Schematic of ERSs integrated MEE model
are mentioned
results are mentioned in Table
in Table 1.
,1.(i 1.2,3...7) and L  0, (i  1,2...7)
Fig. 2 Schematic of ERSs integrated MEE model
Model Formulation
The energy performance of the proposed MEE model is analyzed through the nonlinear
T in0Table
results are mentioned (14)
i i
from the steam obtained
developedfrom sixth effect with aofsteam split fraction
The set ofof m (13)
1 ,T(i iBoundary
and2Li ,13
mathematical equations by employing the 1 st principle thermodynamics.
T Tiof
Table Ti T , ...
(i 72)th,3T
...7)T752  52(T(T8proposed
8)
) LL i 1  L i , (iL 1,,(2i...
7) 1,2...7)
and Fig.
the
energy total2steam
equations Schematic
are coming
created with some of
out ERSs
from
appropriate the integrated
seventh
assumptions Table
that effect. MEE
the system 1 Boundary
operated inmodel
steady values
i T
i  Ti1:1 , (i 
i2for i 7stages
,3,values
... )7,3... of
52
Tof77)Tiand
and (Li ) i ith
T8for i1,2
ERSs
1L stages
1...
i L ,integrated
i of  1,2...7MEE
(i proposed )ERSs (14)
state ideal condition as no plant complexities are considered, the concentration of the solution are T  0 (i  2 L  0 , (i  7 )
(  CTable
T iTVC ) 1 [Boundary
90Ti: 125 0, ;(ofi70 2:,120
3...7th); 66 : 90Li ;60 , (:iintegrated
75
 1,;255 7): 65 ; 52 : 63 ], ( i 1
i i

mulationModel values T and

integrated MEE
0ERSs ...MEE
Fig. 2 Schematic of ERSs integrated MEE model T i  0, (i  2,3...7) and Li  0, (i  1,2...7)
constant throughout the process. Due to the integration of TVC the fresh steam enter into the
Formulation i and Li for i stages of proposed

Fig. 2 Schematic of ERSs integrated MEE L model

th
first combined with the steam out from the 7 stage of the MEE. The intermediate steam is split

mulation with split fraction of y and (1-y) and send as heating medium for the 1 st and 2 nd stage respectively
performance of the proposed MEE model
The energy performance of the proposed MEE model is analyzed through Table Table is1 Boundary
analyzed
1 Boundary ( kg / s )Tof( Tof
values
through
values [C1i )and
 [ 90 : 125 ;th70 : 120 ; 66 : 90 ; 60 : 75 ; 55 : 65 ; 52 : 63 ], ( i  1, 2 ... 6 )
T/. 5
the :L5i for
;nonlinear
3 i: 6stages ; 4 : of
th 8; proposed
6 . 5 : 10 ERSs i
.5; 9 integrated
: 12 ; 6 : 13 MEE ; 6 : 14 , ( i  1
si )and
 [1L .5i :for5 ; 3i : 6stages
; 4 : 8; 6of. 5 :proposed
10 .5; 9 : 12ERSs ; 6 : 13integrated
; 6 : 14 , ( i MEE
i
ormulation as illustrated in Fig. 1. Applying enthalpy balance around the 1 st and 2nd stage, the obtained
T i (Tof )CVarious
) [90[ 90
L
:are
(
125
kg
; 70 : 120 ;in 66 : 90: 90 ; 60 :to75 ; 55 :with
65 ; 52
1, 2 ... 7 )
: 63 ],the (],i constr
C (i 1, 21.,
the nonlinear mathematical equations developed st by employing the 1st
i

al performance
equations
equations
developed
principle of
are expressed
ofthermodynamics.
as in Eq. (1) - (4).
the yVproposedby employing
(  hc ) The L ( H MEE set
 h ) of
the
(h model
L energy
1
H )  equations
Various
principleis analyzed
are created
approaches
i ( thermodynamics.
 through the nonlinear
(1) with
approaches: reported
125 ;
are The
70reported: in
set
120 theof ;
the literature
66literature ;
to 60
smoothly:smoothly
75 ;
deal55 : deal
65
the ; 52 with
constraints : 63 functions,
rgy performance of the proposed MEE model
0
from is inwhich
analyzed s/)through
(2)penalty 51 .:5the
1functions nonlinear
;:isin
6the 4; simplest .:5ones .5ofand 9; :9considered thehere
:614that ( ithe
(i 1prop
m m m 1 1 1 2 2 1

