See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/283048285

Improved Resistance Prediction for High-Speed Round Bilge Hull Forms

Conference Paper · September 2012

CITATIONS READS
0 5,620

1 author:

Prasanta Sahoo
UNSW Canberra
81 PUBLICATIONS 650 CITATIONS

SEE PROFILE

All content following this page was uploaded by Prasanta Sahoo on 03 May 2017.

The user has requested enhancement of the downloaded file.

Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Improved Resistance Prediction for High-Speed Round Bilge Hull

Forms

Prasanta K Sahoo1 and Jae Won

[email protected]

ABSTRACT

Predicting the resistance of a high-speed monohull has been of interest to Naval Architects for several decades.
Even though considerable amount of research has been carried out in this area, there remains a degree of
uncertainty in the accurate resistance prediction in the early design stage which would be immense benefit to
practising naval architects.

This research paper attempts to investigate a method for enhancing the accuracy of resistance prediction methods
for high-speed round-bilge monohull form vessels for a wide range of volumetric Froude numbers (F n). While a
number of systematic series are in existence, their data are either not readily available or scattered in various
internal reports and publications which makes it difficult for practicing naval architects to exploit the knowledge
base. Various countries around the world are enhancing their own defence capabilities by building conventional
round bilge high-speed craft in the range of 30 to 40 knots for navy, coast guard, maritime police to guard against
national security, terrorism threats and high seas piracy. In this paper the following high-speed systematic series
hull forms have been considered for regression analysis, namely:

 NPL (1969)
 S-NPL (1994)
 SKLAD (1972-1980) and
 AMERC (1984-2000)

Earlier objective of this paper was to obtain a common regression equation for a wide parameter space which
would be encompassing all the above systematic series. As this was not feasible due to lack of data in areas that
were considered crucial, hence separate regression analysis has been carried out for each series. The new
regression equations have been proposed for a broad range of geometrical parameters so that a designer has an
instant tool to make a decision regarding powering prediction in the design stage.

INTRODUCTION

Fuel economy and environmental concerns are two dominant factors in this century that demand that resistance be
accurately predicted in the early design stage, so that there is no undue penalty due to high fuel costs throughout the
life of the vessel. This has its own implication in choosing the most appropriate propulsion system to suit the
vessel’s resistance characteristics.

1
Associate Professor (Ocean Engineering), Florida Institute of Technology, 150 W University Boulevard, Melbourne, Florida 32901, USA
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Numerous studies have been carried out on predicting the resistance of high-speed round bilge vessels. Various
methods, factors and assumptions were employed by various researchers at various times which have influenced
their analysis. Owing to different techniques, these methods would have significant difference among their predicted
results for a particular vessel. Therefore, people who may consider using one of the resistance prediction methods to
predict a vessel’s resistance must carefully choose the most appropriate method that suits the vessel’s geometrical
characteristics.

The aim of this paper was to investigate the NPL, S-NPL, SKLAD, and AMECRC series which have undergone
exhaustive tank testing and combine them in an attempt to find a superior solution in predicting the resistance for a
broad range of round bilge high-speed vessels parameters. While this was not feasible while investigations were
under way, separate regression equations have been developed for each of the series pertaining to their own unique
geometrical characteristics.

NOMENCLATURE

AT transom area LCG longitudinal centre of gravity

AX maximum section area (M) length-displacement ratio, (L / ∇1/3)
B breadth (beam) of hull on DWL PB total engine brake power
c chord-length of foil P.C. propulsive coefficient (PE / PB)
CB block coefficient s separation of foils
CL coefficient of lift SP sinkage pressure force on demihull
DF foil drag (S) wetted surface coefficient (S / ∇2/3)
DWL design waterline T draught at DWL
eP dimensionless power ratio U uniform flow velocity
FC chord-Froude number, (U / √gc) V speed
Fn Froude number, (V / √gL) W weight of the vessel
volumetric Froude number, (V /
Fn ZPROP vertical component of the propulsion force
√g∇1/3)
Fn Froude displacement number αi resistance regression coefficients
G, g acceleration due to gravity Β dynamic trim regression coefficient
IW wave interference factor Ε total resistance and weight ratio
L length on DWL ηT transport efficiency
LF foil lift λ wavelength
LCB longitudinal centre of buoyancy ∆ displacement mass
LCF longitudinal centre of floatation ∇ displacement volume
 dynamic trim angle

