02/28/2012 10:05 4057445183 INTERLIBRARYSERVICES PAGE 03/08

Chem. EnR., 211(S), 1166-ll71 (2011)

K,,rr.a,1J.
DOT 10.1007/sl1814--0l0--0~R7•x

Dynamic simulation and optimization of the operation of boil-otl' gas compressors

in a liquefied natural gas gasification plant
Namjin.fang•, MyonngWool< Shin**,Soo I:Tyoung Choi0 •, nncl En Su1> Yoon••t

*School of Chemical and Biological Engineering, ASRI, Seoul Nati011al University, Seoul 151-742, Koren
"'*Inchcon Tc1minlll Division, Korea Gas Corporation, Incheon 406,130, Korea
••*Division of Chemical Engineering, Chonbuk National University, Jeonju 561-756, Korea
(Received 9 ./11/y 2010 • accepted 17 November 2010)

Ahstnti:t-Wc propose an algoritlun for the optimal O~'!"ation schedule of the BOG cornpressi(),i process bailed cm
simulation of the dynamic behavior of an LNG timk. Tiie algorithm uses an empirical boil-off Mc model to predict
the mnount of BOG generation, and an MTT..J:> fonnuk1tion lo distribute the BOG C(>mprcssors looclti, Finally, a safely
umtlysis is done using n dynamic simulator. To improve 1he IICCLU'DCY, A~-pen Dyruimics with the .Peng-Robinson equoti.on
of state is ured in pl/.lCC of a simplified dyMmic model tL'led before. The dynamic sin, ulation of the LNG umk pre~
,qnrc ~bowed 1l1e re5tdt.~ of oscillation witl1in 11 sn:fo prc~~'llTE: range while the BOG compresoors were opemJcd nonnally.
The perlormuncc of I.he proposed algorilhm was found to be ~upi.Tior lo the !llgorilbm used in mul.inc processes !IS well
as those from previous works in tenns o[ssJcty 11ml energy soviugs,
Key words: l)ynnmic Simulo1ion, Optimi7..ntion, Opcrntion, Boil"°ff Gos, Aspen DynDmics

INI'RODUCJ1.0N scnls the LNG gasiftcation pr<:ic~. The plw1t receiVC$ LNG from
a carrier nnd st.ores it in a ayogenic liq1.1id s!Dte. 'OJc liquid tempem-
The changing economic outlook, energy demands, and environ• rure is apµoximntely -160 •c and the pressure is slightly above
mental factors have resulted in incre..1.-;m demmcl for liqucl'icd natu- the tllrna;pheric pre.-:oore irl the soi11ge tanks [1,3]. DUt: lo hen!. trun~
ral gos 0,NG). Worldwide, LNG trode h0$ increiiscd steadily (over fer from the sunounding cnvironmenl~ part of the LNG in tl1e tank
5% peryet-1r) sin<:e U1c indu.•'llry hegnn. '.11iis 1rend i::i c.."<ix:ctexi to con- hi conlinuornily evnporatoo as boil--0ff g;:1s (BOG), which should be
tinue as natllt<ll gas 1.xcomcs more widely used due to increllsed r1::rnoved by compres..qora lo maintain the tank pre,sure within 11 safe
energy demands from developing oountrics. Some indu~1xy ana- nU1gc. An t::,':ce,'l$ amount of cvnporatcd BOG causes SAfcty prob-
lysts hove prcdicled lhal demand for T..NG will double by the end lems rclated to the increased pres.."Rire of 1hc LNG tru1ks, but the op-
of tbe next decade [1 ]. eration of the compres..qors requires much more energy than is nc-
An LNG gasification plant is mainly oompoocd oftNG storage cessrny. Therefore, the correct operation procedure for BOG com-
tanks, pumps, LNG vaporiz.crs, and tl1e gas pipelines. Fig. 1 rcpre- J)('e$S()t'S is CSSt,"tltial for tl1c e.fficicncy of the openition and safety of

- I.NG
NO

;-~~--D- -·-
: BOG
800
recondoo!er

: 1:oniprcs..,01
I
1
I

L'NCJ
'
1
I __ __
.•.•.J,.

