Practical application of viscous-flow

calculations for the

simulation of manoeuvring ships

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus prof.ir. K.C.A.M. Luyben,
voorzitter van het College voor Promoties,
in het openbaar te verdedigen op maandag 9 mei 2011 om 15:00 uur

door

Serge Leon TOXOPEUS

Maritiem Ingenieur
geboren te Alkmaar
ii

Dit proefschrift is goedgekeurd door de promotor:

Prof. dr. ir. R.H.M. Huijsmans

Samenstelling promotiecommissie:
Rector Magnificus, voorzitter
Prof. dr. ir. R.H.M. Huijsmans, Technische Universiteit Delft, promotor
Prof. dr. ir. M. Vantorre, Universiteit Gent
Prof. Dr.-Ing. A. Cura Hochbaum, Technische Universität Berlin
Prof. dr. A.E.P. Veldman, Rijksuniversiteit Groningen
Prof. B. Pettersen, Norwegian University of Science and Technology
Prof. ir. J.J. Hopman, Technische Universiteit Delft
Dr. S.R. Turnock, University of Southampton
Prof. dr. ir. T.J.C. van Terwisga, Technische Universiteit Delft, reservelid

Copyright c 2011 by S.L. Toxopeus, Wageningen, The Netherlands. All rights reserved.
Published by the Maritime Research Institute Netherlands (MARIN)
ISBN 978-90-75757-05-7 (print)
 iii

Summary
Practical application of viscous-flow calculations for the simulation of
manoeuvring ships

Figure 1: Impression of the flow field and hull surface pressures, KVLCC2, β = −10◦

The present work was initiated in order to improve traditional manoeuvring simula-
tions based on empirical equations to model the forces and moments on the ship. With
the evolution of the capability of viscous-flow solvers to predict forces and moments on
ships, it was decided to develop a practical method to simulate the manoeuvrability of
ships in which viscous-flow solvers are utilised and to investigate whether this improves
the accuracy of manoeuvring predictions.
To achieve this goal, the virtual captive test approach is adopted, because of the effi-
cient use of computational resources compared to other methods. This procedure mimics
the approach for manoeuvring simulations in which experimental PMM is used to obtain
the forces and moments on the ship. This study extends the work of other researchers
by providing extensive verification and validation of the predicted forces and moments on
the hull and a detailed study of the sensitivity of the manoeuvring characteristics of the
ship to changes in the hydrodynamic coefficients in the simulation model.
Changes in the flow solvers were required to be able to calculate the flow around
iv Summary

ships in rotational motion. These changes are discussed as well as the acceleration tech-
niques that were developed to reduce the effort spent on grid generation and during the
computations.
In this thesis, it is demonstrated that good predictions of the loads on the hull in
manoeuvring motion can be obtained for a wide range of ship types. The trends in the
forces and moments as a function of the drift angle or yaw rate are simulated well.
The verification studies provide useful insight into the influence of grid density on
the predicted forces and moments. In several cases, validation of the calculations failed,
indicating modelling errors in the numerical results. In these cases, it was generally seen
that the magnitude of the transverse force was under-predicted, while the magnitude of
the yaw moment was over-predicted. For manoeuvring studies in the early design, the
comparison errors are within acceptable levels. However, improvements remain desired
and may be obtained using finer grids, larger domain sizes, different grid topologies with
refinement in the wake of the ship, other turbulence models or incorporating free surface
deformation.
The manoeuvring prediction program SurSim has been used to simulate the manoeu-
vrability of the HTC. A procedure is proposed to derive the hydrodynamic coefficients
required to model the forces and moments on the bare hull. This procedure is chosen to
enable accurate modelling of the linearised behaviour for course-keeping as well as realis-
tic modelling of the harbour manoeuvring characteristics, and to enable the modelling of
non-linear manoeuvres accurately.
To generate validation data for the manoeuvring predictions presented in this thesis,
free sailing manoeuvring tests for the HTC were performed. This test campaign resulted
in a very valuable data set which can be used for public validation studies. Besides
obtaining general characteristics of the manoeuvrability of a single-screw container ship,
unique information has been obtained on the drift angles and rates of turn combined with
propeller and rudder forces. Furthermore, repeat tests have been conducted for selected
manoeuvres. Based on these tests, the uncertainty in the characteristic manoeuvring
properties has been estimated.
By using hydrodynamic manoeuvring coefficients derived from the CFD calculations,
it has been shown that it is possible to improve the prediction of ship manoeuvres com-
pared to predictions using coefficients based on empirical equations. A considerable im-
provement in the turning circle predictions was obtained. The prediction of the yaw
checking and course keeping and initial turning abilities based on zig-zag simulations im-
proved as well, but further improvements are required for more reliable assessment of the
manoeuvring performance.
The sensitivity of the manoeuvring predictions to changes in the hydrodynamic co-
efficients was studied. It was found that for accurate predictions of the manoeuvrability
using coefficients derived from CFD calculations, accurate predictions of especially the
yawing moment must be made.
 v

Contents

Summary iii

Review of tables viii

Review of figures x

List of Symbols xiii

Acronyms xix

1 Introduction 1
1.1 Problem definition and objectives . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Outline of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Background of manoeuvring simulation 5

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Standard mathematical models . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Virtual captive tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4 RANS coupled to ship motions . . . . . . . . . . . . . . . . . . . . . . . . 10
2.5 Validation cases for captive manoeuvring . . . . . . . . . . . . . . . . . . . 12
2.5.1 Series 60 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5.2 DARPA SUBOFF . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.5.3 KVLCC1, KVLCC2 and KVLCC2M . . . . . . . . . . . . . . . . . 14
2.5.4 HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5.5 Other cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.6 Validation cases for free sailing manoeuvring . . . . . . . . . . . . . . . . . 21
2.6.1 Esso Osaka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6.2 KVLCC1, KVLCC2 . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6.3 KCS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6.4 5415M . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
vi Contents

3 Mathematical model 25
3.1 Coordinate system and nomenclature . . . . . . . . . . . . . . . . . . . . . 25
3.2 Manoeuvring simulation program SurSim . . . . . . . . . . . . . . . . . . 26
3.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.2 Equations of motion . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.3 Hull forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.4 Propeller forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2.5 Rudder forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3 Deriving the hydrodynamic coefficients for the user-defined hull forces . . . 34
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4 Viscous flow solvers 37

4.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2 Parnassos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.3 ReFRESCO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4 Turbulence closure models . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.5 Implementation of rotational motion . . . . . . . . . . . . . . . . . . . . . 38
4.5.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.5.2 Governing equations . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.5.3 Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.6 Grid generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.6.1 Parnassos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.6.2 ReFRESCO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.7 Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

5 Verification and validation of steady motion calculations 45

5.1 Introduction to verification and validation . . . . . . . . . . . . . . . . . . 45
5.2 KVLCC1, KVLCC2 and KVLCC2M . . . . . . . . . . . . . . . . . . . . . 47
5.2.1 Iterative error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.2.2 Discretisation error . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.2.3 Local quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.2.4 Global quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.2.5 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.3 HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.3.1 Iterative error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.3.2 Discretisation error for steady drift motion . . . . . . . . . . . . . . 56
5.3.3 Discretisation error for steady yaw motion . . . . . . . . . . . . . . 62
5.3.4 Local quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3.5 Global quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.3.6 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.4 DARPA SUBOFF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Contents vii

5.4.1 Iterative error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

5.4.2 Discretisation error . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.4.3 Local quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.4.4 Global quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.4.5 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.5 Walrus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.5.1 Iterative error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.5.2 Discretisation error . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.5.3 Global quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

6 Free sailing manoeuvring tests 93

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.2 Ship model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.3 Propulsion and steering system . . . . . . . . . . . . . . . . . . . . . . . . 94
6.4 Experimental facility and measurement system . . . . . . . . . . . . . . . . 95
6.5 Data reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6.6 Test procedures and programme . . . . . . . . . . . . . . . . . . . . . . . . 96
6.7 Uncertainty analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.7.2 Zig-zag tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.7.3 Turning circle tests . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

7 Simulation of ship manoeuvrability 103

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
7.2 Deriving the hydrodynamic coefficients . . . . . . . . . . . . . . . . . . . . 104
7.2.1 Resistance curve, wake fraction and thrust deduction fraction . . . 105
7.2.2 Propeller characteristics . . . . . . . . . . . . . . . . . . . . . . . . 105
7.2.3 Rudder forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7.2.4 Hull forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
7.3 Standard manoeuvres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.3.1 Programme of simulations . . . . . . . . . . . . . . . . . . . . . . . 112
7.3.2 RPM-Speed curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
7.4 Sensitivity study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
7.5 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
7.5.1 Zig-zag manoeuvres . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
7.5.2 Turning circle manoeuvres . . . . . . . . . . . . . . . . . . . . . . . 119
7.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
viii Contents

8 Conclusions and recommendations 123

8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
8.2 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

References 127

Samenvatting 143

Acknowledgements 145

Curriculum Vitae 147

Appendices 149
Table pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Figure pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
 ix

Review of tables

2.1 Main particulars of Series 60 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.2 Main particulars of DARPA SUBOFF submarine . . . . . . . . . . . . . . . . . . 14
2.3 Main particulars of KVLCC1, KVLCC2 and KVLCC2M . . . . . . . . . . . . . . 15
2.4 Main particulars of HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5 Main particulars of the HTC propeller . . . . . . . . . . . . . . . . . . . . . . . . 20
2.6 Particulars of the HTC rudder . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.1 Symbols used for rudder forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

5.1 Properties of grids for different drift angles, KVLCC2M, Parnassos . . . . . . . 49

5.2 Properties of grids for uncertainty analysis, KVLCC2M, Parnassos, β = 12◦ . . 49
5.3 Uncertainty analysis, KVLCC2M, Parnassos, β = 12◦ . . . . . . . . . . . . . . 50
5.4 Measurement uncertainties and validation, KVLCC2M, Parnassos, β = 12◦ . . 54
5.5 Properties of grids for different drift angles, HTC, Parnassos, 10 kn . . . . . . . 55
5.6 Properties of grids for uncertainty analysis, HTC, β = 10◦ , Parnassos . . . . . 57
5.7 Uncertainty analysis, HTC, β = 10◦ , 10 kn, Parnassos . . . . . . . . . . . . . . 57
5.8 Properties of grids for uncertainty analysis, HTC, 18 kn, ReFRESCO . . . . . . 59
5.9 Uncertainty analysis, HTC, β = 10◦ , 18 kn, ReFRESCO . . . . . . . . . . . . . 59
5.10 Uncertainty analysis, HTC, β = 30◦ , 18 kn, ReFRESCO . . . . . . . . . . . . . 60
5.11 Properties of grids for uncertainty analysis, HTC, γ = −0.2, Parnassos . . . . . 62
5.12 Uncertainty analysis, HTC, γ = −0.2, 10 kn, Parnassos . . . . . . . . . . . . . . 62
5.13 Uncertainty analysis, HTC, γ = 0.2, 18 kn, ReFRESCO . . . . . . . . . . . . . . 64
5.14 Uncertainty analysis, HTC, γ = 0.4, 18 kn, ReFRESCO . . . . . . . . . . . . . . 65
5.15 Validation, HTC, 10 kn, Parnassos . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.16 Validation, HTC, 18 kn, ReFRESCO . . . . . . . . . . . . . . . . . . . . . . . . 72
5.17 Properties of grids for uncertainty analysis, SUBOFF, β = 0◦ , 18◦ . . . . . . . . 74
5.18 Uncertainty analysis, SUBOFF, β = 0◦ . . . . . . . . . . . . . . . . . . . . . . . . 75
5.19 Uncertainty analysis, SUBOFF, β = −18◦ . . . . . . . . . . . . . . . . . . . . . . 76
5.20 SUBOFF, longitudinal force X, Re = 1.4 × 107 , β = 0◦ . . . . . . . . . . . . . . . 81
5.21 SUBOFF, longitudinal force X, Re = 1.4 × 107 , β = 18◦ . . . . . . . . . . . . . . 81
5.22 SUBOFF, transverse force Y , Re = 1.4 × 107 , β = 18◦ . . . . . . . . . . . . . . . 82
5.23 SUBOFF, yawing moment N , Re = 1.4 × 107 , β = 18◦ . . . . . . . . . . . . . . . 82
5.24 Validation, SUBOFF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.25 Designations and descriptions of Walrus experimental configurations . . . . . . . 85
5.26 Uncertainty analysis, Walrus, β = 0◦ , α = 0◦ , deep water . . . . . . . . . . . . . 87
x Review of tables

5.27 Walrus grid refinement study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

6.1 Experimental uncertainty estimate, zig-zag, 95% confidence interval, all tests . . 100
6.2 Experimental uncertainty estimate, zig-zag, 95% confidence interval, set 1 . . . . 101
6.3 Experimental uncertainty estimate, turning circle, 95% confidence interval, all tests101
6.4 Experimental uncertainty estimate, turning circle, 95% confidence interval, set 1 102

7.1 Summary of zig-zag manoeuvre results, original simulations . . . . . . . . . . . . 104

7.2 Summary of turning circle manoeuvre results, original simulations . . . . . . . . 104
7.3 Propeller No. 5286, open water test No. 45127 . . . . . . . . . . . . . . . . . . . 106
7.4 Hydrodynamic bare hull and added mass coefficients, HTC . . . . . . . . . . . . 110
7.5 Simulation matrix, HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.6 Sensitivity study, HTC, 10 kn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
7.7 Summary of zig-zag manoeuvre results . . . . . . . . . . . . . . . . . . . . . . . . 117
7.8 Summary of turning circle manoeuvre results . . . . . . . . . . . . . . . . . . . . 120
 xi

Review of figures

1 Impression of the flow field and hull surface pressures, KVLCC2, β = −10◦ . . . iii

2.1 Flow chart of manoeuvring simulations . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Body plan of Series 60 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Geometry of DARPA SUBOFF (AFF-1) . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Body plans of KVLCC1 and KVLCC2 . . . . . . . . . . . . . . . . . . . . . . . . 15
2.5 Body plans of KVLCC2 and KVLCC2m . . . . . . . . . . . . . . . . . . . . . . . 16
2.6 Teresa del Mar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.7 Body plan of HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.8 HTC model during oblique motion test using CPMC . . . . . . . . . . . . . . . . 19
2.9 Drawing of the HTC rudder, T =10.3 m . . . . . . . . . . . . . . . . . . . . . . . 21

3.1 Ship-fixed coordinate system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Nomenclature and sign convention for rudder forces . . . . . . . . . . . . . . . . 30
3.3 The axial actuator disk model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.1 Example grid, Parnassos, HTC, γ = 0.556 . . . . . . . . . . . . . . . . . . . . . 41

4.2 Example grid, ReFRESCO, HTC . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.3 Verification of drift sweep procedure, SUBOFF . . . . . . . . . . . . . . . . . . . 43

5.1 Convergence history Y -force, KVLCC2, ReFRESCO . . . . . . . . . . . . . . . 48

5.2 Uncertainty analysis, KVLCC2M, Parnassos, β = 12◦ . . . . . . . . . . . . . . 51
5.3 Comparison with experiments, KVLCC2M, WAKE1 . . . . . . . . . . . . . . . . 52
5.4 Comparison between experiments and calculations, KVLCC hull forms . . . . . . 54
5.5 Convergence history Y -force, HTC, ReFRESCO . . . . . . . . . . . . . . . . . . 56
5.6 Uncertainty analysis, HTC, β = 10◦ , 10 kn, Parnassos . . . . . . . . . . . . . . 58
5.7 Uncertainty analysis, HTC, β = 10◦ , 18 kn, ReFRESCO . . . . . . . . . . . . . 60
5.8 Uncertainty analysis, HTC, β = 30◦ , 18 kn, ReFRESCO . . . . . . . . . . . . . 61
5.9 Uncertainty analysis, HTC, γ = −0.2, 10 kn, Parnassos . . . . . . . . . . . . . . 63
5.10 Uncertainty analysis, HTC, γ = 0.2, 18 kn, ReFRESCO . . . . . . . . . . . . . . 64
5.11 Uncertainty analysis, HTC, γ = 0.4, 18 kn, ReFRESCO . . . . . . . . . . . . . . 65
5.12 Comparison between ReFRESCO and Parnassos, HTC, x = 0.48Lpp . . . . . 66
5.13 Comparison with experiments, HTC, x = −0.48Lpp . . . . . . . . . . . . . . . . . 67
5.14 Comparison between experiments and calculations, HTC, 18 kn, steady drift . . . 69
5.15 Comparison between experiments and calculations, HTC, 18 kn, steady yaw . . . 70
5.16 Uncertainty analysis, SUBOFF, β = 0◦ . . . . . . . . . . . . . . . . . . . . . . . . 75
xii Review of figures

5.17 Uncertainty analysis, SUBOFF, β = −18◦ . . . . . . . . . . . . . . . . . . . . . . 76

5.18 Pressure (top) and friction (bottom) coefficients along the hull, β = 0◦ . . . . . . 77
5.19 Velocities (top) and Reynolds stresses (bottom) at x = 0.978Loa , β = 0◦ . . . . . 78
5.20 Pressure (top) and friction (bottom) coefficients along the hull, β = 2◦ . . . . . . 79
5.21 Velocities (top) and Reynolds stresses (bottom) at x = 0.978Loa , β = 2◦ . . . . . 80
5.22 Comparison between experiments and calculations, SUBOFF, steady drift . . . . 84
5.23 Iterative convergence, Walrus, β = 0◦ , α = 0◦ . . . . . . . . . . . . . . . . . . . . 86
5.24 Convergence history Y -force, Walrus, ReFRESCO . . . . . . . . . . . . . . . . . 86
5.25 Uncertainty analysis, Walrus, β = 0◦ , α = 0◦ , deep water . . . . . . . . . . . . . 88
5.26 Comparison between configuration 2 experiments and calculations, Walrus . . . . 90

6.1 HTC ship model for free sailing tests . . . . . . . . . . . . . . . . . . . . . . . . . 94

6.2 Overview of Seakeeping and Manoeuvring Basin . . . . . . . . . . . . . . . . . . 95
6.3 HTC ship model during free sailing manoeuvring test . . . . . . . . . . . . . . . 97

7.1 Comparison between the original simulations and the free sailing experiments, 18 kn103
7.2 Estimated resistance curve (model scale values) . . . . . . . . . . . . . . . . . . . 105
7.3 Estimated propeller open water curves . . . . . . . . . . . . . . . . . . . . . . . . 106
7.4 Forces on the ship as function of rudder angle, HTC without propeller . . . . . . 107
7.5 Forces on the ship as function of rudder angle and propeller revolutions . . . . . 108
7.6 Forces on the ship as function of rudder angle, 18 kn, β = −10◦ , n = ns . . . . . 109
7.7 Forces on the ship as function of rudder angle, 18 kn, γ = 0.4, n = ns . . . . . . . 110
7.8 Forces on the bare hull as function of yaw rate or drift angle . . . . . . . . . . . . 111
7.9 RPM-Speed relation for HTC, scale 1:30.02, model scale values . . . . . . . . . . 113
7.10 Sensitivity study, HTC, zig-zag, 10 kn . . . . . . . . . . . . . . . . . . . . . . . . 114
7.11 Sensitivity study, HTC, turning circle, 10 kn . . . . . . . . . . . . . . . . . . . . . 115
7.12 Comparison between the simulations and the free sailing zig-zag experiments, 10 kn118
7.13 Comparison between the simulations and the free sailing 20◦ /20◦ zig-zag experi-
ments, 18 kn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
7.14 Comparison between the simulations and the free sailing turning circle experi-
ments, 18 kn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

8.1 Omstroming en druk op het rompoppervlak, KVLCC2, β = −14◦ . . . . . . . . . 143

 xiii

List of Symbols

drift angle at origin (midship) = arctan uv

β [rad]
 
βG drift angle at centre of gravity = arctan uvGG [rad]

βstc drift angle during steady turning circle [rad]

δ (mechanical) rudder angle [rad]

δD error in the experimental value [various]

δinput error caused by errors in the input parameter [various]

δmodel modelling error in the simulated value [various]

δmodel modelling error [various]

δreq required rudder angle [rad]

δSN numerical error in the simulated value [various]

δ̇ rudder turning rate [rad/s]

δ̇max maximum rudder turning rate [rad/s]

 
rLpp
γ non-dimensional yaw rate = V [−]

Λ rudder aspect ratio [−]

λ scale factor [−]

µ dynamic viscosity (= ρν) [kg/(sm)]

∇ displacement volume moulded [m3 ]

ν kinematic viscosity (= µ/ρ) [m2 /s]

Ω vector of rotation [rad/s]

φ roll angle [rad]

φ variable used in verification [various]

φ arithmetic mean or average of realisations of result φ [various]

xiv List of Symbols

φ0 extrapolated value of variable φ for infinitely fine grid [various]

φ1 value of variable φ for finest grid [various]

φexact exact solution for variable φ [various]

ψ yaw angle [rad]

ρ density of the fluid [kg/m3 ]

θ pitch angle [rad]

A0 propeller disc area [m2 ]

AE expanded propeller blade area [m2 ]

AR rudder area [m2 ]

B breadth max. moulded [m]

c bottom clearance [m]

 
∇
Cb block coefficient = Lpp BTm [−]
 
~
τ
Cf friction coefficient = 1
ρV 2 [−]
2 ∞

 
p−p∞
Cp pressure coefficient = 1 2
ρV∞
[−]
2

Cdb flow straightening factor for drift motion [−]

Cdr flow straightening factor for yaw motion [−]

Crue propeller-rudder interaction coefficient [−]

D experimental value [various]

Dp propeller diameter [m]

E comparison error (= S − D) [various]

fi force per unit volume [N/m3 ]

Fs safety factor in uncertainty quantification [-]

 
Fn Froude number = √V [−]
gLpp

G centre of gravity [−]

GM metacentric height [m]

h water depth [m]

J advance coefficient [−]

List of Symbols xv

K roll moment around intersection of waterplane/centreplane [N m]

 
Qp
KQ torque coefficient = n2 Dp5
[−]
 
Tp
KT thrust coefficient = n2 Dp4
[−]

L2 root-mean-square of change of variable between iterations [various]

L∞ maximum change/value of variable between iterations [various]

Loa length overall [m]

Lpp length between perpendiculars [m]

M pitch moment around intersection of midship/waterplane [N m]

m mass [kg]

mij added mass/inertia in direction i due to acceleration in direction j [kg, kgm, kgm2 ]

N yaw moment around intersection of midship/centreplane [N m]

n rate of revolution [1/s]

n surface normal vector [m]

nη number of grid nodes in the normal direction [−]

nξ number of grid nodes in the stream-wise direction [−]

nζ number of grid nodes in the girth-wise direction [−]

NG longitudinal force in centre of gravity [N ]

ns rate of revolution at self propulsion point [1/s]

O origin of forces and moments [−]

p apparent order of convergence [−]

p pressure [N/m2 ]

p roll velocity [rad/s]

p∞ undisturbed far-field pressure [N/m2 ]

P0.7 propeller pitch at 0.7R [m]

q pitch velocity [rad/s]

Qp propeller torque [N m]

R radius [m]
xvi List of Symbols

R resistance [N ]

r yaw velocity [rad/s]

ṙ yaw acceleration [rad/s2 ]

rstc yaw rate during steady turning circle [rad/s]

 
V Lref
Re Reynolds number = ν [−]

S simulation value [various]

sφ standard deviation of realisations of result q [various]

Swa wetted surface area [m2 ]

t thrust deduction fraction [−]

Ta draught moulded at aft perpendicular [m]

Tf draught moulded at fore perpendicular [m]

Tm draught moulded at midship (= (Ta + Tf ) /2) [m]

Tp propeller thrust [N ]

tα/2 Student t-distribution coverage factor [−]

u longitudinal velocity [m/s]

u̇ longitudinal acceleration [m/s2 ]

Uφ discretisation uncertainty for variable φ [various]

UD uncertainty of the experiment [various]

UG uncertainty due to discretisation error [various]

UI uncertainty due to iterative error [various]

Uinput uncertainty due to possible uncertainties in the input parameters [various]

 q 
Uval validation uncertainty = 2 + U2 + U2
UD [various]
SN input

USN numerical uncertainty of the simulation (= UI + UG ) [various]

 √ 
V total velocity = u2 + v 2 + w2 [m/s2 ]

v transverse velocity [m/s]

v̇ transverse acceleration [m/s2 ]

V0 approach speed [m/s]

V∞ undisturbed far-field velocity [m/s]

List of Symbols xvii

Vstc speed during steady turning circle [m/s]

w vertical velocity [m/s]

w wake fraction [−]

X longitudinal force [N ]

x longitudinal position [m]

xB position centre of buoyancy forward of midship [m]

xE earth-fixed longitudinal position [m]

XG longitudinal force in centre of gravity [N ]

xG position centre of gravity forward of midship [m]

xR centre of rotation [m]

Y transverse force [N ]

y transverse position [m]

y+ non-dimensional distance to the wall [−]

y2+ non-dimensional wall distance of first cell away from the wall [−]

yE earth-fixed transverse position [m]

YG longitudinal force in centre of gravity [N ]

Z number of propeller blades [−]

Z vertical force [N ]

z vertical position [m]

resp,max maximum non-dimensional residual of the pressure [-]

xviii List of Symbols

Page intentionally left blank

 xix

Acronyms

AD Advance. 101, 102, 114

AFF Anechoic Flow Facility. 13, 14, 73
ANEP Allied Naval Engineering Publication. 5
APP Aft Perpendicular Plane. 94
ASME American Society Of Mechanical Engineers. 46, 47
AVT Applied Vehicle Technology. 23
CFD Computational Fluid Dynamics. iv, 7–11, 15–17, 33, 53, 75, 77, 79, 87, 89,
103, 116, 119, 125, 144, 161–175
CMT Circular Motion Test. 9, 16
CPMC Computerised Planar Motion Carriage. 9, 95
DARPA Defense Advanced Research Projects Agency. 13, 73, 75, 91, 145
DES Detached Eddy Simulation. 11, 13, 17
DMI Danish Maritime Institute (now FORCE). 22
DMO Defence Materiel Organisation. 145
DOF Degrees Of Freedom. 9, 27, 96
GCI Grid Convergence Index. 47
HSVA Hamburgische Schiffbau-Versuchanstalt. 17, 18, 38, 66, 105–107, 109, 146
HTC Hamburg Test Case. iv, 17, 18, 20, 23, 26, 40, 41, 45, 55, 71, 92–94, 99,
102–106, 109, 114, 117, 119, 123, 124, 126, 144–146
IIHR Iowa Institute of Hydraulic Research. 12
IMO International Maritime Organization. 1, 5, 7, 104, 117, 119, 120
INSEAN Istituto Nazionale Per Studi Ed Esperienze Di Architettura Navale. 16, 22,
146
ISO International Organization for Standardization. 47
IST Instituto Superior Técnico, Portugal. 145
ITA Initial Turning Ability. 114, 117
ITTC International Towing Tank Conference. 21, 85
KCS KRISO Container Ship. 15, 22
KRISO Korean Research Institute of Ships and Ocean Engineering. 14
LED Light-Emitting Diode. 96
LES Large Eddy Simulation. 13
MARIN Maritime Research Institute Netherlands. 7, 9, 11, 22, 23, 26, 27, 38, 93–95,
97, 105, 123, 145, 147
xx Acronyms

MCR Maximum Continuous Rating. 95

MMG Mathematical Manoeuvring model Group. 8, 9, 11
MOERI Maritime & Ocean Engineering Research Institute. 14
MPI Message Passing Interface. 38
NATO North Atlantic Treaty Organization. 1, 5, 23
NMRI National Maritime Research Institute. 51, 146
NNemo Newport News Experimental Model. 10, 20
NSWCCD Naval Surface Warfare Center, Carderock Division. 146
osa Overshoot Angle. 114
PIV Particle Image Velocimetry. 14, 17, 66
PMM Planar Motion Mechanism. iii, 6, 7, 9, 10, 16, 17, 143
PS Port Side. 97, 100–102, 104, 117, 120
QUICK Quadratic Upwind Interpolation for Convective Kinematics. 74
RANS Reynolds-Averaged Navier-Stokes. 10, 11, 13, 14, 17, 38, 47, 55, 73, 81, 109,
126
RMS Root mean square. 46
RNLN Royal Netherlands Navy. 85, 145
RPM Revolutions Per Minute. 22, 93, 95–97, 113
RTO Research and Technology Organisation. 23
SA Spalart-Allmaras. 73, 76–81
SB Starboard. 97, 100–102, 104, 117, 120
SHWG Submarine Hydrodynamics Working Group. 75, 77, 79
SIMPLE Semi-Implicit Method for Pressure Linked Equations. 38
SMB Seakeeping and Manoeuvring Basin. 95
SST Shear Stress Transport. 38, 48, 55, 73, 76–81, 85, 91
STANAG Standardization Agreement. 5
TD Tactical Diameter. 101–103, 114
TNT Turbulent/Non-Turbulent. 38, 73
TUHH Technische Universität Hamburg-Harburg. 38
UMF Uncertainty Magnification Factor. 114, 116
V&V Verification & Validation. 13, 17, 45–47
VIRTUE VIRtual Tank Utility in Europe. 17, 18, 23, 38, 55, 68, 71, 93, 105, 123, 145,
147
VLCC Very Large Crude Carrier. 10, 11, 14–17, 22, 23, 47, 48, 51
 1

Chapter 1

Introduction

An increase in ship sizes can be seen in the last decades. One reason for the increase of
cargo ships is the improved economy of transporting goods with larger ships. For cruise
ships, the increase in size is probably also driven by the need to provide more diverse
amusement for passengers and the competition of cruise operators to provide cruises on
the largest cruise ship in the world. Furthermore, more and more goods are distributed
using ships and therefore an increase of traffic density is observed.
The enlargement of ships and the increased traffic density lead to new challenges in
the design of the ship, one of which is the demand for better manoeuvrability. Since
large ships must operate in existing harbours, they will experience shallow water effects
more severely and the traffic density requires better controllability of the ships. In the
past, the interest in improved manoeuvrability has resulted in requirements posed by the
International Maritime Organization (IMO), while recently also the North Atlantic Treaty
Organization (NATO) started working on manoeuvring criteria for naval ships.
Several methods are available to assess whether a ship’s manoeuvrability complies
with the requirements. When conducted properly, full scale trials will provide information
about ship manoeuvres free of scaling effects or other assumptions. However, due to
weather conditions or current, it may be difficult to obtain accurate trial results and when
the manoeuvrability is deemed insufficient, modifications to the ship will be extremely
expensive. Therefore, assessment of the manoeuvrability is generally made in the design
stage, using model tests or simulations. This approach is much less expensive and provides
more flexibility in the selection of e.g. the steering arrangements, with the drawback that
scale effects will influence the results (model tests) and inaccuracies may be present in
the simulations due to improper selection of the mathematical model or hydrodynamic
coefficients used.
For training purposes of the crew of ships, or feasibility studies regarding entries of
large ships in existing harbours, an increasing demand for full-mission bridge simulations
is observed. Furthermore, the manoeuvrability with new propulsors (e.g. pods), or new
control strategies or operations (joystick control, dynamic tracking, side-by-side opera-
tions) are more and more tried out in simulators before the application in the real ship.
2 Chapter 1 – Introduction

To represent reality as much as possible, the mathematical model of the ship should mimic
the response of the ship to rudder or engine commands as well as possible. This also poses
new requirements on the accuracy of the mathematical models used in the simulation of
ship manoeuvres.
In the last decade, considerable developments have been made in simulation of the
viscous-flow around ships in order to predict the flow in manoeuvring conditions and to
determine the associated forces and moments. With these developments, the possibility
to improve manoeuvring simulations and to partly replace model tests with simulations
emerged.

1.1 Problem definition and objectives

The empirical methods used in manoeuvring simulations to predict the hull forces are only
reliable when the hull under consideration matches the hulls that were included in the
database underlying the empirical formulae, and when the manoeuvring conditions match
those as used when the empirical relations were derived. Application outside the range of
applicability, e.g. for novel hull concepts, requires alternative methods of predicting the
hull forces. For the study which forms the basis of this thesis, the recent developments in
viscous-flow calculations provide an attractive means of improving the accuracy of ship
manoeuvrability predictions.
The objective of the present work is therefore to develop a practical method to simulate
the manoeuvrability of ships in which viscous-flow solvers are utilised to improve the
accuracy of the predicted forces and moments on the hull. In this thesis the feasibility of
the method is demonstrated.
The method presented here will extend the work of other researchers, see chapter 2,
by providing extensive verification and validation of the predicted forces and moments on
the hull and a detailed study of the sensitivity of the manoeuvring characteristics of the
ship to changes in the hydrodynamic coefficients in the simulation model.

1.2 Outline of the thesis

The outline of this thesis is as follows: first, the background of manoeuvring simulation
is presented in chapter 2. An overview of existing mathematical manoeuvring models
is given. Then, based on this overview, the so-called virtual captive tests approach is
selected to predict the forces and moments on a manoeuvring ship. Additionally, test
cases available in literature that can be used to validate manoeuvring predictions are
summarised.
The fast-time simulation program SurSim will be used to predict the ship manoeu-
vres. A description of this program is given in chapter 3. A method to derive the hydro-
dynamic coefficients for the bare hull forces and moments is also proposed. In section 3.1,
the coordinate system and nomenclature used in this thesis are presented.
1.2 – Outline of the thesis 3

The viscous-flow solvers used for the present study are briefly discussed in chapter 4,
together with the grid procedures and boundary conditions used for the current work.
Subsequently, verification and validation of the predicted forces and moments on sev-
eral different ship hulls in manoeuvring motions will be presented in chapter 5 using vali-
dation data available in literature. It is demonstrated that for a wide range of ship types,
accurate predictions of the loads on the hull in manoeuvring motion can be obtained.
Free sailing manoeuvring model tests were performed within the context of this thesis
to provide detailed validation data for the manoeuvring simulations. The tests and an
estimation of the uncertainty in the manoeuvring parameters are discussed in chapter 6.
Coefficients will be derived from these virtual captive tests and will be used in the
simulation program SurSim to model the forces on the hull and subsequently simulate
standard manoeuvres. The modifications of the mathematical model and the results of
the simulations will be presented and discussed in chapter 7.
Finally, conclusions are drawn and recommendations for further work are given in
chapter 8.
For descriptions of the symbols and abbreviations used in this thesis, the reader is
referred to page xiii and page xix, respectively.
4 Chapter 1 – Introduction

Page intentionally left blank

 5

Chapter 2

Background of manoeuvring
simulation

This chapter provides the background of manoeuvring simulations, a review of available

standard mathematical manoeuvring models, and gives information regarding the data
available in literature that can be used to validate predictions of the forces and moments
on a ship in manoeuvring conditions and predictions of ship’s manoeuvrability.

2.1 Introduction
Traditionally, ship manoeuvring studies have focused on assessing compliance with the
manoeuvring standards set by the IMO [69]. However, due to emerging owner and op-
erational requirements, the need has arisen for assessment of manoeuvring capabilities in
operations other than the manoeuvring conditions prescribed by the IMO requirements,
see, for example, Quadvlieg and Van Coevorden [111] or Dand [32]. For naval ships, the
NATO Specialist Team in Naval Ship Manoeuvrability is developing a Standardization
Agreement (STANAG) regarding common manoeuvring capabilities for NATO warships
for specific missions. Örnfelt [103] provides an overview of the NATO efforts towards this
STANAG. Preliminary criteria have been published in NATO Allied Naval Engineering
Publication (ANEP) 70 [96]. In Armaoğlu et al. [8] and Quadvlieg et al. [110], demon-
strations are given of the use of prediction tools to verify compliance with the STANAG
criteria.
The assessment of the manoeuvrability of ships in the design stage can be done ex-
perimentally or numerically, or by combining both. For most engineers, free sailing model
tests are generally the preferred option, since they provide immediate insight into the
manoeuvring characteristics of the ship and no assumptions are made regarding the hy-
drodynamics of the model. However, due to scaling of the model, deviations between
model scale and full scale manoeuvres may occur. Additionally, free sailing model tests
do not give quantitative insight into the forces and moments acting on the hull, which is
required when full mission bridge simulator studies, e.g. for training or feasibility studies,
6 Chapter 2 – Background of manoeuvring simulation

are to be conducted.
Furthermore, sometimes different design variants need to be compared before con-
structing and testing physical models of the ship. In those cases, the use of manoeuvring
simulation programs is preferred. However, to obtain reliable simulations, reliable mod-
els of the forces and moments acting on the ship are required, since these are needed to
calculate the accelerations, velocities and trajectories of the ship during the manoeuvre.
A flow chart of manoeuvring simulations is given in Figure 2.1. It is seen that to obtain
the forces on the ship, either experimental (mostly Planar Motion Mechanism (PMM))
or numerical (mostly empirical) techniques can be used. These forces are fed into a ma-
noeuvring simulation program in the form of coefficients or tables and with the program
the manoeuvring characteristics of the ship can be determined. If needed, modifications
to the design can be made in order to ensure compliance with manoeuvring requirements.
When the results are found to be satisfactory, the mathematical model can be used in a
simulator for e.g. training.

Hull Design

Experiments Calculations Calculations

(PMM) (CFD) (Empirical)

Hull forces

Simulation: Other forces

SurSim (appendages)

ok?
Manoeuvring- Simulator
no
characteristics training
yes

Figure 2.1: Flow chart of manoeuvring simulations

The traditional simulation tools use empirical descriptions of the forces and moments
on the ship’s hull and are generally based on regression analysis of captive manoeuvring
test data for a (preferably wide) range of ships. Due to the lack of resolution of hull
details or application outside the range of the regression database, the prediction of the
manoeuvrability may be unreliable. On the other hand, empirical simulation tools can
provide valuable information regarding the manoeuvring characteristics in a cost-effective
way during the early stages of the design.
Other methods to obtain mathematical models of the forces and moments on the ship
comprise conducting captive model tests for the ship under consideration, or by conducting
a series of free sailing model tests and subsequent system identification. Although these
2.1 – Introduction 7

techniques may yield accurate predictions of the manoeuvring characteristics of the ship,
the construction of the physical model and the use of the experimental facilities can be
costly. Furthermore, the data needs to be analysed and fed into a simulation program,
before the actual manoeuvring characteristics of the design can be assessed.
Therefore, new methods are required to obtain reliable and accurate manoeuvring
simulation models in a cost-effective manner. Such methods should not only be suitable
to predict the yaw checking and turning ability of the ship according to the IMO require-
ments, but also be applicable to operation in confined waterways or harbour manoeuvring
assessment studies, for example.
Recently, viscous-flow calculations provide an attractive means to improve manoeu-
vring simulations. Two approaches are available to use such calculations in the prediction
of the manoeuvring of ships:

virtual captive tests: the forces and moments for a range of forced motions (steady
drift, steady rotation, oscillatory sway or yaw, or combinations thereof) are cal-
culated. From the calculations, hydrodynamic coefficients can be derived which
are subsequently used in the simulation model to predict the forces and moments.
Sometimes, the forces and moments are obtained by interpolation between the data
points. This approach resembles the approach taken when using PMM tests and
the calculations are therefore referred to as virtual captive tests. This procedure is
further discussed in section 2.3.

coupling with body motions: the calculated forces and moments are directly used in
the equations of motions to obtain the accelerations, velocities and position of the
ship. This procedure is further discussed in section 2.4.

In aerodynamics, similar approaches are used. A thorough overview is given by Salas

[120]. In flight dynamics, using the approach of the virtual captive tests is called flying
through the database, while the coupling with the body motions is called flying by the
equations. According to Salas, the first method is the easiest to implement, since it
relies on existing technology. The second method requires new capabilities, but probably
provides more accurate solutions, especially in dynamic conditions.
A concise review of the possible applications of and challenges for Computational Fluid
Dynamics (CFD) in aerodynamics is given by McParlin and Tramel [90]. A series of future
workshops is proposed in which the applicability of CFD to specific flow phenomena is
addressed. Cooperative research efforts are suggested, leading to an understanding of the
capabilities of current CFD techniques and to develop best practice guidelines within the
context of overall aircraft aerodynamics.
Within Maritime Research Institute Netherlands (MARIN), the method of virtual
captive tests is thought to be the most attractive at the moment. The reason for this
is that presently the use of viscous-flow calculations coupled with body motions is too
computationally intensive. For daily practice the turn-around time of manoeuvring sim-
ulations should be in the order of a few days to a week, and in most cases variations of
8 Chapter 2 – Background of manoeuvring simulation

the design have to be considered within this time frame. Furthermore, incorporation of
this method in real-time simulators is not yet feasible1 , which means that for simulator
studies the method of using virtual captive tests is the only viable solution.
When using existing modular mathematical manoeuvring models, see e.g. section 2.2,
it is possible to use the virtual captive tests to improve low-fidelity sub-models and provide
manoeuvring advice within a reasonable time frame. For example, when it is expected
that the modelling of the propeller or rudder performance in the mathematical model is
sufficiently accurate, virtual captive tests need only be conducted for the bare hull. This
greatly simplifies the grid generation process and reduces the number of grid points (and
thereby computation time) required to arrive at an accurate estimation of the loads on
the ship.

2.2 Standard mathematical models

Several standard mathematical models for ship manoeuvrability have been proposed in
the past. The models can be divided into integrated models and modular models. Some-
times, the term tabular model is used. In tabular models, expressions for integrated
forces/moments or modular components as a function of a given parameter are replaced
by look-up tables, see e.g. Eloot and Vantorre [48] and Eloot [47].
The integrated models use single polynomial expressions for each force or moment and
are mainly based on series expansions of the forces and moments around an equilibrium
condition, see Abkowitz [2]. The model proposed by Norrbin [97] is also an integrated
model, but attempts are made to relate the hydrodynamic coefficients in the model to
physical phenomena. Tabular or integrated models are very useful for application to a
specific ship, but it is difficult to compare the coefficients with those of other ships.
Modular models describe each component of the ship separately: empirical formu-
lations are posed for e.g. the bare hull, propellers and rudders. Most of these models
are based on the so-called Mathematical Manoeuvring model Group (MMG) model [98].
The rationale behind the modular models is that this approach will provide the easiest
means to incorporate physical background or more complex methods into the modelling
of the forces on the ship. For example, while the bare hull forces are generally described
using empirical formulae, the rudder forces might be approximated using more advanced
predictions such as lifting line or lifting surface methods, without the need to change
the modelling of other components in the simulation model2 . Another advantage is that
this approach enables a somewhat easier comparison of the coefficients across different
proposed mathematical models.
The first MMG model was proposed by Ogawa and Kasai [98]. Subsequent improve-
ments have been proposed by e.g. Inoue et al. [70] to incorporate changes in loading
condition, Lee and Fujino [86] to adapt the model for twin-screw/twin-rudder ships, or
1
According to Salas [120], real-time CFD simulation will be possible around 2027.
2
Assuming that errors in one sub-model are not compensated by errors in other sub-models.
2.3 – Virtual captive tests 9

Kang and Hasegawa [76] to extend the model to low-speed manoeuvring. For the SIM-
MAN 2008 workshop [130], several contributions were made in which the MMG model
was used to model the manoeuvring of the test case ships.
Other models differing from the MMG model are proposed by e.g. Oltmann [101] and
Hooft and Quadvlieg [66]. Both models utilise cross flow drag coefficients (see e.g. Hooft
[63]) to model non-linear effects in the forces and moments on the ship. Variations of
the model of Hooft and Quadvlieg form the basis of the MARIN in-house manoeuvring
simulation programs SurSim, FreSim (see Hooft and Pieffers [65]) and MPP which have
been used in submissions for the SIMMAN 2008 workshop [149]. SurSim has been used
as a basis for the simulations in this thesis, and is described in detail in section 3.2.
Another attempt to provide a practical but sufficiently accurate, general and physi-
cally sound mathematical model was made by Ankudinov and Jakobsen [6]. Guidelines
for the development of a standard simulation model are given. More information about
manoeuvring models and their applications can be found in Fossen [52] and Eloot [47].

2.3 Virtual captive tests

To obtain the derivatives for a mathematical simulation model, the forces and moments
as a function of the flow around the ship need to be obtained. In the past, this was
done using captive model tests, but with the evolution of viscous-flow solvers, this can
be done numerically as well. In this thesis, the term virtual captive tests will be used
to designate the simulation of captive tests using viscous-flow solvers. Therefore, virtual
captive tests can encompass the simulation of PMM tests, Circular Motion Tests (CMTs)
and Computerised Planar Motion Carriage (CPMC) tests, but also steady drift tests.
Although some work on applying viscous-flow solvers to ships in manoeuvring motion
was published before, several authors published results concerning surface ships in oblique
or rotational motion in the same year, i.e. 1998. Examples are Ohmori et al. [100], Ohmori
[99], Alessandrini and Delhommeau [3] and Cura Hochbaum [27]. At this time, the first
steps towards simulation of captive tests were made, but complete series of calculations
in order to derive coefficients were not yet performed, except by Ohmori. About five
years later, Cura Hochbaum and Vogt [29] and Di Mascio et al. [35] presented work
regarding their progress towards virtual PMM tests. However, manoeuvring simulations
using coefficients derived from the calculations were not yet conducted.
Bellevre et al. [12] study the hydrodynamic derivatives of a submarine. In their
calculations, they simulate steady drift and steady rotational motion and determine the
rudder effectiveness. With coefficients derived from these results, they simulate 6-Degrees
Of Freedom (DOF) manoeuvres and compare the results to experiments and sea trials.
Overall, reasonable agreement is found, but restrictions in the number of grid cells that
could be used hamper the accuracy of the calculations. Furthermore, improvements of
the prediction of the rudder effectiveness were required.
Another example of the application of CFD to actually calculate hydrodynamic coef-
10 Chapter 2 – Background of manoeuvring simulation

ficients and use these to simulate the manoeuvring behaviour of a ship was presented by
Racine and Patterson [112]. For a novel hull form, for which accurate empirical formulae
to describe the forces and moments due to manoeuvres were not available, coefficients
were derived and the stability and trajectory of the vessel were assessed. The hull form
under consideration was the Newport News Experimental Model (NNemo). A sensitivity
study was performed to determine the scope of the required calculations and to reduce
the size of the CFD matrix. Unfortunately, validation of the simulated behaviour could
not be conducted at the time of the study although some of the phenomena found in the
simulations were apparently also found during free running model tests.
A detailed study using virtual PMM simulations is presented by Cura Hochbaum
[28]. Here, simulations of zig-zag and turning circle manoeuvres are conducted for a twin-
screw ferry. The time traces of the PMM simulations are compared to experimental PMM
results. Additionally, the coefficients derived from the numerical study are compared to
the coefficients derived from the experiments. Finally, the manoeuvre results are compared
to the results based on simulations using the experimental hydrodynamic coefficients and
to results obtained using free sailing experiments. The validation is encouraging and
it is demonstrated that the procedure works well, although some improvements in grid
resolution and modelling are proposed.
For the SIMMAN 2008 workshop [130], two participants provided manoeuvring simu-
lations for the KVLCC2 (see section 2.5.3) using coefficients derived from CFD results, i.e.
Cura Hochbaum et al. [30] and Toxopeus and Lee [149]. Cura Hochbaum et al. conducted
CFD calculations for static drift, oscillatory motion and for rudder deflections to arrive
at complete mathematical models for the KVLCC1 and KVLCC2. The agreement of
manoeuvring simulations using the mathematical models with the experiments was very
promising, especially for the KVLCC2. Toxopeus and Lee calculated the hydrodynamic
coefficients using CFD for the bare hull, and used empirical formulae to calculate the
forces due to the propeller and rudder. Comparison of the simulated manoeuvres with
the free sailing experiments showed that the mathematical model needed to be improved,
mainly by extending the range of drift angles and yaw rates used to derive the coefficients.
The present thesis demonstrates a procedure similar to the one used in Toxopeus and
Lee [149], but for a different test case and more attention is paid to the correct modelling
of the forces and moments on the ship.

2.4 RANS coupled to ship motions

When coupling Reynolds-Averaged Navier-Stokes (RANS) calculations to the ship mo-
tions, fewer assumptions about the forces and moments on the ship and its appendages are
made, especially when the calculations are performed for the full scale Reynolds number.
Although this approach is computationally expensive, progress has been demonstrated in
literature. An overview of relevant studies is presented in this section.
Sato et al. [121] conducted a study in which their viscous-flow solver is coupled to
2.4 – RANS coupled to ship motions 11

the equations motions of the ship. The instantaneous forces on the hull are calculated
using CFD, while the forces due to the propeller and rudder are calculated using empirical
formulae based on the MMG model. With their model, they perform zig-zag manoeuvres
for two Very Large Crude Carrier (VLCC) variants. Comparison with the experiments
shows reasonable quantitative agreement.
Pankajakshan et al. [106] apply a coupled procedure to simulate overshoot manoeuvres
for the ONR Body-1 submarine model. The control surfaces and rotating propeller are
included in the viscous-flow calculations. The propeller is incorporated in the simulation
with sliding interfaces. The deflection of the planes is modelled using re-generation of the
grid for each deflection angle, based on interpolation between several grids spanning the
range of deflection of the control surface. The agreement between the simulated results
and the experiments is good.
Jensen et al. [74] show a turning circle simulation for a container ship. The rudder
is modelled with sliding interfaces, while the propeller is modelled using body forces.
Unfortunately, validation was not performed.
Venkatesan and Clark [160] present simulations for an overshoot manoeuvre for ONR
Body-1 (similar to the work by Pankajakshan et al. [106]) and compare the results to
experiments. The propeller is modelled using sliding interfaces, while the control surface
deflection is modelled using mesh deformation. The agreement is promising, but appears
to be slightly less than for Pankajakshan et al.
For the SIMMAN 2008 workshop [130] Carrica and Stern [23] performed Detached
Eddy Simulation (DES) of the KVLCC1 performing zig-zag and turning circles with
moving propeller and rudder. Overlapping grid techniques are used to model the moving
appendages and a level set approach is used to capture the free surface. The simulations
were found to be computationally very intensive and could not be finished before the
workshop. The results are promising and a good demonstration of the current capabilites
of CFD, but some issues remain to be solved. Carrica et al. [22] also conducted free
sailing manoeuvring RANS simulations for the 5415M (see section 2.6.4). Also in this
case, overlapping grids are used to model the hull, bilge keels, stabiliser fins, shafts,
struts and moving rudder and a level set method is used for free surface capturing. The
propeller is modelled through the body-force approach, neglecting local velocity effects.
The agreement between the simulations and free sailing experiments performed at MARIN
[150] was very good, leaving rather limited suggestions for improvements.
Another interesting example of coupling the RANS solution to rigid body motions
is given by Bettle et al. [14]. They study the rising stability of a submarine, i.e. the
development of the roll angle when a submarine needs to surface quickly. Calculations
for full scale Reynolds numbers have been performed. The forces due to the propeller,
ballast system and appendages have been incorporated using coefficient based models.
The simulations detected the underwater roll instabilities and were consistent with results
obtained using fully coefficient-based simulations and with observations during full scale
trials. With the CFD results, the main source of the instabilities, i.e. the rolling moment
generated by the sail, was identified.
12 Chapter 2 – Background of manoeuvring simulation

2.5 Validation cases for captive manoeuvring

In literature, several data sets are available that can be used to validate predictions of
the forces and moments on a ship in manoeuvring conditions. Some of these test cases
also comprise flow field measurements, such that more details about the accuracy of the
viscous-flow simulations can be obtained. This section presents some of the available test
cases.

2.5.1 Series 60

Figure 2.2: Body plan of Series 60

Extensive flow field and force measurements on the well-known Series 60 hull form
were conducted at Iowa Institute of Hydraulic Research (IIHR) and the results were made
available to the public by Longo [89]. For a range of drift angles and Froude numbers,
the forces on the model were measured. Furthermore, wave patterns were measured for
a selected set of drift angles and speeds. For a drift angle of 10◦ , the mean flow was
obtained at several longitudinal stations. During the force measurements, the model was
free to sink, trim and heel and the displacements were recorded. During all other tests,
the model attitude was fixed. The water depth to ship’s draught ratio of h/Tm = 18.7
represented deep water conditions. The main particulars and body plan of the Series 60
are presented in Table 2.1 and Figure 2.2.
Because of the large amount of data obtained during the measurement campaign, this
case is very suitable for validation studies of viscous-flow calculations. Various researchers
have already reported such validation studies, such as Alessandrini and Delhommeau [3],
Cura Hochbaum [27], Campana et al. [21], Di Mascio and Campana [36], Tahara et al.
[140], Toxopeus [141, 142] and Di Mascio et al. [35].
The Series 60 test case is not further considered in this thesis.
2.5 – Validation cases for captive manoeuvring 13

Table 2.1: Main particulars of Series 60

Magnitude
Description Symbol proto model Unit
Scale λ 1:1 1:40 -
Length between perpendiculars Lpp 121.920 3.048 m
Breadth max. moulded B 16.250 0.406 m
Draught moulded fore Tf 6.500 0.163 m
Draught moulded aft Ta 6.500 0.163 m
Displacement volume moulded ∇ 7715 0.121 m3
Wetted surface area bare hull Swa 2528 1.580 m2
Position centre of buoyancy forward of midship xB -1.523 %Lpp
Block coefficient Cb 0.600 -
Length-Breadth ratio L/B 7.503 -
Breadth-Draught ratio B/T 2.500 -
Length-Draught ratio L/T 18.757 -

2.5.2 DARPA SUBOFF

For the Defense Advanced Research Projects Agency (DARPA) SUBOFF submarine hull
form [57, 88] extensive validation data for flow field variables and integral quantities are
available. The test program was initially split into two measurement campaigns. One
series was conducted in a wind tunnel (DTRC Anechoic Flow Facility (AFF)), during
which the flow field around the aft hull and the pressure and shear stress distributions
along the hull length were measured, see Huang et al. [68]. These measurements were
conducted at a Reynolds number of Re = 12 × 106 and at incidence angles of 0◦ and
2◦ . The second series was conducted in a towing basin (DTMB), during which the forces
and moments as a function of the flow incidence angle were measured, see Roddy [116].
The towing tank measurements were conducted at a Reynolds number of Re = 14 × 106 .
The main particulars and an impression of the AFF-1 hullform are given in Table 2.2 and
Figure 2.3.
In literature, several studies concerning calculations on the bare-hull DARPA SUB-
OFF (designated Configuration 3 in Roddy [116] and configuration AFF-1 as defined in
Liu and Huang [88]) can be found, see e.g. Sung et al. [137, 136, 135], Bull [19], Jon-
nalagadda et al. [75], Bull and Watson [20] (looking into scale effects), Yang and Löhner
[166] (comparisons with the experiments, also for AFF-2), Toxopeus [147] (Verification
& Validation (V&V), comparison with the experiments), Toxopeus and Vaz [151] (V&V,
comparison with the experiments) and Vaz et al. [158] (V&V, comparison with the ex-
periments, also for AFF-8).
Extensive work on the appended SUBOFF (AFF-8) has been performed by e.g. Alin et
al. [4, 5], using Large Eddy Simulation (LES), and comparisons were made with results of
DES and RANS calculations. In Fureby [53] an overview is given regarding the application
of LES in engineering studies, with application to the SUBOFF, amongst others.
14 Chapter 2 – Background of manoeuvring simulation

Table 2.2: Main particulars of DARPA SUBOFF submarine

Description Symbol Magnitude Unit

Length overall Loa 4.356 m
Length between perpendiculars Lpp 4.261 m
Maximum hull radius Rmax 0.254 m
Centre of buoyancy (aft of nose) FB 0.4621 Loa -
Volume of displacement (AFF-1) ∇ 0.708 m3
Wetted surface (AFF-1) Swa 5.998 m2

Figure 2.3: Geometry of DARPA SUBOFF (AFF-1)

Wu et al. [165] conducted studies on the SUBOFF moving close to the sea floor. They
found that the bottom effects are proportional to Bc , with B the submarine beam and c
the clearance between the bottom and the submarine hull. The paper shows that RANS
solvers can be used as a practical tool to predict hydrodynamic aspects of submarines.
Etebari et al. [49] present results of a test campaign for the SUBOFF model undergoing
a steady turn. Stereo-Particle Image Velocimetry (PIV) was used to obtain the flow field
around the model. Pressure measurements at two axial cross sections on the model were
conducted and the forces and moments on the model were measured. Three configurations
of the SUBOFF were used: the bare hull, the fully appended model and the bare hull
with towed-array fairing.
Several series of unsteady measurements using the SUBOFF have been conducted, see
e.g. Whitfield [164] or Hosder [67]. Further experiments have been conducted by Granlund
and Simpson [56] on the added mass of the SUBOFF (AFF-2). This work has resulted in
the thesis by Granlund [55].