ABBREVIATIONS

CFD Computational Fluid Dynamics

NPL National Physical Laboratory

KEYWORDS

Monohull, Resistance, High-Speed, Regression

ROUND BILGE MONOHULL SYSTEMATIC SERIES

Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Table1 presents some of the series’ parameters and list of original references. The performance of round bilge hulls
is strongly dependent upon the slenderness ratio, L/1/3, Savitsky et al. (1973). Figure 1 illustrates the L/1/3 range
covered by each of the series while the Figure 2 presents the body plans of the round bilge hull systematic series
described in Table 1.

Table 1: Round Bilge Hull Systematic Series [Bojovic(1998)]

Series (No. of Models) L/1/3 L/B B/T CB LCB Fn

Nordstrom (12*) 5.65-7.72 4.83-6.94 3.16-3.57 0.373-0.41 0.9-2.0
De Groot (31*) 5.23-7.75 0.8-2.7
Marwood-Silverleaf (30*) 5.2-8.2 .45-1.12
Series 63 (5) 4.5-6.4 2.5-5.75 0.9-2.2
Series 64 (27) 8.04-12.4 8.45-18.26 3.0-4.0 0.35-0.55 0.1-1.5
SSPA (9) 6.0-8.0 4.62-8.20 3.0-4.0 0.4 1.0-2.0
NPL (22) 4.47-8.30 3.33-7.50 1.5-2.5 0.397 6.4% L aft Fn=0.3-1.2
S-NPL extended (10) 6.3-9.5 7.0-15.1 1.5-2.5 0.397 6.4%L aft Fn=0.1-1.05
YP (3) 5.57-5.72 3.97-5.17 3.06-5.05 Fn=0.1-0.6
SKLAD (27) 4.5-8.5 4.0-8.0 3.0-5.0 0.35-0.55 1.0-3.0
NRC (24) 6.62-7.93 3.0-5.0 0.35-0.55 Fn=0.2-1.0
D-Series (13) 6.3-6.93 5.68-7.05 3.0-3.75 0.48-.52 Fn=0.15-0.80
VTT (4) 6.586 5.41-6.25 4.39-6.90 0.45-0.60 0.6-3.8
MARIN HSDHF (40) 4.3-13.1 4.0-12.0 2.5-5.5 0.35-0.55 Fn=0.1-1.2
AMECRC HSDHF (14) 4.3-8.7 4.0-8.0 2.5-4.0 0.40-0.50 Fn=0.1-1.0

Figure1: L/1/3 Ranges for Round Bilge Systematic Series [Bojovic (1998)]

NPL SERIES

Resistance data for high-speed round bilge form obtained at NPL were originally presented in 1969. The work was
extended to examine the effect of the hull parameters on calm water resistance, Bailey (1976). Experimental
investigations involved testing of 22 models where the bare hull models were bereft of any keel or appendages. The
water line length LWL and the block coefficient CB of the models were set at 2.54 m and 0.397 respectively, where
the B, T and the displacement of the vessel were varied. The model was also designed to have the LCB at 6.40% of
LWL aft of amidships. These vessels were divided in to 7 groups according to their slenderness ratio, L/1/3. Figure 3
represent the parent hull form of NPL series and Figure 4 describes the distribution of geometrical parameters of
NPL and S-NPL series.
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

SOUTHAMTON EXTENDED NPL SERIES

Ten slender round bilge models were derived from the NPL series and this extended series was deemed to broadly
represent sleder hull forms suitable for catamaran applications. The calm water resistance testing of the S-NPL has
been described by Molland et al. (1994). The models were tested as monohulls and in catamarans configurations
with different hull spacing. The body plan of the hull forms are shown in Figure 5.