I.NG
HOJf----•
:
;
Mctc:rii,s
st:1U01t
LJlpump Hl'PtullP
''
Sul~mq.;ed
vnporiY.cI

Fig. l. t.\ typical LNG receiving iind gnsiiication proccu [2].

1To whom CQl"l'C$pondcncc should be eddrc11RCd.

E,mnil: csyoon@pslnb.snu.110,kr
1166
02/28/2012 10:05 4057445183 INTERLIBRARYSERVICES PAGE 04/08

l)ym1111ic simuhttion m-.d oplimil.~tion of the opcmlion ofboil-ocr sas compressors i11 n liquc(icd nn1.uml gns gasificntion plant 1167
im LNG terminal.
Sevcr.:il studies have a1temptcd to :;:ilve 1he trade-Off bct·ween
energy usage and safety a~mcc assooiated with this p~. Ooe
ex11mplc is the algorithm for the opcrntioi1 of oompreroors rop:,rted
by Shin ct ru., which uo:;ed an cmpiricitl boil-off rnte model for the
safe ond ef:Ucicnt opam:ion of BOO compressors [4]. 11Jey pn.,..
po=l n mixed integer linear 111i:>g1mnming (Mll.,P) formulation to
detcnnine the optimal operatfon policies within tl1e s11fe opcrntion
prc.<;.,;ure. 111c optimal opemtion policies =w to distribute com-
pre.'<.SOI" loads, tum oo/oJ:I the time :lrnction of the operl'!tion of a BOO
compressor, and determine any rcd.t.mdru,t compl'C&lor operation$
in considenition of safely using a simplified dymunic model [SJ,
In ci1is paper, we propose l1f1 optimal operation scheme for BOG
oompres.~ that finds the optimal pressure betwCJ.'11 the steady state
and de:.il'Cl.i tank pre:ssure using a tum on/oif oompressor opemtfon 91op3
To improve \he ac:cmicy of the scheme, a dynamic simulotion and
o safety analysis are conducted to test lhc prupo~ opemlion pnxx.'"
dure using the rigot'ous model of Aspen Dynamics in~c~id of tl-ie
simplified dynm11ic mooet used in pn.,"Vious works [4,5J.

