RECENT ADVANCES IN THE HYDRODYNAMIC DESIGN OF FAST MONOHULLS

Transcription

1 RECENT ADVANCES IN THE HYDRODYNAMIC DESIGN OF FAST MONOHULLS by Gregory Grigoropoulos, Assoc. Professor NTUA ABSTRACT During the last decade, significant effort has been spent to improve the performance of fast monohulls. Thus, in spite of the development of various alternative multi-hull and dynamically supported vehicles, they continue to dominate the market both for commercial and recreational vehicles, as well as for naval combatants. In this work, recent advances in the hydrodynamic design of fast monohulls are presented and discussed. Their merit is compared to the current state of development of the other advanced hull forms. 1 Introduction Fast monohulls still dominate the market of (car-) passenger ships and naval combatants, although fast multi-hulls, hydrofoils and air-supported vehicles possess a remarkable portion, especially at the higher speed range. An overview of the speedlength relationship of commercial ships is provided in Fig.1, from Levander, (2001), enriched by Grigoropoulos and Loukakis (2002). Fast monohulls operate in the upper region of this figure, at speeds corresponding to Froude numbers in excess of This is also the case of modern naval ships, which tend to reduce their size at the same speed, since modern efficient weapons can be accommodated in smaller platforms. A common characteristic of this type of hull forms is the large immersed transom stern. 50 Fn=1.00 Fast Cat Fn= Fast Mono Fn=0.35 Speed [knots] Displacement Type Ferries Fn= Length WL [m] Fig.1: Fast designs of the 90 s (Levander, 2001, Grigoropoulos and Loukakis, 2002). 1

2 Systematic resistance series constitute still a major source of ready to use information for the designer. Although nowadays the presentation of systematic results for a series of hull forms is not as customary as it was three or four decades ago, four systematic series of high-speed monohulls have been developed in the last decade. The series, which have advantageous resistance performance in the semi-displacement or pre-planing speed regime (Froude Number range = ), are: VWS D-Serie, Berlin (Kracht, 1996) SKLAD series, Zagreb (Gamulin, 1996) AMECRC systematic series (Bojovic, 1997) NTUA series of double-chine hull forms (Grigoropoulos & Loukakis, 1999) The latter two of the series have also improved seakeeping characteristics. All four series are presented in Section 2. The above series extend the scope of the following older systematic series of fast monohulls: KTH/NSMB round-bilge & hard-chine series (Nordström 1936, Clement 1964). EMB Series 50 (Davidson and Saurez, 1948) HSVA C' Series (Kracht and Grim, 1960) Series 62 single chine (Clement and Blount, 1963) Series 63 (Beys, 1963) Series 64 (Yeh, 1965) SSPA Methodical Series (Lindgren and Williams, 1968) Series 65 (Holling and Hubble, 1974) NPL Series of round-bilge hulls (Bailey, 1976) Japanese Series of planing craft (Nagai, Tanaka and Yoshida, 1976). BK Series of semi-planing hull forms (Yegorov, et al, 1978) MBK Series of semi-planing hull forms (Yegorov, et al, 1978) TU Delft Series of Deep-V hulls based on Series 62 (Keuning & Gerritsma, 1982) NSMB Series of round-bilge, semi-displacement hulls (Oossanen & Pieffers, 1984) NRC Series of naval ships (Schmitke et al, 1979 and Murdey & Simoes Re, 1985). US Naval Academy Series (Compton, 1986) VTT Series (Lahtiharju et al, 1991) Series based on testing of models smaller than 1 m in length have been excluded from the above list. Based on one or more of the aforementioned series, or on an extended data base of model tests, a series of regression methods are available to the designer of a fast monohull to estimate its resistance during the preliminary design phase, the more important of which are: Van Oortmerssen method (Van Oortmerssen, 1971) Mercier and Savitsky method (Savitsky and Brown, 1976) Series 62/65 regression method (Hadler et al, 1978) Tang method (Ping-zhong et al, 1980) Jin s regression method for round bilge hulls (Jin et al, 1980) Holtrop method (Holtrop, 1984) 2

3 Compton method (Compton, 1986) Japanese Series of planing craft (Nagai and Yoshida, 1993). Radojcic s regression equations for various systematic series (Series 62, TU Delft Series, Series 65, NPL Series) including the recent SKLAD Series (Radojcic et al, 1999) and NTUA Series (Radojcic et al, 2001). The derived simple formulae permit an easy selection of optimal design parameters within the permissible range of the series. However, these readily applicable relations should be used only within the restricted ranges of the design parameters in the series of the database. On the other hand, Fung (1991) developed a speed-independent regression model on the basis of a large database of 526 test conditions of numerous ship types. Following an extensive statistical analysis, he claims his method to be useful for the optimisation of the hull form parameters during the early stages of ship design, as it is statistically reliable within a wide range of speeds and hull form parameters. More experimental and theoretical background, as well as physical insight, form the basis of semi-empirical methods, which address the resistance of prismatic hull forms. The assumption of a prismatic hull form, which greatly simplifies the modelling of planing hulls, is reasonable in the case of pure planing hulls for the underwater part of the hull. However, the assumption breaks down at lower speeds when the fore body comes into contact with the water. Such methods are: Shuford method (Shuford, 1958), modified by Brown and Klosinski (1980) Savitsky method (Savitsky, 1964), including Blount and Fox correction factor for pre-planing regime (Blount and Fox, 1976) Lyubomirov method (Yegorov et al, 1978) Almeter (1993) describes and compares the older methods. He comments, however, on the validity of semi-empirical equations that: Data can be found to justify almost any equation. Furthermore, Almeter (1999) proposes a simplified method to estimate the resistance of a planing craft on the basis of a parent one with different proportions and loadings. The method is valid for craft with similar body plans. Its major advantage is that it is independent of hull proportions, which allows the prediction to be based on a parent craft with different length to beam and beam to draft ratios. According to this method, R/W ratio is plotted versus log10 An, where An is the W so-called Almeter number A n, for various Clement numbers LCG BmV Cn, where LCG is measured from the transom. Both, parameters of 9 / 4 3/ 4 LCG Bm the method, An and Cn are non-dimensional. Almeter claims that, for log(an) > -1.0, the vessel sails in displacement mode and its RT/W ratio is highly affected by its displacement expressed in terms of Cn number, while at planing speeds (log(an) < -1.0) the RT/W ratio is predominantly a function of An. The method is implemented in the case of NTUA series, the only hard chine series described in this work. 3

