Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
TOWARD VIRTUAL REALITY BY COMPUTATIONAL PHYSICS H.Miyata (University of Tokyo, Tokyo) Abstract Twenty years' research works by the author and his colleagues at the University of Tokyo are reviewed and it is described and discussed that the research is moving toward virtual reality technology by use of computational physics, especially computational fluid dynamics techniques. Computer technology is first utilized for the purpose of elucidating physical phenomena, especially nonlinear ones such as turbulence and shock waves and then for the purpose of designing or inventing new systems including ships. 1. Introduction Naval architecture plays one of the most important parts of moving technology and has made great progress in the 20th century. In the last 20 years of the 20th century the computer technology had accelerated the progress of industrialized world and it had been also true in the field of navel hydrodynamics. The computer has become 10,000 times faster or cheaper in the previous 15 years. The ability of a supercomputer in 1985 is that of a personal computer, which we can purchase with 2000 US dollars in 2000. The purpose of engineering research may be classified into two, that is, to elucidate physical phenomena that influences the mechanism of engineering products and to invent some new system of advanced availability. In the field of naval hydrodynamics we had two major physical phenomena to which elucidation of the structure and mechanism was of substantial importance. They are nonlinear waves including wave breaking and turbulent flow including large-scale separated flow. Since both are typically nonlinear phenomena, the theoretical fluid dynamics could not make important contribution. And noticeable attack was started at the end of 1970's. A great amount of efforts have been devoted for the elucidation of such physical phenomena especially in the flow about a ship hull. Although experimental work has made important contribution, the technology of computational fluid dynamics has made much more important contribution with the aid of advanced computer technology. After significant progress was made for the elucidation of physical phenomena the technology of CFD has been used to facilitate design of ships hull-form and other artifacts. At the beginning the purpose was to develop a numerical towing tank. This was achieved by solving the steady flow about a ship with free- surface. In parallel with the advance of computer technology this technique has been extended to the simulation of ship maneuvering and motion in waves. The CFD solution is combined with the solution of the equation of motion. Since arbitrary sea condition can be realized in the numerical towing tank, the motion simulation technique can be developed to a more advanced technology of realizing all motions of a ship in actual sea conditions. It may be safe to say that we are working towards a virtual reality technique based on computational physics. All attitudes, forces and moments as well as all physical phenomena are realized in the computer with a simulation system composed of our application software.
In this article 20 years trend, from the elucidation of physical phenomena to the challenge for the realization of a kind of virtual reality technique, is described with the outputs I have derived with coworkers from 1979. 2. Free-surface shock wave In the last two years of five years experience as a designer at the Ship Initial Design Office of IHI I was in charge of the hull- form design of FUTURE-32 standard design bulk carrier (32,000DWT). The speed requirement was very severe and the fulfillment of the guaranteed trial speed written on the specification was supposed to be marginal. However, the trial result of the first ship was quite contrary. She was faster than the guaranteed speed by 0.6 knot on the ballast condition, which means the propulsion horsepower was 15% excessive. The 6 cylinder engine would have been installed instead of the 7 cylinder one. This experience gave me a motivation for the research of nonlinear ship waves. The cause of such discrepancy between experiment and trial results was supposed to consist in physical phenomenon of ship waves, partially because the discrepancy is much more prominent on the ballast condition with breaking wave phenomenon. Immediately after the delivery of two FUTURE-32 ships, I was invited to the University of Tokyo, where one of the prominent works there was wave pattern picture analysis to know the wave system. My research work for nonlinear ship wave was started with the observation of ship wave pattern pictures. The first insight was an analogy to the shock wave. The nonlinearity of the bow wave phenomenon was exaggerated by a model ship with an extreme bow shape as shown in Fig.1 t13~23~3~. Fig.1 Free-surface shock wave about a wall- sided model with a blunt bow, at En = 0.12, 0.16, 0.24, 0.26. Contrary to the theory of linear dispersive wave the wave pattern is dependent on Froude number and hull-form configuration. The variation of the apex angle of bow eve is rather systematic as shown in Fig.2 and the steepness of the wave slope obviously exceeds the limit of the linear wave system. Fig.2 Free-surface shock wave about series models of various half-entrance angle 5. <,10- <,15. <at En = 0.2774.
