Dynamic Analysis of ROV Operation

Fredrik Rødne Jenssen

Marine Technology
Submission date: June 2015
Supervisor: Svein Sævik, IMT
Co-supervisor: Kjell Larsen, IMT

Norwegian University of Science and Technology

Department of Marine Technology
 NTNU Trondheim
Norges teknisk-naturvitenskapelige universitet
Institutt for marin teknikk

MASTER THESIS SPRING 2015

for
Stud. tech. Fredrik Rødne Jenssen

Dynamic Analysis of ROV Operation

Dynamisk analyse av ROV operasjon

The background for this project is related to the ROV operations performed from the Snorre
B semi-submersible platform where water entry limits the weather window for such
operations. The ROV is launched through a separate deck opening by means of a winch and
a cable running over a banana shaped sheave. Due to space limitations, the distance from the
sheave to the winch is quite small leading to cable fatigue problems. The distance from the
water entry point to the pontoon, further limits the weather window before clashing occurs.
The main purpose of the master thesis work is to perform parameter studies with a numerical
model in order to evaluate the ROV operation criteria, possibly also the amount of inherent
conservatism in the current procedure. Previously, the basis for modelling the ROV launch
operation has been established as part of the project work, such that the effect of different
means for optimizing and improving the procedure can be investigated and documented prior
to in-field testing. The project part included:

1. Familiarization with respect to the Snorre B ROV operation details including the basis
for the operation criteria presently applied . This was obtained by the candidate having
a summer job in Statoil, Stavanger.
2. Literature study, including relevant standards for lift operations, theoretical basis for
computational tools like Sima (Simo/Riflex), familiarization with the Sima tool,
hydrodynamic coefficients.
3. Establish a model that can be used to evaluate the lift operations and to be used as basis
for comparisons/calibration with regard to field experience. This requires input data in
terms of environmental and platform motion characteristics as well as geometry
details.

The master thesis is to continue from the above including:

4. Establish strategies for re-modelling with respect to the effect of different measures that
can be implemented to extend the weather window.

5. Perform sensitivity studies with respect to governing parameters to document the

performance of the different methods and procedures with respect to statistic
variability and using the current operation criteria as a reference.

6. Conclusions and recommendations for further work

All necessary input data is assumed to be provided by Statoil.

i
 NTNU Fakultet for ingeniørvitenskap og teknologi
Norges teknisk-naturvitenskapelige universitet Institutt for marin teknikk

The work scope may prove to be larger than initially anticipated. Subject to approval from the
supervisors, topics may be deleted from the list above or reduced in extent.

In the thesis the candidate shall present his personal contribution to the resolution of problems
within the scope of the thesis work

Theories and conclusions should be based on mathematical derivations and/or logic reasoning
identifying the various steps in the deduction.

The candidate should utilise the existing possibilities for obtaining relevant literature.

Thesis format
The thesis should be organised in a rational manner to give a clear exposition of results,
assessments, and conclusions. The text should be brief and to the point, with a clear language.
Telegraphic language should be avoided.

The thesis shall contain the following elements: A text defining the scope, preface, list of
contents, summary, main body of thesis, conclusions with recommendations for further work, list
of symbols and acronyms, references and (optional) appendices. All figures, tables and
equations shall be numerated.

The supervisors may require that the candidate, in an early stage of the work, presents a written
plan for the completion of the work.
The original contribution of the candidate and material taken from other sources shall be clearly
defined. Work from other sources shall be properly referenced using an acknowledged
referencing system.

The report shall be submitted in two copies:

- Signed by the candidate
- The text defining the scope included
- In bound volume(s)
- Drawings and/or computer prints which cannot be bound should be organised in a separate
folder.
Ownership
NTNU has according to the present rules the ownership of the thesis. Any use of the thesis has to be
approved by NTNU (or external partner when this applies). The department has the right to use the
thesis as if the work was carried out by a NTNU employee, if nothing else has been agreed in
advance.
Thesis supervisors:
Prof. Svein Sævik, NTNU
Prof. II Kjell Larsen, Statoil

Deadline: June 10, 2015

Trondheim, January , 2015

Svein Sævik

ii
Preface

This report is the result of a Master thesis carried out during the spring semester of 2015 at the
Department of Marine Technology, Norwegian University of Science and Technology
(NTNU).

