| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 927
Study on the Prediction of Flow Characteristics
Around a Ship Hull
K.-S. Min' J. Cl~oi, D. Yur ~ K. Chunk, B. Chan2, S. Chunk, B. Han
(Hyundai Heavy ludustrie s Korea)
ABSTRACT
A long temm R & D :' og am on She predicts m
of flow characteristics around c ship hull has ~ en
e tablished with She aim of preparing c suffcie tly
accurate computational medhod for practical
application This program is c mposed of 3 tage
st dies The Ist tage st dy on She Investigation of th
state of art of CFD technology has been carried out
For this purpose, 4 CFD cod s w re selected Using th
selected codes, mmmerical computations have bean
pert:3rrned for 5 different ships in 4 different ship types
together with She experimental measmements Th
c mparisons between the e perime tat results Ed th
results of c mputation show that present CFD
technology need to be farther improved for practical
application
INTRODUCTION
b or r to improve She hyd odynamic
perfommance of c ship, it is necessary to have th
k owledge on She -I w characteristics around c ship
hull Ed to utilize th k owledge in d sign Two
medhod are cvaibble to predict the flow characteristics
one is the experimental method, that is, model test,
Ed th other is She computational method so called
c mputational fluid dynamics(CFD) The e. p rime tat
medhod is to mecsme She items rented to She flow
characteri tics by model te ts This haditiorul method
has long I en used Ed has one definite cdvmtage f It
c mparctively accurate results could be obtained b
geneeal, however, His medhod cd longer time Ed
higher co t Fmfhemmore, there are some it ms which
could not be measmed by model te ts Those me
disadvantages of the e perimental method Th
c mputational method is to calculate She characteristics
using nowaday's high y developed computers Ed CFD
codes This medhod not only is ec Domi 91 in time Ed
cost, but also could estimate s me characteristics
which me not possible to be me tared by model tests
Those are cdvmtages of His method b general,
however, She c mputaticrurl medhod does not have
sufficient accuracy Ed reliability y t for th practical
mplicatims This is She fatal disadvantage of She
computation medhod
b order to improve She CFD technology for
She prediction of -I w characteristics around c ship hull,
rented tudies have I ecu actively carried out in She
world Ed I terrutional CFD Workshops have Hen
h id f ee times(Larssffn 1930, Lsrssen et al 1991,
Kodsmc 1994) Followmg She worldwide trend Ed
necessity, Hy ndai Maritime Research Imtit te HMR
has established th long temm R & D Frogmm on this
subject Ed has carried out She tudy The ultimate goal
of She ~eier;h program is to prepare c suffcie tly
accurate predicts m method for th practical
mplicati ms Ed to efhciently utilize the method for She
act cl design Ed pe fommance srLthsis
This prog cm is composed of 3 stage tudies
as follows:
She I t tage st d:: Irrve tigation of th state of art
of CFD technology
the 2nd tage tudy: Improvement of CFD
technology to the level of
practical application
the 3rd stage study: Study for the act cl utili ction
This pmer deals with the Ist tage st dy of
HMRl's long temm R & D prog cm Ed mcludes She
followmg contents:
Selection of CFD cod s
Calculati m of flow characteristics
Mecsmement of flow characteristics
Comparison betw en te t results Ed She results of
calculation
Evalucti m
Th refcre, the primary pmpose of d is study is
m the validation of prese t CFD technology After She
f 11 evuluatim of th Ist stage tudy, be f the
duection of She CFD study shall be established
For She sake of universal validity in She result
of th tudy, 5 differe t ships in 4 differ t ship type
have been selected as the object ships such as very
k ge cmde oil carrier Ed k ge si e bulk carrier as She
OCR for page 928
f 11 slow sped ship type, LPG carrier es the f 11
medium sped ship type, I xge si e contcirux c xrier as
