Passenger ship seakeeping optimization by the overall motion sickness incidence

Ocean Engineering 76 (2014) 8 6 - 9 7

ELSEVIER

Contents lists available at ScienceDirect

Ocean Engineering

journal homepage: www/.elsevier.com/locate/oceaneng

Passenger ship seakeeping optimization by the Overall Motion

Sickness Incidence

A. Scamardella, V. Piscopo *

The University of Napies "Parthenope", Department of Science and Technology, Napoli 80143, Italy

CrossMark

A R T I C L E I N F O Article history:

Received 5 M a r c h 2013 Accepted 1 December 2013 Available online 22 December 2013

Keywords:

Overall M o t i o n Sickness Incidence Parametric h u l l modelling Hull f o r m o p t i m i z a t i o n NPL series

A B S T R A C T

Wellness and comfort onboard are ones of the most important factors in ship design and may be considered the key criterion to achieve the best seakeeping performances for passenger vessels. In this respect it is possible to achieve appreciable seakeeping improvements by only varying several hull form parameters, even if ship main dimensions and displacement have been already fixed. In the paper a new index, namely the Overall Motion Sickness Incidence (OMSI), defined as the mean MSI value on the main deck, is proposed and assumed as parameter to be minimized in a single-objective optimization procedure. Parametric modelling is used to generate several hull forms derived by the NPL systematic series and, despite of classical methods where the optimization procedures are carried out in regular head waves, various heading angles and two operating scenarios are considered in a seaway, described by the JONSWAP Spectrum. The optimum hull is finally generated and relevant vertical accelerations at some critical points on the main deck, as well as heave, pitch and roll speed polar plots, are compared with the parent ones.

© 2013 Elsevier Ltd. AU rights reserved.

1. Introduction

Ship calm and rough water performance prediction is one of the most important concerns for naval architects, already at the earliest design stage. From this point of view seakeeping analysis has been widely applied after the development, in 1955, of the first practical strip theory, mainly based on the evaluation of the hydrodynamic characteristics of various hull sections by Lewis (1929) conformal mapping technique or Frank Close-Fit approach. In this respect any vessel w i t h good seakeeping qualities has to perform i n a seaway the expected mission that, i n turn, depends on ship service and typology. For instance, naval ships seakeeping performances strictly depend on operational ability and use of weapon and sensor systems, while for passenger or cargo vessels wellness and comfort onboard, as well as crew safety, are the main factors to be accounted and optimized. Anyway, even if seakeeping qualities are not the only dominant aspect in the ship design process, it is always possible to achieve considerable improve-ments, even when displacement and main dimensions have been fixed. For this reason seakeeping optimization, in terms of habit-ability, operability and survivhabit-ability, has become a popular research topic for the last three decades.

•Corresponding author. Tel.: + 3 9 081 5476590; fax: + 3 9 081 5476414.

E-mail address: vincenzo.piscopo@uniparthenope.it (V. Piscopo).

0029-8018/$ - see f r o n t matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.Org/10.1016/j.oceaneng.2013.12.005

Bales (1980) optimized a destroyer-type hull form, in head seas and at various speeds, on the basis of analytical predictions, subsequently deriving by some regression formulas correlating relevant performances to form parameters, the optimum hull. Grigoropoulos and Loukalds (1988) developed a numerical method, based on a nonlinear direct search algorithm to minimize RAO peak values in head regular waves. Similar studies have been also carried out by Hearn et al. (1991), who developed an inverse design procedure, based on the optimum hull nonlinear direct search process. Kukner and Sariöz (1995) optimized the seakeep-ing qualities of a high speed vessel, generatseakeep-ing by the Lackenby method (Lackenby, 1950), several derived hulls having different form parameters as regards the parent ones. Peacock et al. (1997) defined a mathematical model based on a multi-objective research algorithm for displacement monohulls. Sariöz and Sariöz (2006) proposed a new optimization procedure, based on a nonlinear problem solved by direct search techniques. Campana et al. (2009) proposed a new optimizarion technique for the heave morion of the S175 containership, adopted by the ITTC Seakeeping Commit-tee as a benchmark test, considering two different oprimization procedures, namely a filled function based algorithm and a Pardcle Swarm Oprimization method. Diez and Peri (2010) presented a new approach for the robust optimization of a bulk carrier conceptual design, subjected to uncertain operating and environ-mental conditions, so extending the standard deterministic for-mulation for design optimization to take into account the uncertainty related to both design variables, operating conditions

A. Scamardella, V. Piscopo / Ocean Engineering 76 (2014) 86-97 87

and computational results of the simulations. Finally Özüm et al. (2011) investigated the seakeeping qualities of fast ships, system-atically varying both main dimensions and hull form parameters. Anyway, in almost all cases, optimization procedures were based on the assumption that the optimum hull is found when the absolute vertical acceleration i n regular head waves due to combined pitch, heave and roll motions, is minimized, instead of analysing statistical responses in a seaway, so neglecting sea spectra and operating scenarios, and consequently reducing com-putational efforts.

