Hydro-Impact, Fluid Structure Interaction and Structural Response of Modern Yacht

'UNIVERSITY OF SOUTHAMPTON

FACULTY OF ENGINEERING, SCIENCE AND MATHEMATICS SCHOOL OF ENGINEERING SCIENCES

HYDRO-IMPACT, FLUID-STRUCTURE

INTERACTION AND STRUCTURAL RESPONSE

OF MODERN RACING YACHT

by

June Lee

UNIVERSITY OF SOUTHAMPTON

ABSTRACT

FACULTY OF ENGINEERING, SCIENCE AND MATHEMATICS SCHOOL OF ENGINEERING SCIENCES

Doctor of Philosophy

HYDRO-IMPACT, FLUID-STRUCTURE INTERACTION AND STRUCTURAL RESPONSE OF MODERN RACING YACHT

by June Lee

In recent years, faster, lighter and bigger are the key issues in modern racing yacht for extreme performance. As a result, many yachts have experienced various structural failures caused by the hydro-impact phenomenon by slamming.

The structural failure by hydro-impact originates from the facts that the external hydro-impact load and fluid structure interaction effect is somewhat misled and when applying the load into current structural design, the 'dynamic' load is typically, ma-nipulated as 'static' way with fluid structure interaction effect is, generally, ignored. In this thesis, the hydro-impact load by slamming, its fluid structure interaction effect and dynamic response of the local structure of the yacht are studied.

Firstly, to acquire insight into the hydro-impact phenomenon, a series of drop tests and seakeeping-slamming tests are carried out with various sensing instruments of pressure transducers, accelerometer and 'slam patch system' - a specific application form of generally known pressure panel - are installed. The slam patch system is designed and implemented to investigate the hydro-impact loads and fluid structure interaction effect of slamming. Afterward, the measured hydro-impact loads are sum-marised via statistic manipulations with regard to pressure and duration time.

Secondly, the manipulated loads are recycled as input data to simulate the tran-sient response of local structure of the yacht. The applicability of this study which is based on fluid structure interaction and dynamics of structures is provided to com-pare it to the commonly used static structural design criteria given by various organi-sations.

Throughout this study, the dynamic and fluid structure interaction effect by hydro-impact phenomenon on local composite structure can be easily visualised and

Contents

1.2 Objective and Methodology 6

1.2.1 Objective 6

1.22 Methodology 7

2 EXPERIMENT ON HYDRO-IMPACT OF SLAMMING 11

2.1 Measurement Methodology 12

2.1.1 Comparison of Measurement Methods 12

2.1.2 Consideration of Fluid-Structure Interaction 14

2.1.3 Experiment Setup 16

2.2 Slam Patch System 18

2.2.1 Background of Slam Patch 18

2.2.2 Design of Slam Patch 19

2.2.3 Dynamics of Slam Patch in Dry Mode 20

2.2.4 Fluid-Structure Interaction of S/P in Wet Mode 21

2.3 Modal Testing 23 . . . . . . .. .

. ...

... .

2.4.1 Drop Test

2.4.2 Seakeeping-Slamming Test 24

2.4.3 Initial Assessment 26

3 RESULTS OF HYDRO-IMPACT EXPERIMENT 39

3.1 Stochastical Process 40

3.1.1 Method of Estimation 40

3.1.2 Quantification of Uncertainty 41

3.1.3 Model Diagnostics 41

3.1.4 Probability Distribution Function in This Study 42

3.1.5 Curve Fittings in This Study 43

3.2 Consideration of Extreme Values 45

3.2.1 Basics of Extreme Value Distribution 45

3.2.2 Inference Process for Extreme Value 47

3.3 Results of Drop Test 49

3.3.1 At Pressure Transducers 49

3.3.2 At Slam Patches 51

3.3.3 Extreme Values 52

3.4 Results from Seakeeping-Slamming Test 53

3.4.1 The Whole Series of Runs 53

3.4.2 On a Specific Independent Variable 55

3.4.3 Extreme Values 56

4 DESIGN IMPACT PRESSURE BY RULES AND REGULATION 81

4.1 The International Organization for Standardization (ISO) 85

4.1.1 Design Pressure 85

4.1.2 Acceleration 86

4.1.3 Factors 88

4.1.4 Calculated Results using ISO Standard 90

4.2 Lloyd's Register (LR) 96

4.2.1 Non-Displacement Mode Craft 97

4.2.2 Displacement Mode Craft 98

4.2.3 Acceleration 102

4.2.4 Factors 103

4.2.5 Calculated Results using LR Standard 105

4.3 Bureau Veritas (BV) 111 24 . .. . . . .. . . . . . . . .

...

4.3.3 Calculated Results using BV Standard 114

4.4 Det Norske Veritas (DNV) 119

4.4.1 Design Pressure 119

4.4.2 Acceleration 120

4.4.3 Factors 121

4.4.4 Calculated Results using DNV Standard 122

4.5 American Bureau of Shipping (ABS) - HSC Guide 126

4.5.1 Design Pressure 126

4.5.2 Acceleration 127

4.5.3 Factors 128

4.5.4 Calculated Results using ABS HSC Standard 128

4.6 American Bureau of Shipping (ABS) - ORY Guide 134

4.6.1 Design Pressures 134

4.6.2 Calculated Results using ABS ORY Standard 134

4.7 Comparisons of Rules and Slamming Experiment 137

5 NUMERICAL CALCULATION OF TRANSIENT RESPONSE 149

5.1 General 149

5.1.1 Theoretical Background 149

5.1.2 History in This Study 151

Scaling 153

5.2.1 Generals of Scaling 153

5.2.2 Consideration of FRF and Structural Damping 154

5.3 Simulation Methodology 158

5.3.1 Simulation Cases 158

5.3.2 Assumption for FSI 159

.4 Verification Example 1 - Dry Mode 160

5.4.1 Analytical Method 160

5.4.2 Numerical Method 161

5.4.3 Comparison of Results 161

5.5 Verification Example 2 - Wet Mode 164

5.5.1 Plate on the Water Surface 164

5.5.2 Numerical Method 165

5.5.3 Comparison of Results 165

5.6 Response of Open 60' to Various Loads 167

5.2

...

...

...

...

5.6.4 Transient Response in Dry Mode 171

5.6.5 Transient Response in Wet Mode 173

6 CONCLUSION AND FURTHER WORK 197

6.1 Conclusion 197

6.2 Further work and Recommendation 201

A CALCULATION PROCEDURE FOR FLUID STRUCTURE INTERACTION

IN ANSYS 203

B MECHANICAL PROPERTY CALCULATION OF LAMINATE OF SIP 209

C S/P IN DRY AND WET MODES 213

D MODAL TESTING OF S/P IN DRY AND WET MODES 217

E ANALYSIS OF TRANSIENT RESPONSE OF PLATE 227

F NUMERICAL PROCEDURE OF TRANSIENT CALCULATION 231

F.1 Transient Analysis 231

F.2 Solution 233

F.3 Sparse Direct Solver 235

F.4 Nodal and Reaction Load Calculations 236

G EQUIVALENT LAMINATE PROPERTIES 239

References 243

Paper presented at The Modern Yacht Conference, 2007, RINA 249

Paper submitted to The Journal of Sailboat Technology, 2008, SNAME 256

. . . . .

. . .

....

List of Figures

1.1 Midship fracture in America's Cup 9

1.2 Keel failure in Sydney-Hobart Race 9

2.1 Locations of PIT and S/P in drop and S-S tests 28

2.2 Instrument networking 28

2.3 Schematic view and prototype of S/P 29

2.4 Locations of P/T and S/P in drop test 30

2.5 Drop test setup 31

2.6 Locations of P/T and S/P in S-S test 32

2.7 Seakeeping-slaming test 33

2.8 Limit of pressure capacity of S/P 34

2.9 Signal patterns at P/T and S/P in drop test 34

2.10 Time signals of PIT and S/P at low drop height 35

2.11 Spectrums of S/Ps at low drop height 36

2.12 Time signals of S/Ps after filtering at low drop height 36

2.13 Signals in time and frequency of S/P in S-S test 37

3.1 Families of extreme value distribution 58

3.2 Hydro-impact signals at P/Ts in drop test 58

3.3 Relationship of Ha vs. Pp at P/Ts in drop Test 59

3.4 Relationships based on Ha at P/Ts in drop Test 59

3.5 Relationships based on Vi at P/Ts in drop test 60

3.6 Relationships based on V;2 at P/Ts in drop Test 61

3.7 Relationships of Pp vs. Q,, Tavs. Pp, Td vs. Qi at P/Ts in drop test 62

3.8 Hydro-impact signals at S/Ps in drop test 63

3.9 Relationships based on Pp at S/Ps in drop Test 64

...