fhe fme medimm peed ship t pe md r~l ship as fhe bU
fne high sp cd ship type Those ships wxe actually —= 0
built md delivered in 1 990's dx,
For fhe computatiorul aruly is, total 4 CFD
codes w ~e selected, thct is 3 w 11 Im wn commemial
codes mdHMR cod Thethecommxcialcodesare
STAR CD, FL NT md SH FFLOW b iticlly,
CFDSH F IOWA code was mclud d How ver
c mputation by CFDSH F IOWA cod has not be
prog~essed satisfactorily, md it was decided not to
mclude m this pep x
E ormous cmou t of c mputatiom md
measmements have been carried out Due to th limited
space, however, only the brief summxy shall be
p~esented md discussed in fhis pep r
NUMERICAL METHODS
The d tails md f xmulations of th mmmericcl
medhodologies for CFD me w 11 mow md
e tff~sively docmmented in mmy litemt ~es Hence,
only mcm fectmes of fhe methodologies will be
described in this p me'
Fu t of cll, fhe d cl coordirute y t ms have
ben cdopted es shown in Fig I The global coordincte
sy tem(x,y,z) is defm d to represe t th flow pattxm
around hull as positive x in the flow di~ection, positive
y starboard, md positive z upwxd wh ~e the origin is
et th bow md undist rbed fie smface; while th local
coordirute syst m(x',y',E) to enlur~e fhe usefulness of
calcuLted wake pattems in fhe propeller desigm whe~e
fhe xigin is et fhe center of propelle' The phy ical
qu mtities m the pep r me n mdimensicrurli cd by ship
lengfh betwen pxp ndicuLus~s), ship peed(lJo),
md flui d dem it p)
~, ~ ~ ~
:_ . _
Fig I Coordincte Sy tem
V'seous BOw
The mmmericcl procedure p~esented m fhis
pmer d als with me mpressible flow The basic
eq ctions f tt govem the flow xc to describe th
(1)
' + U. ' =
bt ~ dx~ x,
+ dx (R d ' u,ul) (2)
wh ~e U. P. Re, u~u, xc fhe velocity, piezomehic
pressme, Rey olds No, md Rey old shess
~e pectively
b ordx to p~edict t rbule t flows vie fhe
quctions, it becomes necessary to mcke closme
assmmptiom about the Rey olds t~ess, becmse fhe
quctions do not comtit te c closed set The turbulence
mod is md process of exp~essi g fhe Rey olds shesses
m temms of th Imow qumtities cm be categorized
mto Rey olds t~ess model RSM md eddy viscosity
mod I dVt~
b RSM, fhe partial differenticl eq cti ms f x
fhe Rey olds st~esses xc fxmuLted md solved The
RSM mclud s the effect of some impxtmt factors,
such as fhe t~eamlme curvat ~e md th body fmce et,
m the chxactxistics of the t rbulence, but r qui~es
cdditional computation to solve the particl differential
quctions for fhe each components of fhe st~ess
Furfh xmme it is still necessary to model some of t xms
m th u equatims Altxnatively algeb~aic shess
mod l(ASM using th algebrcic equati ms imtecd of
fhe particl diffxe tial eq ctions, on the cssumption f tt
cor~ctive md diffusion txms m RSM is Imexly
d pendent on fhe t rbulence kinetic ffUxgy k c m be
mpliedto sevexal e gmexi g felds
b the E M, based on fhe Boussmesq's
hypothesis, fhe Rey old t~esses me ~ep~ese ted es
m m velocity g~adients The E M are ckssed i to
zero, one, md two quction models according to fhe
m mber of partial diffe~e tial eq ctiom
Th mo t wid Iy used model in engmeermg
mplicati ms is fhe k~ model m conjm tion wifh fhe
wall fm ti m m teed of fhe fne meshes near the wall
surtcces
b sol mg flow pattems arommd 3 dimem iom~l
bodies, it is convenient to use fhe boundxy fitted
Coordinate sy tem The particl diffxenticl quction to
h msfomm fhe phy iccl domcm i to fhe computatiorul
domcm mu t be solved, govexnmg qucti ms clso be
h msfommed This t m f xmation m be divided i to
two way: one is thct both ge metriccl md phy ical
variables a~e trm fommed The oth r is that only
geometrical variables me tr m fommed From fhe
OCR for page 929
viewpoi ts of mmmxical procedmes, the pintial
h inwiformation is simpl x, but invokes m m xical xrors
whff~ the dismepmcy betwen fhe flow t~eam md
phy ical coordi die is k ge
To solv fhe gowxnmg quations, the flow
domain is subdi id d mto z fmte mmmbx of cells md