In the paper a new index, namely the Overall Motion Sickness Incidence, defined as the mean Motion Sickness Incidence (MSI) value on the main deck, is proposed and chosen as a parameter to be minimized in a single-objective optimization procedure, accounting for both operating scenarios and sea spectra. The proposed method is applied to the hull form optimization of a

Table 1

Parent h u l l main dimensions and foi-m parameters.

Displacement A 2781 t Draft to baseline T 4.00 m Waterline length i w L 100.00 m Waterline beam BvvL 17.00 m Prismatic coefficient Cp 0.700 Block coefficient CB 0.400

IVlidship section coefficient O f 0.662 Waterplane area coefficient Cwp 0.781

LCB f r o m FP ( + v e a f t ) LCB 55.00 %1.WL Vertical centre of buoyancy KB 2.791 m Vertical centre of gravity KG 7.000 m

I

F i g . l . Parent h u l l forms.

passenger vessel, derived by the round bilge NPL systematic series: several alternative hulls have been generated, for fixed displace-ment and Froude number, by the Lackenby method, varying both some hull form parameters and the longitudinal centre of buoy-ancy position. Furthermore, despite of classical optimization procedures, based on RAO peak values minimization in regular head waves, various heading angles have been considered in a seaway, described by a JONSWAP Spectrum for all statistically relevant combinations of significant wave height and zero-crossing periods. The optimum hull is finally derived by means of the Pareto Principle, and relevant vertical accelerations at some critical points on the main deck, as well as heave, roll and speed polar plots, are compared w i t h the parent ones.

2. Seakeeping operability performance assessment

2.1. The Overall Motion Sickness Incidence index

Motion sickness generally indicates discomfort on a moving environment, having the peak of different associated symptoms i n vomiting. It is related w i t h motion perception i n the vestibular system, providing the brain w i t h information about self-motion not in accordance w i t h those ones furnished by visual and/or proprioceptive systems, the first one located in the inner ear, the second one i n the skin, muscles and joints (Pérez Arribas and Lopez, 2007). Sickness is generally caused by these conflicting signals, according to sensory rearrangement (Reason and Brand, 1975) and neural mismatch (Benson, 1999) theories. Following the studies sponsored by the US Navy in the early 1970s, to investigate ship motion effects on humans, the flrst mathematical model for sickness was developed by O'Hanlon and McCauley (1974). A series of experiments were carried out on over 500 subjects, while seated w i t h their heads against a backrest and eyes opened in an enclosed verrically oscillating cabin. During these experiments they were exposed to the effects of 25 combinations of 10

Table 3

Alternative hull forms adimensional parameters f o r fixed Cp=0.700 and L C B = 5 5 ^ . Hull Hull Hull Hull H u l l

1 - C B = 0 . 3 9 0 2 - C B = 0 . 3 9 5 0 - C B = 0 . 4 0 0 3 - C B = 0.405 4 - C B = 0 . 4 1 0

CB 0.390 0.395 0.400 0.405 0.410

C M 0.650 0.656 0.662 0.675 0.681

Cwp 0.779 0.780 0.781 0.783 0.784

Table 2

Alternative hull forms adimensional parameters for fixed C B = 0 . 4 0 0 .

Hull 1 - L C B = 53% Hull 2 - L C B = 53% Hull 0 - L C B = 53% H u l l 3 - L C B = 53% LCB f r o m FP ( + v e f w d ) % U L Cp CM CWP L C B f r o m FP ( + v e fwd)%i:„L Cp CM Cv/p L C B f r o m FP ( + ve fwd)%UvL Cp C M CWP 53% H u l l 1 - L C B = 55% 55% 0.600 0.750 0.734 H u l l 1 - L C B = 5 7 % 57% 0.600 0.750 0.724 53% 0.650 0.700 0.772 Hull 2 - L C B = 5 5 % 55% 0.650 0.707 0.760 H u l l 2 - L C B = 5 7 % 57% 0.650 0.701 0.744 53% 0.700 0.662 0.800 Hull 0 - L C B = 55% 55% 0.700 0.662 0.781 Hull 0 - L C B = 5 7 % 57% 0.700 0.662 0.762 53% 0.750 0.621 0.821 H u l l 3 - L C B = 5 5 % 55% 0.750 0.628 0.801 H u l l 3 - L C B = 57% 57%

88 A. Scamardella, V. Piscopo / Ocean Engineering 76 (2014) 86-97 frequencies (from 0.083 to 0.700 Hz) and various magnitudes

(from 0.278 to 5.500 ms"^ RIVIS) up to two hours and it was found the main siclmess cause is the motion vertical component and the maximum sensitivity occurs at 0.167 Hz. The Motion Sickness Incidence (MSI) is defined as the percentage of passengers who vomit after 2 h of exposure to a certain motion:

M S / = 1 0 0 0.5+ er/ logio(0.798Vm^/g)-/<M5;

0.4 (1)

where the factor /^MSI is defined as follows: / t o = 0.654 + 3 . 6 9 7 ( o g , o ( i - ^ +2.320 ' 0 ^ 1 0

