Large-scale LNG sloshing model tests

Date March 2009

Author Bunnik, Tim and René Huljsmans Address Delft University of Technology

Ship Hydromechanics Laboratory

Mekelweg 2, 2626 CD Deift

TU Do Ift

Deift University of Technology

Large-scale LNG Sloshing Model Tests

by

Tim Bunnik and René Huijsmans

Report No. 1639-P

2009

International Journal of Offshore and Polar Engineering, volume ig, Nr. 1, ISOPE, ISSN:

1053-5381,

March 2009

Tiansactions of The ISQPE

VOLUME 19' NUMBER 1

ISOPE

MARCH 2009

International

Deift University of Technology

Ship HydromechafliCS laboratory

Library

I

_{Mekelweg. 2}

26282 CD Deift

Phone: +31 (0)15 2786873

EmaII: p.wdeheer@tudeIft.nI

Journal' of

Offshore and

Polar Engineerin

ENGINEERING

OCEAN

OFFSHORE

MARINE

ARCTIC

ANTARCTIC 'ENVIRONMENT MECHANICS MATERIALS'

ENERGY/RESOURCES

Published quarterly as its principal periodical

by The International Society of Offshore and Polar Engineers

ISOPE

INTERNATIONAL JOURNAL OF OFFSHORE AND POLAR ENGINEERING

Vol. 19

No. 1

MARCH 2009

THE EDITORIAL BOARD

Editors: un s Chung (ISOPE), J Wardenier (Delft Univ of Tech, The Netherlands), R M W Frederking (Nat Research Council,

Canada), W Koterayama (Kyushu Uñiv, Japan)

Associate Editors: H 0 Brandes (Univ of Hawaii, USA), J Chaplin (Univ of Southampton, UK), A H Clement (Ecole Centrale

de Nantes, France), M Fujikubo (Osaka Univ, Japan), S Grilli (Univ of Rhode Island, USA), R E Hobbs (Imperial College of Sci and Tech, UK), M Kashiwagi (Osaka Univ, Japan), I Langen (Univ of Stavanger, Norway), T Matsui (Fukui Univ, Japan), AJNA Sarmento (Instituto Superior Tecnico, Portugal), V A Squire (Univ of Otago, New Zealand), M S Triantafyllou (M.I.T., USA).

Managing Editor: Paulette Whitcomb

ISOPE Board of Directors: D Angelides, R Ayer, M Isaacson, R H Knapp, W Koterayama, S Naito, H [Park, AJNA Sarmento President: W Koterayama (2008-2009), R H Knapp (2008-2009)

Executive Director: un s Chung

CONTENTS

Nonlinear Response

of

Offshore Structures to High Seas

1

Cheung H. KiAi

Large-scale LNG Sloshing Model Tests

8

Im Bunnik and René Huijsnans

Observation

of

Coherent Turbulent Structure Under Breaking Wave ...15

Zhi-Cheng Huang, Shih-Chun Hsiao and Hwung-Hweng Uwung

Dynamic Response of Oscillating Flexible Risers Under Lock-in Events ...23

Carlos Alberto Riveros, Tomoaki Utsunomiya, Katsuya Maeda and Kazuaki Itoh

Status Identification and Optimum Adjustment of Performance of Moored Floating Breakwaters ...31

Eva Loukogeorgaki and Demos C. Angelides

Submerged Plate Breakwater Composed of Horizontal Porous Plate and Slightly Inclined Solid Plate

42 S. T Kee

Roll Response of Various Hull Sectional Shapes Using a Navier-Stokes Solver

46

Yi-Hsiang Vu and Spyms A. Kinnas

A Preliminary Test on Agent-based Docking System for Autonomous Underwater Vehicles

52

Son-Cheol Vu

Short Cluster Airgun Array for Shallow to Deep Crustal Survey ...