fhes quatiom xc ch mged mto zlgeb~zic fomm viz th
disxeti ztion process such zs Finite Diffe~ence
Medhod FDM, Fmite Volmme Method(FVM), Fimte
Analytic Method FAM, Fimte El me t
Medhods FEM) md so fo th
However, the govemmg quations zre
nowinear md cwwpled fomms of fhe contim ity md
mome tum eq ztions A d, hff~ce it is necess xy to us
m iterativ proced re: SltdPLE(S ml implicit Method
for F'essme Lmked Eqwrtiom), A tifcial
c mpressibility Methods, md FISO Pres w~e implicit
wifh Splitti g Operator) etc The SltdPLE is one of th
mo t widely us d procedmes, but is less cc momical
f m th more ~ecent F150, especially m mmsteady
problems, be mse with S MFLE th iterati m is
~equnedateiKhtimestep
Before solvmg fhe quations, th grid mside
flow domain must be geruxated Th grid pattern c m
be categwi cd i to tmct ~ed md mmshw t red grid
The tmctmed g ids do not ~equi~e sp cial atte tion to
defme th com ctivity betweff~ cells bee mse the~e is
one to one cone pondence betwen zdjacent faces of
neighbori g cells Th un tructmed grids zllow mesh
ml mat h on th mte fiKe betw ff~ zdjiKent cells
md w block, fhus locally enhar~mg th mmmxical
~esolutions whe~e ~equired
Potential nOw
The fl w is zssumed stedy, inotational md
mc ompres sib le Th p ote t ial ~ of the dist rbe d
vlocities(~) is deftned by quatim (3) md will
sati fy fhe Lcpkce eq ztion (4):
U = V; (3)
V2¢ = 0 (4)
On fhe hull boundary the nommal v locity
mu t be zero, md on the f~ee s fiKe boundiny z
simibr relation holds This kmematic condition may be
w itteniKi:
~h~+¢yLy h7=0
(s)
whxeh(x,y)iseq ztionforthewz yrufiKe
A dy zmic fie smiace c mdition may be
obtcmed fi m the contmuity of fhe shesses ivxoss th
fiee s rtace This condition degenerates to fhe simple
tatements that th pressme mu t be atmosphxic zt fhe
s rtace, md wifhout fhe generality this prew re may be
s t zero Neglecting rmfiKe temion md mplymg fhe
Bxnwwlii equdtion the dynamic fie wrtace bommdary
condition may be w itten:
gh+2(~+~+~2 U2)=0
(6)
Finally, the v locity is undiswrbed dt mfmite:
V¢=Uo ~ x:+~
(7)
Th se fie w fiKe bommdary conditions zre
wnlinear imd fhey hav to be zpplied zt im imtially
mDcoow su fiKe iteration procedwre, usmg fhe
solution on Imeari cd bwwndiny conditimms, xc
geruxally zdopted
SELECTION OF CODES
For fhe rake of univxwl validbty of this tudy,
totcl 4 codes w re selected, fwrt is 3 w 11 k ow
commemial code(STAb CD, FLU NT, SH FFLOW)
imd HMK code The chxiKteri tics of eiKh code hcs
been summarized m T wle I
Tzble I
Chara>teri om _
Govemimd
equ dion _
TurbMenoe model _
Near wall
FF di~ _
di6ereh dMn
Grid 6Z tem _
Variable la~out
Velooit~ Fremure t
oouFlmd _
NS: Full NS
Zonal: Fote tial/loteg z~Full NS
bE:Stmdxdkd MKE:Modiftedkd
WF: Wzll fwoction
5: Shw tmed
SG: Stzggxed
Charactxi tics of CFD Codes
EMKI
NF
KE
WF
FVM
s
FG
FIMFLE
FLUENT
NF
MKE
WF
FVM
us
NG
UMFLE
US: Un tmctmed
NG:Non stcgge~ed
OCR for page 930
EXPEF IMENTAL METE[OD
The mod I te ts were c mducted et fhe dep
water Towmg TY k of Bl~Rl Th si 4 of fhe tY k is
210x14x6 m in lengfh, width md depth, re pectively,
wifh maximum cYri4ge peed of 11m/s Th co tents
of fhe model tests, mecYring it ms, md dab
acquisition c mditiom to mvestigate fhe flow
charm teristics around c ship Y e show in Tcble 2
Tcble 2 Model Tests md Deb A 4uisition Condbtions
_ bme(6eo)
40
_
_ 40
15
All fhe tests, except the resi tance test, w4'e
cYried out 4t fhe fw4d model conditim with 4ro
smkage md trim, md w4'e cond cted 4t th desig
sp cd Durmg th mesm ments of waw4 elevation md
local ~41Ocity, the mod I was moved mto fhe po t by
300 mm for the gmges to be ac 4ssible
Ti st desenphou
Duri g the resistm 4 tet the model ww
provided with no cppendcges md flv4e m ~4 ticcl
moti m except Dex oy r mod I ship b the ewe of
Destroy r model ship, fhe resi tance te ts w4'e
perfommed et th conditiom wifh 4ppff~dages md
wifhout cpp nd4ges The towmg pomt was located 4t
LCB md b B.