4 Regression analysis has also been applied in conjunction with numerical methods to improve the reliability of their prediction. Hanhirova et al (1995) used Michell s Integral as treated by Tuck (1987) and applied regression analysis to derive two 13- parameter mathematical models for the difference ΔCW = CR CW, for low and high L/B ratios, respectively. It should be noted that wave resistance derived by Michell s Integral is not affected whether the ship/model is free to trim and sink or fixed in position, being proportional to its breadth squared. Doctors and Day (1997) proposed a similar hybrid method. They modelled the hollow cavity in the water flow observed behind the transom stern, taking into account that its length varies with the speed of the vessel. Then, they used the traditional thinship or perturbation analysis of Michell (1898) and Lunde (1951). Slenderness of the hull is increased by means of the transom stern hollow. They also included two form factors in their analysis, one accounting for an empirical frictional correction and the other for wave resistance, to be considered as accurate in the limit of a very thin hull form. The authors determine these factors by matching numerical results to experimental ones. They suggested that their hybrid method should be used to reduce model tests required to investigate a realistic range of trims and displacements. Both hybrid methods apply to multihulls as well. The arsenal of the naval architect in the prediction of the hydrodynamic performance of fast monohulls encompasses potential flow CFD methods, such as: Tulin s linear 3-D potential flow method for slender hulls at high Froude numbers suitable for fast displacement ships (Wang et al, 1995), the combined potential-flow, boundary-layer viscous-flow zonal approach used by SHIPFLOW panel code (1997), although the analytical results of CW underpredict in a consistent way the experimental values for CR (Sahoo et al, 1999), 3-D vortex lattice methods, where in accordance with Wagner s theory (1932), the hull is considered to be the underside of a wing. Lai and Troesch (1995) applied a 3-D boundary condition and assumed a jet region of zero pressure on the leading and side edges of the wing, added mass methods based on Wagner s theory, as the one proposed by Payne (1988) and recently updated by Singleton (2004), and Zarnick s low aspect ratio strip method, as extended by Akers (1999) to predict dynamic panel pressures. In addition, there are some recent attempts to use free-surface viscous codes (Reynolds Averaged Navier-Stokes Equation solvers) as reported by Capponetto (2000). These codes give the designer a better understanding of the planing phenomena, especially if the actual free surface is used instead of the simplistic symmetry plane in the boundary conditions. Although such CFD methods provide promising results (Pemberton et al, 2001), they require experience to obtain reliable results in the prediction of the performance of high-speed craft. There are used, however, in comparative studies when modifications on hull forms of existing vessels or members of systematic series are evaluated. This procedure, which is quite common in the hydrodynamic design of advanced monohulls, encompasses the local and global modification of the hull form and the 4

5 fitting of a suite of appendages to improve the efficiency of their calm water performance. The Enlarged Ship Concept (Keuning and Pinkster, 1995) was the only method dealing with the global hull form parameters in the conceptual design and aiming at improving both the resistance, seakeeping and manoeuvring characteristics of the vessel. A review of these methods and their effect on the behaviour of the vessels is presented in Section 3. On the way to the development of NTUA series Grigoropoulos and Loukakis (1995) compared experimentally the parent hull form of the series against four other equivalent hull forms, i.e. with the same length, beam, displacement and trim (Fig.2). These results show that in the pre-planing region there can be differences in resistance due to the shape of the underwater part of equivalent hull forms, but these differences are of moderate significance only. On the same Figure, the prediction of resistance based on the NPL method is shown. The round bottom hull forms have, as expected, lower resistance in the lower speed range and higher at higher speeds Series 62 with spray rails Delft Series with spray rails Double-cline based on Series 62 with spray rails NTUA Series without spray rails Rounded variant of NTUA Series with spray rails NPL Series Prediction (Bailey, 1976) R Μ / Δ Fig.2: Experimental calm water resistance over displacement ratio (RM/Δ) for five equivalent high-speed hull forms and prediction by NPL Series. The best case with respect to the fitting of spray rails is plotted. During the last decade, some novel ideas were proposed with excellent performance in a prescribed range of displacements and speeds. These novel types of monohulls are described in a Section 4 of the paper. Finally, in an attempt to evaluate the relative merit of fast monohulls compared to other types of advanced vehicles, their advantages and disadvantages are presented and discussed in Chapter 5 of the paper. Fn 5