This nonlinear wave system appeared in the near-field of a ship advancing in deep water was named "Free-surface shock waves". The typical characteristics are 1) the formation of lines of discontinuity, 2) the steepness and unsteadiness of wave front, 3) the satisfaction of shock condition, 4) the systematical change of the angle of wave front, 5) the nondispersive propagation and 6) the dissipation of wave energy into momentum less far behind. These ware clarified through a series of physical experiment t4], see Fig.3. Linear dispersive wave is observed in the far-field but nonlinear waves are prominent. Both linear dispersive and nonlinear dissipative phenomena coexist. 3. Thin, long-protrudent bulb Analytical explanation of this nonlinear phenomenon was supposed to be made only by numerical analysis technique. Special computational method must be developed for such objectives, which needs remarkable time. However, sound understanding of the physical phenomenon by itself can be useful to the hull- form design. The free-surface shock wave (FSSW) is ruled by the Froude number based on draft (or equivalent draft). Therefore, when the Froude number based on draft exceeds 1.0, the FSSW becomes dominant with strongly interacting breaking wave. For the reduction of FSSW at high Froude numbers the entrance angle at the bow should be small. With the decrease of the bow apex angle the angle of shock decreases and consequently the wave resistance value is reduced. This design criterion is simple and straightforward as well as analogical to the design criterion for a supersonic body. This idea was first applied to the hull-form of "USUKI PIONEER" (26,000D WT bulk carrier) designed in 1983, Fig.4, which was equipped with two sails for the fuel saving purpose. Fig.3 Results of detailed wave profile measurement, wedge model of half- entrance angle 20° , Fd~draft- based)= 1. 1. Fig.4 "USUKI PIONEER" 26000 D WT bulk carrier the first ship designed with a thin, long-protrudent bulb in 1983. Thin, long-protrudent bulb could reduce wave resistance by 10 to 20% in comparison with a design with a conventional bulbous bow. In the 1980's this design was disseminated in the world. The thin, long-protrudent bulb is especially useful to ships of which service Froude number is smaller than 0.22. Nonlinear waves like FSSW also occur near the stern, especially that of high-speed, ships as shown in Fig.5, which is the case of Sutorechia-maru passenger boat (Lpp=lOOm, Fn=0,30~. A similar idea with the thin, long- protrudenbt bulb is also effective for the
reduction of nonlinear stern waves. Due to many reasons and restrictions the installation position of such stern bulb is near the waterline as shown in Fig.6. About 5% of energy saving is attained by this Stern-End-Bulb (SEB) and applied mostly to car carriers, truck ferries and container carriers for which the draft at the stern does not remarkably varyt5~6~. Fig.5 Wave picture from airplane, En ·a 0.30, nonlinearity of waves is observed especially around the stern. Fig.6 Stern-End-Bulb for a car ferry. 4. TUMMAC The research work to develop a finite- difference method for the computation of the flow with free-surface about a ship in steady course was started in 1979 and four years was needed until the TUMMAC-tW method was completed in 1983 t7] t83. The key algorithm is based on the MAC- method and the technique to implement the nonlinear free-surface condition is an extension from the SUMMAC-method. Most efforts were focused on the implementation of the body- boundary condition in the framework of the rectangular grid system. The free-slip condition could be successfully satisfied in the body- boundary cells of arbitrary configuration. The computed wave pattern agreed very well with the measured result, especially when wave breaking is scarcely observed. The comparison of wave pattern is made of a tanker hull on ballast condition in Fig.7, which indicates that some discrepancy is present and it is mostly attributable to the fact that the TUMMAC-tW method cannot cope with the breaking wave. However, the agreement was of the satisfactory level and was also applied to the diffraction wave problem of a wedge-shaped bow model and a tanker bow shape. The diffraction wave shows phase-dependent variation of the FSSW formation as shown in Fig.8~9~. Fig.7 Wave contours about a bow of a tanker model at ballast condition, measurement (above) and computed results by TUMMA C- tw code (below).