For the global wa~4 elevation mewmv4ments,
fhe longit ddrul cut method ww utili cd Fom wa~4
gmges of ccpm itY e t pe wifh 50 mm interval w4'e
tied up m om4 umt, so f 4t four Imes of wa~4 elevation
date were obbmed in c smgle ma This holder ww
mo~4d clong c tmss CttY hed et the sidewall of the t mk
Triggermg signal is provided by m opticcl wit h 4t
4 62 m be4d of FF to id ntify the location For th
local waw4 elevati m mewmv4ments, fhe prolx4 of servo
edle type ww cttsched 4t c ha~4rsi g mechmism
md incliru4d by 45° to be Y 4ssible To mex re th
waw4 elevation along the h 11 x rface, th ee pers ms
~ecdthe waw4 profile mdthe aver4ge was bkff
For the local velocity mecY r ments, c rYke
wifh fl~4 Sbole Fitot tulx4s ww ux4d The 5 hole Fitot
t be h4d sphericcl tip wifh 6 mm in d64meter md th
mgle betweff~ w4s of cc ter hole md side holes ww
30° Esch tube was com eted to c p~v4ssmv4 h msd cer
Th kiru4matic calibrati m r mge of th Fitot t bes ww
+30° of pit h md yaw mgle if the flow mgle to lx4
mesmv4d was out of the calibration r mge, fhe date was
d6xorded The mexr ments w4'e cYried out across
fhe cente pb~4 to confum th symmeby of flow
Th verticcl holes to fhe h 11 smiace wifh 3
mm m dYmeter were piemed for the meemv4ments of
fhe hull presx re The holes were on fhe keel Ime md
m th sbtion 2 md stati m I with 20 mm spacing
Th flow liru4s on the h 11 were lsualized
usmg pal t The paint was Ym cpproprYte ml tme of
dy, oil pcmt, wax, Ymd fhl mer Th optimum mixmg
mte will lx4 obtamed fiom the try md error
Uueertehty eu~dysls
Th m erbmty ar3~ly is for fhe resistY e tex
we perfommed by fhe ~ec mmff~dation of fhe 22 4
ITTC resist m 4 c mmittee For the oth r texs, the bie
Ymd precision errors of th g mges; Ymd precision en ors
of fhe meemv4me ts w4'e m~4stigated Th accurm y
of the model geomeby LB,d) is (1, 1, 1) mm The
bie md precision encrs of fhe gmges are listed m
Tcble 3 The resid al flow of fhe tmk is 0 001 m/s in
case of fhe wa~4 elevati m meeur me ts of the hull
eye The mcmacy is wifhm +1 5 mm The a~4mge
value of fhe stmdYd devimims of fhe mecYr me ts
for fhe local waw4 elevm i m md fhe hull p~essme are
0 2 mm md 9 764 N/m, respectively
Tcble3 EnorSom 4soflmtmmentstions
b shmmentmi m BYS limit Frecision
mdx
Velocity (m/s) 0 001 0 0015
Dy 4mometer (N) O I
Thermometer (°C) 0 24 0 16
Wa~4 probe (mm) 0 3 0 7
Traw4rse (mm) O I I
F'esY re t msd cer psi) 0 00625 0 0057
SELECTION OFTEtE OBJECT SB~PS
Th object ships selected for mecYr ments
Ymd m merical Y~ly es comprise 300,00 TDW V CC,
170,000 TDW bukk carrier, 6,300 m LPG carrier,
4,200 TEU co talner cY rier md 5,000 tom 4 d xroyer
Th se fl~4 ships have representmive hull forms of f ii
sl w xpeed f ii medmm sped, fme medbum speed
Ymd flru4 higb sped ship, rexpectively
Fig 2 shows th body pEms md side profiles
of fhe ships TY le 4 md TY le 5 sh w the prmcipal
OCR for page 931
parti tlars of fhe objected ships md model propellers,
t Spectively
300,000 TDW VLCC is fhe KTTC Korec
Towing Tmk Co ference) t mdard ship for the tudy
of flow characteri tics arommd the hull Kim et al 1999,
Vmetcl 1998cb,Choietal 1999) This shipisveb
simibr to th ship which was selected w the one of th
test ewes for the Gothenbmg 2000, c workshop m
CFD m ship hyd ody tmics Of her ships are either
HH Hy ndti Hec y Indushies) stmdard ships m
actualshipsmarmiactut d ttHH inl990's
Model ships w t mcde of wood in mder to
gener tte turbulent flow, fhe st ds of cylind ical
shape(3 2 mm m diameter, 2 5 mm in height md 25mm
mtervul) w t loc tted et St 19 5 md middle of fhe