The background for the work is an initiative by Olav Bruset, Platform Manager at Snorre B,
Statoil. During the spring semester of 2014, he organized a summer internship at Statoil in
Stavanger for two students at the Marine Technology Department. The intention was to
investigate ROV operations at Snorre B and identify limitations of the system. This work was
continued by evaluating the launch and recovery operation in the Specialization Project course
during the fall semester of 2014. Literature study and familiarization with the computer
program SIMO were the main intention. In addition, a preliminary model of the ROV and the
launch and recovery system was made.

In the master thesis work, the preliminary model has been modified and validated to give a
better representation of the operation. Much time was used to obtain usable results from
SIMO. Motion measurements of the ROV were conducted by borrowing an IMU from NTNU,
which the ROV operators at Snorre B mounted onto the ROV and logged data from.
Preparations for the measurements took longer time than planned and resulted in limited
extent of results and references for verification. Also post processing the results proved to be
more difficult than anticipated, partly due to uncertainties in the IMU orientation data.
However, the measurements obtained have given an indication of the ROV wave zone
behaviour and been used extensively in the discussion part of the report. In addition, the
process has been very educational.

The work has been carried out under supervision of Svein Sævik and Kjell Larsen, who have
provided much appreciated guidance and support. I would also like to thank Olav Bruset for
initiating the work, organizing the summer internship and follow up the project. In addition, I
want to thank Professor Martin Ludvigsen and Mauro Candeloro for lending out the IMU and
Saipem for organizing and conducting the measurements at Snorre B.

Trondheim, June, 2015

Fredrik Rødne Jenssen

iii
Sammendrag

Den fjernstyrte undervannsfarkosten (ROV) på den halvt nedsenkbare plattformen Snorre B er

avgjørende for å utføre bore- og produksjonsaktiviteter. Dagens værkriterie for å sjøsette
ROVen er satt til fire meter signifikant bølgehøyde. Dette fører til begrensninger for
undervannsoperasjoner. Værkriteriet er basert på forenklede beregninger beskrevet i anbefalte
metoder fra DNV. Disse er antatt å overvurdere kreftene på grunn av antagelser og
usikkerheter.

I dette arbeidet har sjøsetting og opptak av ROVen blitt vurdert på bakgrunn av analyser i SIMO
og bevegelsesmålinger utført på ROVen. I tillegg er de forenklede beregningene brukt som
sammenlikningsgrunnlag. Hensikten har vært å identifisere kritiske faktorer for fastsettelsen av
værkriteriet. Videre har tiltak for å redusere effekten av de kritiske faktorene blitt evaluert.

En foreløpig modell av ROVen og sjøsettingssystemet ble laget som en del av prosjektoppgaven

i høstsemesteret 2014. Denne modellen har blitt endret og verifisert for å gi en bedre
representasjon av operasjonen. Videre er modellen brukt til å forstå bakgrunnen for effekter
som påvirker værkriteriet. Basert på resultatene, ble en modell av skiveopphenget montert på
Snorre B inkludert ved hjelp av to metoder. Akselerasjons-, vinkelhastighet- og
retningsmålinger ble utført ved å montere en Inertial Measurement Unit (IMU) på ROVen.
Dette ble brukt for å vurdere den faktiske oppførselen av ROVen, og til en viss grad verifisere
SIMO resultatene.