Figure 2: Round Bilge Systematic Series- Body Plans [Bojovic (1998)]

Figure 3: NPL Series Parent Hull Body Plan [Bailey(1976)]

Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Figure 4: Range of Parameters Covered in NPL and S-NPL Series [Bojovic (1998)]

Fig. 5: Southampton Extended NPL Series [Molland et al (1994)]

Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Fig. 6: SKLAD Series Parameter Space [Radojcic et al (1999)]

Fig. 7: SKLAD Series Parent Hull Plan [Radojcic et al. (1999)]

SKLAD SERIES
The research on SKLAD series of models were carried out at the Brodarski institute, in the former Yugoslavia, over
the period from 1972 to 1980. Twenty seven high speed round-bilge, transom-stern, semi-displacement hulls were
developed and used for the research. The displacement volume was kept constant at 0.230 m 3, so that the length of
the models varied from 2.7 to 6m. The models were divided in to three groups each, according to their block
coefficients CB, L/B ratio and B/T ratio. The ranges of varied parameters are outlined in Table 2 and the series’
parameter space is illustrated in Figure 6. The parent hull form of SKLAD series has been shown in Figure 7.

Constant values were taken for the position of the longitudinal centre of buoyancy so that LCB = 8.8%, 9.3% and
9.2% of the LWL aft of amidships for CB=0.35, 0.45 and 0.55 respectively. The round-bilge models had a stern-
knuckle (chine) for approximately 20% LDWL and a built-in spray rail for approximately 40% LDWL. The after body
bottom was both flat (12 degree deadrise angle) and hooked (wedge is incorporated) and provided enough space for
propellers, a low shaft angle, and low dynamic trim. Forward sections were of a deep V-form with a small angle of
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

entrance that ensured good seakeeping qualities and reduced resistance. All models were without appendages and
were ballasted to even trim at rest and towed horizontally at the centre of buoyancy.

Table 2: SKLAD Series Parameters and Range [Radojcic et al. (1999)]

Parameters Range
L/B 4.0 – 8.0
B/T 3.0 – 5.0
CB 0.35 – 0.55
L/1/3 4.5 – 8.5
LCB %LWL aft of 8.8, 9.3 and 9.2 for
midship each CB

Each group had a constant prismatic coefficient CP of 0.715 and a maximum section-area coefficient of 0.621. The
series models were tested over the volumetric Froude number range Fn form 1.0 to 3.0. A comprehensive
regression analysis have been performed by the authors, Radojcic et al (1999), to determine the residuary or total
resistance coefficients.

Fig. 8: AMECRC Parent Hull Body Plan [Sahoo and Doctors (1999)]

AMECRC SERIES

The AMECRC systematic series is based on the High-Speed Displacement Hull Form (HSDHF) systematic series
developed at the Maritime Research Institute Netherlands (MARIN). The HSDHF project was major research
project on combatant vessel design, jointly sponsored by the Royal Netherlands Navy, the United States Navy, the
Royal Australian Navy and MARIN. It was initiated by the growing belief that a significant improvement in the
performance of transom stern, round bilge monohulls could be obtained, especially with regard to their Seakeeping
characteristics.
The parent hull form (Figure 8) of AMECRC series was based on the parent hull form of HSDHF series and
subsequently 13 more models were developed by systematic variation of L/B, B/T and CB, Sahoo and Doctors
(1999). Table 3 presents the geometrical parameters range of HSDHF and AMECRC series and Figure 8 depicts the
parent hull form.. The ‘parameter space’ or series ‘cube’ of the AMECRC series is given in Figure 9. All the 14
models have the same length of 1.6 m and the influence of change of the series parameters on the hull shape are
illustrated in Figures 9 and 10 and parameter range in Table 4.
The models were tested in the Ship Hydrodynamics Centre at the Australian Maritime College. All models were
constructed with a water line length of 1.6 m. Calm water tests were conducted at speeds from 0.4 to 4 m/s,
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

corresponding to Froude Number (Fn) from 0.1 to 1.0. During testing, the models were free to sink and trim, and
resistance, trim and rise of centre of gravity were recorded.