bulb for fhe ship mod is havmg bow bulbs For fhe
d shoy r model with t bulb, t rbulent stimulltors
were located tt 50 mm offthe bow
Th sccle r ttioO of fhe model ships is 47 56,
36815, 14959, 37441, md 276 for 300,00 TDW
V CC, 170,000 TDW buk carrier, 6,300 m LFG
carrier, 4,200 TEU contamer carrier, md 5,000 torme
d shoy r, t spect~vely
Tcble 4 Frmcipcl Farticukr s of the Object Ships
300,000 TDW 170,000 TDW 6,300 m' 4,200 TEU 5,000 tocce
VLCC bulk carrier LPG mrrier coct dner csuier destroyer
L~(m) 32000 27800 9840 25990 13800
LWL (m) 325 50 283 00 99 69 265 80 138 00
B (m) 58 00 45 00 15 70 3220 1740
T(ml 208 165 60 115 47
5 (mi) 27320 19044 1 2208 8 107424 2201 9
V (m ) 312737 5 174274 7030 59526 5273
Lc~ (m,fwd +) 11 136 8 765 0 948 4 277 2048
C~ 08101 08443 07584 06185 04673
C~ 0 8118 0 8465 0 7723 0 6499 0 6070
kD ~
D
10 ~
t D
20 ~
(c) 300,000 TDW V CC
0~)172,000 TDW Bulk Carrier
OCR for page 932
a:
t'°
RESULTS AND DISCUSSIONS
(c) 6,300 m LPG Carrier
(e) 5,000 tom Ck ss Desh oy r
Fig 2 BodyPkms mdSid Profiles
The selected CPD codes md fhe flow
characteri tics to be ccicuhted or measut d have been
smmmari ed m Table 5 C mputttiom md
measut me ts have ben conducted acccrd6ng to Table
5 for each of 5 d6ffet t ships, md c vc t amou t of
mfcrmation for fhe flow characteri tics have beert
pt pared mong fh m, fhe followmg characteri tics
shcil be presented selectively:
resi tance
profile wave elev tti m
loccl resi tance
velocitydi tribution
limitmg stt tmlit
pt ssme di tribution
wake( tt the prop ller plane)
Tcble 5 CPD Cod s md Characteristic to Be
Cal tk ted or Mec tred
Chsrscteristics H C P
Viscous O O O
Resishace Wave X X X
Oversil X X X
Protile X X X
ElW~s~vtleon Locsl X X X
Glob~ X X X
Locsl resishace O X X
Limiting stresm ice O O O
Velocih distriLctioc O O O
Presscre distriLctioc O O O
Bocads y Isyer O O O
Wske O O O
_ _
b Tcble 5, fhe symbols of H. C, P. S. md M represent
HMk, STAR CD, PLUfiNT, SH PPLOW codes, md
mod I te t, t sp ctively
OCR for page 933
Resi tance md self propulsion te ts w t
c mducted accordmg to fhe ITTC St mdmd Proced t
Fomm factors w t detemmined usmg fhe resist mce
values mea tred tt th low speed t giot Resistance
itself hcs beert n mdimensimurlized to resist mce
ccefft ients Thet fot, fhe total t sistance ccefficie t
for fhe mod I ship (C t c m be represe ted by the sum
of viscous resistmce coefficie t(Cv~ md wave
t sistance coefhcie t as shmwn in equ tti m (g)
C~, = C~, + CW = (I + k) CDU + CW (g)
b eqtwtion (S), k md CFM represe t fomm factm md
fi ictiot~l resi tance coefficient, respectively
b c mputatimurl analy is, iscous resistance
c tld be comidet d to be composed of two tt ss
c mponents which are pemerldicular md t mgential to
fhe ship hull, t sp ctively Whert twm tress coefficients
perperldicular to md t mgential to fhe hull me denoted
by Cv~ md CVF, vise ts resist mce coefhcie t c m be
obtamed as follows:
c~U = c>7 +c~ = 1(cy~ +c~
Kind ot Ship
300,000 TDW
VLCC
1 70,000 TDW
bulk mrrier
6,300 m3 LPG
mrrier
4,200 TEU
cochmer
mrrler
5,000 tocce
destroyer
(withoct
sppecdsges)
b equ tti m (9), c,~ md cv~ me the pemend6 thr shess
coeffcie t md th tmgential tt ss coefficient actmg
m fhe mmit section of the hull, md c m be e pressed as
follows:
4~ = 11Cp g~dA
cvf = llcf g~d4
,/~ )4X (9)
(1 o)
(11)
Th resist mce ccefficients obt tit d m this way hcve
been summari cd m Table 6 Table 7 shows fhe
comparison of t sistance characteristics betw ert te t
t tlt md computed result As show t m Tables 7 md S.