Fig.8 Computed wave view and unsteady free-surface shock wave in regular incident waves. Because the TUMMAC-twmethod has the advantages that the effort of grid generation is not required and that the accuracy is of satisfactory level, the code was distributed to major shipbuilding companies and employed as a tool of hull-form design. The wave resistance value obtained by integrating the surface pressure distribution dose not qualitatively coincide with the experimental value. But, hull- form designers soon noted that the accuracy was good in the relative relation of wave resistance value due to the modification of hull- form. It is often the case that the hull-form is designed by the succession of improvement. This code is extended to the version for catamaran and the advantage of small wave dissipation in the far-field is available for the wake-wash problem. An example of wave pattern of a catamaran is shown in Fig.9. 5 ~ computed by TM4DFM modeled "non, 2001 _ win ~e god da. generated by ~ new cell generator, eal200 1 ) 0.5 O 0 1 X A - mmatrical hull Neumann be - C on SSTH70NS1300 beam ins~.helr outelddar- 15 10 atFn-0646 wave height {Intents 0 0025) K-15 2 Fig.9 Wave contours of a catamaran fast ship computed by TUMMAC- IV. 5. Wave breaking simulation One of the most nonlinear fluid- dynamical problems is wave breaking. Since the CFD technology is based on the Navier-Stokes equation, it can cope with any problem of high nonlinearity, although the resolution of the micro-mechanics and very fast phenomenon necessitate top-end ability and capacity of computer. 2D wave breaking simulation was achieved with relatively small amount of efforts. The deforming free-surface configuration is represented by a succession of segments and then not only overturning motion but also impingement of the wave front on the forward free-surface below is well simulated as show in Fig.10 t111. One application example of the 2D wave breaking simulation is a problem of a circular cylinder placed horizontally in the vicinity of the free-surface. The complicated flow field of vortices strongly interacting with the free- surface is realistically simulated as show in Fig.11 L121
Fig.10 Simulation of 2D wave breaking in front of a steadily advancing rectangular body. Fig.11 Contours of vorticity about a circular cylinder set horizontally and advancing steadily near the free- surface. Quite different technique was necessary to achieve 3D wave breaking simulation. The water surface of three-dimensionally breaking motion cannot be treated by a succession of segment. Therefore, based on the idea of the VOF (volume-of-fluid) method a new technique was developed to implement free-surface condition on an interface of complicated topology. We called it density-function method, for which it was revealed later that a similar method called level-set method was being developed almost simultaneously. The first application of the TUMMAC method with the density-function technique on the free-surface was a flow about a vertically placed rectangular cylinder t13~. When Froude number is high the fluid flow is completely nonlinear with wave breaking, vortex shedding, spray and air-entrainment, Fig.12. The air flow above the water surface is simultaneously computed and some features of spray and air- entrainment are observed as in Figl3. When constant values of physical parameter such as density, Reynolds number and so on are varied, a variety of two-phase flow can be treated, such as an oil flow on water layer, a water flow on a liquefied sand bed, that is scoring problem, and so forth. Fig.12 Contours of wave height about a vertical cylinder (0.lm long set in a uniform stream 0.9 second after start of acceleration. Fig.13 Side view of Fig.12, on the longitudinal plane on the side of the pillar.
The density function technique was later employed in another code of finite-volume method for viscous and wave flow problem in the framework of boundary-fitted curvilinear coordinate system. One successful simulation result is show in Fig. 14 for a bow wave problem of a semi-planning boat t141. The 3D breaking waves are well realized showing the difference of wave formation depending on the Froude number. Some features of FSSW about a wedge Fig 16 model were also simulated with the same free- surface condition in t15] t16], and discontinuous and dissipative features are realized as shown in Figs.14 and 15. Fig.14 Overturning wave features simulated with the density-function method, Fn=1.0 (above) and 0.6 (below). Fig.15 Velocity vectors on the free-surface; g discontinuity of flow is present on the free-surface. Bernoulli constant value across the bow wave crest on the lines parallel to the centerline and the disturbed free-surface. Fig.17 Simulation of continuous casting of iron plate production process. Vorticity contours of the flow in the Sagami bay influenced by the Kurosio current computed by TUMMAC- v! code.