... .

3.13 Relationships of Pp vs. Wf throughout the locations in the whole series

of S-S test 66

3.14 Histogram, PDF and CDF of P/Ts in the whole series of S-S test 67

3.15 Histogram, PDF and CDF of S/Ps and ACC. in the whole series of S-S

test 68

3.16 Definition of duration time 69

3.17 Wf vs. Td in the whole series of S-S test 70

3.18 Pp vs. Td in the whole series of S-S test 70

3.19 Wf vs. Tr in the whole series of S-S test 71

3.20 Pp vs. Tr. in the whole S-S test 71

3.21 Wh vs. Qi in the whole S-S test 72

3.22 Pp vs. Ch in the whole S-S test 72

3.23 Qi vs. Td in the whole S-S test 73

3.24 Wf vs. Pp on the fore area of fore body at Wf = 0.2 m 73 3.25 Histogram, PDF and CDF of Pp on the fore area of fore body at Wh = 0.2m 74 3.26 Tiff vs. Pp on the rear area of fore body at Wf = 0.2 m 75

3.27 Histogram, PDF and CDF of Pp on the rear area of fore body at Wh= 0.2m 76 3.28 Wh vs. Pp on the fore area of fore body at Wh = 0.1 m 77 3.29 Histogram, PDF and CDF of Pp at LOC 1, 6 at WI, = 0.1 m 78

3.30 Extremes on the fore body at Wh = 0.2 in 79

3.31 Estimation of Td, T. and Qi at LOC 6 (S/P 1) at W,. = 0.2 m 80

4.1 Body plan of Open60' designed by Le Groupe Finot-Conq 140

4.2 Definition of deadrise angle by ISO and LR 140

4.3 Relationship of panel configuration and kAR for sailing craft in ISO . 140

4.4 Relationship of displacement and bottom pressure for sailing craft in ISO 141

4.5 Design bottom impact pressure of plating by ISO for unlimited service . 141

4.6 Distributions of Pf at various boat speed in LR 142 4.7 Design bottom impact pressure of plating by LR for unlimited service 143

4.8 Definition of deadrise angle for yacht by BV 143

4.9 Longitudinal location of calculation points of av in BV 144

4.10 Design bottom impact pressure of plating by BV for unlimited service 144

4.11 Relationship of design area and slamming pressure by BV sailing yacht 145

4.12 Definition of deadrise angle for round bottom by DNV 145

4.13 Design bottom impact pressure of plating by DNV for unlimited service 146 4.14 Relationship of significant wave height and acceleration by DNV HSLC 146 4.15 Design bottom impact pressure of plating by ABS HSC for unlimited

. .

. .

4.16 Relationships of design area and design head reduction factor by ABS

ORY 147

4.17 Impact pressures by five organisations and model test result 148

5.1 Relationship between duration time and natural period in 1 DOF

sys-tem to half sine impulse 175

5.2 Scaling of FRF 176

5.3 Response to S-S test impulse with various damping 177

5.4 Mode shapes of plate on water cavity 178

5.5 Model of Open 60' 179

5.6 Boundary conditions of Open 60' 180

5.7 Deflection and stress to loading of ISO 181

5.8 Deflection and stress to loading of BV 182

5.9 Mode shapes of Open 60' 183

5.10 Mode shapes of Open 60' at B/C 1 184

5.11 Mode shapes of Open 60' at B/C 2 185

5.12 Response in dry mode to impulse of S-S 1/100 loading with B/C 1 . . 186

5.13 Response in dry mode to impulse of S-S 1/1000 loading with B/C 1 . . 187

5.14 Response in dry mode to impulse of S-S 1/100 loading with B/C 2 . . 188

5.15 Response in dry mode to impulse of S-S 1/1000 loading with B/C 2 . . 189

5.16 Response in wet mode to impulse of S-S 1/100 loading with B/C 1 . . 190

5.17 Response in wet mode to impulse of S-S 1/1000 loading with B/C 1. . 191

5.18 Response in wet mode to impulse of S-S 1/100 loading with B/C 2 . . 192

5.19 Response in wet mode to impulse of SS 1/100 loading with B/C 2

-Beyond duration time of impulse 193

5.20 Response in wet mode to impulse of S-S 1/1000 loading with B/C 2. . 194

5.21 Response in wet mode to impulse of SS 1/1000 loading with B/C 2

-Beyond duration time of impulse 195

C.1 FEA model of S/P 213

C.2 Mode shape of S/P in dry mode 214

C.3 Mode shape in wet mode (displacement) 215

C.4 Mode shape in wet mode (pressure) 216

D.1 Point FRF of S/P No.1 in dry mode 217

D.2 Point FRF of S/P No.2 in dry mode 218

. . . . . . . . . . . .

...

. ...

...

D.6 Point FRF of S/P No.6 in dry mode 220

D.7 Direct FRF of S/P No.1 in dry mode 220

D.8 Direct FRF of S/P No.2 in dry mode 221

D.9 Direct FRF of S/P No.3 in dry mode 221

D.10 Direct FRF of S/P No.4 in dry mode 222

D.11 Direct FRF of S/P No.5 in dry mode 222

D.12 Direct FRF of S/P No.6 in dry mode 223

D.13 Direct FRF of S/P No.1 in wet mode 223

D.14 Direct FRF of S/P No.2 in wet mode 224

D.15 Direct FRF of S/P No.3 in wet mode 224

D.16 Direct FRF of S/P No.4 in wet mode 224

D.17 Direct FRF of S/P No.5 in wet mode

DA& Direct FRF of S/P No.6 in wet mode 225

. . . . v . . . . . 225

List of Tables

7.1 Stational location of P/T and S/P in drop and S-S tests . . 17

2.2 Major measurement instruments and network . . 17

2.3

Mechanical properties of laminate of S/P ... ..

2.4 Natural frequencies of S/P in dry mode with various rod lengths . , 20

2.5 Natural frequencies of S/P in wet mode with various rod lengths ., 21

2.64 Test matrix of S-S test . , 4. 4., 26

3.1 Pressure estimation by Gamma PDF at PIT 1 and S/P1 .44 .4,, 54

3.2 Extreme pressure estimation at Wh = 0.2 m 56

3.3 Estimation of parameter a, b, c for Eq.(3.20) . i.

3.4 Variable estimation for extreme engineering . , .

!.l'...

4.3 Design category vs. significant wave height . . . S4

4.4 Dynamic load factor ri.CG and upper limit according to craft type by ISO 8:7

4.5 Design category factor ,koc by ISO 88

46 Pressure calculation results at V = 12.36 knots by ISO 91 4.7 Pressure calculation results at V =415 knots by ISO. 92 4.8 Pressure calculation results at V = 20 knots by ISO . ...Al 93

4.9 Pressure calculation results at V = 25 knots by ISO 94

4.10 Pressure calculation results at V = 30 knots by ISO . 95

4.11 Pressure determined for Open60' by ISO 95

4.12 Definition of girth distance Go by LR . . 97

4.13 hull form wave pressure factor kr by LR . . . . 100

4.14 Variable determination for at 0.5 Lwy of LR . 101

4.15 Minimum significant wave height H113 by LR 4. . ., 102

4.16 Forebody impact pressure factor ff by LR ,s*, 103

4.17 Stiffening type factor of by LR .

.