60

Satoshi Shimizu, Kaoru Tsukuda, Hidenori Shibata, Ayumi Mizota and Seiichi Mitira

Disk-type Underwater Glider for Virtual Mooring and Field Experiment

66 Masahiko Nakamura, Wataru Koterayama, Masaru Inada, Kenji Marubayashi, Takashi Hyodo, Hiroshi Yoshi,nura and

Yasuhiro Morii

Automatic Measurement

of

Dissolved Inorganic Nitrogen Ions in Coastal Field Using Simplified

Flow Injection Method

71 Rei Arai, Koji Tada, Naoki Nakatani, Taketoshi Okuno, Koichi Ohta

Solid-Liquid Flow Experiment with Real and Artificial Manganese Nodules in Flexible Hoses

77

Chi-Ho Yoon, Yong-C'han Park, Jon g,nyung Park, Young-Ju Kim, Jun g-Seok Kang and Seok-Ki Kwon

Proceedings of ISOPE-2008 VancouverConference 22 Call for PapersISOPE-20l0 Beijing 30

ISOPE Membership 41 Call for PapersOMS-2009 Chennai 59

Proceedings of 7th OMS-2007 Lisbon Symposium 65 Proceedings of 8th PACOMS-2008 Symposium 76 Information for Authors inside back cover Publications Order Information back cover and page 80 Copyright ©2009 The International Society of Offshore and Polar Engineers. All rights reserved. This Journal is registered with the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, USA. For multiple copying for promotional or commercial purposes such as creating new collective works or resale, written permission from ISOPE is required.

INDEXED by Science Citation Index Expanded and Engineering Index.

The International Journal of Offshore and Polar Engineering (IJOPE) is published quarterly (March, June, September, December) by The International Society of Offshore and Polar Engineers, P.O. Box 189, Cupertino, California 950l50l89 USA. CHANGES

OF ADDRESS must be received at ISOPE six weeks before they are to be effective. Please send old labels and new address.

SUBSCRIPTION PRICE: complimentary to members; US$150 annually ($60 a single issue) to nonmembers.

The Society, its directors and officers, the Journal; and members of the Editorial Board assume no responsibility for the statements expressed by this journal's contributors or authors. Individual readers, and nonprofit libraries acting for them, are permitted to make a single copy of an article without charge for use in research or teaching. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. For such uses, written permission is required; write to ISOPE (Attention: IJOPE)1 P.O. Box 189, Cupertino, California 95015-0189 USA. international Standard Serial Number (ISSN): 1053-5381

International Journal of Offshore and Polar Engineering (ISSN 1053-5381) Copyright © by The International Society of Offshore and Polar Engineers Vol. 19, No. 1, March 2009, pp. 8-14

Large-scale LNG Sloshing Model Tests

Tim Bunnik

Mann, Wageningen, The Netherlands

René Huijsmans

Technical University of Deift, Deift, The Netherlands

The motionof the LNG fluid inside gas carriers'is'normally restricted by the loading condition of the vessel,l;e. the vessel is operated at either near empty condition orat fully loaded condition. In this way, the resonance or sloshing effects or the fluid on the ship's hull are limited. However, nowadays the LNG carriers are considered to be operating at Intermediate loading conditions as well. Subsequently they will 'be sailing with partially filled LNG tanks. In this condition the LNG fluid is more likely to be induced into resonance due to wave action and roll motions. This resonanceor sloshing behavior of the LNG fluid will lead to hlgh.impact pressureson the thermally insolated ship's hull. Due to the different physical properties of the LNG fluid with respect to water In terms of density and viscosity, little Is known of the behavior of the LNG fluid In resonance condition. Day-to-day practice In retrieving sloshing loading data is based on relatively small-scale model tests, e.g. 1:20 or 1:30. A disadvantage here is that, because any air pressure effects are nOt modeled, hydroelastic phenomena cannot be modeled at this scale. A way around this.is to reduce the ambient air pressure when running scaled model teals. This by itself is not trivial to do. In this paper, however, we will describe a study of model test experiments on a large-scale 2-dimensional section (large-scale 1:10) of an LNG carrier in various loading conditions without depressurization. Using high-speed video observations the wave front formed by the bore of the LNG in resonance Is related to measured impacts on the tank hull. Also measured Is loading on a hydroelastic panel as part of the hull, with the correctly scaled structural properties. Significant influence of the stiffnesson the pressure pulse was observed.