Table 3: Parameters for HSDHF and AMECRC Systematic Series [Sahoo and Doctors (1999)]

HSDHF AMECRC
L/B 4 – 12 4 -8
B/T 2.5 – 5.5 2.5 – 4.0
CB 0.35 – 0.55 0.396 – 0.50

Fig. 9: AMECRC Systematic Series Parameter Space [Sahoo and Doctors (1999)]

Fig. 10: Change in Hull Shape of AMECRC Series [Sahoo and Doctors (1999)]
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

REGRESSION ANALYSIS

The general purpose of multiple regression is to analyse the relationship between several independent variables or
predictor variables, called x, and a dependent or criterion variable, y. The experiment data model doesn’t have not
only several independent variables but also have several valid cases.

It is easy to deal with data that have some linear or nonlinear relationship between them. A straight or a curve line
can be fit through those points. From the equation of that line the same parameter can be predicted within the valid
range. Simple regressions like this can be done with the help of Microsoft Excel program.

Hull form parameters

 Analysed hull form parameters should be in non-dimensional form.
 Data should give a uniform coverage of the space defined by variations of the analysed hull form parameters. The
density of the uniform coverage is not vital factor, but it is desirable to obtain a full combination of at least three
values of each varied parameter.
 All the parameters that may have a significant effect on the dependent variable should be included in the analysis,
and any parameters that are not included should either be constant or should have an insignificant influence on the
dependent variable.
 The extreme values of all varied parameters should be carefully defined.

Table 4: Systematic Series Parameter Range [Sahoo and Doctors (1999)]

Model L/B B/T CB Model L/1/3
Disp.(kg)
1 8 4 0.396 6.321 8.653
2 6.512 3.51 0.395 11.455 7.098
3 8 2.5 0.447 11.454 7.098
4 8 4 0.447 7.158 8.302
5 4 4 0.395 25.344 5.447
6 8 2.5 0.395 10.123 7.396
7 4 2.5 0.396 40.523 4.658
8 4 2.5 0.5 51.197 4.308
9 8 2.5 0.5 12.804 6.839
10 8 4 0.5 8.002 7.998
11 4 4 0.5 32.006 5.039
12 8 3.25 0.497 9.846 7.464
13 6 3.25 0.45 15.784 6.379
14 6 4 0.5 14.204 6.606
Assumptions Regarding Regression Analysis Application

 The principal parameters of the hull whose performance is being predicted must fall within the range of
parameters’ values covered by the data.
 All parameters that are constant in the analysed data set, must have that same constant value in the proposed
design (comment: if a certain parameter is constant, it does not reduce the prediction accuracy, only prevents the
investigation of the effect of that parameter).
Selection of Independent Variables

Independent variables are generated as functions of varied hull parameters and/or speed.
 Independent variables should be in non-dimensional form. When there is theoretical evidence as to the form that
the independent variables should take, an attempt should be made to utilise that form.
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

 When a regression equation has two highly correlated variables as useful independent variables, it is wrong to
further include their product them as an independent variable because it will lead to some instability in the
equation. And is unnecessary as it will not add significantly to the accuracy of the equation.
 It is possible to have two highly correlated independent variables which if one is included in the regression
equation without the other is not effective, but if both are included then the equation is more accurate.
 It also possible to have two highly correlated independent variables in a regression equation which both have
significantly non-zero coefficients, but which predominantly explain the variance of each other rather than the
variance of the dependent variable. Each of these could become insignificant, if the other is removed from the
equation.

Production of Good Regression Analysis

 Each independent variable used in the regression equation should have a high significance level, generally not
lower than 95%.
 It should not be possible to improve the accuracy of the equation by introducing extra independent variables.
 It should not be possible to exclude an independent variable from the equation without significantly reducing the
accuracy.
 The regression equation should not contain more than ten independent variables, Fairlie-Clark (1975). Fung
(1991) concluded that residual error starts to stagnate after inclusion of 1 to 17 terms. More terms in a regression
equation may contribute to a better fit to the data, yet give a poorer interpolation result, Savitsky et al. (1976).