6 2 Systematic Series of High-Speed Light-Displacement Monohulls VWS D-Series The series originates from a twin-screw round bilge hull form, and refers to relatively broad and short ships. Kracht (1992, 1996) reported on the resistance, wake and propulsion tests carried out with the 13 models of the series. All models had a common LBP = 6.00 m. For each CP value three models with common 10 3 C = 3.0 and varying B/T have been constructed, while a forth model had B/T =3.75, as the parent one, and a 10 3 C = 3.5. Especially for Cp = a fifth model with B/T =3.75 and 10 3 C = 4.0 was built. The body plan of the parent model is shown in Fig.3. Its form parameters are given in Table 1, while the variation of CP and C coefficients and B/T ratio within the series is shown in Table 2. WL 13 WL 12 C L WL WL WL 9 WL 8 WL 7 WL 6 WL 5 WL 4 WL 3 WL 2 WL 1 basis BL A.P. St. 2 St. 20 F.P. Fig.3: Body plan of the parent model of D-Series (model 2521). Table 1: Form Parameters of parent hull form of D-Series Parameter Value Prismatic Coefficient CP = /(AM LBP) B/T ratio (amidships) 3.75 Slenderness coefficient 10 3 C = 10 3 /LBP Sectional Coefficient CX at maximum Section (Section 9) Longitudinal Centre of Buoyancy LCB/LBP (fwd of transom) Table 2: Form Parameters varied to generate D-Series CP B/T C Tests have been carried out at three displacements and speeds corresponding to Fn = The effect of appendages (bossing, V-bracket and rudder) was investigated at the intermediate displacement. At the same displacement, wake and 6

7 self-propulsion tests have been carried out. Finally, the effect of trim by bow and by stern has been investigated for the intermediate displacement of the last three models of the series. The results presented are: model test raw data for the naked hull and the hull with appendage resistance, open water propeller characteristics, self propulsion and wake tests as well as residual resistance coefficients CR, running trim and dynamic CG rise, velocity field at the propeller disk, propulsive performance coefficients, wake fraction w, thrust deduction factor t and relative rotative efficiency ηr. The resistance curve in terms of the non-dimensional coefficient CTL = RTm / (Δ Fn 2 ) for the naked parent hull of the series is given in Fig.5, along with the respective of SKLAD Series. SKLAD series SKLAD series were a designer oriented series developed at Brodarski Institute (Zagreb, Croatia) during the 70s and first published in 1996 (Gamulin). The body plan of the parent model of the series is shown in Fig.4 and its characteristics are presented L in Table C 3. B =0.450 WL B WL =6 =4 WL B WL T AP FP Fig.4: Body plan and bow and stern profiles of the parent model of SKLAD series The series consists of 27 models with all combinations of the form parameters shown in Table 4, split in three groups according to the CB values. Each group has constant CP, CX, CWP and position of LCB ( LBP, LBP and LBP for CB = 0.35, 045 and 0.55, respectively). The models were derived from the parent and the basic forms for CB = 0.35 and 0.55, so that the model displacement was always constant (M = m 3 ), while LWL/BWL, BWL/T and CB were constant in each group. They were tested at level keel, for speeds corresponding to volumetric Froude numbers Fn = and displacements in the range of the non-dimensional coefficient M = LWL/ 1/3 = 4.50 to Table 3: Form Parameters of the parent hull form of SKLAD series Parameter Value Length between perpendiculars LBP = LWL (m) Prismatic Coefficient CP = /(AM LBP) Sectional Coefficient CX at maximum Section LCB/LBP (positive forward of midship section) Half-angle of entrance ie 12.0 o 7

8 Table 4: Form Parameters varied to generate SKLAD series LWL/BWL BWL/T CB The results are presented in the form of constant value curves for residual resistance coefficients CR, running trim, dynamic rise of the centre of gravity (CG) and running wetted surface, on CB, LWL/BWL axes. Graphs are provided for each BWL/T and testing speed, corresponding to Fn = (step 0.25) and 3.00 The last two results were non-dimensionalized by 1/3 and 2/3, respectively, while CR was derived on the basis of the running wetted surface. Since 22 ships have been constructed using the hull form of the series, a reliable relation for the ship to model correlation allowance DCF, as a function of ship Reynolds number ReS is provided: DCF = ( ln ReS) C TL = R Tm /(Δ Fn 2 ) VWS Series Parent Model D1 SKLAD Series Parent Model Prediction for SKLAD Parent Model by NPL Series Prediction for VWS Parent Model via NPL Series Fig.5: Experimental results for the parent models of the VWS and SKLAD Series and predictions for these models using NPL Series. Fn An allowance of 3.5% per screw (including shaft and bracket and rudder) is superimposed on the naked hull total resistance. The effect of LCB shift has been investigated only for the members of the parent model group (2 nd group). Finally, values for the propulsive coefficients w, t and ηr are provided. In Fig.5 the experimental results for the parent models of VWS and SKLAD Series are compared with the respective predictions using NPL Series. Since the size and the displacement of the tested models for these series are different, the nondimensional resistance coefficient CTL = RTm (Δ Fn 2 ) has been used for the comparison. As it can be concluded from this figure, there are significant differences in 8