alions aresome
equations created developed with
appropriate some
assumptions A (appropriate
byUemploying T  T ) that yV (the
the)assumptions
 hcsystem  0 st operated
1 principle L
that
of
steady(thermodynamics.
kg
the( kg /from
system
state which
[
s)Various
is [
.penalized
penalty
operated 5The:; 53with
functions
3set ;steady
is the
6of
:respect :48simplest
; the
to: 8
6;. 56violation
10:and
ones
;.5the
considered
12: 12;here
6; :that
613 ; 6proposed , objective , 21.,
iL
the10 In: 13
deal;work : 14the ,fitness
1 1 m 1 m m m

y )V (   hc )  L ( H  h )  L ( h  H ) st function approaches areof reported in literature constraints.

to smoothly this with the
tical
ondition equations
ideal condition
as no plant developed as(1 no plantby employing
complexities arethe 1
considered, 0
principle
function the of
concentration
is thermodynamics.
i (3)
penalized with The
respect set to the violation of the constraints. In this w
witharecomplexities T  T )  (1  yare )V ( considered, the concentration is of the solution are
m m m 2 2 2 3 3 2

tions areofcreated theoutput

solution some U A (appropriate
constant throughout
assumptions
the
hc )  0 Various
process. Various
Due to that the
approaches
thetheapproaches system
(4)
integration
function areoperated reported
represented
constraints
are reported in assteady
inin Eq.
functions, thethe
in literature
(15).
from which penalty
literature to to smoothly
functions isdeal
smoothlythe deal with
simplest
with the ones
theconstrain
constra
quations are The created with the 1some appropriate assumptions
each other and fed asthat the system operated in steady
2 2 1 2 m m m

oughout the process. Due to the anintegration TVC firstofof T TVC

function the is
fresh represented
steam enter assolution
in isEq.
into the (15).
TVC
st nd
vapor from and 2 stage are combined with heating and considered here that the
ondition of asmedium
no plant complexities are the considered, F proposed objective function is penalized with (15)
from andthe which concentration
penalty of
functions the the are
simplest ones and considered here that the propos
TVC the  f ( z)  p 
thefresh steam enter withinto combined withequations
the steam
from which penalty functions is the simplest ones and
In thisconsidered here that isthe prop
for 3 stage rd
of the MEE average temperature the Energy m

ledcondition as no plant complexities are considered, the concentration of the solution are
avg
th respect to the violation of the constraints. work the fitness function
with the
out for
oughout the process. (Due steam
this
from stage is
the 7th out
represented
from
stage as
 L )(to
in Eq.
of the
 the
the
(5)- (6).
7
MEE.
hc ) integration
stage
The intermediate
 L ( H  h )  L (h of
of the
TVC
 Hfunction
MEE.
steam isThe
the is fresh split intermediate
penalized with
where F is the fitness
(5)steam enter with into steam
respect function
the
asp inisTVC is split
value,
to(15). theFthenegative sign indicates it is a maximization problem, f(z) is
violation
factor, f (and
violation z) ω ispof of the constraints. In violation.
this work
throughout splitthe process.
fraction of yL and Due (1-y) toand
3
the send
1
integrationas heating
avg
of
)0
function
mediumTVC
avg
the
for the is
fresh
3
1stpenalized
andsteam
the
3 represented
objective 3
enter with
function, into
4
respect Eq.
thetheTVC
4
to
penalty 3 mthe totalthepenaltyconstraints.
for constraints’In this wo
ction of y and (1-y) and U send as heating
L )(  hc )medium for the 1(6)st and 2 nd stage respectively
m
( L th
ed with the steam out from the
A (T  T ) 
7 stage of 0
thefunction MEE.
function The
is intermediate
represented as steam
in Eq. is
(15).split
Fasis represented as in Eq.value, is(15).
2nd stage respectively as illustrated in Fig. 1. Applying enthalpy balance
3 3 2 3 3 1 avg avg

dbined with(7)
in Fig. the
In similar steam
way for the out 4 and from 5 stages the th
of the MEE th the energy equations
7obtained stageequations
th
of the where
areMEE.
deduced is
The the
stin Eqs. fitness
Objective function
intermediate
2ndndcurrent function steam splitnegative sign indicates it is a maximization p
action y 1.
ofaround and Applying
- (8). (1-y)
the 1st and and2nd enthalpy
send stage, asthe balance
heating mediumaround arethe
for the11 st and
expressed and 2
asIninthe stage,
stage the obtained
respectively
investigation, there F areFtwo f(objectives:
z)(z)pand to minimize the SC and to maximize (15)the SE
the objective
for stthe 1 function, pmERSs isrespectively
the penalty factor, m pmω is thetheratiototal penalty for constra
st nd
e fraction
expressed of(1)
Eq. yasand- (4).(1-y) and send as heating medium
in Eq.( L  L
(1) )( 
-