INTRODUCTION

Sloshing is a phenomenon of great engineering importance in the fields of naval architecture, ocean engineering and civil

engi-neering. Severe sloshing can occur in a large oil storage tank, a

reservoir and a fuel tank. Especially, an excessive sloshing motion

in an LNG tanker can rupture the pipeline in a tank and the

tank itself. The results of several research programs to investi-gate sloshing in Liquid Natural Gas (LNG) carriers are presented in Abramson et al. (1974). In the study, the history of sloshing-related problems in LNG carriers is discussed, including a list of recorded tank damages for LNG sloshing when the filling height is low and high relative to the tank length. In both cases, impact loads occur and induce extremely high pressures. Classification societies, containment system designers and ship operators have conducted thorough studies of these damages. In every instance the sloshing of the cargo was identified as the cause of the dam-age (Shin et al., 2003). Simple but effective plans were proposed to counter the sloshing impact in the fully loaded condition: The height of the chamfer at the topside was increased and the insula-tion box at the tank top was reinforced to withstand the sloshing impact.

There is a considerable number of investigations on the slosh-ing problem, both numerically as well as experimentally, the early

ones from Chester (1968) and Chester and Bones (1968). 'For the case of small-amplitude excited motions, Faltinsen (2002),

Received July 15, 2007; revised manuscript received by the editors May 6, 2008. The original version (prior to the final revised manuscript) was'presented at the 17th International Offshore and Polar Engineering Conference (ISOPE-2007), Lisbon, July 1-6, 2007.

KEY WORDS: Large-scale oscillation tests, sloshing.

Huijsmans et al. (2004) and Yamamoto et al. (1995) showed results of oscillation experiments where the computed impact

pressures were compared with measured results.

A very steep wave front such as a hydraulic jump has been

observed in experiments. The bore traveled back and forth

between the tank walls (Hill, 2003). Many applications are given to the 2-D sloshing problem while nowadays 3-D geometries can be modeled as well. The violent sloshing problem is determined through a highly nonlinear free-surface motion. In these gravity-driven flows, viscosity effects generally play a minor role, but in the case of liquid LNG, the fluid properties are not so clear. The top layer of the LNG fluid consists of liquid LNG with a rich con-tent of gas bubbles. So the density viscosity and vapour pressure may play a vital role. The presence of bubbles makes the poten-tial flow-type of modeling inadequate. To overcome these kinds of deficiencies of potential flow solvers, volume of fluid (VOF) solvers are used nowadays (Wemmenhove, 2005). The LNG tank

in this study is a closed tank top. In this paper the main results

presented deal With the measurement of the impact pressures and

wave heights due to a rolling motion of the tank from

experi-ments at model scale. Besides the VOF method, Smooth Particle Hydrodynamics (SPH) are applied as well (Nam and Kim, 2006). Hydroelastic effects on cargo tankshave been studied'by Lee et al. (1999) and Xiong et al (2006). When the waves are overturning

and hitting the water surface, air bubbles may be present in the fluid. In this case, a direct numerical 'solution based on

poten-tial flow with the nonlinear free surface conditions would break down.

Wemmenhove (2007) is working on an extension of the VOF method for a second (air) phase, including compressibility. The test results are being used to validate this method.

Fig. I LNG containment system

SLOSHING TESTS

Test Setup and Instrumentation

A 2-D slice of an LNG containment system was build at scale 1:10. The reason for this choice of scale was to try to have as real-istic as possible interaction effects between gas and fluid. Going

to an even larger scale was practically not possible due to the limitations in the dimensions and weight of the tank. The con-tainment system was based on the type of tank used in a No96 LNG carrier (prismatic tank). Fig 2 and Table 1 give the main

dimensions of the containment system (model scale values). The objective of the tests was to provide validation material for a 2-phase VOF method (Wemmenhove, 2005). In this, only rigid

geometries can be modeled, so the model was built as rigid as

possible to avoid interaction effects between the impact forces and

the structural response of the containment system. This system

was made of stainless steel and supported by stiffeners. The walls

were made smooth; no insulation was modeled. The front and

backside of the containment system were made from thick perspex

Fig. 2 Main dimensions of LNG containment system (model

scale)

Table 1 Main dimensions of LNG containment system (model scale)

to allow visualization of the fluid flow during the tests. Fig. 3

shows the containment system under construction in the model workshop.

Hydroelastic effects were modeled by placing 6 measurement panels in the sidewalls containing pressure transducers. Two types

of measurement panels were modeled: Stiff panels with

mini-mal hydroelastic effects, and flexible panels to study hydroelastic effects. Fig. 4 shows the details of one of the stiff measurement panels, which has a natural frequency of approximately 3000 Hz. The flexible panel was built so that it had a natural frequency of

approximately 400 Hz.