9 the prediction of the resistance of the two parent models in the pre-planing region, where NPL predicts intermediate and quite similar for both model values. Thus, the development of systematic series of modern hull forms is quite significant for the profession. AMECRC systematic series These series were developed at the Australian Maritime Cooperative Research Centre (AMECRC). The series consists of 14 semi-displacement round-bilge, transomstern models with straight entrance waterlines and buttock lines, based on the NSMB Series (Oossanen and Pieffers, 1984), with which they share the parent hull. The hull forms of the series can be used as workboats, launches or corvettes. Following the policy of NSMB series, AMRCRC publishes the description of the series and some regression formulae correlating residual resistance with the varied hull form parameters (Bojovic, 1997), avoiding the presentation of the complete resistance results. Only for the parent model, selected on the basis of its superior seakeeping qualities, MARIN published the hull geometry (Fig.6) and the test results BL Fig.6: Body plan of the parent model of HSDHF and AMECRC series Table 5: Form Parameters common to all models of AMECRC series Form Parameter Value Prismatic Coefficient CP = /(AM LBP) Waterplane area coefficient CWP Transom Area / Maximum Sectional Area AT/AX Transom Beam / Maximum Beam BT/BX LCB/LBP (forward of transom) All members of the series share the form parameters of Table 5. Their common waterline length LWL =1.60 m was quite small, due to the size of AMECRC towing tank. The parameters of the parent model of both series and the range of their variation are given in Table 6. It is obvious that the form parameters of the parent model do not have intermediate values within the range of their variation within AMECRC series. In 9

10 addition, to calm water resistance tests performed for speeds 0.4 to 4.0 m/sec (respective Fn = 0.10 to 1.00), seakeeping tests in regular head waves for speeds corresponding to Fn = 0.285, and have been carried out. Table 6: The parent model and the range of parameters in AMECRC and NSMB series Parameter Parent Model AMECRC Series NSMB series LWL/BWL BWL/T CB Bojovic (1997) provides multi-parametric plots of the non-dimensional parameters CR and RR/W (W = weight) using iso-lwl/bwl, BWL/T and CB curves per speed, expressed in terms of Fn and Fn, respectively. Furthermore, a multiple regression analysis and two non-linear estimation techniques are applied on the results. The NTUA Series of double-chine hull forms The NTUA series is based on a proposal of Savitsky et al (1972) for a novel High-Speed Planing Hull for Rough Water with wide transom, warped planing surface, double chine and very fine bowlines. Blount and Hankley (1976) verified the improved performance characteristics of the hull form in rough seas by full scale testing of two high-speed craft, one with the novel hull form and the other with a traditional hard chine. Actually, the CG acceleration data of that craft could be compared favorably with traditional hard-chine craft at twice the sea intensity! During the 90s, various versions of this hull form became popular, especially in the European short-sea shipping. Exploiting the advance of structural technology, large ships around 100 m in length, with quite light (around 1000 mt) displacement were constructed. Propelled by modern engines (Diesels with very high power density and Gas Turbines) and using extensively water jets, they usually operate at speeds corresponding to Fn greater than 0.40 and mostly around During the same period, a systematic Series of double-chine, wide transom hull form with warped planing surface has been developed at the Laboratory for Ship & Marine Hydrodynamics (LSMH) of the National Technical University of Athens (NTUA). The Series are appropriate for the preliminary design of fast monohull ships operating at high but pre-planing speeds (respective Fn =0.55 to ). Furthermore, the parent model of the series has been tested at DERA (now QINETIQ) premises up to a speed corresponding to Fn = 1.80, demonstrating a very satisfactory performance in the planing regime, too. The NTUA double-chine series ended up consisting of five (small) models with LOA/BM = 4.00, 4.75, 5.50, 6.25 & 7.00 and five larger versions of the previous models to accommodate the very light displacements. Each small model (and/or its larger version) was tested at six displacements corresponding to a volume of displacement coefficient CDL= /(0.1. LWL) 3 = 1.00, 1.61, 22.3, 3.00, 3.61 and 4.23, to cover the needs of both large and small fast ships. The lines plan of parent form of the Series with L/B = 5.50 is shown in Fig.7. The deadrise angle ranges from 10 o in the stern region to around 20 o amidships rising to 70 o in the bow region. 10

11 Fig.7: Lines plan of the parent hull form of the NTUA systematic series (the body plan has been scaled up by a factor of three). R T / Δ L/B = 5.5, C DL = 1.0 R T / Δ DYNAMIC TRIM DYNAMIC C.G.-RISE/L WL R / (Δ*Fn 2 ) R/(Δ*Fn 2 ), DYNAMIC TRIM [deg], C.G.-RISE/L WL [%] Fn Fig.8: Resistance, C.G. rise and dynamic trim of the parent hull form in the preplaning region. Grigoropoulos and Loukakis (2002) presented the resistance characteristics for the series (residuary resistance coefficient CR, running trim and dynamic CG rise). CR values were estimated on the basis of static LWL and wetted surface, as it was found to be sufficient for the series. In addition, existing full-scale data and Laboratory seakeeping experiments (see e.g. Grigoropoulos and Loukakis, 1995) in head waves indicate excellent rough water performance characteristics for the Series. Thus, it was 11