4.20 Service area restriction notation factor

pf

4.21 Craft type notation factor Cf by LR . . . _ . . 104

4.22 Pressure calculation results at V = 12.36 knots byLR ,

-4.23 Pressure calculation results at V = 15 knots byLR 11087

4.24 Pressure calculation results at V = 20 knots byLR

4.25 Pressure calculation results at V = 25 knots byLR 109

4.26 Pressure calculation results at V = 30 knots byLR

_

427 Pressure determined for Open 60' byLR .

00'

4.28 Longitudinal bottom slamming pressure distribution factor In Day . ,, 112

4.29 Limit value of acG for motor yacht by BV , . , . 112

4.30 Foci for motor yacht by BV .. ....

4.31 Soc for motor yacht by BV 4.32 Foc for sailing yacht by BV . .

4.33 Soc for sailing yacht by BV ...r. ... . . 114

4.34 Pressure calculation results at V = 12.36 knots by BV motor yacht .. , 115

4.35 Pressure calculation results at V = 15 knots by BV motor yacht

- .,

4.36 Pressure calculation results at V = 20 knots by BV motor yacht .. . 116

4.37 Pressure calculation results at V = 25 knots by BV motor yacht . . 117

4.38 Pressure calculation results at V = 30 knots, by BV motor yacht ,. _ .t . 117

4.39 Pressure calculation results by BV sailing yacht . . . 118,

4.40 Pressure determined for Open 60' by BV

4.41 Acceleration factor fg by DNV . . _{S ^:.} 1: ..,j F v, n, 121 4.42 Service notation by DNV HSLC . _{a "n 4 :1} Ili 121

4.43 Hull type factor kh by DNV . 122

4.44 Acceleration calculation results by DNV .

4.45 Pressure calculation results at V > 12.36 knots by DNVRO Patrol! . . . 123

4.46 Pressure calculation results at V > 12.36 knots by DNV RO Cargo ,, , . 124

4.47 Pressure calculation results at V > 12.36 knots by DNV RO Yacht ,. . 125

4.48 Pressure determined for Open 60' by DNV

4.49 Applicable craft type and length by ABS - HSC . 126

4.50 Vertical acceleration factor k, by ABS HSC . 128

4.51 Vertical acceleration factor Ft, by ABS HSC . .

4.52 Pressure calculation results at V= 12.36 knots by ABSHSC

4.53 Pressure calculation results at V= 15 knots by ABSHSC . 130.

4.54 Pressure calculation results at V= 20 knots by ABSHSC . . 131

4.55 Pressure calculation results at V= 25 knots by ABS HSC

106 113 113 114 118 122 125 128 129

4.57 Pressure determined for Open 60' by ABS 133

4.58 Coefficient F for shell plating by ABS ORY 135

4.59 Design heads h for plating by ABS ORY 135

4.60 Pressure calculation result of ABS ORY 136

4.61 Pressure calculation results by ABS 136

4.62 Design impact pressure on S/P by five organisations 137

4.63 Impact pressure on PIT and S/P by S-S test 137

4.64 Design impact pressure on bottom plating of Open 60' for unlimited

service by five organisations 139

5.1 Natural frequencies of simply supported plate 162

5.2 Natural frequencies of clamped plate 162

5.3 Maximum deflections of simply supported plate under various load type 163

5.4 Natural frequencies of plate in vacuo (Hz) 166

5.5 Natural frequencies of plate on water cavity (Hz) 166 5.6 Loading matrix for transient response calculations 167

5.7 Raw materials for Open 60' structure 168

5.8 Laminate schedule of Open 60' bottom structure 168

5.9 Mechanical properties of Open 60' structure 169

5.10 Summary of static response to ISO and BV loadings 169

5.11 Modal analysis of forebody of Open 60' at B/C 1 170

5.12 Modal analysis of forebody of Open 60' at B/C 2 170

5.13 Transient response to S-S impulse loadings in dry mode - deformation 171

5.14 Transient response to S-S impulse loadings in dry mode - stress 171

5.15 Comparison of loading between BV and S-S tests 172

5.16 Comparison of response to the loading of BV and S-S tests in dry mode 172

5.17 Transient response to S-S impulse loadings in wet mode - deformation 173

5.18 Comparison of response to the loading of BV and S-S tests in dry and

wet modes 173

B.1 Mechanical properties of raw material for S/P 209

. - . . . . . .

..

....

..

...

Nomenclatures

Receptance Deadrise angle

ficg Deadrise angle at LCG

Taylor quotient f Stiffening type factor,

Displacement

0 Running trim angle

OD Deadrise angle

Geometric scale factor

A Scale factor

Poisson's ratio Density

Pc Density of composite

Pf Density of fibre

firn Density of matrix

Trim angle Angular velocity

urn _{Angular natural frequency} Wran Natural frequencies

V Volumetric displacement

acG Design vertical acceleration av Vertical acceleration in g

av Total vertical acceleration

A Area

AD Design area

Am Midship area

AR Reference area

Awet Wetted area

Atop Waterplane area

ABS American Bureau of Shipping

ACV Air Cushion Vehicle

AP Afterward Perpendicular

Shorter dimension of panel between two closest stiffeners in mm. Breadth overall

Bc Chine beam

.BWL Waterline breadth

BV Bureau Veritas

Speed of sound

Damping coefficient matrix Damping matrix

Ch Block coefficient

Craft type notation factor Midship coefficient Prismatic coefficient

'Cvp Vertical prismatic coefficient'

C,.

CDF Cumulative Distribution Function Draught in in

fl

c DNV Det Norske Veritas

Young's modulus

Ef

Si

Young's modulus of matrix Frequency

fd Hull form pressure factor

fj

Acceleration factor

External force, Design head reduction factor External load vector

FD Design area factor

Froude number

Vertical acceleration distribution factor FEA Finite Element Analysis

FFT Fast Fourier Transform FP Foreward Perpendicular

FRF Frequency Response Function

PSI Fluid Structure Interaction

9 go Gravitational constant Waterplane

Girth distance

Cwt. Girth length at waterline

GZmax<60° Maximum righting moment lever at the heel angle less than 600

GC Centre of Gravity

GEV Generalised Extreme Value CRP Glass Reinforced Plastic

It Height, thickness, design head

H113 Significant wave height of 1/3

Hd Drop height

H f Hull notation factor Significant wave height

HSC High Speed Craft

HSLC High Speed and Light Craft

IMOCA International Monohull Open Class Association

1SAF International SAiling Federation

ISO The International Organization for Standardization Stiffness coefficient

ki Longitudinal bottom slamming pressure distribution factor Slamming area factor

k3 Shape and deadrise factor kAR Area pressure distribution factor

kDC Design category factor

Hull type factor

Longitudinal distance factor

Longitudinal pressure distribution fad& Structural component and boat type factor

ksis

km, Gyradius on y axis Stiffness matrix

Spring coefficient matrix

KG Keel to centre of gravity

KM Keel to metercentre

1 Longer dimension of panel between two closest stiffeners in mrn

lu Long dimension (unsupported length) of stiffeners in mm

Length

Lower triangular matrix

L11 L0A Lli-L Lx LCG LOC LR m LDC Al

ncv

Pp

Longitudinal Centre of Gravity Location

Lloyd's Register Mass

Loaded displacement mass Mass matrix

Mass coefficient matrix Number of hulls

Dynamic load factor, acceleration at CG Number of hulls

Element shape function for pressure Element shape function for displacements Unit normal vector to interface

Offshore Racing Yachts Excitation pressure, prototype Excitation pressure

Acoustic pressure

Virtual change in pressure

Bottom slamming pressure at LCG Bottom slamming pressure at any section Design pressure for bottom stiffening Motorcraft bottom pressure

Motorcraft bottom base pressure Minimum motorcraft bottom pressure Design pressure for bottom plating Sailing craft bottom pressure Sailing craft bottom base pressure Minimum sailing craft bottom pressure Bottom impact pressure by slamming Bottom impact pressure