2.5.3 KVLCC1, KVLCC2 and KVLCC2M

The KVLCC1 and KVLCC2 (Korean Research Institute of Ships and Ocean Engineering
(KRISO)3 Very Large Crude Carrier) have been subjects of manoeuvrability studies since
long. These ships have the same main particulars and bow shape, but slightly different
sterns, and should represent typical 300000 t tanker hull forms from around 1997. The
KVLCC1 has a fine stern end bulb and the stern frames have more barge type lines,
while the stern frames of the KVLCC2 are more U-shaped. During the design of the hull
3
Now known as Maritime & Ocean Engineering Research Institute (MOERI)
2.5 – Validation cases for captive manoeuvring 15

Table 2.3: Main particulars of KVLCC1, KVLCC2 and KVLCC2M

Magnitude
Description Symbol KVLCC1 KVLCC2 KVLCC2M Unit
Length between perpendiculars Lpp 320.000 320.000 320.000 m
Breadth max. moulded B 58.000 58.000 58.000 m
Draught moulded fore Tf 20.800 20.800 20.800 m
Draught moulded aft Ta 20.800 20.800 20.800 m
Displacement volume moulded ∇ 312631 312635 312650 m3
Wetted surface area bare hull Swa 27370 27257 27279 m2
Position centre of buoyancy forward of midship xB 3.494 3.497 3.527 %Lpp
Block coefficient Cb 0.810 0.810 0.810 -
Length-Breadth ratio L/B 5.517 5.517 5.517 -
Breadth-Draught ratio B/T 2.788 2.788 2.788 -
Length-Draught ratio L/T 15.385 15.385 15.385 -

Figure 2.4: Body plans of KVLCC1 and KVLCC2 (dotted: KVLCC1, continuous: KVLCC2)

forms, it was anticipated that the manoeuvrability of these ships would be different. The
KVLCC1 and KVLCC2 hull forms were two of the subjects of study during the CFD
Workshop Gothenburg 2000 [83] and the SIMMAN 2008 workshop [130]. The KVLCC2
was studied during the CFD Workshop Gothenburg 2010 as well.
Experiments have been conducted by Kim et al. [77]. For the KVLCC1 and KVLCC2
hull forms, they performed wave elevation and flow field measurements around the 1 : 58
scaled ship models. Furthermore, resistance and propulsion tests were conducted. The
attitude of the model was fixed, except during the resistance and propulsion tests. During
these tests, the trim of the model was modified to arrive at a level running trim during
the actual tests. The water depth to ship’s draught ratio of h/Tm = 13.5 represented
deep water conditions. The publication also contains results for the KRISO Container
Ship (KCS) (see 2.6.3). Similar flow field measurements have been conducted in a wind
16 Chapter 2 – Background of manoeuvring simulation

Figure 2.5: Body plans of KVLCC2 and KVLCC2m

(dotted: KVLCC2m, continuous: KVLCC2)

tunnel, using double-body models. Tests were only conducted for straight-ahead sailing.
The results were published by Lee et al. [85].
Captive model tests for the bare hull KVLCC2 were conducted by Istituto Nazionale
Per Studi Ed Esperienze Di Architettura Navale (INSEAN) in preparation for the SIM-
MAN 2008 Workshop [130], see also Fabbri et al. [50]. A set of PMM tests was performed,
comprising amongst others the measurement of the forces and moments for steady drift
motion and oscillatory yaw motion. During the tests, the model was free to heave and
pitch. The model tests were conducted for deep water (water depth to ship’s draught
ratio h/Tm = 8.3) as well as for restricted water depths using a false bottom: h/Tm = 1.2,
1.5 and 3.0. The scale factor was 1 : 45.71.
Additional measurements for the KCS, KVLCC1 and KVLCC2 have been done by
Ueno et al. [153]. Using CMT, they obtained the forces and moments on the hull with
rudder and propeller as a function of the drift angle and rotation rate. The models were
free to trim, sink and heel. For the KVLCCs, the water depth to draught ratio was
h/Tm = 10.6, representing a deep water condition. The scale factor was 1 : 110.
The KVLCC2M hull form was conceived as a variant of the KVLCC2 but with a
slightly modified aft ship to reduce the complexity of the flow and therefore simplify
viscous-flow computations. The most visible difference is the fairing of the hull around
the propeller shaft position. The KVLCC2M was the main subject of the manoeuvring
studies in the CFD Workshop Tokyo 2005 [58]. Model tests for the KVLCC2M have
been extensively described by Kume et al. [82]. These tests comprised measurements of
forces, surface pressures and stern flow fields for the 1 : 64.4 scaled model at several drift
angles. During the tests, the movement of the model with respect to the carriage was
fully constrained. The water depth to ship’s draught ratio of h/Tm = 24.8 represented
deep water conditions.
Full scale ships of the KVLCCs do not exist. The main particulars of these ships and
2.5 – Validation cases for captive manoeuvring 17

body plans are given in Table 2.3, Figure 2.4 and Figure 2.5 respectively.
Several authors have studied the manoeuvrability of these VLCCs, such as the partic-
ipants of the Tokyo CFD Workshop [58], Toxopeus [143] (KVLCC2M, V&V), Simonsen
and Stern [126, 127] (KVLCC2, deep/shallow water, V&V), Broglia et al. [17] (KVLCC2,
blockage during PMM), Carrica and Stern [23] (KVLCC1, DES coupled with body mo-
tions), Cura Hochbaum et al. [30] (KVLCC1 / KVLCC2, virtual PMM), Toxopeus and
Lee [149] (KVLCC2, virtual PMM), Muscari et al. [95] (KVLCC2, RANS coupled with
body motions), Stern et al. [132] (KVLCC2, DES at large drift angles) and Phillips et al.
[107] (KVLCC2, virtual PMM). More information about the KVLCC results obtained at
SIMMAN 2008 can be found in Stern et al. [131].

2.5.4 HTC

Figure 2.6: Teresa del Mar (source: www.shipphotos.es/teresadelmar.htm)

The Hamburg Test Case (HTC) is a model of the container ship built by Bremer
Vulkan in 1986 as Ville de Mercure, and subsequently named Teresa del Mar, see Fig-
ure 2.6. Teresa del Mar was sold in 2010 and renamed Maria. After the Ville de Mercure
a number of other container ships with the same hull form were built. One of the sister
vessels, Catalina del Mar, is still sailing.
Captive model experiments were conducted on the HTC within the VIRtual Tank
Utility in Europe (VIRTUE) project by Hamburgische Schiffbau-Versuchanstalt (HSVA)
in order to provide additional material for CFD validation. These tests comprised force
measurements for the bare hull, the hull with rudder and the hull with propeller and
rudder. Furthermore, PIV measurements were conducted for the model equipped without
18 Chapter 2 – Background of manoeuvring simulation

rudder, with the model sailing at steady rotational motion. These experiments were
reported in VIRTUE deliverable D3.1.3, see Vogt et al. [161]. The scale of the model λ
was 1:24 during the HSVA tests. The water depth to ship’s draught ratio of h/Tm = 14
represented deep water conditions.
The HTC has been studied numerically by several authors, such as Drouet et al. [37],
Gao and Vassalos [54] and Toxopeus [144, 145, 148].

Hull form
In Table 2.4, the main particulars of the HTC are presented. Both model scale and full
scale (prototype) values are given.

Table 2.4: Main particulars of HTC

HTC
Description Symbol proto model Unit
Scale λ 1:1 1:24 1:30.02 -
Length between perpendiculars Lpp 153.700 6.404 5.120 m
Breadth max. moulded B 27.500 1.1458 0.916 m
Draught moulded fore Tf 10.300 0.4292 0.343 m
Draught moulded aft Ta 10.300 0.4292 0.343 m
Displacement volume moulded ∇ 28342 2.0500 1.048 m3
Wetted surface area bare hull Swa 5567 9.6640 6.177 m2
Position centre of buoyancy forward of midship xB -0.571 %Lpp
Block coefficient Cb 0.650 -
Length-Breadth ratio L/B 5.582 -
Breadth-Draught ratio B/T 2.673 -
Length-Draught ratio L/T 14.922 -

The body plan of the HTC is presented in Figure 2.7. A photograph of the HTC
during the model tests at HSVA is presented in Figure 2.8.
The model was tested fixed in all degrees of motion. It was not equipped with bilge
keels. For the measurements of the model with rudder, the rudder forces were measured
separately. Turbulence was stimulated in all tests. Therefore, in the viscous-flow calcula-
tions it was assumed that the flow was fully turbulent.
2.5 – Validation cases for captive manoeuvring 19

Figure 2.7: Body plan of HTC

Figure 2.8: HTC model during oblique motion test using CPMC
(photograph by HSVA)
20 Chapter 2 – Background of manoeuvring simulation

Propeller

The particulars of the propellers of the prototype HTC and of the model test propellers
(scaled to prototype values) are specified in Table 2.5.

Table 2.5: Main particulars of the HTC propeller

DESIGNATION SYMBOL MAGNITUDE UNIT

Prototype Model Scale
Propeller Id. - HSVA 2208 MARIN 5286R
Diameter Dp 6.105 6.101 6.100 m
Pitch at 0.7R P0.7 4.884 4.994 4.642 m
Pitch ratio at 0.7R P0.7 /Dp 0.800 0.818 0.761 -
Expanded blade area ratio AE /A0 0.569 0.580 0.568 -
Number of blades Z 4 4 4 -
Direction of rotation - clockwise when looking ahead -

Rudder

For the model tests, the rudder was divided into a movable and a fixed (headbox) part in
order to allow turning of the rudder without touching the hull surface. The particulars of
the rudder and a drawing are presented in Table 2.6 and Figure 2.9.

Table 2.6: Particulars of the HTC rudder

Description Symbol Magnitude Unit

Projected total rudder area AR,tot 39.76 m2
Projected movable rudder area AR 29.03 m2
Rudder span hR,tot 9.20 m
Rudder span of movable part hR 7.20 m
Average rudder chord cR 4.32 m
Total area ratio AR,tot /Lpp T 2.51 %
Movable area ratio AR /Lpp T 1.83 %

2.5.5 Other cases

Other publications presenting data that can be used to validate the forces and moments
predicted with viscous-flow calculations for manoeuvring purposes are e.g. the prolate
spheroid [24], the NACA 0012 profile (2D case) [1], ONR Body 1 [138, 10], DTMB 5415
[130, 119, 93, 167, 118] and recently the NNemo [112, 33, 117, 55, 108].
2.6 – Validation cases for free sailing manoeuvring 21

Figure 2.9: Drawing of the HTC rudder, T =10.3 m, dimensions in mm

2.6 Validation cases for free sailing manoeuvring

For validation of the predicted manoeuvres, results from free sailing manoeuvring tests or
full scale trials are required. This section presents some of the test cases that are available
in literature for which free sailing manoeuvring data is present.

2.6.1 Esso Osaka

The Esso Osaka received ample attention due to the existence of well-documented trials
in deep and shallow water published by Crane [26], and it was recommended by the 22nd
International Towing Tank Conference (ITTC) that this ship was used for validation of
force predictions and manoeuvring simulations. However, due to the fact that the hull
form is rather outdated and some doubts arose regarding the scatter in the results during
analysis of different model test campaigns [25], the interest in this validation case has
diminished in the last few years.
Recent studies in which the viscous-flow around the Esso Osaka for captive conditions
was simulated were published by e.g. El Moctar [46], Simonsen [122], Simonsen and Stern
[123, 124, 125] and Van Oers and Toxopeus [155].
22 Chapter 2 – Background of manoeuvring simulation

2.6.2 KVLCC1, KVLCC2

Information about the hull forms and the captive tests conducted for these KVLCCs is
given in section 2.5.3. Free sailing manoeuvring tests, comprising zig-zag and combined
turning-circle/pull-out manoeuvres, were conducted at MARIN to supply validation data
to the SIMMAN 2008 workshop. The models were manufactured by INSEAN and kindly
provided to MARIN for these tests. The tests were reported by Lee [87]. All tests were
conducted at model self propulsion point, with constant propeller Revolutions Per Minute
(RPM). Several tests were repeated in order to obtain an estimate of the uncertainty in
the results. With a model scale of 1 : 45.714, the water depth to ship’s draught ratio of
h/Tm = 11 represented deep water conditions.

2.6.3 KCS
During the SIMMAN 2008 workshop, it was concluded that the existing free sailing ma-
noeuvres with the KCS were not sufficiently accurate to be used as reliable validation
data. Therefore, MARIN decided to conduct a new set of free sailing manoeuvres with
the KCS. The tests comprised zig-zag and combined turning-circle/pull-out manoeuvres,
with variations in GM values and for fully loaded and ballast condition, see Overpelt [104].
With a model scale of 1 : 39.89, the water depth to ship’s draught ratio of h/Tm = 17.5
represented deep water conditions.

2.6.4 5415M
In order to develop new hull form concepts for advanced naval mono-hull ships, a co-
operative research programme, called ”Thales”, was initiated between the Royal Nether-
lands Navy, the Italian Navy and the Danish Navy. The programme comprised develop-
ment of design requirements and procedures, determination of assessment procedures and
selection of design tools. To be able to select the best design tools, benchmark data were
necessary with which design tool data could be compared. Therefore, a model test pro-
gramme was conducted to generate this data. Model tests were conducted by INSEAN
(Italy), DMI (Denmark) and MARIN (The Netherlands). Additional information was
obtained from DTMB (USA).
For the model tests, the representative high speed displacement hull form of the US
Navy DDG 51 destroyer (in literature often designated DTMB 5415) was selected. This
hull form was fitted with a representative twin-propeller / twin-rudder arrangement and
centre-line skeg design comparable to European design practice. Furthermore, stabiliser
fins were fitted to the model. To distinguish this modified geometry and appendage
arrangement to the DTMB 5415, it is designated 5415M. Standard zig-zag experiments,
spiral tests and combined turning circle / pull-out tests were conducted. During these
tests, forces on the appendages were measured in order to validate manoeuvring simulation
programs in more detail. Several repeat runs for some of the manoeuvres were conducted
to assess the uncertainty in the manoeuvring parameters. The tests have been reported
2.7 – Conclusion 23

by Toxopeus and Lee [150]. With a model scale of 1 : 35.48, the water depth to ship’s
draught ratio of h/Tm = 29 represented deep water conditions.
The 5415M was selected as one of the test cases of the SIMMAN 2008 workshop. The
free sailing manoeuvre data are used to validate manoeuvring predictions in Carrica et
al. [22], Bhushan et al. [16], and in the SIMMAN 2008 proceedings [130]. Furthermore,
the roll decay and seakeeping tests performed with this free sailing model are used as
validation data in the NATO Research and Technology Organisation (RTO) Applied
Vehicle Technology (AVT)-161 working group.

2.7 Conclusion
Several different methods are given in literature to simulate ship manoeuvres using viscous
flow calculations. In this thesis, the virtual captive test approach is adopted, since this
is at present the most attractive approach when considering the computational costs.
Furthermore, this approach can directly be used to improve mathematical models for
manoeuvring simulators.
A survey of validation data existing in literature was made. For free sailing manoeu-
vres as well as for captive tests, data can be found that can be used to validate numerical
predictions. Noteworthy is the fact that validation data for captive steady drift motions
are much more abundant than validation data for steady rotation or combined motion.
Especially for more extreme conditions with large turning rates and drift angles (e.g. non-
dimensional turning rates γ of 0.8 - 1.0 and drift angles between 20◦ and 30◦ ), such as
occur during tight turning circles, not much validation data can be found in literature.
Furthermore, to the knowledge of the author, cases in which extensive captive test data
and free sailing manoeuvring test data are available have only been published for the Esso
Osaka and the KVLCCs (although the captive tests do not cover the complete range of
rotation rates experienced during turning circle manoeuvres). In the work leading up
to this thesis, much work was done on simulating the flow around the HTC for captive
conditions and simulations of standard free sailing manoeuvres were conducted within
the VIRTUE project. To generate validation data for these manoeuvres, MARIN decided
to perform free sailing manoeuvring tests for the HTC. Details about these tests will be
presented in chapter 6.
The tools used during this study and the procedure to predict the manoeuvrability of
the ship are discussed in the following chapters.
24 Chapter 2 – Background of manoeuvring simulation

Page intentionally left blank

 25

Chapter 3

Mathematical model

The in-house manoeuvring simulation program SurSim will be used to simulate ship
manoeuvres within the present study. This chapter presents the system of coordinates
adopted in this thesis, details of SurSim and proposes the required steps to derive the
hydrodynamic coefficients for the bare hull forces and moments.

3.1 Coordinate system and nomenclature

PS
y,r
x,u
d
b
SB

y,v V
SB PS

x,u
y,v
q,q f,p

z,w z,w

Figure 3.1: Ship-fixed coordinate system

In ship manoeuvring, two reference frames are used. One frame is attached to the
ship and is called the ship-fixed coordinate system. The motion of the ship-fixed refer-
ence frame is described relative to an earth-fixed inertial reference frame. In the ship-fixed
coordinate system, x is directed forward, y to starboard and z vertically down, see Fig-
ure 3.1. For captive tests or simulation of captive motions, it is customary to use the
intersection of the waterplane, midship and centre-plane as origin O. Therefore, all forces
and moments are given with respect to O, with the longitudinal force X directed forward
positive and the transverse force Y positive when directed to starboard. A positive drift
26 Chapter 3 – Mathematical model

angle β corresponds to the flow coming from starboard. A positive non-dimensional yaw
rate γ corresponds to a turn to starboard when sailing at positive forward speed.
For free sailing manoeuvring tests or simulations, it is however customary to use the
centre of gravity G as the origin of the coordinate system. Therefore, the results of all
manoeuvres are presented with respect to G. In the results, the drift angle βG is given
relative to the centre of gravity is well.
In the earth-fixed coordinate system, xE is directed North, while yE is directed to the
East.
All forces and moments are presented non-dimensionally. The longitudinal force X
and transverse force Y are made non-dimensional1 using 21 ρV 2 Lpp T , the vertical force
Z using 21 ρV 2 Lpp B, the heeling moment K by 12 ρV 2 Lpp T 2 , the pitch moment M by
1
2
ρV 2 L2pp B and the yaw moment N by 21 ρV 2 L2pp T . This method of non-dimensionalisation
has been applied to all force components presented in this thesis. V is the speed of the
ship through the water.
Several subscripts are used to identify separate force components. In this thesis, a
subscript H indicates forces on the hull, R forces on the rudder, P forces on the propeller
and T total forces. Furthermore, f indicates a force contribution due to friction and p a
force contribution due to dynamic pressure (the hydrostatic force component is neglected,
since this force is cancelled by the displacement mass when free surface deformation is
neglected).
For a complete list of symbols used in this thesis, please refer to page xiii.

3.2 Manoeuvring simulation program SurSim

3.2.1 Introduction
The MARIN in-house developed manoeuvring simulation program SurSim is dedicated
to the simulation of the manoeuvrability of mainly twin-screw ferries, cruise ships and
motor yachts. Information about SurSim can be found in [149], [66] and [64]. An example
of a detailed validation study of SurSim can be found in the SIMMAN proceedings [130].
SurSim is able to predict the heel motion of ships during manoeuvres. Due to the
relatively high GM value used during the model tests for the HTC, large heel angles during
manoeuvres are not expected. To simplify the simulations, it was therefore decided to
ignore heel motion for the present study.
SurSim is of the modular type, meaning that the force contributions of the different
components of the ship and their interactions (e.g. hull forces, propeller forces and rudder
forces) are modelled separately. E.g.:

YT = YH + YP + YR (3.1)
1
Unless otherwise specified: for submarines forces and moments are generally made non-dimensional
using L2pp instead of Lpp T as reference area.
3.2 – Manoeuvring simulation program SurSim 27

in which the subscript T denotes the total forces, H the bare-hull forces, P the propeller
forces including hull-propeller interaction and R the rudder forces including hull-propeller-
rudder interaction. In this thesis, the results obtained with the bare-hull forces estimated
using the original empirical formulae are designated ”SurSim”. The results with bare-hull
coefficients obtained from the viscous-flow calculations will be designated ”CFD”.
The following sections describe the equations of motions and the different force com-
ponents. It must be noted that for each force component, user-defined models can also
be used.

3.2.2 Equations of motion

Assuming constant added mass coefficients, the equations of motions in three DOF for
surge, sway and yaw are (see e.g. Hooft and Nienhuis [64] or the complete derivation in
Fossen [52]):
(m + muu ) · u̇ = m · r · v + XG

(m + mvv ) · v̇ + mvr · ṙ = −m · r · u + YG (3.2)

mrv · v̇ + (Iz + mrr ) · ṙ = NG

in which muu = −Xu̇ , mvv = −Yv̇ , mvr = −Yṙ , mrv = −Nv̇ and mrr = −Nṙ are so-called
added mass coefficients. The longitudinal force XG , transverse force YG and yaw moment
NG are the excitation and damping forces on the ship, acting in the centre of gravity G.

The yaw moment in the centre of gravity G is obtained by:

NG = N − xG · Y (3.3)

3.2.3 Hull forces

The hull forces are modelled by using the so-called slender-body method to determine
the linear manoeuvring derivatives and the cross-flow drag method to determine the non-
linear parts of the forces and moments. The linear contributions in these models can be
considered to be representative of the lift generated on the hull while the non-linear ones
represent the drag. The following paragraphs describe the various components of the hull
forces.

Ship resistance

The straight ahead sailing resistance curve of the hull must be given as input to the
program. For this, the MARIN in-house program DESP can be used, which is based on
an improved version of the method of Holtrop and Mennen [62]. The resistance should
be predicted for the appropriate Reynolds numbers, in order to be able to compare the
simulation results directly with full scale trials or with model experiments.
28 Chapter 3 – Mathematical model

Slender body method

The slender-body method is a semi-empirical method to determine the linear force com-
ponents on the hull. According to this method, the linear manoeuvring coefficients are
determined by the rate of change of fluid momentum along the length of the ship. Only
a few empirical parameters based on careful validation with experiments are used. Since
this method utilises the full description of the hull form, the influence of changes in lo-
cal details of the hull can be investigated in the early design stage within a short time
frame (i.e. a few minutes when the hull form input has been prepared). A more detailed
description of the slender-body method can be found in e.g. Toxopeus [145].

Cross-flow drag method

The non-linear contributions to the forces and moments are calculated using e.g. the
so-called cross-flow drag theory, presented by e.g. Hooft and Quadvlieg [66] or Hooft [63].

User-defined hull forces

In SurSim, it is possible to override the hull forces by using a user-defined mathematical

model. This option has been used to introduce the hydrodynamic coefficients obtained
from the viscous-flow calculations into the simulations.
When setting up a mathematical model to describe the forces on a ship due to ma-
noeuvring motion, the intended use of the model determines the structure of the model
itself. E.g. when simulator studies incorporating harbour manoeuvres are to be conducted,
the model should be able to accurately describe the forces and moments on the ship dur-
ing transverse motions, turning-on-the-spot and sailing astern. In the present work, it is
assumed that the manoeuvring model should be valid for a wide range of applications,
including low speed and harbour manoeuvres.
In general, mathematical manoeuvring models for the bare hull consist of three differ-
ent components: (added) mass coefficients, damping coefficients and spring coefficients.
In earlier work by Vassalos et al. [156], Ishiguro et al. [71] and Lee and Shin [84], for
example, or more recently, Bulian et al. [18] it was found that the sensitivity of the ma-
noeuvrability to changes in the added mass coefficients is small. Therefore, it is assumed
that the added mass coefficients can be approximated reliably by using the empiric formu-
lae existing in SurSim. Due to this assumption, no calculations are required to obtain the
added mass coefficients. When only horizontal manoeuvres are considered and neglecting
the heel angle, spring coefficients do not have to be taken into account.
The following non-dimensionalised mathematical model for the longitudinal force X,
transverse force Y and yawing moment N is adopted:

X 0 = Xu0 0 |u0 | · cos β · |cos β| + Xβγ

0
· sin β · γ (3.4)

Y 0 = Yβ0 · |cos β| · sin β + Yγ0 · cos β · γ + Yβ|β|

0 0
· sin β · |sin β| + Yγ|γ| · γ · |γ|
3.2 – Manoeuvring simulation program SurSim 29

0 0
+ Yβ|γ| · sin β |γ| + Y|β|γ · |sin β| γ + Yab0 · cosay β · sinby β · sign sin β (3.5)

N 0 = Nβ0 · cos β · sin β + Nγ0 · |cos β| · γ + Nu0 0 γc · |cos β · γ cn | · signγ

0 0
+ Nγ|γ| · γ · |γ| + Nβ|β| · sin β · |sin β|
0 0
+ Nββγ · sin2 β · γ + Nβγγ · sin β · γ 2 · sign cos β
0
+ Nab · cosan β · sinbn β · sign (cos β · sin β) (3.6)

The coefficients Yab0 and Nab 0

are used to describe the relation between the transverse
force and yawing moment for drift angles β in the range of approximately 30◦ < β < 60◦ .
Similarly, Nu0 0 γc describes the relation between the yaw moment and intermediate yaw
rates γ. The integer values ay , by , an , bn and cn should be adapted to match the correct
order of the relationship. It should be noted that the orders should not be chosen too
high (i.e. 3 or below), to avoid unexpected behaviour at large drift angles.
At zero speed, the non-dimensional yaw rate γ and subsequently the non-dimensio-
0
nal Nγ|γ| contribution will become infinite and therefore due care has to be taken when
implementing this mathematical model in a simulation program. This can be solved by
0
using the Nγ|γ| term in a fully dimensional form.

3.2.4 Propeller forces

The propeller forces are estimated using information from Strom-Tejsen [134], or on the
Wageningen B-Series descriptions, see Oosterveld and Van Oossanen [102]. Alternatively,
it is possible to select a user-defined model for the propeller forces. In this case, the thrust
Tp and torque Qp of the propeller with diameter Dp and rotation rate n are obtained as
follows (subscript p indicates that a variable applies to the propeller):

up = u · (1 − w) (3.7)
up
J= (3.8)
n · Dp
KT = KT 0 + KT 1 · J + KT 2 · J 2 + KT 3 · J 3 + K T 4 · J 4 + KT 5 · J 5 (3.9)
KQ = KQ0 + KQ1 · J + KQ2 · J 2 + KQ3 · J 3 + KQ4 · J 4 + KQ5 · J 5 (3.10)
Tp = KT · ρn2 Dp4 (3.11)
Qp = KQ · ρn2 Dp5 (3.12)

The coefficients KT i and KQi in equations (3.9) and (3.10) are to be specified by the
user and can be obtained from propeller open-water tests. It should be noted that this
user-defined model is not valid for four-quadrant manoeuvre simulations. Based on the
formulae above, the longitudinal force on the ship due to the propeller is given by:

XP = (1 − t) · Tp (3.13)
30 Chapter 3 – Mathematical model

The transverse force YP which is generated by a propeller rotating in an asymmetrical

wake is relatively small and difficult to describe by simple empirical equations. Therefore
this force and the related yawing moment are set to zero: YP = 0 and NP = 0.

3.2.5 Rudder forces

To model the rudder forces, formulae based on publications by e.g. Whicker and Fehlner
[163] (rudders in open water), Inoue [70] (rudders behind ships) and Söding [128] (more
theoretically based) are used. In SurSim, the nomenclature and sign convention as pre-
sented in Figure 3.2 is used. An explanation of the used symbols is given in Table 3.1. A
subscript r indicates that a variable applies to the rudder. In this section, the calculation
of the rudder forces is presented. However, for simplicity, some details such as heel effects
and the angle between the rudder and the vertical axis have not been taken into account
in the formulae in this thesis, although they are modelled in SurSim.
d

dh
de=d-dh

Hx

vr Hy

Lru

ur
Vrr
Dru

Figure 3.2: Nomenclature and sign convention for rudder forces

One of the most complicated aspects in determining the rudder forces is the determi-
nation of the flow velocity and direction at the rudder location as a consequence of the
drift angle, yaw rate and propeller action. Therefore, the formulae describing the flow
velocity are discussed first. To calculate the longitudinal velocity as a function of the
propeller loading, some basic formulae are required. The full analysis can be found in
e.g. Kuiper [81], but the basic principles are stated here to demonstrate the physics and
to modify the formulae for four-quadrant manoeuvres. A cross section of the flow to be
described is given in Figure 3.3.
3.2 – Manoeuvring simulation program SurSim 31

Table 3.1: Symbols used for rudder forces

Symbol Description Unit

δ (Mechanical) rudder angle [rad]
δh Hydrodynamic inflow angle at rudder stock [rad]
δe Effective inflow angle with respect to the rudder = δ − δh [rad]
xr Longitudinal position of rudder [m]
ur Longitudinal inflow velocity at rudder stock [m/s]
vr Transverse inflow velocity at rudder stock [m/s]
p
Vrr Rudder inflow velocity = u2r + vr2 [m/s]
Lru Rudder lift force [N ]
Dru Rudder drag force [N ]
Hx Ship-fixed longitudinal component of rudder force [N ]
Hy Ship-fixed transverse component of rudder force [N ]
AR Rudder area [m2 ]
Λ Rudder aspect ratio [−]

propeller
plane

up+uar up+ua up
p¥ p+Dp p p¥

p 2 p 2 p 2
S1= /4 D1 S= /4 Dp S0= /4 D0

strea
mline

Figure 3.3: The axial actuator disk model

Ahead speed with positive thrust Using axial momentum theory while assuming
inviscid and incompressible flow, the first relation is derived using the conservation of
mass:
π π π
up · D02 = (up + ua ) · Dp2 = (up + uar ) · D12 (3.14)
4 4 4
Conservation of momentum leads to:
π π
ρu2p · D02 − ρ (up + uar )2 · D12 + Tp = 0 (3.15)
4 4
Combining these two relations leads to the following equation for the propeller thrust:
π
Tp = Dp2 · ρ (up + ua ) · uar (3.16)
4
Applying Bernoulli’s law respectively in front and aft of the propeller disc, two addi-
tional equations are found:
1 1
p∞ + ρu2p = p + ρ (up + ua )2 (3.17)
2 2
32 Chapter 3 – Mathematical model

1 1
p∞ + ρ (up + uar )2 = p + ∆p + ρ (up + ua )2 (3.18)
2 2
Subtracting equations (3.18) from (3.17) and using Tp = ∆p· π4 Dp2 , the propeller thrust
is found to be:
π 2 1
 
Tp = Dp · ρ up + uar · uar (3.19)
4 2
such that uar = 2 · ua , following from equation (3.16).
Solving equation (3.19) for uar results in:
v
π π 8Tp
u
Tp − Dp2 · ρup · uar − Dp2 · ρu2ar = 0 ⇒ uar = −up ± tu2p +
u
(3.20)
4 8 ρπDp2

For positive inflow velocity up at the propeller and positive thrust of the propeller,
the axial induced velocity uar is larger than zero. Therefore, equation (3.20) reduces to:
v
8Tp
u
u
uar = tu2 + − up (3.21)
p
ρπDp2

Ahead speed with negative thrust When the forward speed up is positive but Tp
is negative and assuming that ur = up + uar > 0 (we still do not have flow reversal),
equation (3.21) is valid with the additional requirement that uar > −up , resulting in:
8Tp
> −u2p (3.22)
ρπDp2

If the thrust becomes too negative, we get flow reversal due to the propeller action.
In this case, it is assumed that the flow velocity ur at the rudder becomes zero. This
can be obtained by modifying equation (3.21), such that it reads for all conditions with
positive ahead speed up ≥ 0:
v !
8Tp
u
u
uar = tmax u2p + , 0 − up (3.23)
ρπDp2

Astern speed with negative thrust For astern speed and negative thrust, it is as-
sumed that there is no induced velocity from the propeller at the rudder location. In that
case, the change in axial velocity at the rudder becomes:

uar = 0 (3.24)

Astern speed with positive thrust For astern speed but with positive thrust, it is
assumed that the inflow velocity at the propeller position up equals zero. This results in
an axial induced velocity behind the propeller which depends on the propeller thrust and
the local undisturbed velocity u0 . The induced velocity is approximated using:

u0 = u · (1 − wp ) (3.25)
3.2 – Manoeuvring simulation program SurSim 33

v 
8Tp
u
u
uar = max t
 + u0 , 0 (3.26)
ρπDp2
and the velocity at the rudder position is calculated with:

ur = u0 + uar (3.27)

Effective interaction between propeller slipstream and rudder In general, the

propeller slipstream at the rudder location will not completely cover the rudder. Further-
more, the slipstream contraction as used in the actuator disc theory presented above will
not have fully developed yet at the rudder location. Therefore, corrections to the induced
velocity at the rudder need to be made to arrive at the effective induced velocity. This
results in the following empirical prediction of the velocity at the rudder location:

ur = up + Crue · uar (3.28)

Flow straightening Due to the drift angle of the ship and the rotation rate during
manoeuvres, the rudder will also experience a transverse velocity component vr . However,
due to the presence of the ship, the undisturbed flow will be rectified or straightened.
Therefore, the flow-straightening coefficients Cdb for the drift angle and Cdr for the yaw
rate are introduced:
vr = Cdb · v + Cdr · xr · r (3.29)

It must be noted that with this linear equation the effect of the hull on the flow is
strongly simplified (see e.g. Ogawa and Kasai [98]). It can be argued that, for a ship
in manoeuvring motion, vortices that are generated upstream may travel to the rudder
position, resulting in irregular inflow velocities and flow directions. This effect can be
studied using dedicated model tests or with CFD calculations, but is considered to be
outside the scope of the present study.

Rudder forces First the lift and drag coefficients are determined, see eq. (10) in [98]:

6.13 · Λ
CL = (3.30)
2.25 + Λ

CL2
CD = (3.31)
π·Λ
The hydrodynamic rudder angle describes the angle between the flow velocity at the
rudder position and the ship’s longitudinal axis and follows from:
vr
δh = arctan (3.32)
ur
34 Chapter 3 – Mathematical model

while the velocity at the rudder is calculated with:

q
Vrr = u2r + vr2 (3.33)

The lift and drag on the rudder are calculated using:

1
Lru = ρVrr2 AR · CL · cos δe · sin δe (3.34)
2
1
Dru = ρVrr2 AR · CD · sin2 δe (3.35)
2
This results in the longitudinal and transverse forces and the yawing moment on the
ship due to the rudder:

XR = HX = −Dru · cos δh − Lru · sin δh (3.36)

YR = HY = (1 + aH ) · (Lru · cos δh − Dru · sin δh ) (3.37)

NR = YR · xr − XR · yr (3.38)
The coefficients Crue , Cdb , Cdr and aH are ship-dependent coefficients, see section 7.2.

Rudder angle The rudder angle is determined by the required rudder angle δreq and
the actual rudder angle δ, mimicking a simplified steering machine. The required rudder
angle is set depending on the type of manoeuvre (e.g. auto-pilot, zig-zag, turning-circle).
To obtain the actual rudder angle for a new time step, the difference between the required
rudder angle and the actual rudder angle at the current time step is determined first:

∆δ = δreq − δ (3.39)

When |∆δ| is less than a certain threshold (e.g. 3◦ ), then the rudder rate is calculated by:

δ̇ = Cr · ∆δ (3.40)

with Cr a rudder rate constant obtained from Cr = δ̇max /3. For |∆δ| above the threshold,
the rudder rate is given by:
δ̇ = δ̇max · sign∆δ (3.41)
The actual rudder angle at the new time step is obtained by integrating the rudder
rate in time.

3.3 Deriving the hydrodynamic coefficients for the

user-defined hull forces
In this section, the procedure to derive the coefficients for the user-defined hull force model
as presented in section 3.2.3 is described. The damping coefficients in the mathematical
model are derived in four steps:
3.4 – Conclusion 35

1. The linear coefficients for simple motions (slope of force or moment curves at β = 0◦
resp. γ = 0) are found as follows: For steady drift manoeuvres, the obtained forces
or moments are divided by cos β · sin β and the coefficients are taken from the
intersection at β = 0◦ of a linear or polynomial trend line through the data points.
For steady rotation, the same procedure is applied to the forces and moments divided
by γ.

2. Non-linear coefficients for pure transverse motion (β = 90◦ ) and pure rotation
(V = 0) are found using empirical relations (based on the work of Hooft [63],
e.g.). Currently, due to the unsteady nature of these manoeuvres and the com-
plexity of the flow around the hull, these motions are not solved using viscous flow
calculations.

3. Other non-linear components for simple motions can be determined by subtracting

the contributions from the coefficients found in steps 1 and 2 from the calculated
total bare hull forces; the non-linear components for the simple motions can then
be determined using curve fitting.

4. The cross-terms, based on combined motions, are found in a similar way to step 3.
The known contributions of the coefficients from steps 1-3 are subtracted from the
calculated bare hull forces and the remainder is used to fit the cross-terms.

This approach is chosen to enable accurate modelling of the linearised behaviour for
course-keeping (step 1), realistic modelling of the harbour manoeuvring characteristics
(step 2), and accurate modelling of non-linear manoeuvres (steps 3 and 4). To ensure
appropriate responses for astern manoeuvres, it is assumed that the forces and moments
on the hull during astern manoeuvres are identical to those during ahead manoeuvring.
However, if different forces and moments are desired for astern motion, these can be ob-
tained by selecting the linear derivatives based on the sign of the longitudinal ship velocity,
for example, as follows for the coefficient Yβ0 , with Yβ,ahead
0
the appropriate coefficient for
0
ahead speed and Yβ,astern for astern speed:

Yβ0 = Yβ,ahead
0 0
· max(0, sign(cos β)) + Yβ,astern · max(0, −sign(cos β)) (3.42)

3.4 Conclusion
In this chapter, the mathematical model used by SurSim to simulate the manoeuvra-
bility of ships has been described. In the program, it is possible to provide user-defined
hydrodynamic coefficients for the bare hull forces and moments. This makes the program
well suited for the present work to investigate whether the use of viscous-flow calculations
can help in improving the prediction accuracy of simulations.
A procedure to derive the coefficients is proposed. This procedure is chosen to en-
able accurate modelling of the linearised behaviour for course-keeping as well as realistic
36 Chapter 3 – Mathematical model

modelling of the harbour manoeuvring characteristics, and to enable accurate modelling

of non-linear manoeuvres.
In the following chapter, the solvers used in the work for this thesis to calculate the
viscous flow around a hull in manoeuvring motion will be discussed.
 37

Chapter 4

Viscous flow solvers

This chapter provides short background information regarding the viscous-flow solvers
used for the study discussed in this thesis. The grid procedures and the boundary condi-
tions used for the calculations are also presented in this chapter. Since the present work
focussed on the practical application of viscous-flow calculations and not on developing
the solvers or numerical procedures themselves, the theoretical details of the solvers will
not be discussed.

4.1 Background
Two different flow solvers have been used within this study. Initially, the calculations
were done with Parnassos, which is a solver optimised to calculate the flow around
ships in straight-ahead motion. For ships at drift angles or sailing in rotational motion,
it was found that considerable adjustments to the computational parameters (e.g. under-
relaxation) were required to reach the desired convergence levels. In the mean time, the
developments on the more general purpose solver ReFRESCO were started. With this
solver, it appeared to be easier to calculate the flow around ships in manoeuvring mo-
tions, especially for larger drift angles and rotation rates, and when using more advanced
turbulence models. Therefore, results obtained with both solvers are discussed in this
thesis.

4.2 Parnassos
Parnassos is one of MARIN’s in-house incompressible viscous-flow solvers. It is based
on a finite-difference discretisation of the Reynolds-averaged continuity and momentum
equations, using fully collocated variables and discretisation. The equations are solved
with a coupled procedure, retaining the continuity equation in its original form. Generally,
the governing equations are integrated down to the wall, i.e. no wall functions are used.
In Parnassos, multi-block structured grids are used. The implementation of the code is
optimised for solving the flow around ships with the mean flow directed along the ship’s
38 Chapter 4 – Viscous flow solvers

longitudinal axis. This makes the solver especially suitable for calculating and optimising
the resistance or propulsion of ships within a very short time frame, see e.g. Raven et al.
[114] or Van der Ploeg and Raven [154].
More detailed information about the solver can be found in Hoekstra and Eça [60],
Hoekstra [59], Raven et al. [113] or Eça and Hoekstra [40].

4.3 ReFRESCO
ReFRESCO is a MARIN spin-off of FreSCo [157], which was developed within the
VIRTUE EU Project together with Technische Universität Hamburg-Harburg (TUHH)
and HSVA. ReFRESCO is an acronym for Reliable and Fast Rans Equations solver for
Ships, Cavitation and Offshore. It solves the multi-phase unsteady incompressible RANS
equations, complemented with turbulence models and volume-fraction transport equations
for each phase. The equations are discretised using a finite-volume approach with cell-
centred collocated variables. The implementation is face-based, which permits grids with
elements with an arbitrary number of faces (hexahedrals, tetrahedrals, prisms, pyramids,
etc.). The code is parallelised using Message Passing Interface (MPI) and sub-domain
decomposition. Low order and higher-order spatial and temporal discretisation schemes
are available in the code. The equations are solved in a segregated approach, and the
pressure/velocity coupling is solved using the Semi-Implicit Method for Pressure Linked
Equations (SIMPLE) algorithm. The code is targeted, optimised and highly validated for
hydrodynamic applications, in particular for obtaining current, wind and manoeuvring
coefficients of ships, submersibles and semi-submersibles, see [159, 158, 51, 79]. Automatic
wall functions are available. In the present work, however, y2+ values below 1 are obtained
and all equations are integrated down to the wall.

4.4 Turbulence closure models

Several turbulence closure models are available in both Parnassos and ReFRESCO.
The most commonly used are the one-equation turbulence models proposed by Menter
[92] (designated MNT) or by Spalart and Allmaras [129] (designated SA), or two versions
of the two-equation k − ω turbulence model, i.e. the Shear Stress Transport (SST) version
[91] and the Turbulent/Non-Turbulent (TNT) version [78]. The Spalart correction of the
stream-wise vorticity (proposed by Dacles-Mariani et al. see [31]) can be activated.

4.5 Implementation of rotational motion

4.5.1 Background
For ship manoeuvres, not only oblique flow is of interest, but also the flow around the ship
when it performs a rotational motion. In RANS, this can be solved in several ways, such
4.5 – Implementation of rotational motion 39

as moving the grid in a rotational motion through a stationary flow (inertial reference
system), or by letting the flow rotate around the stationary ship (non-inertial reference
system). The latter is adopted in Parnassos and ReFRESCO. This approach has been
used by several authors, see for example section 3.2 in Batchelor [11] or section 1.15 in
Wesseling [162] and the applications to ships of e.g. Alessandrini and Delhommeau [3],
Cura Hochbaum [27] or Ohmori [99]. Using this system, the grid is attached to the hull
form and rotates with the ship. However, each water particle now should experience
centrifugal and Coriolis forces due to the rotation of the coordinate system. These forces
have to be added to the momentum equation as source terms.

4.5.2 Governing equations

Originally, the momentum equations in Parnassos read in a Cartesian coordinate system
(see equation (2.1) in Hoekstra [59])1 :
 
ρuj ui,j + p,i − µui,jj + ρ u0i u0j = fi (4.1)
,j

with fi the force per unit volume that is exerted on a discrete flow volume.
Assuming a steady flow, the rotational motion is simulated by implementing the
centrifugal and Coriolis force as additional force terms, such that the modified momentum
equation reads:
      
ρuj ui,j + p,i − µui,jj + ρ u0i u0j = fi − ρ 2Ω × u −ρ Ω× Ω×r (4.2)
,j i i

with fi a remaining force term per unit volume (e.g. propeller forces), Ω the vector of
rotation, u = (u1 , u2 , u3 ) = (u, v, w) the velocity vector and r = (x − xR ) the radius of
rotation with xR the position of the centre of rotation. In the equation above,
 the
 Coriolis
force is represented by −2ρΩ × u while the centrifugal force is −ρΩ × Ω × r .
In ReFRESCO the Coriolis  and
 centrifugal
 contributions are added to the external
force fi : i.e. fi = −ρ 2Ω × u − ρ Ω × Ω × r .
i i

4.5.3 Boundary conditions

At the outer boundaries in Parnassos calculations, it is assumed that the velocities and
pressure correspond to a solution for potential flow, except for the normal velocity, which
is left free to account for the displacement effect of the boundary layer. The pressure is
based on Bernoulli’s law, but a correction for the rotational motion is however required.
Using section 3.5 of Batchelor [11], it can be shown that the pressure follows from:
p − p∞ 1 
 2 
= Ω × r − |u|2 (4.3)
ρ 2
1
In Hoekstra [59] the index notation and summation convention as given by Aris [7] is used. In this
notation, whenever a subscript appears twice in a term, summation over the range of that index is implied.
A comma in the subscript denotes differentiation.
40 Chapter 4 – Viscous flow solvers

Such a boundary condition is not available in ReFRESCO. Generally, undisturbed

velocities are prescribed at the outer boundary. The modification of the pressure is not
required at the outer boundaries in ReFRESCO since for such boundaries the pressure
is extrapolated from the interior domain.

4.6 Grid generation

4.6.1 Parnassos
For Parnassos, the grids are generated with in-house tools, see Eça et al. [43]. Use is
made of block-structured grids with H-O topology2 , with grid clustering near the bow and
propeller plane in order to resolve the large gradients in those regions. In Parnassos,
the flow must be more-or-less parallel to far-field boundaries or perpendicular to inflow or
outflow planes. This means that to incorporate drift angles or rotational motion the outer
boundaries of the computational domain must be adjusted to match the flow direction.
Furthermore, calculations for larger drift angles and rotation rates require larger domain
sizes. For this, an automated procedure has been made. First, a high-quality base grid
is generated around the hull form. Surface grids on the boundaries of the base grid are
made first and based on these surface grids the interior 3D grid is generated. The outer
boundary of the base grid is located close to the hull, such that the quality of the grid
can be controlled easily by adjusting the surface grids.
When the base grid with the desired grid quality is obtained, it is used to generate grids
for each drift angle or rotation rate (or combination thereof). This is done by rotating the
base grid according to the desired drift angle and subsequently generating the surrounding
grid blocks to match the outer boundaries of the domain. Since the surrounding blocks
are of relatively simple shapes, no user interaction is required to obtain the desired grid
qualities and automatic scripts can be used. This procedure takes just several minutes on
a normal PC, even for fine grids. In Figure 4.1 an example grid for the HTC for a steady
yaw calculation is given. The base grid is raised with respect to the surrounding blocks
to illustrate the procedure.
In Parnassos calculations, the solution from a potential flow calculation is used to
set the pressure and tangential velocity at the far-field boundary and the velocities at
the inlet. With this approach, the size of the computational domain can be smaller than
the size of the computational domain for a calculation with undisturbed velocities at the
far-field or inlet boundaries, as was demonstrated by Eça and Hoekstra [42].

4.6.2 ReFRESCO
For best performance of ReFRESCO, multi-block structured O-O grids3 are used in
general. Calculations for ships at drift angles or rotation rates are conducted by setting the
2
In an H-O grid, one set of grid lines is generally aligned with the incoming flow direction.
3
In O-O grids, grid lines are aligned perpendicular or parallel to the surface.
4.6 – Grid generation 41

Figure 4.1: Example grid, Parnassos, HTC, γ = −0.556 (coarsened for presentation)

boundary conditions to the proper inflow velocities. Unlike the grids used in Parnassos,
the computational domain does not need to be changed for each new calculation. For
simple geometries and grid topologies, the grids for ReFRESCO can be made using
in-house tools. For more generic applications, use is made of commercial grid generation
tools, such as ANSYS ICEM CFD, Numeca HEXPRESS, or PDC GridPro. In these
tools, the grid can be generated and boundary conditions can be defined. However, when
calculations are to be conducted for a range of drift angles or rotation rates, the re-
definition of inflow or outflow boundaries to accommodate each computational condition
requires user interaction. To avoid this, a new boundary condition has been implemented,
designated BCAutoDetect. With this type of boundary condition, it is possible to use a
single grid for different inflow angles, by automatically determining whether faces on the
exterior domain are inflow faces or outflow/pressure faces.
For each face on the exterior boundary, the angle between the velocity u and the
surface normal n is calculated, using:
!
u n
α = arccos · (4.4)
|u| |n|

Surface normals on the exterior boundary are always directed outward of the domain.
Based on the projected velocity, the boundary condition on a face is either set to inflow
(α ≥ αad ) or outflow/ pressure (α < αad ), with αad an adjustable angle. By default αad
is chosen to be 87◦ . This appeared to give slightly better convergence properties than an
angle of 90◦ .
To further facilitate the use of a single grid for all computations, the far field boundary
is generated as a cylindrical or spherical surface. An example grid for the HTC is given
42 Chapter 4 – Viscous flow solvers

Figure 4.2: Example grid, ReFRESCO, HTC (coarsest grid)

in Figure 4.2.
In order to efficiently generate results for many drift angles, a procedure was imple-
mented to automatically increment the drift angle during a single simulation. Simulations
begin at a pre-set drift angle, until a specified number of iterations is reached, or when
the non-dimensional residuals are less than a specified convergence criterion. Next the
drift angle is incremented by ∆β, by changing the inflow conditions, and the solution is
continued from the solution at the previous drift angle. Starting the calculations from a
converged solution at a slightly different drift angle saves time compared to performing
each calculation separately. This procedure is repeated until the desired maximum inflow
angle is reached. In Figure 4.3, it is demonstrated that this approach (continuous lines)
provides the same results as those obtained with multiple single-drift angle calculations
(markers). This procedure was designated drift sweep and the application has already
been presented in e.g. Vaz et al. [158] and Bettle et al. [15].

4.7 Boundary conditions

The calculations presented in this thesis were all conducted without incorporating free-
surface deformation. Based on the speeds used during the experiments for each test case
and the range of drift angles or rotational rates studied, the effects of speed and free-
surface deformation on the forces on the manoeuvring ship are likely to be small and
assumed to be smaller than the uncertainties due to e.g. discretisation errors or errors
in the experimental results. Therefore, symmetry boundary conditions were applied on
the undisturbed water surface. On the hull surface, no-slip and impermeability boundary
4.8 – Conclusion 43

Re = 14 × 106 , ∆β = 5◦ Re = 14 × 106 , ∆β = 5◦
0.015 0.005
istep = 1500 0.004 istep = 1500
0.01 istep = 750 istep = 750
istep = 3000 0.003 istep = 3000
0.005 steady 0.002 steady
0.001
0 0

N
Y

-0.001
-0.005 -0.002
-0.01 -0.003
-0.004
-0.015 -0.005
-20 -10 0 10 20 -20 -10 0 10 20
β [deg] β [deg]

Figure 4.3: Verification of drift sweep procedure, SUBOFF

(istep indicates the maximum number of iterations to perform for one drift angle)

conditions are used. The velocities are set to zero (u = 0).

The velocity components at the inflow plane and on the external boundary in Par-
nassos calculations are taken from a potential flow calculation. For the inflow boundary,
the three velocity components are taken from the potential flow solution, while in the
external boundary, the tangential velocities and the pressure are set. During the viscous-
flow calculation, the velocity normal to the external boundary is updated to allow for the
displacement effect of the boundary layer. Because the velocity and pressure behind the
ship are unknown, Neumann boundary conditions are applied on the outflow plane.
For ReFRESCO, the boundary conditions on the exterior domain are determined
using the BCAutoDetect boundary condition, which automatically applies inflow (Dirich-
let) or outflow (Neumann) conditions on the cell faces. If a cylindrical domain is used,
boundary conditions need to be set on the bottom/lower surface. For shallow water cal-
culations, the boundary condition will be set to moving-wall/fixed slip (u = V ∞ ). For
unrestricted calculations, the boundary condition on the lower surface can be set to either
a pressure boundary condition or to a symmetry boundary condition.