12 decided to slowly construct an extensive series both for resistance and seakeeping, bearing in mind the absence of systematic seakeeping results for other series and the lack of series suitable for large fast ships operating at very light displacements. As it is obvious in Fig.8, the series have very good resistance trend as speed increases, together with negligible squat and very small dynamic trim angle in the region of Fn = Furthermore, the resistance characteristics of the Series compare favorably with other hull forms appropriate for the pre-planing Fn range, while the proposed hull form possess also a wide transom, present in all modern designs of fast monohull ships. Prediction methods and the experimental results of NTUA series On the basis of the experimental results for NTUA series, some of the methods reviewed in the paper are evaluated. Thus, in order to investigate the applicability of Almeter s method, already described, the experimental results of hull forms of NTUA series with different L/B ratios but similar Cn numbers are plotted in Fig.9 in the form of R/W vs. log10an. Although the hull forms differ, their R/W curves are quite similar in accordance with Almeter s suggestion that Cn number is a critical test parameter Cn = 0.068, CDL = 1.61 L/B = 7.00 Cn = 0.070, CDL = 1.61 L/B = 5.50 Cn = 0.070, CDL = 1.61 L/B = 6.25 Cn = 0.070, CDL = 1.61 L/B = 6.25 Cn = 0.073, CDL = 2.23 L/B = 4.00 Cn = 0.077, CDL = 2.23 L/B = 4.75 R/W Log 10 A n Fig.9: Plot of R/W vs. log10an according to Almeter s method (1999) for the member hull forms of NTUA series at similar Cn numbers. Furthermore, as it can be deduced from Fig.10, the R/W ratio for hull forms with varying Cn number rises quite smoothly in the planing regime (log10an < 1.50), while Cn affects it significantly at lower (displacement and semi-displacement) speeds. However, the hump at An = 1.00, described by Almeter (1999) is not present. It should be noted, however, that, since NTUA are suitable for large and rather light monohulls, Cn range is , while the respective ranges of Series 62 (Clement and Blount, 1962) and TU Delft Series of Deep-V hulls based on Series 62 (Keuning and 12

13 Gerritsma, 1982) are and , where the larger values correspond to small craft situations R/W Cn = 0.042, CDL = 1.00, Cn = 0.070, CDL = 1.61 Cn = 0.094, CDL = 2.23 Cn = 0.126, CDL = 3.00 Cn = 0.149, CDL = 3.62 Cn = 0.169, CDL = Log 10 A n Fig.10: Plot of R/W vs. log10an according to Almeter s method (1999) for the parent hull form of NTUA Series (L/B = 5.50) at different Cn numbers. 450 C TLm = R Tm /(Δ Fn 2 )x NTUA Experimental Results Regression Method of Radojcic et al (2001) VTT Hard Chine Hulls (Lahtiharju, 1991) Savitsky Method (1964) BOAT-3D Prediction (Payne, 1988) NPL Series Prediction (Bailey, 1976) VTT Round Bilge Hulls (Lahtiharju, 1991) Fig.11: Non-dimensional resistance coefficient CTLm for the parent model of NTUA Series at CDL = 1.61 predicted by Savitsky, BOAT-3D, VTT for hard-chine and round-bilge hulls, NPL and Radojcic methods and derived from model tests. Fn 13

14 In another test of widely available prediction methods, the experimental results of the NTUA series have been compared with the predictions of six such methods. In Fig.11, Savitsky method, BOAT-3D code based on the added mass method of Payne (1988), NPL series and three regression methods are compared with the experimental results of the parent model of NTUA series for a light displacement corresponding to CDL = As it was expected, regression method of Radojcic et al (2001), based on NTUA series provides very accurate predictions. On the other hand, the semi-empirical method of Savitsky and the added mass method of BOAT-3D, in close agreement among themselves slightly overestimate the experimental results. Both of these methods are suitable for higher (planing) speeds. Among the other three methods, which are based on model tests of different than the double-chine hull forms, NPL series, based on interpolation of model tests of round bottom hull forms, provides consistent predictions for the higher speeds, while in the lower speed region (Fn < 0.50) it demonstrates the better performance of the round bottom hull forms, as it was also the case for the model tests of Fig. 2. On the other hand, the VTT methods, based on the regression analysis of their own results, as well as on results of older series, predict consistently lower resistance values throughout the pre-planing speed region, but they also predict lower resistance values for hard chine hull forms than for round bilge ones, which is not the usual case. Therefore, the rather obvious conclusion that one should be very careful in assessing the expected resistance of his design, when using prediction methods based on different type hull forms, can be drawn. 3 Modifications to improve the Performance of Fast Monohulls Depending on their speed, high-speed monohulls experience dynamic lift to a certain degree. At the lower speed range, corresponding roughly to Fn between 0.4 and 0.9, their underwater shape is rounded with straight entrance waterlines and buttock lines and a transom stern. At higher speeds, they are designed with one or more hard chines and straight sections to take advantage of the extra dynamic lift available at these speeds. In this section some methods for improving their performance, as the fitting of stern wedges or flaps and of spray rails are reviewed and discussed. As Savitsky (1964) has demonstrated, the performance of prismatic planing hulls in calm water is dominated by the displacement and its longitudinal distribution expressed by LCG, the breadth over chine and the deadrise angle. In the case of nonprismatic hulls with varying deadrise, the respective longitudinal distribution of breadth over chine and deadrise angles should be taken into account. For any given combination of these design parameters the hull is planing at any speed with a specific dynamic trim. Thus, the problem of optimizing the design of a planing hull form is reduced to finding out the optimum combination of these parameters, resulting in reduced horsepower requirements. The achieved dynamic trim angles in this case, are closely associated with the specific hull form, so that it could be said that, instead of seeking for reduced resistance, the designer aims at the specification of the associated dynamic trim over speed curve. Since the displacement and the LCG are usually predetermined by the owner s requirements, the main task of the designer is to determine an optimized combination 14