Prnn Generalised force

Magnitude of impulse

Pp Peak pressure

Ps Hydrostatic pressure + hydrodynamic wave pressure

Ps/ Slamming pressure

Ps/p Pitching slamming pressure

Hydrodynamic wave pressure

PDF Probability Density Function PIT Pressure Transducer

qmn Normal coordinate

Qi Impulse quantity

RIB Rigid Inflatable Boat Spacing of stiffeners in win Interface surface

Sic Service type notation factor

Sr Reference area

SES Surface Effect Ship

S/P

S-S Seakeeping-Slamming

SSC Special Service Craft

STA Station

SWATH Small Waterplane Twin Hull

Time

ti Impulse duration time Draught, Natural period Draught Duration time Tf Force Transmissibility Draught TX. Rise time Try Transmissibility Local draught

Nodal displacement vector

it Nodal velocity vector Nodal acceleration vector Nodal vector

Nodal displacement vector at time t Nodal velocity vector at time tn Nodal acceleration vector at time

Un i-1 Nodal displacement vector at time 4,4.1

Un.-1-1 Nodal velocity vector at time t, 4.1

1-11;f-1 Nodal acceleration vector at time tn+1

Upper triangular matrix

V Boat speed in knots, Volume

Fibre volume fraction Impact velocity

vi2 _{Impact velocity squared}

Speed of model boat, Matrix volume fraction

VCB Vertical centre of buoyancy Deflection

WI Wave frequency, Fiber weight fraction

Wh Wave height

Wm Matrix weight fraction

"777L71. Mode shape

Chapter 1

INTRODUCTION

1.1

Background

1.1.1

Motivation

In several decades, sailing market has claimed higher speed, higher performance racing yacht to win in numerous races. Exotic, lighter and stiffer materials have been introduced to yacht design and building industries and the higher performance rac-ing yacht has been achieved through lighter and stiffer structure. Meanwhile, as the environment of race gets harsher and aspiration of sailor to speed gets higher, struc-tural failures of the racing yacht also have been increased because of the tendency to adopt the extreme lightness into the structure, the typical example is the carbon-honey sandwich structure used in several ORMA 60 trimaran. Severe failures in structural member of racing yacht have been reported from various media sources.

For example, in America's cup campaign, 'Young America' experienced a catas-trophic hull failure in their amidships, shown at Figure 1.1, by a few series of large waves [log,

'She was sailing in New Zealands Hauraki Gulf at 9 knots, in about 20 knots of wind, when, while tacking from port to starboard, she hit a series of three large waves. The impact of the first wave slowed the boat from 9 to 3, and caused all onboard computers to crash. By the time the third wave passed, the boats side decks had buckled, the topsides had torn, and the hull had folded up... About a week before the accident, the team suspected delamination and repaired it... After the start of the race, the crew noticed the repaired area were starting to show abnormal movement...'

Hobart Race, various severe failures at hull, appendages and rigs by various reasons have been reported [71] [72] [74].

In the Jacques Vabre Race,

'As if starting the biennial doublehanded transatlantic race in 25+ knot, on-the-nose winds and rough seas weren't enough for the 19 monohulls (Saturday start) and 16 inultis (Sunday start), Sunday night a cold front rolled in, blasting primarily the multihull fleet with 35 to 45-knot winds and reported 20-ft seas. In the wee hours of Monday, EPIRBs went off on the 60-ft trimarans Sodebo and Orange Project. The former snapped off its port ama and capsized. The latter suffered a broken main beam and also went upside down. Three hours later, at 0615, a report came in that Foncia had also capsized. With the retirement of Brossard earlier on Sun-day with a cracked main hull, that takes four of the ten 60-ft multis out of the running. The six co-skippers of the three capsized boats were rescued, some more banged and bruised than others but all okay. At this writing, their smashed boats were all either under tow back to land, or about to be...'

Another example is the keel failure in Sydney-Hobart Race, shown at Figure 1.2

[74],

'...Wharington described the series of events: "we were sailing on port tack, and we landed on a rogue wave, the hydraulic keel rams broke and the keel swung over to one side. The boat lay over on its side. ... But then the thing broke loose again and we were unable to restrain it and it started chopping out the inside the boat. It began to be unsafe and we felt the need to get off the boat. We were concerned about whether the keel would keep on working its way through and destroy the framing around it, and fall out the bottom"...

It will be interesting to see the hydraulics manufacturer's defence to this lawsuit. It could be argued that heading into an environment as hostile and unpredictable as the Bass Strait and Tasman Sea invalidates the manufacturer's warranty. Russel bowler, one half of the Farr Yacht Design partnership, certainly has some sympathy for the designers and builders of 'Konica Mi-nolta': "They are very competent group of engineers. But you're talking about sailing a 100ft boat in the worst place in the world. And there is some wisdom in backing off at the right time. The crew knew they needed to get into flatter water and they reckon they were just five miles away from getting there when the boat broke"...'

structural response. However, the exact mechanism of the response of racing yacht to slamming is still beyond designer's expectation since most of the structural design criteria are based on static-wise approach, not based on the real dynamic situation be-cause of the uncertainty and complexity of the matter. The static approach can easily be found in typical powerboat design Ill] [12] [40].

The ignorance to this dynamic approach in yacht design are shown at following examples. One example of failure by wave impact can be found in recent TransAt

Jacques Vabre Race 2005 which says,

...Sodebo (boat name, ORMA 60 trimaran) is more complex. Thomas Coville's boat suf-fered a lateral impact from a large wave against its windward float, resulting in its virtually instantaneous rupture. It's likely here that the honeycomb core, with its sensitivity to impact, was the cause of the rupture. A foam core would probably have avoided this problem, but this would have been heavier and, at the time, wouldn't have allowed prepreg construction...'

Another one is such that French composite engineer Herve Devaux who is closely involved in the engineering of Juan Kouyoumdjian's second Volvo 70 for ABN AMRO, 'The stresses applied by the appendages (keels, daggerboards, foils, rudders) are also relatively well met by calculation, though there remains some uncertainty as to dynamic loads -this can only be resolved through measurements taken at sea. As for the stresses caused by the sea, those provoked by the hull dynamics are well managed, while those due to the impact of the waves are more random; progress has been made but unfortunately a better understanding is only achieved through breakage. i.e. empiricism...'

Finally, designer Juan Kouyoumdjian testifies structural dynamics of sailing yacht failure as,

'...Most of the problems have arisen because these boats are too stiff. You can't fight against these deformations. You have to design a structure to accept them. To do that, you must quan-tify them and that's the problem we have. remember saying very early on that, these had to be flexible boats. We had discussions in the office about cars having suspension and asking where

we can put in some energy-absorbing structure. Obviously we haven't found the answer yet...'

It is not easy to quantify the value of this uncertainty - the dynamic response, be-cause of the existence of strong non-linearity of the load, fluid structure interaction (FSI) effect and dynamic approach in structural design. This is the assignment

bur-1.1.2

Academic History

Hydro-impact by skimming hasbeen well studied in planing boat field than ship or yacht design area.

The pioneering research works were carried out by von Karman [81] followed by Wagner [33], Ochi et al [48], Stavovy and Chuang [76] [15], Korobkin et al [5] [2], Heller and Jasper [30], Allen and Jones [66], Fridsma [27], Savitsky pi and Faltinsen

et al [53] [58] [59].

The studies by von Karman and Wagner are based on the two dimensional plate analogy, they idealised the impact problem as a two dimensional wedge entry on the calm water surface to estimate the hydro-impact pressure known as the linear slam-ming theories. Wagner's theory is simple and useful such that measured peak pres-sure is typically a little lower than his theory and gives conservative estimation for practical use. Ochi et al also suggest analytical approach on slamming phenomenon where the role of air entrapment and compressibility are found. Korobkin et al devel-ops two, three dimensional and shallow water impact theories. All these works done by is to find the hydrodynamic pressure in theoretical way.

Since the phenomenon itself is so complicated, the subject has been studied in ex-perimental way such that full-scale and model-scale tests have been carried out in

parallel to the two dimensional drop tests.