REGRESSION ANALYSIS TECHNIQUE

Several other techniques have been tried to predict RR/ in terms of L/B, B/T, L/1/3 and CB. The method used by
Radojcic (1997) has been adopted here. In his method principal hull form and loading parameters were transformed
in to another set of variables with a range from -1 to 1.
The method has two sets of similar equations, one for AMECRC and SKLAD series hull forms where the parameter
space has varying L/B, B/T, L/1/3 and CB; and the other for S-NPL series hull form which has all but CB is fixed.
The equation developed for S-NPL is similar to the equation developed by Radojcic (1997) for NPL hull forms
which have 27 terms, and the equation developed for AMECRC and SKLAD series have 48 terms. They are as
follows:
S-NPL hull series:
 L ( L / B) min  ( L / B) max 
  
x1   
B 2
 ( L / B) max  ( L / B) min 
  (1)
 2 

 B ( B / T ) min  ( B / T ) max 
  
x2   
T 2
 ( B / T ) max  ( B / T ) min 
  (2)
 2 

 L ( L / 1/ 3 ) min  ( L / 1/ 3 ) max 

 1/ 3  

x3   
2
 ( L /  ) max  ( L / 1/ 3 ) min 
1/ 3

  (3)
 2 
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

RR   a0  a1 x1  a2 x2  a3 x3  a4 x12  a5 x2 2  a6 x3 2  a7 x1 x2  a8 x1 x3  a9 x2 x3  a10 x12 x2  a11 x12 x3  a12 x2 2 x1

 a13 x2 2 x3  a14 x3 2 x1  a15 x3 2 x2  a16 x13  a17 x2 3  a18 x33  a19 x13 x2  a20 x13 x3  a21 x2 3 x1
 a22 x2 3 x3  a23 x33 x1  a24 x33 x2  a25 x12 x2 2  a26 x12 x3 2  a27 x2 2 x3 2
(4)
AMECRC & SKLAD hull series:
 L ( L / B) min  ( L / B) max 
  
x1   
B 2
 ( L / B) max  ( L / B) min 
  (5)
 2 
 B ( B / T ) min  ( B / T ) max 
  
x2   
T 2
 ( B / T ) max  ( B / T ) min 
  (6)
 2 
 L ( L / 1 / 3 ) min  ( L / 1 / 3 ) max 
  
 1 / 3 2 
x3   
 ( L / 1 / 3 ) max  ( L / 1 / 3 ) min 
  (7)
 2 
 
 (C )  (CB ) max 
 CB  B min 
x4   
2
 (CB ) max  (CB ) min 
  (8)
 2 
R R   a 0  a1 x1  a 2 x 2  a3 x3  a 4 x 4  a5 x12  a 6 x 2 2  a 7 x3 2  a8 x 42  a9 x1 x 2  a10 x1 x3  a11 x1 x 4  a12 x 2 x3
 a13 x 2 x 4  a14 x3 x 4  a15 x1 2 x 2  a16 x1 2 x3  a17 x12 x 4  a18 x 2 2 x1  a19 x 2 2 x3  a 20 x 2 2 x 4  a 21 x3 2 x1  a 22 x3 2 x 2
 a 23 x3 2 x 4  a 24 x 42 x1  a 25 x 42 x 2  a 26 x 42 x3  a 27 x13  a 28 x 2 3  a 29 x3 3  a30 x 4 3  a31 x13 x 2  a32 x13 x3  a33 x13 x 4
 a34 x 2 3 x1  a35 x 2 3 x3  a36 x 2 3 x 4  a37 x3 3 x1  a38 x3 3 x 2  a39 x3 3 x 4  a 40 x 4 3 x1  a 41 x 4 3 x 2  a 42 x 4 3 x3  a 43 x1 2 x 2 2
 a 44 x12 x3 2  a 45 x12 x 4 2  a 46 x 2 2 x3 2  a 47 x 2 2 x 4 2  a 48 x3 2 x 4 2
(9)
FORWARD STEPWISE REGRESSION PROCEDURE