15 of longitudinal distribution of breadths and deadrise angles, resulting in reduced calm water resistance. When this objective cannot be achieved, stern wedges or adjustable trimming flaps should be used to reduce the running trim by stern of a planing hull. The stern wedges are simple constructions and they can produce high lift forces, resulting in an improved hydrodynamic performance of the vessel in a limited speed range. On the contrary, the trimming flaps permit the fine tuning of the dynamic trim to its optimum value, corresponding to the minimum resistance for a given speed. However, their constructional details do not allow for very heavy loading. Grigoropoulos and Loukakis (1995, 1996) fitted spray rails and stern wedges, respectively, on the parent model of the NTUA series without clear (positive or negative) effect. They tested the model at speeds corresponding to Fn up to 1.10 using stern wedges with lengths 2, 5, 7.5 and 10% of LWL. The optimum wedge length was found to fall in the range of 2% to 5% of LWL. At each wedge length, the model was fitted with different span-beam ratio wedges, concluding that the full span wedges are the most efficient. One year later, Grigoropoulos (1997) combined the effect of stern wedges with spray rails on the same hull form without achieving significant improvement in the performance of the model. It seems that the model without the wedges runs at a nearly optimum trim. Grigoropoulos and Loukakis (2001) drew similar results when they investigated the effect of static trim on the performance of the parent model of NTUA series. OPV-1 Knuckle OPV-2 Knuckle Spray rail 2 Spray rail 1 Knuckle - Spray rail 1 OPV-3 Spray rail 1 Spray rail 2 Fig.12: Body plans of the tested models (Grigoropoulos and Damala, 1999). On the contrary, the investigations of Grigoropoulos and Damala (1999) were more successful. They investigated experimentally the combined effects of spray rails and stern wedges on the calm water performance of three high-speed round-bottom hull forms. The calm water performance of the three models of offshore patrol vessels (OPV), depicted in Fig.12, has been optimized via stern wedges and one or two spray 15

16 rail series in the bow region (Müller-Graf, 1991). The three models have been tested at a series of speeds up to Fn = 1.00, 0.75 and 0.60, respectively. The extensive investigation aimed at determining the particular effects of the aforementioned appendages on the resistance of these models, in conjunction with the modification of their displacement, trim angle and vertical location of the centre of gravity. The experimental results were thoroughly analysed to investigate the necessity of fitting one or two series of spray rails in combination with stern wedges, and to specify the most efficient design parameters, such as the form and location of these appendages. The following major conclusions were drawn: The efficiency of spray rails, when fitted according to the guidelines provided in the literature (see e.g. Lindrgren and Williams, 1968) is restricted at Fn > The lift generation seems to dominate the influence of the wedges on the resistance of high-speed vessels. On the other hand, wedges reduce the trim by stern at speed, thus partly counteracting the effect of lift. Finally, stern wedges improve the propulsive performance of fast monohulls, leading to even higher attainable maximum speeds. Karafiath and Fisher (1987) focused their interest in the effect of stern wedges on the propulsive efficiency of large naval ships. Since analytical results of the hydrodynamics of the wedge effect on semi-displacement hull form were not available, they combined experimental data with analytical results derived using a potential flow code (Dawson, 1979) to conclude that a properly designed stern wedge may lead to a 6% reduction in the delivered power. However, they claimed that the modification of the flow field around the after body of the ship by the wedge, and not the trim change, causes the principal changes in powering performance. More recently, Cusanelli and Karafiath (2001) reviewed the efforts in David Taylor Model Basin, since 1989 to design stern wedges (ending at the transom) and flaps (extending aft of the transom) for improving the performance of destroyers and frigates as well as 52-m long patrol coastal boats. In the later case, reduced span flaps were fitted. Other appendages that are fitted mainly on hard chine, high-speed monohulls are the T-foils, i.e. foils with reverted T cross section that are fitted in the bow region of the bottom, as well as interceptors in the stern region. Both of these devices, which are met in hard-chine hulls, are active. The T-foils are trim tuning and anti-pitching devices, fitted when fine bows result in poor pitch damping, while the action of interceptors is quite similar to that of the trimming flaps. In the previous paragraphs local modifications and add-ons to improve the performance of a fast monohull in calm water are presented. Local modifications, however, cannot affect the seakeeping characteristics of ships. Only a variation of the main (global) hull form parameters during the conceptual design can significantly affect the seakeeping performance of a ship. Along these guidelines Keuning and Pinkster (1995) proposed the Enlarged Ship Concept (ESC). The authors used an existing and quite successful design of patrol boat (Stan Patrol 2600) as base design, which they lengthened by 25% and 50%, whilst keeping all other design parameters, such as beam, speed, payload, etc. constant. Although the calculated building cost is increased by 6% in the latter case, the enlarged ship has the following advantages over the base boat: 16

17 The Fn is reduced for the same speed. The L/B and L/ 1/3 ratio are increased, which is beneficial both for calm water resistance and seakeeping. The pitch radius of gyration is increased. The position of the prime working areas on board is optimized with respect to vertical motions. Keuning and Pinkster (1997) and Keuning et al (2001) further refined the concept by proposing two modifications of the bow shape, over some 25% of the length, both below and above the still waterline, the TUD 4100 and the Axe bow, in order to improve the seakeeping behaviour. The aim of this bow modification was to reduce the non-linear hydrodynamic forces in particular at the fore ship (Fig.13). Fig.13: Bow refinement of ESC for improving seakeeping performance. Both bows (TUD 4100 to a lesser extent) result in: Reduction of the flare of the bow sections Narrowing and increase of length of waterlines Deepening of the fore foot Increase of freeboard. Keuning et al (2001) studied the behaviour (i.e. heave and pitch motions) in both head- and following irregular waves of the three systematic bow shape variations. They also investigated the manoeuvring characteristics for these variations. Since the 17