1 Irrü (2) while m2 and are the 2nd and 4th spectral moments of the ship vertical motion spectrum S^{a}e) as a function of the encounter frequency coe, as follows:

m 2 = ƒ a>]Sz(o}e)doie

m4= w'lSzio}e)da)e

Jo

(3)

(4)

(age, gender), to the motion sickness dose value MSDV:

VI = k„,MSDV (5)

The latter parameter depends on weighted RMS vertical accel-eration and exposure time T^:

M S D V = v W J ^ (6) where is evaluated according to the following formula:

/•oo

m4w = / oASz{o]e)G^(a)e)dcOe

Jo

having denoted by C(oje) the frequency weight function, derived by experimental observations and maximum in the range 0.111-0.271 Hz: C(a)e) = 0.111 1 if We < 0.111 Hz if 0.11) <Q>e < 0.271 Hz if We > 0.271 Hz (8)

After the first work's by O'Hanlon and McCauley (1974), Lawther and Griffin (1998) conducted similar studies on car ferries operating in the English Channel and analysed the consequent sickness among passengers. Data were collected for 17 voyages of about 6 hours in duration, involving over 4900 passengers. They obtained results similar to those ones of O'Hanlon and McCauley in the strongest correlation between MSI and vertical acceleration, both in magnitude and duration of exposure, also founding that roll and pitch motions, even i f not provoking sickness in them-selves, when combined w i t h heave, may produce more seasick-ness than predicted by classical models. They proposed a new index, namely the Vomiting Incidence (VI), actually embodied in British Standard 6841 (BSl 1987b) and ISO Draft International Standard 2631-1:1997, proportional by means of a constant km varying in accordance to the exposed population characteristics

1 i . i 5" 9

—-—

—

—

V-Fig. 2. Probable fractions of time at various ship-wave heading angles (2nd scenario).

A. Scamardella. V. Piscopo / Ocean Engineering 76 (2014) 86-97 89

In the proposed optimization procedure, the model by O'Han-lon and McCauley was applied, as it is still nowadays the most cited one, even i f the factor ;(MSI was computed by the formula proposed by Lloyd (1998):

I'MSI.Uyod '- - 0 , 8 1 9 + 2.32 (9)

Besides as the ship vertical acceleration, as well as the MSI, significantly vary along the ship breadth and length (Ascione et al., 2005), a new index, namely the OMSI (Overall Motion Sickness Incidence), defined as the mean MSI over the deck for any assigned sea-state and heading angle, may be introduced:

Wtn.Tz)jn 7 (10) OMSI

Finally, denoting by (x,y,Zdeci<)f the coordinates of the i t h of Nc remote control location points on main deck, OMSI is finally defined as follows for any assigned sea-state and heading angle:

1

0MS/,H„3,r,,,,, = ^ Z MSIiH„,.T.),.,.,.ix.y.z,„,),

2.2. Derived tiull forms generation

(11)

The parent hull for the optimization procedure was chosen between those ones of the NPL systematic series due to both its relatively simple lines and wide variability of hull form para-meters. Main data and hull sections are shown in Table 1 and Fig. 1, respectively.

Derived hulls have been systematically generated, at the same Froude number and displacement, changing by the Lackenby

90 A. Scamardella, V. Piscopo / Ocearx Engineering 76 (2014) 86-97 method, several form parameters, namely the block, prismatic and

midship section coefficients, as well as the longitudinal centre of buoyancy position. The parametric transformarion involves mov-ing the stations fore and aft unchangmov-ing the section shapes, until the required parameters specifications are met. In any case it might be also possible to obtain derived hull forms by indepen-dently varying both sectional area and waterplane curves (Grigoropoulos, 2010; Pérez and Clemente, 2011). A key quality of the adopted transformation is that it maintains the fairness of the hull, described by means of rational B-splines, to a very high degree of accuracy. It is noticed that the parent hull may be reasonably close to the final one, as it is not possible to make major changes to the search parameters before the hull starts becoming unreasonably distorted, which implies that the varia-tions of hull coefficients may be of the order of about 10%. In this respect the method is not intended to be used for gross design modifications. Because of the nonlinear nature of the

transformation and the requirement that fairness have to be maintained, it is clear that long fore and aft overhangs distort proportionally more than the underwater body. The shorter the overhangs of the vessel are, the less noticeable this effect w i l l be. Two different hull form series have been generated: the first one varying the prismatic and midship coefficients, the second one the block coefficient, at same wateriine length and breadth. For the second derived group the immersion was varied accordingly:

T v „ r = ^ T (12)

In Table 2, the first derived hull forms are presented: the prismatic coefficient has been varied i n the range 0.60-0.75 w i t h 0.05 step, for three different values of LCB, namely 53, 55 and 57% of wateriine length from the ship forward perpendicular. Data in bold refer to parent hull.