PROPOSED ALGORID-IM
FOR OPTIMAL OPERATlON

The proposed pmoo1urc for dc1cnnining the optimnl operation Fig. 2. Flowchart of the rm.•pos~cl algorithm.
of BOG comprc~sor., is illustrated in Fig. 2, '11,e proposed algo-
rithm invCllves the three stq.,s of lo□d c1Jlc111A1.ion, MILP lbmmla-
tion, and dynamic simulation and safoty onlllysis. The rcd.lllldant compre~ m-e on standby at idle level in order
t. Load Calculation fcx safo opemtion. Also, tl1e c,p:."T!ltors noo:l n;; know the time to tum
An empi1iccl model is proposed for the estimation of the rate of on backup comprcssms, becouse 1hc backup compros.sors need the
boil-off gas genemtion in mi LNG storage tonk [4]. sb.utup time which is 30 min i11 tl1e cooling system and tl,c tank pi-es-
sure will increase in this period. In the previous works, we c.<llcu-
(1) latcd the turning on time 1JSing a simplified dynamic model (5]. rn
this popcr:, the snfct.y opernt:ioi1 scheme is proposed by using 11 dy-
where CR is the rollover coefficient ~I) by LNG cit'culatior1 B,, namic simulator. Fust, initial time and limitation pi-e.<:.<n1rc arc as•
is 0,e boil-o!Tra1c fium the specifications Q,- 1) impplied hy thr; lank sumed. '11ien, 1fa: dymrrnic simulatiori i::i performed during Slsrtur
mmmfoctunx, P,, is the LNG daisity (kg/m~), and VL is LNG vol- lime, If the pressure is lower or highc'.'r lhan limitntioo pres..'llll"C, the
ume (m3). K,, K~, and K3ste C011-ection faclocs fortbc offre1. of the initial time will be modified. And then, the procedure is "'--peated,
lank pressure tmm the LNG vapor pressure, lhe LNG temperature, The simulation procedure is also shown in Fig. 2 using the ti£()i-
and the ambient temperstut'e, respectively. ous model from Aspen Dynamic..'!. First, the gencnttcxl BOG is com-
Bm;cd on this model, the total compressor load cnn be cnlcu- puted by tl1e cnlculotioo module. Nert, the pt'C..~tre of the t1U1k (P..,.)
latcd for the desirable highest tank pressure (Pi.). l11e target com- is changoo by the O\'.lC1'8tion of the compressors and updated to the
! ~ lood F0 cm be ctilculal-cd u~ng 1he cnpiriail model equations, BOG colculotion module, und finally the gcncrutod BOG is recom-
which C!ln tc Stm1maii'7..ed as follows: puted, 1be !'Out.inc of this calculetion is rc..p:atcd.
F0=F(P,.) (2)
PUOCILSS DESCRJPTJON OF Tiffi UOG
whcre Fis n lbnction dcfmcd by Eq. (I). CO:tm>RESSION .PROCESS
2. MILP Fonnula1ion
The :tvill,P foimulation i:, pmposed for tl:1e optimal distribution The Korea Gas ~01voro1io11 (KOGAS) has three LNG receiv-
of the compressor load, which minimizes 1he totnl ovc..-rugc powt::r ing terminrus in Koren. 111e receiving t:aminal foci] itie~ arc kcatoo
com,umption. 1l1e details pertaining to 1l1is fonnulation am aveil- in Pyongtack, Jncheon, ond Tongyoung. TI'lis paper focuses on the
ablc in U1e litenilure LS]. Pyongb.1ek LNG 1enninal hi.x.:uusc of the flV3ilobility of it.'l opera--
3. D:m,imic Simulation and Safety AnalyRis ti.on data, process data, and 1he stabiliz.ation of its operation system.
A dynamic simulation and S1Jfot_y nnntytliB Wt.TC performed to J"ll& "Jhi; Pyongtaek LNG tem1irn1l opc:,"Tlltes ten LNG sturn1:,>c ti.mks
diet the pre.<i.'lll'C change in the lank. In the dynmnic simulatioo. ti.Sing with a vok1me of 100,00Jm1 each ond aBOR (boil-off rate) of0.1%/
1hc proJXJ.'5Cd algorithm, lhc pressure of the tank O!;Ci lleted witl1in a day according to the ~ificoti0tlS. 111cm are six BOO comprc:s-
safe presstu-e rnnge (P,::a:P~1,) when the BOG compt'ffi.'54.11'$ WCl'C SOIS witl1 n nmx:imurn c11pacity af 1O tomih cad!. TI1e design pccs-
operated n01mnl!y; here, P, is the steady ::-talc picssm:. suro of a storn.ge umk at this focili1y is -2.S-230 g/cm'· gQuge, 1l1c

Korcnn ,J. Oillm. Eng.(Vol. 28, Nn. 5)

02/28/2012 10:05 4057445183 INTERLIBRARYSERVICES PAGE 05/08

1168 N. J~ngctal.

Tobie I. Opcnition spcclficntion~ or .BOG comprc~so"

., [ '
Mixing LNG Air-COl)lcd cooling pen 1 I
Cooling system 0.,1~ t••P•-.••- •••• ... \~••, • , .., '' ,
(in thcl" stage) (in the I" and 2,., stage)
Flow rate fNm 1/h] 12,000 12,000
P [kg/crn 1] 10.56 10.56
·-1,-,,. ◄ 1
Shaft power [kW] 909 966 -~t.•
Declared power [kW] 1,100 1,100 .... -.p.,,.!~
...... ·,~l'(l:C,