Since relatively smaller effort is required for the grid generation the TUMMAC method is applied to other problems, which includes complicated 3D configuration. The TUMMAC- tW technique can be applied to a variety of fluid flow problem which contain free-surface and a body of 3D configuration in any engineering field. One example is an application to the continuous casting process of steel plate manufacturing as shown in Fig.17. The complicated flow field made of liquid iron jet, 3D vorticities and free-surface wave are well simulated t101. The versatility of ()FD technique is thus demonstrated. A case of oceanographical flow near a Sagami bay is shown in Figl8. It is noted that the ocean current of Pacific ocean (Kuroshio) often causes 3D flow field at the interface with the bay. Similar two-phase flow technique with density function treatment was recently applied to a flow with bubbles L171. 108 micro bubbles are placed in the turbulent boundary layer and the mechanism to reduce frictional resistance was investigated by numerical simulation as show in Fig.19. This application indicates that such CFD technique with density function method in the framework of rectangular grid system has broad possibility of application to various environmental problems. Fig.19 Air bubbles in a turbulent boundary layer with vorticity contours of the How. 6. Separated flow The problem of turbulent flow is of global importance and the turbulent flow is often accompanied by flow separation. In the problems of moving body the emphasis is put on not only the fluid motion in the boundary layer but also that in the wake region. In the problems of naval hydrodynamics such viscous fluid flow region is also connected to the moving interface. Since a turbulent flow is intrinsically of time-dependent nature, the time-marching simulation method such as that of the MAC method must be employed. First method is developed in 1985 for the solution of viscous flow with free-surface in the framework of a finite-difference method with a curvilinear, boundary-fitted coordinate system [181. The viscous fluid motion at the stern of a ship is accompanied by separation phenomenon and a part of this structure is called wake, which is important for the propulsion efficiency. First output of such effort appeared in 1987 by a finite-difference method as shown in Fig.20. Fig.20 Vorticity contours about a Wigly hull advancing steadily, wave and viscous flow are simultaneously solved. It was soon noted that the finite- volume method is more suitable for this problem mostly from the viewpoint of robustness. A number of researchers attacked this problem with special efforts of choosing and tuning the turbulence model. It is the consensus we have that the grid number is still insufficient for such high-Reynolds number flow with free-
surface and some tuning or compromise is required in the choice of grid spacing and turbulence model. However, the simulation of viscous flow about a ship can be successfully used for practical design purpose. The mechanism of flow separation and the structure of a separated flow about a more blunt body were not well elucidated. The separated flows about a circular cylinder and that about a sphere were important target of scientific research. Because they are 3D and unsteady phenomena, experimental approach found substantial difficulties. From rather scientific viewpoint the separated flows about a sphere and a body of revolution with conical after body are studied by use of TUMMAC method and a finite-volume method, respectively. The computer-graphic pictures of the flow are shown in Figs.21, 22 and 23~19~20~. Fig.21 Isosurfaces of second derivative of pressure of a flow past a sphere. Fig.22 Isosurfaces of pressure at the after - end of a body of revolution with conical afterbody. Fig.23 Same as Fig.22. The most interesting thing is that the structure of a separated flow contains some common features independent of the object configuration and furthermore some part is also common to the structure of a turbulent boundary layer. The separated flow past a blunt body is composed of ring-shaped vortices with mostly by lateral vorticity component and horseshoe vortices with mostly longitudinal vorticity component. These two types of vortices are periodically shed. The presence of horseshoe-type vortex is quite common to all turbulent flow. One application of the simulation of separated flow is the case of automobile ~21~. With the understanding that small-scale vortices are not resolved due to the limited number of grid points numerical tests were conducted whether the method, which resolves only middle-scale vortices (5% of automobile length and greater), can discriminate automobile configuration of smaller drag coefficient. Since grossly 80% of aerodynamical drag is by the pressure distribution made by the time-dependent vortex shedding behind the body, the resolution of the separated from field is important, see Fig.24. Fig.24 Streamlines behind an automobile model with critical geometry.