A. Scamardella, V. Piscopo / Ocean Engineering 76 (2014) 86-97 91

In Table 3 the second derived hull forms group is shown: block coefficient has been varied in the range 0.39-0.41 w i t h 0.005 step, for fixed values of Cp=0.700 and L C B = 5 5 % L w l .

Seakeeping analysis has been performed by a commercial Strip Theory code, suitable for both monohull and catamaran forms. In this respect the software validation documents include some comparisons for the N P L systematic series, between theoretical and experimental RAO curves, the last ones derived by the towing tank measurements carried out at the University of Southampton (IVIolland et al., 1995 and Couser, 1996). At slow speeds (F,, < 0.50) theoretical values show very good correlation w i t h the experi-mental ones, which implies that the results by the strip theory may be considered accurate enough for the analysed vessel (F„=0.32).

Particularly in the analysis the linear strip theory model, based on the work by Salvesen et al. (1970), is used to calculate both the coupled heave and pitch response of the vessel and the roll one,

1

Fig. 6. MSI distribution for 7^=9.5 s. (1) Heading a n g l e = 4 5 d e g ; MSlmax=0.l M S I , „ a x = 0 7 5 8 % . (4) Heading a n g l e = 1 8 0 deg; M S i , „ „ = 0 . 7 9 9 % .

assuming in the latter case that the vessel behaves as a simple, damped, spring/mass system, and the added inertia and damping are constant w i t h frequency. As concerns pitch and roll inertias of the vessel, the pitch gyradius is 25%, while the roll one is 35%. The total roll damping has been set equal to 0.075.

2.3. Operating scenarios and sea spectra

The ship operational effectiveness refers to the vessel ability of accomplishing its mission whatever conditions it may encoun-ter during its life. As previously said, in classical optimization methods, mainly based on generation of ship variants derived by a parent hull, operating scenarios and sea spectra are not defined. The classical rule of thumb, in fact, is based on finding the optimum hull minimizing heave and pitch RAOs, accepting as the worst scenario that one in head regular waves. Anyway, even if regular wave prediction indicates the optimum hull f o r m

3

92 A. Scamardella, V. Piscopo / Ocean Engineering 76 (2014) 86-97

superiority, a more realistic comparison should be based on responses predicted in irregular seas, covering all headings and speeds. In fact, according to Sariöz and Sariöz (2005a, 2005b), the ship seakeeping assessment may take into account all phenom-ena w h i c h might curtail or prevent vessel operational perfor-mances, and so it may not be based on a single sea state, but it is necessary to consider the aggregate sea conditions the vessel may encounter during its life.

Furthermore, it is not always true the worst scenario happens at head seas, as OMSI peak values may be recognized also at transverse headings, depending on wave zero-crossing period, as i t w i l l be subsequently shown. So for a more reliable optimization procedure, it is necessary to account, in the evaluation of the OMSI index, for all heading angles and wave zero-crossing periods, according to the following equation:

OMSI = ^

I

S

Z

having denoted by Nj and N,, the number of considered sea states and heading angles, each one w i t h a certain probability of occurrence pj and p,„ respectively. In the optimization procedure all statistically relevant combinations of significant wave heights and zero-crossing periods, each one w i t h a certain j o i n t prob-ability of occurrence, have been considered for Mediterranean Sea Region. Mediterranean Sea Region starisdcal data, relative to summertime, have been associated to the JONSWAP Spectrum (Piscopia et al., 2002; Benassai et al., 2004). The zero-crossing period has been varied in the range 3.5-9.5 s w i t h 1 s step, which implies the modal period varied between 4.5 s and 12.5 s. The significant wave height was varied in the range 0.5-5.5 m w i t h 1 m step, which implies that the non-dimensional ratio Hi/a/Lwr is comprised between 0.005 and 0.055 m. The analysis has been carried out accounting for headings comprised between 45° and 180° w i t h 5° step, disregarding the range comprised between

Fig. 7. OMSI d i s t r i b u t i o n versus heading angles for d i f f e r e n t values of and H , ; 3 = 2 . 5 m .

0°-45°, to avoid ship manoeuvring problems, with a ship speed of 20 kn. Two different scenarios have been analysed: in the first one it was assumed all heading angles have the same probability of occurrence, while in the second one a specific probability density function has been defined, according to Fig. 2.

In all cases, for each sea state and heading angle, the MSI was evaluated i n correspondence of 125 remote locations points on the main deck, according to Fig. 3. As concerns the computational effort, for each sea spectrum and vessel speed, considering 28 equally spaced headings in the range 45°-180°, the computational time is about 10 min, which implies that the total time amount for each candidate hull is about 90 min, including the

post-3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0

T , [ s ]

—c—H1/3=0-1m —D—H1/3=2-3m —o—H1/3=4-5m —•—H1/3=1-2m —•—H1/3=3-4m — ^ H 1 / 3 = 5 - 6 m

Fig. 8. OMSI distribution versus for d i f f e r e n t values o f significant wave height H , / 3 .