----m•:i.11

Tuhlc 2. Compo.,itions of LNG nnd BOG in pyon~tncl< LNGtc"'

minol (mole%) ·
LNGt (Specific LNG2 (Specific BOG
Composition
Gravity: 0.4746) Gravity: 0.455)
" "
Methane 85.11 89.26 99.64 Fig. 3. Pre(,11.~ ch:mi:,rcs for the vmiou~ 1..·11mp1~1sor opcn1tioo op•
Ethane 9.32 8.64 0.02 tions using tl,c Pcng-Roblnsori EOS of Aspen Dynamic.'!.
Prop11nc '1.58 1.44
iso-Butllrtc 0.56 0.27 lions of 1hc snt1.1mtcd liquid density arc inoccurate using this equa-
n-Butanc 0.42 0.35 tion, wiU, 11n 11vemgc 1.,11ur of 1rppmximalely 10% due to ni1roge;,i,
iso-Pcn1/lne interactions wi1h meth1n1c, ctlmue, or other gn..~ in the LNG mix-
n-Pcntane ltm: [7,8]. llle nitrogen internetlon effect, however, is di:,regardcd.
Nitrogen 0.01 0.04 0.34 in this study.
c~)rbon dioxide Shin ct Ill. [4] reported.changes in the pi-cs.-,urein anLNGtmkfor
comprcs..'IOI" flow mte.5 fo::,m Oto 30 tom.vh. In tl1is paper, the change
in the pressure with various compressor operation options is prc-
BOG comprcs.•urs require 30 mjn of idle Cl[J<,TJlion for s1m1.up and rlicttXl, 11s represented in Fig. 3, using the Peng-Rohinoon cquation
bnvc load levels 0, 50. 75, and 1009!\ at which the electric t.,'1.immt of state of Aspe1l Dynmnics. 1h:) 1-e.5Ults show some di:O.'ercno::s in
is 65, 115, 127, vnd 145 A, rcspcctivciy, when operated at 6,fOO V, the rate of the pres.'ll.l1~
. change, For example, the simufation resu)1s
1be design spccifiClltions of the BOG compn::,"50\'S arc $h0\-vn in af Uic tank p ~ rc11ch 190 g,/cm 2g wiU1in 36 min when lhc com-
'fable I. 11lc nonunl operation range oftbc tank pre.'5.."1.lre i$ between ~ loads are not operated (F=O 1omv11) as compared 1o the simpli-
50 '?;/1.,m2ond 170 &lcrrl gauge. If lhe pres.c;i.n:e decreases to 40 g/ fi~ dynamic model based on the ideal gns law, which requires 40
1.,m2{¥luge, 11 vacuum breaker is ac1iv~tcd, nnd if it h,Ot'C(1oes to l ':X) 'l;I mm.
1.,,n2 f?f!Uge, the BOG is sent 1D n flm-e stnd.;, Toe hcight !lild climnetes· To test the operation prooo:lure g<.."tlcrated by the propo..c:ed algo-
of the LNG stor<1gc tm1k are 55.16m and 57.04m, respectively. ri!hm, the situation fer a high boil-Off rate is c.cosidcttxi. LNG stornge
Monufocturing supports two types of LNG cc.m1fX')sitioo in lhe tcch- !Jinks /ll'e filled ro% in 1l1is co~. th\1s, U1c total 1.:NG vollime is 5X.Xl,C<Xl
niCltl spcciftcntions: LNG 1 (Sp.x:ific Gravity: 0.4746) and LNG 2 m:.1 The LNG tcmj:ff!ltw:e is - l 59 •c, ond the density is 459. 7 kg/
(Si:rofic Gravity: 0.455) in the LNG storage tnnk,, as shown in m • TI1e temperature o:fthc gas phose is -140°C_, ruid the ambient