The containment system was oscillated on a large oscillator

capable of generating combined sway and roll motions of the con-tainment system. Fig. 5 shows the concon-tainment system mounted on the oscillator.

The tests were carried out with water and air at atmospheric

pressure. During the tests the following items were measured: Pressures in the panels.

Water height in the containment system (12 locations). Motions and accelerations of the containment system.

A data acquisition system was connected to the containment

system measuring at high frequency (10 kl-lz) to capture also the short (fast) impacts. Further, video recordings were made with a

high-speed camera operating at 100 or 200 Hz. The recordings

were synchronized with the other measurements by measuring the start pulse of the camera. Digital camera recordings were made of the total setup.

1rr

4

'4-7

Fig. 3 LNG containment system model under construction in

model workshop

International Journal of Offshore and Polar Engineering, Vol. 19, No. 1, March 2009, pp. 8-14 9

H 2.697 m BI 1.948 m Cl 0.438 m C2 0.838 m B2 1.948 m B3 1.948 m B4 1.948 m

10

Fig. 4 Details of measurement panel with pressure transducers

Large-scale LNG Sloshing Model Tests

Test Program

The test program consisted of the following variations: Filling rates of 10%, 25%, 70% and 95%.

Regular and irregular motions.

Sway only, roll only and combined sway and roll motions. Depending on the filling rate, different measurement panels were instrumented and the high-speed camera was focused on a differ-ent part of the containmdiffer-ent system. The high-speed camera was connected to the model.

The roll motions of the LNG carrier were obtained using

the-oretical roll and sway Response Amplitude Operators valid for

beam seas. The following (full-scale) input spectra were used:

The time traces of the sway and roll motions were generated with

a random phase model and input to the control software of the

oscillator. Results

Unless indicated otherwise, the results discussed here are

related to the stiff panel and for nondepressurised conditions. 10% filling rate. The fluid flow at a 10% filling rate is charac-terized by:

Bores that develop in the containment system due to shallow water effects.

Air captured during impacts on the vertical sidewalls.

Fig. 6 shows a snapshot of a digital camera recording of a

typical bore, using an irregular motion.

Fig. 7 shows a high-speed camera image of the flow at 10%

filling rate just prior to impact, with entrapped air.

The results show that for regular motion tests, there is a large spreading in the measured peak pressures. In these tests, the con-tainment system was oscillated with a periodic motion, with

sub-sequent periods showing no visual difference in the measured motions. Fig. 8 shows the measured sway motion (10 periodic

events plotted on top of each other). Fig. 9 shows the measured impact pressure in these 10 successive events.

Fig. 5 Empty containment system mounted on oscillator Fig. 6 Bore at 10% filling rate, irregular motions Hs (m) Tp (s) 3.0 10.6 8.1 10.6 11.1 13.4 12.1 16.2 I

I

E C 0 0 E >., U) 0.4 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 0.4o _0.5

Fig. 7 Air captured during impact

Due to inertial effects, the imported motions were not

com-pletely sinusoidal.

It can be seen that the peak pressure varies between 100 and

550 hPa. The total test lasted 20 mm, meaning that approximately 400 of these impacts were generated. These peak pressures were collected and used to generate a probability density function, as shown in Fig. 10. This shows again the large amount of spreading in the peak pressure.

Instead of looking only at the peak pressures, it is also possible to look at the pressure integrated over the duration of the impact,

as shown in Fig. 11. This is also a more relevant parameter for

the structural response of the containment system. Fig. 12 shows the probability density function of the integrated impact pressure.

The spreading in this parameter is much smaller.