(4).
hc )  L ( H  h )  L ( h  H )  0 (7) and
ndthe 2 stage As fSE isdefined  as
d in Fig. 1. Applying Uenthalpy A (T  T )  ( L balance  L )(  hc ) around the 1 and st (8) 2
of
stage, the obtained
proposed integrated MeE. of difference between the
i i 1 i i i i i i 1 i 1 i

0
ated in Fig. where,1. Applying enthalpyof the 4 andbalance
i i i 1
around
where
i

Objective Fthe is
i
1the and
i 1

function
fitness 2nd of
amount
i
stage,
feed liquor
function
i
thevalue,
andobtained
product liquor with respect to the amount of SC by the MEE. Hence, this
negative sign indicates itwith
is ais maximization prob
e expressed as in Eq. (1) - (4).
th th

kwhere F is the fitness is function value, negative sign indicates itindicates a maximization
it is a The p
i denotes the stage number 5 stage. Further, employment of feed split

s are expressed operation yV as ( 

m inmEq. (1)
to the 6 and  7hc
th
)
m - (4).
stages  withL a (split
1 H
th
1 
fractionh 1 )
of  L
(1-k)
2 (
and h 2  H
respectively.
1 ) The 0 energy (1)
problem where turned F is
in to the fitness
a single (1) functionoptimization
objective value, negative
problem sign several constraints.
balance equation get modified and rendered as in Eqs. (9) - (12). the objective function,
theInH the currentfunction,
objective investigation,
problem
(2)
pmthe
for
maximizationpismMEE the
is the ispenalty
there penalty
formulated
problem,
(1) are(2) factor,
two
as
f(z) factor,
in the
is Eq. andand
objectives:
(16).
objective ω is ω theistothe
function, total
minimize
pmtotalis penalty
thepenalty the for
penalty SC constraint
for
factor, and to m
constra
yV L( )(
U )Ahc ( T )  L T () H  yV h ()  L  (hch  )  0)  0 (9) and ω is the total penalty for constraints’ violation.
(1) :As
(L m  hc1
m L
1 ( H
m m  h )  L 1
1 (h  H1 )  (1 
m 1k ) L (mh  H
2 )  (
21 
m k ) L Cp 1T  0
yVmU(A m(T  Thc ( L) L L )(1 (H (1h1k))LCpLObjective
2 ( h0 of2  the H 1 ) proposed 0(10)
function ERSs integrated Maximize MeE. f ( z )SE  SEis  defined as the ratio of differen
L  L (16)
6 5 5 5 6 6 6 7 7 6 f f 6 f

Objective function
f 1
) m hc1)  T
(1  y )mV(Lm U ((11mkA)L1(TLhcm)( m T ) 1)) LL(2H yV (Hhm)(2kLm(hh2hc ) )m kL)LCp 3( 0Th30  H 2 ) (11)  0 (3) Objective function(2) (3)
6 6 5 6 6 5 5 5 f
V
amount
)  0 of(12) feed liquor andthere product liquor
(2) objectives:with
: g ( z respect to the amount of SC bytothe M
hc H 0
7
U A (T  T )  yV (  Cp
U A (T  T1 ) 1 m( Lm (1  k1) L  L )( m hc )m kL In
f 6 6
hcTmthe
6

 0 current
7 7

investigation,
7 f f

are
7

two
f

, z ,..., to
z ) minimize thethe SC andand maxi
(1These )V mU(nonlinear
 yfourteen 2 Am 2
7
(Thc 7
mT) 2)ofLthe2(1(MEE
6 7
H )Vhm2(as)mequality
 2ytreated 7
3hc
L (Inhproblem the current
3 )H0with 2) 
f
0 investigation,
(4)
6
In the currentthere
6 6
Subjected
are(4)
(3)
investigation,
f
two to
thereobjectives:
are two to0 minimize
objectives: to minimize the SC
SC
i
and
1 2
to ma n