A. Scamardella, V. Piscopo / Ocean Engineering 76 (2014) 86-97 93

In Figs. 4-6 the MSI/lVISImax distribution on parent hull half-deck is shown for different heading angles, namely 45°, 90°, 135° and 180° and zero-crossing periods T^, namely 3.5, 6.5 and 9.5 s with a signiflcant wave height Hi/3 = 1.0m. As i t could be pre-dictable, a great variability is found along the ship breadth and length, when heading angles and zero-crossing periods vary. The IVISI transverse variation is less remarkable for lower Tz values than for higher ones, while the iVISImax values, especially at transverse headings, significantly decrease when Tz increases. The MSI long-itudinal variability is, obviously, negligible in almost all cases at transverse headings, while i t is noticeable at following or head seas, which implies that for a more reliable optimization proce-dure, the OMSI index is suitable as global parameter to be minimized.

It is noticed that horizontal accelerations have not been considered in the evaluation of MSI, as the Froude number, equal to 0.32 for the analysed ship, is quite low. In fact, according to the findings of a recent EU project COMPASS (Turan, 2006), the role of horizontal accelerations for the occurrence of MSI is appreciable onboard high speed vessels (F„ > 0.50), so confirming the results of some experiments carried out in the past on a standard catamaran, a Deep/V monohull and a wave piercing catamaran. A similar conclusion has also been drawn by Tamura and Arima (2006), while investigating the ride comfort of a high-speed passenger craft.

In Fig. 7 parent hull OMSI is plotted versus heading angles for different values of zero-crossing periods. In Fig. 8 OMSI is plotted versus Tz, for different values of significant wave height, founding it is maximum when Tz lies in the range 5-6 s, significandy decreasing for higher and lower values. Similar results, as shown in Fig. 9, may be found when the probability function of Fig. 2 is considered. The preliminary analysis clearly shows the only head sea condition is not sufficient to find the optimum hull and several sea-states, characterized by a certain combination of wave zero-crossing periods and significant wave heights, have to be considered.

Table 4

OMSI f o r the first group o f alternative h u l l forms for fixed C B = 0 . 4 0 0 .

Hull 1 - L C B = 5 3 % Cp=0.600 Hull 2 - L C B = 53% Cp=0.650 Hull 0 - L C B = 5 3 % Cp=0.700 H u l l 3-- L C B = L C B = 5 3 % Cp = 0 . 7 5 0

OMSI%

—

Hull 1- L C B = 5 5 % Cp=0.600 Hull 2- L C B = 5 5 % Cp=0.650 Hull 0- LCB = 55% C p = 0 . 7 0 0 H u l l 3- L C B = L C B = 5 5 % Cp^ =0.750

OMSI% 8.039 8.303 8.479 8.355

Hull 1- L C B = 5 7 % Cp=0.600 Hull 2- L C B = 57% Cp=0.650 H u l l 0- L C B = 5 7 % Cp=0.700 H u l l 3- L C B = L C B = 5 7 % Cp= =0.750

OMSI% 8.118 8.191 8.001

-Table 5

OMSI f o r the second group o f alternative hull forms f o r fixed Cp=0.700 and L C B = 55%LVVL.

Hull 1 - C B = 0 . 3 9 0 Hull 2 - C B = 0 . 3 9 5 Hull 0 - C a = 0 . 4 a o Hull 3 - C f l = 0 . 4 0 5 Hull 4 - C f l = 0.410

OMSI% 8.316 8.403 8.479 8.576 8.656

Table 6

Weighted OMSI f o r the first group o f alternative hull forms f o r fixed C s = 0 . 4 0 0 - e f f e c t i v e heading angle d i s t r i b u t i o n .

Hull 1 - L C B = 53% Cp=0.600 Hull 2 - L C B = 53% Cp=0.650 H u l l 0 - L C B = 5 3 % Cp=0.700 H u l l 3 -- L C B = L C B = 5 3 % Cp = 0 . 7 5 0

OMSI% _ 4.304 4.203 4.051

Hull 1 - L C B = 55% Cp=0.600 Hull 2 - L C B = 5 5 % Cp=0.650 Hull 0 - L C B = 55% Cp=0.700 H u l l 3 - L C B = L C B = 5 5 % Cp = 0 . 7 5 0

OMSI% 4.503 4.411 4.255 3.964

Hull 1 - L C B = 5 7 % Cp= 0.600 Hull 2 - L C B = 57% Cp=0.650 Hull 0 - L C B = 5 7 % Cp=0.700 H u l l 3 - L C B = L C B = 57% Cp: =0.750

OMSI% 4.769 4.577 4.202 _

processing phase, carried out by a dedicated programme devel-oped in MATLAB MathWorks.

3. Seakeeping optimization assessment

3.1. Influence of zero-crossing period and heading angles

It is well known that onboard passenger can occupy good places, generally near the ship centre of gravity, and not so good places (Esteban et al., 2005), as combined vertical accelerations generally increase going to bow or stern, or towards ship sides, depending on heading angles and zero-crossing periods.