Slamming theory on high performance powerboat starts from the work of Heller and Jasper. They instrumented and obtained data on aluminum hulled torpedo boat (YP110) and used this data as the basis of the empirical aspects of load calculation. Allen and Jones measured pressure, acceleration and structural response on two plan-ing craft in full-scale and combined the results with theoretical models to derive the method to predict the design pressures. Savitsky and Fridsma did extensive model test on planing craft in waves. Chuang, Faltinsen at al, carried out a series of drop test. So far, the works of them are concentrated mainly to find hydro-impact pressure and have become the foundation of typical powerboat design methodology. On the other hand, structure - oriented approaches are also carried out such as Spencer [41], Koelbel [40] and Joubert [56]. Koelbel recommended a modification of the design pro-cedure of the traditional design methodology.

pioneering works in early 1970s such as Sellars [26] [25], Bishop and Price [64] [49] [62] and Haszpra [29].

Sellars et all studied the influence of structural elastics on maximum slamming pressures. R.E.D. Bishop et al estimated slamming responses by the methods of linear dynamics. Haszpra summarised the hydro-elastic problem from similitude, experi-ment methodology and testing technique. Recently, Faltinsen et al [37] [23] [50] [21] [7], Bereznitski [9], Scolan and Korobkin [82], Rosen [69] [70] [68] and Kapsenberg et al [43] shows a representative research work on the hydro-impact, FSI and structural response.

Faltinsen eta] carried out the study on the influence of hydro-elastics on local slam-ming loads and structural responses. Bereznitski examined hydro-elastic effect with parameters of stiffness, air entrapment and deadrise angle. Kapsenberg et al found the hydro-impact pressure and global response on passenger ship model with an ar-ray of pressure transducers and strain gauges. Rosen investigated the hydro-impact and structural responses of planing craft in waves. This area is now play the one of the key roles in structural design of ship and boat.

However, the research work on the subject in the yacht design field is relatively rare since slamming of racing yacht is relatively new research field by the demands of high performance racing yacht.

Joubert investigated the pressure load through the inference of hull plating defor-mation 1551 [55] [57]. Wilson et al studied the slamming phenomenon on Open 60' using pressure panel system [54] [24]. Reichard [65], Hentinen and Holm [31], carried out some full-scale measurements by instrumenting sailing boats with pressure trans-ducers. Battley and Lake [8] investigated the slamming effect on sandwich structure.

Nevertheless, the works on this hydro-impact problems are still under research in the world. Especially, in yacht industries, the typical methodology of structural design in terms of structural dynamics under impact is still based on static approaches and most of the published research about slamming on yachts has concentrated on the def-inition of suitable equivalent static pressure whereas ship design area is attempting to find the real responses under real hydro-impact load theoretically or experimentally.

1.2

Objective and Methodology

Dynamic behaviour of sailing yacht in rough weather is being new design

paradigm in design community. This is a typical dynamic (transient) hydro-elastic problem in terms of hydro-impact load. This dynamic response would be important factor for serious high performance racing yacht to manage the weight of boat and structural integrity at the same time to win races.

This study challenges the dynamic approach with the modeled hydro-impact load from experiments to investigate the difference between the static and dynamic ap-proach and to realise the situation that the boat encounters in the ocean. To quantify the hydro-impact load and associated characteristics, seakeeping-slamming test and others will be prepared. In addition, comparing the test and dynamic approach results to the rules of structural design will be carried out to suggest a recommendation to structural design of yacht with regard to the structural dynamics with fluid-structure interaction effect.

1.2.1

Objective

The objectives in this study is to satisfy the demands of the common questions issued by design community like,

What is the maximum hydro-impact load by the slamming to a given yacht? Which part of the yacht structure is most loaded, to which degree and how fre-quently?

How this can be calculated and implemented into design process?

The objectives and process of this research is constituted in three phases to follow the pre-mentioned questions such that,

Clarify the hydro-impact by slamming through experiments with Open 60'

model ;

Define the variables,

Establish adequate experiment setup,

Ka)

Select organisations,

Evaluate and compare the results of the rules.

3. Verify the dynamic behaviour of the structure under the various loads ;

Assess the dynamic responses of yacht's structure based on the experiment results and rules,

Suggest structural design recommendation with regard to structural

de-sign.

1.2.2

Methodology

To fulfill the objectives, following methods will be implemented. 1. Instrument ;

Pressure transducer and Slam patch system are considered, Design and analyse the Slam Patch system,

Networking the instrumental system.

2. Carry out experiments and assess the slamming phenomenon ;

Carry out drop and seakeeping-slamming tests and measure the hydro-impact pressure, acceleration and the related,

Measure the hydro-elastic effect,

Assess the characteristics of hydro-impact in stochastic way - magnitude of pressure, duration time, impulse shape and quantity, FSI effect, etc.

3. Evaluate five rules and regulations given by five organisations;

Calculate the design of by ISO, LR, BV, DNV and ABS with regard to the yacht structure design,

Compare the calculation results, Compare the methodologies given.

4. Simulate the structural response to various toad cases.; Calculate the response to five organisations' load,

Calculate the response to the hydro-impact load by experiments,

Ka) (b) Ka) (b) tc) KO

Compare the responses,

Suggest design recommendations in terms of structure integrity.

The simulation of the structure response will be based on commercial FEM program and before carry out the simulation, couple of verifications will be performed.

I

;Sal

I. .s..ThILY.

IL

- T. skandia7"---,----_ n ta

;:to-Figure 1.1: Midship fracture in America's Cup

artair

Yl

.z

Chapter 2

EXPERIMENT ON

HYDRO-IMPACT OF SLAMMING

The objective of this experiment is to investigate the hydro-impact characteristics of the slamming phenomenon. In this study, the hydro-impact is characterised as a pressure impulse which has typical characteristics of,

Peak pressure, Duration time, Impulse shape,

Fluid structure interaction (PSI> effect.

First of all, adequate measuring method and related hardware to measure the char-acteristics are studied. To include the FSI effect and to find the trend of pressure over unit area, a special measurement system is implemented which is designated 'Slam Patch system (SIP)', a special application form of pressure panel. Details of method-ology and hardware are described in following sections.

2.1

Measurement Methodology

2.1.1

Comparison of Measurement Methods

There are four different methods to measure external pressure in naval architecture which are by,

Pressure Transducer (P/T), Pressure panel,

Strain gauge, Accelerometer.

Each method measures different characteristics of external pressure. However, for the sake of structural design of a yacht, finding a pressure over an unit area is im-portant to determine the scantling of the related structure, for example, hull-bulkhead joint. The issue remains, when using these methods, is how to exactly measure the pressure or how to convert the measured signal into what is useful to structural

de-sign of yacht?

The simplest way to acquire pressure is by a PR- or array of P/Ts in model tests or full-scale tests. Alternatively, at model-scale, pressure can be measured by isolating a small section of structure and measure the pressure on this small section with a pres-sure panel. In this case, the prespres-sure panel is a S/P. This S/P is sized to correspond to a specific unit area. The panel area of the hull bottom which is in-between stiffeners is the typical example. On the other hand, a strain gauge or an accelerometer can be attached to the structure in the model tests or full-scale tests to deduce pressure by measuring the structural response. In other words, pressure can be determined indi-rectly by measuring the structural deflections or accelerations to the actual or scaled structure or directly using SIP and PIT. It is important to recognise that each of these methods measures different physical quantities and that there is no necessary relation-ship among the quantities.

The different physical quantity of each measurement methods are,,

That the PIT measures pressure on a small area without hydro-elastic effect since PIT has its own high natural frequency and small diaphragm area. The transducer measures peak pressure when the peak pressure occurs at the

loca-That the S/P measures instantaneous pressure with hydro-elastic or PSI effect over an arbitrary wide area which is larger compared to P/T's diaphragm. The load cell in S/P is more likely to catch the peak pressure. Furthermore, the mea-sured signal may contain not only the external pressure but also the natural behaviour component of S/P which is coupled with external pressure, if it is designed to do so.

That deflection or acceleration measured by strain gauge or accelerometer de-pends on the mechanical properties and scantling of the structure. Therefore, for the model test, the model structure must be adequately scaled down from full-scaled one with a known Frequency Response Function (FRF) or for a known structure, the measured signals must be manipulated to deduce the pressure. In either case, it is not an easy task to carry out.

It is also important to study the pros and cons of each method to select adequate measurement method and hardware. The following is a summary of pros and cons of each measurement method.