fhet me comidemble diffet t betwen c mputed
md measmed results Futh mmore, th t tlts of
computatim me not comiste t Farti tlarly. fhe
computed t sults of wave t si tance by SE FFLOW
code differ from the te t result very mt h Even if fhe
conhibution of wave t si tance to total resi tance tt
d sign sp cd is genemlly less thtt 2% for f 11 slow
peed ships, the reas m of fhis ddscrepff~cy should be
mvestigated md improved Fm practical purpose, my
predLction method should have at lea t +3%, or
prefembly +2% accutacy md pt cision
Table 6 Comparisons of Resist mce Chmacteristics between
Model Tests md Numerical Calcul tti ms
Method
M
H
F
M
C
F
M
E
F
M
E
F
M
H
F
s
CVM
I 00 00
95 34
98 93
105 03
9323
10000
9917
115 41
78 08
89 99
10000
83 97
95 44
103 99
90 80
10000
10793
98 65
109 87
103 39
10000
145 96
96 71
100 72
10066
1 cw 1 Cl,,, 1
10'0 10'0
j 1126 19
10000
776 60
10000
13491
10000
104 40
lOtOO
98~52
lOtOO
IOS46
100000
343 33
10000
12099
lOtOO
93 67 1
97 64
OCR for page 934
Table 7 Resist mce Characteri tics Model Scale)
Kind of Ship Test Condition Medhod C Mxto3 C Mxto3 CwxlO C Mxto3
Model test 3 841 0 042 3 883
LM =6 728 m E~1 3 662
300, 0 TDW :1 6 m/s STAR CD 3 174 3 800
Fn =0 142 FLllENT 4 034
SE FFL W 3 581 0 473 4 054
Model test 3 737 0 047 3 784
LM=7 551 m E~1 3 706
170,000 TDW 1 272 m/s
bulkcarne R :097 10 STAR CD 3006 4312
Fn=0 147 FLllENT 2 918
SE FFL W 3 363 0 365 3 728
Model test 3 705 2 475 6 180
LM=6578m E~1 3 112
LFG car ~er ~t 995 m/s STAR CD 2 976 3 536
Fn =0 249 FLllENT 3 853
SE FFL W 3 364 3 339 6 703
Model test 3 038 0 24 3 277
LM=6 942 E~1 3 279
4,200 TEU 2 018 m/s
co tamercarrier R ~ 40 10 STAR CD 2821 2997
Fn =0 239 FLllENT 3 338
SE FFL W 3 141 0 824 3 965
Model test 3 340 2 545 5 885
5,000 tom~e LM =5 0 m E~1 4 875
d~sp=nt :2 98t m/s STAR CD 2 952 3 230
(w6-ho t Fn =0 425 FLllENT 3 364
SE FFL W 3 362 2 384 5 746
OCR for page 935
Wave profile
Fig 3 shows th comparis m of wave profile
between measured result md computed result by
SH FFL W code As shown in Fig 3, two results
cg e w 11 except in the forward part md fihe Oft part
As w 11 know, however, even fihe linearized potential
fiheob predicts wave profile verb w 11 For the regi ms
apart fi m th hull, particularly m fihe wake legion, th
c mputed result shows the exaggerated wave elevation
for all kinds of ship
Oo _
0
I~ ;
for
Local resistcoce
(c) 300,000 TDW V CC
(c) 300,000 TDW VLCC
O 6,300 m LEG Carrier
0 03 04 06 08 ~
K) 4,200 TEU Co tamer Carrier
Fig 3 Comparison of Profile Wave Elevati m
Fig 4 shows th local resistance coeffcie 33
predicted by HtdRI cod Since it was not possible to
be calculated by ocher codes, comparison could not be
made How ver, it could be d duced Fiat fihe
c mponent d e to pe pff~dicular to ship hull is
dominent m th foreyard md oft pants while th
component due to Fiction is domment m fihe middle
Begun
~ . ~ i, ~ ~ ~ ~ ~ ~ ~
K) 6,300 m LEG Carrier
(d 4200 TE
(e) 5,000 tom Desh oy r
Fig 4 Longitudinal Distribution of Local Viscous
Resistmce Components byHfifRI Code
OCR for page 936
Velocity distribution
Fig. 5 shows the axial velocity distribution for
30O,000 TDW VLCC at the longitudinal position of
station l~x=0.95~. Due to rapid change in hull shape,
there exists low speed region in aft part. This region
could be well measured by experiment. In general,
however, the computations do not show this region
clearly. Particularly, the compution by HMRI code
shows comparatively thicker boundary layer thickness
and slower velocity gradient.
o
z
1~ 1 - ' 5--
-0.06R /
=,
-0.08
1
0L
-0.02
1 ~,,-
-0.08 ~
-0.02
-O.OX
o
-0.02.
-0.04
-0.06
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 0.02 0.04 0.06 0.08 -0.08
y
(a) Experiment
--—, . it,, O S- I
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 0.02 0.04 0.06 0.08 -0.08
y
(b) HMRI
z
-0.02
-0.04
-0.06
. .
. . . . . . . . . . .
0 0.02 0.04 0.06 0.08
y
(c) STAR-CD
,,,,/
<./ -~',o.1 A/ /
/
0 0.02 0.04 0.06 0.08
y
(d) FLUENT
P~
l ~
l ~ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 0.02 0.04 0.06 0.08
y
(e) SHIPFLOW
Fig. 5 Axial Velocity Contours of 30O,000 TDW
VLCC at the Station 1 (x=0.95)
OCR for page 937
Limiting streamline
Fig. 6 shows the measured and the computed
limiting streamlines for 6,300m3 LPG carrier. It is
clearly shown in Fig. 6 that the measured streamlines
are directed downward in the forward part and upward
in the aft part. However, the angle of computed
streamlines with respect to free-surface is not as steep
as that of measured in the forward part. In the aft part,
the computed stremlines are concentrated to the
propeller shaft region rather than direced
upward particularly, those from HMRI and
SHIPFLOW codes). Furthermore, the computations by
STAR-CD and FLUENT codes show flow separation
in the wide region of aft part. The results from
SHIPFLOW code also show this phenomenon weakly.