Basert på analysene i SIMO er impulslaster i løftekabelen (umbilical) funnet som den mest
kritiske faktoren. Disse oppstår som følge av at drag-krefter på ROVen i bølgesonen gir slakk i
løftekabelen. Opptak av ROVen og fare for kollisjon med plattformskroget er funnet å være av
mindre betydning for værkriteriet. Også bevegelsesmålingene indikerer at slakk i kabelen kan
ha oppstått. De horisontale ROV-akselerasjonene fra målingene i bølgesonen var i samme
størrelsesorden som resultatene fra SIMO. Likevel fører mangel på posisjonsreferanser og
usikkerheter til at det er vanskelig å verifisere bevegelsesmålingene.

På bakgrunn av impulsbelastningene i kabelen er et effektivt skiveoppheng for å redusere

stivheten i systemet antatt å være det beste alternativet for å redusere dimensjonerende krefter
i systemet. De to modellene av skiveopphenget i SIMO reduserte kabelstrekket under
impulsbelastninger, men forkortet ikke perioden med slakk kabel. Imidlertid gjør forenklinger
av systemet og usikkerheter rundt stivheten i opphenget det vanskelig å vurdere effekten av et
slikt system. Forutsatt at målte impulsakselerasjonen er på grunn av slakk i kabelen, kan
målingene indikerer begrenset påvirkning i stivheten med dagens skiveoppheng.

iv
Summary

At the semi-submersible rig Snorre B, the work class remotely operated vehicle (WROV) is
essential to perform drilling and production activities. Currently, the weather criterion for
launching the ROV is set to four meters significant wave height. This leads to restrictions for
subsea operations dependent on the ROV. The weather criterion is based on simplified
calculations from DNV’s Recommended Practice, which is suspected to overestimate the forces
due to assumptions and uncertainties.

In this work, the launch and recovery operation has been evaluated based on time domain
analysis in SIMO and motion measurements of the ROV. In addition, simplified calculations
have been conducted for comparison. The intention has been to determine critical factors
when setting the weather criterion for the operation. Further, measures to reduce the effect of
the critical factors have been evaluated.

A preliminary model of the ROV and the launch and recover system was made in the project
thesis work during the fall semester in 2014. This model has been modified and verified to give
a better representation of the operation. Further, the model has been used to understand the
background for effects influencing the weather criterion. Based on the results, a representation
of the sheave suspension mounted at Snorre B was included by two methods. Accelerations,
angular velocities and orientation measurements were done by mounting an inertial
measurement unit on the ROV. This was used to evaluate the actual behaviour of the ROV, and
to some extend verify the SIMO results.

Based on the SIMO analysis, large snap loads in the umbilical during launch is identified as a
critical factor. This is a consequence of slack umbilical due to drag force from vertical water
particle velocity. The recovery phase and pontoon impact risk are of less importance for the
weather criterion. Occurrence of snap load in the umbilical is also indicated from the motion
measurements. Horizontal accelerations in the wave zone were in the same magnitude as the
results from SIMO. However, lack of position references and uncertainties in the measurements
makes them difficult to verify.

Because of the snap loads, an effective sheave suspension system seems to be the best option to
reduce the dimensioning forces. The two sheave suspension stiffness models in SIMO reduced
the umbilical tension during snap loads, but did not reduce the duration of slack umbilical.
However, simplifications of the system and uncertainties in input parameters makes it difficult
to evaluate the effect. Assuming measured impulse acceleration is due to a snap load, the
measurements can indicate limited effect of the current sheave suspension system.

v
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Structure of the report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Problem description 5
2.1 Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Snorre B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.2 ROV system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.3 Launch and recovery system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Phases during launch and recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.1 Lift off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.2 Lifting below deck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.3 Lifting through the wave zone . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.4 Operating below the wave zone . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3 Theory of the analysis 13