4.8 Conclusion
In this chapter, the viscous-flow solvers used in this thesis have been presented. For
ship manoeuvres, not only the flow around the ship in oblique motion is of interest, but
also the flow around the ship when it performs a rotational motion. To compute the
flow around the ship in rotational motion, the flow solvers had to be modified. For this
work, the rotational motion was incorporated by using a non-inertial reference system
and supplementing the equations of motions with body forces representing the centrifugal
and Coriolis contributions to the flow.
44 Chapter 4 – Viscous flow solvers

To generate the grids for a range of drift angles and yaw rates, different approaches
were adopted depending on the flow solver used. For Parnassos, automated scripts
were developed with which the desired grids could be generated rapidly. For the more
generic solver ReFRESCO, all computations are conducted using one grid, but changing
the inflow angles and boundary conditions depending on the desired manoeuvring motion.
For this, a new boundary condition was developed, which removes the need to pre-process
each grid for each new manoeuvring condition.
To further improve the efficiency of the ReFRESCO calculations, a so-called drift
sweep procedure was developed to automatically calculate the forces and moments on the
ship for a range of drift angles.
The following chapter presents verification and validation of the viscous-flow calcula-
tions from which the hydrodynamic coefficients for the bare hull forces in the mathematical
model can be derived.
 45

Chapter 5

Verification and validation of steady

motion calculations

In this chapter, the verification and validation (V&V) of viscous-flow calculations for ship
hulls in steady manoeuvring motion are discussed. Results of several test cases will be
presented to demonstrate that the flow solvers can be applied to compute the viscous-flow
around various hull forms and that accurate predictions of the forces and moments on the
manoeuvring ship can be obtained. The results for the HTC will be used in chapter 7
to derive hydrodynamic coefficients for the user-defined mathematical model for the hull
forces in SurSim, as presented in chapter 3.

5.1 Introduction to verification and validation

In order to assess whether computational results are reliable and accurate, verification
(solving the equations right [115]) and validation (solving the right equations) studies are
required. Two quantities are used: errors and uncertainties, designated by respectively
δ and U . An error is defined as the difference between a simulated or measured value
and a comparison value (sometimes called the true value), and it is therefore a quantity
that has a sign and magnitude. An uncertainty defines an interval containing the true
value within a certain degree of confidence. In verification, uncertainties in the results are
assessed, while the aim in validation is to show the suitability of the selected model for the
problem. Verification can be divided in two parts: code verification in which it is verified
whether a code correctly solves the equations of the model; and solution verification in
which the numerical uncertainty in a prediction is quantified. This thesis concentrates
on solution verification and validation. Code verification is outside of the scope of this
thesis and interested readers are referred to e.g. Eça et al. [44] in which the viscous-flow
solver Parnassos is subjected to code verification. Similar efforts have been undertaken
to verify ReFRESCO.
In solution verification, the numerical uncertainty USN in a solution φ is estimated.
The exact solution φexact is unknown. The numerical uncertainty is governed by a nu-
46 Chapter 5 – Verification and validation of steady motion calculations

merical error multiplied by a factor of safety Fs . The numerical error consists of three
components [115]: the round-off error; the iterative error and the discretisation error. The
viscous-flow solvers used in this study use double precision and therefore the round-off
error becomes negligible compared to the other errors. This means that the numerical
uncertainty USN is obtained by1

USN = UI + UG (5.1)

The uncertainty UI due to the iterative process depends on whether the equations are
sufficiently resolved. In simple flow problems, a reduction of the iterative error to machine
accuracy is sometimes possible, but for complex flows this may be too time consuming.
Estimations of UI based on changes in the solution during the iterative process are then
required. The discretisation uncertainty UG is obtained by estimating the discretisation
error δG and multiplying this with the factor of safety. When the differences between
solutions on progressively finer grids reduce (i.e. we are converging towards the exact
value), the discretisation error can be estimated with Richardson extrapolation (RE) and
the use of the Grid Convergence Index (GCI) [115]:

δG ≈ δRE = φi − φ0 = α · hpi (5.2)

in which φi stands for the value of a considered quantity, φ0 is the estimate of the exact
solution, α is a constant, h is the typical cell size and p is the observed order of accuracy.
Additional or alternative error estimators are also used. With this estimator, the uncer-
tainty follows from: UG = |δG | · Fs . See Eça et al. [44, 45] for a complete discussion of the
procedure.
The aim of validation is to establish the comparison error E and the validation un-
certainty Uval and to obtain an interval that contains the modelling error δmodel . An error
δ is supposed to have a magnitude and sign, while an uncertainty U is used to designate
an interval containing an error of unknown magnitude and sign. The comparison error is
defined by the difference between the simulated value S and the experimental data value
D:
E =S−D (5.3)
This error contains all errors in the experiment as well as in the simulation. The validation
uncertainty is estimated by:
q
Uval = UD2 + USN
2 2
+ Uinput (5.4)

in which UD is the uncertainty in the experimental result and Uinput is the uncertainty in
the input parameters (e.g. fluid properties, geometry). Evaluation of Uinput is assumed to
be outside the scope of the present workq and is taken as zero for simplicity. Therefore
Uval can be calculated through Uval = UD2 + USN 2
. According to the American Society
1
According to the ASME V&V-20 procedure, Root mean square (RMS) addition cannot be applied
to the uncertainties, because of the dependency of e.g. UG on UI .
5.2 – KVLCC1, KVLCC2 and KVLCC2M 47

Of Mechanical Engineers (ASME) V&V-20 procedure [9], the modelling error δmodel is
defined by:
δmodel = E − (δSN + δinput − δD ) (5.5)

By comparing E with the validation uncertainty, the following statements can be

made:

|E| >> Uval : the comparison error is governed by the modelling error (δmodel ≈ E), which
indicates that the model should be improved to reduce the comparison error.

|E| < Uval : the modelling error is within the noise imposed by the uncertainties contained
in the validation uncertainties. Conclusions about the modelling error can only be
drawn when the uncertainties are reduced. In this case, it is said that the model
and its solution are validated at a level of Uval .

In this thesis, solutions with |E| < Uval are indicated with a check-mark (✔), while
solutions with |E| >> Uval are not validated and therefore indicated with a cross (✘).

In the past, extensive procedures for V&V have been presented, such as the Interna-
tional Organization for Standardization (ISO) Guide to the Expression of Uncertainty in
Measurement [72], Stern et al. [133] and recently the ASME Standard for Verification and
Validation in Computational Fluid Dynamics and Heat Transfer [9]. In these procedures,
guidelines are given, but references to methods which can be used to obtain improved
uncertainty estimates are provided. The procedure based on a least-square version of the
Grid Convergence Index (GCI) as proposed by Eça et al. [44, 45] is followed in this thesis.

5.2 KVLCC1, KVLCC2 and KVLCC2M

The KVLCC hull forms have been described in section 2.5.3. In this section, RANS
calculations for these hulls will be presented and compared to the available experimental
validation data. Only the conditions representing deep water will be considered.
With Parnassos, viscous flow calculations have been conducted for the KVLCC2M
at several drift angles. These have been reported previously by Toxopeus [143]. The
calculations were done with an undisturbed water surface, i.e. neglecting the generation
of waves. The Reynolds number was set to Re = 3.945 × 106 . The Menter one-equation
turbulence model was used.
Calculations with Parnassos were performed for the KVLCC1 and KVLCC2 hull
forms, see Toxopeus and Lee [149]. Also for these calculations an undisturbed water
surface and the Menter one-equation turbulence model were used. Contrary to the model
tests, it was assumed that the model attitude was fixed, i.e. dynamic trim, sinkage and
heel were not taken into account. The grid consisted of 3.356 × 106 nodes. The Reynolds
number was set to Re = 3.27 × 106 .
48 Chapter 5 – Verification and validation of steady motion calculations

ReFRESCO-SST, Re = 3.7 × 106 , β from 0◦ to 32◦

10+0
10−2
10−4 0◦ 4◦ 8◦ 12◦ 16◦ 20◦ 24◦ 28◦ 32◦
10−6
∆Y

10−8
10−10
10−12
10−14
0 25000 50000
iteration number

Figure 5.1: Convergence history Y -force, KVLCC2, ReFRESCO

The flow around the KVLCC2 hullform was also studied using ReFRESCO. One grid
with 5.388 × 106 cells was used for all drift angles, while the computations for different
rotation rates were made with a finer grid of 12.721 × 106 cells. Similar to the Parnassos
calculations, an undisturbed water surface and a fixed model attitude were assumed. The
Reynolds number was set to the slightly different value of 3.7×106 and the SST turbulence
model was used. The relation between the drift angle and the forces and moments on the
hull was obtained with the drift sweep procedure as introduced in section 4.6.2.
An overview of all experimental and computational results of the forces and moments
on the KVLCCs for captive conditions is given in the tables on page 151 through page 156.

5.2.1 Iterative error

In the Parnassos calculations a reduction of the maximum difference in the pressure
coefficient between consecutive iterations to 5 × 10−5 was adopted as the convergence
criterion. For all cases, the adopted convergence criterion results in a reduction of the
difference in the (total) non-dimensional force and moment components between consecu-
tive iterations of well below 5 × 10−5 . This is more than two to three orders of magnitude
smaller than the discretisation uncertainty and therefore the iterative uncertainty UI can
be neglected, see Eça et al. [41].
For ReFRESCO, all calculations were run until the maximum non-dimensional resid-
ual of the pressure resp,max (the so-called L∞ norm) between successive iterations had
dropped well below 1 × 10−5 or when further iterative convergence was not obtained. The
changes in the non-dimensional integral quantities (forces and moments) were well below
1 × 10−7 . An example of the convergence history of the transverse force Y is given in
Figure 5.1. In this calculation, the drift angle is increased after a predefined number of
iterations from 0◦ to 30◦ in steps of 2◦ . Each increase of the drift angle is visible as a
5.2 – KVLCC1, KVLCC2 and KVLCC2M 49

spike in the convergence history. For each drift angle, the changes in Y force reduce to
below 1 × 10−10 . The convergence of the other force components is similar.

5.2.2 Discretisation error

Several grid topologies have been used for the Parnassos calculations of the flow around
the KVLCC2M [39]. The results presented in this thesis were all obtained on structured
grids with H-O topology with extra grid clustering close to the bow and propeller plane.
For each grid, the variation in the number of grid nodes in the stream-wise, normal and
girth-wise (nξ , nη and nζ ) directions is presented in Table 5.1, which also includes the
maximum y + value for the first cells adjacent to the hull, designated y2+ , that was obtained
during the calculations. It is seen that y2+ < 1 which is required when the flow is to be
resolved down to the wall.

Table 5.1: Properties of grids for different drift angles, KVLCC2M, Parnassos

Fn β xin xout |y|max zmax nξ nη nζ Nodes y2+

0◦ 0.73 -0.92 0.18 0.18 449 81 45 1.6×106 0.32
3◦ 0.74 -0.93 0.42 0.36 449 95 2×45 3.8×106 0.40
- 6◦ 0.75 -0.94 0.49 0.36 449 95 2×45 3.8×106 0.55
9◦ 0.76 -0.95 0.55 0.36 449 95 2×45 3.8×106 0.69
12◦ 0.76 -0.95 0.61 0.38 449 95 2×45 3.8×106 0.80

For a drift angle of 12◦ , a series of geometrically similar grids has been generated
in order to investigate the discretisation error. The grid coarsening has been conducted
in all three directions. For some of the grids, however, the distance of the first node
to the hull surface has been maintained in order to capture the velocity gradients in
the boundary layer. This might introduce scatter in the results due to non-geometric
similarity. Table 5.2 shows the number of nodes and y2+ values for these grids.

Table 5.2: Properties of grids for uncertainty analysis, KVLCC2M, Parnassos, β = 12◦

id β nξ nη nζ hi Nodes y2+ Comment

1 12 449 95 2×45 1.00 3838950 0.80
2 12 409 87 2×41 1.10 2917806 0.75
3 12 361 81 2×37 1.24 2163834 0.62
4 12 329 74 2×33 1.37 1606836 0.71
5 12 297 65 2×29 1.51 1119690 0.94
6 12 249 65 2×25 1.81 809250 0.78
7 12 225 48 2×23 2.00 496800 1.15 based on grid 1, coarsened by 2×2×2
8 12 177 41 2×19 2.55 275766 1.25 based on grid 3, coarsened by 2×2×2
9 12 145 33 2×15 3.11 143550 1.73 based on grid 5, coarsened by 2×2×2
10 12 121 33 2×13 3.73 103818 1.32 based on grid 6, coarsened by 2×2×2
11 12 113 24 2×12 4.00 65088 2.17 based on grid 1, coarsened by 4×4×4
50 Chapter 5 – Verification and validation of steady motion calculations

For grid 5, it was not possible to reach the required convergence criterion. Therefore
the results for this grid are dropped from further analysis.
For a drift angle of 12◦ , the predicted values of the friction (index f ) and pressure
(index p) components as well as the total force and moment coefficients for each force or
moment variable φ are given in Table 5.3 with their estimated uncertainties Uφ . Based
on an analysis of the results for each grid, it was decided to use the 6, 7 or 8 finest grids
for the uncertainty analysis. The number of grids ng used depended on the scatter in the
results for the coarsest grids. In the table, φ1 indicates the solution obtained on the finest
grid, φ0 the extrapolated solution and p the apparent order of convergence.

Table 5.3: Uncertainty analysis, KVLCC2M, Parnassos, β = 12◦

Item φ0 φ1 Uφ p
Xp - −2.32 × 10−3 32.3% 1

Xf −1.57 × 10−2 −1.54 × 10−2 2.1% 1.45

X - −1.78 × 10−2 7.8% 2

Yp −6.45 × 10−2 −6.26 × 10−2 4.8% 1.25

Yf −1.84 × 10−3 −1.70 × 10−3 13.4% 1.34
Y −6.67 × 10−2 −6.43 × 10−2 5.6% 1.13
Zp 3.39 × 10−1 3.20 × 10−1 7.7% 0.52
Zf - 1.20 × 10−3 13.2% 2 1 Oscillatory convergence
Z 3.41 × 10−1 3.21 × 10−1 7.9% 0.51 2 Monotonic divergence
Kp - 3.24 × 10−3 9.3% 1 3 Oscillatory divergence
Kf −2.16 × 10−4 −1.74 × 10−4 30.7% 0.44
K - 3.07 × 10−3 6.7% 1

Mp - −3.97 × 10−2 3.8% 2

Mf 1.09 × 10−3 1.08 × 10−3 0.7% 1.71

M - −3.86 × 10−2 4.0% 2

Np - −2.55 × 10−2 9.2% 2

Nf - 2.94 × 10−4 25.4% 2

N - −2.53 × 10−2 9.6% 2

The absolute value of the uncertainty in the pressure components is larger than in the
friction components. The uncertainty in the longitudinal friction component Xf is about
one-third of the uncertainty in the longitudinal pressure component Xp . For the other
forces and moments, the uncertainty in the friction component is at least one order of
magnitude smaller than the uncertainty in the pressure component. Since most integral
forces and moments are dominated by the pressure component, this results in relatively
large uncertainties in the overall forces and moments.
In Figure 5.2, the convergence of the side force and yaw moment coefficients with grid
refinement is presented. It is seen that upon grid refinement, the estimated value for Y
(indicated by cfd) comes closer to the experimental value (indicated by exp). Considerable
scatter is visible in the data and therefore it is not easy to establish whether data points are
located in the asymptotic range of convergence. This is typical for this type of calculation,
5.2 – KVLCC1, KVLCC2 and KVLCC2M 51

as already observed previously by e.g. Eça et al. [38] and Hoekstra et al. [61].
Looking at the yawing moment N , the maximum difference between the estimated val-
ues for all grids is 5.1%. Because the difference between the estimated values is relatively
small and scatter on the data is present, monotonic divergence is found and extrapolation
to zero step size could not be made. This results in a relatively large uncertainty of 9.6%.

-0.054 -0.0225
-0.056 p=1.1 -0.023 U=9.6%
-0.058 U=5.6% -0.0235 exp
-0.024
-0.06 exp -0.0245
-0.062 -0.025

N
Y

-0.064 -0.0255
-0.066 -0.026
-0.068 -0.0265
-0.027
-0.07 -0.0275
-0.072 -0.028
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.5 1 1.5 2 2.5 3 3.5 4
relative step size relative step size

Figure 5.2: Uncertainty analysis, KVLCC2M, Parnassos, β = 12◦

5.2.3 Local quantities

The experimental data of the wake field in the complete WAKE1 plane (near the propeller
plane, perpendicular to the incident flow) have kindly been made available by National
Maritime Research Institute (NMRI). In Figure 5.3 comparisons of the axial velocity fields
between the experiments (dotted lines) and the Parnassos calculations (solid lines) for
0◦ , 6◦ and 12◦ drift angle are made.
This figure shows that in most parts of the plane, the viscous-flow calculations corre-
spond well with the experiments. Even for 12◦ drift angle, the strength and position of the
vortex generated at the starboard bilge (its centre is located at y = 0.11Lpp , z = 0.03Lpp )
is quite accurately captured by the calculations.
In the port side area (windward), discrepancies are found for the 12◦ drift case, how-
ever. In the calculations, the contour lines are straightened while they retain their hook-
shape in the experimental results. Also just behind the propeller hub for 0◦ drift angle,
the hook-shape in the measurements appears more pronounced than in the calculations.
This can be attributed to the turbulence modelling, as was also observed by Eça et al.
[38].
Overall, it is concluded that the flow field around the aft ship is quite accurately
predicted.

5.2.4 Global quantities

By comparing the viscous flow calculation results for the KVLCCs, the influence of the
stern shape on the forces and moments on the hull can be investigated. In Figure 5.4, a
comparison is graphically presented.
52 Chapter 5 – Verification and validation of steady motion calculations

0.00
0.9

0.8
0.8
0.7
0.95

0. 0.2
4

0.9
0.7

0.95
0.02

0.4
0.

0.4
6
0.7

0.6
7
0.
0.
0.8

0.6

4
z/Lpp

0.8
0.04 0.9

0.80.6
0.40.7
0.8
0.

9
0.

0.2
95

0.7
0.06 95
0.

0.08 0.9

pp, β=0° Experiments, KVLCC2M, x=-0.48Lpp, β=0°

0.95
Parnassos, KVLCC2M, x=-0.48L
0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

y/Lpp
0.00
0.95

0.6

0.9
7
0. 8
0.4 0.4 0.
0.70.8

0.8
0.4

0.6
0.02 0. 0.7
6

0.95
0.9

0.7

0.6
0.4

0.9
0.7
0.9

0.7
0.9

0.8
0.6
z/Lpp

0.04
0.70.80
0.9

0.9
0.8
5

0.4
0.6

0.4

0.9
.8

0.9
0.9
0.4
4
0.7.60.9
0.7 0.

0.06
0 .9
0

5
0.9
0.9

0.9

0.9
0.08 5

Parnassos, KVLCC2M, x=-0.48Lpp, β=-6° Experiments, KVLCC2M, x=-0.48Lpp, β=-6°

0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

y/Lpp
0.00
0.6

0.
0.7

0.2
9
0.9

0.9
0.95

00 0.4 0.8
0.6.4.4
0.02
0.8
0.8

0.7

0.8
0.
7
0.
8

0.8
0.9

0.9
8
0.6
0.6

0.
0.6

0.6
z/Lpp

0.04
0.7
0.

0.6
0.9

0.95
7
0.9

0.9
0.

0.8
0.4 0.7 0.8

0.9
0.9

0.9
0.6

0.7

5 0.95
0.95
0.9

0.8
0.9

0.06 0.9
0.6

0.95

0.08 0.95

Parnassos, KVLCC2M, x=-0.47Lpp, β=-12° Experiments, KVLCC2M, x=-0.47Lpp, β=-12°

0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

y/Lpp

Figure 5.3: Comparison with experiments, KVLCC2M, WAKE1

(solid lines: calculations, dotted lines: experiments)
5.2 – KVLCC1, KVLCC2 and KVLCC2M 53

Based on this comparison, it is seen that a trend may be found in the yaw moment N
against the drift angle due to the change in hull form: the yaw moment for the KVLCC1
is slightly higher than for the KVLCC2 and KVLCC2M hulls. This trend is visible in the
experiments and in the CFD results. However, no trend can be seen in the transverse force
Y or the force or moment against the yaw rate. Furthermore, one would expect the results
for the KVLCC2M to be closer to the KVLCC2 results than to the KVLCC1 results.
Therefore, it is concluded that the differences between the CFD results for the Y force
against the drift angle and Y and N against the yaw rate are within the accuracy of the
calculations and are not representative for the differences between the hull forms. Based
on the above observations, it is expected that the three hulls will have similar manoeuvring
behaviour. This has been demonstrated during the free sailing manoeuvring tests with
the KVLCC1 and KVLCC2, see MARIN Report No. 21571-1-SMB [87], in which it was
found that indeed there were no significant differences in the manoeuvring performances,
other than a small difference in the directional stability derived from pull-out tests.
A comparison between the Parnassos and ReFRESCO results for the KVLCC2
shows only marginal differences, even though a different turbulence model was used. Only
in the N moment against γ, a consistent difference between the results of the two solvers is
seen. The differences are within the uncertainty of the predictions and a reduction of the
uncertainties is required to investigate whether the differences are caused by modelling
errors. Therefore, no conclusions can be drawn about the influence of either the solver,
the grid layout or the turbulence model on the results.

5.2.5 Validation
The uncertainties in the measurements of the forces and moments are specified by Kume
et al. [82] and summarised in Table 5.4. The values for β = 12◦ were calculated by in-
terpolation between the uncertainties for β = 9◦ and β = 18◦ . Using the measurement
uncertainties and assuming that the simulation numerical uncertainty USN is only influ-
enced by the discretisation uncertainty UG (i.e. USN = UI + UG = UG ), Table 5.4 can
be constructed. The uncertainty Uinput due to uncertainties in the input parameters (e.g.
fluid properties,qgeometry) is assumed to be zero and therefore Uval can be calculated
through Uval = UD2 + USN 2
.
It is seen that for the longitudinal force X and yaw moment N the comparison error |E|
is smaller than the validation uncertainty Uval which means that the solution is validated
at levels of 8.6% and 10.4% respectively. These levels are judged to be good. If lower
validation levels are desired then the numerical uncertainty needs to be reduced. For the
transverse force Y validation is not achieved (|E| > Uval ), which indicates modelling errors.
It is seen that the magnitude of the Y force is under-predicted. Changes in turbulence
model or domain size or the inclusion of free surface may lead to improvements of the
comparison error.
54 Chapter 5 – Verification and validation of steady motion calculations

0 0.02
−0.01 Exp (KVLCC2)
Exp (KVLCC2M)
KVLCC1 (Parnassos)

X
−0.02
−0.1 0 KVLCC2 (Parnassos)
KVLCC2 (ReFRESCO)
−0.03 KVLCC2M (Parnassos)
Y

−0.2

N
−0.04
−0.02
0 10 20 30 40
β

−0.3 −0.04

−0.4 −0.06
0 10 20 30 40 0 10 20 30 40
β β
0.04 0.01

0.03 0

0.02 −0.01
Y

0.01 −0.02

0 −0.03

−0.01 −0.04
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
γ γ

Figure 5.4: Comparison between experiments and calculations, KVLCC hull forms (Lines rep-
resent CFD results)

Table 5.4: Measurement uncertainties and validation, KVLCC2M, Parnassos, β = 12◦

Measurement uncertainty Validation

β UD X (%D) UD Y (%D) UD N (%D) X Y N
0◦ 3.3 - - D× 103 -17.5 -70.8 -25.4
9 ◦ 3.6 3.2 4.3 S × 103 -17.8 -64.3 -25.3
18 ◦ 2.9 3.1 3.7 E = S − D (%D) 1.5 -9.2 -0.5
12 ◦ 3.4 3.2 4.1 UD (%D) 3.4 3.2 4.1
USN (%S)
q
7.8 5.6 9.6
Uval = UD 2 + U 2 (%D) 8.6 6.0 10.4
SN
Validated? ✔ ✘ ✔
5.3 – HTC 55

5.3 HTC
The HTC has been described in section 2.5.4. In this section, RANS calculations for this
hull will be presented and compared to available experimental validation data. The captive
model tests have been conducted for speeds of 1.05 m/s and 1.89 m/s, corresponding to
10 kn and 18 kn on full scale. On model scale (λ = 24), this results in Reynolds numbers
of Re = 6.29 × 106 and Re = 12 × 106 , respectively. All calculations presented in this
thesis have been performed at model scale Reynolds numbers. To clearly distinguish the
different speeds used for the calculations or model tests and avoid confusion with the
speeds for other scale factors and the free sailing tests, calculations for the lower speed
will be identified with 10 kn, while the calculations for the higher speed will be identified
with 18 kn.
Calculations using Parnassos were made for the HTC hull form, using the Menter
one-equation turbulence model. Free surface and appendages were not modelled in the
calculations. These calculations were previously documented in VIRTUE Deliverable 3.1.1
[152]. Table 5.5 shows the conditions for the 10 kn computations. The domain sizes and
variation in the number of grid nodes in the stream-wise, normal and girth-wise (nξ , nη
and nζ ) directions are given, together with the y2+ values that were obtained.

Table 5.5: Properties of grids for different drift angles, HTC, Parnassos, 10 kn

Fn β xin xout |y|max zmax nξ nη nζ Nodes y2+

0◦ 0.70 -0.75 0.16 0.16 329 81 51 1.2×106 0.56
2.5◦ 0.71 -0.76 0.35 0.31 377 95 2×51 3.7×106 0.74
- 5◦ 0.71 -0.79 0.38 0.31 377 95 2×51 3.7×106 0.81
10◦ 0.72 -0.91 0.45 0.31 377 95 2×51 3.7×106 0.90
15◦ 0.73 -1.11 0.52 0.32 377 95 2×51 3.7×106 0.90

Calculations have been conducted with ReFRESCO as well. Also in this case, free
surface and appendages were not modelled. All calculations were conducted for the 18 kn
condition and the SST turbulence model was selected for turbulence closure. One single
grid was used for all calculations. The steady drift calculations were performed with the
drift sweep procedure as introduced in section 4.6.2.
An overview of all experimental and computational results of the forces and moments
on the HTC for captive conditions is given in the tables on page 157 and page 158.

5.3.1 Iterative error

In the Parnassos calculations a reduction of the maximum difference in the pressure
coefficient between consecutive iterations to 5 × 10−5 was adopted as the convergence
criterion. For all cases, the adopted convergence criterion results in a reduction of the
difference in the (total) non-dimensional force and moment components between consecu-
tive iterations of well below 4 × 10−5 . This is more than two to three orders of magnitude
56 Chapter 5 – Verification and validation of steady motion calculations

smaller than the discretisation uncertainty and therefore the iterative uncertainty UI can
be neglected, see Eça et al. [41].
For ReFRESCO, all calculations were run until the maximum non-dimensional resid-
ual of the pressure resp,max (the so-called L∞ norm) between successive iterations had
dropped well below 1 · 10−5 or when further iterative convergence was not obtained. The
changes in the non-dimensional integral quantities (forces and moments) were well below
1 × 10−7 . An example of the convergence history of the transverse force Y is given in
Figure 5.5. In this calculation, the drift angle is increased after a predefined number of
iterations from 0◦ to 30◦ in steps of 2.5◦ . Each increase of the drift angle is visible as the
spikes in the convergence history. For each drift angle, the changes in Y force reduce to
below 1 × 10−8 . The convergence of the other force components is similar.

ReFRESCO-SST, Re = 12 × 106 , β from 0◦ to 30◦

10+0
10−2
10−4
◦
10−6 0 5◦ 10◦ 15◦ 20◦ 25◦ 30◦
∆Y

10−8
10−10
10−12
10−14
0 25000 50000 75000 100000
iteration number

Figure 5.5: Convergence history Y -force, HTC, ReFRESCO

5.3.2 Discretisation error for steady drift motion

For Parnassos, a series of geometrically similar grids has been generated for a drift
angle of 10◦ , in order to investigate the discretisation error. The grid coarsening has
been conducted in all three directions. For each grid, the variation in the number of grid
nodes in the stream-wise, normal and girth-wise (nξ , nη and nζ ) directions is presented
in Table 5.6, which includes also the y2+ values that were obtained during the calculations
conducted for the 10 kn condition.
For a drift angle of 10◦ , the predicted values φ1 of the friction (subscript f ) and
pressure (subscript p) components as well as the total force and moment coefficients are
presented in Table 5.7 with their estimated uncertainties Uφ . Based on an analysis of
the results for each grid, it was decided to use the eight finest grids for the uncertainty
analysis. The number of grids ng used was chosen based on the scatter in the results for
the coarsest grids.
5.3 – HTC 57

The absolute value of the uncertainty in the pressure components is larger than in the
friction components, as was already found during the uncertainty study for the KVLCC2M
hull form, see section 5.2. The uncertainty in the longitudinal friction component Xf is
about one-third of the uncertainty in the longitudinal pressure component Xp . For the
other forces and moments, the uncertainty in the friction component is at least one order
of magnitude smaller than the uncertainty in the pressure component. Since most integral
forces and moments are dominated by the pressure component, this results in relatively
large uncertainties in the overall forces and moments. In Raven, Van der Ploeg and

Table 5.6: Properties of grids for uncertainty analysis, HTC, β = 10◦ , Parnassos

id β nξ nη nζ hi Nodes y2+ Comment

1 10 377 95 2×51 1.00 3653130 0.90
2 10 361 91 2×49 1.04 3219398 0.86
3 10 297 77 2×41 1.25 1875258 0.97
4 10 257 65 2×35 1.47 1169350 1.19
5 10 185 48 2×26 2.00 461760 1.48 based on grid 1, coarsened by 2×2×2
6 10 177 46 2×25 2.08 407100 1.51 based on grid 2, coarsened by 2×2×2
7 10 145 39 2×21 2.50 237510 1.76 based on grid 3, coarsened by 2×2×2
8 10 129 33 2×18 2.94 153252 2.28 based on grid 4, coarsened by 2×2×2
9 10 89 23 2×13 4.17 53222 3.08 based on grid 2, coarsened by 2×2×2
10 10 73 19 2×11 5.00 30514 4.06 based on grid 1, coarsened by 2×2×2

Table 5.7: Uncertainty analysis, HTC, β = 10◦ , 10 kn, Parnassos

Item φ0 φ1 Uφ p
Xp −5.33 × 10−4 −3.40 × 10−3 111.1% 0.67
Xf - −1.22 × 10−2 2.6% 2

X −1.40 × 10−2 −1.57 × 10−2 14.1% 0.94

Yp −3.68 × 10−2 −4.31 × 10−2 19.5% 0.69
Yf −1.18 × 10−3 −1.12 × 10−3 7.2% 1.80
Y −3.76 × 10−2 −4.42 × 10−2 19.9% 0.65
Zp 7.86 × 10−2 8.56 × 10−2 5.8% 0.22
Zf - 3.40 × 10−4 4.4% 2 1 Oscillatory convergence
Z 7.86 × 10−2 8.60 × 10−2 5.8% 0.21 2 Monotonic divergence
Kp 2.07 × 10−2 1.96 × 10−2 7.0% 1.46 3 Oscillatory divergence
Kf −1.79 × 10−3 −1.76 × 10−3 3.3% 1.97
K 1.89 × 10−2 1.79 × 10−2 7.5% 1.42
Mp −2.24 × 10−3 −1.55 × 10−3 59.7% 0.75
Mf - 3.03 × 10−4 1.6% 2

M −1.92 × 10−3 −1.25 × 10−3 72.6% 0.76

Np −2.45 × 10−2 −2.44 × 10−2 1.3% 3.45
Nf 2.78 × 10−5 2.48 × 10−5 54.8% 1.98
N −2.45 × 10−2 −2.44 × 10−2 1.2% 3.48
58 Chapter 5 – Verification and validation of steady motion calculations

Eça [113], an extensive study to improve the uncertainty and accuracy of the pressure
resistance component is presented.

-0.013 0.022
-0.014 p=0.9
0.02
p=1.4
-0.015 U=14.1% U=7.5%
-0.016 exp 0.018 exp
-0.017 0.016
-0.018

K
X

-0.019 0.014
-0.02 0.012
-0.021
-0.022 0.01
-0.023 0.008
0 1 2 3 4 5 6 0 1 2 3 4 5 6
relative step size relative step size
-0.035 -0.022
p=0.6 p=3.5
-0.04 U=19.9% -0.0225 U=1.2%
exp -0.023 exp
-0.045
-0.0235
N
Y

-0.05
-0.024
-0.055 -0.0245
-0.06 -0.025
0 1 2 3 4 5 6 0 1 2 3 4 5 6
relative step size relative step size

Figure 5.6: Uncertainty analysis, HTC, β = 10◦ , 10 kn, Parnassos

In Figure 5.6 the forces and moments on the ship are shown for the different grids.
The scatter in the results is much smaller than found for the KVLCC2M results. For a
relative step size below 3, the results appear to converge. The observed convergence rate
p, however, is found to be small for both X and Y (p = 0.9 and 0.6 respectively). Due to
the slow convergence, the difference between the extrapolated value φ0 for zero step-size
and the value φ1 is large and hence the uncertainty is relatively large.
Noteworthy is the fact that based on the trends with the current grids, the estimations
(indicated by cfd) for X, Y and N for increasing numbers of grid nodes do not converge
to the experimental values (indicated by exp). This may be caused by either modelling
errors or by uncertainties in the experimental values.
A similar grid study was conducted for ReFRESCO. Using an automatic procedure
in GridPro, a series of geometrically similar grids was generated, using grid coarsening in
all three directions for each block in the grid. For each grid, the number of cells in the
grid ncells and the number of faces on the hull surface nhull are given in Table 5.8, which
includes also the y2+ values that were obtained during the calculations at drift angles of
0◦ and 30◦ and a non-dimensional rotation rate of 0.4. For each grid, the full range of
drift angles between 0◦ and 30◦ was calculated.
For drift angles of 10◦ and 30◦ , the uncertainty estimates are presented in Table 5.9
and Table 5.10 respectively. Based on an analysis of the results for each grid, it was
decided to use the four finest grids for the uncertainty analysis. The number of grids ng
5.3 – HTC 59

used was chosen based on the scatter in the results for the coarsest grid.
In Figure 5.7 and Figure 5.8 the forces and moments on the ship are graphically
presented for the different grids. The scatter in the results is relatively large and the
results are less consistent than those obtained with Parnassos. For a relative step size
of 1.6 or below, the results appear to converge. The observed convergence rate p, however,
is found to be large for all force components, hence the uncertainty is relatively large due
to the use of a larger safety factor Fs . Finer grids are required to reduce the uncertainty
in the calculations.

Table 5.8: Properties of grids for uncertainty analysis, HTC, 18 kn, ReFRESCO

id hi ncells nhull y2+ (β = 0◦ ) y2+ (β = 30◦ ) y2+ (γ = 0.4)

1 1.00 5234432 31872 0.27 0.52 0.42
2 1.27 2564378 19706 0.35 0.64 0.49
3 1.60 1288960 12540 0.46 0.81 0.60
4 1.99 659352 8040 0.59 0.85 0.77
5 2.58 303462 4698 0.78 1.23 0.99

Table 5.9: Uncertainty analysis, HTC, β = 10◦ , 18 kn, ReFRESCO

Item φ0 φ1 Uφ p
Xp - −3.15 × 10−3 44.0% 1

Xf −1.17 × 10−2 −1.14 × 10−2 4.0% 1.23

X - −1.45 × 10−2 6.6% 1

Yp - −3.73 × 10−2 32.5% 2

Yf −1.05 × 10−3 −1.02 × 10−3 8.9% 3.62

Y - −3.84 × 10−2 31.0% 2

Zp 8.45 × 10−2 8.65 × 10−2 6.5% 6.78

Zf - 2.58 × 10 −4 13.4% 2 1 Oscillatory convergence
Z 8.48 × 10−2 8.67 × 10−2 6.4% 6.73 2 Monotonic divergence
Kp - 1.60 × 10 −2 14.6% 2 3 Oscillatory divergence
Kf 1.54 × 10−3 1.50 × 10−3 7.6% 3.36
K - 1.75 × 10−2 11.6% 2

Mp - −1.37 × 10−3 39.4% 1

Mf - −2.74 × 10−4 498.4% 1

M −1.49 × 10−3 −1.64 × 10−3 27.2% 5.88

Np - −2.23 × 10−2 7.0% 2

Nf −7.58 × 10−5 −5.07 × 10−5 62.3% 0.64

N - −2.23 × 10−2 6.9% 2
60 Chapter 5 – Verification and validation of steady motion calculations

Table 5.10: Uncertainty analysis, HTC, β = 30◦ , 18 kn, ReFRESCO

Item φ0 φ1 Uφ p
Xp - 6.48 × 10−3 109.5% 1

Xf −1.22 × 10−2 −1.18 × 10−2 9.7% 2.92

X - −5.33 × 10−3 90.0% 1

Yp −2.27 × 10−1 −2.34 × 10−1 9.4% 3.62

Yf −3.67 × 10−3 −3.56 × 10−3 9.4% 3.16
Y −2.31 × 10−1 −2.38 × 10−1 9.1% 3.63
Zp - 2.38 × 10−1 4.7% 2

Zf - 1.13 × 10−3 8.1% 1 1 Oscillatory convergence

Z - 2.39 × 10−1 4.7% 2 2 Monotonic divergence
Kp 1.10 × 10−1 1.12 × 10−1 4.0% 6.56 3 Oscillatory divergence
Kf 4.81 × 10−3 4.67 × 10−3 9.1% 3.17
K 1.15 × 10−1 1.16 × 10−1 3.5% 7.07
Mp −7.86 × 10−3 −6.96 × 10−3 36.3% 6.52
Mf - −3.74 × 10−4 343.1% 1

M −8.02 × 10−3 −7.34 × 10−3 26.2% 5.25

Np −7.25 × 10−2 −7.45 × 10−2 7.6% 6.16
Nf −5.97 × 10−5 −3.77 × 10−5 179.5% 4.63
N −7.26 × 10−2 −7.45 × 10−2 7.5% 6.18

-0.0135 0.021
U=6.6% U=11.6%
-0.014 0.02
exp exp
-0.0145 0.019
-0.015
0.018
K
X

-0.0155
-0.016 0.017
-0.0165 0.016
-0.017 0.015
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size
-0.025 -0.0205
U=31.0% -0.021 U=6.9%
-0.03 exp exp
-0.0215
-0.035 -0.022
-0.04 -0.0225
N
Y

-0.045 -0.023
-0.0235
-0.05 -0.024
-0.055 -0.0245
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size

Figure 5.7: Uncertainty analysis, HTC, β = 10◦ , 18 kn, ReFRESCO

5.3 – HTC 61

0 0.124
-0.002 U=90.0% p=7.1
0.122
-0.004 exp U=3.5%
-0.006 0.12 exp
-0.008 0.118
K
X

-0.01 0.116
-0.012 0.114
-0.014
-0.016 0.112
-0.018 0.11
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size
-0.21 -0.068
-0.22 p=3.6 -0.07 p=6.2
-0.23 U=9.1% -0.072 U=7.5%
-0.24 exp -0.074 exp
-0.076
-0.25
N
Y

-0.078
-0.26 -0.08
-0.27 -0.082
-0.28 -0.084
-0.29 -0.086
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size

Figure 5.8: Uncertainty analysis, HTC, β = 30◦ , 18 kn, ReFRESCO

62 Chapter 5 – Verification and validation of steady motion calculations

5.3.3 Discretisation error for steady yaw motion

For a non-dimensional rotation rate of γ = −0.2, a series of geometrically similar grids has
been generated in order to investigate the discretisation error for Parnassos. The base
grids used are the same as those used for the uncertainty estimation for a steady drift angle
of 10◦ . The grid coarsening has been conducted in all three directions. Table 5.11 presents
the number of nodes and y2+ values for these grids obtained for the 10 kn condition.

Table 5.11: Properties of grids for uncertainty analysis, HTC, γ = −0.2, Parnassos

id γ nξ nη nζ hi Nodes y2+ Comment

3 -0.2 297 77 2×41 1.00 1875258 0.80
4 -0.2 257 65 2×35 1.18 1169350 0.98
6 -0.2 177 46 2×25 1.67 407100 1.48 based on grid 2, coarsened by 2×2×2
8 -0.2 129 33 2×18 2.35 153252 1.93 based on grid 4, coarsened by 2×2×2
9 -0.2 89 23 2×13 3.33 53222 2.78 based on grid 2, coarsened by 4×4×4
10 -0.2 73 19 2×11 4.00 30514 3.69 based on grid 1, coarsened by 5×5×5

The predicted values of the friction (index f ) and pressure (index p) components as
well as the total force and moment coefficients are presented in Table 5.12 with their
estimated uncertainties. Based on an analysis of the results for each grid, it was decided
to use the four finest grids (grids 3, 4, 6 and 8) for the uncertainty analysis.

Table 5.12: Uncertainty analysis, HTC, γ = −0.2, 10 kn, Parnassos

Item φ0 φ1 Uφ p
Xp −1.83 × 10−3 −2.07 ×10−3 25.8% 5.46
Xf - −1.17 × 10−2 3.3% 1

X −1.35 × 10−2 −1.38 × 10−2 5.1% 6.05

Yp - −5.76 × 10−3 4.2% 1

Yf −2.09 × 10−4 −2.21 × 10−4 29.5% 2.54

Y - −5.98 × 10−3 6.1% 1

Zp 6.14 × 10−2 6.04 × 10−2 2.2% 0.71

Zf - 1.62 × 10−4 4.8% 1 1 Oscillatory convergence
Z 6.16 × 10−2 6.06 × 10−2 2.2% 0.69 2 Monotonic divergence
Kp −2.31 × 10−3 −2.14 × 10−3 16.0% 5.00 3 Oscillatory divergence
Kf −1.77 × 10−4 −1.94 × 10−4 25.3% 2.68
K −2.49 × 10−3 −2.33 × 10−3 12.6% 5.60
Mp - −1.24 × 10−3 6.2% 1

Mf - 2.93 × 10−4 2.7% 1

M - −9.50 × 10−4 8.8% 1

Np - 7.34 × 10−3 10.7% 2

Nf −1.06 × 10−4 −1.11 × 10−4 9.4% 2.24

N - 7.22 × 10−3 10.9% 2

Similar to what was found for steady drift, the absolute uncertainty in the pressure
5.3 – HTC 63

components is larger than in the friction components. Compared to the calculations for
steady drift, the relative uncertainties for the rotational motion results are in most cases
smaller.
In Figure 5.9 the longitudinal force X, transverse force Y , heel moment K and yawing
moment N are given for the different grids. It is seen that the results do not differ much
between the individual results, but convergence is not always found due to scatter. For a
relative step size below 3, reasonably consistent results are however found.

-0.013 -0.001
p=6.1 -0.0012 p=5.6
-0.0135
U=5.1% -0.0014 U=12.6%
-0.014 exp -0.0016 exp
-0.0145 -0.0018

K
X

-0.015 -0.002
-0.0155 -0.0022
-0.0024
-0.016 -0.0026
-0.0165 -0.0028
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
relative step size relative step size
-0.002 0.009
U=6.1% 0.0085 U=10.9%
-0.003 exp exp
0.008
-0.004 0.0075
-0.005 0.007
N
Y

-0.006 0.0065
0.006
-0.007 0.0055
-0.008 0.005
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
relative step size relative step size

Figure 5.9: Uncertainty analysis, HTC, γ = −0.2, Parnassos

(Experimental values for 18 kn condition)

Using the grids summarised in Table 5.8, the uncertainties in the ReFRESCO results
for non-dimensional rotation rates of γ = 0.2 and γ = 0.4 were determined. Based on
an analysis of the results for each grid, it was decided to use the four finest grids for the
uncertainty analysis, which is presented in Table 5.13 and Figure 5.10 for γ = 0.2 and
Table 5.14 and Figure 5.11 for γ = 0.4.
The convergence with grid refinement for rotational motion is slightly better (more
realistic apparent orders of convergence) than the grid sensitivity for steady drift presented
in Table 5.9.

5.3.4 Local quantities

In both the Parnassos and ReFRESCO calculations for drift angles larger than 10◦ ,
or as well as for large rotation rates, an area of flow separation was found at the leeward
side of the bulbous bow, see Figure 5.12 for γ = 0.4. This area is probably caused by the
64 Chapter 5 – Verification and validation of steady motion calculations

Table 5.13: Uncertainty analysis, HTC, γ = 0.2, 18 kn, ReFRESCO

Item φ0 φ1 Uφ p
Xp - −2.39 × 10−3 30.1% 2

Xf −1.14 × 10−2 −1.10 × 10−2 4.2% 1.02

X −1.48 × 10−2 −1.34 × 10−2 12.7% 0.54
Yp 8.53 × 10−3 7.41 × 10−3 19.3% 1.61
Yf - 1.70 × 10−4 15.6% 2

Y 8.71 × 10−3 7.58 × 10−3 19.1% 1.59

Zp - 6.56 × 10−2 3.7% 1

Zf 1.29 × 10−4 1.31 × 10−4 2.5% 5.22 1 Oscillatory convergence

Z - 6.57 × 10−2 3.7% 1 2 Monotonic divergence
Kp 1.43 × 10−3 1.98 × 10−3 35.1% 1.31 3 Oscillatory divergence
Kf −2.29 × 10−4 −1.63 × 10−4 51.2% 0.74
K 1.22 × 10−3 1.82 × 10−3 41.7% 1.28
Mp - −8.67 × 10−4 76.5% 2

Mf −2.83 × 10−4 −2.74 × 10−4 3.8% 0.91

M - −1.14 × 10−3 56.2% 2

Np −5.96 × 10−3 −7.46 × 10−3 25.4% 0.85

Nf −1.17 × 10−4 −1.11 × 10−4 7.3% 0.70
N −6.07 × 10−3 −7.57 × 10−3 24.9% 0.85

-0.0115 0.004
-0.012 p=0.5 p=1.3
U=12.7%
0.0035 U=41.7%
-0.0125
-0.013 exp 0.003 exp

-0.0135 0.0025
K
X

-0.014 0.002
-0.0145
-0.015 0.0015
-0.0155 0.001
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size
0.01 -0.0055
0.009 p=1.6 -0.006 p=0.8
0.008 U=19.1% -0.0065 U=24.9%
0.007 exp -0.007 exp
-0.0075
0.006
N
Y

-0.008
0.005 -0.0085
0.004 -0.009
0.003 -0.0095
0.002 -0.01
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size

Figure 5.10: Uncertainty analysis, HTC, γ = 0.2, 18 kn, ReFRESCO

5.3 – HTC 65

Table 5.14: Uncertainty analysis, HTC, γ = 0.4, 18 kn, ReFRESCO

Item φ0 φ1 Uφ p
Xp −3.37 × 10−3 −3.16 × 10−3 20.5% 2.19
Xf −1.20 × 10−2 −1.16 × 10−2 5.1% 0.96
X −1.52 × 10−2 −1.47 × 10−2 4.3% 1.66
Yp 1.61 × 10−2 1.50 × 10−2 9.7% 2.02
Yf 3.39 × 10−4 3.52 × 10−4 4.9% 1.98
Y 1.64 × 10−2 1.53 × 10−2 9.3% 2.02
Zp - 7.30 × 10−2 7.2% 2

Zf 1.60 × 10−4 1.64 × 10−4 2.9% 2.01 1 Oscillatory convergence

Z - 7.32 × 10−2 7.2% 2 2 Monotonic divergence
Kp 3.33 × 10−3 3.55 × 10−3 8.0% 2.03 3 Oscillatory divergence
Kf - −3.75 × 10−4 8.2% 2

K 2.95 × 10−3 3.17 × 10−3 9.3% 2.01

Mp - −1.23 × 10−3 125.7% 2

Mf −2.96 × 10−4 −2.87 × 10−4 3.9% 0.81

M - −1.52 × 10−3 100.6% 2

Np −1.68 × 10−2 −1.82 × 10−2 9.3% 1.35

Nf - −2.15 × 10−4 3.7% 1

N −1.71 × 10−2 −1.84 × 10−2 9.1% 1.36

-0.012 0.0048
-0.0125 p=1.7 0.0046 p=2.0
-0.013 U=4.3% 0.0044 U=9.3%
-0.0135 exp 0.0042 exp
-0.014 0.004
-0.0145 0.0038
K
X

-0.015 0.0036
-0.0155 0.0034
-0.016 0.0032
-0.0165 0.003
-0.017 0.0028
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size
0.017 -0.016
0.016 p=2.0
-0.017 p=1.4
0.015 U=9.3% U=9.1%
0.014 exp
-0.018 exp
0.013 -0.019
0.012
N
Y

0.011 -0.02
0.01 -0.021
0.009
0.008 -0.022
0.007 -0.023
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size

Figure 5.11: Uncertainty analysis, HTC, γ = 0.4, 18 kn, ReFRESCO

66 Chapter 5 – Verification and validation of steady motion calculations

0.02

0.00

0.8
0.2
0.4 00
0..47.7
00

0.4
0.9
5
1 8
0.

.9

.6.06.
.9 0

2
.60.2
0.02

0.9
0.95
00.6

0.8

9
0.4

0.
1
0.4
0.8
0.2 00.7
z/Lpp

00.7.7
0.04

.70.95

00.8.42
00.4.8 1

0..66
0.9
0.06

0.02
0.9

0.
95
0.08

1
ReFRESCO fine, x=0.48Lpp, γ=0.4 Parnassos MNT norud, x=0.48Lpp, γ=0.4
0.
95
0.10
0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

y/Lpp
Figure 5.12: Comparison between ReFRESCO and Parnassos, HTC, x = 0.48Lpp , 18 kn
(solid lines: ReFRESCO, dotted lines: Parnassos)

neglect of the free water surface and may lead to errors in the prediction of the forces on
the hull.
The experimental data of the wake field in a plane 0.48Lpp aft of midship have kindly
been made available by HSVA2 . In the upper half of Figure 5.13 a comparison of the
non-dimensional axial velocity fields between the experiments (dotted lines) and the cal-
culations (solid lines) for a non-dimensional yaw rate γ of 0.4 is presented. The non-
dimensional axial velocity along a horizontal line in this plane at z = 0.05Lpp is given
in the lower half of the figure. This figure shows that in most parts of the plane, the
viscous-flow calculations correspond reasonably well with the experiments, especially at
port side (windward). At the leeward side, the wakes of three different vortices are found
in the Parnassos results: around y = 0.01Lpp the vortex generated at the stern, around
y = 0.04Lpp the vortex generated at the bow and at y = 0.07Lpp the vortex generated
at the starboard bilge. In the ReFRESCO results and in the experiments, only the
wakes of two vortices are seen. Based on an analysis of the flow along the length of the
ship calculated by ReFRESCO, it was observed that the bow vortex merges with the
vortex generated at the bilge. Apparently, this effect is not captured by Parnassos. The
difference between the experiments and the calculation is mainly caused by the relatively
coarse grid used for the Parnassos calculations.
Overall, it is concluded that the flow field at the propeller plane is predicted reasonably
well.

2
The actual immersion of the PIV probe, which served as reference for the position of the field of view
during the wake field measurements, was unknown, because the immersion was measured before the final
adjustment of the model condition when the model still had a considerable trim by the stern due to the
weight of the PIV equipment. Based on the apparent location of the propeller hub and the ship hull, a
vertical shift by about 0.011Lpp has been applied.
5.3 – HTC 67

0.02

0.00

0.
0.8

0.7
0.9
0.
95
1 0.6

0.
8
0.6 8

95
0.
7 0.

1
0.7

0.9
0.7
0.02

0.9 0.9
0.9

0.8

0.9
0.7
0.

0.8
95

0.6

0.9
1
0.6 1
z/Lpp

0.8

0.8
0.04 0.6

0.07.6
0.70.8

0.95
1
0.6

0.9
0.8 1
0.9
0.06 5

0.9
0.9

5
0.08 1

Parnassos MNT norud, x=-0.48Lpp, γ=0.4 Experiments_shifted, x=-0.48Lpp, γ=0.4

0.10
0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
0.02 y/Lpp

0.00
0.9

0.8
0.95
0.95

0.6
0.7 0.9 1 0.8

0.9

0.9
00..8
8 0.6
0.02

1
1

0.9
0.6 0.7

00.
0.8

.7 70.6
1

9
0.95
0.9

0.95

0.
z/Lpp

0.95
0.8
0.

0.04 6
1

0.
95

0.9
.4
0.9

00.6
1

0.

0.8 0.8
7

0.