18 proposed bow modifications were suspected to increase the sensitivity of the ships with the sharper and deeper bows to broaching in following waves, they also studied this aspect of the behaviour in waves. The results of the comparison between these three designs, with the modified bow shapes, lead to the conclusion that the seakeeping performance of AXE 4100 hull form is superior to that of TUD 4100, which in turn is better than the ESC The comparison was made in terms of significant or extreme vertical acceleration in the bow region and slamming. The authors consider extreme values as more critical for limiting the operation of the vessel. On the contrary, manoeuvring characteristics and broaching tendency of the modified hull forms are inferior. 4 New trends in the design of fast monohulls Ten years ago, Paragon Mann ( designed a series of hybrid slender, wave piercing hull forms ranging 16 to 50 m in length, denoted as VSV, which are appropriate for very high speeds. This hull form possesses a high for its length L/B ratio (L/B > 5.50), a sharp bow profile with a very low entrance angle and wedgeshaped waterlines. The hull form was tested in the 380-ft towing tank of U.S. Naval Academy Hydromechanics Laboratory (Schleicher et al, 1997) against a conventional planing hull form. Although the two models had the same waterline length and payload, the wave piercing hull form displaced 20% less and exhibited lower resistance that the conventional hard-chine one, especially in the pre-hump region and no porpoising tendency. The LCG position was critical for both the improved calm water resistance and dynamic stability characteristics. Furthermore, VSV hull form had reduced seakeeping responses at high speeds, low radar signature for military operations and very good manoeuvring characteristics. Two VSV hulls with LOA = 16 and m, operating at speeds in the 50 kn range are currently produced by Halmatic. On the other hand, in the case of large monohulls (frigates, littoral combat ships, cruisers and passenger ferries), impact loads (bow slamming) result in severe distress of the structure in the bow region. Thus, the designers proposed to reverse the inclination of the bow stem profile resulting in a wave-piercing configuration, which reduces significantly bow fatigue due to wave loads. Furthermore, this tumblehome hull form design offers significant power savings due to reduced calm water resistance, while it is also critical to meeting low Radar Cross Section (RCS) signature objectives. The concept has already been incorporated in the new four-year US Navy Project awarded in 2001 to Northrop Grumman led Gold Team, denoted as DD(X) and aiming at the design of a high-performance, low operational cost frigate. Furthermore, a similar concept incorporating a protruding bow bulb, developed by Chantiers de l Atlantique and Principia Marine has been adopted for the 240-m, 38- knot fast ferry ROPAX 2000 designed within the VRSHIPS-ROPAX Project (Fig. 14). The vessel will be propelled by a combination of a pair of waterjets and a pair of pods. 18

19 Fig.14: Artistic view and preliminary body plan of the ROPAX 2000 high-speed ferry. 5 Pros and cons of advanced fast monohulls In this Section, advanced fast monohulls of displacement, semi-displacement and planing type are compared against other types of modern vehicles. Furthermore, the hard-chine hull form is compared against the other competitive type of monohulls, the round bilge hull form. All comparisons refer to similar-sized ships and they are based on the following characteristics (Repetto, 2001): Platform stability, deck area, volume space and draft weight and trim sensitivity. Range of speed and propulsion configuration Seakeeping and maneuvering characteristics Global and local strength, slamming loads Survivability (stealth characteristics and vulnerability) Acquisition and operating costs. Modern vehicles, monohulls and multihulls are classified into three major categories according to the way their weight is supported, i.e. hydrostatic buoyancy, hydrodynamic lift and powered lift. These categories form the corners of the classical sustentation triangle (Fig.15a), while along the sides and inside the triangle hybrid hull forms can be found. Thus, advanced monohulls (semi-displacement and planing) are located along the side connecting hydrostatic buoyancy with hydrodynamic lift. The sustentation triangle evolves to a pyramid if aerodynamic lift is also taken into consideration (Fig. 15b). On the basis of the above classification of ship types advanced monohulls are firstly compared with displacement ships, including SWATH, conventional displacement catamarans and trimarans. SWATHs, which operate at speeds exceeding the limitation of Fn = 0.50, inherent to the surface displacement ships, are superior with respect to seakeeping and offer a large deck. However, due to their increased wetted surface, they higher installed power requirements and their draft (which is 19

20 large) and trim, are sensitive to displacement changes. Their acquisition and operating cost is higher than that of monohulls. In order to increase the attainable speed of SWATHs, the stern is modified to that of a planing catamaran. Hydrostatic Buoyancy (displacement hull) Aerodynamic Lift WIG Hydrodynamic Lift Powered Lift Hydrodynamic Lift Hydrostatic buoyancy Powered Lift (a) (b) Fig.15: Classical sustentation pyramid and triangle (Repetto, 2001). Displacement catamarans operate at higher speeds at the same cost, offer larger deck area, reduced roll motions, higher initial stability, better maneuverability and survivability. However, they suffer from structural problems in the transversal box connection and have higher vertical responses. Displacement trimarans, on the other hand are expected to further increase the advantages of catamarans, while they reduce their disadvantages. At the hydrodynamic corner of the sustentation triangle the hydrofoils are located. Both types of them the surface piercing and the fully submerged achieve higher cruising speeds, higher level of comfort up to wave heights, which prevent foil borne mode, and excellent maneuverability. Even in hull borne mode of operation in very rough seas, foils reduce both vertical and lateral motions. On the other hand, their principal disadvantage is their limited payload capability and their large draft. On the side of the triangle connecting the hydrostatic and the hydrodynamic lift, semi-displacement and planing monohulls and multihulls are located. The latter combine the characteristics of multihulls with high-speed. Furthermore, since the length of planing catamarans is comparable with their breadth, the combined pitch roll dynamic response (cork screw motion) emerges to be a major cause of discomfort for the passengers in rough seas, which has to be reduced by the use of active motion controls. On the same side of the triangle wave piercers are found. This type of unconventional hull forms encompasses displacement and hybrid monohulls, as well as hybrid catamarans. The fitting of a tumblehome bow offers improved calm water resistance characteristics both at intermediate (displacement mode) and high (planing) speeds, reduced structure loading due to impact loads (slamming) and low radar cross 20