The relative relation between the estimated drag by CFD and the measured drag for some series of automobile configuration was very good. Furthermore the lift was well estimated with same level of accuracy and the similar technique was successfully applied to the aerodynamical noise problem, in which it is elucidated that one of the causes of wind noise is breakdown of vortices ~22~. The time-marching technique of the MAC-type algorithm provides good resolution of the periodically- repeated viscous phenomenon of vortex shedding. It may be said that the accordance with physics is, in a sense, better for the separated flow of thick boundary layer and it is a little more difficult and sensitive to have good agreement in the case of thinner boundary layer accompanied with relatively gentle separation as the case of ship flow. 7. Moving technology The moving technology is to complete a moving system by implementing factors of substantial importance, they are, support and structure, power and speed, stability and control. Overall performance is given by synthesizing these factors. In the case of ships for sea transportation the weight is hydrodynamically supported, the power is transmitted to the hydrodynamical propulsion system and the total performance including stability, speed, ride- comfort, maneuverability, are mostly determined by hydrodynamical mechanism. Therefore the total system is hydrodynamically designed. Due to the progress of CFD both wave and viscous fluid flow are well elucidated and resultant forces can be utilized in the design process. The simulation of a flow about a ship advancing in steady straight course provides better hull form of smaller resistance and when a model for propulsion system is introduced the propulsion performance can be estimated. A new challenge was pursued in the latter part of 1990's towards the technique to cope with maneuvering and motion simulation. This can be achieved by combining the CFD technique with the solution of the equation of motion. The forces and moments are obtained by integrating the surface pressure and they are put into the equations of motion. For the introduction of the motion into the CFD computation two methods are supposed to be most successful. One is a method in which the grid system is fixed to the ship and the grid system moves giving accelerations due to motions to the Navier-Stokes equation as external forces. Another method is to use the moving grid system, in which the grids are deformed in accordance with the ship motion in six degrees of freedom. For more complicated system like catamaran or hydrofoils these method was still useless within the 20th century, when we worked hard to develop new ship systems operating at high-speed, one is super-slender twin-hull (SSTH) system (Fig.25) and the hydrofoil catamaran (HC) system in the 1990's(Fig26~. Fig.25 Fast catamaran ferry with super- slender twin-hull (SSTH).
Fig.26 Experiment ship "Exceller" for hydrofoil catamaran system. 8. Racing yacht design When the author was assigned to be a technical director and chief designer of Nippon Challenge for the 30th Americas Cup yacht race, I decided to fully utilize the CFD techniques for the development and design, Fig.27. The experience of sailing boat design is poor in Japan and this must be compensated by the information technology especially CFD. In parallel with the formation of a strong design team a new system of digital design by CAD and CFD is developed for the special sailing boat as shown in Fig.28. The design system for the international Americas Cup Class sailing boat was composed with the newest CAD and CFD softwares. Each element of the boat. that is. .. . . ... . .. . . . , , sails, keel with bulb and rudder are modeled by equations derived from experiments and CFD computations and the characteristics of these elements are put into the equation of motion for the PPS system and then CFD computation is made only for the hull. A new simulation technique called PPS(performance prediction simulation) is developed using the moving method by grid deformation. The sailing boat is set in an O-O type boundary-fitted coordinate system with six-degrees of freedom. The motion of the boat is treated by the grid deformation except that the steady advance motion is treated at the outer boundary and the hull surface is allowed to slip for the rolling motion as seen in Fig.29. Fig.27 America's cup yacht during the race i n 2000 | Initial Condition ~ Initial GRID Generation {~ ~ I ; ~ , . . it.'.': '< i.: ~ 'a ~'1 ~ , ~ ill ~~ Time LevelT-0 ~ . ~ i ~ i. Reed Restart Data Update Density Function and Generate Wave I 1 | SOLVE EQiJATIOIS OF MOTION I I I GRID Generation Move Water Surface? i, SOLVE NS-| EGUATION Fig.28 Block diagram for the simulation of ship motion.