Table 2. 1i1c composition oftl1c BOG is also presented [6]. tempemturt: is 30 •c. The LNG is pumped out, imd 20% is recy•
clcd. into the ronks. The initial timk pressure is t 50 g,lcm' gauge.
PEIU~ORMANCE EVALUATION To maintnin the initial tonk pressure, two-!llld-a-hnlfloods of com-
prcs.<:01-s ore ncccss11I)'.
A simplified model is oltcn t19'."d for lhe initi11.l step or the proces.~ The proJ.Xlsed method slorts from ll tnrget steady stntc tank pres-
simulation to lower the computation 1ime; how1.,-v<.,:. U1c obtained sure P"S: 170 g,/cm2 g~ugc. 1lle ste.cidy slate tank pressure (P,) is
solution moy not be occuratc. 11,m:fore, 11 ri g~irous model is usually 167.9 g,/crn 2 gm1ge when the oompressor is operated under n .full
considett:d to imµovc 1l1c sinmlntion pe1fotmo11cc nnd increase the oc- load. Let P,,=170 g!cm2 guugc. Thll'l, F."'8.978 tons/h. Optimiw-
curncy of !he results despite the longer cor:npulalkm lime. In this (Xlpcr, tion for this load rerults in the use of one Cc)mprcssor at full level
ml e.quntion of st.ate model is uc:a:l to predict the prossurc change.~ for 0.8978 1111d 0.1022 at tl'lC idle level.
in All LNG tank for occurncy. Amollg the various cq1mtions nvail- 1l1e dynomic simuhition wos perfotmcd using n general LNG
obk:, those gencmlly u=l are the Peng-Rohinson cqm1tion of state composition and the Peng-Robinson equnticm of state. 111e litera-
and the Soave-Rcdlich-Kwong equotion o.f Slate. The Peng-Robin- ture rc:polts thnt the composition of metl=c in LNG varies from
son equntfon of stale is commonly 11'lC(l w:i the c..·qualion of s1ote lo 80 to 97% [9-11 J. ln tiudition, the boil,>ff gns ot 150 !l,-lcm2 gimge
predict the pressure behavior of BOG Using the Pt."llg-Robinf.Oll in U1c tmlk is mainly composed of me1hiUJC and cihane, as shown
cqualion of state, lhc K-value nnd vapor density values can be pn:- in Table 3. 11J.is is calculalcd 11.~ing the flash model in A:;qxn Plus.
dictccl quite aocurntcly. Several studies have :rr..portcd the pn:dictcd The compositi.on af the BOO supporto1 by 1l1c mooufnc1.urcr, shown
tl1eimodymunic data for on LNG- mixture, spccincally tbe bubble in Tnble 2, was not used in the dynamic simulation, as the compo-
point prc...'lSurc !IJld the liquid density, coinparcd wi1h expe1imentf1l sition of the BOO in 1hc toolmic.nl ~pecifa:ations applie- to the &,...'ncr..~
datn using 1hc Peng-Robinson equation of state. However, prroic- conclition of the de..<rign sp::cificotions for 1l1e LNG stornge tank i1t
M11y, 2011
02/28/2012 10:05 4057445183 INTERLIBRARYSERVICES PAGE 05/08