3.0 3.5 4.0 4.5 6.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0

T , [ s ]

—0—H1/3=0-1 m —o—H1/3=2-3 m —o—H1/3=4-5m

—•—H1/3=1-2m — ^ H 1 / 3 = 3 - 4 m — ^ H 1 / 3 = 5 - 6 m Fig. 9. Weighted OMSI d i s t r i b u t i o n versus Tz f o r d i f f e r e n t values of the signiflcant wave height Hi;3.

94 A. Scamardella. V. Piscopo / Ocean Engineering 76 (2014) 86-97

Table 7

W e i g h t e d OMSI f o r the second group of alternative h u l l forms f o r f i x e d Cp=0.700 and LCB=55%LwL-effective heading angle distribution.

H u i n - C B = 0 . 3 9 0 H u l l 2 - C 6 = 0.395 Hull 0 - C B = 0 . 4 0 0 H u l l 3 - C B = 0 . 4 0 5 H u l l 4 - C B = 0 . 4 1 0

0MS1% 4.244 4.249 4.255 4.254 4.253

Table 8

-Parent and o p t i m u m hull main dimensions.

Parent h u l l Optimum hull

Displacement A 2781 2781 f D r a f t to baseline T 4.00 4.00 m Waterline length LwL 100.00 100.00 m Waterline beam BwL 17.00 17.00 m Prismatic coefficient Cp 0.700 0.710 Block coefficient CB 0.400 0.400 M i d s h i p section coefficient CM 0.662 0.657

Waterplane area coefficient CWP 0.781 0.766

%iwL

L C B f r o m FP ( + v e a f t ) L C B 55.00 57.00 %iwL

Overall Motion Sickness Incidence (1st scenario) OMSI 8.479 7.937 %

Overall Motion Sickness Incidence (2nd scenario] OMSI 4.255 4.124 %

I

Fig. 10. Parent and o p t i m u m h u l l f o r m s .

3.2. Changing of Cp, LCB and Cg

As previously said, in the first operating scenario it was assumed all heading angles have the same probability of occur-rence. Tables 4 and 5 show the OMSI values for both derived hull f o r m groups. Tables 6 and 7 show, for the second operating scenario, the same results. In both cases it seems the optimum hull may be found shifting after the longitudinal centre of buoy-ancy, increasing the prismatic coefficient, decreasing the block and midship section ones, obviously paying attention to both max-i m u m allowable trmax-im and unwanted bare hull resmax-istance max-increase. 3.3. Optimum hull genemtion

After the investigation of changing different hull form para-meters, i t was found the optimum hull may be generated by the following steps:

1. decrease the block coefficient, for fixed wateriine length, breadth and displacement, increasing the immersion according to Eq. (12);

2. shift after the centre of buoyancy, as far as possible, depending on both equilibrium and trim considerations;

3. increase the prismatic coefficient, decrease the midship section and waterplane area ones.

It is noticed that in all cases these variation cannot penalize the bare hull resistance. Table 8 shows parent and optimum hull main data, while i n Fig. 10 optimum hull forms are compared w i t h

< — • . y -'sS ' N I 'N _{\ \} \ • / '

/ : r

Fig. 11. Sectional area d i s t n b u d o n for parent (black line) and o p t i m i z e d (red line)

hull forms. (For interpretation o f the references to color i n this figure legend, the reader is referred to the web version of this article.)

Table 9

Remote location points w h e r e vertical accelerations have been evaluated.

Description Units 1 2 3 4 5 6

Longitudinal position ( - f v e f w d ) m - 4 5 . 0 0.0 30.0 - 4 5 . 0 0.0 30.0 Offset f r o m centreline m 0.0 0.0 0.0 7.5 7.5 7.5 Vertical position m 9.8 9.8 9.8 9.8 9.8 9.8

parent ones. Black and red lines refer to parent and optimum hulls, respectively.

Some differences between hull forms arise, as i t could be predictable. The optimum hull, in fact, has more V-shaped sections in the forebody, while the longitudinal centre of buoyancy is shifted about 2 % L w l afterward. The relevant OMSI is equal to 7.937% w i t h 6.4% reduction as regards the parent hull value, while for the second scenario i t is equal to 4.124% w i t h 3.0% reduction, which implies that slight hull form variations may produce appreciable seakeeping improvements, without altering the bare hull resistance. In fact, by the slender body method based on the works of Couser et al. (1998), the wave coefficients at i ' = 2 0 k n are 0.669 X 10"^ and 0.663 x 10"^ for parent and o p t i m u m hulls.

A. Scamardella, V. Piscopo / Ocean Engineering 76 (2014) 86-97 95 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 fe [Hz] 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 f. [Hz] 0.2 0.3 0.4 0.5 0.6 felHz] 0.3 0.4 0.5 0.6 f . t H z ] 0.7 0.8 0.9 1.0 0.0 0.1 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 f. [Hz]

d

^

_—

Fig. 13. (a) Vertical accelerations (4-5-6)-Heading a n g l e = 4 5 deg. (b) Vertical accelerations ( 4 - 5 - 6 ) - H e a d i n g angle

A. Scamardella. V. Piscopo / Ocean Engineering 76 (2014) 86-97 97

respectively. In Fig. 11 sectional area curves for both parent and optimum hulls are shown: an appreciable redistribution of sub-merged area has been found, w i t h a decrease of the forward volume and a consequent increase of the afterward one. The adopted reference system has the origin at amidships.