3.1 Wave zone analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.1 Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.2 Motions and forces in the wave zone . . . . . . . . . . . . . . . . . . . . . . . 15
3.1.3 Hydrodynamic properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Simplified analysis method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.1 Main assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.2 Force calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.3 Accept criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

vii
 3.3 SIMO time domain analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3.1 Coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3.2 Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3.3 Force models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4 Procedure 27
4.1 DNV Simplified method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1.1 Calculation method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1.2 Input data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.1.3 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2 SIMO Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2.1 Platform motions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2.2 Launch and recovery system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.2.3 ROV system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2.4 Sea state representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.2.5 Verification of SIMO model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.3 Motion measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.3.1 Equipment and installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.3.2 Test procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.3.3 Post processing measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.3.4 Uncertainties and validation of measurements . . . . . . . . . . . . . . . . . 42

5 Results and discussion 43

5.1 Simplified method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.2 Motion measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.2.1 Above the water surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2.2 Wave zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2.3 Acceleration spikes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.3 Time domain analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.3.1 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.3.2 Winch speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.3.3 Horizontal translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.3.4 ROV recovery phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3.5 Weather condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.3.6 Snap loads in SIMO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.3.7 Stiffness and damping in the launching system . . . . . . . . . . . . . . . . . 62
5.4 Comparison of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

viii
 5.4.1 Steady accelerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.4.2 Vertical acceleration spikes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.4.3 Umbilical tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.5 Weather criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6 Conclusions and further Work 69

6.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.2 Recommendations for Further Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Bibliography 73

A Motion measurements i

B SIMO analysis vii

C Slender elements properties xvi

D IMU Calibration certificate xviii

ix
List of Figures

2.1 Snorre B, view from north-west. Photo by Harald Pettersen, Statoil. . . . . . . . . . 6

2.2 The WROV system as seen from behind. . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Drawing of the sheave suspension in front of the winch (JMC Engineering, 2006b). 8
2.4 The blue cursor frame on top of the TMS illustrating clearance to the sheave. . . . 10
2.5 Drawing of the ROV below the deck opening. Adapted from Bruset (2014) . . . . . 11

4.1 Visualization of the ROV and LARS model. . . . . . . . . . . . . . . . . . . . . . . . . 32

4.2 Time domain analysis without waves. . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Position of IMU as seen in front of the ROV. Photos received form the ROV operators. 38
4.4 Origin of the coordinate systems from the IMU and the ROV as modelled in SIMO 40

5.1 Wave zone accelerations during launch 2015-04-14 at 02:50 . . . . . . . . . . . . . . 47

5.2 Acceleration spike during launch 2015-04-13 at 01:47 . . . . . . . . . . . . . . . . . . 48
5.3 Winch stop below wave zone during launch 2015-04-13 at 01:47. . . . . . . . . . . . 50
5.4 Sensitivity analysis of added mass coefficient on umbilical tension. . . . . . . . . . 51
5.5 Sensitivity analysis of drag coefficient on umbilical tension. . . . . . . . . . . . . . . 52
5.6 Depth dependent coefficients sensitivity for the lower element frame. . . . . . . . . 53
5.7 Umbilical tension in H s = 4 m and T p = 8 s without horizontal drag and added
mass coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.8 ROV horizontal and vertical position during 30 wave seeds in H s = 4 m and T p = 7 s. 56
5.9 Pontoon horizontal and vertical position relative to the deck opening position dur-
ing 30 wave seeds in H s = 5 m and T p = 9 s. . . . . . . . . . . . . . . . . . . . . . . . . 57
5.10 Umbilical tension during recovery for 30 wave seeds in H s = 4 m and T p = 8 s. . . . 58
5.11 Vertical wave kinematics compared to umbilical tension. . . . . . . . . . . . . . . . 60
5.12 ROV vertical motion during snap load in H s = 4 m and T p = 8 s. . . . . . . . . . . . 61
5.13 Tension and vertical acceleration variation due to change in stiffness. . . . . . . . . 63
5.14 ROV acceleration in global y-direction versus position relative to wave elevation in
H s = 4 m and T p = 8 s with winch speed 0.4 m/s. . . . . . . . . . . . . . . . . . . . . 65

xi
5.15 ROV motion and wave kinetics in H s = 4 m and T p = 8 s with winch speed 0.4 m/s. 66