Lo: - ~ ~ ; ~ ~ ~ Fig. 29 Grid system for a racing boat. 1 In spite of the small grid number of 45,000 this simulation could provide very useful results t243. For the design of racing yacht the 1% difference of resistance is meaningful. The ~ .. . .. .. . . accuracy ot the relative magn~tucte clue to small modification was 70% correct in case the experimentally verified difference of resistance was 1%. The difference of 0.1°trim angle due to the difference of hull form was rigorously estimated which is very important information because such small difference of attitude influences on the balance and maneuverability of the yacht. Since the equation of motion and the Nervier-Stokes equation are simultaneously solved in the time-marching procedure, not only the performance in a steady motion but also that of unsteady motion can be simulated by this PPS technique. Actually a simulation of course changing maneuver can be made with a simple control system as shown in Fi~.30. rat ~" . · ~ O the etI~ect~veness of our design method was not wholly verified bv the race in _ ... . .. ,, 2()uo, although Nippon was the second at the end of Round Robins and assumed to be one of the fastest boats. The experiment after the race shows that our design was superior to the boat of the defender Team New Zealand, as shown in Fig.31. On almost all close-hauled conditions our boat must have been faster when same sails, and boat maneuver were given. Two boats designed by our new system are now sailed by the UK team, Fig.32. Fig.30 Computer-graphic view of an IACC . . . . , class yacht In course keeping motion. Ct (Roenianoo codfic~en~h~ UPRIGHT condition tot 067 . 006 0~3 002 0~ o 5 7 t tl '3 V'~o' |°~4 o 4 - - ~ - h 't | 1N=4O ~ Fig.31 Resistance coefficient curve of the race boat of Nippon 2000 (blue) and of Team New Zealand. Fig.32 GBR52 Idaten~ex.JPN52) won the 150years Jubilee Regatta in August 2001.
9. Ships in waves Another method of simulating ship motion is to use a grid system fixed to the ship. The grid system moves giving accelerations by the motions to the Nervier-Stokes equation as external forces as shown in Fig.33. For the grid system an O-H type boundary-fitted grids are used so that the wave making is more conveniently made. The first simulation was made for a Series 60 hull advancing in regular heading waves, see Fig.34.The hull is set free to pitching and heaving motions. For the motion problem the robustness is of significant importance because the high pressure caused by motion may locally break the conservation law locally and subsequently break the solution. The regeneration of grids at each time step is carefully made and the density-function method Is emptoyect. ~ ne simulation shows that this technique can cope with the pitching motion very close to the slamming motion and the motion amplitude agrees well with the experimental results. ~ Ace. : ,~r~ Region of computation O-H grid system :: ~ ~ ~ : Fig.33 Ship-fixed coordinate system in a space-fixed coordinate system for the simulation of ship motion. Fig.34 Series 60 model in heaving and pitching motion in heading waves, wave contours (left) and pressuure distribution~right). For the rolling motion special technique is devised. The free-surface mane rotates instead ot the grids nor the hull, and clustering are made near the free-surface at each time step. Motion simulation with three degrees of freedom pitch, heave and roll is performed for an IACC sailing boat advancing in oblique(15013) regular heading waves. It is noted that the coupling motion of rolling and pitching gives quite different polar diagram due to the difference of wave length. However the IACC boat is not equipped with sails, keel and rudder, the motion is not realistic, see Fig.35. When the model equations and data of the sails and appendages are added the performance turns to be much more realistic due to the motion damping effect of the four lifting surfaces as seen in Fig.36. Considering that the motion experiment of sailing boat with sails is very difficult, such a "virtual" sailing simulation seems to be very useful. Nonlinear motion often occurs in the following wave conditions the extreme case is broaching, which may be caused by nonlinear (often sudden) change of hydrodynamical forces.
By generating waves by a numerical wavemaker of acceleration type set at the exit boundary, the motion simulation is made in the following wave condition with three degrees of freedom, that is, pitch, roll and heave as seen in Fig.37. When we can increase the degree of freedom of the motion, more realistic motion will be observed in the computer together with the interesting time history of a lot of physical values in the near future. Fig.35 PITCHING,HEAVING and ROLLING motion of AC boat in oblique incident wave. Fig. 36 PITCHING,HEAVING and ROLLING motion of AC boat with KEEL and RUDDER in oblique incident wave. Fig.37 PITCHING,HEAVING and ROLLING motion of Series 60 model in oblique following wave. 10. Design by virtual reality The performance of a ship has been classified into four, that is, resistance, propulsion, maneuver and motion in waves. However all these characteristics are going to be examined by CFD simulations. They are judged by the forces, moments, attitude and motions, which are evaluated by the numerical simulations. A designer, which uses a set of CFD simulation codes, can be a specialist of all of resistance, propulsion, maneuver and motion In waves. It may be safe to say that the information technique of CFD is a kind of techniques of system integration. The complicated design process for better performance of a ship has been simplified by use of CFD simulation, which provides all hydrodynamical properties simultaneously. This seems to be a great tread of design technology. The performance of the designed vehicle is examined through the simulation in the computer and a lot of information is fed back to the designer. A designer can have a great number of design experiences, learn a lot of things and become a talented expert one by this system assisted by the cheap computers in a relatively short period.