H 2 is ) ofturned
0 ERSstoinintegrated to a single objective h ( zoptimization problem with several c
1equations m , z ,..., z )  0
(1  y )V m (decision
number of unknown
 m variableshc )  L ( H 2  hsupplied
nd m such as2 fresh steam 2 ) of (VLthe h 3proposed
constraints
of 3), (output
the temperaturefourteen
proposed ERSs integrated
maximize the SE MeE. (3)
ofMeE. AsAs
the proposed SE SE is
ERSs defined
is defined
integrated as theAs
as
MeE. the ratio
SE ratio
i
of of
is defined difference
1
asdifferenc
2 n

vapor from eachthe

stage (T1 U st
and
A ( T 2  T stage
)  ( 1  are y ) V combined
(   hc )with
 0 each other and fed as the (4)
heating
0
(i=1,2,…,6)),
2 2 from and
U 2 A12 (T1 the
output
2 liquor
) and
flow
(1 2nd
rates (L
ym)Vstage
(i=1,2,…,7)).
aremproblem
amount Eqs. (13)
of feed
- (14) for the
illustrates
liquor MEE
where,
and
i,the is
theformulated
is ratio number
product of difference liquor aswith
(4)
between in Eq.
of inequality/equality
respect
the (16).
constraints
amount toof the andliquor
feed n number
amount of unknown
of
and product SCSC byby
liquor decision
the MEE
m (m  hc
m
 T21st m )  0 of feed
The output vapor i
combined with each i

the 3 rd
stage of the st MEE with
ndheating an average temperature amount of T liquor
and
variables. the Butand
Energy
in thisproductwork,equations
there liquor
is no of with
inequality respect
constraint to the
is present amount
only of
equality constraints theare ME
vapor from the 1 and
other and fed as
st 2
the stage
nd are
medium combined
for the 3rd with
problemstage each
of the
is other
MEE
turned avg withand in fed
with
to aas
respect the
single
toheating
the amount
objective
SC by
optimization
the MEE. Hence,
L
this
problem
 L
problem
with
is turned
several cons
utisvapor
ethe from theas 1in and 2(5)- stage andare the combined problem withforeach thisisstage turned
other is and
considered. in
in to fedtoa single
a single
as theobjective
heating objective
optimization optimization
problem withf several problem
1 constraints. with The several co
3represented
rd an Eq.with (6).
average temperature of Tavg Energy equations
stage
rd of the MEE an average temperature
problem offor T the
avg andMEE the is Energy
problem formulated for the
Maximize
equations MEE as isin Eq. : f(16).
formulated
( zas)  in
SE(16).
Eq.

for the 3 stage of the MEE with an average temperature of Tavg and the Energy equations
represented as in Eq. (5)- (6). problem for the MEE
Solution is
Algorithm formulated as in Eq. (16). V
e is represented ( L3  L1as )(inavgEq.  hc (5)- )(6).  L3 ( H 3  h3 )  L 4 ( h4  H 3 )  0In this section, a newly developed (5)a physics-based metaheuristic approach 0
named Flow Direction
age is represented as in Eq.avg(5)- (6). Subjected toto: solve
g i ( z1the , zconstrained
2 ,...,
L f Lz nL )1  0
( L3  L1U )(A3 (T2hc  T3 ))  (LL33(HL3 1)(h3avg ) Lhc(h )H0 )  0 (5) Algorithm (FDA) [17] has
Maximize been applied
(5) (6) : f ( zf )( zpoint SEwith  lowestf height. L
optimization
1
problem. (16)FDA is
( L3  L31 avg )( avg avg hc avg )  L3 ( H 3  h3 ) 4 Lavg44(h4 3H 3 )  0inspired by the flow direction Maximize of the:outlet
(5) hi ()z1 , SE z 2 ,..., V )  0 The performance of this
z n0 V
ay for the 4th and U 3 A53 (thT2stages  T ) of ( L the  LMEE )(avg the  hcenergy )  0 equations (6) are deduced
algorithm has been tested as (6) in
onlyEqs. against the benchmark function0 and some other applications.
U 3 A3 (T2 3 T3 ) 3( L3 1 L1 )( avg avg hc avg )  0 However, it is also imperative Subjected (6)
to to :togthe
observe
:i (gzperformance
1 , z 2 ,..., zofn )this
z 0
algorithm
ay for the 4 th
and 5 th
stages of the MEE the energy where, equationsi, is the are number
industrialdeduced optimization
Subjected
ofasinequality/equality
in Eqs.
problems. Hence, FDA i ( z1 ,constraints
is
z 2 ,...,
adopted in n)  0
this
and towards n number such real time
study to optimize the steam
of unk
In similarthway for the th 4th and 5th stages of the MEE the energy equations h ( z , z ,..., z )  0
r way forarethe 4 and 5 stages of the MEE the energy equations
variables. But in this work, there is(7) are deduced as in Eqs.
no inequality i h 1( z 2, zconstraint
2 ,..., z n )  0
n is present only equality
( Li  L i 1 )( i  hci )  Li ( H i  hi )  Li 1 (hi 1  H i )  0
deduced as in Eqs. (7) - (8). i 1