However, this separation phenomenon has not been
shown in the results of paint test. The method of paint
test is not sufficient to validate the separation and
further study is necessary.
(a) Experiment
(b) HMRI
(c) STAR-CD
(d) FLUENT
(e) SHIPFLOW
Fig. 6 Limiting Streamlines for 6,300 m3 LPG Carrier
OCR for page 938
Pressure distribution
Pressure measurements on the hull suface
have not been carried out yet, except on the keel line.
However, computations have been performed for all
CFD codes. Fig. 7 shows the comparison of pressure
distribution on the keel line for 3 selected ships and Fig.
8 shows the computed pressure contours on the hull for
17O,OOO TDW bulk carrier. In general, the
characteristics of pressure distribution is in accordance
with the direction of the limiting streamlines. It is
shown in Fig. 8 that rapid pressure changes occur in the
forward and aft parts where hull shape changes rapidly
and that the rate of pressure change is greater in the
forward region than in the aft region.
(a) HMRI
(b) STAR-CD
(c) FLUENT
(d) SHIPFLOW
Fig. 7 Pressure Contours on the Hull for 17O,OOO TDW Bulk Carrier
OCR for page 939
0.00 0.25 0.50
X
(a) 30O,OOO TDW VLCC
1.0
0.8
0.6
0.4
~4
0.2
0.0
-0.2
-0.4
o Fop.
HMRI
----- STAR-CD
------ FLUENT
-''-''-''-'-- SEUPFLOW
0.75 1.00
o Exp
HMRI
----- STAR-CD
------ FLUENT
-''-''-''-'-- SEUPFLOW
0.00 0.25 0.50 0.75 1.00
X
(b) 6,300 m3 LPG Carrier
1.0
0.8
0.6 ~
0.4 l
0.2 ~ '1
0.0 ~ _ +,,=_ __ _ A_
-0.2
0 4 0.00
o Exp
HMRI
----- STAR-CD
------ FLUENT
-''-''-''-'-- SEUPFLOW
0.25 0.50 0.75 1.00
X
(c) 4,200 TEU Container Carrier
Fig. 8 Pressure Distribution on the Keel Line
Wake~at the propeller planet
Fig. 9 and 10 show the axial velocity
contours(wake) and velocity vectors on the propeller
plane for 17O,OOO TDW bulk carrier and 4,2000 TEU
container carrier. It is not easy to make a definite
comparison between the measured and the computed
results with these figures. Therefore, the radial
distribution of mean axial velocity, that is, the
circumferentially averaged radial distribution of axial
velocity has been prepared, because this information is
used in the actual propeller design. It has been found
that there are rather big differences between the test
results and the computed results which could not be
accepted in the practical purpose.
/,f: / Jo,_
, 0-2037 ,i ~-
,;,,/,,>/ ,,
'A ,~
; ~ Itt I~g
, ~ ~ ~ /
, a::, ~ ~
~.J
O.'
O.4
O.`
O.2
r/R
0.(
-O..
-O.`
-O.4
-O.,
j j~
, \, \\
;\ ,iw ,
1.0 ~
., I
0.8 ~
''.
0.6 -\
\,
0
0.4 .-
\ 0.2 ' ,;
r/R
. -02 \
,. -0.6
-0.8
-1.0
,
1'''' ~
i,- / chub {'
,, / ' /
, ~
/ 0.2
(b) HMRI
0 ~ (IS
r/Ro.o ~~~ - ! ~ ~ \ r/Ro.o ~~,~· ~~o~ \;'¢\x~
-02 'I -02 a: ".
-0.6 ~ a/ ~ ~ ~ <~ 0.
-0 8 -0
0.2
(a) Experiment
(c) STAR-CD
~ :~ ::::
(d) FLUENT
1.0~
0.8 lti~ > I\,,
r | ~ 2
-0 2 7> ~ ,.~, :
-0.4
-0.6 .
-0.8 ~:~
~2
-1.0
(e) SHIPFLOW
Fig. 9 Axial Velocity Contours and Velocity Vectors
on the Propeller Plane for 17O,OOO TDW Bulk
Carrier
\
.
,. ,~
il
OCR for page 940
l.or
1.0 _
0.~
n ,
O.
O.'
r/R
0,1
n
n.
-0.1_
-0.O
-1.0
· ~~
i
,1'1, Oo At,. .. \ ~
(a) Experiment
n'
o.
O., ~
r/R
0.~
-O.2
-O. .