The ultimate goal of CFD research is to establish a system of virtual reality, which provide not only excellent image by computer- graphics, but also all performance characteristics of a designed object to a designer. A very simple procedure is as shown in Fig.38. Every possibility is examined and evaluated through the succession of CFD computation. When ship motion and added resistance which determines the fuel consumption in the life-cycle are of special importance, the system shown in Fig.39 will be useful, in which simulation of ship motion is employed to estimate the added resistance. A significant number of ships encounter serious accident and often suck into the deep sea on stormy condition. The performance of a ship in an extreme condition, such as the survival condition with 30m wave height and 70knots wind speed, is very difficult to predict but the structural design must be made with sound understanding of the characteristics of a ship in such condition. With the virtual reality technique by the computational physics this will be done in the future. The extreme sea condition can be realized in the computer and the designed ship is set on this sea surface and all characteristics are derived from the simulation. ~ r ~ : | WAVE SPECTRUM RESPONSE FUNCTIONS | RESISTANCE |: ~ . ~ . ·: I ., ADDED RESISTANCE I (PROBA8lLITY) l :: :: : :: ~ ~ '1,: - j FUEL CONSUMPTION in Me LIFSCYCLE .~ :~ ~~ . ~ ~~ . . .. ~ WIN PROBABILITY in the LONG-RUNNING RACE ~ i ~ i ~ ~ ~~ ~ ~ . ~ ~ ~ ~ 1 1 1 ~ 1 1 ~ 1 ~ 1 1 ~ 1 ~ 1 1 ~ ~ 1 ~ ~ 1 1 _ 11 1 11 1 _11 11 1 ~ Fig.38 Hull-form design procedure by Fig.39 Flow chart for life-cycle oriented hull- form design. Concluding remarks Seventy years from 1980 to 2000 will be said a special period in the long history of technology. The progress of computer technology influenced a lot onto a wide variety of the engineering field. It was also true in our field of naval architecture. We must still continue to fully digest the new technology and this may contrarily reveal that the technology of system engineering such as naval architecture is very valuable thing, for which we must keep making good progress. References [1l T. Inui, H. Kajitani and H. Miyata ~ Experimental investigations on the wave making in the near-field of ships, J. Kansai Soc. Nav. Archit. Jpn. 173 (June 1979), 95-107. [2] H. Miyata, T. Inui and H. Kajitani: Free full use of CAD and CFD simulation. surface shock waves around ships and their effects on ship resistance, J. Soc. Nav. Archit. Jpn. 147 (June 1980), 1-9. Nav. Archit. Ocean Engng. 18 (1980), 1-9.
t3] N. Kawamura, H. Kajitani, H. Miyata and Y. Tsuchiya: Experimental investigation on the resistance component due to Wee surface shock waves on series ship models, J. Kansai Soc. Nav. Archit. Jpn. 179 (Dec. 1980), 45-55. t10] N. Suzuki and H. Miyata: Practical use of a fluid flow simulation with solidification in the design of continuous casting processes, Industrial and Environmental Applications of Fluid Mechanics, ASME- FED Vol. 145 (1992), 97- 101, t4] H. Miyata and T. Inui: Nonlinear ship waves, Advances in Applied Mechanics 24, t11] H. Miyata: Finite-difference simulation Academic Press (1984), 215-288. of breaking waves, J. Computational Physics, 65-1 (July 1986), 179-214. t5] H. Miyata, Y. Tsuchiya and T. Inui: Resistance reduction by stern-end-bulb (first report), J. Soc. Nav. Archit. Jpn. 148 (Dec. 1980), 10-16. L6] H. Miyata, Y. Tsuchiya and T. Inui: Resistance reduction by stern-end-bulb (second report), J. Soc. Nav. Archit. Jpn. 149 t13] H. Miyata, M. Katsumata, Y. G. Lee and (June 1981), 1-10. H. Kajitani: A finite-difference simulation method for strongly interacting two-layer flow, J. Soc. Nav. Archit. Jpn. 163 (June t7] H. Miyata and S. Nishimura: Finite- 