where, considered. is thethe number of of inequality/equality constraints andand n number of of unkno

( Li  Li 1U)(iAii (Tihc )  Li ((HLii  Lhii)1 )(
1 i Ti ) 
Li i 1 (hc hi iwhere,
)1   0H i, i)  i,0is number (7) (8)
inequality/equality constraints n number unk
( Li  Li 1 )(i  hci )  Li ( H i  hi )  Livariables. 1 (hi 1  H But i )  0in this work,
(7) where, i, is the number
there is no (7) of
inequalityinequality/equality
constraint constraints
is present and n number
only of
equality co
notes the stage Ui A number
i (Ti 1  Tiof ) the ( Li 4 Li and th
1 )(i 5
th
 hcstage.i)  0
Solution
variables. Further, Algorithm
Butemploymentin thisunknown work,decision there
of feed (8) is no
split
variables. inequality
But in this work, constraint
there is noisinequality
present only equality
constraint
th U A ( T  T )  ( L  L )(   considered.
hc ) 
considered. 0 (8) (8)
is present only equality constraints are considered.
the 6the
notes andstage 7thnumber stages
i i i 1
with
of thea 4split
i thi
and
th
fraction
i 1 th i
5 stage. th
of i
In(1-k) andemployment
this section,
Further, k respectively.
a newly developed
of feed The split energya physics-based metaheuristic approach named
denotes
ation the stage number of the 4 and 5 stage. Further, employment of feed split
the 6getth modified
and 7th stages and with rendered a split as fractionin Eqs. (9) Solution
of (1-k) -Algorithm
(12).and Algorithm k (FDA)
respectively. [17] has Thebeen energy applied to solve the constrained optimization pr
to the 6th and 7th stages with a split fraction Solution of inspired(1-k) and Algorithm
by athe k respectively.
flowdeveloped The
direction aofphysics-based energy the outlet point with lowest height. named
49
The perfo
ation get modified and rendered as in Eqs. (9) In - (12).
this section, newly (9) metaheuristic approach FloF
5 )(5  hc
Lquation get5) L6 (H 6  hand
modified 6 ) rendered
L7 (h7  Has 6) in(1Eqs.  k )L(9) f (In h-f this )  (1 k )L
 H 6section,
(12). CpT developed
a fnewly 0 a physics-based metaheuristic approach named
Algorithm algorithm (FDA) has been
[17] tested
has been (9) only applied against the benchmark
to solve the constrained optimization probl function and some othe
L )( UhcA) (TL (HT ) h )( L  L(hL )(H  )hc (1  )k )L(1(hAlgorithm k) L H Cp ) (1T k(FDA) )L0 Cp[17] T  0 has been applied
(10) to solve the constrained optimization pro
TECHNICAL PAPERS Surendra
economy Pratap
of the Singh,
MEE unit. The Nishant Nilekar,
pseudo code Kapil Kumar
of the proposed algorithmGoyal
for the proposed model are
given in Table 2.
Solution Algorithm Table 2 Pseudo Code of FDA
Define objective function f(z)
In this section, a newly developed a physics-
Define input parameters (Lf, T1, Ai, xi ,∀i =1 to7) // Table 2
based metaheuristic approach named Flow Set the initial algorithm parameters
Direction Algorithm (FDA) [17] has been applied LB & UB // Table 1
to solve the constrained optimization problem. Number of Population (N=100)
FDA is inspired by the flow direction of the outlet Search space Dimension (D=14)
point with lowest height. The performance of Max iteration (maxite=1000)
this algorithm has been tested only against the Initialization
Generate the initial flows
benchmark function and some other applications.
Evaluation of objective function value for each flow and keep the best one as the outlet
However, it is also imperative to observe the Creation of β numbers of neighbor with ∆ neighborhood radius for each individual flow
performance of this algorithm towards such real Calculate the objective function value of each neighbor and identify the best neighbor
time industrial optimization problems. Hence, FDA if fitness of best neighbor is less than the current flow
is adopted in this study to optimize the steam Update the flow velocity
economy of the MEE unit. The pseudo code of the V  randn * S 0 // S0 slope vector between the neighbor and current position of the flow
proposed algorithm for the proposed model are  Flow fitness(i)  Neighbor fitness( j )  // for ith flow and jth neighbor and d dimension
given in Table 2. S 0 (i, j, d )   
 Flow x(i, d )  Neighbor x( j, d ) 
*Data available 
from Star Paper Mill, Saharanpur, U.P, India
The operating parameters considered here to Result Analysis Generate the new position
validate the model is a pulp and paper mill data, The performance of theFlow energy
newX (integrated
i)  FlowX (i)MEE
 Flow fitness(i)  Neighbor fitness( j)
 V *  model has been evaluated through