1
0.9
0.06
1

0.08 0.
95

ReFRESCO fine, x=-0.48Lpp, γ=0.4 Experiments_shifted, x=-0.48Lpp, γ=0.4

0.10
0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
1.2
y/Lpp

1.0

0.8

0.6
u

0.4

0.2

0.0 Parnassos MNT norud

x=-0.48Lpp, z=-0.05Lpp, γ=0.4 Experiments_shifted
-0.2
0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
1.2
y/Lpp

1.0

0.8

0.6
u

0.4

0.2

0.0 ReFRESCO fine

x=-0.48Lpp, z=-0.05Lpp, γ=0.4 Experiments_shifted
-0.2
0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

y/Lpp

Figure 5.13: Comparison with experiments, HTC, x = −0.48Lpp , 18 kn (solid lines: calcula-
tions, dotted lines: experiments)
68 Chapter 5 – Verification and validation of steady motion calculations

5.3.5 Global quantities

In Figure 5.14 and Figure 5.15 the results for a speed of 18 kn are presented.
For steady drift, good agreement with the experiments is seen for the transverse force
Y , heel moment K and yaw moment N results based on the viscous flow calculations
using both Parnassos and ReFRESCO. Only the results for the longitudinal force X
deviate from the measured results. Within the VIRTUE project, a similar comparison
was made using results from several solvers [152]. There, it was seen that the results from
several solvers deviated from the measured X force, but the same solvers showed excellent
agreement at a lower speed corresponding to the 10 kn condition. In the calculations using
these solvers, the deformation of the free surface was neglected (double body assumption).
The discrepancy between the calculations and measurements presented in Figure 5.14 can
therefore be caused by the neglect of free surface deformation, and therefore speed effects,
or due to uncertainty in the measurements.
For rotational motion, the prediction of the yaw moment N is good, but deviations are
seen for the other forces and moment. The magnitudes of the Y force and heel moment K
during pure rotation are, however, very small and of less significance than the other forces
or moments. Furthermore, it can be expected that the uncertainty in the measurements
for the Y force and K moment is relatively high, since the hydrodynamic contribution in
these forces or moments is found by subtracting the relatively high centrifugal components
from the measured components.
Comparing the Parnassos and ReFRESCO calculations, some differences can be
seen. For steady drift, the differences are small and of the same order of magnitude as
the uncertainty in the predictions. In general, the Y force and N moment predicted by
Parnassos as a function of the drift angle β are slightly closer to the measurements than
the ReFRESCO results. This may be caused by the use of a different turbulence model.
For steady yaw, the differences are somewhat larger and for these calculations, the
ReFRESCO calculations are slightly closer to the measurements. This may be caused
by either a much finer grid for ReFRESCO (i.e. 5.2 million versus 1.9 million for Par-
nassos), or by the different turbulence model.
5.3 – HTC 69

−0.005 0.12
−0.005 Exp (HSVA)
Parnassos
0.1
ReFRESCO
−0.01
−0.01 0.08

X
K
X

0.06
−0.015

−0.015 0.04
−0.02
0 5 10 15 20 25 30
0.02 β

−0.02 0
0 10 20 30 0 10 20 30
β β
0 0

−0.05 −0.002

−0.1 −0.004
M
Y

−0.15 −0.006

−0.2 −0.008

−0.25 −0.01
0 10 20 30 0 10 20 30
β β
0.3 0

0.25
−0.02

0.2
N

−0.04
Z

0.15

−0.06
0.1

0.05 −0.08
0 10 20 30 0 10 20 30
β β

Figure 5.14: Comparison between experiments and calculations, HTC, 18 kn, steady drift
70 Chapter 5 – Verification and validation of steady motion calculations

−3
−0.012 x 10
6

5
−0.014
4
X

−0.016

K
3

2
−0.018
1

−0.02 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
γ γ
−3
0.025 x 10
0

0.02
−1
0.015
Y

−2
0.01

−3
−0.005 Exp (HSVA)
0.005 Parnassos
ReFRESCO
0 −0.01
−4
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
X

γ γ
−0.015
0.09 0

−0.02
−0.0050 5 10 15 20 25 30
0.08 β
−0.01

0.07
Z

−0.015

−0.02
0.06
−0.025

0.05 −0.03
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
γ γ

Figure 5.15: Comparison between experiments and calculations, HTC, 18 kn, steady yaw
5.3 – HTC 71

5.3.6 Validation
The uncertainties in the measurements of the forces and moments have not been specif-
ically determined. However, an estimate of the uncertainty UD can be made based on
results obtained for similar test conditions during measurements at the start of VIRTUE
and measurements conducted later in the project. Some tests were repeated and for
each test, the uncertainty was estimated using an arbitrary factor of safety of 1.25,
i.e. UD = 1.25 × abs (φtest1 − φtest2 ). The overall data uncertainty for the HTC ex-
periments is taken from the average of the data uncertainties obtained for several re-
peat tests. With these estimated data uncertainties and assuming that the simulation
numerical uncertainty USN is only influenced by the discretisation uncertainty UG (i.e.
USN = UI + UG = UG ), Table 5.15 and Table 5.16 can be constructed.
The Parnassos result for β = 10◦ shows that for the longitudinal and transverse
forces X and Y the comparison errors |E| are smaller than the validation uncertainties
Uval which means that the solution is validated at levels of 13.3% and 21.1% respectively.
These levels are judged to be relatively high and indicate that the numerical uncertainty
needs to be reduced if lower validation levels are desired. For the heel moment K and
the yawing moment N validation is not achieved (|E| > Uval ), which indicates modelling
errors. Changes in turbulence model or domain size or the inclusion of free surface may
lead to improvements of the comparison error.
The ReFRESCO result for Y for β = 10◦ is validated at a level of 27.2%, which is
high. The numerical uncertainty USN needs to be reduced to reduce the validation level.
The yaw moment N is validated at an acceptable level of 8.3%. The other components
are not validated and indicate modelling errors. For the higher drift angle β = 30◦ , Y and
N are validated at levels of 12.8% and 9.5% respectively, which are reasonable values.
For the steady rotation case, γ = 0.2 with Parnassos (the results have been ob-
tained by mirroring the γ = −0.2 results for comparison with the ReFRESCO results),
the comparison errors |E| are larger than the validation uncertainties for all forces and
moments. Since experimental results for the 10 kn condition with γ = 0.2 were not avail-
able, the experimental values for the 18 kn condition were used for comparison. This
mainly introduces an error in the longitudinal force X due to a difference in the frictional
resistance and therefore a large comparison error in X is to be expected. For the other
components, the influence of the different speed is expected to be of less influence and
therefore the validation should still provide insight into the accuracy of the calculations.
Even though a comparison between two speed conditions is made, it is concluded based
on the validation that the large values of the comparison errors indicate that modelling
errors are present in the simulation results.
For ReFRESCO, the comparison errors |E| for γ = 0.2 are smaller than for Par-
nassos. Together with the higher numerical uncertainties USN , this leads to validation
of X, Y , K and N at levels of 11.4%, 20.6%, 66.8% and 22.2%, respectively. These levels
are high and USN needs to be reduced to obtain lower validation levels.
The ReFRESCO results for γ = 0.4 for Y , K and N are validated at levels of 12.7%,
72 Chapter 5 – Verification and validation of steady motion calculations

10.5% and 10.5% respectively, which are judged to be reasonable. The comparison errors
|E| for all components are found to be small. For X, modelling errors appear to be
present, which might be improved by incorporating free surface deformation.

Table 5.15: Validation, HTC, 10 kn, Parnassos

Validation, β = 10◦ Validation, γ = 0.2

X Y K N X Y K N
D × 103 -16.6 -46.3 20.6 -22.9 -15.1 -7.8 -1.1 8.8
S × 103 -15.7 -44.2 17.9 -24.4 -12.5 -5.4 -2.1 7.3
E = S − D (%D) -5.8 -4.4 -13.2 6.6 -17.2 -31.1 82.4 -16.8
UD (%D) 1.4 9.1 4.3 5.3 1.4 9.1 4.3 5.3
USN (%S)
p 14.1 19.9 7.5 1.2 6.1 6.1 12.6 10.9
Uval = UD 2 + U2 (%D) 13.3 21.1 7.8 5.4 5.2 10.0 23.4 10.5
SN
Validated? ✔ ✔ ✘ ✘ ✘ ✘ ✘ ✘

Table 5.16: Validation, HTC, 18 kn, ReFRESCO

Validation, β = 10◦ Validation, γ = 0.2

X Y K N X Y K N
D × 103 -16.8 -46.3 20.9 -24.1 -15.1 -7.8 -1.1 8.8
S × 103 -14.5 -38.4 17.5 -22.3 -13.4 -7.6 -1.8 7.6
E = S − D (%D) -13.5 -17.2 -16.1 -7.3 -11.2 -3.4 59.8 -13.5
UD (%D) 1.4 9.1 4.3 5.3 1.4 9.1 4.3 5.3
USN (%S)
p 6.6 31.0 11.6 6.9 12.7 19.1 41.7 24.9
Uval = UD 2 + U2 (%D) 5.9 27.2 10.6 8.3 11.4 20.6 66.8 22.2
SN
Validated? ✘ ✔ ✘ ✔ ✔ ✔ ✔ ✔
Validation, β = 30◦ Validation, γ = 0.4
X Y K N X Y K N
D × 103 -17.5 -240.3 110.0 -70.4 -16.7 -16.1 -3.1 18.5
S × 103 -5.3 -238.0 116.4 -74.5 -14.7 -15.3 -3.2 18.4
E = S − D (%D) -69.6 -1.0 5.8 5.9 -11.7 -4.8 3.1 -0.6
UD (%D) 1.4 9.1 4.3 5.3 1.4 9.1 4.3 5.3
USN (%S)
p 90.0 9.1 3.5 7.5 4.3 9.3 9.3 9.1
Uval = UD 2 + U2 (%D) 27.4 12.8 5.6 9.5 4.0 12.7 10.5 10.5
SN
Validated? ✘ ✔ ✘ ✔ ✘ ✔ ✔ ✔
5.4 – DARPA SUBOFF 73

5.4 DARPA SUBOFF

The DARPA SUBOFF hull form has been described in section 2.5.2. In this thesis,
only the bare hull (AFF-1) configuration is considered. This section presents RANS
calculations for the bare hull and a comparison is made with the experimental validation
data.
In earlier work, calculations were made using Parnassos with Menter’s one-equation
turbulence model. First, an H-O grid topology was used and large comparison errors in the
longitudinal force X were found, see Toxopeus [147]. Subsequently, an alternative grid-
layout, i.e. an axi-symmetric grid, was used. It was found that this reduced the comparison
error considerably, but not completely. By using the axi-symmetric grid together with the
TNT version of the k − ω turbulence model, the comparison error became smaller than
the validation uncertainty. This work demonstrated the large influence of the grid layout
and the turbulence model.
Inspired by this earlier work, new calculations for the DARPA SUBOFF bare hull
form were made using ReFRESCO. Some of these calculations were compared to the
Parnassos results. This comparison was presented in Toxopeus and Vaz [151] and is not
repeated here for brevity. For all ReFRESCO calculations, use was made of Menter’s
SST version of the two-equation k − ω turbulence model. Some additional calculations
were performed with the one-equation Spalart-Allmaras (SA) turbulence model.
The governing equations were integrated down to the wall, i.e. no wall-functions are
used. The flow was solved using a steady approach. Two sets of calculations were con-
ducted: a first set for Re = 12 × 106 at β = 0◦ and β = 2◦ for comparison with the flow
field, pressure and friction measurements in the wind tunnel by Huang et al. [68], and a
second set for Re = 14 × 106 at a range of drift angles between 0◦ and 18◦ for comparison
with force measurements in the towing tank by Roddy [116].
It should be noted that for the SUBOFF results the coordinate system and non-di-
mensionalisation of the forces and moments differs from those adopted in the rest of this
thesis: all forces and moments are made non-dimensional using a reference area of L2pp
instead of Lpp T and the yaw moment is given relative to the centre of buoyancy, instead
of with respect to midship. Additionally, the coordinate x indicates the distance along
the ship length from the bow, positive aft.
An overview of all experimental and computational results of the forces and moments
on the SUBOFF AFF-1 configuration for captive conditions is given in the table on
page 159.

5.4.1 Iterative error

All calculations were run until the maximum non-dimensional residual of the pressure
resp,max (the so-called L∞ norm) between successive iterations had dropped well below
1 · 10−5 or when further iterative convergence was not obtained. The changes in the non-
dimensional integral quantities (forces and moments) were well below 1 × 10−7 . This is
74 Chapter 5 – Verification and validation of steady motion calculations

more than two to three orders of magnitude smaller than the discretisation uncertainty
and therefore the iterative uncertainty UI can be neglected, see Eça et al. [41].

5.4.2 Discretisation error

The results were all obtained on structured axi-symmetric grids with O-O topology. For
each grid, the variation in the number of grid nodes in the stream-wise, normal and
girth-wise (nξ , nη and nζ ) directions is given in Table 5.17.

Table 5.17: Properties of grids for uncertainty analysis, SUBOFF, β = 0◦ , 18◦

id β nξ nη nζ hi hi Nodes Comment
(β = 0 ) (β = 18◦ ) ×10−3
◦

1 0 275 119 129 1.00 - 4222

2 0,18 241 105 113 1.14 1.00 2859
3 0,18 201 87 93 1.39 1.22 1626
4 0,18 171 75 81 1.60 1.40 1039
5 0,18 138 60 65 2.00 1.75 538
6 0,18 121 53 57 2.29 2.00 366 based on grid 2, coarsened by 2×2×2
7 0,18 101 31 47 2.78 2.43 147 based on grid 3, coarsened by 2×2×2
8 0,18 86 38 41 3.20 2.80 134 based on grid 4, coarsened by 2×2×2
9 0,18 61 27 29 4.57 4.00 48 based on grid 2, coarsened by 4×4×4

For Re = 14 × 106 , the discretisation error has been investigated. In Table 5.18 and
Figure 5.16, the results for β = 0◦ are presented. The graphs show that scatter exists in
the data: the data points are not exactly aligned along the curve. Reasons for this might
be e.g. the non-evenly spaced cell nodes, the use of numerical limiters or lack of perfect
geometrical similarity between the grids.
For this high Reynolds number, i.e. when convection dominates, and when using an
unstructured-grid Quadratic Upwind Interpolation for Convective Kinematics (QUICK)
scheme for convective fluxes, it is expected that ReFRESCO will be second order ac-
curate [45]. The observed order of convergence p depends on the force component under
consideration. For the friction force, a value just below 1 is found, indicating that the
convergence with grid refinement follows a linear order of accuracy. For the other com-
ponents, a much higher order is found, which is most probably caused by scatter and
insufficiently fine grids.
In the previous study by Toxopeus and Vaz [151], the convergence appeared to be
better. However, comparing the old (FreSCo) results with the new (ReFRESCO)
results with a finer grid added, it is seen that now the four finest grids show a more
consistent trend than the four finest grids in the previous study. The present results are
therefore judged to be more reliable. The overall uncertainty U in X is 4.5% which is
judged to be small.
In Table 5.19 and Figure 5.17, the results for β = 18◦ are presented. In this case, the
apparent order of convergence p ranges from 0.74 for Xf to 6.07 for Xp . This indicates
5.4 – DARPA SUBOFF 75

Table 5.18: Uncertainty analysis, SUBOFF, β = 0◦

Item φ0 φ1 Uφ p 1
−4 −4 Oscillatory convergence
Xp −1.22 × 10 −1.29 × 10 12.2% 6.71 2
−3 −4 Monotonic divergence
Xf −1.00 × 10 −9.67 × 10 4.4% 0.97 3
−3 −3 Oscillatory divergence
X −1.11 × 10 −1.10 × 10 4.5% 5.05

-0.0001 -0.00084
-0.00015 p=6.7 -0.00086 p=1.0
-0.0002 U=12.2% -0.00088 U=4.4%
-0.00025 -0.0009
-0.0003 -0.00092
Xp

Xf
-0.00035 -0.00094
-0.0004 -0.00096
-0.00045 -0.00098
-0.0005 -0.001
-0.00055 -0.00102
0 1 2 3 4 5 6 0 1 2 3 4 5 6
relative step size relative step size
-0.0007
-0.0008 p=5.0
-0.0009 U=4.5%
exp
-0.001
-0.0011
X

-0.0012
-0.0013
-0.0014
-0.0015
0 1 2 3 4 5 6
relative step size

Figure 5.16: Uncertainty analysis, SUBOFF, β = 0◦

that a finer grid needs to be used to obtain a solution closer to the so-called asymptotic
range, where the order of convergence will be equal to or lower than the order of the
discretisation scheme.
For the components of the transverse force Y and yaw moment N the apparent orders
of convergence are between 0.62 and 1.21, which may indicate that for the transverse force
and yawing moment the grid density is closer to the asymptotic range. The uncertainty
in X is found to be relatively large. The large value is caused by the fact that for the
overall force X monotonic convergence was not obtained. However, the value is acceptable
from an engineering viewpoint. The uncertainty in the overall transverse force or yawing
moment is judged to be small.

5.4.3 Local quantities

Sung [139] has written instructions for the Submarine Hydrodynamics Working Group
(SHWG), see www.shwg.org, for the post-processing of the DARPA SUBOFF CFD cal-
culations in order to make consistent comparisons between results of different calculations
of different solvers, institutions, Reynolds numbers, grid topologies, etc. The figures pre-
76 Chapter 5 – Verification and validation of steady motion calculations

Table 5.19: Uncertainty analysis, SUBOFF, β = −18◦

Item φ0 φ1 Uφ p
Xp 2.70 × 10−4 2.41 × 10−4 31.7% 6.07
Xf −1.13 × 10 −3 −1.06 × 10−3 7.9% 0.74
X - −8.22 × 10−4 11.3% 2
1 Oscillatory convergence
Yp 5.21 × 10−3 5.36 × 10−3 3.6% 1.17 2 Monotonic divergence
Yf 3.26 × 10−4 3.01 × 10−4 10.3% 0.62 3 Oscillatory divergence
Y 5.53 × 10−3 5.66 × 10−3 3.0% 1.21
Np 3.45 × 10−3 3.40 × 10−3 1.9% 0.81
Nf - 1.75 × 10−5 8.2% 2

N 3.47 × 10−3 3.41 × 10−3 2.0% 0.79

-0.0007 0.0074
-0.0008 U=11.3% 0.0072 p=1.2
exp 0.007 U=3.0%
-0.0009 0.0068 exp
-0.001 0.0066
0.0064
X

-0.0011 0.0062
-0.0012 0.006
0.0058
-0.0013 0.0056
-0.0014 0.0054
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
relative step size relative step size
0.0036
p=0.8
0.0035
U=2.0%
0.0034 exp
0.0033
N

0.0032
0.0031
0.003
0.0029
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
relative step size

Figure 5.17: Uncertainty analysis, SUBOFF, β = −18◦

sented in this section were made according to these instructions. The experimental values
are obtained from flow field and pressure measurements conducted by Huang et al. [68].
These experiments were conducted at a Reynolds number of 12 × 106 . The first and sec-
ond cases defined in the instructions are comparisons of the pressure Cp and friction Cf
coefficients along the hull, see Figure 5.18.
These graphs show that the differences in pressure coefficient between the results are
negligible. A very small difference between the SST and SA results is found at the stern,
which explains the difference in the longitudinal pressure coefficients Xp . For the skin
friction coefficient, it is seen that the results with the SA model are in general slightly
closer to the experimental data than the SST results. The differences between the results
explain the differences in forces found in Table 5.20.
5.4 – DARPA SUBOFF 77

1.5
Exp
ReFRESCO−SST
1 ReFRESCO−SA
Cp

0.5

−0.5
0 0.2 0.4 0.6 0.8 1
x/L
oa
−3
x 10
7
Exp
6 ReFRESCO−SST
ReFRESCO−SA
5

4
Cf

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Loa

Figure 5.18: Pressure (top) and friction (bottom) coefficients along the hull, β = 0◦

The predicted distribution of the pressure coefficient is close to the experiments. The
trends in the predicted distribution of the friction coefficient correspond well to the trends
found in the experiments. Although some discrepancies at the bow and stern area are
found, it is concluded that the prediction of the pressure and skin friction coefficients is
good. It is noted that the discrepancies at the bow and stern were also present in the
calculations by Bull [19] and Yang and Löhner [166] and in all results submitted for a
collaborative CFD study within the SHWG [146].
The difference between the SST and SA results for the streamwise Vx and radial
velocities Vr at x = 0.978Loa in the aft part of the hull, see Figure 5.19 (top) is considered
to be negligible. Comparing the computed results with the experiments, it is observed
that the trends in the development of the boundary layer are very well predicted by both
solvers, but quantitative discrepancies are seen. Especially the magnitudes of the radial
velocities are different. It is seen that in the experiments the radial velocity changes sign
between (r − R0 )/Rmax = 2 and (r − R0 )/Rmax = 0.8, suggesting outward radial flow in
the far field. This may be caused by the use of an open-jet wind tunnel.
In this study, also the correlation between the measured and the predicted Reynolds
0 0
shear stresses is investigated by comparison of −VVx2Vr . Following the eddy-viscosity as-
0
78 Chapter 5 – Verification and validation of steady motion calculations

sumption, the Reynolds stresses are defined by:

− Vx0 Vr0 = 2 · νt · Sij (5.6)

with eddy viscosity νt and strain rate tensor Sij

!
1 ∂ui ∂uj
Sij = + . (5.7)
2 ∂xj ∂xi

In Figure 5.19 (bottom), the Reynolds shear stresses for the aft-most longitudinal sta-
tion are presented. It is observed that the curve representing the SST results corresponds
very well with the measurements. The results using the SA turbulence model are also
close to the measurements, but under-predict the peak of the distribution.

2
Exp
ReFRESCO−SST
1.5 ReFRESCO−SA
(r−R0)/Rmax

0.5
Vr Vx
0
−0.2 0 0.2 0.4 0.6 0.8 1
Velocity/V0
−3
x 10
1.5

1
−Vx’Vr’/V20

Exp
ReFRESCO−SST
ReFRESCO−SA
0.5

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
(r−R0)/Rmax

Figure 5.19: Velocities (top) and Reynolds stresses (bottom) at x = 0.978Loa , β = 0◦

Figure 5.20 presents comparisons of the pressure Cp and friction Cf coefficients along
the hull at the leeward plane of symmetry. Figure 5.21 shows the axial Vx , tangential Vα
and radial Vr velocities (top part of the figure) and Reynolds shear stress (lower part of
the figure), given for the leeward symmetry plane located at x = 0.978Loa . These graphs
show that the distribution of the pressure coefficient along the length of the ship and the
velocity distribution at the stern is quite well represented. The difference between the
5.4 – DARPA SUBOFF 79

SST and SA results is considered to be small. However, with the SA turbulence model,
the radial velocity Vr appears to be too negative compared to the SST results and the
results obtained using other solvers and turbulence models during the SHWG CFD study
[146]. The distribution of the Reynolds shear stress shows reasonable correspondence with
the measurements.

1.5
Exp
ReFRESCO−SST
1 ReFRESCO−SA
Cp

0.5

−0.5
0 0.2 0.4 0.6 0.8 1
x/Loa
−3
x 10
7

5 ReFRESCO−SST
4 ReFRESCO−SA
Cf

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Loa

Figure 5.20: Pressure (top) and friction (bottom) coefficients along the hull, β = 2◦ (leeward
meridian)

5.4.4 Global quantities

Experimental force measurement results are available for the straight-ahead condition and
for oblique motion and were published by Roddy [116]. The experiments were conducted
in the towing basin of the David Taylor Research Center. During the tests, the model was
supported by two struts. The speed in the experiments resulted in a Reynolds number
of 14 × 106 . For the straight-flight condition the experimental value of the longitudinal
force was found to be:

X = average(Xtest1 , Xtest2 ) = average(−1.061, −1.051) × 10−3

(5.8)
= −1.056 × 10−3
80 Chapter 5 – Verification and validation of steady motion calculations

2
Exp
ReFRESCO−SST
1.5 ReFRESCO−SA
(r−R0)/Rmax

0.5

Vr Vt Vx
0
−0.2 0 0.2 0.4 0.6 0.8 1
Velocity/V0

−3
x 10
1.5

1
−Vx’Vr’/V20

Exp
ReFRESCO−SST
ReFRESCO−SA
0.5

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
(r−R0)/Rmax

Figure 5.21: Velocities (top) and Reynolds stresses (bottom) at x = 0.978Loa , β = 2◦ (leeward
meridian)

The longitudinal force components obtained from the calculations for β = 0◦ are given
in Table 5.20. As can be expected for submarine hull forms, the largest part (about 90%)
of the total resistance is caused by friction. This means that for the bare hull, the form
factor is relatively low, i.e. (1 + k) = X/Xf = 1.13 for SST and 1.07 for SA. From the
experiments, the form factor is estimated to be (1 + k) = X/Xf (ITTC) = 1.13. This is a
normal value for a bare hull submarine.
The comparison error E between the ReFRESCO prediction of X and the mea-
surement is about 3.7%, which is judged to be good for practical applications when also
the uncertainty in the experimental data is taken into consideration. It is found that
the total resistance predicted using the SST turbulence model is slightly higher than the
experimental value, while the SA results are slightly lower. The skin friction coefficient
predicted using SA is lower than the coefficient found using SST, as can also be observed
in Figure 5.18. In the aft ship, the pressure coefficient predicted using SA is marginally
higher than the pressure predicted with the SST model. This explains the lower pressure
resistance found in the SA results.
Figure 5.22 presents the force and moment components obtained from the calculations
and the values from the experiments for oblique inflow. In Tables 5.21 through 5.23 the
5.4 – DARPA SUBOFF 81

Table 5.20: SUBOFF, longitudinal force X, Re = 1.4 × 107 , β = 0◦

Solver Grid Integral values ×103

X Xf Xp E (%D)
Exp (DTRC) - -1.061 - - -
Exp (DTRC) - -1.051 - - -
Mean µexp -1.056 - - -
ITTC-57 - -0.936 - -
Schoenherr - -0.919 - -
Katsui - -0.905 - -
Grigson - -0.932 - -
ReFRESCO-SST 4138×103 -1.096 -0.967 -0.129 3.7
ReFRESCO-SA 4138×103 -1.017 -0.950 -0.067 -3.7

results for β = 18◦ are shown.

The comparison error E is about 13%, which is within the uncertainty band of the
measurements. With the SA turbulence model, better agreement is found. The trends in
the transverse force Y and yaw moment N are predicted reasonably well. The deviation
from the measurements may be caused by the modelling error of using a steady RANS
approach, which with increasing inflow angle may be disputable. Furthermore, with the
O-O grid topology the grid density away from the hull reduces considerably, such that
the wake may be insufficiently resolved.
The comparison shows that the turbulence model plays an important role in the
prediction of the forces on the SUBOFF. This was also found during a study in which
Parnassos was used with two different turbulence models: the SST turbulence model
and the Menter one-equation model (MNT) [147]. It was found that by changing the
turbulence model from MNT to SST, the comparison error in X between the experiments
and the simulations reduced from 40.8% to 6.8%. Interestingly, the influence is mostly
visible in X: the other forces or moments only change marginally when different tur-
bulence models are used. This means that either more advanced turbulence models are
required or that other factors such as grid layout or resolution of the grid in the wake of
the submarine affect the results.

Table 5.21: SUBOFF, longitudinal force X, Re = 1.4 × 107 , β = 18◦

Solver Grid Integral values ×103

X Xf Xp E (%D)
Exp (DTRC) - -0.670 - - -
Exp (DTRC) - -0.852 - - -
Mean µexp -0.761 - - -
3
ReFRESCO-SST 4138×10 -0.860 -1.067 0.207 13.0
3
ReFRESCO-SA 4138×10 -0.767 -1.074 0.307 0.8
82 Chapter 5 – Verification and validation of steady motion calculations

Table 5.22: SUBOFF, transverse force Y , Re = 1.4 × 107 , β = 18◦

Solver Grid Integral values ×103

Y Yf Yp E (%D)
Exp (DTRC) - -7.355 - - -
Exp (DTRC) - -7.438 - - -
Mean µexp -7.397 - - -
3
ReFRESCO-SST 4138×10 -5.383 -0.297 -5.086 -27.2
3
ReFRESCO-SA 4138×10 -5.678 -0.307 -5.371 -23.2

Table 5.23: SUBOFF, yawing moment N , Re = 1.4 × 107 , β = 18◦

Solver Grid Integral values ×103

N Nf Np E (%D)
Exp (DTRC) - -2.986 - - -
Exp (DTRC) - -2.939 - - -
Mean µexp -2.962 - - -
ReFRESCO-SST 4138×103 -3.370 -0.017 -3.353 13.8
ReFRESCO-SA 4138×103 -3.383 -0.017 -3.366 14.2

5.4.5 Validation
The uncertainties in the measurements of the forces and moments have not been deter-
mined. However, validation of the solution can still be performed, when a data uncertainty
UD is assumed. To obtain an estimate of the uncertainty in the experimental data, the
uncertainty UD in the experimental data is estimated using the difference between two
measurements for the same condition and a factor of safety of 1.25. For example, the
uncertainty in the longitudinal force X for β = 0◦ is estimated by:

UD = 1.25 × abs (Xtest1 − Xtest2 ) = 1.25 × 10−5 = 1.2% × X (5.9)

For other incidence angles, the same procedure can be applied. With this estimated un-
certainty and assuming that the simulation numerical uncertainty USN is only influenced
by the discretisation uncertainty UG (i.e. USN = UI + UG = UG ), Table 5.24 can be con-
structed. For β = 18◦ , it is seen that for the longitudinal force X the comparison error
|E| is smaller than the validation uncertainty Uval which means that the solution of X is
validated at a level of 27.2%. This level is judged to be high and indicates that especially
the experimental uncertainty needs to be reduced if lower validation levels are desired.
For all other forces and moments validation is not achieved (|E| > Uval ), which in-
dicates modelling errors. The magnitude of the Y force is under-predicted, while the
magnitude of the N moment is over-predicted. Changes in turbulence model or in the
domain size but also using higher grid densities in the wake or time accurate solution
procedures may lead to improvements of the comparison error.
5.4 – DARPA SUBOFF 83

Table 5.24: Validation, SUBOFF

Validation, β = 0◦ Validation, β = 18◦

X X Y N
D× 103 -1.06 D× 103 -0.76 -7.40 -2.96
S × 103 -1.10 S × 103 -0.86 -5.38 -3.37
E = S − D (%D) 3.7 E = S − D (%D) 13.0 -27.2 13.8
UD (%D) 1.2 UD (%D) 24.0 1.1 1.6
USN (%S)
q
4.5 USN (%S)
q
11.3 3.0 2.0
Uval = UD 2 + U 2 (%D) 4.8 Uval = UD 2 + U 2 (%D) 27.2 2.5 2.8
SN SN
Validated? ✔ Validated? ✔ ✘ ✘
84 Chapter 5 – Verification and validation of steady motion calculations

−3
x 10
1.5
Exp
1 ReFRESCO−SST

0.5
X

−0.5

−1

−1.5
0 5 10 15 20 25 30 35 40
β
0.005

−0.005
Y

−0.01

−0.015 Exp
ReFRESCO−SST
−0.02

−0.025
0 5 10 15 20 25 30 35 40
β
−3
x 10
1

−1 Exp
−2 ReFRESCO−SST
N

−3

−4

−5

−6
0 5 10 15 20 25 30 35 40
β

Figure 5.22: Comparison between experiments and calculations, SUBOFF, steady drift
5.5 – Walrus 85

5.5 Walrus
In work conducted for the Royal Netherlands Navy (RNLN) into the influence of the
seafloor on the manoeuvrability of submarines, validation studies were performed by Bet-
tle under the supervision of Toxopeus [15]. The flow around the Walrus bare hull form,
with deck and sail was computed with ReFRESCO. Other appendages were not included
in the study. The SST turbulence model was used and all equations were integrated down
to the wall (y2+ values were below 1). Calculations have been done for a range of clear-
ances c between the sea bottom and the submarine. In this thesis, only the deep water
calculations are considered. For the steady drift calculations, the drift sweep procedure
as introduced in section 4.6.2 was used.
It should be noted that for the Walrus results the coordinate system and non-di-
mensionalisation of the forces and moments differ from those adopted in the rest of this
thesis: all forces and moments are made non-dimensional using a reference area of L2pp
instead of Lpp T . The origin of the right-handed coordinate system is located at the
intersection of the longitudinal axis of symmetry of the hull, midship and centre-plane.
The experiments for the Walrus-class submarine were conducted by the David W.
Taylor Naval Ship Research and Development Center (DTNSRDC) [34]. Experimental
results were obtained for the three configurations of the early Walrus design listed in
Table 5.25.
Table 5.25: Designations and descriptions of Walrus experimental configurations [34]

] Hull with deck Bridge fairwater (sail), X-tail,

sailplanes, sonar dome propeller
1 ✔ ✔ ✔
2 ✔ ✔
3 ✔

It should be noted that the design of the Walrus as used for the DTNSRDC model
tests differs slightly from the real Walrus class submarine design, which was used for the
present study. The largest differences compared to the real design are a slightly smaller
length (0.9%Loa ) and the absence of the Toekan (exhaust diffuser). It is expected that
this discrepancy will have only a small effect on the overall forces and moments.
The calculations were performed at Re = 5.2 × 106 (to resemble the condition of
the free sailing experiments) whereas the DTNSRDC experiments were conducted at two
higher Reynolds numbers: 9 million and 14 million. The main effect of Reynolds number
in this range is to reduce the viscous drag. In order to better compare the results, the
axial force evaluated in the calculations at zero drift angle was scaled to Re = 14 million
using the ITTC 1957 friction line:
Cf,ITTC,Re=14×106
XRe=14×106 = Xp + Xf,Re=5.2×106 · (5.10)
Cf,ITTC,Re=5.2×106
This was done for the overall force on the hull and sail as well as on the hull surface alone.
86 Chapter 5 – Verification and validation of steady motion calculations

5.5.1 Iterative error

2 2
u u
0 v 0 v
w w
-2
p -2
p

log Linf residual

log L2 residual

-4 -4

-6 -6

-8 -8
c = 11.9 B c = 11.9 B
-10 α = 0 , β = 0 -10 α = 0 , β = 0
o o o o

6 6
1.9x10 Cell 1.9x10 Cells
-12 -12
0 2000 4000 6000 8000 0 2000 4000 6000 8000
Iterations Iterations

Figure 5.23: Iterative convergence, Walrus, β = 0◦ , α = 0◦

Excellent convergence was achieved, with L∞ and L2 norms of the residuals dropping
below 10−5 and 10−7 , respectively, in all cases of β = 0◦ . The integrated forces and
moments were unchanging to 7 significant digits (the precision with which this data was
written to the results file) for the last several hundred iterations. Figure 5.23 shows the
convergence histories of a selected calculation with a drift angle of zero.

ReFRESCO-SST, Re = 5.2 × 106 , β from −20◦ to 20◦

10+0
∆Y
10−2
10−4 -20◦ -15◦ -10◦ -5◦ 0◦ 5◦ 10◦ 15◦ 20◦
10−6
10−8
10−10
10−12
10−14
0 5000 10000 15000 20000 25000 30000 35000 40000
iteration number

Figure 5.24: Convergence history Y -force, Walrus, ReFRESCO

For non-zero drift angles, the L2 residuals dropped at least four orders of magnitude.
The effect on the integral quantities is small: the non-dimensional changes in the forces
and moments were well below 1×10−8 . This is more than two to three orders of magnitude
smaller than the discretisation uncertainty and therefore the iterative uncertainty UI can
5.5 – Walrus 87

be neglected, see Eça et al. [41]. An example of the convergence history of the non-
dimensional Y force for the drift-sweep calculation is shown in Figure 5.24. Each spike in
the line indicates the beginning of a new drift angle.

5.5.2 Discretisation error

The results of the discretisation error analysis are presented in Table 5.26 and Figure 5.25.
For this analysis, the results for the coarsest grid were taken as an outlier and the least-
squares method was applied to the four finest grids only. It is seen that the uncertainties
in X and M are small. The uncertainty in the vertical force Z (due to the asymmetry of
the hull with deck and sail) is found to be large. This is mainly caused by scatter in the
results, which can be attributed to the fact that it is difficult to manually generate grids
with ICEM CFD which are exactly geometrically similar. Furthermore, the magnitude of
this out-of-plane force is small compared to the vertical forces experienced when sailing
close to the sea floor or when sailing at a pitch angle.

Table 5.26: Uncertainty analysis, Walrus, β = 0◦ , α = 0◦ , deep water

Item φ0 φ1 Uφ p
Xp −2.09 × 10−4 −2.16 ×10−4 11.2% 7.56
Xf −1.32 × 10−3 −1.31 × 10−3 1.0% 1.98
X - −1.53 × 10−3 1.7% 2
1
−5 Oscillatory convergence
Zp −2.62 × 10 −2.72 × 10−5 9.9% 0.49 2 Monotonic divergence
Zf - −4.86 × 10−6 38.5% 1
3 Oscillatory divergence
Z −3.04 × 10−5 −3.20 × 10−5 10.2% 0.69
Mp −5.47 × 10−5 −5.44 × 10−5 2.4% 5.45
Mf 2.39 × 10−5 2.38 × 10−5 2.1% 3.22
M −3.08 × 10−5 −3.06 × 10−5 2.6% 6.83

5.5.3 Global quantities

Table 5.27 compares the scaled CFD results with experimental values (Re = 14 million)
for the hull-sail (configuration 2) and hull only (configuration 3) configurations. The
computations predict the resistance on the hull to within 3% of the experimental data.
This is expected to be within experimental error bounds based on the scatter in the
experimental data. The total resistance of hull and sail was predicted to be 9% lower
than configuration 2. It was expected that the calculated resistance would be lower than
the experimental value because the computations do not account for the drag on the
sailplanes, which is present in the experimental results for configuration 2.
The unrestricted-water calculations are also compared with configuration 2 experi-
ments for a range of drift angles in Figure 5.26. The agreement is considered to be very
good for all components of forces and moments, with the exception of pitching moment
M . The pitching moment in the calculations follows the same trend as the experiments
88 Chapter 5 – Verification and validation of steady motion calculations

-0.00019 -0.00124
-0.0002 p=7.6 -0.00125 p=2.0
-0.00021 U=11.2% -0.00126 U=1.0%
-0.00022 -0.00127
-0.00128
Xp

Xf
-0.00023
-0.00129
-0.00024 -0.0013
-0.00025 -0.00131
-0.00026 -0.00132
-0.00027 -0.00133
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
relative step size relative step size
-0.00149
U=1.7%
-0.0015
-0.00151
-0.00152
X

-0.00153
-0.00154
-0.00155
-0.00156
0 0.5 1 1.5 2 2.5 3
relative step size

Figure 5.25: Uncertainty analysis, Walrus, β = 0◦ , α = 0◦ , deep water

but is shifted down to smaller values. It is possible that this discrepancy is a result of
the sailplanes being present in the experiments but not in the calculations. The drag

Table 5.27: Walrus grid refinement study (values ×103 )

Total forces and moments

Cells Xp Xf X Form factor X(Re = 14M ) Xexp E = X − Xexp
1+k Config 2 (%Xexp )
470 -0.266 -1.275 -1.546 1.209 -1.295 -8.0
954 -0.236 -1.282 -1.523 1.184 -1.275 -9.4
1909 -0.220 -1.297 -1.521 1.170 -1.274 -1.407 -9.4
3761 -0.221 -1.302 -1.527 1.170 -1.279 -9.1
7509 -0.216 -1.309 -1.530 1.165 -1.281 -8.9

Forces and moments on hull only

Cells Xp Xf X Form factor X(Re = 14M ) Xexp E = X − Xexp
1+k Config 3 (%Xexp )
470 -0.214 -1.101 -1.314 1.194 -1.104 2.4
954 -0.199 -1.107 -1.306 1.179 -1.097 1.8
1909 -0.190 -1.120 -1.311 1.170 -1.101 -1.078 2.2
3761 -0.187 -1.125 -1.312 1.166 -1.102 2.2
7509 -0.189 -1.131 -1.320 1.167 -1.109 2.9
−3
Cf,ITTC,Re=14×106 = 2.832 × 10 (non-dimensional with wetted surface)
Cf,ITTC,Re=5.2×106 = 3.372 × 10−3 (non-dimensional with wetted surface)
5.5 – Walrus 89

and lift on the sailplanes would tend to increase M , consistent with the shift observed
in Figure 5.26. It also appears from the scatter in the data and the differences between
positive and negative drift angles that there is more uncertainty in M (relative to the
scale used for the plot) than in the other integral quantities.
The comparison with experiments showed that CFD gives accurate predictions for
the forces and moments on the submarine for the case of unrestricted water over a range
of drift angles.
90 Chapter 5 – Verification and validation of steady motion calculations

2
CFD, 1x106 Cells, Re = 5.2x106
CFD, 4x106 Cells, Re = 5.2x106 x
CFD, Extrap. to Re=14x106 20
1 x Expt, Re = 9x106 x
6
+ Expt, Re = 14x10
x +
x
+
+
103 X’

103 Y’
++
++
0 x 0 ++
+
+
x +
x
+
x
x
+
-1 + + x

++ ++++ + + x -20
+ + x
+

-2
-20 -10 0 10 20 -20 -10 0 10 20
β (deg) β (deg)
12 1.5

10 1 x

x
8 x
x 0.5 +
x
+
6 +
103 K’
103 Z’

++
++
x
x
0 ++
++
4 +
x
+
+
x -0.5
2 +
x
x
+
+
x
0
++
+++ + -1
+++
++

-2 -1.5
-20 -10 0 10 20 -20 -10 0 10 20
β (deg) β (deg)
1.2 10

1 x x

x 5 x
0.8 +
x
x +
+ +
0.6 +
103 M’

103 N’

x +
x ++
+ 0 +
+
+
0.4 + + +
x + +
+ +
+ +
x
+++++
0.2 ++
+ x
-5
x
0

-0.2 -10
-20 -10 0 10 20 -20 -10 0 10 20
β (deg) β (deg)

Figure 5.26: Comparison between configuration 2 experiments [34] and calculations, Walrus
(deep water, 4 × 106 cells)
5.6 – Conclusion 91

5.6 Conclusion
In this chapter, verification and validation of the predicted forces and moments on several
different ship hulls in manoeuvring motions have been presented using available validation
data from literature. It was demonstrated that for a wide range of ship types, good
predictions of the loads on the hull in manoeuvring motion can be obtained. The trends
in the forces and moments as a function of the drift angle or yaw rate are simulated well.
The results obtained with two different solvers and using different turbulence models were
compared. In general, it is concluded that the differences between the results obtained
with the two solvers are relatively small and within the numerical uncertainties, except
for cases where large differences in grid density were used. In those cases, better results
were obviously obtained on the finer grids. In general, it appears that the combination
of ReFRESCO with SST provides results that are slightly closer to the experimental
values than the combination of Parnassos with MNT. Considering the computational
effort, some differences exist between the two solvers. Parnassos is very fast when a
converged solution can be obtained and needs only two processors for the cases presented
in this section. For ReFRESCO a large number of processors is required to solve the
flow in a similar time-frame as Parnassos. However, for more complicated flows, such
as found at large drift angles or rotation rates, it becomes more difficult to converge a
solution with Parnassos, since this code is optimised for ships sailing in straight-ahead
conditions. The combination of ReFRESCO as a general code and the possibility to use
high-quality multi-block structured grids is more robust for these conditions and requires
much less manual interaction to arrive at a converged solution.
The verification studies provide useful insight into the influence of the grid density
on the predicted forces and moments. Summarizing, the numerical uncertainties in the
forces and moments obtained by the viscous-flow calculations were found to be about 10%
to 15% on average. The uncertainties in Y and K appear to be somewhat higher than in
X or N . In several cases, validation of the calculations failed, indicating modelling errors
in the numerical results. In these cases, it was generally seen that the magnitude of the
transverse force was under-predicted, while the magnitude of the yaw moment was over-
predicted. For manoeuvring studies in the early design, the comparison errors are within
acceptable levels. However, improvements remain desired and might be obtained using
finer grids, larger domain sizes, different grid topologies with refinement in the wake of
the ship, other turbulence models or incorporating free surface deformation. Furthermore,
unsteady phenomena in the flow have been ignored. In future studies, the influence of
instationary flow on the forces and moments needs to be investigated.
For one of the hull forms presented in this chapter, the DARPA SUBOFF, the influence
of a variation of the turbulence model was studied. By changing the turbulence model
from MNT to the more complex SST model, a considerable reduction of the comparison
error in X was found. Unfortunately, the changes in the other forces and moments were
small and therefore this did not lead to the desired overall improvements. Further study
is therefore required to determine the cause of the modelling error.
92 Chapter 5 – Verification and validation of steady motion calculations

The drift-sweep procedure as proposed in section 4.6.2 and implemented in Re-

FRESCO was used and the verification and validation studies shows that the procedure
provides good results in comparison with the Parnassos calculations and with the ex-
periments. With this procedure, the amount of manual interaction for the user decreases
considerably. Furthermore, the convergence for the different drift angles is faster than
when each drift angle would be calculated separately.
In the following chapter, free sailing model tests for the HTC will be presented, which
will be used for validation of the manoeuvring predictions presented in chapter 7.
 93

Chapter 6

Free sailing manoeuvring tests

6.1 Introduction
Unfortunately, results of free sailing manoeuvres for the HTC (see 2.5.4) were not available
at the end of the VIRTUE project and validation of the predicted manoeuvres could not be
performed. Therefore, MARIN decided to perform such manoeuvres outside of the scope
of VIRTUE. The results of this manoeuvring test programme can be used for public
domain comparisons of simulation results and for development of procedural guidelines
for free running model tests when it concerns manoeuvring of single propeller ships and
engine control during manoeuvres.
The purpose of the manoeuvring tests was to determine the yaw checking and course
changing abilities and the turning ability of the ship and provide data for validation of
manoeuvring predictions. To determine the manoeuvring characteristics, standard zig-
zag and combined turning circle/pull-out experiments were conducted with the following
variations:

• with and without bilge keels

• design speed and lower speed

• different procedures for the rudder angle application during zig-zag manoeuvres with
respect to the neutral rudder angle

• propeller rate control to simulate the engine behaviour during manoeuvring

This chapter presents details of the model, the experimental facility and equipment,
the data reduction procedures and the test programme. In this thesis, only the tests
without bilge keels and with constant propeller RPM are considered. More details about
the tests and drawings of the propeller and bilge keels can be found in MARIN Report
No. 23277-3-SMB [105].
94 Chapter 6 – Free sailing manoeuvring tests

6.2 Ship model

Details of the HTC are already given in section 2.5.4. The main particulars of the hull
and a small scale body plan are given in Table 2.4 and Figure 2.7. For the MARIN free
sailing manoeuvres, a wooden ship model was built to a scale ratio λ of 1:30.02. The
model was designated Ship Model No. 8971 and was tested with and without bilge keels.
Bow thruster openings were not modelled. Turbulence on the model was stimulated using
studs at the bow and sand strips at the appendages. All experiments described in this
thesis were carried out for a loading condition corresponding to a draught of 10.3 m on
even keel. The metacentric height GM of the model was adjusted to the full-scale value
of 1.09 m, which is relatively large for a vessel of this size and type. This results in small
heel angles during the manoeuvres such that the influence of heel on the manoeuvrability
can be neglected. A photograph of the model is given in Figure 6.1.

Figure 6.1: HTC ship model for free sailing tests

6.3 Propulsion and steering system

The ship model was fitted with a single propeller - single rudder arrangement. The rate
of rudder application during the tests corresponded to the full-scale value of 4.6◦ /s. The
rudder arrangement consists of a spade type rudder with headbox with a total lateral area
of approximately 39.76 m2 , or 2.51%Lpp T and a movable area of 29.03 m2 , or 1.83%Lpp T .
The rudder is positioned at the Aft Perpendicular Plane (APP). MARIN stock propeller
No. 5286 R was used to propel the ship model during the manoeuvring tests. The direction
of rotation of the propeller was clockwise when looking ahead. The propeller and rudder
properties are presented in Table 2.5, Table 2.6 and Figure 2.9. The longitudinal and
transverse forces on the rudder and the rudder stock moment have been measured using
a three-component force transducer. The propeller thrust and torque have also been
measured, using a two-component transducer in the shaft, just ahead of the propeller.
6.4 – Experimental facility and measurement system 95

Tests have been performed with a constant RPM setting and with modelling of a
concise engine control. In the latter case, a maximum power level that could be delivered
to the propeller corresponding to 100% Maximum Continuous Rating (MCR) at full scale
was assumed. When the power absorbed by the propeller exceeds this value, the propeller
rate of revolutions is decreased, simulating full scale engine behaviour. This method has
been validated against full scale feedback and has been found to provide better agreement
between model tests and full scale trials. The power level absorbed by the propeller
was measured during a speed run corresponding to 18.0 kn at full scale. This power was
assumed to be equivalent to 85% MCR.

6.4 Experimental facility and measurement system

The free sailing model tests were performed in MARIN’s Seakeeping and Manoeuvring
Basin (SMB) [109], which measures 170 m × 40 m, see Figure 6.2. The water depth in the
basin is 5 m. The model is followed along the basin length by a main carriage spanning the
width of the basin. A sub-carriage travels along the main carriage in transverse direction.
The carriage can follow all movements of the model in the horizontal plane. With an extra
mountable turntable, the system has a CPMC capability which includes the possibility
to mimic rotating arm tests.

Figure 6.2: Overview of Seakeeping and Manoeuvring Basin

96 Chapter 6 – Free sailing manoeuvring tests

At two adjacent sides of the basin, segmented wave generators consisting of hinged
flaps are installed. Each flap is controlled separately by a driving motor and has a width
of 60 cm. This set-up makes it possible to generate waves in any direction. The waves can
be long and short crested and multi-directional. The wave generator system is equipped
with an active wave reflection compensation feature and higher order wave synthesis
techniques. Opposite the wave generators, passive sinkable wave absorbers are installed.
A Krypton contact-less optical measurement system (now part of Nikon, see e.g.
www.nikonmetrology.com/optical cmm) is used to determine the position of the model
in six DOF. A target consisting of several infra-red Light-Emitting Diodes (LEDs) glued
in fixed positions on a non-deformable plate is mounted on the model on a location such
that the target is in the Krypton observation area and the location and orientation of the
target can be determined. By specifying in Krypton the position of the target relative
to the centre of gravity, the system calculates the position of the centre of gravity of the
model relative to the sub-carriage based on the measured position and orientation of the
target. The Krypton camera is mounted on the subcarriage and consequently moves with
the carriage. To obtain the x and y position of the vessel in the basin, the position of the
sub-carriage and the relative distance measured with Krypton are combined. All motions
are defined in the basin-fixed system of axes, except roll and pitch which are defined in
the ship-fixed system of axes.
The rudder angle and propeller RPM are actively controlled by the steering system.
The data acquisition consists of recording analogue and digital signals. The analogue
signals (e.g. propeller and rudder forces) are sampled before being recorded by the mea-
surement system, while digital signals (e.g. steering system and Krypton output) are
recorded directly.

6.5 Data reduction

All results of the model tests are presented as prototype (i.e. full scale) values in the tables
and figures in this thesis by applying Froude’s law of similitude to the measured data and
the assumption of salt water on full scale.
The raw data obtained from the measurements has been filtered before presentation
and before parameters were derived. Furthermore, sign conventions according to those
presented in section 3.1 have been applied.

6.6 Test procedures and programme

The manoeuvring tests were performed at the self propulsion point of the model. Correc-
tions for scale effects were not made, since some of the scale effects (such as the relatively
higher resistance or the relatively higher wake fractions at model scale) tend to even
out and because of the lack of worldwide consensus on how to correct for scale effects,
6.6 – Test procedures and programme 97

see ITTC Recommended Procedures and Guidelines on free sailing manoeuvring tests
[73]. Before commencing the manoeuvring tests, the relation between the propeller RPM
and the achieved speed was determined. During these tests, the model was steered on a
straight course with an autopilot. From the recordings, the average rudder angle required
for straight-ahead sailing was obtained. These angles were adopted as the neutral rud-
der angle to compensate the propeller wheel effect for that specific speed. Based on the
RPM-speed relationship, the propeller RPMs to sail at speeds corresponding to 10 kn and
18 kn were derived. These RPMs were used during the remainder of the test programme.
Standard zig-zag and combined turning circle/pull-out manoeuvres have been con-
ducted (see e.g. Bertram [13]). A photograph of the ship model during one of the tests is
shown in Figure 6.3. For the zig-zag tests, the rudder execute angle was given relative to
the neutral rudder angle. For a neutral angle of e.g. 1◦ to starboard, the actual mechanical
steering angles were 11◦ to starboard and 9◦ to port-side for a 10◦ /10◦ zig-zag. In the test
results, the presented rudder angle is however compensated for the neutral angle and will
therefore show rudder angles between 10◦ PS and 10◦ SB.

Figure 6.3: HTC ship model during free sailing manoeuvring test

For this thesis, only a subset of the test programme was used. This subset corresponds
to the conditions for which simulations were performed and is summarised in Table 7.5
on page 112, where also references are given to the table pages and figure pages with
the results. In MARIN Report No. 23277-3-SMB [105] a discussion of all tests is given.
98 Chapter 6 – Free sailing manoeuvring tests

The repeat tests with the ship model equipped with bilge keels were only used to provide
further estimates of the uncertainty in the experimental results.

6.7 Uncertainty analysis

6.7.1 Introduction
It is possible to conduct an analysis of the experimental uncertainty using the repeat
test results. In this section, the procedure proposed by the Guide to the Expression of
Uncertainty in Measurement (ISO-GUM) [72] is followed. In this guide, two methods
of evaluation of the uncertainty are given and the Type A method is adopted for this
thesis. The Type A method evaluates the uncertainty by statistical analysis of a series of
observations. The arithmetic mean or average φ and experimental standard deviation sφ
of each result φ are determined first. Then, based on the number of observations and the
standard deviation, an estimate of the uncertainty is made using a desired level of confi-
dence. Since only a limited number n of identical manoeuvres was realised, an estimate
of the experimental uncertainty U of the mean value φ is made using a coverage factor
based on the t-value from the Student t-distribution. Using the Student distribution, the
uncertainty estimates with 95% confidence (α = 5%) are found.
During the analysis of the model tests, it was observed that scatter existed in the
zig-zag test results for an approach speed of 10 kn, while the scatter for 18 kn was consid-
erably less. Furthermore, some irregularities were observed in the turning circle results for
18 kn. Therefore, a second test series was conducted to generate additional repeat tests to
improve the uncertainty estimates of the experimental results. The tests conducted in the
first series have test numbers consisting of six digits starting with 1, while the numbers of
the tests from the second series start with 2. It should be noted that the second test series
has been performed a few months after the first series, and that the setup of the model
has been redone for the second test series. This means that the estimated uncertainty
values include uncertainties due to the experimental setup. Further uncertainties such
as introduced due to model manufacture tolerances or measurement equipment precision
are assumed to be much smaller than the uncertainties found in the repeat tests and are
therefore not further considered.

6.7.2 Zig-zag tests

Using the repeat tests for 10 kn without bilge keels and 18 kn without and with bilge keels,
uncertainty estimates of the zig-zag test results were made. The results are presented in
Table 6.1. In Table 6.2 the uncertainty estimate based on the first test series only is given.
It is seen that the uncertainties for the 10 kn zig-zag tests are on average slightly
higher than for the 18 kn tests. At 10 kn, the second overshoot values for the 10◦ /10◦
zig-zag appear to be larger than at 18 kn, indicating an improving yaw checking and
course keeping ability for higher approach speeds. However, this does not apply to the
6.8 – Conclusion 99

20◦ /20◦ zig-zag test: at the higher approach speed, the higher turning rate during the
time required to reverse the rudder angle results in larger overshoot angles.
Interesting to see is the difference between the 18 kn results with and without bilge
keels. With bilge keels, the overshoot and initial turning ability values appear to in-
crease (i.e. a deterioration of the yaw checking and initial turning abilities) while also the
uncertainties in the results increase.
Comparing the uncertainty estimates based on the first test series (Table 6.2) and
those based on all tests (Table 6.1), it is observed that due to the larger number of
observations the uncertainty in the mean values of the zig-zag parameters reduced. This
is mainly caused by a reduction in the standard deviations sφ and a reduction of the
Student t coverage factor. From this, it can be concluded that the Student t coverage
factor results in a conservative uncertainty estimate.

6.7.3 Turning circle tests

Using the repeat tests for 18 kn without and with bilge keels, uncertainty estimates of the
turning circle test results were made, see Table 6.3. In Table 6.4 the uncertainty estimate
based on the first test series only is given.
The uncertainties in the advance and tactical diameter values are found to be small, i.e.
3% or less. There does not appear to be a consistent difference between the uncertainties
in the results with or without bilge keels. In general, it is seen that with bilge keels, the
turning ability improves (5% smaller tactical diameter) somewhat.
Similar to the uncertainty analysis for the zig-zag, it is observed that due to the larger
number of observations the uncertainty in the mean values of the turning circle parameters
is generally reduced (see Table 6.3 and Table 6.4). This is mainly caused by a reduction of
the Student t coverage factor while the standard deviation sφ remains roughly the same.
Also in this case, the Student t coverage factor results in a conservative uncertainty
estimation. Only for the steady turning diameter Dstc a small increase in uncertainty is
obtained by including the additional test.