1988), 1-16. difference simulation of nonlinear ship waves, J. Fluid Mechanics 157 (Aug. 1985), 327- 357. [12] H. Akimoto, M. Sugihara and H. Miyata : Vortex motions and forces about a horizontal cylinder advancing Beneath the waves, J. Soc. Nav. Archit. Jpn. 170 (Dec. 1991), 253-263 t8] H. Miyata, S. Nishimura and A. Masuko: Finite difference simulation of nonlinear waves generated by ships of arbitrary three- dimensional configuration, J. Computational Physics 60-3 (Sept. 1985), 391- [15] A. Kanai and H. Miyata: Elucidation of 436. the structure of free surface shock waves about a wedge model by finite-difference method, J. Soc.Nav. Archit. Jpn. Vol.177, t14] H. Orihara and H. Miyata Numerical Simulation Method for Flows About a Semi- Planing Boat with a Transom Stern J. Ship Research, Vol.44, No.3, Sept. 2000, pp. 170- 185 t9] H. Miyata, M. Kanai, N. Yoshiyasu and Y. 147-159. (May 1995) Furuno: Diffraction waves about an advancing wedge model in deep water, J. Ship Research 34-2 (June 1990), 105- 122. [16] A. Kanai and H. Miyata Numerical analysis of the structure of free-surface shock wave about a wedge model, J. Ship Research,
Vol.40, No.4, Decl996, 278-287 t17] A. Kanai and H. Miyata,Direct numerical simulation of wall turbulent flows with microbubbles International Journal for Numerical Methods in Fluids Int. J. Numer. Meth. Fluids 2001; 35: 593-615 [18] H. Miyata, T. Sato and N. Baba: Difference solution of a viscous flow with free- surface wave about an advancing ship, J. Computational Physics, 72-2 (Oct. 1987), 393- 421. t1 9] H. Miyata and Y. Yamada: A finite difference method for 3D flows about bodies of complex geometry in rectangular co- ordinate systems, Int. J. Numerical Methods in Fluids 14 (1992), 1261-1287. [20] H. Miyata, M. Zhu and O. Watanabe Numerical study on a viscous flow with free- surface waves about a ship in steady straight course by a finite-volume method,J. Ship Research 36-4 (Dec. 1992), 332-345. [21] K. Matsunaga, H. Miyata, K. Aoki and M. Zhu: Finite-difference simulation of 3D vertical flows past road vehicles, 1992 SAE Inter. Congress and Exposition, Detroit,Vehicle Aerodynamics, SAE SP-908 (Feb. 1992), 65-84. ., [22] Y. Hanaoka, M. Zhu and H. Miyata: Numerical prediction of wind noise around the front pillar of a car-like body, 1993 SAE, 7th Inter. Pacific Conf. and Exposition on AutomotiveEngg., Phoenics, SAE 931895 (Nov. 1993), 1-11. [23] Y. Sato, H. Miyata and T. Sato :CFD simulation of 3-dimensional motion of a ship in waves: application to an advancing ship in regular heading waves, J. Mar. Sci. Technol. (1999) 4:108-116 t24] H. Akimoto, Development and application of CFD simulation technique for ships in 3D motion, Ph.D thesis, University of Tokyo 1996(in Japanese). L25] H. miyata: Time-marching CFD simulation for moving boundary Problems, 21St Symposium on Naval Hydrodynamics 1996, Tronheim
DISCUSSION Ki-Han Kim Office of Naval Research, USA How did you handle the free-surface "ridding in your computations; surface tracking or surface capturing technique? AUTHOR'S REPLY In most of the 3D cases the capturing technique using density function method is employed. This is similar to the level-set method. DISCUSSION Arthur M. Reed Naval Surface Warfare Center, Carderock Division, USA Prof Miyata used the term "stability" several times. However, it is not what is meant by stability. Did the references refer to ship stability, or was the reference referring to computational stability? AUTHOR'S REPLY For computer simulation stability of the solution is most important rather than computational economy due to the advance of computer technology. We are now very close to capsizing simulation. Therefore the ship stability problem is one of the important targets of CFD simulation.
