Twenty-Third Symposium on Naval Hydrodynamics (2001) / Chapter Skim
Currently Skimming:
Free Surface Viscous Flow Computation Around A Transom Stern Ship by Chimera Overlapping Scheme
Free Surface Viscous Flow Computation Around A Transom Stern Ship by Chimera Overlapping Scheme
Pages 441-456
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 441... ...
ABSTRACT The focus of this p Her is to investigate the c mm ility of a numerical scheme Hat inco porates a free so face bommdffy treatment md a smkage/tam calculation into a viscous flow computation mound a su face ship Utilizmg d is scheme computations me pe to med on two selected h msom tern ships The ff ility to compute file wet, partially wet, or d y t msom is developed, including a new g id topology for h msom .-. - 3 Computational results on two t msom-stff ships are consistent with available model experimental data The sh me of the wave profile along the hull is generally predicted well, while the magnitude of wave elevation still needs to be improved The wave pattern is also well copulated, except that wave cre ts md troughs me not ff Shop ff the model measurements The computed total resister ce is in good con elation wish model measured values Aldhough two different types of grid pacing md di tribution techniques me investigated, a father tudy is needed to dete mule a more effective gad system for free su face viscous flow computations INTRODUCTION Computing file flow around a so face ship moving steadily in calm water hff beff a challengmg research task for yews The complexity of file flow physics ff und a smface hip hff generally required tw sepmate numerical approaches, one to compute the wwe elevation on the fi ee smfff e md a second to compute file viscous boundary layer mound ship hull The basehne equation to compute file fiee so face elevation is usually based on potential flow Leo y The numerical scheme c m use either a complicated Green fmmction or use file p mel method with simple Rmkme sources The governing equations to compute file viscous bommdffy layer on a ship hull su face me based on the Nwie~Stokes equations The mlm en cat medhods to compute viscous flow h we been developed fiom a simpli led boundary iayff mprox im at ion to a vat i ety of different lo mu l at i ons of N wier-Stokes equations in this p me' a combmed numerical techmique based on computing the free su face viscous flow ff und a realistic ship hull is developed it is based on sol ing the Reynolds Averaged Navier-Stokes (R NS)
|
From page 442... ...
The compaisons include tota resistmce coeffcients, smkage/tam, wave prohles, md wave paten~ m the fiee su fa e The numerica treaments for computmg fi ee smfa e elevaion md smkage/tam a~e discussed m detail The computaiona g id topology used to hmdle t msom tem geomet y is descabed, a well a gad spa mg, grid di tabution, md computaiona grid domain A limited grid st dy on flow computaions is briefly discussed NUMERICAL SCHEME Nmmerica caculaion of viscous flow aound a su fa e hip mu t mclude tw major physica mechmisms: computaion of the water elevaion a the fiee su fa e md computaion of fLe viscous bounday layer a und fLe ship hull Until tly, fLe fiee su fa e wave elevaion generaed by a moving ship ha been h mdled by the potentia flow approximaion, which asumes no iscosity The caculaion of fLe iscous bounday layer is pe fo med by fLe doublebody approximaion, which igmores fLe effect of the fiee su fa e Ba ed on this kmd of decoupled nmm~sica treament, the free su fa e viscous flow a ommd a moving hip c m be affordably solved The interation between viscous bounday lay~s a d fiee su fa e wavemaking is a sumed to be minima As for pra tica :~gineering ca cula ions, this a sumption rem ains va id to suppo t ship design effo ts However, due to recent advmces in CFD numenca schemes md computer CPU power, fLis decouplmg of the mmmerica medhodology may not be necessay One cm compute the iscous bounday layer md the free smfae wave elevaion together in m affordable ma ner This is a hieved by solving the R NS equa ions wifh ca ef I trea ment of free su fa e bommda y The fundam:~ta RANS equaions in fLe free su fa e flow computaions remain the same a those used for caculaions of doublebody flow, with fLe exception of the pressure te m which ha to mclude staic head fiom fiee su fa e bounday The numerica heament of double-body RANS computaions cm be gen~saly adopted in fLe free sufae R NS caculaions wifLout my modiEcaions The mmmerica scheme for fLis doublebody R NS computaion in FREE9S ha been m developm:~t for yea a d will not be discussed m this paper Lin et a 1995, Lm et a 2000) The mam numenca effo t h~se is to h mdle the movement of the fi ee su fa e bounda y Sev~sa numenca methods have been developed to compute this hee su fa e bounday, md they cm generaly be clasibed m th ee ways: su fa e htting, su fa e t~ckmg, md su fa e capt rmg The effo ts of al these methods a~e to compute fiee su fa e elevaion by saisfymg bodh kmemaic md dynamic bounday conditions required in the free su fa e The kinemaic bounday condition forces water paticles on the fiee su fa e to remam m the bommday su fa e al the time The dynamic condition sai hes constmt amosphenc pressure on the fiee su fa e bounday F EE9S ha adopted fLe su fa e htting medhod developed by Famer (Fam~s et a 1993)
|
From page 443... ...
is a highspeed, h msom .'. -I dispiacemfft ship Model 5415 is a so face combat mt hull selected by OUR for CFD validation Dimensions for file hullfomms me show m table I md bodyplm plots of tenons 0 to 20 se show in figures I md 2 The only appendages included in either file compm.rons or file model tests me show in figures I md 2 Model 5365 hff a centerline skeg md model 5415 hff m integrated skeg md sonar dome Figmre I DTMB Model 5365
|
From page 444... ...
0.183 0.248 CHIMERA OVERLAPPING GRID Grid Topology Free surface viscous flow computations involve two major physical parameters, Reynolds number and Froude number. To accurately compute viscous flow phenomena due to the effect of Reynolds number, fine grid resolution is needed in the boundary layer.
|
From page 445... ...