This test was repeated (identical motions), but the stiff panel

in the containment system sidewall was replaced with a flexible one (400 Hz natural frequency instead of 3000 Hz). Figs. 13 and

1.5 2

time [s]

2.5 3

Fig. 8 Measured sway motion: 10 successive periodic events plot-ted on top of each other

600 500 400 300 E 200 1 100 0

Fig. 9 Spreading in peak pressure in 10 successive regular

impacts test(t13OOU1 UW 0.14

I 0.12

C 0

t 0.1

_C '0.08 U) C a) D _0.06 .0 ca 0.04 .0 2 - 0.02 120 100 80 60

0

0 40 20 '?32 0

Fig. 11 Integrated pressure pulse

32.5 33

time [s]

33;5 International Journal of Offshore and Polar Engineering, Vol. 19, No. 1, March 2009, pp. 8-14 11

100 200 300 400

pressure [hPa]

Fig. 10 Probability density function of peak pressure P07 in reg-ular sloshing test, 10% filling rate, stiff panel

14 show the probability density function of the peak pressure and the integrated impact pressure.

The peak pressure distribution is very similar to the distribution derived from the test with the stiff panel. However, the integrated pressure pulse is significantly lower (about 10%).

25% filling rate. At a 25% filling rate; the water level is such

that shallow water effects are less noticeable. The mean water

test

37830001 002

325 330 335 300 305 310 315 320

0.6 C

0

t 0.5

C 0.4 U) C m 0.3 .0 .0

2

. 0.1 0.1 C 0 0 0.08

:-

0.04 (5 .0 0 0.02

Fig. 13 Probability density function of peak pressure P07 in reg-ular sloshing test, 10% filling rate, flexible panel

0.6 C

0

0.7

2

0.1 0.12 0.7

t 0.5

C

.' 0.4

U) m 0 0

test 37830 001 002

test 37830 003 002

test 37820 003 002

5 10 15 20

pressure pulse [hPa s]

Fig. 14 Probability density function of integrated pressure pulse P07 in regular sloshing test, 10% filling rate, flexible panel

ii..eIIh

(UUUUI

II. US.!!I 1 1.8USU

I'11

Fig. 15 High-speed camera images at 25% filling rate-photo

12286 (left) and photo 12291 (right)

level is such that the oblique lower corners of the containment system are fully submerged Thus, the impacts are now directly

hitting the vertical wall of the containment system, causing high impacts and significant runup, sometimes even up to the ceiling of the containment system. Fig. 15 shows high-speed camera images just prior to impact, and of a typical, relatively severe impact on the container system wall. The red arrow indicates the position of a pressure measurement. There is 0.05 s betWeen the 2 images.

Fig. 16 relates these images to the pressure measurement and

the water height measurement close to the wall of the

contain-ment system (6-cm distance). This figure shows that the synchro-nization of the high-speed camera images with the measurements

gives very good insight in the physics of the impact and is thus

very useful in validating numerical simulations of these types of sloshing events.

70% filling raze. At a 70% filling rate, the wave runup on the walls is high enough to hit the ceiling of the containment system regularly. The upper chamfer changes the direction of the flow so that a water jet shoots along the ceiling and even hits the opposite

wall, where it mixes with the local flow. This is shown in the

4 snapshots in Fig. 17.

Fig 18 shows the impact pressure measurements associated with this impact, on the ceiling of the containment system. The

impact pressure is characterized by a double peak First the water jet hits the ceiling, then the water mass behind the jet flows past the ceiling. The duration of the first impact is very short; the dura-tion of the second impact is longer, but with a lower magnitude.

-0.5 photo - 12286 test 37820 002 002

r

photo 1 12291 water height

- pressure

223.1 223.2

2233

223.4 223.5 223.6 time [s]

Fig. 16 High-speed camera images related to pressure and water height measurement at 25% filling rate

12 Large-scale LNG Sloshing Model Tests

5 10 15 20 25 30

pressure pulse [hPa s]

Fig. 12 Probability density function of integrated pressure pulse P07 in regular sloshing test, 10% filling rate, stiff panel

100 200 300 pressure [hPa]

400

30 25

international Journal of Offshore and Polar Engineering, Vol. 19, No. 1, March 2009, pp. 8-14 13

Fig, 17 Digital camera images of regular test at 70% filling rate

95% filling rate. At a 95% filling rate, large impacts did not occur because the water does not have the freedom to move

around and build up momentum. However, some very interesting

mixing of air and water was observed close to the ceiling of the Fig. 19 High-speed camera images of regular test at 95% filling containment system, as shown in the 4 snapshots in Fig. 19. rate showing large bubble clouds