 implication a
present in the north side of India and are available Flow x(i)  Neighbor x( j ) 
newly developed optimization approach as mentioned above. In this study,   the considered
from various literature as prescribed in Table 3. else Generate an integer random number
optimization problem is a single objective maximization problem solved in constraint environment
The considered MEE is a seven effect evaporator if Flow fitness( r )  Flow fitness(i)
to opt the SE value. Then Flow thenewXobtained
(i )  FlowXresults are* (compared
(i )  randn with (the
FlowX ( r )  FlowX i )) previously reported results.
operated at constant input temperature T0
(1470C). For solving this problem, else the above-mentioned algorithm, FDA is employed with 100 search
agents, 1000 iterationsFlow andnewX run(ifor 30 times.
)  FlowX The*simulated
(i )  2randn ( BestX  FlowXresult
(i )) of the FDA converges at 600
Table 3 Operating parameters of MEE* iterations and the bestend optimal result is obtained at 23 rd run with the execution time of 138.56
Update the objective function and position of the new flows
Parameters (Nomenclature) Values (unit) seconds.
end if
Specifically, for greedy
Perform this constraint
selection togrey-box optimization
update population if there isproblem with fourteen decision variables,
better solution
Feed liquor concentration (xf) 0.118 thisDisplay
FDA the algorithm
optimum performs
results satisfactorily. The energy efficiency parameters SC and SE obtained
Feed liquor flow rate (Lf) 15.611 Kg/s byThethisoperating
approachparameters
shows a deviation considered of here
42.21% and 75.4%
to validate the respectively
model is a pulp thanand
thatpaper
of a standalone
mill data,
plant operating
deviation
present inofthe MEEand
42.21%
north unit.
side ofTable
75.4%India 4and
indicates
respectively the that
than
are available optimal unknown
of a various
from standalone parameters
plant operating
literature and MEE
theirin
as prescribed dependent
unit. Table3.
Table
Output vapor temperature (T8) 52˚C 4The
indicates theparameters
optimal
thermo-physic
considered MEE unknown is of the parameters
a seven proposed MEEand their
effect evaporator dependent
by employing
operated at thermo-physic
FDA. andparameters
Fig. 3input
constant Fig. 4 shows
temperatureof the
the
T0
Feed liquor temperature (Tf) 65˚C proposed
(147 0C). MEE
convergence by employing
curve of FDA and FDA. Fig. 3 and
variation in Fig. 4 shows
the heat the convergence
enthalpies respectively. curve of FDA
The and unknown
optimal variation
in the heat
process enthalpies
parameter respectively.
values obtained The optimal unknown process parametermodelvalues obtained
byfor the
Minimum temperature increment Table 3for the energy
Operating integrated
parameters of MEE* MEE have stated a flow
energy integrated MEE model have stated by a flow diagram
diagram as shown in Fig. 5. The SE obtained for the standalone MEE is 4.76 with SC 2.25 kg/s
Parameters (Nomenclature) as shown in Fig.
Values 5. The
(unit) SE obtained for
at first stage(δT ) 10˚C the standalone MEE is liquor
Feed 4.76 with SC 2.25(xkg/s
concentration which having an error of 2.59%0.118 as compare to the plant
f)
min