A 1r~nsverse cu1 of 1be grids ~1 S1~1ion 1 To co~ute ~ccur~1ely 1be ~ve system gener~1ed ~round ~ moving ship, especi~Ny 1be .
|
From page 446... ...
h om the hull to the outer bound The mitial spacing no mat to file wal wa si ed to y+ value of 10 The growdh rate of the grid pacing wa a con t mt 10% for file Bust wavelength away fiom file hull md then law, mptoticaly approaches 0% at file outer bound, where the gad spacing is Lw /10 The effect of this di tabution is to concentrate a large Faction of the total the points inside file boundary layer A single tw -sided Monotonic Rational Quadratic Spline distribution fiom the hull to file outer bound wa used for subsequent gads of Model 5365, refened to a G id 2, md for bodh 5415 grids The initial spacing nommal to the wall wa sized to y+ value of I The mitial grow h rate of the grid off file wall is exLemely large md this results in a more unifo mly spaced g id, with fat fewer points Aside the bounda y layer For file double body blocks in all gads a TANH distribution wa used no mat to the hull Gnd I wa intended to be relatively con e a d file dimension on file fiee so face nommal to file hull wa chosen to be 61 The mproximate total number of grid points wa 500,000 Grid 2 md bodh 5415 grids were si ed to 111 pomts md reamed in file other dimensions a well The mproximate total number of grid points wa 1 3 million for each gad EXPERIMENTAL DATA Model 5365 wa tested m the bare hull condition ~ ah the centerline skeg in place Model tests were condu ted at tw different facilities, one at David Taylor Model BE m (Jenkins 1964) , md file other at the National Maritime Instit te (Gadd md Russell 1961)
|
From page 447... ...
-I wave elevations in the hxed condition a Fn = 0 48 is show m Flgn e. 28-29 The computed elevations fiom Gnd 2 how better conelaion with expenmenta data ah m G id I Since the meatt ed data is archer limited, file comparison here is only qu mtitaively -and ted Forces md moments are calclllaed Song wifh computations of flow va iables m the process of flow computations They a e obtamed by integmting the shed stress md pressure on file wa I smfa e The total resist mce of a smfa e ship is flus computed a the x-componem of total mtegraed force a ting on the ship wetted hull su fa e Table 2 how file computed total resista cc compa cd wifh available mea ured da a for Model 5365 computation For file hxed condition, bodh calclllalons on G id I a d Gnd 2 ale mostly over-predi ted compa Id with Jenkms's measurements, while file result from G id 2 computation is closes to file model meatt ed data However, the prediction treed is consi tent wifh expenmenta t end b Table 3 file computed results for sink tam condition conelae well wifh the model test data with file exception of Fn = 0 65 in Tables 4 md 5, file computed Linkage md him forModel 5365 computation using Gnd 2 are listed Song wifh file model measured data it is found that both sinkage md trim cumputatiuts have consistent trends wifh model experimental data An mcreaed positive tam (bowup)
|
From page 448... ...
Although two different types of grid spacing and distribution techniques are investigated, a further study is needed to determine a more effective grid system for free surface viscous flow computations. In addition, more systematic comparisons between numerical flow computations and model experimental measurements are needed so that the applications of numerical flow tools into ship design effort can finally be accomplished.
|
From page 449... ...
_ I O > ~ -0.05 OR 1'-' Fixed Condition · Jenkins -- -- -- - Grid 1 — ~ — ~ , ~ , 0 0.2 0.4 0.6 0.8 Distance from FP, x/L Fig 9. Comparison of Wave Profile for Fr # = 0.35 0.15 1 ~ 0.1 LL —0.05 .= I O > ~ -0.05 Fixed Condition · Jenkins 0.2 0.4 0.6 0.8 1 Distance from FP, x/L Fig 10.
|
From page 450... ...
Comparison of Wave Profile for Fr # = 0.35 0.15 ~ 0.1 LL —0.05 . _ I O > ~ -0.05 Fixed Condition · Jenkins — Grid2 · - - — i...
|
From page 451... ...
0.15 1 ~ 0.1 LL —0.05 . _ I O > ~ -0.05 Fixed Condition · Jenkins Grid 2 0 0.2 0.4 0.6 0.8 Distance from FP, x/L Fig 23.
|
From page 452... ...
Comparison of Wave Profile for Fr # = 0.48 0.15 1 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 ~ 0.1 \- ~ ~ 0.05 .= I O > ~ -0.05 n 0.06 I, O 0 03 Sink&Trim Condition Jenkins Gadd&Russell — Grid2 .
|
From page 453... ...
Comparison of Wave Pattern for Fr # = 0.41 1 1 2
|
From page 454... ...
. = 0 02 04 03 08 1 D stance from FP x/L Fig 33 Comp of Wave Profile forFr#= 0 41
|
From page 455... ...
4 in Figure 30 md 31, we see z comparison of the predicted fiee surface viscous wave field to the model measurements using the whisker probes As characteristic of mmy similar comparisons, the predictions are very smooth in nature whereas the model measurements have z great deal of high frequency content that is not captured m the prediction My observation of mmy model tests conducted in very still calm water is that the flow field around She model is in reality m unsteady flow with z significmt temporal flow variation tendency near She tr msom md that the high frequency content in She wave field is real Could the mfhor's comment on She resolution of this problem of z steady state prediction for phffmmenz that has some temporal variztmn? Kzrafiath G
|
From page 456... ...
; also, Proceedings of She 2nd Co terence for New Ship & Marine Techmology into 21 st C ntury, Hong Kong, June 1995, pp 53-92 [E glish] AUTHOR'S REPLY None received
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.