6.8 Conclusion
Free sailing manoeuvring test on the HTC have been performed. This test campaign
resulted in a very valuable data set which can be used for public validation studies. Besides
obtaining general characteristics of the manoeuvrability of a single-screw container ship,
unique information has been obtained on the drift angles and rates of turn combined with
propeller and rudder forces. From this, important information for the development and
validation of manoeuvring prediction tools is obtained. Furthermore, repeat tests have
been conducted for selected manoeuvres and based on these tests, the uncertainty in the
characteristic manoeuvring properties has been estimated. Even when a small number
of observations is available, it is concluded that the verification procedure proposed in
100 Chapter 6 – Free sailing manoeuvring tests

Table 6.1: Experimental uncertainty estimate, zig-zag, 95% confidence interval, all tests

Zig-zag 10 kn, without bilge keels

δ/ψ = 10◦ /10◦ δ/ψ = 20◦ /20◦
1st overshoot 2nd overshoot Initial turning 1st overshoot
angle [deg] angle [deg] ability [Lpp ] angle [deg]
PS SB PS SB PS SB PS SB
Average φ ◦
9.8 10.1 ◦ ◦
22.3 25.0 ◦ 1.68 1.81 15.1 ◦ 15.4◦
Standard deviation sφ 1.3◦ 1.3◦ 1.6◦ 2.8◦ 0.05 0.25 1.0◦ 1.5◦
Observations n 8 8 8 8 8 8 4 4
tα/2 (n − 1) 2.36 2.36 2.36 2.36 2.36 2.36 3.18 3.18
√
U = t · sφ / n 1.1◦ 1.1◦ 1.3◦ 2.4◦ 0.05 0.21 1.7◦ 2.4◦
U/φ 12% 11% 6% 10% 3% 11% 11% 15%
φ−U ◦
8.6 9.0 ◦ ◦
21.0 22.6 ◦ 1.64 1.60 13.4 ◦ 13.0◦
φ+U ◦
10.9 11.2 ◦ ◦
23.6 27.4 ◦ 1.73 2.01 16.7 ◦ 17.7◦

Zig-zag 18 kn, without bilge keels

δ/ψ = 10◦ /10◦ δ/ψ = 20◦ /20◦
1st overshoot 2nd overshoot Initial turning 1st overshoot
angle [deg] angle [deg] ability [Lpp ] angle [deg]
PS SB PS SB PS SB PS SB
Average φ ◦
10.0 9.7 ◦ ◦
18.9 21.6 ◦ 1.57 1.67 18.6 ◦ 18.2◦
Standard deviation sφ 0.9◦ 0.2◦ 0.6◦ 0.2◦ 0.02 0.01 0.2◦ 0.7◦
Observations n 3 3 3 3 3 3 3 3
tα/2 (n − 1) 4.30 4.30 4.30 4.30 4.30 4.30 4.30 4.30
√
U = t · sφ / n 2.1◦ 0.6◦ 1.4◦ 0.6◦ 0.06 0.03 0.5◦ 1.7◦
U/φ 21% 6% 8% 3% 4% 2% 3% 10%
φ−U 7.9◦ 9.2◦ 17.4◦ 21.0◦ 1.51 1.64 18.1◦ 16.4◦
φ+U 12.1◦ 10.3◦ 20.3◦ 22.2◦ 1.63 1.70 19.1◦ 19.9◦

Zig-zag 18 kn, with bilge keels

δ/ψ = 10◦ /10◦ δ/ψ = 20◦ /20◦
1 overshoot 2 overshoot Initial turning 1st overshoot
st nd

angle [deg] angle [deg] ability [Lpp ] angle [deg]

PS SB PS SB PS SB PS SB
Average φ ◦
11.0 10.2 ◦ ◦
23.4 27.6 ◦ 1.68 1.70 21.2 ◦ 20.8◦
Standard deviation sφ 0.1◦ 0.2◦ 1.2◦ 1.5◦ 0.01 0.08 0.7◦ 0.6◦
Observations n 3 3 3 3 3 3 3 3
tα/2 (n − 1) 4.30 4.30 4.30 4.30 4.30 4.30 4.30 4.30
√ ◦ ◦ ◦ ◦ ◦
U = t · sφ / n 0.4 0.6 2.9 3.7 0.03 0.21 1.8 1.6◦
U/φ 3% 6% 12% 13% 2% 12% 8% 8%
φ−U ◦
10.6 9.6 ◦ ◦
20.6 23.9 ◦ 1.65 1.49 19.4 ◦ 19.2◦
φ+U 11.3◦ 10.8◦ 26.3◦ 31.3◦ 1.71 1.91 22.9◦ 22.4◦
6.8 – Conclusion 101

Table 6.2: Experimental uncertainty estimate, zig-zag, 95% confidence interval, set 1

Zig-zag 10 kn, without bilge keels

δ/ψ = 10◦ /10◦ δ/ψ = 20◦ /20◦
1st overshoot 2nd overshoot Initial turning 1st overshoot
angle [deg] angle [deg] ability [Lpp ] angle [deg]
PS SB PS SB PS SB PS SB
Average φ ◦
10.6 10.7 ◦ ◦
21.6 25.6 ◦ 1.67 1.61 15.1 ◦ 15.4◦
Standard deviation sφ ◦
0.3 1.0 ◦ 1.7 ◦ 4.6 ◦ 0.09 0.03 1.3◦ 1.8◦
Observations n 3 3 3 3 3 3 3 3
tα/2 (n − 1) 4.30 4.30 4.30 4.30 4.30 4.30 4.30 4.30
√
U = t · sφ / n 0.6◦ 2.5◦ 4.2◦ 11.4◦ 0.22 0.07 3.2◦ 4.5◦
U/φ 6% 23% 19% 45% 13% 4% 21% 29%
φ−U ◦
10.0 8.2 ◦ ◦
17.4 14.2 ◦ 1.46 1.54 11.9 ◦ 10.9◦
φ+U 11.2◦ 13.1◦ 25.8◦ 36.9◦ 1.89 1.68 18.3◦ 19.8◦

Table 6.3: Experimental uncertainty estimate, turning circle, 95% confidence interval, all tests

Turning circle 18 kn, δ = 35◦ , without bilge keels

Advance Tactical diam. Diameter
AD/Lpp TD/Lpp Dstc /Lpp
PS SB PS SB PS SB
Average φ 2.72 2.86 2.56 2.79 2.10 2.33
Standard deviation sφ 0.03 0.03 0.02 0.03 0.03 0.03
Observations n 4 4 4 4 4 4
tα/2 (n − 1) 3.18 3.18 3.18 3.18 3.18 3.18
√
U = t · sφ / n 0.04 0.04 0.04 0.04 0.05 0.05
U/φ 2% 2% 1% 2% 2% 2%
φ−U 2.68 2.81 2.53 2.75 2.05 2.28
φ+U 2.77 2.90 2.60 2.84 2.15 2.38

Turning circle 18 kn, δ = 35◦ , with bilge keels

Advance Tactical diam. Diameter
AD/Lpp TD/Lpp Dstc /Lpp
PS SB PS SB PS SB
Average φ 2.72 2.76 2.50 2.57 2.13 2.18
Standard deviation sφ 0.03 0.03 0.01 0.02 0.02 0.00
Observations n 3 3 3 3 3 3
tα/2 (n − 1) 4.30 4.30 4.30 4.30 4.30 4.30
√
U = t · sφ / n 0.07 0.08 0.02 0.04 0.04 0.01
U/φ 2% 3% 1% 2% 2% 1%
φ−U 2.66 2.68 2.48 2.53 2.09 2.17
φ+U 2.79 2.84 2.52 2.61 2.16 2.19
102 Chapter 6 – Free sailing manoeuvring tests

Table 6.4: Experimental uncertainty estimate, turning circle, 95% confidence interval, set 1

Turning circle 18 kn, δ = 35◦ , without bilge keels

Advance Tactical diam. Diameter
AD/Lpp TD/Lpp Dstc /Lpp
PS SB PS SB PS SB
Average φ 2.71 2.86 2.55 2.80 2.08 2.34
Standard deviation sφ 0.02 0.03 0.02 0.02 0.01 0.02
Observations n 3 3 3 3 3 3
tα/2 (n − 1) 4.30 4.30 4.30 4.30 4.30 4.30
√
U = t · sφ / n 0.05 0.08 0.05 0.05 0.02 0.04
U/φ 2% 3% 2% 2% 1% 2%
φ−U 2.66 2.78 2.50 2.75 2.07 2.30
φ+U 2.76 2.94 2.61 2.86 2.10 2.39

this chapter provides good estimates of the uncertainty in the measurements, provided a
Student t coverage factor is used.
In the following chapter, hydrodynamic coefficients will be derived and manoeuvring
simulations will be conducted for the HTC. The model tests results will be used to validate
these simulations.
 103

Chapter 7

Simulation of ship manoeuvrability

7.1 Introduction
When SurSim is used to predict the manoeuvrability of the HTC, results as shown
in Figure 7.1, Table 7.1 and Table 7.2 are obtained1 . It is clearly seen that in order to
reliably assess the manoeuvring behaviour of the ship an improvement of the mathematical
formulae is required: the comparison error for the first overshoot angle during the 20◦ /20◦
zig-zag manoeuvre is −72% of the experimental value and the comparison error in the
Tactical Diameter (TD) is +46%. In earlier studies, it was already seen that the forces
and moments on the bare hull predicted by SurSim were insufficiently accurate, see [148].
To improve the empirical formulations in the mathematical model, CFD calculations will
be used.

600
45

500
30

400
15
−δ [deg] ; ψ [deg]

x [m]

300
E

0
200

−15
100

−30
0
20◦ /20◦ PS 35◦ SB
−45
0 25 50 75 100 125 150 175 200 225 250 −100 0 100 200 300 400 500 600 700
108005 108013 205002 104028 104026
time 108009 zz09.26−20.00 yE [m] 104019 tc09.26−−35.00

Figure 7.1: Comparison between the original simulations and the free sailing experiments, 18 kn
(thick blue lines: simulation, others: experiments)

In this chapter, hydrodynamic coefficients for the modelling of the HTC bare hull
1
In this chapter, modifications to some of the empirical hull-propeller-rudder coefficients will be made.
In the original SurSim predictions here, the same modifications were made. This means that any differ-
ences between these simulations and the simulations presented later in this chapter are only caused by
changes in the mathematical model for the bare hull forces.
104 Chapter 7 – Simulation of ship manoeuvrability

forces and moments as presented in chapter 5 will be derived. Only the results computed
using Parnassos will be considered, to limit the scope of the work. The procedure pro-
posed in section 3.3 is followed. With the obtained coefficients, manoeuvring simulations
will be conducted with SurSim. The results of the simulations will be compared to the
free sailing manoeuvring experiments to demonstrate the feasibility of the approach and
the improvement of the simulations compared to the original fully empiric SurSim sim-
ulations. See section 2.5.4 and chapter 6 for more information on the HTC and the free
sailing manoeuvring experiments.

Table 7.1: Summary of zig-zag manoeuvre results, original simulations, 18 kn (average from
repeat tests)

δ/ψ = 10◦ /10◦ δ/ψ = 20◦ /20◦

st nd
V0 Lpp /V0 1 overshoot 2 overshoot Initial turning 1st overshoot
[kn] [s] angle [deg] angle [deg] ability [Lpp ] angle [deg]
Result IMO Result IMO Result IMO Result IMO
◦ ◦ ◦ ◦ ◦ ◦
Exp PS 10.0 ±2.1 18.9 ±1.4 1.57±0.06 18.6 ±0.5
Exp SB 9.7◦ ±0.6◦ 13.3 21.6◦ ±0.6◦ 29.9 1.67±0.03 2.5 18.2◦ ±1.7◦ 25◦
18 16.6 SurSim 2.8◦ 3.6◦ 2.25 6.2◦
E PS -7.2◦ -15.3◦ 0.68 -12.4◦
◦ ◦
E SB -6.9 -18.0 0.58 -12.0◦

Table 7.2: Summary of turning circle manoeuvre results, original simulations, 18 kn (average
from repeat tests)

V0 Lpp /V0 δ Advance Tactical diameter

[kn] [s] [deg] AD/Lpp T D/Lpp
Result IMO Result IMO
Exp PS 2.72±0.04 2.56±0.04
Exp SB 2.86±0.04 4.5 2.79±0.04 5.0
18 16.6 35◦ SurSim 3.96 3.97
E PS 1.24 1.41
E SB 1.10 1.18

7.2 Deriving the hydrodynamic coefficients

In this section, the coefficients for the mathematical model in SurSim as described in
chapter 3 will be derived. The forces and moments predicted with this mathematical
model will be compared to the available validation data. First, the resistance curve is
estimated, while subsequently, the ship-dependent interaction coefficients between the
hull, propeller and rudder will be derived. Finally, the procedure as presented in section
3.3 will be followed to obtain the hydrodynamic coefficients for the hull forces.
7.2 – Deriving the hydrodynamic coefficients 105

7.2.1 Resistance curve, wake fraction and thrust deduction frac-

tion
The resistance curve, wake fraction and thrust deduction fraction are obtained from ex-
periments conducted by HSVA prior to the VIRTUE project. Since validation of the
simulations will be done using model experiments, the model scale resistance curve must
be used. The resistance curve was obtained for a loading condition (trimmed by the stern)
that did not completely correspond to the loading condition during the VIRTUE captive
experiments for the manoeuvring workpackage. Therefore, a new resistance curve for the
different loading condition was predicted using DESP, see section 3.2.3. Furthermore,
the resistance curve for the free sailing manoeuvring model was obtained, i.e. for a scale
of λ = 30.02. In Figure 7.2, the original resistance curve (HSVA trimmed), the estimated
resistance curve for the captive condition (est. even keel) and the estimated resistance for
scale 1:30.02 are shown.
Since the simulation program SurSim uses the input values corresponding to the
full-scale HTC, the resistance curve is scaled by λ3 to arrive at values suitable for the
simulation. For all comparison between the SurSim results and the captive measurements
the resistance curve for scale λ = 24 is used, while for the comparisons with the free sailing
tests the resistance curve for scale λ = 30.02 is used.

200
1:24 (HSVA trimmed)
180 1:24 (est. even keel)
160 1:24 (EFD captive)
140 1:30.02 (est.)
120
Rm [N]

100
80
60
40
20
0
0 0.5 1 1.5 2 2.5
Vm [m/s]

Figure 7.2: Estimated resistance curve (model scale values)

The wake fraction w and thrust deduction fraction t for the captive condition are
estimated based on DESP predictions. The predicted values are: w = 0.38 and t = 0.22.

7.2.2 Propeller characteristics

Using the Strom-Tejsen propeller model [134] for the HSVA 2208 propeller and open water
test results for MARIN propeller No. 5286, the propeller open-water curves as presented
106 Chapter 7 – Simulation of ship manoeuvrability

in Figure 7.3 are obtained. For all comparisons between the SurSim results and the
captive measurements propeller 2208 is modelled, while for the comparisons with the free
sailing tests propeller 5286 is modelled. The coefficients derived from the open water tests
with propeller No. 5286 are as given in Table 7.3.

0.5
KT (2208)
KQ (2208)
0.4 KT (5286)
KQ (5286)
0.3
KT , 10 · KQ

0.2

0.1

-0.1
0 0.2 0.4 0.6 0.8 1
J

Figure 7.3: Estimated propeller open water curves

Table 7.3: Propeller No. 5286, open water test No. 45127

Coefficient Value Coefficient Value

KT 0 0.366897 KQ0 0.040802
KT 1 -0.345036 KQ1 -0.029636
KT 2 0.068841 KQ2 0.000525
KT 3 -0.710991 KQ3 -0.066086
KT 4 0.948559 KQ4 0.105238
KT 5 -0.428915 KQ5 -0.059477

7.2.3 Rudder forces

Rudder-to-hull interaction

Based on the HSVA captive experiments, the relation between the side force on the rudder
and its effect on the total side force on the ship was validated. The value calculated by
SurSim for (1 + aH ) (see Equation 3.37) appears to correctly model the rudder force on
the ship. For the HTC, a value of (1 + aH ) = 1.255 is found. For the ship without rudder,
a relation between the rudder angle and the force on the ship as shown in Figure 7.4 is
obtained. It is seen that this modelling closely approximates the experimental values,
except for the largest rudder angles, where stall appears to be present in the experiments.
7.2 – Deriving the hydrodynamic coefficients 107

However, during the manoeuvres studied in this thesis, the rudder operates in the propeller
race, and it is known that for those conditions, the stall angle increases considerably, see
Kracht [80] or Molland and Turnock [94], and a lift curve as modelled by SurSim is more
appropriate.

1 8
SurSim SurSim
exp 6 exp
0.5 4

N × 104 [kNm]
Y × 103 [kN]

2
0 0
-2
-0.5 -4
-6
-1 -8
-40-30-20-10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
δ [deg] δ [deg]

Figure 7.4: Forces on the ship as function of rudder angle, HTC without propeller, 18 kn, β = 0◦

Propeller-to-rudder interaction
Based on the HSVA captive experiments, the relation between the propeller thrust and
induced velocity on the rudder was estimated (see Equation 3.28). To correlate the forces
predicted by SurSim to the measured forces on the ship, the following value is used:
Crue = 0.55. A relation between the rudder angle and the force on the ship for different
propeller revolutions n as given in Figure 7.5 is obtained. The revolutions at model self
propulsion are designated ns . It is seen that this modelling closely approximates the
experimental values, except for the largest rudder angles.
108 Chapter 7 – Simulation of ship manoeuvrability

3
SurSim, no prop
SurSim, ns
2 SurSim, 1.4 · ns
exp
Y × 103 [kN]

-1

-2

-3
-40 -30 -20 -10 0 10 20 30 40
δ [deg]
2
SurSim, no prop
1.5 SurSim, ns
SurSim, 1.4 · ns
1 exp
N × 105 [kNm]

0.5
0
-0.5
-1
-1.5
-2
-40 -30 -20 -10 0 10 20 30 40
δ [deg]

Figure 7.5: Forces on the ship as function of rudder angle and propeller revolutions,
18 kn, β = 0◦
7.2 – Deriving the hydrodynamic coefficients 109

Flow straightening
The relation between the drift angle β and rotation rate γ and the effective inflow at the
rudder was validated with the HSVA captive experiments. The last two free parameters
in the SurSim rudder model are the coefficients for the flow straightening for drift (Cdb )
and the flow straightening for rotation (Cdr ), see Equation 3.29. These needed to be
modified to obtain better agreement with the tests. For the HTC, the following values
were adopted: Cdb = 0.9, Cdr = 0.8. Figure 7.6 and Figure 7.7 show the relation between
the rudder angle and the force on the ship for a drift angle of β = −10◦ and for a yaw rate of
γ = 0.4 respectively. With these settings, the rudder forces are modelled reasonably well,
although some discrepancies still remain. Further improvement to the rudder modelling
is however judged to be outside of the scope of the present work, since the main focus is
to demonstrate the influence of the bare hull force model on the manoeuvrability of the
ship.

6 4.5
SurSim SurSim
exp 4 exp
5
3.5
N × 105 [kNm]
Y × 103 [kN]

4 3
2.5
3
2
2 1.5
1
1
0.5
0 0
-40 -30 -20 -10 0 10 20 30 40 -40-30-20-10 0 10 20 30 40
δ [deg] δ [deg]

Figure 7.6: Forces on the ship as function of rudder angle, 18 kn, β = −10◦ , n = ns

7.2.4 Hull forces

The forces and moments on the bare hull are based on a fit through the RANS calcu-
lations with Parnassos for steady drift, steady rotation and combined motion. These
calculations were presented in section 5.3 and an overview of the results can be found in
the tables on page 157 and page 158. The procedure of deriving the coefficients has been
described in section 3.3. Coefficients for pure sideway motion (Yβ|β| , Nβ|β| ) or turning on
the spot (Yγ|γ| , Nγ|γ| ) were based on the default values in SurSim. The hydrodynamic
damping coefficients derived from the viscous-flow calculations and the added mass coeffi-
cients estimated by SurSim are given in Table 7.4, while the transverse forces and yawing
moments as a function of drift or yaw motion are given in Figure 7.8. The (added) masses
are made non-dimensional using 21 ρL2pp T and the (added) inertias by 12 ρL4pp T .
110 Chapter 7 – Simulation of ship manoeuvrability

5 0
SurSim SurSim
exp -0.5 exp
4
-1

N × 105 [kNm]
Y × 103 [kN]

3 -1.5
-2
2
-2.5
1 -3
-3.5
0
-4
-1 -4.5
-40 -30 -20 -10 0 10 20 30 40 -40-30-20-10 0 10 20 30 40
δ [deg] δ [deg]

Figure 7.7: Forces on the ship as function of rudder angle, 18 kn, γ = 0.4, n = ns

Table 7.4: Hydrodynamic bare hull and added mass coefficients, HTC

Coefficient ValueCoefficient Value Coefficient Value

0
Xu|u| -0.0141Xβγ0 0.1025 m0 0.2328
Yβ0 -0.1735Nβ0 -0.1442 0
Izz 0.0134
Yγ0 0.0338Nγ0 -0.0276 m0uu 0.0247
0 0
Yβ|β| -1.1378Nβ|β| -0.0375 m0vv 0.2286
0 0
Yγ|γ| 0.0123Nγ|γ| -0.0386 m0rr 0.0150
0
Yβ|γ| -0.0537 0
Nββγ -0.9037 m0vr 0.0074
0
Y|β|γ 0.1252 0
Nβγγ -0.2679 m0rv 0.0074
0
Yab 0.6747Nab0 -0.0300 m0 + m0uu 0.2575
0
N|u|γc -0.0075 m0 + m0vv 0.4614
0 + m0
Izz 0.0284
ay 3 an 1 rr
by 2 bn 3
cn 2
7.2 – Deriving the hydrodynamic coefficients 111

0.4 0.2
β = −10◦
0.3 β = 0◦ 0.15
β = 6◦
0.2 β = 10◦ 0.1
β = 16◦
0.1 0.05

N [-]
Y [-]

0 0
-0.1 -0.05
β = −10◦
-0.2 -0.1 β = 0◦
β = 6◦
-0.3 -0.15 β = 10◦
β = 16◦
-0.4 -0.2
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 -0.8-0.6-0.4-0.2 0 0.2 0.4
γ γ
0.4 0.2
γ = −0.6
0.3 γ = −0.4 0.15
γ = −0.2
0.2 γ=0 0.1
γ = 0.2
0.1 0.05
N [-]
Y [-]

0 0
-0.1 -0.05
γ = −0.6
-0.2 -0.1 γ = −0.4
γ = −0.2
-0.3 -0.15 γ =0
γ = 0.2
-0.4 -0.2
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
β [deg] β [deg]
Figure 7.8: Forces on the bare hull as function of yaw rate (top) or drift angle (bottom)
(lines: fit, markers: cfd results)
112 Chapter 7 – Simulation of ship manoeuvrability

7.3 Standard manoeuvres

7.3.1 Programme of simulations

Calculations were conducted for speeds corresponding to 10 kn and 18 kn on full scale,
or F n = 0.132 and F n = 0.238 respectively. All manoeuvres were conducted with a full
scale rudder turning rate of δ̇ = 4.6◦ /s. In Table 7.5 the simulations that were performed
are indicated, together with the specification of the table and figure pages on which the
results are presented. In this section, all values given are presented in full scale values.

Table 7.5: Simulation matrix, HTC

Manoeuvre Presentation
Id V0 δ ψ Table page Figure page
[kn] [deg] [deg] Timetrace Track
zz05.14-10.00 10 -10 161 177 178
zz05.14–10.00 10 -10 10 162 179 180
zz05.14-20.00 20 -20 163 181 182
zz05.14–20.00 -20 20 164 183 184
zz09.26-10.00 10 -10 165 185 186
zz09.26–10.00 18 -10 10 166 187 188
zz09.26-20.00 20 -20 167 189 190
zz09.26–20.00 -20 20 168 191 192
tc05.14-35.00 35 169 193 194
tc05.14–35.00 -35 195 196
tc05.14-25.00 10 25 - 170 197 198
tc05.14–25.00 -25 199 200
tc05.14-15.00 15 171 201 202
tc05.14–15.00 -15 203 204
tc09.26-35.00 35 172 205 206
tc09.26–35.00 -35 173 207 208
tc09.26-25.00 18 25 - 174 209 210
tc09.26–25.00 -25 211 212
tc09.26-15.00 15 175 213 214
tc09.26–15.00 -15 215 216

All calculations have been conducted without incorporating heel. The model tests
have been performed with a GM value that is relatively high for this type of ship and
therefore the influence of heel on the manoeuvres is expected to be small. It should be
noted that for speeds close to the design speed of the ship, assessment of the heel angle
for this type of ship will be important when the GM value is small.
7.4 – Sensitivity study 113

7.3.2 RPM-Speed curve

Prior to conducting free sailing model tests for a certain speed, the required RPM to sail
this speed needs to be determined. Normally, this is done by applying a pre-defined RPM
value to the propeller and measuring the speed of the model obtained with this RPM. This
is done for several different RPM settings, resulting in the RPM-Speed relation curve. By
interpolating at a given speed, the required RPM is obtained.
With SurSim, the RPM-Speed curve can be predicted, when the proper resistance
curve and propeller particulars are available. With the resistance curve and the propeller
characteristics as given in sections 7.2.1 and 7.2.2, the RPM-Speed relation as shown in
Figure 7.9 is obtained. For a speed of 0.938 m/s (corresponding to 10 kn on full scale),
an RPM of 389 is found, while for a speed of 1.69 m/s (corresponding to 18 kn on full
scale), an RPM of 700 is found. For the experiments, the values of 400 and 745 are found
respectively, resulting in comparison errors of 2.7% and 6.0%. This is judged to be quite
small considering the possible uncertainty in the wake and thrust fractions and in the
resistance curve used in SurSim.

1000
CFD 1:30.02
Exp. (bilgekeels)
800 Exp. (no bilgekeels)
RPM (model)

600

400

200

0
0 0.5 1 1.5 2
Vm [m/s]

Figure 7.9: RPM-Speed relation for HTC, scale 1:30.02, model scale values

7.4 Sensitivity study

A sensitivity study was conducted in order to determine the influence of estimation errors
in each hydrodynamic manoeuvring derivative on the results for standard manoeuvres.
In the present study, a set of manoeuvres using the mathematical model described in
section 7.2 was simulated during which one of the coefficients was individually multiplied
by a factor of 1.1. This value was chosen based on the uncertainties in the order of 10%
in the predicted forces and moments, as determined in chapter 5. The changes in the
114 Chapter 7 – Simulation of ship manoeuvrability

manoeuvring characteristics can be determined and be expressed as percentages of the

original values. The resulting factors are the Uncertainty Magnification Factors (UMFs).
Zig-zag manoeuvres were conducted to obtain the first and second Overshoot Angle
(osa) and the Initial Turning Ability (ITA) during the 10◦ /10◦ manoeuvre and the first
overshoot angle during the 20◦ /20◦ zig-zag manoeuvre. From turning-circle manoeuvres
with 35◦ steering angle, the Advance (AD) and TD were obtained. From the steady
turning circle results, the yaw rate rstc , velocity Vstc and drift angle βstc were derived as
well.

HTC HTC
40 40
X’u’|u’|×1.1
30 30 N’β×1.1
X’βγ×1.1
N’γ×1.1
20 Y’β×1.1 20
N’γ|γ|×1.1
10 Y’γ×1.1 10
Change %

Change % N’β|β|×1.1
0 Y’β|β|×1.1 0
N’ab×1.1
Y’γ|γ|×1.1
−10 −10
N’|u’|γc×1.1
Y’ab×1.1
−20 −20 N’ββγ×1.1
Y’β|γ|×1.1
−30 −30 N’βγγ×1.1
Y’|β|γ×1.1
−40 −40
osa1 osa2 ita osa1 osa1 osa2 ita osa1
10/10 10/10 10/10 20/20 10/10 10/10 10/10 20/20
HTC
40

30

20 Crue×1.1
10 Chru×1.1
Change %

0 Cdbluff×1.1

−10 Cdblee×1.1
Cdr×1.1
−20

−30

−40
osa1 osa2 ita osa1
10/10 10/10 10/10 20/20

Figure 7.10: Sensitivity study, HTC, zig-zag, 10 kn

Based on the sensitivity study, the results as collected in Figure 7.10, Figure 7.11 and
Table 7.6 were obtained for an approach speed of 10 kn. Another sensitivity study that
was conducted for an approach speed of 18 kn shows similar results. It is clear that for
the HTC deviations in Nβ0 have the largest impact on the accuracy of the prediction of
the yaw checking and course keeping ability, while of all linear coefficients it also has the
largest influence on the turning ability. Nγ0 is also an important coefficient. Yγ0 is the least
important linear coefficient for accurate predictions. Furthermore, it is seen that for the
zig-zag manoeuvres, the linear derivatives are more important compared to the non-linear
derivatives than during the turning circle manoeuvres. It is also found that the 10◦ /10◦
zig-zag manoeuvre is more sensitive to changes in the hydrodynamic derivatives than the
20◦ /20◦ zig-zag manoeuvre.
0
The turning ability is most sensitive to changes in the non-linear coefficients Nββγ
7.4 – Sensitivity study 115

HTC HTC
40 40
X’u’|u’|×1.1 X’u’|u’|×1.1
30
30
X’βγ×1.1 X’βγ×1.1
20 Y’β×1.1
20 Y’β×1.1

Y’γ×1.1 10 Y’γ×1.1

Change %
10
Change %

Y’β|β|×1.1 0 Y’β|β|×1.1
0
Y’γ|γ|×1.1 Y’γ|γ|×1.1
−10 −10
Y’ab×1.1 Y’ab×1.1
−20 −20
Y’β|γ|×1.1 Y’β|γ|×1.1
−30 −30
Y’|β|γ×1.1 Y’|β|γ×1.1
−40 −40
AD TD
rstc Vstc βstc

HTC HTC
40 40
N’β×1.1 30 N’β×1.1
30
N’γ×1.1 N’γ×1.1
20 20
N’γ|γ|×1.1 N’γ|γ|×1.1
10

Change %
10
Change %

N’β|β|×1.1 N’β|β|×1.1
0 0
N’ab×1.1 N’ab×1.1
−10
−10
N’|u’|γc×1.1 N’|u’|γc×1.1
−20 N’ββγ×1.1
−20 N’ββγ×1.1

−30 N’βγγ×1.1 −30 N’βγγ×1.1

−40 −40
AD TD
rstc Vstc βstc

HTC HTC
40 40

30 30

20 20 Crue×1.1
Crue×1.1
Chru×1.1 10 Chru×1.1
Change %

10
Change %

0 Cdbluff×1.1 0 Cdbluff×1.1

Cdblee×1.1 −10 Cdblee×1.1

−10
Cdr×1.1 Cdr×1.1
−20 −20

−30 −30

−40 −40
AD TD
rstc Vstc βstc

Figure 7.11: Sensitivity study, HTC, turning circle, 10 kn

0
and Nβγγ . The sensitivity of the steady turning circle results to changes in the non-
0
linear coefficient Nβγγ is large: when this coefficient is increased by 10%, the rate of
turn increases and the speed drops such that an unrealistic situation is reached and the
simulation is aborted (indicated by a change of 100%).
Similar conclusions were found by Lee and Shin [84] who studied zig-zag manoeuvres
for a chemical carrier and two oil tankers and Bulian et al. [18] who conducted a sensitivity
study for the Esso Osaka.
The hull-propeller-rudder interaction coefficients Crue , Chru (= 1 + aH ), Cdbluf f (= Cdb )
and Cdr also have a relatively large influence on the yaw checking and course keeping
ability of the ship. From these coefficients, the results are most sensitive to changes in
Chru .
116 Chapter 7 – Simulation of ship manoeuvrability

The sensitivity study demonstrates that for accurate predictions of the manoeuvra-
bility using coefficients derived from CFD calculations, accurate predictions of especially
the yawing moment must be made. It should be noted however, that the sensitivity of the
results depends on the individual ship, due to different balancing between coefficients.
Changes in coefficients are not necessarily independent of changes in other coefficients.
For example, in the procedure used to derive the coefficients, a change in the linear
0
coefficient Yγ0 will lead to changes in the non-linear coefficients Yβ|γ| 0
and Y|β|γ , since these
0
are derived after subtracting the linear contribution Yγ ·cos β·γ and non-linear contribution
0
Yγ|γ| · γ · |γ| from the total force Y 0 .

Table 7.6: Sensitivity study, HTC, 10 kn, changes in percentages of the original values
(Blue values indicate UMFs larger than 50%, while red values indicate UMFs larger
than 100%)

Zig-zag manoeuvres Turning circle manoeuvres

10/10 20/20 35◦
Variation osa1 osa2 ita osa1 AD TD rstc Vstc βstc
Xu0 0 |u0 | ×1.1 -7.25 -8.41 -1.86 -4.63 -1.84 -1.45 3.05 2.53 0.28
0 ×1.1
Xβγ -0.14 -0.67 0.00 -0.25 0.00 -0.36 -1.02 -1.90 1.18
0
Yβ ×1.1 -7.04 -7.52 0.62 -3.45 -0.37 -1.45 2.37 -1.27 -0.28
0
Yγ ×1.1 -3.73 -3.07 0.62 -1.72 0.00 -0.36 0.85 -0.63 -0.14
0
Yβ|β| ×1.1 -2.90 -5.44 0.00 -2.71 -0.74 -3.99 8.97 -5.06 -0.52
0
Yγ|γ| ×1.1 -0.28 -0.26 0.00 -0.15 0.00 0.00 0.34 0.00 -0.03
0
Yab ×1.1 2.00 3.58 0.00 1.62 0.37 2.17 -3.05 1.90 0.56
0 ×1.1
Yβ|γ| -0.41 -0.51 0.00 -0.30 0.00 -0.36 0.85 -0.63 -0.10
0 ×1.1
Y|β|γ -0.62 -1.25 0.00 -0.74 0.00 -0.72 1.86 -1.27 -0.24
0
Nβ ×1.1 35.06 33.85 -1.86 20.97 -3.68 -6.52 4.74 -4.43 2.85
0
Nγ ×1.1 -10.42 -10.20 1.86 -7.24 2.57 2.90 -1.69 1.27 -1.08
0
Nγ|γ| ×1.1 -2.42 -3.33 0.62 -3.10 1.84 3.26 -2.54 2.53 -1.56
0
Nβ|β| ×1.1 0.55 0.96 0.00 0.54 0.00 -0.72 0.68 -0.63 0.38
0
Nab ×1.1 0.00 0.13 0.00 0.05 0.00 0.00 0.17 0.00 0.14
0
N|u0 |γc ×1.1 -0.41 -0.61 0.00 -0.54 0.37 0.72 -0.51 0.63 -0.28
0
Nββγ ×1.1 -1.73 -4.80 0.00 -3.89 3.68 11.59 -10.32 10.76 -6.81
0 ×1.1
Nβγγ 1.31 3.20 0.00 2.86 -2.94 -8.70 -100.00 -100.00 -100.00
Crue ×1.1 -9.45 -12.25 -2.48 -6.35 -2.21 -2.54 -0.51 -8.23 10.63
Chru ×1.1 -18.84 -22.20 -3.11 -12.06 -2.57 -1.81 0.85 -2.53 3.99
Cdbluf f ×1.1 -6.97 -8.16 0.62 -4.28 0.74 1.09 -0.17 1.27 -1.60
Cdblee ×1.1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Cdr ×1.1 -7.04 -6.85 1.24 -4.09 0.74 1.09 -0.17 1.27 -1.49
7.5 – Validation 117

Table 7.7: Summary of zig-zag manoeuvre results (average from repeat tests)

δ/ψ = 10◦ /10◦ δ/ψ = 20◦ /20◦

st nd
V0 Lpp /V0 1 overshoot 2 overshoot Initial turning 1st overshoot
[kn] [s] angle [deg] angle [deg] ability [Lpp ] angle [deg]
Result IMO Result IMO Result IMO Result IMO
Exp PS 9.8◦ ±1.1◦ 22.3◦ ±1.3◦ 1.68±0.05 - 15.1◦ ±1.7◦
Exp SB 10.1◦ ±1.1◦ - 25.0◦ ±2.4◦ - 1.81±0.21 - 15.4◦ ±2.4◦ -
◦ ◦
10 29.9 CFD 14.5 31.3 1.61 20.3◦
E PS 4.7◦ 9.0◦ -0.07 5.2◦
E SB 4.4◦ 6.3◦ -0.20 4.9◦
Exp PS 10.0◦ ±2.1◦ 18.9◦ ±1.4◦ 1.57±0.06 18.6◦ ±0.5◦
Exp SB 9.7 ±0.6 13.3 21.6 ±0.6 29.9 1.67±0.03 2.5 18.2◦ ±1.7◦ 25◦
◦ ◦ ◦ ◦

18 16.6 CFD 16.2◦ 32.8◦ 1.64 23.7◦

◦ ◦
E PS 6.2 13.9 0.07 5.1◦
E SB 6.5◦ 11.2◦ -0.03 5.6◦

7.5 Validation
A comparison between the simulations based on the improved hydrodynamic derivatives
and the free sailing experiments is made in Table 7.7, Figure 7.12 and Figure 7.13 for the
zig-zag manoeuvres and in Table 7.8 and Figure 7.14 for the turning circles.

7.5.1 Zig-zag manoeuvres

The predicted ITA is within the measurement accuracy, when averaging the results for the
manoeuvres started to port side and starboard. For the zig-zag manoeuvres at 10 kn, it is
seen that the overshoot angles are considerably over-predicted, especially for the 10◦ /10◦
zig-zag. The comparison error E is about 5◦ , which is judged to be large. When looking
at the predicted rate of turn r (see e.g. figure page 179), the physics appear to be very
well predicted, since the shape of the rate of turn time trace resembles the measured time
trace quite closely. However, the slightly larger value of the rate of turn means a larger
build-up of momentum and subsequently larger overshoot angle. The simulations of the
20◦ /20◦ zig-zag manoeuvres show a better agreement with the experiments, in particular
for 18 kn, see Figure 7.13. In this case, the increase of overshoot angles with increasing
speed is captured as well.
Considering the IMO criteria [69] for the yaw checking and course keeping ability,
the experiments indicate that the HTC complies with the criteria, although the margin is
small: the first overshoot angle for the 10◦ /10◦ zig-zag manoeuvre is close to the criterion.
According to the simulations, the HTC does not comply with the criteria for the yaw
checking and course keeping ability: both the first and second overshoot angles are larger
than the limits. Although the simulations provide conservative values, the distinction
between whether or not the HTC complies with the IMO zig-zag criteria is not predicted
reliably.
118 Chapter 7 – Simulation of ship manoeuvrability

Compared to the original simulations with the original bare hull mathematical model
of SurSim, see Table 7.1, a considerable improvement is obtained: all comparison errors
reduce in magnitude and now the predictions of the overshoot angles are conservative
instead of too optimistic in case of the original SurSim predictions.

45 45

◦ ◦
30
10 /10 PS 30
10◦ /10◦ SB

15 15
−δ [deg] ; ψ [deg]

−δ [deg] ; ψ [deg]
0 0

−15 −15

−30 −30

−45 −45
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
203013 203008 203004 106011 106002 203011 203007 203003 106010 106005
203010 203006 time 106007 zz05.14−10.00 203009 203005 time 106006 zz05.14−−10.00

45 45

30 30
20◦ /20◦ PS
15 15
−δ [deg] ; ψ [deg]

−δ [deg] ; ψ [deg]

0 0

−15 −15

−30 −30
20◦ /20◦ SB
−45 −45
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
204003 106013 106004 204002 106012 106003
time 106009 zz05.14−20.00 time 106008 zz05.14−−20.00

Figure 7.12: Comparison between the simulations and the free sailing zig-zag experiments, 10 kn
(thick blue lines: simulation, others: experiments)

45 45

30 30
20◦ /20◦ PS
15 15
−δ [deg] ; ψ [deg]

−δ [deg] ; ψ [deg]

0 0

−15 −15

−30 −30
20◦ /20◦ SB
−45 −45
0 25 50 75 100 125 150 175 200 225 250 0 25 50 75 100 125 150 175 200 225 250
108005 108013 108012 108003
time 108009 zz09.26−20.00 time 108008 zz09.26−−20.00

Figure 7.13: Comparison between the simulations and the free sailing 20◦ /20◦ zig-zag experi-
ments, 18 kn (thick blue lines: simulation, others: experiments)
7.6 – Conclusion 119

7.5.2 Turning circle manoeuvres

The prediction of the turning ability, and especially of the advance AD and tactical
diameter T D is more impressive. The simulation results are very close to the average of
the manoeuvres started to port side and starboard. The comparison errors for AD and
T D are small, i.e. less than 0.1 × Lpp . When the manoeuvre reaches the steady turning
circle condition, some deviations from the experiments are seen for steering angles of 35◦ :
in general, the drift angle β and the speed loss (V0 − Vstc ) are over-predicted. This results
in a slightly smaller turning diameter Dstc . These observations apply to the manoeuvres
conducted with an approach speed of 10 kn as well as for 18 kn.
The distinction whether or not the HTC complies with the IMO turning ability criteria
is accurately made using the predictions. This is however not a surprise considering the
margin to the limiting values.
Compared to the simulations with the original bare hull mathematical model of Sur-
Sim, see Table 7.2, a considerable improvement is obtained with the new hydrodynamic
coefficients based on CFD calculations: all comparison errors reduce in magnitude and
now the predictions of the tactical diameter values are very close to the experimental
ones.

7.6 Conclusion
Using hydrodynamic manoeuvring coefficients derived from CFD calculations of the forces
on the bare hull, it has been shown that it is possible to improve the prediction of ship ma-
noeuvres compared to predictions using coefficients based on empirical equations, which
was the objective of the present study. In this chapter, it has been demonstrated that
a considerable improvement of the turning circle predictions was obtained. The predic-
tion of the yaw checking and course keeping and initial turning abilities based on zig-zag
simulations improved as well, but further improvements are required for more reliable
assessment of the manoeuvring performance.
The sensitivity of the manoeuvring predictions on changes in the hydrodynamic co-
efficients was studied. It was found that the linear coefficients mostly determine the
sensitivity of the results of the zig-zag manoeuvres, while non-linear coefficients affect
mostly the turning circle results. Hull-propeller-rudder coefficients were also found to
be important in the sensitivity study. The study demonstrates that for accurate predic-
tions of the manoeuvrability using coefficients derived from CFD calculations, accurate
predictions of especially the yawing moment must be made.
120 Chapter 7 – Simulation of ship manoeuvrability

Table 7.8: Summary of turning circle manoeuvre results (average from repeat tests)

V0 Lpp /V0 δ Advance Tactical diameter

[kn] [s] [deg] AD/Lpp T D/Lpp
Result IMO Result IMO
Exp PS 2.76 2.70
Exp SB 2.77 - 2.85 -
35◦ CFD 2.72 2.76
E PS -0.04 0.06
E SB -0.05 -0.09
Exp PS 3.24 3.30
Exp SB 3.10 - 3.44 -
10 29.9 25◦ CFD 3.09 3.27
E PS -0.15 -0.03
E SB -0.01 -0.17
Exp PS 3.86 4.11
Exp SB 3.73 - 4.40 -
15◦ CFD 3.78 4.15
E PS -0.03 0.04
E SB 0.10 -0.25
Exp PS 2.72±0.04 2.56±0.04
Exp SB 2.86±0.04 4.5 2.79±0.04 5.0
35◦ CFD 2.81 2.77
E PS 0.09 0.21
E SB -0.05 -0.02
Exp PS 3.02 3.14
Exp SB 3.35 - 3.42 -
18 16.6 25◦ CFD 3.15 3.27
E PS 0.13 0.13
E SB -0.20 -0.15
Exp PS 3.64 4.00
Exp SB 4.10 - 4.43 -
15◦ CFD 3.83 4.15
E PS 0.19 0.15
E SB -0.27 -0.28
7.6 – Conclusion 121

450
400
400
350
350
300
300
250
250
xE [m]

x [m]
200 200

E
150 150

100 100

50 50

0 0
◦
−50 35 PS −50 35◦ SB
−500 −400 −300 −200 −100 0 100 −100 0 100 200 300 400 500
205003 104021 104029 205002 104028 104026
y [m] 104027 tc09.26−35.00 y [m] 104019 tc09.26−−35.00
E E

500
450

400
400
350

300
300
250
xE [m]

x [m]

200
E

200
150

100 100
50

0 0
−50 25◦ PS 25◦ SB
−600 −500 −400 −300 −200 −100 0 100 −100 0 100 200 300 400 500 600
104023 104022
y [m] tc09.26−25.00 y [m] tc09.26−−25.00
E E

600

600
500
500
400
400

300
xE [m]

x [m]

300
E

200
200

100
100

0 0
15◦ PS 15◦ SB
−700 −600 −500 −400 −300 −200 −100 0 100 −200 −100 0 100 200 300 400 500 600 700 800
104025 104024
yE [m] tc09.26−15.00 yE [m] tc09.26−−15.00

Figure 7.14: Comparison between the simulations and the free sailing turning circle experiments,
18 kn (thick blue lines: simulation, others: experiments)
122 Chapter 7 – Simulation of ship manoeuvrability

Page intentionally left blank

 123

Chapter 8

Conclusions and recommendations

8.1 Conclusions
The present work was initiated in order to improve traditional manoeuvring simulations
based on empirical mathematical equations to model the forces and moments on the ship.
With the evolution of viscous-flow solvers and their promising results in predicting the
forces and moments on ships, it was decided to develop a practical method to simulate the
manoeuvrability of ships in which viscous-flow solvers are utilised and investigate whether
this improves the accuracy of the predicted simulations.
Several different methods are given in literature to simulate ship manoeuvres using
viscous-flow calculations, see chapter 2. In this thesis, the virtual captive test approach is
adopted, because of the efficient use of computational resources compared to other meth-
ods. Furthermore, this approach can directly be used to improve mathematical models
for manoeuvring simulators. The present study extends the work of other researchers by
providing extensive verification and validation of the predicted forces and moments on
the hull and a detailed study of the sensitivity of the manoeuvring characteristics of the
ship to changes in the hydrodynamic coefficients in the simulation model.

Concerning the forces and moments acting on a ship hull in manoeuvring motions, it is
noteworthy that validation data for captive steady drift motions are much more abundant
than for steady rotation or combined motion. Especially for more extreme conditions with
large turning rates and drift angles, such as occur during tight turning circles, not much
validation data can be found in literature. Furthermore, cases in which both extensive
captive test data and free sailing manoeuvring test data are available are scarce.
In the work leading to this thesis, much effort was spent on simulating the flow around
the HTC for captive conditions and simulations of standard free sailing manoeuvres were
conducted within the VIRTUE project. To generate validation data for these manoeu-
vres, MARIN decided to perform free sailing manoeuvring tests for the HTC. This test
campaign resulted in a very valuable data set which can be used for public validation
studies, see chapter 6. Besides obtaining general characteristics of the manoeuvrability of
a single-screw container ship, unique information has been obtained on the drift angles
124 Chapter 8 – Conclusions and recommendations

and rates of turn in combination with propeller and rudder forces. From this, impor-
tant information for the development and validation of manoeuvring prediction tools is
obtained. Furthermore, repeat tests have been conducted for selected manoeuvres and
based on these tests, the uncertainty in the characteristic manoeuvring properties has
been estimated. Even when a small number of observations is available, it is concluded
that the verification procedure proposed in this thesis provides good estimates of the
uncertainty in the measurements, provided a Student t coverage factor is used.
The manoeuvring prediction program SurSim has been used to simulate the manoeu-
vrability of the HTC, see chapter 3. In the program, it is possible to provide user-defined
hydrodynamic coefficients for the bare hull forces and moments. This makes the program
well suited for the present work to investigate whether the use of viscous-flow calcula-
tions can help in improving the prediction accuracy of simulations. A procedure to derive
the coefficients is proposed. This procedure is chosen to enable accurate modelling of
the linearised behaviour for course-keeping as well as realistic modelling of the harbour
manoeuvring characteristics, and to enable accurate modelling of non-linear manoeuvres.
For ship manoeuvres, not only the flow around the ship in oblique motion is of interest,
but also the flow around the ship when it performs a rotational motion. To compute the
flow around the ship for such conditions, the flow solvers used in the present study had
to be modified. For this work, the rotational motion was incorporated by using a non-
inertial reference system and supplementing the equations of motions with body forces
representing the centrifugal and Coriolis contributions on the flow.
To generate the grids for a range of drift angles and yaw rates, different approaches
were adopted depending on the flow solver used. For Parnassos, automated scripts
were developed with which the desired grids could be generated rapidly. For the more
generic solver ReFRESCO, all calculations are conducted using one grid, but changing
the inflow angles and boundary conditions depending on the desired manoeuvring motion.
To facilitate this, a new boundary condition was developed, which removes the need to
pre-process each grid for each new manoeuvring condition.
In chapter 5, it was demonstrated that for a wide range of ship types, good predictions
of the loads on the hull in manoeuvring motion can be obtained. The trends in the forces
and moments as a function of the drift angle or yaw rate are simulated well. Two different
solvers and using different turbulence models were applied. In general, it is concluded
that the differences between the results obtained with the two solvers are relatively small
and within the numerical uncertainties, except for cases where large differences in grid
density were used. In those cases, better results were obviously obtained on the finer
grids.
The verification studies provide useful insight into the influence of the grid density on
the predicted forces and moments. In several cases, validation of the calculations failed,
indicating modelling errors in the numerical results. In these cases, it was generally seen
that the magnitude of the transverse force was under-predicted, while the magnitude of
the yaw moment was over-predicted. For manoeuvring studies in the early design, the
8.2 – Recommendations 125

comparison errors are within acceptable levels. However, improvements remain desired
and may be obtained using finer grids, larger domain sizes, different grid topologies with
refinement in the wake of the ship, other turbulence models or incorporating free surface
deformation.
The verification and validation studies show that the drift-sweep procedure proposed
in this thesis and implemented in ReFRESCO provides good results in comparison
with the Parnassos calculations and with the experiments. By using the procedure,
the amount of manual interaction for the user decreases considerably. Furthermore, the
convergence for the different drift angles is faster than when each drift angle would be
calculated separately.

By using hydrodynamic manoeuvring coefficients derived from CFD calculations of

the forces on the bare hull, it has been shown that it is possible to improve the prediction of
ship manoeuvres compared to predictions using coefficients based on empirical equations,
which was the objective of the present study. In chapter 7, it has been demonstrated
that a considerable improvement of the turning circle predictions was obtained. The
prediction of the yaw checking and course keeping and initial turning abilities based on
zig-zag simulations improved as well, but further improvements are required for more
reliable assessment of the manoeuvring performance.
The sensitivity of the manoeuvring predictions on changes in the hydrodynamic co-
efficients was studied. It was found that the linear coefficients mostly determine the
sensitivity of the results of the zig-zag manoeuvres, while non-linear coefficients influence
mostly the turning circle results. Hull-propeller-rudder coefficients were also found to
be important in the sensitivity study. The study demonstrates that for accurate predic-
tions of the manoeuvrability using coefficients derived from CFD calculations, accurate
predictions of especially the yawing moment must be made.

8.2 Recommendations
To improve the accuracy of manoeuvring simulations and to reduce the uncertainty in the
simulation results due to small changes in the predicted forces and moments on the hull,
the following recommendations are made:

• One of the most promising prospects of the use of viscous-flow solvers is the ability
to estimate the hull forces for full scale conditions. This will eliminate possible
scale effects and improve the correspondence between the predictions and the actual
prototype results. Therefore, calculations for prototype Reynolds numbers should
be made.

• The overall accuracy of force and moments predictions on ship hulls should be im-
proved by investigating in more detail the influence on the predictions of turbulence
models, domain size, grid density and topology.
126 Chapter 8 – Conclusions and recommendations

• From the set of forces and moments predicted by ReFRESCO as presented in

chapter 5 hydrodynamic coefficients for the HTC can be derived. These should be
compared to the coefficients derived from the Parnassos results. By using the
coefficients, manoeuvring predictions can be made, which can be compared to the
simulation results obtained with the Parnassos coefficients. This will give more
insight into the sensitivity of the simulation results to changes in the viscous-flow
results.

• In this thesis, only the bare hull is considered in the viscous-flow calculations. How-
ever, it is possible to include the propeller influence by e.g. using an actuator disc
model to model the propeller thrust, or by calculating the flow around the propeller
with a potential flow code and to introduce the calculated forces on the propeller
as a force field in the RANS calculations. Furthermore, it is possible to include the
rudder in the grid, such that the forces due to the rudder can be obtained as well.
With such computations, the empirical modules for the propeller and rudder can be
substituted by results obtained with the viscous-flow solvers and probably improve
the manoeuvring predictions.

• To prepare for future developments (i.e. increase in computing power), the coupling
of the RANS equations and the equations of motions should be implemented in
ReFRESCO. This will avoid the simplifications made in the virtual captive test
approach, such as quasi-steadiness, and therefore result in a better reliability of
the simulations. In a first stage, this can be done for the bare hull only, with the
appendage forces predicted using e.g. a coupling with SurSim, but fully appended
in a later stage.
 127

References

[1] I. H. Abbott and A. E. von Doenhoff. Theory of Wing Sections: Including a Sum-
mary of Airfoil Data. Dover Publications, Inc., New York, 1959.

[2] A. Abkowitz. Lectures on ship hydrodynamics - steering and maneuverability.

Technical Report Hy-5, Hydro- og Aerodynamisk Laboratorium, Lyngby, Denmark,
1964.

[3] B. Alessandrini and G. Delhommeau. Viscous free surface flow past a ship in drift
and in rotating motion. In 22nd Symposium on Naval Hydrodynamics, pages 491–
507, Fukuoka, Japan, August 1998.

[4] N. Alin, R. E. Bensow, C. Fureby, T. Huuva, and U. Svennberg. Current capabilities

of DES and LES for submarines at straight course. Journal of Ship Research,
54(3):184–196, September 2010.

[5] N. Alin, C. Fureby, and S. U. Svennberg. LES of the flow past simplified submarine
hulls. In 8th International Conference on Numerical Ship Hydrodynamics, Busan,
Korea, September 2003.