Besides, in order to compare parent and optimum hulls in terms of vertical accelerations, six remote control points on the ship main deck have been chosen. Table 9 shows the relevant coordinates, respect to a reference system having the origin in correspondence of the baseline at amidships.

The first three remote location points are located at centreline, while the other three ones towards the ship side. In Fig. 12a-d the R.M. S. of vertical acceleration at remote control points 1-2-3 is plotted versus the encounter frequency/., in the range between 0 and 1 Hz, w i t h 0.1 Hz step. Similarly Fig. 13a-d show the same results at remote control points 4-5-6. Relevant results, referring to a seaway described by a JONSWAP Spectrum with H , /3 = 1.0 m and 7^=6.5 s, suggest that the vertical acceleration components due to heave and roll motions are consistently lower for the optimum hull than for the parent one, while the component due to pitch motion is substantially unchanged. The most noticeable reduction of vertical acceleration is recognized in the heading range from beam to head seas, while only at quartering sea a slight incrementis found, b u t i t doesn't penalize the OMSI due to the low values of relevant accelerations. In all cases the curves show a peak between 0.10 and 0.20 Hz, which is consistent with the peak frequency equal to 0.15 Hz of the JONSWAP spectrum. Finally in Fig. 14a-f heave, roll and pitch speed polar plots are shown for parent and optimum hulls respectively for H 1 / 3 = 1.0 m and 7^= 5.5 s.

In any case it seems that it is not an easy task to estimate the real improvements, in terms of comfort onboard, only by compar-ing the vertical acceleration curves of parent and optimum hulls at some critical points. In this respect the OMSI seems to be a more suitable index to estimate the effective improvements in terms of comfort onboard.

4. Conclusions

In the paper a new index, namely the Overall Motion Sickness Incidence, defined as the mean MSI value on the ship main deck, was proposed and chosen as parameter to be minimized in a single-objective optimization procedure, to improve the wellness onboard of passenger ships. Despite of classical optimization techniques, where seakeeping performances are optimized by minimizing pitch and heave RAOs peak values in regular head waves, the OMSI index permits to account for both heading angles in a seaway and operating scenarios.

The NPL parent hull has been chosen as test case and two groups of alternative hull forms have been generated, the first one varying the prismatic coefficient and longitudinal centre of buoy-ancy position, the second one the block coefficient. The performed analysis clearly shows that both heading angles and operaring scenarios may not be neglected i n a more reliable optimization procedure, as well as the assumed seaway. The proposed method may also be applied to Motion Induced Interruptions and extended to multi-objective optimization analyses.

Acloiowledgements

The work has been financed by the Department of Science and Technology of the University of Naples "Parthenope", under the Research Project PROGETTO INSIST "Innovazione Tecnologica nei Sistemi di Trasporto" POR Campania FSE 2007/2013 CUP B25B09000040007 - Research Stream: "Ottimizzazione della tenuta al mare di mezzi navali ai fini del miglioramento

del Comfort e della Sicurezza dei passeggeri e del personale imbarcato".

References

Ascione I.. Benassai C , Pozzi S., 2005. Wave spectral hindicasting i n the gulf o f Naples for m a r i t i m e operation management. I n : Proceedings o f International Conference on M a r i t i m e Transportation ICMRT 2005. Ischia, September 2005, vol. I , pp. 506-516.

Bales N.K., 1980. Optimizing the seakeeping performance o f destroyer-type hulls. In: T h i r t e e n t h ONR Symposium on Naval Hydrodynamics, Tokyo, 1980. Benassai, C , Chianese, E., Sansone, E., 2004. Wave spectral evaluation on the

coasdines of southern Italy. Coastal Env. I, 301-310,

Benson, A.J.. 1999. M o t i o n Sickness. Aviation Medicine, t h i r d ed. B u t t e r w o r t h -Heinemann, Oxford, pp. 4 5 5 - 4 7 1 .

British Standard 6841:1987. Guide to Measurement and Evaluation o f Human Exposure to W h o l e - b o d y Mechanical Vibration and Repeated Shock. Campana, E,R, Liuzzi, C., Lucidi, S., Peri, D., Piccialli, V., Pinto. A., 2009. New global

optimization methods for ship design problems. Eng, Optim. 10 (4), 533-555. Couser P.R., 1996, A n Investigation of the Performance o f High-speed Catamarans i n Calm Eater and Waves, PhD Thesis. University o f Southampton. March 1996. Couser, PR., Wellicome, J.F.. Molland, A.F„ 1998. A n improved method for the

theoretical prediction o f the wave resistance o f transom-stern hulls using a slender body approach. Int. Shipbuild. Prog. 45 (444), 1998.