This is the method used in the paper for regression analysis. The method starts with a single independent variable in
the regression model. At each succeeding step, additional independent variables are introduced based on
significance testing using the t-test. Those variables that possess the greatest statistical significance relative to the
dependent variable are added first. This procedure is repeated until no significant independent variables can be
found outside the regression model. Also, if a previously added independent variable becomes insignificant due to
the addition of new variables, the previously added variable is removed from the regression model. The acceptance
and rejection of each independent variable is purely based on the F-test (criteria F-to-remove and F-to-enter in the
software).
This model has been used for all of the analyses described in this paper. F-values, mentioned above, where selected
so that the final regression model contains no variables with a statistical significance (p-level) greater than 0.05
(5%). Specifically, p-level represents the probability of error that is involved in accepting observed results as valid.
When conducting regression analysis results could become unstable if highly correlated independent variables are
included in the regression model. Control over this matter was achieved by setting the tolerance level from 0 – 0.001
(0% - 0.1%). That means that variables whose tolerance was under this level were considered redundant with the
contribution of other independent variables already in the equation.
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

FINAL REGRESSION MODEL AND RESULTS

For the purpose of developing the regression equations, initially RR/ has been derived from the experimental data.
AMECRC and S-NPL systematic series had their own published experimental data, where SKLAD systematic series
had a regression equation developed by Radojcic et al. (1999) based on CR.
Linear and non-linear interpolation analyses have been performed to arrange the results for the specific Froude
numbers where needed. Therefore an uncertainty is expected in organizing the final experimental results for
regression analysis. Thus assuming the uncertainty is very small, it has been ignored in this regression analysis.
Regression equation has been obtained for a wide range of Fn. The ranges of Fn are for AMECRC systematic
series 1 to 2, S-NPL systematic series 1 to 2.5 and SKLAD systematic series 1 to 3. All three series have increments
of 0.25 for Fn. The regression equation obtained for all range of Fn have a higher degree of accuracy with R2 =
0.9999 or higher while obtaining the coefficients of the equations. The predicted values of RR/ for all the existing
models are quite close to the actual experiment values. Regression equation obtained by Bojovic (1998) for
calculating the wetted surface area coefficient (CS) has been reproduced.
It is to be noted that regression coefficients which do not play a significant role in the regression equation have been
ignored in the equations as shown. Table 5 depicts the regression coefficients for the AMECRC Series which needs
to be read in conjunction with final equation (10). Table 6 represents the regression coefficients for the SKLAD
series as per the equation (11).

Equation for AMECRC Series:

RR
 a0  a3 x3  a4 x4  a6 x2 2  a7 x3 2  a8 x42  a10 x1 x3  a13 x2 x4  a14 x3 x4

 a15 x12 x2  a16 x12 x3  a17 x12 x4  a18 x2 2 x1  a19 x2 2 x3  a20 x2 2 x4 
a22 x3 2 x2  a23 x3 2 x4  a25 x42 x2  a27 x13  a32 x13 x3  a33 x13 x4 
a35 x2 3 x3  a36 x2 3 x4  a37 x33 x1  a39 x33 x4  a40 x4 3 x1  a41 x43 x2
(10)
 a43 x12 x2 2  a44 x12 x3 2  a48 x32 x4 2

Equation for SKLAD Series:

RR
 a 0  a1 x1  a 2 x 2  a3 x3  a 4 x 4  a5 x1 2  a 6 x 2 2  a 7 x3 2  a8 x 42  a9 x1 x 2  a10 x1 x3  a11 x1 x 4

 a12 x 2 x3  a13 x 2 x 4  a14 x3 x 4  a15 x1 2 x 2  a16 x1 2 x3  a17 x12 x 4  a18 x 2 2 x1  a19 x 2 2 x3  a 20 x 2 2 x 4
 a 21 x3 2 x1  a 22 x3 2 x 2  a 23 x3 2 x 4  a 24 x 42 x1  a 25 x 42 x 2  a 26 x 42 x3
(11)

Equation for S-NPL Series:

RR
 a0  a1 x1  a 2 x2  a3 x3  a 4 x12  a5 x2 2  a6 x3 2  a7 x1 x2  a8 x1 x3  a9 x2 x3  a10 x12 x2

 a11 x12 x3  a12 x2 2 x1  a13 x2 2 x3  a14 x3 2 x1  a15 x3 2 x2  a16 x13  a17 x2 3  a18 x33
(12)

Equation for NPL Series, Radojcic (1997):