which having an error Feed

ofliquor
2.59% as compare
flow ratedue
to the plant operated MEE model
(L f) to the incorporation of different 15.611
[15]; whereas, due
Area Evaporator units operated MEE model [15]; whereas, ERSs, Kg/s
the SE enhancement
to the incorporation Output ofvapor
different ERSs,
temperature (T 8the
) SE enhancement percent increased 52˚C 71.1% with the
(Ai) (i=1,2, …,7) 540, 540, 660 percent increased Feed 71.1% liquorwithtemperature
the implication
(Tf) of same algorithm. The comparative analysis of the SE
implication of same algorithm. The comparative analysis of the SE and SC65˚C is demonstrated in Fig.
660, 660, 660 and SCMinimum
is demonstrated
temperature in Fig. 6. at first stage( Tmin )
increment 10˚C
6.
690 m2 Area Evaporator units (Ai) (i=1,2, …,7) 540, 540, 660, 660, 660, 660, 690 m 2

*Data available from Star Paper Mill, Saharanpur, U.P, India

Result Analysis
The performance of the energy integrated MEE
model has been evaluated through implication
a newly developed optimization approach as
mentioned above. In this study, the considered
optimization problem is a single objective
maximization problem solved in constraint
environment to opt the SE value. Then the
obtained results are compared with the previously Fig.3 3Convergence
Convergence curve
Fig. curveofofFDA
FDA Fig.Fig. 4 HeatEnthalpy
4 Heat Enthalpy Variation
Variation of Vapor and Liquor
of Vapor and Liquor
reported results. For solving this problem, the during MEE operation
during MEE operation
above-mentioned algorithm, FDA is employed
with 100 search agents, 1000 iterations and run
for 30 times. The simulated result of the FDA
converges at 600 iterations and the best optimal
result is obtained at 23rd run with the execution Fig. 5 Flow sheet of
time of 138.56 seconds. the proposed MEE by
employing FDA
Specifically, for this constraint grey-box
optimization problem with fourteen decision
variables, this FDA algorithm performs
satisfactorily. The energy efficiency parameters
SC and SE obtained by this approach shows a
Fig. 5 Flow sheet of the proposed MEE by employing FDA

50

Table 4 Optimal Parameters obtained by FDA for the Proposed MEE model
Process Parameters 1 st 2nd 3rd 4th 5 th 6th 7th Output
Vapor Enthalpy (kJ/kg) 2697.3 2681 2636 2622.7 2611.7 2605.6 2600.7 2593.7
 Energy Analysis of Multiple Effect Evaporator in an Indian Pulp & Paper Industry TECHNICAL PAPERS
Table 4 Optimal Parameters obtained by FDA for the Proposed MEE model [7] P. Sharan, T. Neises, C. Turchi, Optimal feed flow
sequence for multi-effect distillation system
Process Parameters 1st 2nd 3rd 4th 5th 6th 7th Output integrated with supercritical carbon dioxide
Vapor Enthalpy (kJ/kg) 2697.3 2681 2636 2622.7 2611.7 2605.6 2600.7 2593.7 Brayton cycle for seawater desalination, J.
Clean. Prod. 196 (2018) 889–901. https://doi.
Latent heat of Vapor (kJ/kg) 2207.8 2236.9 2313.1 2334.1 2351.5 2360.9 2368.4 2379 org/10.1016/J.JCLEPRO.2018.06.009.
Condensate Enthalpy (kJ/kg) 493.5 448.99 328.2 283.14 264.31 248.36 235.49 NA [8] A.S. Hassan, M.A. Darwish, Performance of
Liquor enthalpy (kJ/kg) 314.44 249.25 239.67 228.41 221.83 214.83 201.33 NA thermal vapor compression, Desalination.
335 (2014) 41–46. https://doi.org/10.1016/j.
SC (kg/sec) 1.4142 desal.2013.12.004.
SE 8.1366 [9] H.T. El-Dessouky, H.M. Ettouney, Multiple-
effect evaporation desalination systems:
Thermal analysis, Desalination. 125 (1999)
259–276. https://doi.org/10.1016/S0011-
9164(99)00147-2.

[10] R. Bhargava, S. Khanam, B. Mohanty, A.K. Ray,

Selection of optimal feed flow sequence for
a multiple effect evaporator system, Comput.
Chem. Eng. 32 (2008) 2203–2216. https://doi.
org/10.1016/j.compchemeng.2007.10.012.

[11] S. Khanam, B. Mohanty, Development of

a new model for multiple effect evaporator
system, Comput. Chem. Eng. 35 (2011)
1983–1993. https://doi.org/10.1016/j.
compchemeng.2010.11.001.