It is expected that the mixing of air and water has a significant

impact on the wall pressures. The tests at this high filling rate

are extremely useful for validation of 2-phase flow simulations These effects were captured accurately by means of a large num-(simulation of air and water together). _{ber of measurement devices, including pressure transducers, water}

height sensors, motion and acceleration sensors, and cameras CONCLUSIONS (including high speed). Significant effects of the stiffness of the panel are observed. This leads to the conclusion that hydroelastic

Large-scale oscillation tests (1: 10) with a 2-D section of an _{effects cannot be disregarded. The results of the tests are being}

LNG containment system were successfully performed at various

used to validate a VOF method extended with 2-phase flow and

filling rates (10%, 25%, 70% and 95%). The tests were carried _{compressibility.}

out to obtain validation material for a VOF method, extended for a 2-phase flow. The tests showed some very interesting physics

of sloshing, sUch as: _{ACKNOWLEDGEMENTS}

Shallow water bores

Runup against the vertical walls of the containment system Stichting Technische Wetenschappen (STW) and the partici-Impacts on the ceiling of the containment system pants of the ComFlow-2 JIP are acknowledged for their financial Captured air during impacts support and the permission to publish these test results. Heavy mixing of water and air

REFERENCES

test37820 003 004

250 _{Abramson, HN, Bass, RL, Faltinsen, 0, and Olsen, HA (1974).}

200

-"Liquid Slosh in LNG Carriers," 10th Symp Naval Hydrodyn, Cambridge, MA, USA, pp 37 1-388.

Chester, W (1968). "Resonant Oscillations of Water WavesI.

150-

Theory," Proc R Soc London, Ser A, Vol 306, pp 5-22.

0) Chester, W, and Bones, JA (1968). "Resonant Oscillations of

0

0

- Water WavesH. Experiment:' Proc R Soc London, Ser A,

Vol 306, pp 23-39

U)

U,

50-

Faltinsen, OM, and Timokha, AN (2002). "Asymptotic Modal

cL Approximation of Nonlinear Resonant Sloshing in a

Rectan-gular Tank with Small Fluid Depth," J Fluid Mech, Vol 470,

0)

pp 319-357.

Hill, DF (2003). "Transient and Steady-state Amplitudes of Forced Waves in Rectangular Basins," Phys Fluids, Vol 15,

267.5

268 268.5 269

time [s] No 6, pp 1576-1587.

Huijsmans, RHM, Trischler, G, Gaillarde, G, and Dallinga, R

In: Offshore and Polar Eng Conf, Toulon, France, ISOPE,

Vol 3, pp 424-432.

Lee, DY, and Choi, HS (1999). "Study on Sloshing in Cargo Tanks Including Hydroelastic Effects," Marine Sci and Tech, Vol 4,

No 1, pp 27-34.

Nam, B, and Kim, Y (2006). "Simulation of 1\vo-dimensional

Sloshing Flows by SPH Method," Proc 17th In: Offshore and Polar Eng Conf, San Francisco, ISOPE, Vol 3, pp 342-347.

Shin, Y, Kim, JW, Lee, H, and Hwang, C (2003). "Sloshing Impact of LNG Cargoes in Membrane Containment Systems in the Partially Filled Condition," Proc 13th In: Offshore and

Polar Eng Conf, Honolulu, ISOPE, Vol 3, pp 509-515. Tan, M, Xiong, YP, Xing, iT, and Toyoda, M (2006). "A

Numeri-cal Investigation of Natural Characteristics of a Partially Filled

Tank Using a Substructuie Method:' HydroelasilcitY 2006:

Hydroelasticity in Marine Techn Beijing, pp 180-190.

Wemmenhove, R, Veidman, AEP,

and Loots, E (2005).

"Modelling Two-Phase Flow with Offshore Applications," Conf Offshore Mech & Arci Eng, Paper 67460, Halkidiki,

Greece.

Wemmenhove, R, Veldman, AEP, Luppes, R, and Bunnik, THJ (2007). "Numerical Simulation of Sloshing in LNG Tanks with a Compressible Two-Phase Model," Conf Offshore Mech & Arct Eng, Paper 29294, San Diego.

Yamamoto, S, Kataoka, F, Shioda, S, and Ashitani, Y (1995).

"Study on Impact Pressure Due to Sloshing in Midsized LNG

Carrier," In: J Offshore and Polar Eng, ISOPE, Vol 5, No I,

pp 92-98.