[6] V. K. Ankudinov and B. K. Jakobsen. Physically based maneuvering model for

simulations and test evaluations. In MARSIM International Conference on Marine
Simulation and Ship Manoeuvring, number B4, Terschelling, The Netherlands, June
2006.

[7] R. Aris. Vectors, Tensors and the Basic Equations of Fluid Mechanics. Prentice-
Hall, Inc., Englewood Cliffs, N.J., 1962.

[8] E. Armaoğlu, R. Eggers, F. H. H. A. Quadvlieg, and P. van Coevorden. On the

manoeuvrablity of naval surface ships with respect to a new STANAG. In RINA
WARSHIP 2010: Advanced Technologies in Naval Design and Construction, Lon-
don, UK, June 2010.

[9] ASME. Standard for verification and validation in computational fluid dynamics
and heat transfer. American Society of Mechanical Engineers (ASME) V&V 20
Committee, November 2009.
128 References

[10] P. Atsavapranee, T. Forlini, D. Furey, J. Hamilton, S. Percival, and C.-H. Sung. Ex-
perimental measurements for CFD validation of the flow about a submarine model
(ONR Body-1). In 25th Symposium on Naval Hydrodynamics, St. Johns, Newfound-
land and Labrador, Canada, August 2004.

[11] G. K. Batchelor. An Introduction to Fluid Mechanics. Cambridge University Press,

1967. ISBN 0 521 66396 2.

[12] D. Bellevre, A. Diaz de Tuesta, and P. Perdon. Submarine maneuverability assess-

ment using CFD. In 23rd Symposium on Naval Hydrodynamics, pages 820–832, Val
de Reuil, France, September 2000.

[13] V. Bertram. Practical Ship Hydrodynamics. Butterworth-Heinemann, 2000.

[14] M. C. Bettle, A. G. Gerber, and G. D. Watt. Unsteady analysis of the six DOF
motion of a buoyantly rising submarine. Computers & Fluids, 38(9):1833–1849,
October 2009.

[15] M. C. Bettle, S. L. Toxopeus, and A. G. Gerber. Calculation of bottom clearance

effects on Walrus submarine hydrodynamics. International Shipbuilding Progress,
57(3–4):101–125, 2010.

[16] S. Bhushan, T. Xing, P. Carrica, and F. Stern. Model- and full-scale URANS simu-
lations of Athena resistance, powering, seakeeping, and 5415 maneuvering. Journal
of Ship Research, 53(4):179–198, December 2009.

[17] R. Broglia, R. Muscari, and A. Di Mascio. Numerical analysis of blockage effects

in PMM tests. In 26th Symposium on Naval Hydrodynamics, Rome, Italy, Septem-
ber 17-22 2006.

[18] G. Bulian, R. G. Nicolosi, and A. Francescutto. On the effect of uncertainty mod-

eling in the hydrodynamic derivatives of a ship manoeuvring mathematical model.
In 7th ICHD International Conference on Hydrodynamics, pages 359–368, Ischia,
Italy, October 2006.

[19] P. W. Bull. The validation of CFD predictions of nominal wake for the SUBOFF
fully appended geometry. In 21st Symposium on Naval Hydrodynamics, pages 1061–
1076, Trondheim, Norway, June 1996.

[20] P. W. Bull and S. Watson. The scaling of high Reynolds number viscous flow
predictions for appended submarine geometries. In 22nd Symposium on Naval Hy-
drodynamics, pages 1000–1014, Fukuoka, Japan, August 1998.

[21] E. F. Campana, A. Di Mascio, and R. Penna. CFD analysis of the flow past a ship
in steady drift motion. In 3rd Osaka Colloquium on advanced CFD applications to
ship flow and hull form design, pages 151–159, Osaka, Japan, May 1998.
References 129

[22] P. M. Carrica, F. Ismail, M. Hyman, S. Bhushan, and F. Stern. Turn and zigzag
maneuvers of a surface combatant using a URANS approach with dynamic overset
grids. In SIMMAN Workshop on Verification and Validation of Ship Manoeuvring
Simulation Methods, pages F17–22, Copenhagen, Denmark, April 2008.

[23] P. M. Carrica and F. Stern. DES simulations of KVLCC1 in turn and zigzag maneu-
vers with moving propeller and rudder. In SIMMAN Workshop on Verification and
Validation of Ship Manoeuvring Simulation Methods, pages F11–16, Copenhagen,
Denmark, April 2008.

[24] C. J. Chesnakas and R. L. Simpson. Full three-dimensional measurements of the

cross-flow separation region of a 6:1 prolate spheroid. Experiments in Fluids, 17(1–
2):68–74, June 1994.

[25] D. Clarke and A. P. Mesbahi, editors. 24th International Towing Tank Conference,
Edinburgh, United Kingdom, September 2005.

[26] C. L. Crane. Maneuvering trials of the 278,000 DWT Esso Osaka in shallow and
deep water. In Transactions of the SNAME, volume 87, pages 251–283, 1979.

[27] A. Cura Hochbaum. Computation of the turbulent flow around a ship model in
steady turn and in steady oblique motion. In 22nd Symposium on Naval Hydrody-
namics, pages 550–567, Fukuoka, Japan, August 1998.

[28] A. Cura Hochbaum. Virtual PMM tests for manoeuvring prediction. In 26th Sym-
posium on Naval Hydrodynamics, Rome, Italy, September 17-22 2006.

[29] A. Cura Hochbaum and M. Vogt. On the prediction of hydrodynamic forces on

a manoeuvring ship. In MARSIM International Conference on Marine Simulation
and Ship Manoeuvring, pages RC–20–1–7, Kanazawa, Japan, August 2003. The
Society of Naval Architects of Japan and Japan Institute of Navigation.

[30] A. Cura Hochbaum, M. Vogt, and S. Gatchell. Manoeuvring prediction for two
tankers based on RANS simulations. In SIMMAN Workshop on Verification and
Validation of Ship Manoeuvring Simulation Methods, pages F23–28, Copenhagen,
Denmark, April 2008.

[31] J. Dacles-Mariani, G. G. Zilliac, J. S. Chow, and P. Bradshaw. Numeri-

cal/experimental study of a wingtip vortex in the near field. AIAA Journal, 33:1561–
1568, September 1995.

[32] I. Dand. Low-speed manoeuvring criteria: some considerations. In MARSIM Inter-

national Conference on Marine Simulation and Ship Manoeuvring, pages KN–4–1–
8, Kanazawa, Japan, August 2003. The Society of Naval Architects of Japan and
Japan Institute of Navigation.
130 References

[33] J. A. DeMoss. Drag measurements on an ellipsoidal body. Master’s thesis, Virginia

Polytechnic Institute and State University, Blacksburg, Virginia, August 2007.

[34] E. M. Dempsey. Experimental investigation of the stability and control character-

istics of the Walrus class submarine. Technical report, David W. Taylor Naval Ship
Research and Development Center, November 1980.

[35] A. Di Mascio, R. Broglia, and R. Muscari. Unsteady RANS simulation of a ma-

noeuvring ship hull. In 25th Symposium on Naval Hydrodynamics, pages 9–18, St.
Johns, Newfoundland and Labrador, Canada, August 2004.

[36] A. Di Mascio and E. F. Campana. The numerical simulation of the yaw flow of a free
surface ship. In 7th International Conference on Numerical Ship Hydrodynamics,
pages 6.1–1–10, Nantes, France, July 1999.

[37] A. Drouet, E. Jacquin, B. Alessandrini, P. Ferrant, L. Gentaz, C. Monroy, J. M.

Rousset, P. E. Guillerm, and P. Perdon. Simulation of unsteady ship maneuvering
on calm water and in waves using free-surface RANS solver. In 27th Symposium on
Naval Hydrodynamics, volume 2, B29, Seoul, Korea, October 5–10 2008.

[38] L. Eça and M. Hoekstra, editors. Workshop on CFD Uncertainty Analysis, Lisbon,
Portugal, October 2004. Instituto Superior Técnico.

[39] L. Eça and M. Hoekstra. On the influence of grid topology on the accuracy of ship
viscous flow calculations. In 5th Osaka Colloquium on Advanced CFD Applications
to Ship Flow and Hull Form Design, pages 1–10, March 2005.

[40] L. Eça and M. Hoekstra. The numerical friction line. Journal of Marine Science
and Technology, 13(4):328–345, November 2008.

[41] L. Eça and M. Hoekstra. Evaluation of numerical error estimation based on grid
refinement studies with the method of the manufactured solutions. Computers &
Fluids, 38(8):1580–1591, September 2009.

[42] L. Eça and M. Hoekstra. On the numerical accuracy of the prediction of resis-
tance coefficients in ship stern flow calculations. Journal of Marine Science and
Technology, 14(1):2–18, March 2009.

[43] L. Eça, M. Hoekstra, and J. Windt. Practical grid generation tools with applications
to ship hydrodynamics. In 7th International Conference on Grid Generation and
Computational Field Simulations, Hawaii, February 2002.

[44] L. Eça, G. N. V. B. Vaz, and M. Hoekstra. Code verification, solution verification

and validation in RANS solvers. In 29th International Conference on Ocean, Off-
shore and Arctic Engineering OMAE, number OMAE2010-20338, Shanghai, China,
June 2010.
References 131

[45] L. Eça, G. N. V. B. Vaz, and M. Hoekstra. A verification and validation exer-

cise for the flow over a backward facing step. In Fifth European Conference on
Computational Fluid Dynamics, Lisbon, Portugal, June 2010. ECCOMAS.

[46] O. M. El Moctar. Numerical computations of flow forces in ship manoeuvring.

Schiffstechnik / Ship Technology Research, 48:98–123, 2001.

[47] K. Eloot. Selection, Experimental Determination and Evaluation of a Mathematical

Model for Ship Manoeuvring in Shallow Water. PhD thesis, Universiteit Gent, 2006.

[48] K. Eloot and M. Vantorre. Development of a tabular manoeuvring model for hull
forces applied to full and slender ships in shallow water. In MARSIM Interna-
tional Conference on Marine Simulation and Ship Manoeuvring, pages RC–18–1–9,
Kanazawa, Japan, August 2003. The Society of Naval Architects of Japan and Japan
Institute of Navigation.

[49] A. Etebari, P. Atsavapranee, J. B. Carneal, A. S. Percival, D. J. Grant, C.-H. Sung,

and I.-Y. Koh. Experimental measurements on a SUBOFF model in a turning
maneuver. In 27th Symposium on Naval Hydrodynamics, volume 2, B30, Seoul,
Korea, October 5–10 2008.

[50] L. Fabbri, L. Benedetti, B. Bouscasse, F. L. Gala, and C. Lugni. An experimental

study of the maneuverability of a blunt ship: the effect of the water depth. In 9th
NuTTS Numerical Towing Tank Symposium, Le Croisic, France, October 2006.

[51] F. Fathi, C. M. Klaij, and A. H. Koop. Predicting loads on a LNG carrier with
CFD. In 29th International Conference on Ocean, Offshore and Arctic Engineering
OMAE, number OMAE2010-20122, Shanghai, China, June 2010.

[52] T. I. Fossen. Guidance and Control of Ocean Vehicles. John Wiley and Sons, 1994.

[53] C. Fureby. Towards the use of large eddy simulation in engineering. In 46th AIAA
Aerospace Sciences Meeting and Exhibit, number AIAA 2008-605, Reno, Nevada,
January 2008. American Institute of Aeronautics and Astronautics.

[54] Q. Gao and D. Vassalos. Computational hydrodynamic derivatives by numerical

PMM. In Marine CFD, Southampton, UK, March 2008. Royal Institution of Naval
Architects.

[55] K. Granlund. Steady and Unsteady Maneuvering Forces and Moments on Slender
Bodies. PhD thesis, Virginia Polytechnic Institute and State University, February
2009.

[56] K. Granlund and R. L. Simpson. Modeling unsteady maneuvers on slender bodies. In

AIAA Atmospheric Flight Mechanics Conference and Exhibit, number AIAA 2007-
6721, Hilton Head, South Carolina, August 2007. American Institute of Aeronautics
and Astronautics.
132 References

[57] N. C. Groves, T. T. Huang, and M. S. Chang. Geometric characteristics of DARPA

SUBOFF models (DTRC Model Nos. 5470 and 5471). Report No. DTRC/SHD-
1298-01, March 1998.

[58] T. Hino, editor. CFD Workshop Tokyo, Tokyo, Japan, March 2005. National Mar-
itime Research Institute.

[59] M. Hoekstra. Numerical Simulation of Ship Stern Flows with a Space-Marching

Navier-Stokes Method. PhD thesis, Delft University of Technology, Faculty of Me-
chanical Engineering and Marine Technology, October 1999.

[60] M. Hoekstra and L. Eça. PARNASSOS: an efficient method for ship stern flow
calculation. In 3rd Osaka Colloquium on advanced CFD applications to ship flow
and hull form design, pages 331–357, Osaka, May 1998.

[61] M. Hoekstra, L. Eça, J. Windt, and H. C. Raven. Viscous flow calculations for
KVLCC2 and KCS models using the PARNASSOS code. Gothenburg 2000 Work-
shop on Numerical Ship Hydrodynamics, 2000.

[62] J. Holtrop and G. G. J. Mennen. An approximate power prediction method. Inter-

national Shipbuilding Progress, 29:166–170, July 1982.

[63] J. P. Hooft. The cross flow drag on a manoeuvring ship. Ocean Engineering,
21(3):329–342, April 1994.

[64] J. P. Hooft and U. Nienhuis. The prediction of the ship’s manoeuvrability in the
design stage. SNAME Annual Meeting, 102:1–24, November 1994.

[65] J. P. Hooft and J. B. M. Pieffers. Maneuverability of frigates in waves. Marine

Technology, 25(4):262–271, October 1988.

[66] J. P. Hooft and F. H. H. A. Quadvlieg. Non-linear hydrodynamic hull forces de-

rived from segmented model tests. In MARSIM International Conference on Ma-
rine Simulation and Ship Manoeuvrability, pages 399–409, Copenhagen, Denmark,
September 1996.

[67] S. Hosder. Unsteady skin-friction measurements on a maneuvering Darpa2 Suboff

model. Master’s thesis, Virginia Polytechnic Institute and State University, June
2001.

[68] T. T. Huang, H.-L. Liu, N. Groves, T. Forlini, J. Blanton, and S. Gowing. Mea-
surements of flows over an axisymmetric body with various appendages in a wind
tunnel: the DARPA SUBOFF experimental program. In 19th Symposium on Naval
Hydrodynamics, pages 312–346, Seoul, South Korea, August 1992.

[69] IMO. Resolution MSC. 137(76): Standards for ship manoeuvrability. International
Maritime Organisation, London, 2002.
References 133

[70] S. Inoue, M. Hirano, and K. Kijima. Hydrodynamic derivatives on ship manoeu-

vring. International Shipbuilding Progress, 28(321):112–125, 1981.

[71] T. Ishiguro, S. Tanaka, and Y. Yoshimura. A study on the accuracy of the recent
prediction technique of ship’s manoeuvrability at early design stage. In MARSIM
International Conference on Marine Simulation and Ship Manoeuvring, pages 547–
561, Copenhagen, Denmark, September 1996.

[72] Guide to the expression of uncertainty in measurement. International Organization

for Standardization (ISO), 1995.

[73] ITTC. Testing and extrapolation methods manoeuvrability - free running model
tests. Recommended Procedures and Guidelines, 7.5-02, 06-01, September 2008.

[74] G. Jensen, M. Klemt, and Y. Xing-Kaeding. On the way to the numerical basin for
seakeeping and manoeuvring. In 9th Symposium on Practical Design of Ships and
Other Floating Structures (PRADS), number 74, Lübeck-Travemünde, Germany,
September 2004.

[75] R. Jonnalagadda, L. K. Taylor, and D. L. Whitfield. Multiblock multigrid incom-

pressible RANS computation of forces and moments on appended SUBOFF con-
figurations at incidence. In 35th Aerospace Sciences Meeting and Exhibit, number
AIAA 1997-624, Reno, Nevada, January 1997.

[76] D. Kang and K. Hasegawa. Prediction method of hydrodynamic forces acting on

the hull of a blunt-body ship in the even keel condition. Journal of Marine Science
and Technology, 12(1):1–14, March 2007.

[77] W. J. Kim, S. H. Van, and D. H. Kim. Measurement of flows around modern

commercial ship models. Experiments in Fluids, 31(5):567–578, November 2001.

[78] J. C. Kok. Resolving the dependence on free-stream values for the k − ω turbulence
model. National Aerospace Laboratory, The Netherlands, Report NLR-TP-99295,
July 1999.

[79] A. H. Koop, C. M. Klaij, and G. N. V. B. Vaz. Predicting wind shielding for FPSO
tandem offloading using CFD. In 29th International Conference on Ocean, Offshore
and Arctic Engineering OMAE, number OMAE2010-20284, Shanghai, China, June
2010.

[80] A. M. Kracht. On propeller-rudder interaction. In Intemational Symposium on

Propulsors and Cavitation, Hamburg, Germany, June 22–25 1992.

[81] G. Kuiper. Resistance and propulsion of ships. Delft University of Technology, July
1991.
134 References

[82] K. Kume, J. Hasegawa, Y. Tsukada, J. Fujisawa, R. Fukasawa, and M. Hinatsu.

Measurements of hydrodynamic forces, surface pressure, and wake for obliquely
towed tanker model and uncertainty analysis for CFD validation. Journal of Marine
Science and Technology, 11(2):65–75, June 2006.

[83] L. Larsson, F. Stern, and V. Bertram. Benchmarking of computational fluid dy-

namics for ship flows: The Gothenburg 2000 workshop. Journal of Ship Research,
47(1):63–81, March 2003.

[84] H.-Y. Lee and S.-S. Shin. The prediction of ship’s manoeuvring performance in
initial design stage. In Practical Design of Ships and Mobile Units (PRADS), pages
633–639, The Hague, The Netherlands, September 1998.

[85] S.-J. Lee, H.-R. Kim, W.-J. Kim, and S.-H. Van. Wind tunnel tests on flow charac-
teristics of the KRISO 3,600 TEU containership and 300k VLCC double-deck ship
models. Journal of Ship Research, 47(1):24–38, March 2003.

[86] S. K. Lee and M. Fujino. Assessment of a mathematical model for the manoeuvring
motion of a twin-propeller twin-rudder ship. International Shipbuilding Progress,
50(1/2):109–123, 2003.

[87] S. W. Lee. Free sailing manoeuvring tests on KVLCC1 and KVLCC2. MARIN
Report No. 21571-1-SMB, Wageningen, The Netherlands, December 2007.

[88] H.-L. Liu and T. T. Huang. Summary of DARPA SUBOFF experimental program
data. Report No. CRDKNSWC/HD-1298-11, June 1998.

[89] J. F. Longo. Effects of Yaw on Model-Scale Ship Flows. PhD thesis, University of
Iowa, May 1996.

[90] S. C. McParlin and R. W. Tramel. Taxonomy of flight mechanics issues for aircraft,
and underlying fluid dynamics phenomena. In 47th AIAA Aerospace Sciences Meet-
ing Including The New Horizons Forum and Aerospace Exposition, number AIAA
2009-744, Orlando, Florida, January 2009. American Institute of Aeronautics and
Astronautics.

[91] F. R. Menter. Two-equation eddy-viscosity turbulence models for engineering ap-

plications. AIAA Journal, Vol. 32(8):1598–1605, August 1994.

[92] F. R. Menter. Eddy viscosity transport equations and their relation to the k − 
model. Journal of Fluids Engineering, 119(4):876–884, December 1997.

[93] R. W. Miller. PMM calculations for the bare and appended DTMB 5415 using the
RANS solver CFDShip-Iowa. In SIMMAN Workshop on Verification and Validation
of Ship Manoeuvring Simulation Methods, pages F41–49, Copenhagen, Denmark,
April 2008.
References 135

[94] A. F. Molland and S. R. Turnock. Wind tunnel investigation of the influence of

propeller loading on ship rudder performance. Transactions RINA, 135:105–120,
1993.

[95] R. Muscari, R. Broglia, and A. Di Mascio. Trajectory prediction of a self propelled

hull by unsteady RANS computations. In 27th Symposium on Naval Hydrodynamics,
volume 2, B31, Seoul, Korea, October 5–10 2008.

[96] NATO. ANEP-70 Guidance for naval surface ships mission oriented manoeuvring re-
quirements. Naval Group 6 Specialist Team on Naval Ship Manoeuvrability, Septem-
ber 2003.

[97] N. H. Norrbin. Theory and observations on the use of a mathematical model for
ship maneuvering in deep and confined waters. In 8th Symposium on Naval Hydro-
dynamics, pages 807–904, Pasadena, California, August 1970.

[98] A. Ogawa and H. Kasai. On the mathematical model of manoeuvring motion of

ships. International Shipbuilding Progress, 25(292):306–319, December 1978.

[99] T. Ohmori. Finite-volume simulations of flows about a ship in manoeuvring motion.

Journal of Marine Science and Technology, 3(2):82–93, June 1998.

[100] T. Ohmori, M. Fujino, and H. Miyata. A study on flow field around full ship forms
in manoeuvring motion. Journal of Marine Science and Technology, 3(1):22–29,
March 1998.

[101] P. Oltmann and S. D. Sharma. Simulation of combined engine and rudder ma-
noeuvres using an improved model of hull-propeller-rudder interactions. In 15th
Symposium on Naval Hydrodynamics, pages 83–108, Hamburg, Germany, Septem-
ber 1984.

[102] M. W. C. Oosterveld and P. van Oossanen. Further computer analyzed data of the
Wageningen B-screw series. International Shipbuilding Progress, 22(251):251–262,
July 1975.

[103] M. Örnfelt. Naval mission and task driven manoeuvrability requirements for naval
ships. In 10th International Conference on Fast Sea Transportation (FAST), pages
505–518, Athens, Greece, October 2009.

[104] B. Overpelt. Manoeuvring tests on KCS. MARIN Report No. 23991-1-SMB, Wa-
geningen, The Netherlands, September 2010.

[105] B. Overpelt and S. L. Toxopeus. Manoeuvring tests on HTC. MARIN Report No.
23277-3-SMB, Wageningen, The Netherlands, January 2011.
136 References

[106] R. Pankajakshan, M. G. Remotigue, L. K. Taylor, M. Jiang, W. R. Bridley, and D. L.

Whitfield. Validation of control-surface induced submarine maneuvering simulations
using UNCLE. In 24th Symposium on Naval Hydrodynamics, pages 83–97, Fukuoka,
Japan, July 2002.

[107] A. B. Phillips, S. R. Turnock, and M. Furlong. Evaluation of manoeuvring coef-

ficients of a self-propelled ship using a blade element momentum propeller model
coupled to a Reynolds averaged Navier Stokes flow solver. Ocean Engineering,
36(15–16):1217–1225, November 2009.

[108] J. E. Poremba. Hydrodynamics and maneuvering simulations of a non-body-of-

revolution submarine. Master’s thesis, Pennsylvania State University, December
2009.

[109] F. H. H. A. Quadvlieg. A new combined seakeeping and manoeuvring basin for

the third millennium maritime research. In MARSIM International Conference on
Marine Simulation and Ship Manoeuvring, Orlando, Florida, May 2000.

[110] F. H. H. A. Quadvlieg, E. Armaoğlu, R. Eggers, and P. van Coevorden. Prediction

and verification of the maneuverability of naval surface ships. In SNAME Annual
Meeting and Expo, Seattle/Bellevue, WA, November 2010.

[111] F. H. H. A. Quadvlieg and P. van Coevorden. Manoeuvring criteria: more than IMO
A751 requirements alone! In MARSIM International Conference on Marine Sim-
ulation and Ship Manoeuvring, pages RB–1–1–8, Kanazawa, Japan, August 2003.
The Society of Naval Architects of Japan and Japan Institute of Navigation.

[112] B. J. Racine and E. G. Paterson. CFD-based method for simulation of marine-

vehicle maneuvering. In 35th AIAA Fluid Dynamics Conference and Exhibit, num-
ber AIAA 2005-4904, Toronto, Ontario, Canada, June 2005. American Institute of
Aeronautics and Astronautics.

[113] H. C. Raven, A. van der Ploeg, and L. Eça. Extending the benefit of CFD tools in
ship design and performance prediction. In 7th ICHD International Conference on
Hydrodynamics, pages 573–580, October 2006.

[114] H. C. Raven, A. van der Ploeg, and A. R. Starke. Computation of free-surface vis-
cous flows at model and full scale by a steady iterative approach. In 25th Symposium
on Naval Hydrodynamics, August 2004.

[115] P. J. Roache. Verification and Validation in Computational Science and Engineer-

ing. Hermosa Publishers, Albuquerque, New Mexico, 1998.

[116] R. F. Roddy. Investigation of the stability and control characteristics of several

configurations of the DARPA SUBOFF model (DTRC Model 5470) from captive-
model experiments. Report No. DTRC/SHD-1298-08, September 1990.
References 137

[117] R. F. Roddy, R. M. Curphey, and K. P. Delaney. Advances in understanding of

maneuvering characteristics for non-body-of-revolution submersibles. In 27th Sym-
posium on Naval Hydrodynamics, volume 2, B36, Seoul, Korea, October 5–10 2008.

[118] N. Sakamoto. URANS and DES Simulations Of Static And Dynamic Maneuvering
For Surface Combatant. PhD thesis, University of Iowa, Iowa City, Iowa, May 2009.

[119] N. Sakamoto, P. M. Carrica, and F. Stern. URANS simulations of static and dy-
namic maneuvering for surface combatant. In SIMMAN Workshop on Verification
and Validation of Ship Manoeuvring Simulation Methods, pages F50–55, Copen-
hagen, Denmark, April 2008.

[120] M. D. Salas. Digital flight: The last CFD aeronautical grand challenge. Journal of
Scientific Computing, 28(213):479–505, September 2006.

[121] T. Sato, K. Izumi, and H. Miyata. Numerical simulation of maneuvering motion. In

22nd Symposium on Naval Hydrodynamics, pages 724–737, Fukuoka, Japan, August
1998.

[122] C. D. Simonsen. Rudder, Propeller and Hull Interaction by RANS. PhD thesis,
Technical University of Denmark, Department of Naval Architecture and Offshore
Engineering, May 2000.

[123] C. D. Simonsen and F. Stern. Verification and validation of RANS maneuvering

simulation of Esso Osaka: effects of drift and rudder angle on forces and moments.
Computers & Fluids, 32(10):1325–1356, December 2003.

[124] C. D. Simonsen and F. Stern. Flow pattern around an appended tanker hull form
in simple maneuvering conditions. Computers & Fluids, 34(2):169–198, February
2005.

[125] C. D. Simonsen and F. Stern. RANS maneuvering simulation of esso osaka with
rudder and a body-force propeller. Journal of Ship Research, 49(2):98–120, June
2005.

[126] C. D. Simonsen and F. Stern. Flow structure around maneuvering tanker in deep
and shallow water. In 26th Symposium on Naval Hydrodynamics, number M1, Rome,
Italy, September 17–22 2006.

[127] C. D. Simonsen, F. Stern, and K. Agdrup. CFD with PMM test validation for
manoeuvring VLCC2 tanker in deep and shallow water. In MARSIM Interna-
tional Conference on Marine Simulation and Ship Manoeuvring, number M04, Ter-
schelling, The Netherlands, June 2006.

[128] H. Söding. Forces on rudders behind a manoeuvring ship. In Third Symposium on

Numerical Ship Hydrodynamics, 1980.
138 References

[129] P. R. Spalart and S. R. Allmaras. A one-equatlon turbulence model for aerodynamic

flows. In 30th Aerospace Sciences Meeting & Exhibit, Reno, Nevada, January 1992.
American Institute of Aeronautics and Astronautics.

[130] F. Stern and K. Agdrup, editors. SIMMAN Workshop on Verification and Validation
of Ship Manoeuvring Simulation Methods, Copenhagen, Denmark, April 2008.

[131] F. Stern, K. Agdrup, S. Y. Kim, A. Cura Hochbaum, K. P. Rhee, F. H. H. A.

Quadvlieg, P. Perdon, T. Hino, R. Broglia, and J. Gorski. Lessons learnt from the
workshop on verification and validation of ship manoeuvring simulation methods
- SIMMAN 2008. In International Conference on Marine Simulation and Ship
Maneuverability, MARSIM, pages M–24–1–10, Panama City, Panama, August 2009.

[132] F. Stern, F. Ismail, T. Xing, and P. Carrica. Vortical and turbulent structures
using various convection schemes with algebraic reynolds stress-DES model for the
KVLCC2 at large drift angles. In 27th Symposium on Naval Hydrodynamics, volume
2, B34, Seoul, Korea, October 5–10 2008.

[133] F. Stern, R. V. Wilson, H. W. Coleman, and E. G. Paterson. Comprehensive

approach to verification and validation of CFD simulations - part 1: Methodology
and procedures. Journal of Fluids Engineering, 123(4):793–802, December 2001.

[134] J. Strom-Tejsen and R. R. Porter. Prediction of controllable pitch propeller perfor-

mance in off-design conditions. In Third Ship Control Systems Symposium, 1972.

[135] C.-H. Sung, T.-C. Fu, and M. Griffin. Validation of incompressible flow computation
of forces and moments on axisymmetric bodies undergoing constant radius turning.
In 21st Symposium on Naval Hydrodynamics, pages 1048–1060, Trondheim, Norway,
June 1996.

[136] C.-H. Sung, T. C. Fu, M. J. Griffin, and T. T. Huang. Validation of incompress-

ible flow computation of forces and moments on axisymmetric bodies at incidence.
In 33rd Aerospace Sciences Meeting and Exhibit, number AIAA 1995-528, Reno,
Nevada, January 1995.

[137] C.-H. Sung, M. J. Griffin, J. F. Tsai, and T. T. Huang. Incompressible flow computa-
tion of forces and moments on bodies of revolution at incidence. In 31st Aerospace
Sciences Meeting and Exhibit, number AIAA 1993-787, Reno, Nevada, January
1993.

[138] C.-H. Sung, M.-Y. Jiang, B. Rhee, S. Percival, P. Atsavapranee, and I.-Y. Koh.
Validation of the flow around a turning submarine. In 24th Symposium on Naval
Hydrodynamics, Fukuoka, Japan, July 2002.

[139] C.-H. Sung, B. Rhee, and E. Ammeen. SHWG CFD benchmark study on DARPA
SUBOFF. Part 1. Bare hull. SHWG Correspondence, pages 1–26, November 2005.
References 139

[140] Y. Tahara, J. F. Longo, and F. Stern. Comparison of CFD and EFD for the Series
60 CB =0.6 in steady drift motion. Journal of Marine Science and Technology,
7(1):17–30, August 2002.

[141] S. L. Toxopeus. Simulation and validation of the viscous flow around the Series 60
hull form at 10◦ drift angle. In 7th NuTTS Numerical Towing Tank Symposium,
Hamburg, Germany, October 2004.

[142] S. L. Toxopeus. Validation of calculations of the viscous flow around a ship in

oblique motion. In The First MARIN-NMRI Workshop, pages 91–99, Tokyo, Japan,
October 2004.

[143] S. L. Toxopeus. Verification and validation of calculations of the viscous flow around
KVLCC2M in oblique motion. In 5th Osaka Colloquium on Advanced CFD Appli-
cations to Ship Flow and Hull Form Design, pages 200–209, Osaka, Japan, March
2005.

[144] S. L. Toxopeus. Calculation of hydrodynamic manoeuvring coefficients using

viscous-flow calculations. In 7th ICHD International Conference on Hydrodynamics,
pages 493–502, Ischia, Italy, October 2006.

[145] S. L. Toxopeus. Validation of slender-body method for prediction of linear ma-

noeuvring coefficients using experiments and viscous-flow calculations. In 7th ICHD
International Conference on Hydrodynamics, pages 589–598, Ischia, Italy, October
2006.

[146] S. L. Toxopeus, editor. SHWG collaborative CFD excercise - Bare hull DARPA
SUBOFF submarine at straight flight and drift angle. MARIN Report 21668-1-CPM,
November 2007.

[147] S. L. Toxopeus. Viscous-flow calculations for bare hull DARPA SUBOFF submarine
at incidence. International Shipbuilding Progress, 55(3):227–251, December 2008.

[148] S. L. Toxopeus. Deriving mathematical manoeuvring models for bare ship hulls using
viscous flow calculations. Journal of Marine Science and Technology, 14(1):30–38,
March 2009.

[149] S. L. Toxopeus and S. W. Lee. Comparison of manoeuvring simulation programs

for SIMMAN test cases. In SIMMAN Workshop on Verification and Validation
of Ship Manoeuvring Simulation Methods, pages E56–61, Copenhagen, Denmark,
April 2008.

[150] S. L. Toxopeus and S. W. Lee. Free sailing tests for European Destroyer. MARIN
Report No. 22367-5-CPM, Wageningen, The Netherlands, February 2009.
140 References

[151] S. L. Toxopeus and G. N. V. B. Vaz. Calculation of current or manoeuvring forces

using a viscous-flow solver. In 28th International Conference on Ocean, Offshore and
Arctic Engineering OMAE, number OMAE2009-79782, Honolulu, Hawaii, May 31–
June 5 2009.

[152] S. L. Toxopeus (editor). Calculation and validation of hydrodynamic derivatives.

Virtue Deliverable D3.1.1, Revision 4, April 2009.

[153] M. Ueno, Y. Yoshimura, Y. Tsukada, and H. Miyazaki. Circular motion tests

and uncertainty analysis for ship maneuverability. Journal of Marine Science and
Technology, 14(4):469–484, December 2009.

[154] A. van der Ploeg and H. C. Raven. CFD-based optimization for minimal power and
wake field quality. In 11th International Symposium on Practical Design of Ships
and Other Floating Structures (PRADS), Rio de Janeiro, Brazil, September 2010.

[155] B. J. van Oers and S. L. Toxopeus. On the relation between flow behaviour and the
lateral force distribution acting on a ship in oblique motion. In 10th International
Cooperation on Marine Engineering Systems ICMES, London, UK, March 2006.

[156] D. Vassalos, S. H. Park, and B. S. Lee. Developing a manoeuvring design capability

for naval vessels. In The Sixth International Symposium On Practical Design Of
Ships And Mobile Units (PRADS), pages 1.616–1.628, Seoul, Korea, September
1995.

[157] G. N. V. B. Vaz, F. A. P. Jaouen, and M. Hoekstra. Free-surface viscous flow compu-

tations. Validation of URANS code FreSCo. In 28th International Conference on
Ocean, Offshore and Arctic Engineering OMAE, number OMAE2009-79398, Hon-
olulu, Hawaii, May 31–June 5 2009.

[158] G. N. V. B. Vaz, S. L. Toxopeus, and S. Holmes. Calculation of manoeuvring forces

on submarines using two viscous-flow solvers. In 29th International Conference on
Ocean, Offshore and Arctic Engineering OMAE, number OMAE2010-20373, Shang-
hai, China, June 2010.

[159] G. N. V. B. Vaz, O. Waals, F. Fathi, H. Ottens, T. Le Souef, and K. Kwong. Current

Affairs - model tests, semi-empirical predictions and CFD computations for current
coefficients of semi-submersibles. In 28th International Conference on Ocean, Off-
shore and Arctic Engineering OMAE, number OMAE2009-80216, Honolulu, Hawaii,
May 31–June 5 2009.

[160] G. Venkatesan and W. Clark. Submarine maneuvering simulations of ONR Body

1. In 26th International Conference on Offshore Mechanics and Arctic Engineering
OMAE, number OMAE2007-29516, San Diego, California, June 10–15 2007.
References 141

[161] M. Vogt, A. Cura Hochbaum, M. Schneider, and J. Dautel. Report on model tests.
Virtue Deliverable D3.1.3, December 2007.

[162] P. Wesseling. Principles of Computational Fluid Dynamics. Springer-Verlag, 2000.

ISBN 3-540-67853-0.

[163] L. F. Whicker and L. F. Fehlner. Free-stream characteristics of a family of low-

aspect-ratio all-movable control surfaces for application to ship design. Technical
Report DTNSRDC Report No. 933, David W. Taylor Naval Ship R&D Center,
Bethesda, MD., May 1958.

[164] C. C. Whitfield. Steady and unsteady force and moment data on a DARPA2 sub-
marine. Master’s thesis, Virginia Polytechnic Institute and State University, August
1999.

[165] B.-S. Wu, F. Xing, X.-F. Kuang, and Q.-M. Miao. Investigation of hydrodynamic
characteristics of submarine moving close to the sea bottom with CFD methods.
Journal of Ship Mechanics, 9(3):19–28, June 2005.

[166] C. Yang and R. Löhner. Prediction of flows over an axisymmetric body with ap-
pendages. In 8th International Conference on Numerical Ship Hydrodynamics, Bu-
san, Korea, September 2003.

[167] H. Yoon. Phase-Averaged Stereo-PIV Flow Field And Force/Moment/Motion Mea-

surements For Surface Combatant In PMM Maneuvers. PhD thesis, University of
Iowa, Iowa City, Iowa, December 2009.
142 References

Page intentionally left blank

 143

Samenvatting
Praktische toepassing van viskeuze omstromingsberekeningen voor
het simuleren van manoeuvrerende schepen

Figure 8.1: Omstroming en druk op het rompoppervlak, KVLCC2, β = −14◦

Het werk dat beschreven is in deze scriptie is gestart om traditionele manoeuvreer-

simulaties gebaseerd op empirische modellen van de krachten en moment op het schip te
verbeteren. Omdat de nauwkeurigheid van viskeuze-omstroming berekeningen om romp-
krachten te voorspellen steeds toeneemt, is besloten om een praktische methode voor het
simuleren van de scheepsmanoeuvreerbaarheid te ontwikkelen, waarbij viskeuze reken-
technieken gebruikt worden. Het doel is te onderzoeken of hiermee de nauwkeurigheid
van manoeuvreerpredicties verbeterd kan worden.
Om dit doel te bereiken, is de virtual captive test aanpak gebruikt, vanwege het ef-
ficiënte gebruik van rekenkracht van deze aanpak in vergelijking met andere methoden.
Deze procedure bootst de traditionele manoeuvreersimulaties na, waarbij experimentele
PMM wordt gebruikt om de krachten en momenten op het schip te bepalen. De huidige
studie breidt het werk van andere onderzoekers uit, door middel van een uitgebreide
verificatie en validatie van de voorspelde krachten en momenten op de romp en een gede-
tailleerde gevoeligheidsstudie van de manoeuvreereigenschappen van het schip op veran-
144 Samenvatting

deringen in de hydrodynamische coëfficiënten in het simulatiemodel.

Om de stroming rond schepen in gierbeweging te kunnen berekenen, waren aanpassin-
gen in de rekenprogramma’s nodig. Deze veranderingen worden in dit werk beschreven,
samen met technieken om de benodigde tijd voor gridgenerering te verkleinen en om
sneller geconvergeerde oplossingen te krijgen.
De mogelijkheid om voor een brede range scheepstypes goede voorspellingen van de
krachten op het romp te krijgen zal worden gedemonstreerd. Aangetoond zal worden dat
de trends in de krachten en momenten als functie van de drifthoek of giersnelheid goed
overeenkomen met metingen.
De verificatiestudies leveren bruikbare informatie met betrekking tot de invloed van
de griddichtheid op de voorspelde krachten. In een aantal gevallen waren de afwijkingen
in de berekeningen groot, wat aantoont dat er modelleerfouten in de numerieke resultaten
aanwezig zijn. In deze gevallen bleek vaak dat de dwarskracht te klein voorspeld werd,
terwijl het giermoment te groot was. Voor manoeuvreerstudies zijn deze afwijkingen in
het algemeen binnen acceptabele grenzen. Toch blijven verbeteringen wenselijk en deze
kunnen verkregen worden door het gebruik van fijnere grids, grotere domein groottes,
een andere grid topologie met verfijningen van het grid in het zog van het schip, andere
turbulentie modellen of het meenemen van het vrije vloeistofoppervlak.
Het predictieprogramma SurSim is gebruikt om de manoeuvreerbaarheid van de HTC
te simuleren. Een procedure wordt voorgesteld om de hydrodynamische coefficienten die
nodig zijn om de krachten op de romp te beschrijven af te leiden. Deze procedure is
gekozen om nauwkeurig zowel het lineaire gedrag tijdens koershouden als het realistis-
che gedrag tijdens havenmanoeuvres te modelleren en tevens niet-lineair gedrag goed te
beschrijven.
Ter validatiemateriaal van de manoeuvreersimulaties zijn vrijvarende manoeuvreer-
proeven met een model van de HTC uitgevoerd. Deze proeven hebben een waardevolle
set gegevens opgeleverd die algemeen gebruikt kan worden voor validatie studies. Naast
het verkrijgen van de manoeuvreereigenschappen voor een enkelschroef container schip, is
door dit proevenprogramma unieke informatie beschikbaar gekomen met betrekking tot
de belastingen op de schroef en het roer tijdens manoeuvres. Ook zijn door middel van
herhalingsproeven de onzekerheden in de manoeuvreereigenschappen bepaald.
In dit proefschrift zal worden aangetoond dat manoeuvreersimulaties verbeterd kun-
nen worden, als er hydrodynamische coëfficiënten worden gebruikt die met behulp van
CFD zijn bepaald. Vooral in de voorspellingen van de draaicirkel zijn grote verbeterin-
gen zichtbaar ten opzichte van simulaties met coëfficiënten gebaseerd op empirie. Ook de
voorspelling van het gedrag tijdens koershouden, op basis van de zig-zag manoeuvres, is
verbeterd, maar een verdere verbetering is nodig om meer betrouwbare voorspellingen te
krijgen.
De gevoeligheid van de manoeuvreereigenschappen op veranderingen in de hydrody-
namische coëfficiënten is onderzocht. Hieruit blijkt dat vooral een nauwkeurige berekening
van het giermoment nodig is om betrouwbare voorspellingen te krijgen.
 145

Acknowledgements

The work for this thesis was sponsored by the background research and development of
MARIN. I thank MARIN and in particular my former manager Frans Quadvlieg and
current manager Henk Prins for their support. I also appreciate the time and effort spent
by my supervisor René Huijsmans from Delft University of Technology during this work.
Thanks to him I had the opportunity to finalise this work with a PhD degree.
The work presented in this thesis regarding the DARPA SUBOFF and the Walrus
submarine was funded through TNO Defence, Security and Safety within the framework
of Programma V705 carried out for DMO of the Royal Netherlands Navy. Their support
is greatly acknowledged.
Another part (the work on the HTC) was funded by the Commission of the Euro-
pean Communities through the Integrated Project VIRTUE under grant 516201 in the
sixth Research and Technological Development Framework Programme (Surface Trans-
port Call). I thank the members of the manoeuvring Work Package of VIRTUE for their
discussions and insights during the project.

When starting work at MARIN, I was introduced to the subject of ship manoeu-
vring by my former colleague Jan Hooft. Thanks to him, I became interested in ship
manoeuvring simulations.
Furthermore, this work could not have been done without the support of various
colleagues at MARIN. I therefore acknowledge the help of (in random order): Jaap Windt
and Chris Klaij for their help in grid generation; Auke van der Ploeg for his efforts
in getting difficult Parnassos jobs to converge; Martin Hoekstra and Luı́s Eça (from
Instituto Superior Técnico, Portugal (IST)) for their advice regarding uncertainty ana-
lysis; Hoyte Raven for his insight into hydrodynamics in general and discussions about the
scope of work; Guilherme Vaz for his development and support of ReFRESCO; Bram
Starke for discussions on viscous-flow and help with getting started with Parnassos;
Erik van Wijngaarden for his textual suggestions about my thesis and our discussions
about being a PhD student; my colleagues at the Helpdesk for keeping the computers and
cluster alive. And all others who put up with me when I enthusiastically wanted to show
them my ColourFul Diagrams.
Also the work of Mark Bettle from University of New Brunswick is appreciated. He
conducted the calculations for the Walrus submarine presented in this thesis. It was a
pleasure to have you stay at MARIN and thanks for the great job, Mark!
146 Acknowledgements

Bas Overpelt managed and analysed the free sailing experiments for the HTC. Kees
Boers supervised also some of the tests. Thanks to both for their efforts and hard work.
Experimental data existing in the public domain has been used to validate the viscous-
flow calculations presented in this thesis. This contribution of valuable data of INSEAN,
NMRI, HSVA and Naval Surface Warfare Center, Carderock Division (NSWCCD) is
greatly appreciated.
Finally, I dearly thank my wife Aisah for her continuous encouragement and support
of my PhD work and her valuable advice on planning. Without you, I would still be
writing this thesis!
 147

Curriculum Vitae

Serge Toxopeus was born on April 23rd , 1972 in Alkmaar, The Netherlands. He finished
high school at the C.S.G. Jan Arentsz in Alkmaar in 1990 and subsequently started to
study Maritime Engineering at Delft University of Technology, Faculty of Mechanical En-
gineering and Marine Technology (currently part of the Faculty of Mechanical, Maritime
and Materials Engineering). Serge graduated in 1996 at the Ship Hydromechanics De-
partment with a Master’s thesis on the dynamic stability and manoeuvring of planing
ships.
In the same year, he became an employee of the Maritime Research Institute Nether-
lands (MARIN) in Wageningen as Project Leader, Consultant, Project Manager and
Senior Project Manager (depending on the year) in the Manoeuvring Department. His
main tasks were the development and improvement of mathematical manoeuvring models
for surface ships and submarines and the assessment of the manoeuvrability of ships using
simulations and model tests. He was actively involved in commercial projects for about
eight years.
Around 2002, the desire arose to improve the manoeuvring simulations using advanced
techniques and therefore Serge started a PhD study regarding the practical application of
viscous-flow calculations for the simulation of manoeuvring ships. Since 2005, most of his
time was spent on this PhD study. The work was partly funded as background research
by MARIN and partly through the EU project VIRTUE. This thesis is the consolidation
of the work done in the past eight years.
To be able to fully concentrate on his research, Serge moved to the Research and
Development department in 2007. His present position is Senior Researcher.
148 Curriculum Vitae

Page intentionally left blank

 149

Appendices
150 Appendices

Page intentionally left blank

Table pages – Experimental and computational results, KVLCC1 151

Experimental and computational results of forces, KVLCC1

Id Re [-] Fn [-] β γ Solver Turb. Grid Xp Xf X Yp Yf Y Zp Zf Z
01 3.3×106 - 0.0 0.00 Parnassos MNT 1.5×106 -0.003 -0.015 -0.018 - - - 0.090 0.000 0.090
02 3.3×106 - 0.5 0.00 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 -0.002 -0.000 -0.002 0.080 0.000 0.080
03 3.3×106 - 1.0 0.00 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 -0.003 -0.000 -0.004 0.080 0.000 0.080
04 3.3×106 - 2.0 0.00 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 -0.006 -0.000 -0.006 0.081 0.000 0.081
05 3.3×106 - 4.0 0.00 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 -0.012 -0.001 -0.013 0.084 0.000 0.085
06 3.3×106 - 6.0 0.00 Parnassos MNT 3.356×106 -0.004 -0.015 -0.019 -0.023 -0.001 -0.024 0.089 0.000 0.089
07 3.3×106 - 9.0 0.00 Parnassos MNT 3.356×106 -0.001 -0.016 -0.018 -0.042 -0.002 -0.043 0.102 0.000 0.102
08 3.3×106 - 12.0 0.00 Parnassos MNT 3.356×106 -0.003 -0.016 -0.019 -0.063 -0.002 -0.065 0.117 0.000 0.117
09 3.3×106 - 15.0 0.00 Parnassos MNT 3.356×106 -0.002 -0.016 -0.018 -0.086 -0.002 -0.088 0.136 0.001 0.137
10 3.3×106 - 0.0 0.13 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 0.004 -0.000 0.004 0.081 0.000 0.082
11 3.3×106 - 0.0 0.26 Parnassos MNT 3.356×106 -0.002 -0.015 -0.017 0.011 -0.000 0.011 0.082 0.000 0.082
12 3.3×106 - 0.0 0.39 Parnassos MNT 3.356×106 -0.001 -0.015 -0.017 0.017 -0.000 0.017 0.092 0.000 0.092
13 3.3×106 - 0.0 0.52 Parnassos MNT 3.356×106 0.000 -0.015 -0.015 0.022 -0.000 0.022 0.097 0.000 0.097
14 3.3×106 - 0.0 0.65 Parnassos MNT 3.356×106 0.002 -0.015 -0.013 0.029 -0.000 0.029 0.105 0.000 0.105
15 3.3×106 - -4.0 0.39 Parnassos MNT 3.356×106 -0.007 -0.015 -0.023 0.033 0.000 0.033 0.096 0.000 0.096
16 3.3×106 - -12.0 0.52 Parnassos MNT 3.356×106 -0.020 -0.017 -0.037 0.104 0.001 0.105 0.124 0.000 0.125
152 Appendices

Experimental and computational results of moments, KVLCC1

Id Re [-] Fn [-] β γ Solver Turb. Grid Kp Kf K Mp Mf M Np Nf N
01 3.3×106 - 0.0 0.00 Parnassos MNT 1.5×106 - - - -0.013 0.000 -0.013 - - -
02 3.3×106 - 0.5 0.00 Parnassos MNT 3.356×106 0.001 -0.000 0.001 -0.013 0.000 -0.012 -0.002 0.000 -0.002
03 3.3×106 - 1.0 0.00 Parnassos MNT 3.356×106 0.003 -0.000 0.002 -0.013 0.000 -0.012 -0.003 0.000 -0.003
04 3.3×106 - 2.0 0.00 Parnassos MNT 3.356×106 0.005 -0.001 0.005 -0.013 0.000 -0.012 -0.006 0.000 -0.006
05 3.3×106 - 4.0 0.00 Parnassos MNT 3.356×106 0.011 -0.001 0.010 -0.013 0.000 -0.012 -0.011 0.000 -0.011
06 3.3×106 - 6.0 0.00 Parnassos MNT 3.356×106 0.018 -0.002 0.017 -0.013 0.000 -0.012 -0.016 0.000 -0.016
07 3.3×106 - 9.0 0.00 Parnassos MNT 3.356×106 0.035 -0.002 0.032 -0.015 0.000 -0.015 -0.023 0.000 -0.022
08 3.3×106 - 12.0 0.00 Parnassos MNT 3.356×106 0.052 -0.003 0.050 -0.014 0.000 -0.013 -0.028 0.000 -0.028
09 3.3×106 - 15.0 0.00 Parnassos MNT 3.356×106 0.076 -0.003 0.073 -0.015 0.000 -0.015 -0.034 0.000 -0.034
10 3.3×106 - 0.0 0.13 Parnassos MNT 3.356×106 0.004 -0.000 0.004 -0.013 0.000 -0.012 -0.004 0.000 -0.004
11 3.3×106 - 0.0 0.26 Parnassos MNT 3.356×106 0.008 -0.000 0.008 -0.013 0.000 -0.012 -0.009 0.000 -0.009
12 3.3×106 - 0.0 0.39 Parnassos MNT 3.356×106 0.012 -0.000 0.011 -0.013 0.000 -0.012 -0.015 0.000 -0.014
13 3.3×106 - 0.0 0.52 Parnassos MNT 3.356×106 0.015 -0.000 0.015 -0.013 0.000 -0.013 -0.021 0.000 -0.021
14 3.3×106 - 0.0 0.65 Parnassos MNT 3.356×106 0.019 -0.001 0.018 -0.014 0.000 -0.013 -0.027 0.001 -0.026
15 3.3×106 - -4.0 0.39 Parnassos MNT 3.356×106 -0.001 0.001 0.000 -0.010 0.000 -0.009 -0.005 0.000 -0.005
16 3.3×106 - -12.0 0.52 Parnassos MNT 3.356×106 -0.028 0.002 -0.026 -0.003 0.000 -0.002 -0.001 0.000 -0.001
Table pages – Experimental and computational results, KVLCC2 153