[12] P. Sharan, S. Bandyopadhyay, Energy

optimization in parallel/cross feed multiple-effect
evaporator based desalination system, Energy.
111 (2016) 756–767. https://doi.org/10.1016/J.
Fig. 6 Variation in SC and SE at different operating condition
ENERGY.2016.05.107.

[13] M.A. Sharaf, A.S. Nafey, L. García-Rodríguez,

Conclusion
Thermo-economic analysis of solar thermal
Energy integrated MEE simulated in different solution platforms provides a better energy performance power cycles assisted MED-VC (multi effect
than that of the standalone MEE system. The energy consumption in the paper industry can be distillation-vapor compression) desalination
reduced to a notable extent with integration of different ERSs. The principle of process integration and processes, Energy. 36 (2011) 2753–2764.
numerical optimization are the keys to reduce the energy utilization. A steady state nonlinear energy https://doi.org/10.1016/j.energy.2011.02.015.
modeling of hybrid MEE based on 1st principle of thermodynamics is evolved in this work and the
[14] S. Khanam, B. Mohanty, Energy reduction
model is validated by comparing it with the base MEE model operated at plant with available data. The schemes for multiple effect evaporator systems,
performance analysis of the proposed model is carried out with the implication of a new metaheuristic Appl. Energy. 87 (2010) 1102–1111. https://doi.
approach named FDA, and the obtained results illustrate that there is more than 70% of enhancement org/10.1016/j.apenergy.2009.05.003.
in the steam economy as compared to the base and plant operated MEE model with a high reduction
in energy consumption which leads to reduction in economic and environmental issues. [15] R. Bhargava, S. Khanam, B. Mohanty, A.K.
Ray, Mathematical model for a multiple effect
Reference evaporator system with condensate-, feed- And
[1] O.P. Verma, G. Manik, S.K. Sethi, A comprehensive review of renewable energy source on energy optimization product- flash and steam splitting, Indian J.
of black liquor in MSE using steady and dynamic state modeling, simulation and control, Renew. Sustain. Chem. Technol. 15 (2008) 118–129.
Energy Rev. 100 (2019) 90–109. https://doi.org/10.1016/j.rser.2018.10.002. [16] R. Bhargava, S. Khanam, B. Mohanty, A.K. Ray,
[2] D. Kaya, H. Ibrahim Sarac, Mathematical modeling of multiple-effect evaporators and energy economy, Simulation of flat falling film evaporator system
Energy. 32 (2007) 1536–1542. https://doi.org/10.1016/j.energy.2006.09.002. for concentration of black liquor, Comput.
Chem. Eng. 32 (2008) 3213–3223. https://doi.
[3] O.P. Verma, T.H. Mohammed, S. Mangal, G. Manik, Minimization of energy consumption in multi-stage org/10.1016/j.compchemeng.2008.05.012.
evaporator system of Kraft recovery process using Interior-Point Method, Energy. 129 (2017) 148–157.
https://doi.org/10.1016/j.energy.2017.04.093. [17] O.P. Verma, G. Manik, T.H. Mohammed, Energy
management in multi stage evaporator through
[4] G. Gautami, S. Khanam, Selection of optimum configuration for multiple effect evaporator system, a steady and dynamic state analysis, Korean J.
Desalination. 288 (2012) 16–23. https://doi.org/10.1016/j.desal.2011.12.005. Chem. Eng. 34 (2017) 2570–2583. https://doi.
org/10.1007/s11814-017-0180-4.
[5] T. Chen, Q. Ruan, Modeling and energy reduction of multiple effect evaporator system with
thermal vapor compression, Comput. Chem. Eng. 92 (2016) 204–215. https://doi.org/10.1016/j. [18] H. Karami, M.V. Anaraki, S. Farzin, S. Mirjalili, Flow
compchemeng.2016.05.011. Direction Algorithm (FDA): A Novel Optimization
Approach for Solving Optimization Problems,
[6] S. Pati, O.P. Verma, Performance optimization of multiple stage evaporator using interior-point method and
Comput. Ind. Eng. 156 (2021) 107224. https://
metaheuristic approaches in environment of real-time plant complexities, Int. J. Green Energy. (2021) 1–18.
doi.org/10.1016/J.CIE.2021.107224.
https://doi.org/10.1080/15435075.2021.1884868.

51