1. Pressure Transducer Pros

Has relatively very high natural frequency, Measures instantaneous pressure,

Measures pure hydro-impact pressure without PSI effect,

Propagation of pressure can be detected with array of transducers, Easy to install.

Cons

Measures pin-point pressure, so may miss the peak pressure, Pressure over area is deduced with array of transducers,

Mathematical process needed to deduce the pressure over an area, Expensive.

2. Slam Patch

Z.

Measures pressure over arbitrary area, Can exclude or include FSI effect by design,

Mathematical process may not needed to deduce pressure over area, Propagation of pressure can be detected with array of S/P,

Cheap. Cons

Accuracy problem in manufacturing,

Propagation of pressure within panel area can not be detected, Difficult to install.

3. Strain gauge, accelerometer Pros

Structural response or acceleration is measured, Easy to install,

Cons

Deduce pressure from structural deflection or acceleration, Pressure over area is difficult to infer,

Need sophisticated filtering process.

In this study, two methods will be implemented to measure the hydro-impact pres-sure. PIT and S/P are the main instruments and additionally accelerometer is used to

extract extra information from the tests.

2.1.2

Consideration of Fluid-Structure Interaction

Every structure has its own natural characteristics such as the natural frequency and natural mode in vacuo or dry mode. It is also generally known that when a struc-ture is in contact or immersed in the water or wet mode, the natural characteristics of the structure change. Since the boat is in service on the water surface, part of hull structure is always immersed in the water and has natural characteristics in the wet mode. When the natural characteristics of a structure in dry mode transform to the different natural characteristics in wet mode, it is known as FSI or hydro-elastic effect.

dry mode drop and natural modes change into the ones in the wet mode.

However, with regard to the hydro-impact phenomenon, the nature of the phe-nomenon is an impulse which is strongly dynamic and transient. As the load is

dy-namic, the response of structure is also dynamic. One distinct difference between static and dynamic environment is 'frequency dependency' or 'relativity' between structure and excitation in terms of time such that the response depends on the duration time of excitation as the base of structure's natural behaviour. For example, when the duration time or frequency content of external load matches with one of the natural frequencies of the structure in a way, the response is magnified by the resonance phenomenon. The response of the structure under the hydro-impact is governed by the PSI effect and its relativity.

This FSI effect problem has been recently recognised in the yacht design commu-nity especially, for the trans-ocean racing yacht. At the moment, when designing the yacht's structure, there are no standard for the calculation of the transient response. Instead, idealised ultimate load of static is applied directly into the structure in the dry mode and hence the dynamic response is omitted and substituted by the static calculation. However, recent structural damages suggests that the design community should speculate as to the reasons for the damage and lead to the conclusion that they are the structural dynamics and FSI effect. Since the high frequency structural wave dissipates instantly within a short distance and low frequency structural wave travels relatively longer distances, in terms of global and local hydro-elastics, low frequency would dominates the global response and low and high frequencies would determine the local response.

To understand the dynamic response of a structure with the PSI effect included at a local basis, the structure is on the semi-infinite water and the transient load by hydro-impact experiment is then applied. For this purpose, the commercial Finite El-ement Analysis (FEA) program ANSYS is available to analyse the structure under the condition. In this program, the water or fluid domain is represented by the acoustic wave equation or its matrix form which then calculates the acoustical pressure. This

2.1.3

Experiment Setup

For the simplicity, the experimental setup is described in advance without the de-tails of the S/P. To investigate the characteristics of hydro-impact by slamming, drop test and seakeeping-slamming test are arranged on the model of an Open 60' racing yacht. The foreward bottom area of the model is selected and P/Ts and S/Ps are in-stalled. P/Ts are installed around the S/ Ps and where a S/P cannot be attached shown at Table 2.1 and Figure 2.1. Static calibration is carried out on P/Ts and S/Ps to de-termine the calibration factors. On the other hand, since the structure of the model is not scaled down from the full-scale boat and limitation makes it difficult to carry

out dynamic calibration, an accelerometer is excluded in measuring the hydro-impact signal to deduce the pressure. Instead, the accelerometer is used as an extra sensor to confirm the consistency of structural vibration and to deduce impact velocity on the basis of the manufacturer's calibration factor.

Modal testing is carried out to find the shape of FRF of force transmissibility. Fur-thermore, in the case of the seakeeping-slamming test, pendulum tests are performed to estimate the vertical centre of gravity and pitch radius of gyration to fit the model to the full-scale Open 60' yacht.

Drop test is arranged in the towing tank of the University of Southampton in such a way that the rear ends of the boat are fixed by strings and fore end by electro-magnetic release. By changing the vertical location of electro-magnetic, the drop height can be adjusted. Thus this is not a vertical drop test, instead it is rotational drop test which is more realistic at sea. A potentiometer and accelerometer are attached to measure the drop height, impact velocity and acceleration. The view of the test setup is shown at Figure 2.5.

For the seakeeping-slamming test, the boat is towed in the towing tank of the Southampton Solent University with the wave of up to 0.2 m which corresponds to 1.4 m full-scale wave. The wave frequencies are from 0.4Hz to 0.9Hz with increment of 0.05Hz where the resonance in pitch/heave motion exist as revealed in previous study [191. Three different boat speeds are arranged on the upright position of the boat at which the maximum hydro-impact pressure is expected.

The locations where P/Ts and S/Ps are installed are shown at Figure 2.1. Red and black dots are S/Ps' and P/Ts' respectively. The stations where the sensors are

Table 2.1: Stational location of PIT and S/P in drop and S-S tests

Station No. 2.25, 2.75, 3.25, 3.5, 4.25 2.5-3, 3-3.5, 3.5-4

Table 2.2 and Figure 2.2 are the summary of major instruments and their network. The measured signals are stored in computer and then processed in MATLAB using Fast Fourier Transformation (FFT) method. The further details of model tests are de-scribed in the later sections.

Table 2.2: Major measurement instruments and network

P/T

S/P

Instrument Model Manufacturer Range

I/O card PCM-DAS08 Measurement Computing

Data acquisition Labview 8.0 National Instruments

Data acquisition Turboad Wolfson Unit

PIT RDP A105 RDP ± 50 psi

S/P (L/C) Kistler 9712BE250 Kistler + 250 N

Accelerometer Endevco 2256-100 Endevco + 50 g

2.2

Slam Patch System

In this section, details of S/P are written with regard to design, dynamics and FSI effect.

2.2.1

Background of Slam Patch

The purpose of the S/P is to measure the external hydro-impact pressure over an arbitrary specified area. To achieve this goal, the system must give a constant gain of 1 within the frequency range of interest. For example, if the frequency content of an external pressure is up to 100Hz, the system must provide a constant gain of 1 within this frequency range of 100Hz such that the external pressure is identically measured without any additional effect which is mostly caused by resonance (and any of broad band excitation) of the measuring system. If there are any resonances in the system within the frequency range of interest, the external signal is magnified or distorted by the resonance and it is difficult to procure the identical signal. This is where the difficulty lies to design an adequate S/P system.

There are two ways to design the S/P to measure two different characters that, 1. When measuring the pure external pressure as P/T behaves,

The first resonance must be far beyond the frequency content of the external signal,

Within the range, the gain must be 1 to measure identical signal. When measuring total response as the accelerometer behaves,

If the S/P system is a model of the local structure of full-scale in term of

FRF, i.e. FRFs/p = FRFFu.SaIe, then measured signal is total response of the

local structure as accelerometer does. However, in this case, the system has a specified area of interest, the signal is limited in this area unlike the accelerometer,

Since the FRF is known, it is possible to separate the components of the signal into two components by filtering - the external pressure and coupling component (resonance component if resonance is exist in the system). In this study, both aspects are concerned, mainly concentrated to the former one. Structural design of the system is carried out based on statics and dynamics. After construction of the system, modal testing is carried out in dry and wet modes to verify

2. (a)

2.2.2

Design of Slam Patch

Since MARIN indicated a pressure of approximately 200 kPa and the designers of the Open 60' are using 350kPa including the materials safety factor [39] such that 250kPa is selected and scaled down to model to design SIR The panel area of S/P where the pressure will be applied is decided as 80mm long, 80inm wide and small bridge structure is taken to install load cell. The materials for the construction of the system are E-glass woven roving reinforcement and polyester matrix. After static structural analysis, following specification is determined for the construction.