Diez, M., Peri, D„ 2010. Robust optimization for ship conceptual design. Ocean Eng. 37 (2010), 9 6 6 - 9 7 7

Esteban, S., Giron - Sierra, J.M., De - Andres Toro, B., De La Cruz, J„ Riola, J.M., 2005. Fast ships models for seakeeping i m p r o v e m e n t studies using flaps and t - f o i l . M a t h . Comput. Modell. 41 (2005), 1-24.

Grigoropoulos, G.J., Loukakis, T.A.. 1988. A n e w method for developing h u l l forms w i t h superior seakeeping qualities. In: Proceedings o f CADMO 88, Southampton.

Grigoropoulos, G.J., 2010. Hull f o r m optimizarion in calm and rough water. CompuL-Aided Des. 42 (11), 977-984. (2010).

Hearn, G.E., Hills. W., Sariöz, K„ 1991,Practical seakeeping for design: a ship shape approach. Trans. R. Inst. Nav. Archit..

lSO-2631/1,1997 Evaluation of Human Exposure to W h o l e - b o d y Vibration. Part 1, second ed. 1997-05-01.

Kukner, A., Sariöz, IC, 1995. High speed hull f o r m oprimisation f o r seakeeping. Adv. Eng. Software 22, 179-189.

Lackenby, H., 1950. On the systemaric geometrical variarion o f ship forms. Trans. INA, 92.

Lawther, A., G r i f f i n , M.J., 1998. Motion sickness and m o r i o n characterisrics o f vessels at sea. Ergonomics 3 1 , 10.

Lewis, F.M., 1929. The inertia o f the water surrounding a v i b r a t i n g ship. Trans. SNAME 37. 1-20.

Lloyd. A.R.J.M., 1998. Seakeeping Ship Behaviour in Rough Water. A.R.J.M. Llyod, Gosporth, Hampshire p, 1998.

M o l l a n d A.E, Wellicome J.R, Couser RR., 1995. Experimental Measurements o f t h e Seakeeping Characteristics o f Fast Displacement Catamarans i n Long-crested Head-seas. University o f Southampton, Ship Science Report No. 89, December 1995.

O'Hanlon. J.R, McCauley, M,E., 1974. M o t i o n sickness as a f u n c d o n o f the frequency and accelerarion o f vertical sinusoidal m o t i o n . Aerosp. M e d . 45. 3 6 6 - 3 6 9 . Oziim, S., § e n e r , B.. Yilmaz, H„ 2011. A parametric study on seakeeping assessment

of fast ships In conceptual design stage. Ocean Eng. 38, 1439-1447. Peacock, D„ Smith, W.R, Pal, RK.. 1997 M i n i m a l ship m o r i o n h u l l - f o r m design f o r

high speed using multi-criteria oprimizarion techniques. I n : Proceedings o f Fourth International Conference on Fast Sea Transportation (FAST '97). vol. 2. Sydney, Australia.

Pérez, R, Clemente, J.A., 2011, Constrained design o f simple ship hulls w i t h B-spline surfaces. Comput.-Aided Des. 43 (2011), 1829-1840.

P é r e z Arribas, R L . Lopez, Pineiro, 2007. Seasickness prediction i n passenger ships at the design stage. Ocean Eng. 34 (2007), 2 0 8 6 - 2 0 9 2 .

Piscopia R., Inghilesi R., Panizzo A., Corsini S., Franco L , 2002, Analysis o f 12-year wave measurements by the Italian Wave N e t w o r k . In: Proceedings of 2 8 t h Internadonal Conference on Coastal Engineering, Cardiff, July 2002, 121-133. Reason, J.T., Brand, J.J., 1975. M o t i o n Sickness. Academic Press, L o n d o n .

Salvesen, N., Tuck, E,0., Faltinsen, O.M.. 1970. Ship motions and sea loads. Trans. SNAME 78, 250-287,

Sarioz, K., Sariöz, E.. 2005a. Effect of criteria on seakeeping performance assess-ment. Ocean Eng. 32 (10), 1161-1173.

Sariöz. K., Sariöz, E., 2005b. Habitability assessment o f passenger vessels based o n ISO criteria. Mar. Technol. 42 (1), 4 3 - 5 1 .

Sariöz, K., Sariöz, E., 2006. Practical seakeeping performance measures for high speed displacement vessels. Nav. Eng.). 118 (4), 2 3 - 3 6 .

Tamura, Y , Arima, M . (2006). Measurement and analysis o f ship m o r i o n o f a h i g h -speed passenger craft. I n : Proceedings o f the Sixteenth International Offshore and Polar Engineering Conference, May-June 2006, San Francisco, California, USA. Internadonal Society of Offshore and Polar Engineers (ISOPE) 4, 514-519, Turan, 0., 2006. A Rational Approach for Reduction o f M o t i o n Sickness

and Improvement of Passenger Comfort and Safety i n Sea Transportarion (COMPASS). COMPASS Project Technical Report, Universities o f Glasgow and Strathclyde, Glasgow.