RR
 a0  a1x1  a2 x2  a3 x3  a5 x2 2  a12 x2 2 x1  a13 x2 2 x3  a16 x13  a19 x13 x2  a20 x13 x3  a22 x23 x3  a23 x33 x1

 a26 x12 x32  a27 x2 2 x32 (13)
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Table 7: AMECRC Series Regression Coefficients

Fn
ai 1.00 1.25 1.50 1.75 2.00

a0 0.023279 0.043599 0.061922 0.066116 0.060799

a3 0.000000 -0.052000 -0.012971 -0.048972 -0.052922

a4 0.000000 -0.004931 0.000000 0.000000 0.000000

a6 0.000000 0.000000 0.000000 0.002113 -0.002020

a7 -0.099520 0.000000 0.000000 0.000000 0.000000

a8 0.000000 -0.000432 0.000000 0.000000 0.000000

a10 0.000000 0.021886 0.000000 0.072744 0.031722

a13 -0.003880 0.000000 0.523622 0.299613 0.000000

a14 0.000000 0.000000 0.000000 0.000000 -0.012989

a15 0.000000 0.003712 0.018875 0.006999 -0.006071

a16 -0.036998 0.000000 -0.026311 0.000000 0.000000

a17 0.000000 0.010202 0.007989 0.008120 0.008268

a18 0.000000 0.000000 0.000000 0.000401 0.000000

a19 -0.001972 0.000000 0.000000 0.000000 0.000000

a20 0.000000 0.002972 0.000000 0.000000 0.007980

a22 -0.001931 0.000000 0.000000 0.000000 0.007229

a23 0.000000 -0.018931 -0.046854 -0.024202 -0.025657

a25 -0.000610 -0.000843 -0.002327 -0.002226 0.000000

a27 -0.001214 0.000000 0.000000 0.000000 0.000000

a32 0.000000 0.000000 -0.003590 -0.069911 0.000000

a33 0.000000 0.000000 0.000000 0.000000 0.002056

a35 0.000000 0.000000 -0.000625 0.000000 0.000000

a36 0.000000 0.006318 -0.022273 0.000000 0.000000

a37 -0.002368 0.000000 0.000000 0.000000 0.000000

a39 0.002468 0.002279 -0.006240 0.000000 0.039232

a40 0.000000 0.000000 0.000000 0.002961 -0.003173

a41 0.000000 0.000000 -0.539618 -0.299514 0.000000

a43 0.000000 0.000000 0.004550 0.000000 0.000000

a44 0.098373 0.000000 -0.001300 0.000000 0.000000

0.000000 -0.007258 0.000000 -0.004089 0.000000

a48
Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Table 10: SKLAD Series Regression Coefficients

Fn
ai 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3
a0 0.513250 0.032440 0.290520 -0.945600 0.154870 0.440900 -0.848500 -1.026700 -0.070970

a1 -3.306930 0.000000 0.459060 10.679900 0.000000 -2.522410 6.844500 8.281800 -0.151850

a2 -1.235130 0.000000 -0.040812 3.011100 -0.269920 -0.826620 2.113600 2.557500 0.000000

a3 5.562080 -0.060170 2.491560 -11.914900 1.349990 4.726010 -10.093300 -12.212900 -1.006460

a4 1.124810 0.000000 0.681920 -2.018900 0.277620 0.919760 -2.041200 -2.469800 -0.358190

a5 0.108200 0.000000 -6.292570 -12.501300 -2.358710 -0.641400 -3.248100 -3.930200 2.069300

a6 0.104930 0.002860 -0.354660 -1.193600 -0.007070 -0.236280 0.013600 0.016400 0.450970

a7 -0.390920 -0.018758 -3.807000 -7.143100 1.011810 2.478460 0.926100 1.120600 3.330640

a8 -0.125760 0.004260 0.296720 0.981800 0.051650 -0.031980 0.179000 0.216600 -0.029371

a9 -0.332970 -0.000290 -3.444920 -6.180200 -0.985530 -1.335920 0.000000 0.000000 1.963290