Configuration,

Static design load : 35.7kPa in model-scale (250 kPa in full-scale), Configuration : pressure panel part + bridge part,

Scantling of pressure panel : 80 x 80 x 2 (mm) (square plate x minimum

thickness),

Scantling of bridge : 25 x 25 x 15 x 2 (mm) (length x width x height xi minimum thickness).

Materials,

Reinforcement : 293 g/m2 E-glass woven roving, Matrix,: polyester resin.

Lamination,

Laminating method.: hand layup + vacuum,

Laminate orientation : 8 laminates with ±45, 0/90, -1145, 0/90, 0/90, ±45, 0/90, ±45 (symmetry),

Fibre weight fraction (IITf) : 0.5, Matrix weight fraction (Wm): 0.5.

Mechanical properties refer to following Table 2.3

Table 2.3: Mechanical properties of laminate of S/P

2.2.3

Dynamics of Slam Patch in Dry Mode

Numerical calculation using ANSYS is carried out to find the natural behaviour of S/P with suitable boundary condition including a connecting rod. The mechanical properties are the same as the static calculation. Since, ANSYS does not provide an element for woven laminate, a woven laminate is assumed as a combination of two unidirectional laminates using the SHELL99 element and the connecting rod is mod-eled by the SOLID92 element. However, shear modulus and Possion's ratio in the minor direction i.e. C, x.iy:, are needed using SHELL99 element and for this purpose, the properties are assumed by Halpin-Tsai equation [16]. Modeling of S/P and mode shapes are provided at Appendix C. Table 2.4 are natural frequencies up to four modes with various connecting rod lengths.

Table 2.4: Natural frequencies of S/P in dry mode with various rod lengths

It can be seen from the calculation result that the natural frequency of the panel part which is a square plate is high enough to avoid any expected resonance of up to 200Hz. Since it was expected in the first design stage of S/P that the natural behaviour of the panel part may play a role in the behaviour of S/P but it is shown that the critical part in S/P design is on the length of connecting rod and the S/P actually behaves like cantilever beam-column. So it can be said that the natural behaviour of the connecting rod is a decisive factor in the design of S/P. Furthermore, the stiffeners used in S/P to prevent the bending mode of panel is found by additional calculation, with or without

Mode No. L = 3mm L = 5mm L = 7mm Mode

1 574.7 525.7 485.8 Bending of rod

751.2 655.6 582.5 Bending or rod

3 790.3 675.6 595.2 Twisting of rod

4 1360.5 1360.4 1360.3 Bending of panel

Mechanical properties E-glass Polyester Unidirectional laminate

Young's modulus, E (GPa) 72 2.5 15.8

Shear modulus, G (GPa) 29 3.4 4.09

Poisson's ratio, v 0.2 0.38 0.345

Density p (kg 17-n3) 3800 900 1455

2.2.4

Fluid-Structure Interaction of S/P in Wet Mode

Let's assume a FSI system, where a structure is in contact with the water, floating or submerged. In the FSI system, the natural frequencies of the structure change - re-ducing because of the effect of the water. The water acts as an added mass [44] and this is generally known as one aspect of FSI effect. Another aspect can be detected in terms of the mode shapes changing. Other aspect involved in hydro-impact phenomenon is such that when the external impulse is applied on the structure in the wet mode, the response of structure in FSI system is dependent on a relativity between the natural frequencies of the structure in wet mode and duration time or frequency content of external impulse as in the general dynamics.

Since S/P works in the wet mode, the reducing of the natural frequencies must be known to satisfy the purpose. In other words, if the purpose of S/P is to measure the pure hydro-impact pressure, the dropped natural frequencies must be beyond the frequency content of the external impulse. If it is to measure the FSI effect between an hydro-impact pressure and a structure, the S/P must represent the structure at full-scale by an adequate scaling law.

In this study, the main purpose of the S/P is to measure the pure hydro-impact pressure, since it is difficult to make S/P represent a local structure using an adequate scaling law, especially the damping factor of the structure. For this reason, it can be said that the S/P in this case is simply a bigger PIT. The target remained is that the first natural frequency of SIP can be relatively high enough to the maximum frequency of hydro-impact pressure to avoid any resonance.

The same method and cases in dry mode calculation are recalculated with the wa-ter in contact to the S/P. As mentioned earlier, the fluid domain in this case uses the acoustic wave equation. Table 2.5 is the calculation result and modeling and mode shapes can be found in Appendix C.

Table 2.5: Natural frequencies of S/P in wet mode with various rod lengths

Mode No.

L = 3mm L = 5mm L 7mm

1 417.3 382.1 353.0 Bending of rod

By comparing Table 2.4 with Table 2.5, it can be concluded that the natural frequen-cies at each mode reduce consistently, but mode shapes are found to be unchanged. Although only the natural frequencies are found in these calculations, the shape of Frequency Response Function (FRF) is not known. In S/P design, since the frequency range of interest must be remained in the gain of 1 to measure identical external pres-sure, the shape of FRF must be known which is sensitive to damping factor.

2.3

Modal Testing

The aim of modal testing is to find the natural characteristics of S/P, especially FRF. Impactor testing is carried out which measures response force at S/P and driving force at the impactor when an impact is applied on the S/P. The input is an impul-sive force and the output force is measured at the load cell on the S/P Since input and output are both forces, the FRF is the force transmissibility (Tf). In the case of dry mode, point transmissibility' and transfer transmissibility2, especially the direct transmissibility3 are measured to compare the results to the one in wet mode, because, in wet mode, only direct transmissibility measurement is possible. The input and out-put data are processed in MATLAB and the FRFs are generated. Modal testing results are in Appendix D.

2.3.1

Dry Mode

Figure D.1 to Figure D.12 in Appendix D are impactor test results. Upper graphs are force transmissibilities in linear scale and lower graphs are phase angles. The fre-quency range is up to 2150Hz.

From the figures, within the range of 0 to 500Hz, it can be said that the S/P has an approximate gain of 1. However, the first natural frequencies of all S/P are a bit higher than the calculated results. This is probably due to the manufacturing error of S/P and connecting rod length - actual values are a bit higher than the design. 2.3.2 Wet Mode

Direct point measurement carried out because of space constraint and existence of the water. To carry out the point measurement, the impacting must be carried out un-der the water which is impossible in this case. Figure D.13 to Figure DIM in Appendix D are the direct point test results in wet mode. Like the test results in dry mode, upper one is force transmissibility in linear scale and below one is phase angle. Black bold lines in graphs are average FRF. Consistently, natural frequencies drop and it can be said that at least, up to 250Hz, the gain is 1 or near around. So it can be concluded that within this 250Hz range, measured pressure signal is identical to the external pressure.

2.4

Slamming Test

Drop tests have been carried out by many people around the world to study an re-lationship among any parameters [20] [28] [51] [831. Most of the parameters are drop height, impact velocity and the result is pressure.The objective of this drop test is also consistent with them but the object is the pressure on Open 60' yacht's hull.

The main objectives of the drop test are, To find the practicability of the SIP,

To find the reconciliation of S/P and P/T,

To find the relationship between hydro-impact pressure and drop height or im-pact velocity,

To confirm the test configuration before carrying out the seakeeping-slamming test.

The model is hung by strings at its stern and an electro-magnetic release system at its bow. The locations of P/Ts and S/Ps are shown at Figure 2.4 and the Figure 2.5 is the test scene.

By powering off the electro-magnet, the boat rotates at its stern and drops into the water. Strings are used to exclude any noise which can be detected when frame structure is used to hold the model at the stern. When the fore body of the boat drops on the water surface, corresponding pressure, drop acceleration and displacement are measured through sensors - S/P, PIT, accelerometer and potentiometer. The main parameters in this drop test are drop height from water surface to bottom of model. The drop heights start from 0 to around 40cm.