-0.1
-0.8 ---
-1.0 ~
~ 0.2~> ~ - f
1 o.o ,~.
./ ~ i I'm\
-0.4
-0.6
-0.8
-1.0
D ~~ 0.6
; r/R
; .k',/\i~ O (
,~.~ .~
,~ 0.4
it,/ -0.6
-O.
0.2
-1.0
1.0
0.8 ~ i' i: \
0.4 _ 'I ' it's\
-0.2 , i I ,
-0.4 ' 1 '
-0.6 ~; i,'
-0.8 ~'~ ~ ~ ~2
-1.0
(e) SHIPFLOW
0.t
ji/. ~
~ ,~
1- .
,.
Am\
" . /-1
I
· \ Ace j
~ j
(b) HMRI
..
1.0 ~ .
l Leo,
I · ~ \ \,' ~ ~
' 1 ' ~ ' '' \\
,1 \ . /\ ~ v.
I .! \,< . \ \ \ '
l ,~ Ad. i. \ . \. \. WE
. Ado.\> >.~
I l. ..~ \ \. . '. I. \ . \ .
. to \; by \. ~
\y I, \1\ ~ joey ,-
~.\-I\` Jo \ \ \.
~~ ~s1~
1 >!'''l''1''~
Hi/:>,, I , ! ; ~ j \ j \ j j ~
. . '3,,K,~ V\ ht ~ \ Vow V ', v
-A ~ , i; i; \; \/; ~ . \ ./
-. ,' ,,'1N 1 /' \ i ';; 1 ' ~ 1
. =, I/ ;~Y \; i ~ ;/'
~ \/ ~ ~ j \; V i ~ i
i 1/, j ~ j i; \/j,;,
= ~ 1 1: i ~ i ~
(c) STAR-CD (d) FLUENT
0.~
r/R
0.0
Fig. 10 Axial Velocity Contours and Velocity Vectors
on the Propeller Plane for 4,200 TEU Container Carrier
CONCLUSIONS
The flow characteristics around a ship hull
were investigated through the numerical and
experimental methods. For the numerical analysis, four
CFD codes (STAR-CD, FLUENT, SHIPFLOW and
HMRI) were used. These numerical results were
compared with those of model tests. The object 5 ships
were actually built in 1 990's.
In general, the comparison between numerical
analyses and model tests showed rather large
discrepancies and lack of consistency both
quantitatively and qualitatively. Further improvements
in the accuracy and the precision of the codes are
necessary in order to be used for the practical
applications in the prediction of flow characteristics
and hull form design. And it is also necessary to know
the flow characteristics in the full scale of a ship by the
numerical and experimental methods.
Among the four selected codes, only
SHIPFLOW can treat the free surface effect using the
potential flow theory based on the Rankine panel
method. Large differences between analyses and
model tests were especially shown for wave pattern
around the stern of a ship. The effects of the free
surface to the flow characteristics around a ship will be
more clearly concluded later after further analyses
using the free-surface viscous code based on RANS
equations.
As was mentioned in the Chapter of
Introduction, the 2nd and the 3rd stage studies will be
carried out based on the results of present study.
REFERENCES
Choi, J.E., Sea, H.W., Han, B.W., 'Experimental Study
on the Flow around a Full Slow-Speed Ship', Proc. of
JAKOM'99~ 4th Japan-Korea Joint Workshop on Ship
& Marine Hydrodynamics, Fukuoka, 1999.
Kim, W.J., Kim, D.H., Van, S.H., "Calculation of
Turbulent Flows around VLCC Hull Forms with Stern
Frameline Modification", Proc. of the 7th International
Conference on Numerical Ship Hydrodynamics,
Nantes, 1999.
Larsson, L. (editor), "SSPA-ITTC Workshop on Ship
Boundary Layers", SSPA Publication No. 9O, 1980.
Larsson, L., Patel, V.C., and Dyne, G., "Ship Viscous
Flow - Proceedings of 1990 SSPA-CTH-IIHR
Workshop", Flowtech International AB, Gothenburg,
Sweden, 1991.
Kodama, Y.(editor), "Proceedings of CFD Workshop
Tokyo 1994", Tokyo, Japan, 1994.
Van, S.H., Kim, W.J., Kim, D.H., Lee, C.J.,
"Experimental Study on the Flow Characteristics
around VLCC with Different Stern Shape", Proc. of the
3rd International Conference on Hydrodynamics
(ICHD), Seoul, 1998.
Van, S.H., Kim, W.J., Yim, G.T., Kim, D.H., Lee, C.J.,
"Experimental Investigation of the Flow
Characteristics around Practical Hull Forms", Proc. of
3rd Osaka Colloquium on Advanced CFD Applications
to Ship Flow and Hull Form Design, Osaka, 1998
Representative terms from entire chapter:
ship hull