a10 0.397760 0.123820 12.973620 24.520100 3.479220 0.000000 4.356800 5.271700 -5.405990

a11 0.172350 0.000000 1.101930 1.534300 0.615690 0.241520 0.914000 1.105900 0.394460

a12 0.113140 0.048200 2.683190 5.922500 -0.045110 0.922390 -1.043000 -1.262000 -2.727690

a13 0.058230 -0.001020 0.002800 -0.039800 -0.031620 0.264600 -0.182300 -0.220600 -0.007310

a14 0.124810 -0.037890 1.168150 1.892700 0.643570 1.017340 -0.476500 -0.576600 -1.168100

a15 -0.011950 0.000000 -0.162990 -0.285900 -0.177010 -0.155620 0.261000 -0.315800 -0.178380

a16 0.042470 0.000000 -1.883950 -3.649800 -0.529810 -0.080960 -0.530500 -0.642000 1.088780

a17 0.001910 0.000000 0.000000 0.000000 -0.032470 -0.081400 -0.035400 -0.042800 -0.056290

a18 0.000000 -0.003120 0.168550 0.325700 0.027300 -0.049170 0.010700 0.013000 -0.160160

a19 0.037150 0.000000 -0.221780 -0.624300 0.007380 -0.016400 0.041700 0.050400 0.296080

a20 0.000090 0.000000 0.004360 0.003400 -0.004810 -0.144000 -0.008300 -0.010100 -0.012530

a21 0.078230 0.000000 0.556630 0.879400 0.180780 0.416740 -0.083700 -0.101300 -0.302420

a22 -0.000490 0.006590 0.172920 0.377300 0.104600 0.025110 0.218600 0.264500 0.121380

a23 0.043660 -0.009150 0.598090 1.080300 0.506520 0.777610 0.425400 0.514700 0.147860

a24 -0.031620 0.000000 -0.195200 -0.271500 -0.076760 -0.005090 -0.094400 -0.114200 0.000320

a25 -0.009960 -0.001580 -0.012480 -0.013000 0.006180 -0.025520 0.031000 0.037600 0.015740

0.018660 -0.003190 0.274470 0.506000 0.078040 0.068100 0.087800 0.106200 -0.068550

a26
 Eigth International Conference
On High-Performance Marine Vehicles
Duisburg,Germany
26-28 September 2012

Table 11: SNPL- Series Regression Coefficients

Fn
ai 1.00 1.25 1.50 1.75 2.00 2.25 2.5
a0 0.132000 0.246610 0.312000 0.383000 0.265840 0.462320 0.535910

a1 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

a2 0.000000 0.001790 -0.010900 0.000000 -0.157760 0.000000 0.000000

a3 0.121800 -0.113300 -0.117000 -0.161700 0.277923 -0.180480 -0.189950

a4 0.000000 0.000000 0.000000 -0.012150 1.657930 -0.173220 -0.004773

a5 0.000000 0.021110 0.024387 0.038207 0.070020 0.036020 0.036920

a6 -0.456300 -0.001440 0.043920 0.000000 0.000000 0.000000 0.000000

a7 0.069900 0.000000 0.000000 -0.010200 0.616710 0.000000 0.000000

a8 0.571800 0.000000 0.000000 0.151102 -1.104850 0.325690 0.221180

a9 0.142000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

a10 0.091300 0.000000 -0.007209 0.000000 0.251379 0.249980 0.442020

a11 0.000000 0.000000 -0.040987 0.000000 0.000000 0.000000 -0.784770

a12 0.000000 0.024990 0.000000 0.000000 0.000000 0.000000 0.047540

a13 0.052380 -0.271220 0.000000 0.000000 0.000000 0.024170 0.000000

a14 0.000000 0.000000 -0.006230 0.018770 0.000000 -0.227270 0.000000

a15 0.000000 -0.107380 -0.014560 0.019200 0.000000 0.000000 0.000000

a16 -0.041000 0.000000 0.000000 0.003200 0.000000 0.357990 1.010080

a17 0.020000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.000000 0.168090 0.000000 -0.031070 -0.086020 0.000000 0.000000

a18

Table 12: NPL- Series Regression Coefficients

View publication stats