21 section signature. Thus, as explained in the previous section, wave piercers have become popular, recently. At the power lift corner of the sustentation triangle Air Cushion Vehicles (ACV) are located. Among the pros of these vehicles are the ability to operate at very high speed, their low vulnerability to underwater explosions, their small draft and underwater signatures and their amphibious operation. On the contrary, they are affected by winds, they are sensitive to trim and have high acquisition and maintenance costs, due to seals and lift fan systems and specific electronic equipment for ridecontrol devices. A popular type of hybrid hull forms is the Surface Effect Ship (SES), a crossover between displacement catamaran and ACVs, which operates at speeds in excess of 40 kn, with reduced underwater signature levels, good platform stability, shallow draft and large deck area. It also has better calm water transport efficiency ET for volumetric Froude numbers Fn > 2 (Blount, 1993). The non-dimensional calm water transport efficiency ET is defined as ET = ΔV/P, where P is the total power used for propulsion and dynamic lift. However, this type is not amphibious and suffers from significant speed loss in head seas, higher production and maintenance costs. The top of the sustentation pyramid is allocated to Wing in Ground (WIG) vehicles, which can be seen as a crossover between an ACV and an aircraft. These vehicles, which operate at speeds in the range of 50 to 250 kn, have a very high transport efficiency expressed as the amount of fuel used per passenger per kn. Many of them designated as ekranoplans are in operation in the Russia and Ukrania since the early 60s. However, only a sufficiently large WIG, weighting around t, would fulfill all expectations concerning efficiency and seaworthiness, while a power several times larger than that required for cruising is necessary for taking-off the craft. On the basis of the above discussion, monohulls can provide a suitable solution to most of the operational requirements, because of their inherent flexibility of the hull type. Thus, they constitute a very competitive and popular type of advance vehicles. On the other hand, among high-speed monohulls hard-chine and round bilge hull forms compete strongly each other. The evolution of hard chine hulls lead to the development of deep-vee hull forms, studied and experimentally evaluated extensively by Serter since early 60s (Serter, 1982), and the double chine ones with warp which were described in the section on NTUA Series. Proceeding now to the comparison of the two major monohull competitors, the hard chine and the round bilge hull form, the former one (deep-vee and double chine) possesses better seakeeping qualities resulting in reduced power requirements in confused seas, better maneuvering, dynamic stability and course-keeping characteristics. Its calm water performance is, in general, inferior to that of an equivalent round bilge hull at the lower speed range up to Fn less than about 0.45 (see Fig.2), while it becomes superior at higher speeds (Blount, 1995). Finally, hard chine hull provides more internal space than the round bilge one and it can easily be fitted with water jets for operation in shallow waters. 21

22 6 Discussion - Conclusions In this paper recent (last decade) advances in the hydrodynamic design of highspeed monohulls are presented. Firstly the outline of three new round-bottom and one double-chine hull form Series is presented. All series are wide-transom and appropriate both for commercial and naval ship applications ranging from relatively small fast ships (yachts, patrol boats) to large ships (ferries, corvettes). The former operate at speeds well in excess of Fn = 0.50, while the latter operate at speeds around the hump (Fn = 0.40 to 0.60). Since, the performance of these vessels in the above speed range is very sensitive to local modifications and to appendages fitted, the effect of stern wedges, adjustable trimming flaps, ands spray rails was presented. The combination of these appendages adjusts the trim of the vessel to the optimum one, corresponding to the minimum resistance, while they also reduce the running displacement by offering lift, especially at the higher speed range. Similar is the effect of T-foils and interceptors. In addition, the Enlarged Ship Concept to improve the calm and rough water behaviour of a vessel by varying its global hull form characteristics is presented and discussed. The new trends in the design of high-speed monohulls are expressed in terms of two novel monohull designs presented, one appropriate for small patrol boats and pleasure craft and the other for large surface combatants and ferries. The former combines a very slender bow with wide transom, while the latter possesses a tumblehome bow form, with or without a bow bulb, reducing resistance in calm water resistance and slamming in rough water. They both constitute promising monohull solutions. Finally, in order to support the view that monohulls are a very attractive choice for most of the design dilemmas about the type of the hull form to be used, the advantages and the disadvantages of high-speed monohulls over the other competitive advanced hull forms are presented. In addition, the major monohull types are compared each other and their relative merits are summarized. Only the widely recognized characteristics of the various hull forms were taken into consideration for the comparisons presented in the section. Nomenclature AM the sectional area of the maximum ship section 0.5 LCG B V 2 m, Almeter number chine beam, the volume of displacement coefficient An Bm CDL = 0.1L 3 WL 4 3/4 Cn = LCG 9 / B m, Clement number CR = R ( 1 2 R 2 WS V ), residuary resistance coefficient CFS frictional resistance coefficient for the ship CTLm = RTm / (Δ Fn 2 ), non-dimensional resistance coefficient CTm = R ( 1 2 WS V ), total resistance coefficient for the model Tm 2 CTS = R ( 1 2 WS V ), total resistance coefficient for the ship TS 2 22