2.4.2

Seakeeping-Slamming Test

The objective of seakeeping-slamming test is to acquire the hydro-impact char-acteristics in the real situation. Expected hydro-impact characters of slamming and related parameters are,

Hydro-impact characteristics, Hydro-impact pressure (Pp),

Shape of hydro -impact pressure in time domain. Parameters,

Boat speed (V,,,), Wave frequency (Wf),

Wave height (147h),

Location of measurement (LOC).

Finding the relationship between the hydro-impact pressure and its associated pa-rameters are the primary objective of this experiment, however, finding the maximum or extreme behaviour of the hydro-impact phenomenon is another objective. For this purpose, a set of experiments is arranged in the towing tank of Southampton Solent University. Before carrying out the tests, the displacement of the model is scaled down from the full-scale boat as in the drop test. Furthermore, to correspond the pitch radius of gyration of model to the full-scale boat, pendulum tests in roll and pitch motions are carried out and the locations of adjustable ballast weights are decided by measuring the associated oscillation time [45]. The wave and carriage speed are also confirmed again before the test.

The model is tested in an upright position in which the maximum pressure is likely to occur. The maximum wave height achievable in this towing tank is limited within 0.2m which corresponds to 1.4m high at scale. This wave height of 1.4 m in full-scale may be not enough compared to the testimonies from sailors or even contradic-tory to their descriptions. One of the reports says the slamming starts from Force 6 in Beaufort Wind Scale at the boat speed of 10 knots [lit The likely venue of slamming are Cape Hope in Africa and Cape Horn in South America where current and wave direction are opposite.

The profile of Force 6 is,

Mean wind speed = 12.35 m/s, Mean significant wave height = 3 m, Mean wave period = 6.69 s.

the actual wave height and wave period can be changed, and Force 6 wave contains various wave spectrum, acquiring the exact wave profile under a boat speed is diffi-cult assignment. In this study, it is assumed that the slamming begins when there is resonance in pitch/heave motion, not by the wave shape, i.e. the boat is excessively excited in pitch/heave motions when a certain waves are encountered. On the other hand, because of the limit with the experiment facilities, realistic generation of exper-iment environment belongs to further work.

For this reasons, the range of parameters are selected as Table 2.6 and the coin, binations are taken into experiment on the upright boat position and head sea. The increment of wave frequency in model-scale is 0.05 Hz.

Table 2.6: Test matrix of S-S test

The experimental setup is the same as the drop test and extra instruments of wave probe and potentiometers for heave/pitch motions are added and connected to MT-BOAD system. The locations of measurements are shown at Figure 2.6 and Figure 2.7 is the experiment scene.

2.4.3

Initial Assessment

The capacity limit is found at S/P during the drop test. Figure 2.8 shows the limit. However, in the S-S test, the signals are well below the limit.

Below Figure 2.9 is typical signal patterns during the drop test. Corresponding lo-cations of P/Ts and S/Ps are shown at Figure 2.4 where the impact location is station 3 The signals from S/Ps must show similar patterns to the ones of P/Ts if the design and construction are carried out properly. However, flexibility of the S/P system and its gain in the FRF make it difficult to obtain the similar patterns and consequently. This makes the peak pressure at S/Ps higher than the P/Ts up to 40%.

In Figure 2.10 are the signals at relatively low drop height at the same location and

Model-scale Full-scale

Boat speed 1.5, 2, 2.5 m/s 7.71, 10.28, 12.85 knots

Wave frequency 0.4 0.9 Hz 0.15 0.34 Hz

Encounter frequency 0.4245 1.0238 Hz 0.1604 0.3869 Hz

of SIP in the Figure 2.10 is filtered from 200Hz and the filtered peak pressure at S/P shown at the Figure 2.11 and Figure 2.12 reveals a similar magnitude as PIT.

situation also occurs in S-S test, especially when the motion is rigorous as shown at Figure 2.13.

In this initial assessment of the signals, it can be concluded that,

In the frequency domain of P/T, most of power measured are concentrated within 200Hz range,

In the drop test, even the drop height is small, S/P overestimates the peak pres-sure up to 40% higher compared to the P/T's. This is becasue of the flexibility of the SIP, i.e. the resonance frequency range is exicited,

In the S-S test, the overestimation of peak pressure shows a similar pattern as in the drop test,

At least, the frequency range of gain 1 is expected to be extended up to 500Hz or even further if the case is to measure the pure hydro-impact signal,

In this case of S/P, it can represent a local structure with a specific FRF.

Similar

1..

Figure 2.1: Locations of PIT and S/P in drop and S-S tests

Labview (S/R : 3125 Hz) Turboad (S/R : 1553 Hz)

Signal Process (MATLAB)

Figure 2.2: Instrument networking

(a) Global view (b) Local view

S P 1

S P2

PTI

ACC POT

Connecting rod Bridge pat

t.

Figure 2.3: Schematic view .and prototype of ,S/P.

(a) Schematic view

SwF.' 'Sal tiC- 42%. Ca'''. a

a

e

i/13 1

+Pir

Pr!" 2

Fir 3

PiT 4

Figure 2.4: Locations of PIT and S/P in drop test 3 1 2

STA. 3

STA. 4

STA. 5

vE.

I

4`4,.4%.

SP)

-7/

Figure 2.5: Drop test setup

7-.7Stil VI

tor

SfP 1

SIP 2

S/P 3

FiT 1

PiT 2

PiT 4

Figure 2.6: Locations of P/T and S/P in S-S test

STA. 2

STA, 3

STA, 4

STA. 5

II

'Pr

a

ale

Figure '2.7:'2.7: Seakeeping-slamirtg test

.r

411

nito

3grt.tz..

20 15 a. E. 10 Li 5 0 30 2 10 20 8 ta. 15 10 5 00 0 oo

TIME DOMAIN OF PIT

150 100

50

Zi3 0

FREQUENCY DOMAIN OF PIT

500 1000 1500 2000

Frequency Hz]

FREQUENCY DOMAIN OF SIP

SIP 1

S/P 2 SIP 3

500 1000 1500 Frequency [Hz]

Figure 2.9: Signal patterns at P/T and S/P in drop test

2000

45

DROP HEIGHT VS PEAK PRESSURE AT 3/P1

40 *el au 0 o o0 04 0 0 ePie e000 0 o OD 00 at, CD 00 0 35 0 8 30 Lu ,cc 25 0.128 0.13 0.132 0.134 Time [s] TIME DOMAIN OF SIP

0.138 S/P 1 S/P 2 S/P 3 oa COS

fl

Figure 2.8: Limit of pressure capacity of S/P

0.13 0.14 0.15 Time MI P/T 2 P/13 PIT 4 0.136 50

6 2 0 14. 1 2 4010

SLOUTM1IAI 1111160.11IPACT 3101141. AT Pik

(a) At PIT

5E00E111140 11Y0140411PACr SIGNAL AT VP.

1026 4030 SEQUENCE (b) At S/P -34P 6+14 -ti/P 3 4:123 4000 4035 4010 4046 SEQUENCE

3

ZO

POWER SPECTRA!. 00 05115 ESTIMATE OF SRI

Foil%

Filming

820 FREQUENCY

POWER SPECTRAL DENSITY ESTIMATE OF WITS

0.11re 111111110 --- Mat ----api SIP 2 SIPS 401 6E0 HE 1603 IOU 1402 1693 FNECUENC,

Figure 2.11: Spectrums of S/Ps at low drop height

FILTERED SETWEROAL SISMAI OE STI's

E03 1011/ 1200 1./01/6013

FREQUENCY [1111 POWER SPECTRAL DENSITY ESTIMATE OF 942

4091 4010 4020 430 .4093

SEQUENCE

Figure 2.12: Time signals of S/Ps after filtering at low drop height

40 35 30 25 10 -5 -60 -80 -140 -180 -180 200

TIME HISTORY AT SLAM PATCH

(a) Time domain

POWER SPECTRUM OF SLAM PATCH

- CASE 1 -CASE 2 CASE 1 CASE 2 600 800 1000 FREQUENCYIHz1 (b) Frequency domain 1200 1400 1600 0.1 0.2 0.3 0.4 0.5 0.8 TIME Is] 0 -20 400 200