Dynamic simulation and optimization of the operation of hoil-offgns comprcs!!Ors in a liquefied nntt1ml gus Basification plunl J 169

Tobie l Compositions of the LNG and DOG as nn c:uin1pl0

Component LNG (mot%) BOG at 150 g,/cm' g (mol%)
N2 0_[)4 0
Cl 89,26 98.84
C2 8.64 1.16
C3 l.44 ()
n-C4 0 0
i-C4 0,27 0
n-C5 0 0
i-CS 0,35 0

0.0 1Q,0 ~()0 40,0 AQ,0 10a

1·,ma mlrtt<I~•

Fig. Ci. Dynamic slmuh1tion J'csult, to predict the pressure to turn

on II bnckup 1."0mprcs,qor.

Gff1phDOV

min when a eotnprcssor foils without n oockt~) comprES50!; as shown

in Fig. 6. This is E:arlier tbo11 tl1c time p1roicto:l aco:,rtling to the ideal
gas lnw.111is proreclure ii; represented in Fig. 7, where ~ ' : l ( J f I
is the working compressor tllld compt.'CS30!" 2 is the backiJp com-
Fig. 4. Simplified BOG comr,rc.~~i•.•n proCQSS,
pressor Addiiiom1lly, tllii power consumption of this rndl1od i$ corn-
part-d. with 1h9t of a conventional method in 'fable 4. 111c results
!he Pymgtnek tennimil facility. fmm I.he f.lllWt.T cxmsumpl:ion im, lower 1}1un in a prevkl11~ paper
A simplified depiction of the BOO oomprcs..,zjon proce..,;;s is shown [4] because IJ1c compn~srors rn-c operated in a high-presstu-c range.
in Fig. 4. Tt is composed of tl1c gas phase of nn LNG storage tnnk In addition, the sa:li.'1.y of tlie LNG tank when operating thi:: bcckup
9nd a comprc:sso1·, for 1hc simulation, tl1c stomge tank is divided compr=r is gwmmtccd lllrough the tisc of n rigorous dynamic
into the two phcses (liquid and gas), of which tl1e gas phu..~ has bocn simulation.
modcb:i Some fe.<Jlttt'CS of the process ..~re rernovoo to kci.-p it simple
TI1c rc::sull~ of dynamic simulation arc shown in Fig, 5. At sl.lr• CONCLUSION
tup, com[)l~5ror 1 is op<n1tcd undcr a :full lond until the pressun:?
reaches 166.5 &,lcm' g. which requires 36.5 min. Compressor I is An algorithm wo~ J))'Of)O~ for 1he optimal operation of n BOG
then opt.Tnled idfo for 3.6 min until tl1e pr=t1re re1Jchc::1 170 g,/cm~ corn~= at un J.N(t so~·iJication plom. Ba~ul m an ei11piric3!
g. Ne;,.,1, oomprcsror I 11ll1s for 42,6 min under !I foll lo~d and then boil-<:iffra1e model, 1hc total compre,"l.o;or load con be colcnlatt-d for
idles for 2.4 min. 171e operation ptDCO:lurc is Il.,i:,catro, end the p1~- a given target tank pii::s..'ll.llc Nexi. an :MILP formulatkm Wll~ applied
si.tm o..~illatcs within c snfc mogc. The bsckup comprc.=r 2 is 1utnul for the optimal distribution of tl1c comprcs.',Of lood and for the op-
on at a pre=-e of 163 ~,lcm 2 g 16.8 min after the startup and l"D• crntion schedule. Finally, dynamic simulation and safety flI1nlysis
mains on s1nndby, ']110 prcssum n::.ichES ot 190 s,lcm 2 g within 30 were patbrmcd to prn:lict tlie pr·cssun: chm1ge using o rigorous model

0,0 20.0 60.0 80,0 10¢.0 120.0 1'0.0 160 0

TIM.ft tnln111C(\

Fig, S. Dynmnic ~iurnlation remit for the LNG ~tor:igc tnnk pru~~ure,

l<o1·ciin ,J. Chem. EnR,(Vol. 28, No. 5)

02/28/2012 10:05 4057445183 INTERLIBRARYSERVICES PAGE 07/08

1170 N. ,Tm1g ct nl.

1110

1~0

!20

r.,: ti)(;

1
:,>
l,
l!U

60
~
.1(1

:Jl~

0
0 I:! j}1 .36 ~ t'itl ·rz 84 ~)I,' f(l~ 120 t<•~• lttti 1fi0
ltmln)

Fig. 7. Opcr11tino schedule for n BOG c:ompnisso1:

Table 4. Compnrison of opcr11tion mcthm.1~

··- .... -- ...
Mcthncl Ct)mprcssor loads (tons/h) Targct pressu~~ (g/cm 2i:i) Power consumptio11 (kW) Energy ~nvc (%)
Conventional 10, 10, 5, 0 150 3,102
Propo$Cd by shin ct ol. [4J 10,0 167.5 1,386 55.3
Optimal 10/0,0 170 1,358 56.22

in U1c dy11on:tlc simulator of Aspen J)ynomics. A C!lsc study was K1 con-ection factor for the offset of the tnnk pressure from Ilic
:

conducted ot 1hc Pyongtack LNG teiminal. 111c dynamic simulo.- LNG vapor pre;surc
1.ion was done to test the gi:,ncrntc:d proco:lLu~ and pn:dict 111c prrs- K, : correction foctor for the LNG lc:.,mpcratwc
Sll1'C d1angc using t11e Peng-Robinson equation of state of Aspen K) : correction factor for Lhc ambient temperature
Dymunics. To opply V!\rious ooncrntmlions ofLNQ lhc concentro- Ps : ste~dy state tank pressure 1,lq~,lcm 1g]
tion of tl,e BOO w<1s calculnted u..~ng the flash module in Aspen P,,, : desired highest tank pressure [kg/cmigl
Ph.IS. lbc rc:,"Ults of the pressure chnngc for the LNG storogc lunk P,,,,k : pressure of LNG tank [kg,/cm'g I
were similar to those of a simplified dynamic model wi1hin a safe COMP : BOG compressors
mngc. _f-Jmvt-'V\::I; the sol'ely-limil pr-."'>.,urc und time to tum Ori 11 bm.:kup OASPHASE: gas phmie in the LNO- t11nk
comprc5SJr both difl'eredftom th<:..."1C values in 1he simplified dynamic GENBOG : generated BOG
model, TI1e pen:onmince of the proposed algoritlnn i$ superior to TOLIQUID : material sl1'eam to liquid phase
1h8t of both the routine process and tl1c results of previotl!l work in TOCOMP : m:itcrial slreom linking LNG timk to BOG compt~-
1enns of &ucty nnd energ_v savings. 1110 proposed algorithm can sors
incrci1sc !lie efficiency when opcrnting nnd controlling the LNG TORECOND : mmainl st:rcnn:i linking BOO compressor,; to mcon-
g;isification prncess. The dynamics of nn LNG storage ttmk u5ing dcn:-:er
the ngorous model can re !!dop1ed for 1111 LNG tcrminols, piping
networks nnd LNG industiies. REFERtNCES

ACKNOWLEl>GEMENT l . S. l~ B. Lcmmcrs nnd B. V. F1uor Hiu1rlcrn. 1-fydroro,-bon Procr.ss-

ir~, 84(7), 55 (2005).
Thb paper was :SUl)JXlt'tci by ASRJ (Automntion nnd Systems 2. D. I-1 Kirn, l H. J..l!c, H Y Kim und Y S. B1iek,J. Km-e.m1 Inst.
lk~Ut.:h lnstitute) ond by l3K2 l llL Seoul N:itional Univci-sity. Gas, S(3), 22 (200 I).
3. M. W. Shin, G Lee, D. Shin, S. H. Choi imd E. S, Yoon, Pmcecd-
NOMENCLATURE ings ofthe :l"' Infen1<1tia1a/ Symposium 011 Advanced Control of
Industrial Pt'Ocesses (AdCONIP'0.'i), 166 (2005),
F : predicted boil offrnte [tons/hr] 4. M. W. Shin, D. Shin. S. H. Choi ll!ld E. S. Yoon, Korrrm .!. Chem.
FO : total output compressor rate [tons/hr] Eng., 25(1 ), 7 (2008).
CR : rollover coefficient by LNG circulation 5. M W. Shin, D. Shin, S. H. Choi, E. S. Yoon 1.md C. Han, ind Hng.
J33 1
: boil off rate from the specification Qi- ] Chem. Res., 46(20), 6540 (2007).
P,. : LNG density l_kg/m~] 6. D. H Kim, Y. P11d(. I. K. Yoon, .T. M. Hn ru1d Y S. l?flrk,A strKiy on
VL : LNG volume [m3] Iha oprimimtim cfproces.• OJ'emti011fa· lhe qffeoicmcy Improvement

Mny, 2011
02/28/2012 10:06 4057445183 INTERLIBRARVSERVICES PAGE 08/08

Dynnmic simul11tion and optimi,:ation oflhc operation ofboil-ofT'gas comprca11cm1 in a liquefied natural sas sasification plant 1171

efI.NG raCdivr'Hg tannfna( Ka'f!.a Gos Capcmttcn R&D Cenler 437(199!)).

R~pon (2001 ). 9. J. J.,. Woodword,J. Hazard. MaU!1:, 140(3), 478 (2007).
7. K. Nasrifur and M ~an. Fluid Pm• Equilibria, 200(1), 10. J. A. Alderman, Proce.t.v Safety P1"Qgre,"-~, 7A(3), 144 (2005).
203(2002). 11. 13. Eisentrout, !.NG Journal, Fc:brtlaf)', 28 (2007).
8. K. Nasrifur an1 M Moshn:gbi1111,Fn.1idPlwe Bquili'brio, 158-1(,0,

Korenn .J, 01CJJI. Eng.(VoL 28, No. S)