Experimental and computational results of forces, KVLCC2

Id Re [-] Fn [-] β γ Solver Turb. Grid Xp Xf X Yp Yf Y Zp Zf Z
01 3.3×106 - 0.0 0.00 Parnassos MNT 1.5×106 -0.003 -0.015 -0.018 - - - 0.090 0.000 0.090
02 3.7×106 0.142 0.3 0.00 Exp (INSEAN) - - - - -0.017 - - -0.001 - - -
03 3.7×106 0.142 0.5 0.00 Exp (INSEAN) - - - - -0.014 - - -0.003 - - -
04 3.7×106 0.142 0.5 0.00 Exp (INSEAN) - - - - -0.018 - - -0.002 - - -
05 3.7×106 0.142 0.9 0.00 Exp (INSEAN) - - - - -0.031 - - -0.019 - - -
06 3.3×106 - 1.0 0.00 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 -0.004 -0.000 -0.004 0.080 0.000 0.080
07 3.7×106 0.142 1.0 0.00 Exp (INSEAN) - - - - -0.016 - - -0.003 - - -
08 3.7×106 0.142 1.8 0.00 Exp (INSEAN) - - - - -0.019 - - -0.009 - - -
09 3.7×106 0.142 2.0 0.00 Exp (INSEAN) - - - - -0.015 - - -0.008 - - -
10 3.3×106 - 2.0 0.00 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 -0.007 -0.000 -0.007 0.081 0.000 0.081
11 3.7×106 0.142 3.8 0.00 Exp (INSEAN) - - - - -0.020 - - -0.017 - - -
12 3.7×106 0.142 3.9 0.00 Exp (INSEAN) - - - - -0.015 - - -0.017 - - -
13 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.015 - - -0.015 - - -
14 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.015 - - -0.015 - - -
15 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.016 - - -0.016 - - -
16 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.016 - - -0.015 - - -
17 3.3×106 - 4.0 0.00 Parnassos MNT 3.356×106 -0.004 -0.015 -0.019 -0.015 -0.001 -0.016 0.084 0.000 0.084
18 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.016 - - -0.015 - - -
19 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.014 - - -0.015 - - -
20 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.015 - - -0.015 - - -
21 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.017 - - -0.016 - - -
22 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.017 - - -0.016 - - -
23 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.017 - - -0.015 - - -
24 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - -0.016 - - -0.016 - - -
25 3.3×106 - 6.0 0.00 Parnassos MNT 3.356×106 -0.004 -0.015 -0.019 -0.025 -0.001 -0.026 0.089 0.000 0.089
26 3.7×106 0.142 6.0 0.00 Exp (INSEAN) - - - - -0.015 - - -0.024 - - -
27 3.7×106 0.142 6.0 0.00 Exp (INSEAN) - - - - -0.020 - - -0.025 - - -
28 3.3×106 - 9.0 0.00 Parnassos MNT 3.356×106 -0.004 -0.016 -0.020 -0.045 -0.001 -0.047 0.101 0.000 0.101
29 3.3×106 - 15.0 0.00 Parnassos MNT 3.356×106 -0.002 -0.016 -0.019 -0.093 -0.002 -0.095 0.136 0.001 0.137
30 3.3×106 - 0.0 0.13 Parnassos MNT 3.356×106 -0.003 -0.015 -0.018 0.006 -0.000 0.006 0.081 0.000 0.081
31 3.7×106 0.142 0.0 0.20 Exp (INSEAN) - - - - -0.015 - - -0.009 - - -
32 3.3×106 - 0.0 0.26 Parnassos MNT 3.356×106 -0.002 -0.015 -0.017 0.012 -0.000 0.012 0.082 0.000 0.082
33 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.015 - - -0.020 - - -
34 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.016 - - -0.020 - - -
35 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.016 - - -0.019 - - -
36 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.015 - - -0.018 - - -
37 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.017 - - -0.019 - - -
38 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.016 - - -0.019 - - -
39 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.014 - - -0.019 - - -
40 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - -0.015 - - -0.020 - - -
41 3.3×106 - 0.0 0.39 Parnassos MNT 3.356×106 -0.002 -0.015 -0.017 0.018 -0.000 0.018 0.092 0.000 0.093
42 3.3×106 - 0.0 0.52 Parnassos MNT 3.356×106 -0.001 -0.015 -0.016 0.026 -0.000 0.026 0.097 0.000 0.097
43 3.7×106 0.142 0.0 0.60 Exp (INSEAN) - - - - -0.011 - - -0.031 - - -
44 3.3×106 - 0.0 0.65 Parnassos MNT 3.356×106 0.001 -0.015 -0.014 0.033 -0.000 0.033 0.106 0.000 0.106
45 3.7×106 0.142 -4.0 0.40 Exp (INSEAN) - - - - 0.011 - - 0.025 - - -
46 3.3×106 - -4.0 0.39 Parnassos MNT 3.356×106 -0.008 -0.015 -0.023 0.038 0.000 0.038 0.095 0.000 0.096
47 3.3×106 - -9.0 0.39 Parnassos MNT 3.356×106 -0.026 -0.016 -0.042 0.071 -0.001 0.070 0.107 0.000 0.107
48 3.3×106 - -12.0 0.52 Parnassos MNT 3.356×106 -0.023 -0.017 -0.039 0.113 0.002 0.115 0.129 0.001 0.129
49 3.7×106 - 0.0 0.00 ReFRESCO SST 5.388×106 -0.003 -0.015 -0.018 -0.000 0.000 -0.000 0.080 0.000 0.081
50 3.7×106 - 2.0 0.00 ReFRESCO SST 5.388×106 -0.003 -0.015 -0.018 -0.007 -0.000 -0.007 0.080 0.000 0.081
51 3.7×106 - 4.0 0.00 ReFRESCO SST 5.388×106 -0.003 -0.015 -0.018 -0.014 -0.001 -0.015 0.083 0.000 0.083
52 3.7×106 - 6.0 0.00 ReFRESCO SST 5.388×106 -0.003 -0.015 -0.018 -0.025 -0.001 -0.026 0.088 0.000 0.089
53 3.7×106 - 8.0 0.00 ReFRESCO SST 5.388×106 -0.003 -0.015 -0.018 -0.038 -0.001 -0.039 0.095 0.000 0.096
54 3.7×106 - 10.0 0.00 ReFRESCO SST 5.388×106 -0.003 -0.015 -0.018 -0.052 -0.002 -0.053 0.104 0.000 0.105
55 3.7×106 - 12.0 0.00 ReFRESCO SST 5.388×106 -0.002 -0.015 -0.018 -0.067 -0.002 -0.069 0.115 0.000 0.115
56 3.7×106 - 14.0 0.00 ReFRESCO SST 5.388×106 -0.002 -0.016 -0.017 -0.084 -0.003 -0.086 0.127 0.001 0.127
57 3.7×106 - 16.0 0.00 ReFRESCO SST 5.388×106 -0.001 -0.016 -0.016 -0.101 -0.003 -0.104 0.140 0.001 0.141
58 3.7×106 - 18.0 0.00 ReFRESCO SST 5.388×106 0.001 -0.016 -0.015 -0.119 -0.003 -0.122 0.155 0.001 0.156
59 3.7×106 - 20.0 0.00 ReFRESCO SST 5.388×106 0.002 -0.016 -0.014 -0.139 -0.004 -0.143 0.171 0.001 0.172
60 3.7×106 - 22.0 0.00 ReFRESCO SST 5.388×106 0.004 -0.016 -0.013 -0.160 -0.004 -0.164 0.189 0.001 0.190
61 3.7×106 - 24.0 0.00 ReFRESCO SST 5.388×106 0.005 -0.016 -0.011 -0.181 -0.004 -0.185 0.207 0.001 0.208
62 3.7×106 - 26.0 0.00 ReFRESCO SST 5.388×106 0.007 -0.017 -0.009 -0.202 -0.005 -0.207 0.227 0.001 0.228
63 3.7×106 - 28.0 0.00 ReFRESCO SST 5.388×106 0.010 -0.017 -0.007 -0.224 -0.005 -0.229 0.248 0.001 0.249
64 3.7×106 - 30.0 0.00 ReFRESCO SST 5.388×106 0.012 -0.017 -0.005 -0.246 -0.006 -0.251 0.269 0.001 0.271
65 3.7×106 - 32.0 0.00 ReFRESCO SST 5.388×106 0.014 -0.017 -0.003 -0.267 -0.006 -0.273 0.291 0.001 0.293
66 3.7×106 - 0.0 0.10 ReFRESCO SST 12.721×106 -0.003 -0.015 -0.018 -0.002 0.000 -0.002 0.081 0.000 0.081
67 3.7×106 - 0.0 0.20 ReFRESCO SST 12.721×106 -0.002 -0.015 -0.017 -0.006 0.000 -0.006 0.081 0.000 0.081
68 3.7×106 - 0.0 0.30 ReFRESCO SST 12.721×106 -0.002 -0.015 -0.017 -0.011 0.000 -0.011 0.081 0.000 0.082
68 3.7×106 - 0.0 0.40 ReFRESCO SST 12.721×106 -0.001 -0.015 -0.016 -0.016 0.000 -0.016 0.083 0.000 0.083
69 3.7×106 - 0.0 0.50 ReFRESCO SST 12.721×106 0.000 -0.016 -0.016 -0.025 0.001 -0.024 0.090 0.000 0.091
69 3.7×106 - 0.0 0.60 ReFRESCO SST 12.721×106 -0.002 -0.017 -0.019 -0.030 0.000 -0.030 0.095 0.000 0.096
70 3.7×106 - 0.0 0.65 ReFRESCO SST 12.721×106 -0.002 -0.017 -0.019 -0.028 0.000 -0.028 0.095 0.000 0.096
154 Appendices

Experimental and computational results of moments, KVLCC2

Id Re [-] Fn [-] β γ Solver Turb. Grid Kp Kf K Mp Mf M Np Nf N
01 3.3×106 - 0.0 0.00 Parnassos MNT 1.5×106 - - - -0.013 0.000 -0.013 - - -
02 3.7×106 0.142 0.3 0.00 Exp (INSEAN) - - - - - - - - - - -0.000
03 3.7×106 0.142 0.5 0.00 Exp (INSEAN) - - - - - - - - - - -0.001
04 3.7×106 0.142 0.5 0.00 Exp (INSEAN) - - - - - - - - - - -0.001
05 3.7×106 0.142 0.9 0.00 Exp (INSEAN) - - - - - - - - - - -0.010
06 3.3×106 - 1.0 0.00 Parnassos MNT 3.356×106 0.002 -0.000 0.002 -0.012 0.000 -0.012 -0.003 0.000 -0.003
07 3.7×106 0.142 1.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.002
08 3.7×106 0.142 1.8 0.00 Exp (INSEAN) - - - - - - - - - - -0.005
09 3.7×106 0.142 2.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.005
10 3.3×106 - 2.0 0.00 Parnassos MNT 3.356×106 0.005 -0.001 0.004 -0.012 0.000 -0.012 -0.005 0.000 -0.005
11 3.7×106 0.142 3.8 0.00 Exp (INSEAN) - - - - - - - - - - -0.012
12 3.7×106 0.142 3.9 0.00 Exp (INSEAN) - - - - - - - - - - -0.012
13 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
14 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
15 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
16 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
17 3.3×106 - 4.0 0.00 Parnassos MNT 3.356×106 0.010 -0.001 0.009 -0.012 0.000 -0.012 -0.011 0.000 -0.011
18 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
19 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
20 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
21 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.012
22 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.012
23 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
24 3.7×106 0.142 4.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.011
25 3.3×106 - 6.0 0.00 Parnassos MNT 3.356×106 0.017 -0.002 0.015 -0.013 0.000 -0.012 -0.015 0.000 -0.015
26 3.7×106 0.142 6.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.017
27 3.7×106 0.142 6.0 0.00 Exp (INSEAN) - - - - - - - - - - -0.017
28 3.3×106 - 9.0 0.00 Parnassos MNT 3.356×106 0.031 -0.002 0.028 -0.013 0.000 -0.013 -0.021 0.000 -0.021
29 3.3×106 - 15.0 0.00 Parnassos MNT 3.356×106 0.072 -0.003 0.069 -0.015 0.000 -0.014 -0.032 0.000 -0.032
30 3.3×106 - 0.0 0.13 Parnassos MNT 3.356×106 0.005 -0.000 0.004 -0.012 0.000 -0.012 -0.004 0.000 -0.004
31 3.7×106 0.142 0.0 0.20 Exp (INSEAN) - - - - - - - - - - 0.007
32 3.3×106 - 0.0 0.26 Parnassos MNT 3.356×106 0.009 -0.000 0.009 -0.013 0.000 -0.012 -0.010 0.000 -0.009
33 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.015
34 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
35 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
36 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
37 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
38 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
39 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
40 3.7×106 0.142 0.0 0.40 Exp (INSEAN) - - - - - - - - - - 0.016
41 3.3×106 - 0.0 0.39 Parnassos MNT 3.356×106 0.014 -0.000 0.013 -0.013 0.000 -0.012 -0.016 0.000 -0.015
42 3.3×106 - 0.0 0.52 Parnassos MNT 3.356×106 0.018 -0.000 0.017 -0.013 0.000 -0.013 -0.022 0.000 -0.021
43 3.7×106 0.142 0.0 0.60 Exp (INSEAN) - - - - - - - - - - 0.026
44 3.3×106 - 0.0 0.65 Parnassos MNT 3.356×106 0.022 -0.001 0.021 -0.013 0.000 -0.013 -0.028 0.001 -0.028
45 3.7×106 0.142 -4.0 0.40 Exp (INSEAN) - - - - - - - - - - -0.016
46 3.3×106 - -4.0 0.39 Parnassos MNT 3.356×106 0.003 0.001 0.004 -0.009 0.000 -0.009 -0.006 0.000 -0.006
47 3.3×106 - -9.0 0.39 Parnassos MNT 3.356×106 0.020 -0.000 0.019 -0.006 0.000 -0.005 0.001 0.000 0.001
48 3.3×106 - -12.0 0.52 Parnassos MNT 3.356×106 -0.029 0.003 -0.027 -0.001 0.000 -0.000 -0.003 0.000 -0.003
49 3.7×106 - 0.0 0.00 ReFRESCO SST 5.388×106 -0.000 -0.000 -0.000 -0.013 -0.000 -0.013 0.000 -0.000 0.000
50 3.7×106 - 2.0 0.00 ReFRESCO SST 5.388×106 0.005 0.001 0.005 -0.013 -0.000 -0.013 -0.005 -0.000 -0.005
51 3.7×106 - 4.0 0.00 ReFRESCO SST 5.388×106 0.010 0.001 0.011 -0.013 -0.000 -0.013 -0.010 -0.000 -0.010
52 3.7×106 - 6.0 0.00 ReFRESCO SST 5.388×106 0.017 0.001 0.018 -0.013 -0.000 -0.013 -0.014 -0.000 -0.014
53 3.7×106 - 8.0 0.00 ReFRESCO SST 5.388×106 0.026 0.002 0.028 -0.013 -0.000 -0.014 -0.018 -0.000 -0.018
54 3.7×106 - 10.0 0.00 ReFRESCO SST 5.388×106 0.037 0.002 0.040 -0.014 -0.000 -0.014 -0.022 -0.000 -0.022
55 3.7×106 - 12.0 0.00 ReFRESCO SST 5.388×106 0.050 0.003 0.053 -0.014 -0.000 -0.015 -0.026 -0.000 -0.026
56 3.7×106 - 14.0 0.00 ReFRESCO SST 5.388×106 0.064 0.003 0.068 -0.015 -0.000 -0.015 -0.030 -0.000 -0.030
57 3.7×106 - 16.0 0.00 ReFRESCO SST 5.388×106 0.079 0.004 0.083 -0.016 -0.000 -0.016 -0.033 -0.000 -0.034
58 3.7×106 - 18.0 0.00 ReFRESCO SST 5.388×106 0.096 0.005 0.101 -0.017 -0.000 -0.017 -0.037 -0.000 -0.038
59 3.7×106 - 20.0 0.00 ReFRESCO SST 5.388×106 0.114 0.005 0.119 -0.018 -0.000 -0.018 -0.041 -0.000 -0.041
60 3.7×106 - 22.0 0.00 ReFRESCO SST 5.388×106 0.133 0.006 0.139 -0.019 -0.000 -0.019 -0.044 -0.001 -0.044
61 3.7×106 - 24.0 0.00 ReFRESCO SST 5.388×106 0.153 0.006 0.159 -0.020 -0.000 -0.020 -0.047 -0.001 -0.047
62 3.7×106 - 26.0 0.00 ReFRESCO SST 5.388×106 0.175 0.007 0.181 -0.021 -0.000 -0.021 -0.049 -0.001 -0.050
63 3.7×106 - 28.0 0.00 ReFRESCO SST 5.388×106 0.197 0.007 0.204 -0.022 -0.000 -0.023 -0.052 -0.001 -0.053
64 3.7×106 - 30.0 0.00 ReFRESCO SST 5.388×106 0.220 0.008 0.228 -0.024 -0.000 -0.024 -0.055 -0.001 -0.056
65 3.7×106 - 32.0 0.00 ReFRESCO SST 5.388×106 0.244 0.008 0.252 -0.025 -0.000 -0.025 -0.057 -0.001 -0.058
66 3.7×106 - 0.0 0.10 ReFRESCO SST 12.721×106 -0.004 -0.000 -0.004 -0.013 -0.000 -0.013 0.004 0.000 0.004
67 3.7×106 - 0.0 0.20 ReFRESCO SST 12.721×106 -0.008 -0.000 -0.008 -0.013 -0.000 -0.013 0.009 0.000 0.009
68 3.7×106 - 0.0 0.30 ReFRESCO SST 12.721×106 -0.011 -0.000 -0.012 -0.013 -0.000 -0.014 0.013 0.000 0.013
68 3.7×106 - 0.0 0.40 ReFRESCO SST 12.721×106 -0.014 -0.001 -0.015 -0.014 -0.000 -0.014 0.017 0.000 0.018
69 3.7×106 - 0.0 0.50 ReFRESCO SST 12.721×106 -0.019 -0.001 -0.020 -0.015 -0.000 -0.015 0.023 0.001 0.024
69 3.7×106 - 0.0 0.60 ReFRESCO SST 12.721×106 -0.021 -0.001 -0.021 -0.014 -0.000 -0.014 0.029 0.001 0.030
70 3.7×106 - 0.0 0.65 ReFRESCO SST 12.721×106 -0.023 -0.001 -0.023 -0.014 -0.000 -0.014 0.032 0.001 0.033
Table pages – Experimental and computational results, KVLCC2M 155

Experimental and computational results of forces, KVLCC2M

Id Re [-] Fn [-] β γ Solver Turb. Grid Xp Xf X Yp Yf Y Zp Zf Z
01 3.9×106 0.142 0.0 0.00 Exp (NMRI) - - - - -0.018 - - - - - -
02 3.9×106 - 0.0 0.00 Parnassos MNT 1.6×106 -0.003 -0.015 -0.017 - - - 0.080 0.000 0.080
03 3.9×106 0.142 3.0 0.00 Exp (NMRI) - - - - -0.018 - - -0.013 - - -
04 3.9×106 0.142 3.0 0.00 Exp (NMRI) - - - - -0.018 - - -0.012 - - -
05 3.9×106 - 3.0 0.00 Parnassos MNT 3.8×106 -0.003 -0.015 -0.018 -0.009 -0.001 -0.009 0.082 0.000 0.082
06 3.9×106 0.142 6.0 0.00 Exp (NMRI) - - - - -0.018 - - -0.026 - - -
07 3.9×106 - 6.0 0.00 Parnassos MNT 3.8×106 -0.003 -0.015 -0.018 -0.022 -0.001 -0.023 0.089 0.000 0.089
08 3.9×106 0.142 9.0 0.00 Exp (NMRI) - - - - -0.017 - - -0.046 - - -
09 3.9×106 - 9.0 0.00 Parnassos MNT 3.8×106 -0.003 -0.015 -0.018 -0.041 -0.001 -0.042 0.100 0.000 0.100
10 3.9×106 0.142 12.0 0.00 Exp (NMRI) - - - - -0.018 - - -0.071 - - -
11 3.9×106 - 12.0 0.00 Parnassos MNT 3.8×106 -0.002 -0.015 -0.018 -0.063 -0.002 -0.064 0.115 0.000 0.115
12 3.9×106 - 12.0 0.00 Parnassos SST 3.8×106 -0.003 -0.015 -0.018 -0.066 -0.002 -0.067 0.115 0.000 0.115
13 3.9×106 - 12.0 0.00 Parnassos MNT2 3.8×106 -0.001 -0.015 -0.017 -0.062 -0.002 -0.064 0.114 0.000 0.114
14 3.9×106 0.142 15.0 0.00 Exp (NMRI) - - - - -0.016 - - -0.098 - - -
15 3.9×106 - 15.0 0.00 Parnassos MNT 3.8×106 -0.001 -0.016 -0.016 -0.083 -0.002 -0.085 0.133 0.001 0.133
16 3.9×106 0.142 18.0 0.00 Exp (NMRI) - - - - -0.014 - - -0.124 - - -
17 3.9×106 - 18.0 0.00 Parnassos MNT 3.8×106 0.003 -0.016 -0.013 -0.109 -0.002 -0.112 0.153 0.001 0.154
18 3.9×106 - 0.0 0.10 Parnassos MNT 6.5×105 -0.005 -0.015 -0.020 0.004 -0.000 0.004 0.080 0.000 0.081
19 3.9×106 - 0.0 0.20 Parnassos MNT 6.5×105 -0.005 -0.015 -0.019 0.009 -0.000 0.009 0.083 0.000 0.083
20 3.9×106 - 0.0 0.25 Parnassos MNT 6.5×105 -0.005 -0.015 -0.019 0.011 -0.000 0.011 0.084 0.000 0.084
21 3.9×106 - 0.0 0.30 Parnassos MNT 8.7×104 -0.007 -0.014 -0.021 0.013 0.000 0.013 0.085 0.000 0.085
22 3.9×106 - 0.0 0.40 Parnassos MNT 8.7×104 -0.004 -0.014 -0.018 0.016 0.000 0.016 0.088 0.000 0.089
23 3.9×106 - 0.0 0.60 Parnassos MNT 8.7×104 -0.003 -0.014 -0.017 0.025 0.000 0.025 0.101 0.000 0.101
24 3.9×106 - -12.0 0.10 Parnassos MNT 8.7×104 -0.026 -0.015 -0.041 0.073 -0.002 0.071 0.116 0.000 0.116
25 3.9×106 - -12.0 0.30 Parnassos MNT 8.7×104 -0.035 -0.016 -0.051 0.087 -0.002 0.085 0.119 0.000 0.119
26 3.9×106 - -12.0 0.60 Parnassos MNT 8.7×104 -0.025 -0.016 -0.041 0.123 0.002 0.125 0.131 0.000 0.132
156 Appendices

Experimental and computational results of moments, KVLCC2M

Id Re [-] Fn [-] β γ Solver Turb. Grid Kp Kf K Mp Mf M Np Nf N
01 3.9×106 0.142 0.0 0.00 Exp (NMRI) - - - - - - - - - - -
02 3.9×106 - 0.0 0.00 Parnassos MNT 1.6×106 - - - -0.013 0.000 -0.012 - - -
03 3.9×106 0.142 3.0 0.00 Exp (NMRI) - - - - - - - - - - -0.006
04 3.9×106 0.142 3.0 0.00 Exp (NMRI) - - - - - - - - - - -0.007
05 3.9×106 - 3.0 0.00 Parnassos MNT 3.8×106 0.007 -0.001 0.006 -0.013 0.000 -0.012 -0.008 0.000 -0.008
06 3.9×106 0.142 6.0 0.00 Exp (NMRI) - - - - - - - - - - -0.014
07 3.9×106 - 6.0 0.00 Parnassos MNT 3.8×106 0.016 -0.001 0.015 -0.013 0.000 -0.013 -0.015 0.000 -0.015
08 3.9×106 0.142 9.0 0.00 Exp (NMRI) - - - - - - - - - - -0.019
09 3.9×106 - 9.0 0.00 Parnassos MNT 3.8×106 0.031 -0.002 0.029 -0.014 0.000 -0.013 -0.020 0.000 -0.020
10 3.9×106 0.142 12.0 0.00 Exp (NMRI) - - - - - - - - - - -0.025
11 3.9×106 - 12.0 0.00 Parnassos MNT 3.8×106 0.050 -0.003 0.047 -0.014 0.000 -0.014 -0.026 0.000 -0.025
12 3.9×106 - 12.0 0.00 Parnassos SST 3.8×106 0.054 -0.002 0.051 -0.014 0.000 -0.014 -0.026 0.000 -0.026
13 3.9×106 - 12.0 0.00 Parnassos MNT2 3.8×106 0.050 -0.003 0.047 -0.015 0.000 -0.014 -0.026 0.000 -0.026
14 3.9×106 0.142 15.0 0.00 Exp (NMRI) - - - - - - - - - - -0.031
15 3.9×106 - 15.0 0.00 Parnassos MNT 3.8×106 0.074 -0.003 0.071 -0.016 0.000 -0.015 -0.033 0.000 -0.032
16 3.9×106 0.142 18.0 0.00 Exp (NMRI) - - - - - - - - - - -0.035
17 3.9×106 - 18.0 0.00 Parnassos MNT 3.8×106 0.103 -0.004 0.100 -0.017 0.000 -0.017 -0.038 0.000 -0.038
18 3.9×106 - 0.0 0.10 Parnassos MNT 6.5×105 0.003 -0.000 0.003 -0.012 0.000 -0.012 -0.003 0.000 -0.003
19 3.9×106 - 0.0 0.20 Parnassos MNT 6.5×105 0.007 -0.000 0.007 -0.012 0.000 -0.012 -0.006 0.000 -0.006
20 3.9×106 - 0.0 0.25 Parnassos MNT 6.5×105 0.008 -0.000 0.008 -0.012 0.000 -0.012 -0.008 0.000 -0.008
21 3.9×106 - 0.0 0.30 Parnassos MNT 8.7×104 0.010 -0.000 0.010 -0.012 0.000 -0.011 -0.010 0.000 -0.010
22 3.9×106 - 0.0 0.40 Parnassos MNT 8.7×104 0.014 -0.000 0.014 -0.012 0.000 -0.012 -0.014 0.000 -0.013
23 3.9×106 - 0.0 0.60 Parnassos MNT 8.7×104 0.022 -0.000 0.022 -0.013 0.000 -0.012 -0.023 0.000 -0.023
24 3.9×106 - -12.0 0.10 Parnassos MNT 8.7×104 0.050 -0.001 0.049 -0.011 0.000 -0.010 0.023 -0.000 0.023
25 3.9×106 - -12.0 0.30 Parnassos MNT 8.7×104 0.018 -0.001 0.017 -0.006 0.000 -0.006 0.013 -0.000 0.013
26 3.9×106 - -12.0 0.60 Parnassos MNT 8.7×104 -0.028 0.002 -0.025 0.000 0.000 0.001 -0.005 0.000 -0.005
Table pages – Experimental and computational results, HTC 157

Experimental and computational results of forces, HTC

Id Re [-] Fn [-] β γ Solver Turb. Grid Xp Xf X Yp Yf Y Zp Zf Z
01 6.3×106 0.132 0.0 0.00 Exp (HSVA) - - - - -0.014 - - 0.000 - - 0.055
02 6.3×106 0.132 0.0 0.00 Exp (HSVA) - - - - -0.014 - - 0.000 - - 0.063
03 1.1×107 0.238 0.0 0.00 Exp (HSVA) - - - - -0.014 - - 0.000 - - 0.062
04 1.1×107 0.238 0.0 0.00 Exp (HSVA) - - - - -0.014 - - 0.000 - - 0.064
05 6.3×106 - 0.0 0.00 Parnassos MNT 3.7×106 -0.002 -0.012 -0.014 - - - 0.058 0.000 0.058
06 8.3×106 - 0.0 0.00 Parnassos MNT 1.2×106 -0.002 -0.011 -0.013 - - - 0.058 0.000 0.059
07 1.2×107 - 0.0 0.00 Parnassos MNT 3.7×106 -0.002 -0.011 -0.013 - - - 0.058 0.000 0.059
08 6.3×106 - 2.5 0.00 Parnassos MNT 3.7×106 -0.003 -0.012 -0.014 -0.008 -0.000 -0.009 0.060 0.000 0.060
09 6.3×106 0.132 5.0 0.00 Exp (HSVA) - - - - -0.015 - - -0.020 - - 0.064
10 6.3×106 0.132 5.0 0.00 Exp (HSVA) - - - - -0.015 - - -0.017 - - 0.067
11 1.1×107 0.238 5.0 0.00 Exp (HSVA) - - - - -0.015 - - -0.021 - - 0.068
12 1.1×107 0.238 5.0 0.00 Exp (HSVA) - - - - -0.015 - - -0.019 - - 0.072
13 6.3×106 - 5.0 0.00 Parnassos MNT 3.7×106 -0.003 -0.012 -0.015 -0.017 -0.000 -0.017 0.065 0.000 0.066
14 1.2×107 - 5.0 0.00 Parnassos MNT 3.7×106 -0.003 -0.011 -0.013 -0.016 -0.000 -0.017 0.066 0.000 0.066
15 6.3×106 0.132 10.0 0.00 Exp (HSVA) - - - - -0.017 - - -0.048 - - 0.097
16 6.3×106 0.132 10.0 0.00 Exp (HSVA) - - - - -0.017 - - -0.045 - - 0.088
17 1.1×107 0.238 10.0 0.00 Exp (HSVA) - - - - -0.017 - - -0.048 - - 0.089
18 1.1×107 0.238 10.0 0.00 Exp (HSVA) - - - - -0.017 - - -0.045 - - 0.091
19 3.7×106 - 10.0 0.00 Parnassos MNT 3.7×106 -0.004 -0.013 -0.017 -0.044 -0.001 -0.045 0.086 0.000 0.086
20 6.3×106 - 10.0 0.00 Parnassos MNT 3.7×106 -0.003 -0.012 -0.016 -0.043 -0.001 -0.044 0.086 0.000 0.086
21 1.2×107 - 10.0 0.00 Parnassos MNT 3.7×106 -0.003 -0.011 -0.014 -0.042 -0.001 -0.043 0.086 0.000 0.086
22 7.4×108 - 10.0 0.00 Parnassos MNT 5.3×106 -0.003 -0.006 -0.009 -0.038 -0.001 -0.039 0.087 0.000 0.087
23 6.3×106 - 15.0 0.00 Parnassos MNT 3.7×106 -0.003 -0.013 -0.016 -0.079 -0.002 -0.081 0.115 0.001 0.115
24 6.3×106 0.132 20.0 0.00 Exp (HSVA) - - - - -0.015 - - -0.134 - - 0.152
25 6.3×106 0.132 20.0 0.00 Exp (HSVA) - - - - -0.015 - - -0.127 - - 0.164
26 1.1×107 0.238 20.0 0.00 Exp (HSVA) - - - - -0.019 - - -0.129 - - 0.159
27 1.1×107 0.238 20.0 0.00 Exp (HSVA) - - - - -0.019 - - -0.123 - - 0.160
28 6.3×106 0.132 30.0 0.00 Exp (HSVA) - - - - -0.008 - - -0.244 - - 0.243
29 1.1×107 0.238 30.0 0.00 Exp (HSVA) - - - - -0.018 - - -0.240 - - 0.258
30 1.2×107 0.238 0.0 0.10 Exp (HSVA) - - - - -0.014 - - 0.003 - - 0.069
31 6.3×106 - 0.0 0.10 Parnassos MNT 1.2×106 -0.002 -0.012 -0.014 0.002 0.000 0.002 0.058 0.000 0.058
32 6.3×106 - 0.0 0.15 Parnassos MNT 1.2×106 -0.002 -0.012 -0.014 0.004 0.000 0.004 0.059 0.000 0.059
33 1.2×107 0.238 0.0 0.20 Exp (HSVA) - - - - -0.015 - - 0.008 - - 0.070
34 6.3×106 - 0.0 0.20 Parnassos MNT 1.9×106 -0.002 -0.012 -0.014 0.006 0.000 0.006 0.060 0.000 0.061
35 1.2×107 - 0.0 0.20 Parnassos MNT 1.2×106 -0.002 -0.011 -0.013 0.005 0.000 0.005 0.060 0.000 0.061
36 6.3×106 - -5.0 0.20 Parnassos MNT 1.2×106 -0.006 -0.012 -0.017 0.025 0.001 0.026 0.067 0.000 0.067
37 6.3×106 - -10.0 0.20 Parnassos MNT 1.2×106 -0.009 -0.012 -0.021 0.051 0.001 0.052 0.088 0.000 0.088
38 6.3×106 - 0.0 0.30 Parnassos MNT 1.2×106 -0.003 -0.012 -0.014 0.009 0.000 0.009 0.064 0.000 0.064
39 1.2×107 0.238 0.0 0.40 Exp (HSVA) - - - - -0.017 - - 0.016 - - 0.079
40 6.3×106 - 0.0 0.40 Parnassos MNT 1.2×106 -0.003 -0.012 -0.015 0.012 0.000 0.013 0.068 0.000 0.068
41 1.2×107 - 0.0 0.40 Parnassos MNT 1.2×106 -0.002 -0.011 -0.013 0.012 0.000 0.012 0.068 0.000 0.068
42 6.3×106 - -6.0 0.40 Parnassos MNT 1.2×106 -0.010 -0.012 -0.022 0.043 0.001 0.044 0.078 0.000 0.079
43 6.3×106 - -10.0 0.40 Parnassos MNT 1.2×106 -0.014 -0.012 -0.026 0.063 0.001 0.064 0.095 0.000 0.095
44 6.3×106 - -15.0 0.40 Parnassos MNT 1.2×106 -0.019 -0.013 -0.032 0.108 0.002 0.110 0.127 0.001 0.128
45 1.2×107 0.238 0.0 0.56 Exp (HSVA) - - - - -0.019 - - 0.024 - - 0.086
46 6.3×106 - 0.0 0.56 Parnassos MNT 1.2×106 -0.003 -0.012 -0.015 0.015 0.001 0.016 0.077 0.000 0.078
47 1.2×107 - 0.0 0.00 ReFRESCO SST 5.2×106 -0.002 -0.011 -0.013 - - - 0.063 0.000 0.063
48 1.2×107 - 2.5 0.00 ReFRESCO SST 5.2×106 -0.002 -0.011 -0.013 -0.006 -0.000 -0.006 0.065 0.000 0.065
49 1.2×107 - 5.0 0.00 ReFRESCO SST 5.2×106 -0.003 -0.011 -0.014 -0.013 -0.000 -0.014 0.069 0.000 0.069
50 1.2×107 - 7.5 0.00 ReFRESCO SST 5.2×106 -0.003 -0.011 -0.014 -0.024 -0.001 -0.025 0.076 0.000 0.076
51 1.2×107 - 10.0 0.00 ReFRESCO SST 5.2×106 -0.003 -0.011 -0.015 -0.037 -0.001 -0.038 0.086 0.000 0.087
52 1.2×107 - 12.5 0.00 ReFRESCO SST 5.2×106 -0.003 -0.012 -0.014 -0.054 -0.001 -0.055 0.099 0.000 0.099
53 1.2×107 - 15.0 0.00 ReFRESCO SST 5.2×106 -0.002 -0.012 -0.014 -0.074 -0.002 -0.076 0.113 0.000 0.114
54 1.2×107 - 17.5 0.00 ReFRESCO SST 5.2×106 -0.001 -0.012 -0.013 -0.098 -0.002 -0.100 0.130 0.001 0.130
55 1.2×107 - 20.0 0.00 ReFRESCO SST 5.2×106 -0.000 -0.012 -0.012 -0.124 -0.002 -0.126 0.149 0.001 0.149
56 1.2×107 - 22.5 0.00 ReFRESCO SST 5.2×106 0.001 -0.012 -0.011 -0.150 -0.003 -0.153 0.169 0.001 0.170
57 1.2×107 - 25.0 0.00 ReFRESCO SST 5.2×106 0.003 -0.012 -0.009 -0.177 -0.003 -0.180 0.191 0.001 0.192
58 1.2×107 - 27.5 0.00 ReFRESCO SST 5.2×106 0.005 -0.012 -0.007 -0.205 -0.003 -0.208 0.214 0.001 0.215
59 1.2×107 - 30.0 0.00 ReFRESCO SST 5.2×106 0.006 -0.012 -0.005 -0.234 -0.004 -0.238 0.238 0.001 0.239
60 1.2×107 - 0.0 0.10 ReFRESCO SST 5.2×106 -0.002 -0.011 -0.013 0.003 0.000 0.003 0.064 0.000 0.064
61 1.2×107 - 0.0 0.20 ReFRESCO SST 5.2×106 -0.002 -0.011 -0.013 0.007 0.000 0.008 0.066 0.000 0.066
62 1.2×107 - 0.0 0.40 ReFRESCO SST 5.2×106 -0.003 -0.012 -0.015 0.015 0.000 0.015 0.073 0.000 0.073
63 1.2×107 - 0.0 0.56 ReFRESCO SST 5.2×106 -0.004 -0.012 -0.016 0.020 0.001 0.021 0.089 0.000 0.090
64 1.2×107 - -5.0 0.20 ReFRESCO SST 5.2×106 -0.006 -0.011 -0.017 0.027 0.001 0.027 0.073 0.000 0.073
65 1.2×107 - -10.0 0.20 ReFRESCO SST 5.2×106 -0.009 -0.011 -0.021 0.054 0.001 0.055 0.096 0.000 0.096
66 1.2×107 - -20.0 0.20 ReFRESCO SST 5.2×106 -0.012 -0.012 -0.025 0.138 0.003 0.141 0.169 0.001 0.170
67 1.2×107 - -5.0 0.40 ReFRESCO SST 5.2×106 -0.009 -0.011 -0.021 0.040 0.001 0.041 0.080 0.000 0.080
68 1.2×107 - -10.0 0.40 ReFRESCO SST 5.2×106 -0.015 -0.011 -0.027 0.065 0.001 0.067 0.100 0.000 0.101
69 1.2×107 - -15.0 0.40 ReFRESCO SST 5.2×106 -0.020 -0.012 -0.032 0.108 0.002 0.110 0.138 0.000 0.138
70 1.2×107 - -20.0 0.40 ReFRESCO SST 5.2×106 -0.024 -0.012 -0.036 0.152 0.003 0.155 0.182 0.001 0.182
71 1.2×107 - -30.0 0.40 ReFRESCO SST 5.2×106 -0.028 -0.013 -0.041 0.254 0.004 0.259 0.284 0.001 0.285
72 1.2×107 - -20.0 0.56 ReFRESCO SST 2.5×106 -0.033 -0.013 -0.047 0.183 0.003 0.186 0.200 0.001 0.201
158 Appendices

Experimental and computational results of moments, HTC

Id Re [-] Fn [-] β γ Solver Turb. Grid Kp Kf K Mp Mf M Np Nf N
01 6.3×106 0.132 0.0 0.00 Exp (HSVA) - - - - 0.000 - - -0.004 - - 0.000
02 6.3×106 0.132 0.0 0.00 Exp (HSVA) - - - - 0.000 - - - - - 0.000
03 1.1×107 0.238 0.0 0.00 Exp (HSVA) - - - - 0.000 - - -0.004 - - 0.000
04 1.1×107 0.238 0.0 0.00 Exp (HSVA) - - - - 0.000 - - - - - 0.000
05 6.3×106 - 0.0 0.00 Parnassos MNT 3.7×106 - - - -0.001 0.000 -0.001 - - -
06 8.3×106 - 0.0 0.00 Parnassos MNT 1.2×106 - - - -0.001 0.000 -0.001 - - -
07 1.2×107 - 0.0 0.00 Parnassos MNT 3.7×10 6 - - - -0.001 0.000 -0.001 - - -
08 6.3×106 - 2.5 0.00 Parnassos MNT 3.7×10 6 0.004 -0.000 0.004 -0.001 0.000 -0.001 -0.006 0.000 -0.006
09 6.3×106 0.132 5.0 0.00 Exp (HSVA) - - - - 0.010 - - -0.004 - - -0.012
10 6.3×106 0.132 5.0 0.00 Exp (HSVA) - - - - 0.009 - - - - - -0.011
11 1.1×107 0.238 5.0 0.00 Exp (HSVA) - - - - 0.009 - - -0.004 - - -0.012
12 1.1×107 0.238 5.0 0.00 Exp (HSVA) - - - - 0.010 - - - - - -0.012
13 6.3×106 - 5.0 0.00 Parnassos MNT 3.7×106 0.009 -0.001 0.008 -0.001 0.000 -0.001 -0.013 0.000 -0.013
14 1.2×107 - 5.0 0.00 Parnassos MNT 3.7×106 0.009 -0.001 0.008 -0.001 0.000 -0.001 -0.013 0.000 -0.013
15 6.3×106 0.132 10.0 0.00 Exp (HSVA) - - - - 0.021 - - -0.005 - - -0.023
16 6.3×106 0.132 10.0 0.00 Exp (HSVA) - - - - 0.021 - - - - - -0.022
17 1.1×107 0.238 10.0 0.00 Exp (HSVA) - - - - 0.021 - - -0.004 - - -0.025
18 1.1×107 0.238 10.0 0.00 Exp (HSVA) - - - - 0.021 - - - - - -0.024
19 3.7×106 - 10.0 0.00 Parnassos MNT 3.7×10 6 0.019 -0.002 0.017 -0.002 0.000 -0.001 -0.024 0.000 -0.024
20 6.3×106 - 10.0 0.00 Parnassos MNT 3.7×106 0.020 -0.002 0.018 -0.002 0.000 -0.001 -0.024 0.000 -0.024
21 1.2×107 - 10.0 0.00 Parnassos MNT 3.7×106 0.020 -0.002 0.018 -0.002 0.000 -0.001 -0.024 0.000 -0.024
22 7.4×108 - 10.0 0.00 Parnassos MNT 5.3×106 0.021 -0.001 0.020 -0.002 0.000 -0.002 -0.025 0.000 -0.025
23 6.3×106 - 15.0 0.00 Parnassos MNT 3.7×106 0.040 -0.003 0.037 -0.002 0.000 -0.002 -0.037 0.000 -0.037
24 6.3×106 0.132 20.0 0.00 Exp (HSVA) - - - - 0.060 - - -0.007 - - -0.048
25 6.3×106 0.132 20.0 0.00 Exp (HSVA) - - - - 0.062 - - - - - -0.047
26 1.1×107 0.238 20.0 0.00 Exp (HSVA) - - - - 0.063 - - -0.006 - - -0.047
27 1.1×107 0.238 20.0 0.00 Exp (HSVA) - - - - 0.064 - - - - - -0.045
28 6.3×106 0.132 30.0 0.00 Exp (HSVA) - - - - 0.110 - - -0.011 - - -0.073
29 1.1×107 0.238 30.0 0.00 Exp (HSVA) - - - - 0.110 - - -0.009 - - -0.070
30 1.2×107 0.238 0.0 0.10 Exp (HSVA) - - - - 0.000 - - - - - -0.004
31 6.3×106 - 0.0 0.10 Parnassos MNT 1.2×106 0.001 0.000 0.001 -0.001 0.000 -0.001 -0.003 0.000 -0.003
32 6.3×106 - 0.0 0.15 Parnassos MNT 1.2×106 0.002 0.000 0.002 -0.001 0.000 -0.001 -0.005 0.000 -0.005
33 1.2×107 0.238 0.0 0.20 Exp (HSVA) - - - - 0.001 - - - - - -0.009
34 6.3×106 - 0.0 0.20 Parnassos MNT 1.9×106 0.002 0.000 0.002 -0.001 0.000 -0.001 -0.007 0.000 -0.007
35 1.2×107 - 0.0 0.20 Parnassos MNT 1.2×10 6 0.002 0.000 0.002 -0.001 0.000 -0.001 -0.007 0.000 -0.007
36 6.3×106 - -5.0 0.20 Parnassos MNT 1.2×106 -0.006 0.001 -0.005 0.001 0.000 0.001 0.005 0.000 0.005
37 6.3×106 - -10.0 0.20 Parnassos MNT 6
1.2×10 -0.012 0.002 -0.010 0.003 0.000 0.003 0.014 0.000 0.014
38 6.3×106 - 0.0 0.30 Parnassos MNT 1.2×106 0.004 0.000 0.004 -0.001 0.000 -0.001 -0.012 0.000 -0.012
39 1.2×107 0.238 0.0 0.40 Exp (HSVA) - - - - 0.003 - - - - - -0.018
40 6.3×106 - 0.0 0.40 Parnassos MNT 1.2×106 0.005 0.000 0.005 -0.001 0.000 -0.001 -0.018 0.000 -0.018
41 1.2×107 - 0.0 0.40 Parnassos MNT 1.2×106 0.005 0.000 0.005 -0.001 0.000 -0.001 -0.018 0.000 -0.018
42 6.3×106 - -6.0 0.40 Parnassos MNT 1.2×106 -0.004 0.001 -0.003 0.004 0.000 0.004 -0.003 0.000 -0.003
43 6.3×106 - -10.0 0.40 Parnassos MNT 1.2×106 -0.007 0.002 -0.005 0.007 0.000 0.007 0.004 0.000 0.005
44 6.3×106 - -15.0 0.40 Parnassos MNT 1.2×106 -0.012 0.003 -0.009 0.010 0.000 0.011 0.008 0.000 0.008
45 1.2×107 0.238 0.0 0.56 Exp (HSVA) - - - - 0.004 - - - - - -0.029
46 6.3×106 - 0.0 0.56 Parnassos MNT 1.2×106 0.009 0.000 0.009 -0.002 0.000 -0.001 -0.029 0.000 -0.029
47 1.2×107 - 0.0 0.00 ReFRESCO SST 5.2×10 6 - - - -0.001 -0.000 -0.001 - - -
48 1.2×107 - 2.5 0.00 ReFRESCO SST 5.2×106 0.004 0.000 0.004 -0.001 -0.000 -0.001 -0.005 -0.000 -0.005
49 1.2×107 - 5.0 0.00 ReFRESCO SST 5.2×106 0.008 0.001 0.009 -0.001 -0.000 -0.001 -0.011 -0.000 -0.011
50 1.2×107 - 7.5 0.00 ReFRESCO SST 5.2×106 0.012 0.001 0.013 -0.001 -0.000 -0.001 -0.017 -0.000 -0.017
51 1.2×107 - 10.0 0.00 ReFRESCO SST 5.2×106 0.016 0.002 0.018 -0.001 -0.000 -0.002 -0.022 -0.000 -0.022
52 1.2×107 - 12.5 0.00 ReFRESCO SST 5.2×106 0.024 0.002 0.025 -0.002 -0.000 -0.002 -0.028 -0.000 -0.028
53 1.2×107 - 15.0 0.00 ReFRESCO SST 5.2×106 0.035 0.002 0.037 -0.003 -0.000 -0.003 -0.034 -0.000 -0.034
54 1.2×107 - 17.5 0.00 ReFRESCO SST 5.2×106 0.048 0.003 0.051 -0.003 -0.000 -0.004 -0.040 -0.000 -0.041
55 1.2×107 - 20.0 0.00 ReFRESCO SST 5.2×106 0.061 0.003 0.065 -0.004 -0.000 -0.004 -0.047 -0.000 -0.047
56 1.2×107 - 22.5 0.00 ReFRESCO SST 5.2×106 0.074 0.003 0.078 -0.005 -0.000 -0.005 -0.054 -0.000 -0.054
57 1.2×107 - 25.0 0.00 ReFRESCO SST 5.2×106 0.087 0.004 0.091 -0.006 -0.000 -0.006 -0.062 -0.000 -0.062
58 1.2×107 - 27.5 0.00 ReFRESCO SST 5.2×106 0.099 0.004 0.104 -0.007 -0.000 -0.007 -0.068 -0.000 -0.068
59 1.2×107 - 30.0 0.00 ReFRESCO SST 5.2×106 0.112 0.005 0.116 -0.007 -0.000 -0.007 -0.074 -0.000 -0.075
60 1.2×107 - 0.0 0.10 ReFRESCO SST 5.2×106 0.001 -0.000 0.001 -0.001 -0.000 -0.001 -0.003 -0.000 -0.003
61 1.2×107 - 0.0 0.20 ReFRESCO SST 5.2×106 0.002 -0.000 0.002 -0.001 -0.000 -0.001 -0.007 -0.000 -0.008
62 1.2×107 - 0.0 0.40 ReFRESCO SST 5.2×106 0.004 -0.000 0.003 -0.001 -0.000 -0.002 -0.018 -0.000 -0.018
63 1.2×107 - 0.0 0.56 ReFRESCO SST 5.2×106 0.006 -0.001 0.005 -0.001 -0.000 -0.001 -0.030 -0.000 -0.030
64 1.2×107 - -5.0 0.20 ReFRESCO SST 5.2×106 -0.007 -0.001 -0.008 0.001 -0.000 0.001 0.004 -0.000 0.004
65 1.2×107 - -10.0 0.20 ReFRESCO SST 5.2×106 -0.014 -0.002 -0.016 0.003 -0.000 0.003 0.013 -0.000 0.013
66 1.2×107 - -20.0 0.20 ReFRESCO SST 5.2×106 -0.047 -0.004 -0.050 0.005 -0.000 0.005 0.032 -0.000 0.031
67 1.2×107 - -5.0 0.40 ReFRESCO SST 5.2×106 -0.006 -0.001 -0.007 0.003 -0.000 0.003 -0.006 -0.000 -0.006
68 1.2×107 - -10.0 0.40 ReFRESCO SST 5.2×106 -0.010 -0.002 -0.011 0.007 -0.000 0.007 0.002 -0.000 0.002
69 1.2×107 - -15.0 0.40 ReFRESCO SST 5.2×106 -0.014 -0.003 -0.017 0.011 -0.000 0.011 0.008 -0.000 0.008
70 1.2×107 - -20.0 0.40 ReFRESCO SST 5.2×106 -0.031 -0.004 -0.035 0.014 -0.000 0.013 0.014 -0.000 0.014
71 1.2×107 - -30.0 0.40 ReFRESCO SST 5.2×106 -0.071 -0.006 -0.077 0.018 -0.000 0.018 0.027 -0.000 0.027
72 1.2×107 - -20.0 0.56 ReFRESCO SST 2.5×106 -0.020 -0.004 -0.025 0.021 -0.000 0.020 0.003 -0.000 0.003
Table pages – Experimental and computational results, SUBOFF 159

Experimental and computational results of forces and moments, SUBOFF

Id Re [-] β γ Solver Turb. Grid Xp Xf X Yp Yf Y Np Nf N
01 1.4×107 0 0.0 Exp (DTRC) - - - - -1.061 - - -0.099 - - 0.098
02 1.4×107 0 0.0 Exp (DTRC) - - - - -1.051 - - -0.082 - - 0.070
03 1.4×107 0 0.0 ReFRESCO SST 4138×103 -0.129 -0.967 -1.096 0.000 0.000 0.000 0.000 0.000 0.000
04 1.4×107 0 0.0 ReFRESCO SA 4138×103 -0.067 -0.950 -1.017 -0.007 0.000 -0.007 0.000 0.000 0.000
05 1.4×107 1 0.0 Exp (DTRC) - - - - -1.060 - - -0.021 - - -0.346
06 1.4×107 1 0.0 Exp (DTRC) - - - - -1.050 - - -0.206 - - -0.160
07 1.4×107 2 0.0 Exp (DTRC) - - - - -1.045 - - -0.133 - - -0.517
08 1.4×107 2 0.0 Exp (DTRC) - - - - -1.069 - - -0.292 - - -0.350
09 1.4×107 2 0.0 ReFRESCO SST 3
4138×10 -0.138 -0.969 -1.107 -0.135 -0.029 -0.164 -0.438 -0.001 -0.439
10 1.4×107 2 0.0 ReFRESCO SA 3
4138×10 -0.065 -0.952 -1.017 -0.168 -0.032 -0.199 -0.000 -0.000 -0.000
11 1.4×107 3 0.0 Exp (DTRC) - - - - -1.060 - - -0.321 - - -0.771
12 1.4×107 3 0.0 Exp (DTRC) - - - - -1.060 - - -0.408 - - -0.598
13 1.4×107 4 0.0 Exp (DTRC) - - - - -1.061 - - -0.449 - - -0.996
14 1.4×107 4 0.0 Exp (DTRC) - - - - -1.046 - - -0.592 - - -0.864
15 1.4×107 4 0.0 ReFRESCO SST 4138×103 -0.132 -0.973 -1.105 -0.330 -0.060 -0.390 -0.899 -0.002 -0.902
16 1.4×107 6 0.0 Exp (DTRC) - - - - -1.059 - - -0.886 - - -1.385
17 1.4×107 6 0.0 Exp (DTRC) - - - - -1.061 - - -0.947 - - -1.226
18 1.4×107 6 0.0 ReFRESCO SST 3
4138×10 -0.121 -0.980 -1.100 -0.600 -0.094 -0.694 -1.350 -0.003 -1.353
19 1.4×107 8 0.0 Exp (DTRC) - - - - -1.070 - - -1.540 - - -1.710
20 1.4×107 8 0.0 Exp (DTRC) - - - - -1.080 - - -1.550 - - -1.570
21 1.4×107 8 0.0 ReFRESCO SST 4138×103 -0.106 -0.989 -1.095 -0.975 -0.128 -1.102 -1.768 -0.005 -1.773
22 1.4×107 10 0.0 Exp (DTRC) - - - - -1.049 - - -2.450 - - -2.008
23 1.4×107 10 0.0 Exp (DTRC) - - - - -1.079 - - -2.337 - - -1.875
24 1.4×107 10 0.0 ReFRESCO SST 4138×103 -0.086 -1.002 -1.088 -1.487 -0.161 -1.648 -2.144 -0.007 -2.151
25 1.4×107 12 0.0 Exp (DTRC) - - - - -0.994 - - -3.460 - - -2.270
26 1.4×107 12 0.0 Exp (DTRC) - - - - -1.060 - - -3.410 - - -2.150
27 1.4×107 12 0.0 ReFRESCO SST 3
4138×10 -0.048 -1.018 -1.066 -2.164 -0.195 -2.359 -2.483 -0.009 -2.492
28 1.4×107 14 0.0 Exp (DTRC) - - - - -0.914 - - -4.586 - - -2.511
29 1.4×107 14 0.0 Exp (DTRC) - - - - -1.014 - - -4.569 - - -2.404
30 1.4×107 14 0.0 ReFRESCO SST 4138×103 0.017 -1.036 -1.019 -3.005 -0.229 -3.234 -2.793 -0.012 -2.805
31 1.4×107 16 0.0 Exp (DTRC) - - - - -0.809 - - -5.860 - - -2.750
32 1.4×107 16 0.0 Exp (DTRC) - - - - -0.947 - - -5.960 - - -2.670
33 1.4×107 16 0.0 ReFRESCO SST 4138×103 0.100 -1.052 -0.952 -3.973 -0.263 -4.236 -3.080 -0.014 -3.094
34 1.4×107 18 0.0 Exp (DTRC) - - - - -0.670 - - -7.355 - - -2.986
35 1.4×107 18 0.0 Exp (DTRC) - - - - -0.852 - - -7.438 - - -2.939
36 1.4×107 18 0.0 ReFRESCO SST 4138×10 3 0.207 -1.067 -0.860 -5.086 -0.297 -5.383 -3.353 -0.017 -3.370
37 1.4×107 18 0.0 ReFRESCO SA 4138×10 3 0.307 -1.074 -0.767 -5.371 -0.307 -5.678 -3.366 -0.017 -3.383
38 1.4×107 20 0.0 ReFRESCO SST 4138×103 0.336 -1.082 -0.745 -6.331 -0.330 -6.661 -3.613 -0.020 -3.633
39 1.4×107 22 0.0 ReFRESCO SST 4138×103 0.487 -1.096 -0.609 -7.691 -0.363 -8.053 -3.859 -0.024 -3.882
40 1.4×107 24 0.0 ReFRESCO SST 4138×103 0.661 -1.109 -0.448 -9.167 -0.395 -9.562 -4.092 -0.027 -4.119
41 1.4×107 36 0.0 ReFRESCO SST 300×103 2.162 -1.072 1.089 -20.679 -0.549 -21.228 -5.135 -0.046 -5.180
42 1.4×107 60 0.0 ReFRESCO SST 300×103 6.196 -0.881 5.315 -39.279 -0.823 -40.102 -6.565 -0.070 -6.635
43 1.4×107 75 0.0 ReFRESCO SST 300×103 6.769 -0.571 6.197 -43.006 -0.928 -43.935 -5.845 -0.060 -5.905
44 1.4×107 90 0.0 ReFRESCO SST 300×10 3 1.859 -0.093 1.766 -43.672 -0.980 -44.653 -0.527 -0.005 -0.531

Note: The forces and moments have been multiplied by 103

160 Appendices

Page intentionally left blank

Table pages – zig-zag 161

ZIG-ZAG 10◦ /10◦ PS, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 106002 106007 106011 203004 203006 203008 203010 203013
Approach speed V0 kn 10.00 10.00 10.00 10.50 10.00 10.00 10.00 10.00 10.00
Steering angle (δ) deg 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00
Check angle (ψ) deg -10.00 -10.00 -10.00 -10.00 -10.00 -10.00 -10.00 -10.00 -10.00
nd
2 execute time s 48.3 50.7 53.4 47.4 50.7 51.1 50.7 50.1 53.4
3rd execute time s 233.3 209.6 215.9 208.0 195.1 215.3 201.6 200.8 204.7
4th execute time s 491.0 444.1 442.8 432.3 411.0 451.7 427.9 - 429.3
th
5 execute time s 744.0 708.0 672.3 666.8 - 701.9 - - -
Reach s 216.9 190.6 196.7 189.1 174.6 197.9 182.3 181.0 185.9
Period s 510.8 498.3 456.5 464.9 360.3 486.6 377.2 - 375.9
Overshoot angle (ψ)
-1st execute deg 14.5 10.3 10.7 10.8 7.9 11.9 8.8 8.6 9.2
nd
-2 execute deg 31.3 23.5 21.1 20.3 20.8 24.6 23.8 21.6 22.7
-3rd execute deg 29.4 28.1 20.2 20.9 26.1 27.2 25.4 - 22.3
-4th execute deg 27.8 - 18.4 18.9 - 23.9 - - -
Overshoot time
-1st execute s 81.2 60.9 60.4 58.5 50.7 62.8 53.5 52.6 53.4
nd
-2 execute s 124.2 104.9 94.7 94.3 92.5 106.9 99.9 98.6 99.1
-3rd execute s 119.2 120.2 93.8 102.3 112.4 113.1 111.5 - 82.4
-4th execute s 115.6 - 94.5 95.5 - 109.4 - - -
Initial turning ability m 247.4 258.9 269.4 242.9 255.7 263.8 257.6 253.3 266.4
-Relative to Lpp - 1.61 1.68 1.75 1.58 1.66 1.72 1.68 1.65 1.73
Maximum heel angle (φ)
-1st execute deg 0.0 1.3 1.0 1.0 1.2 1.3 1.0 1.1 0.9
-2nd execute deg 0.0 -1.6 -1.5 -1.4 -1.3 -1.5 -1.4 -1.4 -1.3
rd
-3 execute deg 0.0 1.3 0.8 1.0 1.1 1.0 1.0 - 1.0
-4th execute deg 0.0 - -0.9 -1.2 - -1.4 - - -
Maximum drift angle (β)
-1st execute deg 4.5 4.3 4.7 4.6 4.2 4.7 4.0 4.8 4.3
-2nd execute deg -8.8 -8.4 -8.9 -9.3 -8.3 -9.1 -8.3 -8.0 -8.4
rd
-3 execute deg 10.5 10.1 10.5 10.1 9.3 10.0 9.6 - 9.7
-4th execute deg -10.3 - -11.1 -10.5 - -10.3 - - -
Maximum rate of turn (r)
-1st execute deg·s-1 -0.41 -0.39 -0.40 -0.41 -0.36 -0.42 -0.37 -0.37 -0.37
nd
-2 execute deg·s-1 0.64 0.58 0.55 0.55 0.56 0.59 0.58 0.57 0.58
-3rd execute deg·s-1 -0.66 -0.60 -0.53 -0.53 -0.60 -0.60 -0.60 -0.58
th
-4 execute deg·s-1 0.64 - 0.50 0.51 0.58
clear columns contain CFD results
gray columns contain experimental results
162 Appendices

ZIG-ZAG 10◦ /10◦ SB, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 106005 106006 106010 203003 203005 203007 203009 203011
Approach speed V0 kn 10.00 10.00 10.00 9.50 10.00 10.00 10.00 10.00 10.00
Steering angle (δ) deg -10.00 -10.00 -10.00 -10.00 -10.00 -10.00 -10.00 -10.00 -10.00
Check angle (ψ) deg 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00
nd
2 execute time s 48.3 46.9 48.2 51.5 57.8 65.0 64.4 49.0 51.8
3rd execute time s 233.3 212.8 211.8 214.2 215.3 220.0 232.9 204.9 200.8
4th execute time s 491.0 469.8 474.5 439.2 446.1 459.4 480.3 - 423.0
th
5 execute time s 744.0 721.6 726.0 666.9 685.2 705.4 - - -
Reach s 216.9 194.4 193.6 194.8 198.0 201.3 217.8 185.1 181.3
Period s 510.8 508.8 514.3 452.6 469.9 485.4 415.9 - 371.2
Overshoot angle (ψ)
-1st execute deg 14.5 11.6 10.8 9.6 9.5 9.2 12.3 9.4 8.6
nd
-2 execute deg 31.3 28.0 28.4 20.3 23.3 25.3 26.9 25.3 22.4
-3rd execute deg 29.4 25.5 25.1 18.8 23.8 25.0 24.1 - 22.2
-4th execute deg 27.8 28.3 - 19.5 25.4 24.2 - - -
Overshoot time
-1st execute s 81.2 62.4 63.5 54.6 55.8 54.6 65.6 57.0 50.7
nd
-2 execute s 124.2 120.2 127.1 96.9 100.9 109.6 113.3 109.4 96.8
-3rd execute s 119.2 115.2 110.1 95.2 107.5 113.4 111.9 - 99.7
-4th execute s 115.6 124.5 - 93.4 109.5 108.8 - - -
Initial turning ability m 247.4 242.6 250.4 249.1 292.5 333.4 334.8 254.7 263.2
-Relative to Lpp - 1.61 1.58 1.63 1.62 1.90 2.17 2.18 1.66 1.71
Maximum heel angle (φ)
-1st execute deg 0.0 -1.5 -1.6 -1.1 -1.3 -1.3 -1.6 -1.5 -1.3
-2nd execute deg 0.0 1.2 1.4 1.0 1.3 1.3 1.1 1.4 1.1
rd
-3 execute deg 0.0 -1.3 -1.3 -1.1 -1.4 -1.2 -1.1 - -1.2
-4th execute deg 0.0 1.2 - 0.8 1.0 1.1 - - -
Maximum drift angle (β)
-1st execute deg -4.5 -5.0 -4.7 -5.1 -5.1 -4.6 -5.1 -4.7 -4.8
-2nd execute deg 8.8 8.7 8.8 8.9 8.2 8.6 8.8 8.6 8.0
rd
-3 execute deg -10.5 -10.4 -10.4 -10.1 -10.2 -10.3 -10.4 - -9.9
-4th execute deg 10.3 11.0 - 10.4 10.0 9.9 - - -
Maximum rate of turn (r)
-1st execute deg·s-1 0.41 0.42 0.41 0.39 0.41 0.39 0.44 0.40 0.38
nd
-2 execute deg·s-1 -0.64 -0.59 -0.59 -0.53 -0.58 -0.58 -0.59 -0.57 -0.56
-3rd execute deg·s-1 0.66 0.59 0.59 0.51 0.59 0.60 0.60 0.58
th
-4 execute deg·s-1 -0.64 -0.59 - -0.51 -0.59 -0.58
clear columns contain CFD results
gray columns contain experimental results
Table pages – zig-zag 163

ZIG-ZAG 20◦ /20◦ PS, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 106004 106009 106013 204003
Approach speed V0 kn 10.00 10.00 9.50 10.00 10.00
Steering angle (δ) deg 20.00 20.00 20.00 20.00 20.00
Check angle (ψ) deg -20.00 -20.00 -20.00 -20.00 -20.00
nd
2 execute time s 51.3 51.5 53.4 53.2 51.7
3rd execute time s 226.8 227.1 226.3 223.3 218.1
4th execute time s 409.8 417.6 413.4 410.4 406.6
th
5 execute time s 585.8 619.7 611.1 606.9 601.9
Reach s 203.3 198.9 195.4 193.3 189.8
Period s 359.0 386.9 376.7 380.9 379.4
Overshoot angle (ψ)
-1st execute deg 20.3 16.6 14.3 14.4 15.0
nd
-2 execute deg 19.8 17.6 14.6 14.8 17.2
-3rd execute deg 17.1 17.5 13.8 14.3 15.5
-4th execute deg 16.5 15.3 12.0 13.1 14.5
Overshoot time
-1st execute s 64.7 57.8 53.5 51.3 54.1
nd
-2 execute s 63.8 64.8 55.5 50.5 59.6
-3rd execute s 57.5 70.1 59.7 62.0 62.5
-4th execute s 55.7 54.2 51.5 52.1 57.3
Initial turning ability m 260.1 261.1 253.2 259.8 262.0
-Relative to Lpp - 1.69 1.70 1.65 1.69 1.70
Maximum heel angle (φ)
-1st execute deg 0.0 2.2 1.4 1.5 2.0
-2nd execute deg 0.0 -1.7 -1.3 -1.3 -1.1
rd
-3 execute deg 0.0 1.2 1.0 0.9 1.2
-4th execute deg 0.0 -1.6 -1.3 -1.3 -1.4
Maximum drift angle (β)
-1st execute deg 8.6 8.7 9.1 8.7 8.6
-2nd execute deg -14.5 -13.4 -14.4 -14.2 -13.1
rd
-3 execute deg 14.6 13.5 14.0 13.5 14.1
-4th execute deg -14.4 -13.6 -14.3 -14.1 -13.0
Maximum rate of turn (r)
-1st execute deg·s-1 -0.75 -0.71 -0.67 -0.69 -0.71
nd
-2 execute deg·s-1 0.84 0.75 0.68 0.69 0.74
-3rd execute deg·s-1 -0.79 -0.71 -0.62 -0.63 -0.72
th
-4 execute deg·s-1 0.78 0.68 0.59 0.61 0.67
clear columns contain CFD results
gray columns contain experimental results
164 Appendices

ZIG-ZAG 20◦ /20◦ PS, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 106003 106008 106012 204002
Approach speed V0 kn 10.00 10.50 9.50 9.50 10.00
Steering angle (δ) deg -20.00 -20.00 -20.00 -20.00 -20.00
Check angle (ψ) deg 20.00 20.00 20.00 20.00 20.00
nd
2 execute time s 51.3 53.4 52.9 54.3 49.9
3rd execute time s 226.8 225.5 219.5 219.2 211.8
4th execute time s 409.8 428.2 420.3 420.9 410.2
th
5 execute time s 585.8 615.9 604.5 603.6 591.2
Reach s 203.3 199.1 190.5 188.9 184.6
Period s 359.0 391.0 377.3 379.5 379.4
Overshoot angle (ψ)
-1st execute deg 20.3 17.4 14.4 14.2 15.4
nd
-2 execute deg 19.8 17.9 16.1 16.3 17.8
-3rd execute deg 17.1 15.8 13.3 13.1 14.5
-4th execute deg 16.5 16.5 14.3 14.5 15.1
Overshoot time
-1st execute s 64.7 58.6 49.2 46.1 51.6
nd
-2 execute s 63.8 67.0 65.8 64.3 66.3
-3rd execute s 57.5 60.5 55.2 49.9 55.2
-4th execute s 55.7 68.1 57.0 61.4 56.8
Initial turning ability m 260.1 282.3 250.1 258.2 251.9
-Relative to Lpp - 1.69 1.84 1.63 1.68 1.64
Maximum heel angle (φ)
-1st execute deg 0.0 -1.7 -1.6 -1.5 -1.6
-2nd execute deg 0.0 1.4 1.1 1.2 1.2
rd
-3 execute deg 0.0 -1.6 -1.2 -1.4 -1.4
-4th execute deg 0.0 1.2 1.1 1.3 1.1
Maximum drift angle (β)
-1st execute deg -8.6 -8.9 -9.0 -9.0 -8.5
-2nd execute deg 14.5 14.0 13.9 14.3 13.8
rd
-3 execute deg -14.6 -13.6 -14.0 -14.0 -13.6
-4th execute deg 14.4 13.9 13.6 13.7 13.8
Maximum rate of turn (r)
-1st execute deg·s-1 0.75 0.73 0.67 0.67 0.72
nd
-2 execute deg·s-1 -0.84 -0.77 -0.71 -0.70 -0.77
-3rd execute deg·s-1 0.79 0.68 0.61 0.63 0.68
th
-4 execute deg·s-1 -0.78 -0.69 -0.64 -0.63 -0.70
clear columns contain CFD results
gray columns contain experimental results
Table pages – zig-zag 165

ZIG-ZAG 10◦ /10◦ PS, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 108002 108007 108011
Approach speed V0 kn 18.00 17.50 17.50 17.00
Steering angle (δ) deg 10.00 10.00 10.00 10.00
Check angle (ψ) deg -10.00 -10.00 -10.00 -10.00
nd
2 execute time s 27.3 27.4 26.9 27.9
3rd execute time s 134.5 112.6 116.7 113.8
4th execute time s 280.5 226.3 234.0 227.2
th
5 execute time s 422.5 352.9 361.3 -
Reach s 125.6 101.0 105.3 101.8
Period s 288.0 240.3 244.7 199.2
Overshoot angle (ψ)
-1st execute deg 16.2 9.4 11.0 9.6
nd
-2 execute deg 32.8 18.5 19.5 18.6
-3rd execute deg 30.2 22.5 22.4 23.2
-4th execute deg 28.8 18.5 19.0 -
Overshoot time
-1st execute s 48.0 29.8 33.5 30.3
nd
-2 execute s 70.4 45.6 49.0 45.3
-3rd execute s 66.8 52.3 53.9 55.3
-4th execute s 65.0 47.1 50.2 -
Initial turning ability m 251.5 241.0 238.1 245.4
-Relative to Lpp - 1.64 1.57 1.55 1.60
Maximum heel angle (φ)
-1st execute deg 0.0 3.5 3.5 3.5
-2nd execute deg 0.0 -4.2 -4.0 -4.2
rd
-3 execute deg 0.0 3.0 3.3 3.0
-4th execute deg 0.0 -3.6 -3.6 -
Maximum drift angle (β)
-1st execute deg 4.6 4.7 5.1 4.8
-2nd execute deg -9.1 -8.3 -8.2 -8.1
rd
-3 execute deg 10.6 9.6 9.7 9.6
-4th execute deg -10.4 -9.5 -9.3 -
Maximum rate of turn (r)
-1st execute deg·s-1 -0.75 -0.69 -0.71 -0.68
nd
-2 execute deg·s-1 1.17 0.98 0.99 0.98
-3rd execute deg·s-1 -1.19 -1.00 -1.01 -1.01
th
-4 execute deg·s-1 1.16 0.97 0.97
clear columns contain CFD results
gray columns contain experimental results
166 Appendices

ZIG-ZAG 10◦ /10◦ SB, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 108001 108006 108010
Approach speed V0 kn 18.00 17.50 17.50 17.50
Steering angle (δ) deg -10.00 -10.00 -10.00 -10.00
Check angle (ψ) deg 10.00 10.00 10.00 10.00
nd
2 execute time s 27.3 28.8 29.1 28.8
3rd execute time s 134.5 113.8 115.1 114.2
4th execute time s 280.5 236.4 235.9 237.0
th
5 execute time s 422.5 - 354.3 355.0
Reach s 125.6 102.6 103.6 102.7
Period s 288.0 207.7 239.2 240.9
Overshoot angle (ψ)
-1st execute deg 16.2 10.0 9.6 9.6
nd
-2 execute deg 32.8 21.6 21.3 21.8
-3rd execute deg 30.2 19.0 19.1 19.2
-4th execute deg 28.8 - 21.5 21.4
Overshoot time
-1st execute s 48.0 29.4 30.1 29.3
nd
-2 execute s 70.4 50.8 50.8 51.9
-3rd execute s 66.8 46.5 49.7 46.5
-4th execute s 65.0 - 52.5 51.4
Initial turning ability m 251.5 257.9 258.5 254.9
-Relative to Lpp - 1.64 1.68 1.68 1.66
Maximum heel angle (φ)
-1st execute deg 0.0 -4.1 -4.1 -4.0
-2nd execute deg 0.0 3.6 3.6 3.6
rd
-3 execute deg 0.0 -3.6 -3.7 -3.7
-4th execute deg 0.0 - 3.1 3.0
Maximum drift angle (β)
-1st execute deg -4.6 -4.5 -4.4 -4.6
-2nd execute deg 9.1 8.5 8.4 8.3
rd
-3 execute deg -10.6 -9.7 -9.2 -9.1
-4th execute deg 10.4 - 9.6 9.4
Maximum rate of turn (r)
-1st execute deg·s-1 0.75 0.72 0.71 0.71
nd
-2 execute deg·s-1 -1.17 -1.00 -0.99 -1.00
-3rd execute deg·s-1 1.19 0.98 0.99 0.99
th
-4 execute deg·s-1 -1.16 -0.97 -0.99
clear columns contain CFD results
gray columns contain experimental results
Table pages – zig-zag 167

ZIG-ZAG 20◦ /20◦ PS, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 108013 108009 108005
Approach speed V0 kn 18.00 17.50 17.50 17.50
Steering angle (δ) deg 20.00 20.00 20.00 20.00
Check angle (ψ) deg -20.00 -20.00 -20.00 -22.00
nd
2 execute time s 29.5 29.9 29.3 29.6
3rd execute time s 132.3 131.2 130.1 130.9
4th execute time s 237.0 237.5 237.6 238.3
th
5 execute time s 337.8 352.0 352.9 353.4
Reach s 119.3 115.1 113.7 114.3
Period s 205.5 217.8 222.7 222.4
Overshoot angle (ψ)
-1st execute deg 23.7 18.9 18.5 18.6
nd
-2 execute deg 21.5 18.3 18.9 18.5
-3rd execute deg 19.0 19.6 19.4 19.0
-4th execute deg 18.5 16.3 16.4 16.1
Overshoot time
-1st execute s 38.8 33.2 33.0 33.4
nd
-2 execute s 36.7 34.1 34.3 34.6
-3rd execute s 33.2 38.6 39.0 38.0
-4th execute s 32.4 30.0 31.3 30.6
Initial turning ability m 269.6 261.3 257.5 258.0
-Relative to Lpp - 1.75 1.70 1.68 1.68
Maximum heel angle (φ)
-1st execute deg 0.0 5.0 5.1 5.0
-2nd execute deg 0.0 -3.7 -4.1 -3.8
rd
-3 execute deg 0.0 3.1 3.0 3.4
-4th execute deg 0.0 -3.1 -3.0 -3.1
Maximum drift angle (β)
-1st execute deg 9.0 9.1 9.0 9.3
-2nd execute deg -14.9 -13.5 -13.8 -13.9
rd
-3 execute deg 14.9 14.0 13.7 14.0
-4th execute deg -14.7 -13.3 -13.4 -13.4
Maximum rate of turn (r)
-1st execute deg·s-1 -1.36 -1.25 -1.25 -1.25
nd
-2 execute deg·s-1 1.50 1.30 1.29 1.29
-3rd execute deg·s-1 -1.40 -1.24 -1.24 -1.22
th
-4 execute deg·s-1 1.39 1.16 1.16 1.16
clear columns contain CFD results
gray columns contain experimental results
168 Appendices

ZIG-ZAG 20◦ /20◦ SB, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 108003 108008 108012
Approach speed V0 kn 18.00 17.50 17.00 17.50
Steering angle (δ) deg -20.00 -20.00 -20.00 -20.00
Check angle (ψ) deg 20.00 20.00 22.00 20.00
nd
2 execute time s 29.5 29.0 30.4 30.1
3rd execute time s 132.3 124.6 126.1 126.8
4th execute time s 237.0 243.3 242.4 244.4
th
5 execute time s 337.8 347.6 347.0 349.6
Reach s 119.3 108.8 110.0 111.2
Period s 205.5 219.1 218.9 220.5
Overshoot angle (ψ)
-1st execute deg 23.7 18.5 17.3 18.6
nd
-2 execute deg 21.5 21.7 20.8 21.8
-3rd execute deg 19.0 16.3 16.5 16.8
-4th execute deg 18.5 19.3 18.6 19.3
Overshoot time
-1st execute s 38.8 30.6 28.3 32.5
nd
-2 execute s 36.7 41.4 39.7 40.8
-3rd execute s 33.2 30.4 31.0 30.9
-4th execute s 32.4 39.4 38.2 39.9
Initial turning ability m 269.6 257.6 264.4 266.1
-Relative to Lpp - 1.75 1.68 1.72 1.73
Maximum heel angle (φ)
-1st execute deg 0.0 -5.7 -5.6 -5.4
-2nd execute deg 0.0 3.2 3.0 3.2
rd
-3 execute deg 0.0 -3.3 -3.3 -3.4
-4th execute deg 0.0 3.2 3.3 3.2
Maximum drift angle (β)
-1st execute deg -9.0 -8.6 -8.8 -8.6
-2nd execute deg 14.9 14.3 14.1 14.4
rd
-3 execute deg -14.9 -13.4 -13.1 -13.2
-4th execute deg 14.7 13.8 13.8 13.9
Maximum rate of turn (r)
-1st execute deg·s-1 1.36 1.26 1.24 1.24
nd
-2 execute deg·s-1 -1.50 -1.34 -1.34 -1.36
-3rd execute deg·s-1 1.40 1.18 1.19 1.19
th
-4 execute deg·s-1 -1.39 -1.21 -1.22 -1.22
clear columns contain CFD results
gray columns contain experimental results
Table pages – turning-circle 169

TURNING CIRCLE 35◦ PS/SB, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 107002 CFD 107001
Approach speed V0 kn 10.00 10.00 10.00 9.50
Steering angle (δ) deg 35.00 35.00 -35.00 -35.00
Advance (AD) m 418.4 424.8 418.4 425.8
Transfer (TR) m -160.1 161.3 160.1 177.3
Tactical diameter (TD) m -424.7 414.3 424.7 437.7
Dstc m 305.9 356.5 305.9 380.2
T90 s 110.2 120.8 110.2 124.9
T180 s 239.5 249.5 239.5 262.6
T360 s 539.6 541.7 539.6 570.3
Tstc s 609.3 567.6 609.3 616.0
rexecute deg·s-1 -1.13 -0.97 1.13 0.96
rstc deg·s-1 -0.59 -0.63 0.59 0.58
rresidual deg·s-1 - 0.13 - 0.07
rresidual /rstc - - -0.21 - 0.12
Vstc kn 3.1 3.8 3.1 3.8
Vstc /V0 - 0.31 0.39 0.31 0.39
φmax,in deg 0.0 -1.0 0.0 1.0
φmax,out deg 0.0 0.8 0.0 -1.0
φstc deg 0.0 -0.2 0.0 0.1
φstc /φmax - - -0.22 - -0.07
Drift angle during turn (βstc ) deg 28.8 20.4 -28.8 -19.0
Dstc /Lpp ratio - 1.99 2.32 1.99 2.47
AD/Lpp ratio - 2.72 2.76 2.72 2.77
TD/Lpp ratio - -2.76 2.70 2.76 2.85
clear columns contain CFD results
gray columns contain experimental results
170 Appendices

TURNING CIRCLE 25◦ PS/SB, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 107005 CFD 107004
Approach speed V0 kn 10.00 10.50 10.00 9.50
Steering angle (δ) deg 25.00 25.00 -25.00 -25.00
Advance (AD) m 474.5 498.6 474.5 476.1
Transfer (TR) m -201.7 206.7 201.7 216.3
Tactical diameter (TD) m -502.3 507.9 502.3 528.1
Dstc m 399.9 447.6 399.9 476.6
T90 s 123.7 134.3 123.7 143.1
T180 s 252.5 272.9 252.5 295.9
T360 s 546.8 583.0 546.8 630.0
Tstc s 601.6 609.0 601.6 649.3
rexecute deg·s-1 -0.98 -0.86 0.98 0.81
rstc deg·s-1 -0.60 -0.59 0.60 0.55
rresidual deg·s-1 - 0.14 - 0.13
rresidual /rstc - - -0.23 - 0.23
Vstc kn 4.1 4.5 4.1 4.5
Vstc /V0 - 0.41 0.43 0.41 0.48
φmax,in deg 0.0 -1.3 0.0 0.5
φmax,out deg 0.0 0.8 0.0 -1.3
φstc deg 0.0 -0.2 0.0 0.1
φstc /φmax - - -0.26 - -0.09
Drift angle during turn (βstc ) deg 21.4 17.1 -21.4 -15.0
Dstc /Lpp ratio - 2.60 2.91 2.60 3.10
AD/Lpp ratio - 3.09 3.24 3.09 3.10
TD/Lpp ratio - -3.27 3.30 3.27 3.44
clear columns contain CFD results
gray columns contain experimental results
Table pages – turning-circle 171

TURNING CIRCLE 15◦ PS/SB, V0 = 10 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 107007 CFD 107008
Approach speed V0 kn 10.00 10.00 10.00 10.00
Steering angle (δ) deg 15.00 15.00 -15.00 -15.00
Advance (AD) m 581.7 593.5 581.7 573.3
Transfer (TR) m -271.8 275.0 271.8 289.8
Tactical diameter (TD) m -638.3 632.3 638.3 676.8
Dstc m 571.3 591.5 571.6 622.0
T90 s 150.0 162.4 150.0 168.2
T180 s 290.0 316.1 290.0 337.2
T360 s 603.1 655.8 603.1 696.3
Tstc s 637.7 685.4 636.9 715.8
rexecute deg·s-1 -0.80 -0.72 0.80 0.68
rstc deg·s-1 -0.56 -0.53 0.57 0.50
rresidual deg·s-1 - -0.04 - 0.03
rresidual /rstc - - 0.08 - 0.06
Vstc kn 5.5 5.3 5.5 5.3
Vstc /V0 - 0.55 0.52 0.55 0.54
φmax,in deg 0.0 -1.0 0.0 0.3
φmax,out deg 0.0 0.8 0.0 -1.0
φstc deg 0.0 -0.2 0.0 -0.1
φstc /φmax - - -0.26 - 0.15
Drift angle during turn (βstc ) deg 14.9 14.0 -14.9 -12.4
Dstc /Lpp ratio - 3.72 3.85 3.72 4.05
AD/Lpp ratio - 3.78 3.86 3.78 3.73
TD/Lpp ratio - -4.15 4.11 4.15 4.40
clear columns contain CFD results
gray columns contain experimental results
172 Appendices

TURNING CIRCLE 35◦ PS, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 104029 104027 104021 205003
Approach speed V0 kn 18.00 18.00 18.50 17.50 18.00
Steering angle (δ) deg 35.00 35.00 35.00 35.00 35.00
Advance (AD) m 431.7 419.8 416.0 414.2 423.6
Transfer (TR) m -160.4 164.8 162.2 158.2 163.7
Tactical diameter (TD) m -425.1 394.7 394.4 389.0 397.7
Dstc m 305.7 320.5 321.1 319.0 330.1
T90 s 62.6 62.5 61.7 62.8 63.0
T180 s 134.3 124.5 123.8 125.4 125.9
T360 s 301.1 262.3 262.9 263.9 265.8
Tstc s 338.4 276.1 277.8 276.5 284.5
rexecute deg·s-1 -2.04 -1.93 -1.96 -1.93 -1.90
rstc deg·s-1 -1.06 -1.30 -1.30 -1.30 -1.27
rresidual deg·s-1 - -0.49 -0.55 -0.55 -0.30
rresidual /rstc - - 0.38 0.43 0.42 0.24
Vstc kn 5.5 7.1 7.1 7.0 7.1
Vstc /V0 - 0.31 0.39 0.39 0.40 0.39
φmax,in deg 0.0 -2.4 -2.7 -2.6 -2.6
φmax,out deg 0.0 3.5 3.8 3.0 3.4
φstc deg 0.0 -1.0 -1.1 -1.1 -0.8
φstc /φmax - - -0.29 -0.28 -0.36 -0.23
Drift angle during turn (βstc ) deg 28.8 21.2 21.5 21.2 20.6
Dstc /Lpp ratio - 1.99 2.09 2.09 2.08 2.15
AD/Lpp ratio - 2.81 2.73 2.71 2.69 2.76
TD/Lpp ratio - -2.77 2.57 2.57 2.53 2.59
clear columns contain CFD results
gray columns contain experimental results
Table pages – turning-circle 173

TURNING CIRCLE 35◦ SB, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 104026 104019 104028 205002
Approach speed V0 kn 18.00 18.00 18.00 18.00 18.50
Steering angle (δ) deg -35.00 -35.00 -35.00 -35.00 -35.00
Advance (AD) m 431.7 443.5 441.4 433.8 438.7
Transfer (TR) m 160.4 177.3 179.9 180.6 175.0
Tactical diameter (TD) m 425.1 427.8 431.1 434.2 424.3
Dstc m 305.7 361.7 357.2 362.0 352.1
T90 s 62.6 66.4 66.1 65.9 64.9
T180 s 134.3 133.1 133.1 133.3 132.0
T360 s 301.1 282.2 281.6 283.4 283.4
Tstc s 338.4 304.5 302.4 306.2 304.8
rexecute deg·s-1 2.04 1.84 1.84 1.84 1.90
rstc deg·s-1 1.06 1.18 1.19 1.18 1.18
rresidual deg·s-1 - 0.34 0.37 0.37 0.51
rresidual /rstc - - 0.29 0.31 0.31 0.43
Vstc kn 5.5 7.3 7.2 7.2 7.1
Vstc /V0 - 0.31 0.40 0.40 0.40 0.38
φmax,in deg 0.0 2.1 2.2 2.0 2.1
φmax,out deg 0.0 -2.9 -3.0 -3.5 -3.2
φstc deg 0.0 0.3 0.3 0.3 -0.1
φstc /φmax - - -0.10 -0.09 -0.08 0.01
Drift angle during turn (βstc ) deg -28.8 -19.9 -19.6 -20.4 -20.1
Dstc /Lpp ratio - 1.99 2.35 2.32 2.36 2.29
AD/Lpp ratio - 2.81 2.89 2.87 2.82 2.85
TD/Lpp ratio - 2.77 2.78 2.80 2.83 2.76
clear columns contain CFD results
gray columns contain experimental results
174 Appendices

TURNING CIRCLE 25◦ PS/SB, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 104023 CFD 104022
Approach speed V0 kn 18.00 18.50 18.00 18.00
Steering angle (δ) deg 25.00 25.00 -25.00 -25.00
Advance (AD) m 484.9 464.6 484.9 515.2
Transfer (TR) m -201.9 204.2 201.9 218.4
Tactical diameter (TD) m -502.5 482.0 502.5 525.5
Dstc m 400.0 412.3 400.0 486.1
T90 s 69.8 68.7 69.8 75.5
T180 s 141.3 136.3 141.3 149.0
T360 s 304.8 284.8 304.8 311.1
Tstc s 334.3 297.6 334.3 333.3
rexecute deg·s-1 -1.76 -1.69 1.76 1.57
rstc deg·s-1 -1.08 -1.21 1.08 1.08
rresidual deg·s-1 - -0.48 - 0.37
rresidual /rstc - - 0.40 - 0.34
Vstc kn 7.3 8.5 7.3 8.9
Vstc /V0 - 0.41 0.46 0.41 0.49
φmax,in deg 0.0 -2.0 0.0 2.1
φmax,out deg 0.0 3.0 0.0 -3.3
φstc deg 0.0 -0.9 0.0 0.2
φstc /φmax - - -0.29 - -0.06
Drift angle during turn (βstc ) deg 21.3 17.5 -21.3 -15.7
Dstc /Lpp ratio - 2.60 2.68 2.60 3.16
AD/Lpp ratio - 3.15 3.02 3.15 3.35
TD/Lpp ratio - -3.27 3.14 3.27 3.42
clear columns contain CFD results
gray columns contain experimental results
Table pages – turning-circle 175

TURNING CIRCLE 15◦ PS/SB, V0 = 18 kn

DESIGNATION UNIT MAGNITUDE

TestNo - CFD 104025 CFD 104024
Approach speed V0 kn 18.00 18.00 18.00 18.50
Steering angle (δ) deg 15.00 15.00 -15.00 -15.00
Advance (AD) m 588.4 560.1 588.4 630.3
Transfer (TR) m -271.8 265.1 271.8 300.5
Tactical diameter (TD) m -638.2 614.2 638.2 681.3
Dstc m 571.2 577.8 571.4 662.7
T90 s 84.0 82.3 84.0 91.3
T180 s 161.8 160.4 161.8 175.7
T360 s 335.8 331.3 335.8 358.4
Tstc s 354.3 336.0 354.3 373.8
rexecute deg·s-1 -1.44 -1.42 1.44 1.28
rstc deg·s-1 -1.02 -1.07 1.02 0.96
rresidual deg·s-1 - -0.62 - 0.40
rresidual /rstc - - 0.58 - 0.42
Vstc kn 9.8 10.5 9.8 10.8
Vstc /V0 - 0.55 0.58 0.55 0.59
φmax,in deg 0.0 -1.4 0.0 1.3
φmax,out deg 0.0 3.2 0.0 -3.2
φstc deg 0.0 -0.3 0.0 -0.7
φstc /φmax - - -0.09 - 0.20
Drift angle during turn (βstc ) deg 14.9 13.5 -14.9 -12.4
Dstc /Lpp ratio - 3.72 3.76 3.72 4.31
AD/Lpp ratio - 3.83 3.64 3.83 4.10
TD/Lpp ratio - -4.15 4.00 4.15 4.43
clear columns contain CFD results
gray columns contain experimental results
176 Appendices

Page intentionally left blank

Figure pages – zig-zag 177

ZIG-ZAG 10◦ /10◦ PS, V0 = 10 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 100 200 300 400 500 600 700 800 900 1000

45
30
15
ψ [deg]

0
−15
−30
−45
0 100 200 300 400 500 600 700 800 900 1000

0.9
0.6
0.3
r [deg/s]

0
−0.3
−0.6
−0.9
0 100 200 300 400 500 600 700 800 900 1000

2.5
2
1.5
1
φ [deg]

0.5
0
−0.5
−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700 800 900 1000

6
5
4
V [m/s]

3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000

15
10
5
βG [deg]

0
−5
−10
−15
0 100 200 300 400 500 600 700 800 900 1000
203010 106011
time 203008 106007
203006 106002
203013 203004 zz05.14−10.00
thick blue lines: simulation, others: experiments
178 Appendices

ZIG-ZAG 10◦ /10◦ PS, V0 = 10 kn, TRACK

4000

3500

3000

2500
x [m]

2000
E

1500

1000

500

−1000 −500 0 500 1000

yE [m]
203010 106011
203008 106007
203006 106002
203013 203004 zz05.14−10.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 179

ZIG-ZAG 10◦ /10◦ SB, V0 = 10 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 100 200 300 400 500 600 700 800 900 1000

45
30
15
ψ [deg]

0
−15
−30
−45
0 100 200 300 400 500 600 700 800 900 1000

0.9
0.6
0.3
r [deg/s]

0
−0.3
−0.6
−0.9
0 100 200 300 400 500 600 700 800 900 1000

2.5
2
1.5
1
φ [deg]

0.5
0
−0.5
−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700 800 900 1000

6
5
4
V [m/s]

3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000

15
10
5
βG [deg]

0
−5
−10
−15
0 100 200 300 400 500 600 700 800 900 1000
203009 106010
time 203007 106006
203005 106005
203011 203003 zz05.14−−10.00
thick blue lines: simulation, others: experiments
180 Appendices

ZIG-ZAG 10◦ /10◦ SB, V0 = 10 kn, TRACK

4000

3500

3000

2500
x [m]

2000
E

1500

1000

500

−1000 −500 0 500 1000

yE [m]
203009 106010
203007 106006
203005 106005
203011 203003 zz05.14−−10.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 181

ZIG-ZAG 20◦ /20◦ PS, V0 = 10 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 100 200 300 400 500 600 700 800 900 1000

45
30
15
ψ [deg]

0
−15
−30
−45
0 100 200 300 400 500 600 700 800 900 1000

0.9
0.6
0.3
r [deg/s]

0
−0.3
−0.6
−0.9
0 100 200 300 400 500 600 700 800 900 1000

2.5
2
1.5
1
φ [deg]

0.5
0
−0.5
−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700 800 900 1000

6
5
4
V [m/s]

3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000

15
10
5
βG [deg]

0
−5
−10
−15
0 100 200 300 400 500 600 700 800 900 1000
106013
time 106009
106004
204003 zz05.14−20.00
thick blue lines: simulation, others: experiments
182 Appendices

ZIG-ZAG 20◦ /20◦ PS, V0 = 10 kn, TRACK

3000

2500

2000
x [m]
E

1500

1000

500

−1000 −500 0 500 1000

yE [m]
106013
106009
106004
204003 zz05.14−20.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 183

ZIG-ZAG 20◦ /20◦ SB, V0 = 10 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 100 200 300 400 500 600 700 800 900 1000

45
30
15
ψ [deg]

0
−15
−30
−45
0 100 200 300 400 500 600 700 800 900 1000

0.9
0.6
0.3
r [deg/s]

0
−0.3
−0.6
−0.9
0 100 200 300 400 500 600 700 800 900 1000

2.5
2
1.5
1
φ [deg]

0.5
0
−0.5
−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700 800 900 1000

6
5
4
V [m/s]

3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000

15
10
5
βG [deg]

0
−5
−10
−15
0 100 200 300 400 500 600 700 800 900 1000
106012
time 106008
106003
204002 zz05.14−−20.00
thick blue lines: simulation, others: experiments
184 Appendices

ZIG-ZAG 20◦ /20◦ SB, V0 = 10 kn, TRACK

3000

2500

2000
x [m]
E

1500

1000

500

−1000 −500 0 500 1000

yE [m]
106012
106008
106003
204002 zz05.14−−20.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 185

ZIG-ZAG 10◦ /10◦ PS, V0 = 18 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 50 100 150 200 250 300 350 400 450 500

45
30
15
ψ [deg]

0
−15
−30
−45
0 50 100 150 200 250 300 350 400 450 500

2
1.5
1
r [deg/s]

0.5
0
−0.5
−1
−1.5
−2
0 50 100 150 200 250 300 350 400 450 500

6
4.5
3
φ [deg]

1.5
0
−1.5
−3
−4.5
−6
0 50 100 150 200 250 300 350 400 450 500

10
8
V [m/s]

6
4
2
0
0 50 100 150 200 250 300 350 400 450 500

20
15
10
βG [deg]

5
0
−5
−10
−15
−20
0 50 100 150 200 250 300 350 400 450 500
108011
time 108007
108002
zz09.26−10.00
thick blue lines: simulation, others: experiments
186 Appendices

ZIG-ZAG 10◦ /10◦ PS, V0 = 18 kn, TRACK

4000

3500

3000

2500
x [m]

2000
E

1500

1000

500

−1500 −1000 −500 0 500 1000

yE [m]
108011
108007
108002
zz09.26−10.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 187

ZIG-ZAG 10◦ /10◦ SB, V0 = 18 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 50 100 150 200 250 300 350 400 450 500

45
30
15
ψ [deg]

0
−15
−30
−45
0 50 100 150 200 250 300 350 400 450 500

2
1.5
1
r [deg/s]

0.5
0
−0.5
−1
−1.5
−2
0 50 100 150 200 250 300 350 400 450 500

6
4.5
3
φ [deg]

1.5
0
−1.5
−3
−4.5
−6
0 50 100 150 200 250 300 350 400 450 500

10
8
V [m/s]

6
4
2
0
0 50 100 150 200 250 300 350 400 450 500

20
15
10
βG [deg]

5
0
−5
−10
−15
−20
0 50 100 150 200 250 300 350 400 450 500
108010
time 108006
108001
zz09.26−−10.00
thick blue lines: simulation, others: experiments
188 Appendices

ZIG-ZAG 10◦ /10◦ SB, V0 = 18 kn, TRACK

4000

3500

3000

2500
x [m]

2000
E

1500

1000

500

−1000 −500 0 500 1000 1500

yE [m]
108010
108006
108001
zz09.26−−10.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 189

ZIG-ZAG 20◦ /20◦ PS, V0 = 18 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 50 100 150 200 250 300 350 400 450 500

45
30
15
ψ [deg]

0
−15
−30
−45
0 50 100 150 200 250 300 350 400 450 500

2
1.5
1
r [deg/s]

0.5
0
−0.5
−1
−1.5
−2
0 50 100 150 200 250 300 350 400 450 500

6
4.5
3
φ [deg]

1.5
0
−1.5
−3
−4.5
−6
0 50 100 150 200 250 300 350 400 450 500

10
8
V [m/s]

6
4
2
0
0 50 100 150 200 250 300 350 400 450 500

20
15
10
βG [deg]

5
0
−5
−10
−15
−20
0 50 100 150 200 250 300 350 400 450 500
108005
time 108009
108013
zz09.26−20.00
thick blue lines: simulation, others: experiments
190 Appendices

ZIG-ZAG 20◦ /20◦ PS, V0 = 18 kn, TRACK

3000

2500

2000
x [m]
E

1500

1000

500

−1200 −1000 −800 −600 −400 −200 0 200 400 600 800
yE [m]
108005
108009
108013
zz09.26−20.00
thick blue lines: simulation, others: experiments
Figure pages – zig-zag 191

ZIG-ZAG 20◦ /20◦ SB, V0 = 18 kn, TIMESERIES

25
20
15
10
δ [deg]

5
0
−5
−10
−15
−20
−25
0 50 100 150 200 250 300 350 400 450 500

45
30
15
ψ [deg]

0
−15
−30
−45
0 50 100 150 200 250 300 350 400 450 500

2
1.5
1
r [deg/s]

0.5
0
−0.5
−1
−1.5
−2
0 50 100 150 200 250 300 350 400 450 500

6
4.5
3
φ [deg]

1.5
0
−1.5
−3
−4.5
−6
0 50 100 150 200 250 300 350 400 450 500

10
8
V [m/s]

6
4
2
0
0 50 100 150 200 250 300 350 400 450 500

20
15
10
βG [deg]

5
0
−5
−10
−15
−20
0 50 100 150 200 250 300 350 400 450 500
108012
time 108008
108003
zz09.26−−20.00
thick blue lines: simulation, others: experiments
192 Appendices

ZIG-ZAG 20◦ /20◦ SB, V0 = 18 kn, TRACK

3000

2500

2000
x [m]
E

1500

1000

500

−800 −600 −400 −200 0 200 400 600 800 1000 1200
yE [m]
108012
108008
108003
zz09.26−−20.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 193

TURNING CIRCLE 35◦ PS, V0 = 10 kn, TIMESERIES

40

30
δ [deg]

20

10

0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

0
−90
−180
−270
ψ [deg]

−360
−450
−540
−630
−720
−810
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

0
−0.2
−0.4
r [deg/s]

−0.6
−0.8
−1
−1.2
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

1.2
0.8
0.4
φ [deg]

0
−0.4
−0.8
−1.2
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

6
5
4
V [m/s]

3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

30
25
20
βG [deg]

15
10
5
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
time
107002
tc05.14−35.00
thick blue lines: simulation, others: experiments
194 Appendices

TURNING CIRCLE 35◦ PS, V0 = 10 kn, TRACK

2500

2000

1500
x [m]
E

1000

500

−1400 −1200 −1000 −800 −600 −400 −200 0 200 400

yE [m]

107002
tc05.14−35.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 195

TURNING CIRCLE 35◦ SB, V0 = 10 kn, TIMESERIES

−10
δ [deg]

−20

−30

−40
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

810
720
630
540
ψ [deg]

450
360
270
180
90
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

1.2
1
0.8
r [deg/s]

0.6
0.4
0.2
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

1.2
0.8
0.4
φ [deg]

0
−0.4
−0.8
−1.2
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

6
5
4
V [m/s]

3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

0
−5
−10
βG [deg]

−15
−20
−25
−30
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
time
107001
tc05.14−−35.00
thick blue lines: simulation, others: experiments
196 Appendices

TURNING CIRCLE 35◦ SB, V0 = 10 kn, TRACK

1200

1000

800
x [m]

600
E

400

200

0 100 200 300 400 500 600 700 800 900

yE [m]

107001
tc05.14−−35.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 197

TURNING CIRCLE 25◦ PS, V0 = 10 kn, TIMESERIES

40

30
δ [deg]

20

10

0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

0
−90
−180
−270
ψ [deg]

−360
−450
−540
−630
−720
−810
0 200 400 600 800 1000 1200 1400 1600 1800 2000

0
−0.2
−0.4
r [deg/s]

−0.6
−0.8
−1
−1.2
0 200 400 600 800 1000 1200 1400 1600 1800 2000

1.2
0.8
0.4
φ [deg]

0
−0.4
−0.8
−1.2
−1.6
−2
0 200 400 600 800 1000 1200 1400 1600 1800 2000

6
5
4
V [m/s]

3
2
1
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

30
25
20
βG [deg]

15
10
5
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time
107005
tc05.14−25.00
thick blue lines: simulation, others: experiments
198 Appendices

TURNING CIRCLE 25◦ PS, V0 = 10 kn, TRACK

1600

1400

1200

1000
x [m]
E

800

600

400

200

−1000 −800 −600 −400 −200 0 200

yE [m]

107005
tc05.14−25.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 199

TURNING CIRCLE 25◦ SB, V0 = 10 kn, TIMESERIES

−10
δ [deg]

−20

−30

−40
0 200 400 600 800 1000 1200 1400 1600 1800 2000

810
720
630
540
ψ [deg]

450
360
270
180
90
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

1.2
1
0.8
r [deg/s]

0.6
0.4
0.2
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

1.2
0.9
0.6
0.3
0
φ [deg]

−0.3
−0.6
−0.9
−1.2
−1.5
−1.8
−2.1
0 200 400 600 800 1000 1200 1400 1600 1800 2000

6
5
4
V [m/s]

3
2
1
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

0
−5
−10
βG [deg]

−15
−20
−25
−30
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time
107004
tc05.14−−25.00
thick blue lines: simulation, others: experiments
200 Appendices

TURNING CIRCLE 25◦ SB, V0 = 10 kn, TRACK

1200

1000

800

600
x [m]
E

400

200

0 100 200 300 400 500 600 700 800 900

yE [m]

107004
tc05.14−−25.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 201

TURNING CIRCLE 15◦ PS, V0 = 10 kn, TIMESERIES

40

30
δ [deg]

20

10

0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

0
−90
−180
−270
ψ [deg]

−360
−450
−540
−630
−720
−810
0 200 400 600 800 1000 1200 1400 1600 1800 2000

0
−0.2
−0.4
r [deg/s]

−0.6
−0.8
−1
−1.2
0 200 400 600 800 1000 1200 1400 1600 1800 2000

1.2
0.9
0.6
0.3
0
φ [deg]

−0.3
−0.6
−0.9
−1.2
−1.5
−1.8
−2.1
0 200 400 600 800 1000 1200 1400 1600 1800 2000

6
5
4
V [m/s]

3
2
1
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

30
25
20
βG [deg]

15
10
5
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time
107007
tc05.14−15.00
thick blue lines: simulation, others: experiments
202 Appendices

TURNING CIRCLE 15◦ PS, V0 = 10 kn, TRACK

1200

1000

800

600
x [m]
E

400

200

−800 −700 −600 −500 −400 −300 −200 −100 0

yE [m]

107007
tc05.14−15.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 203

TURNING CIRCLE 15◦ SB, V0 = 10 kn, TIMESERIES

−10
δ [deg]

−20

−30

−40
0 200 400 600 800 1000 1200 1400 1600 1800 2000

810
720
630
540
ψ [deg]

450
360
270
180
90
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

1.2
1
0.8
r [deg/s]

0.6
0.4
0.2
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

1.2
0.9
0.6
0.3
0
φ [deg]

−0.3
−0.6
−0.9
−1.2
−1.5
−1.8
−2.1
0 200 400 600 800 1000 1200 1400 1600 1800 2000

6
5
4
V [m/s]

3
2
1
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

0
−5
−10
βG [deg]

−15
−20
−25
−30
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time
107008
tc05.14−−15.00
thick blue lines: simulation, others: experiments
204 Appendices

TURNING CIRCLE 15◦ SB, V0 = 10 kn, TRACK

1000

800

600
x [m]
E

400

200

−200
−100 0 100 200 300 400 500 600 700 800
yE [m]

107008
tc05.14−−15.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 205

TURNING CIRCLE 35◦ PS, V0 = 18 kn, TIMESERIES

40

30
δ [deg]

20

10

0
0 100 200 300 400 500 600 700

0
−90
−180
−270
ψ [deg]

−360
−450
−540
−630
−720
−810
0 100 200 300 400 500 600 700

0
−0.5
r [deg/s]

−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700

2
φ [deg]

−2

−4
0 100 200 300 400 500 600 700

10
8
V [m/s]

6
4
2
0
0 100 200 300 400 500 600 700

30
25
20
βG [deg]

15
10
5
0
0 100 200 300 400 500 600 700
104021
time 104027
104029
205003 tc09.26−35.00
thick blue lines: simulation, others: experiments
206 Appendices

TURNING CIRCLE 35◦ PS, V0 = 18 kn, TRACK

1200

1000

800
x [m]

600
E

400

200

−900 −800 −700 −600 −500 −400 −300 −200 −100 0

yE [m]
104021
104027
104029
205003 tc09.26−35.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 207

TURNING CIRCLE 35◦ SB, V0 = 18 kn, TIMESERIES

−10
δ [deg]

−20

−30

−40
0 100 200 300 400 500 600 700

810
720
630
540
ψ [deg]

450
360
270
180
90
0
0 100 200 300 400 500 600 700

2.5
2
r [deg/s]

1.5
1
0.5
0
0 100 200 300 400 500 600 700

3
2
1
φ [deg]

0
−1
−2
−3
−4
0 100 200 300 400 500 600 700

10
8
V [m/s]

6
4
2
0
0 100 200 300 400 500 600 700

0
−5
−10
βG [deg]

−15
−20
−25
−30
0 100 200 300 400 500 600 700
104028
time 104019
104026
205002 tc09.26−−35.00
thick blue lines: simulation, others: experiments
208 Appendices

TURNING CIRCLE 35◦ SB, V0 = 18 kn, TRACK

1200

1000

800
x [m]

600
E

400

200

0 100 200 300 400 500 600 700 800 900

yE [m]
104028
104019
104026
205002 tc09.26−−35.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 209

TURNING CIRCLE 25◦ PS, V0 = 18 kn, TIMESERIES

40

30
δ [deg]

20

10

0
0 100 200 300 400 500 600 700 800 900 1000

0
−90
−180
−270
ψ [deg]

−360
−450
−540
−630
−720
−810
0 100 200 300 400 500 600 700 800 900 1000

0
−0.5
r [deg/s]

−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700 800 900 1000

4.5
3.6
2.7
1.8
φ [deg]

0.9
0
−0.9
−1.8
−2.7
−3.6
0 100 200 300 400 500 600 700 800 900 1000

10
8
V [m/s]

6
4
2
0
0 100 200 300 400 500 600 700 800 900 1000

30
25
20
βG [deg]

15
10
5
0
0 100 200 300 400 500 600 700 800 900 1000
time
104023
tc09.26−25.00
thick blue lines: simulation, others: experiments
210 Appendices

TURNING CIRCLE 25◦ PS, V0 = 18 kn, TRACK

1000

800

600
x [m]
E

400

200

−900 −800 −700 −600 −500 −400 −300 −200 −100 0

yE [m]

104023
tc09.26−25.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 211

TURNING CIRCLE 25◦ SB, V0 = 18 kn, TIMESERIES

−10
δ [deg]

−20

−30

−40
0 100 200 300 400 500 600 700 800 900 1000

810
720
630
540
ψ [deg]

450
360
270
180
90
0
0 100 200 300 400 500 600 700 800 900 1000

2.5
2
r [deg/s]

1.5
1
0.5
0
0 100 200 300 400 500 600 700 800 900 1000

3
2
1
φ [deg]

0
−1
−2
−3
−4
0 100 200 300 400 500 600 700 800 900 1000

10
8
V [m/s]

6
4
2
0
0 100 200 300 400 500 600 700 800 900 1000

0
−5
−10
βG [deg]

−15
−20
−25
−30
0 100 200 300 400 500 600 700 800 900 1000
time
104022
tc09.26−−25.00
thick blue lines: simulation, others: experiments
212 Appendices

TURNING CIRCLE 25◦ SB, V0 = 18 kn, TRACK

1000

800

600
x [m]
E

400

200

0 100 200 300 400 500 600 700 800 900

yE [m]

104022
tc09.26−−25.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 213

TURNING CIRCLE 15◦ PS, V0 = 18 kn, TIMESERIES

40

30
δ [deg]

20

10

0
0 100 200 300 400 500 600 700 800 900 1000

0
−90
−180
−270
ψ [deg]

−360
−450
−540
−630
−720
−810
0 100 200 300 400 500 600 700 800 900 1000

0
−0.5
r [deg/s]

−1
−1.5
−2
−2.5
0 100 200 300 400 500 600 700 800 900 1000

4
3.2
2.4
φ [deg]

1.6
0.8
0
−0.8
−1.6
−2.4
0 100 200 300 400 500 600 700 800 900 1000

10
8
V [m/s]

6
4
2
0
0 100 200 300 400 500 600 700 800 900 1000

30
25
20
βG [deg]

15
10
5
0
0 100 200 300 400 500 600 700 800 900 1000
time
104025
tc09.26−15.00
thick blue lines: simulation, others: experiments
214 Appendices

TURNING CIRCLE 15◦ PS, V0 = 18 kn, TRACK

1000

900

800

700

600
x [m]

500
E

400

300

200

100

−800 −700 −600 −500 −400 −300 −200 −100 0

yE [m]

104025
tc09.26−15.00
thick blue lines: simulation, others: experiments
Figure pages – turning-circle 215

TURNING CIRCLE 15◦ SB, V0 = 18 kn, TIMESERIES

−10
δ [deg]

−20

−30

−40
0 100 200 300 400 500 600 700 800 900 1000

810
720
630
540
ψ [deg]

450
360
270
180
90
0
0 100 200 300 400 500 600 700 800 900 1000

2.5
2
r [deg/s]

1.5
1
0.5
0
0 100 200 300 400 500 600 700 800 900 1000

2.4
1.6
0.8
φ [deg]

0
−0.8
−1.6
−2.4
−3.2
−4
0 100 200 300 400 500 600 700 800 900 1000

10
8
V [m/s]

6
4
2
0
0 100 200 300 400 500 600 700 800 900 1000

0
−5
−10
βG [deg]

−15
−20
−25
−30
0 100 200 300 400 500 600 700 800 900 1000
time
104024
tc09.26−−15.00
thick blue lines: simulation, others: experiments
216 Appendices

TURNING CIRCLE 15◦ SB, V0 = 18 kn, TRACK

1000

800

600
x [m]
E

400

200

−200
−100 0 100 200 300 400 500 600 700 800
yE [m]

104024
tc09.26−−15.00
thick blue lines: simulation, others: experiments