A.1 Launch from 2015-04-02 at 19:58 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

A.2 Launch from 2015-04-04 at 14:19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
A.3 Launch from 2015-04-05 at 03:47 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
A.4 Launch from 2015-04-09 at 22:36 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
A.5 Recovery from 2015-04-12 at 20:32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
A.6 Launch from 2015-04-13 at 01:47 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
A.7 Recovery from 2015-04-14 at 00:15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
A.8 Launch from 2015-04-14 02:50 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

B.1 Umbilical tension in H s = 4 m and T p = 7 s. . . . . . . . . . . . . . . . . . . . . . . . vii

B.2 Umbilical tension in H s = 4 m and T p = 8 s. . . . . . . . . . . . . . . . . . . . . . . . viii
B.3 Umbilical tension in H s = 4 m and T p = 9 s. . . . . . . . . . . . . . . . . . . . . . . . viii
B.4 Umbilical tension in H s = 4 m and T p = 10 s. . . . . . . . . . . . . . . . . . . . . . . . ix
B.5 Umbilical tension in H s = 4 m and T p = 11 s. . . . . . . . . . . . . . . . . . . . . . . . ix
B.6 Umbilical tension in H s = 4 m and T p = 12 s. . . . . . . . . . . . . . . . . . . . . . . . x
B.7 Umbilical tension in H s = 4 m and T p = 13 s. . . . . . . . . . . . . . . . . . . . . . . . x
B.8 Umbilical tension in H s = 4 m and T p = 14 s. . . . . . . . . . . . . . . . . . . . . . . . xi
B.9 Umbilical tension in H s = 5 m and T p = 8 s. . . . . . . . . . . . . . . . . . . . . . . . xi
B.10 Umbilical tension in H s = 5 m and T p = 9 s. . . . . . . . . . . . . . . . . . . . . . . . xii
B.11 Umbilical tension in H s = 5 m and T p = 10 s. . . . . . . . . . . . . . . . . . . . . . . . xii
B.12 Umbilical tension in H s = 5 m and T p = 11 s. . . . . . . . . . . . . . . . . . . . . . . . xiii
B.13 Umbilical tension in H s = 5 m and T p = 12 s. . . . . . . . . . . . . . . . . . . . . . . . xiii
B.14 Umbilical tension in H s = 5 m and T p = 13 s. . . . . . . . . . . . . . . . . . . . . . . . xiv
B.15 Umbilical tension in H s = 5 m and T p = 14 s. . . . . . . . . . . . . . . . . . . . . . . . xiv
B.16 Umbilical tension in H s = 5 m and T p = 15 s. . . . . . . . . . . . . . . . . . . . . . . . xv

xii
List of Tables

2.1 Properties of the ROV system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Properties of the launch and recovery system. . . . . . . . . . . . . . . . . . . . . . . 8

4.1 Wave height and sheave motion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.2 Properties of the umbilical. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3 ROV properties at the chosen load case. . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4 Hydrodynamic coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.5 Static force, vertical added mass and calculated velocities and accelerations. . . . . 29
4.6 Position of LARS components relative to the body coordinate system of Snorre B. . 30
4.7 Simple wire coupling input. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.8 Tensioner input. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.9 Mass properties of the ROV system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.10 Hydrodynamic properties of the ROV system. . . . . . . . . . . . . . . . . . . . . . . 33
4.11 Depth dependency of the horizontal slender element properties. . . . . . . . . . . . 34
4.12 Zero-up-crossing period for the JONSWAP spectrum. . . . . . . . . . . . . . . . . . . 35
4.13 IMU sensor specifications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

5.1 Static and hydrodynamic loads from simplified calculations. . . . . . . . . . . . . . 43

5.2 Snap load calculation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.3 Received log files with corresponding wave condition. . . . . . . . . . . . . . . . . . 45
5.4 Maximum positive acceleration in wave zone with approximate duration. . . . . . 48
5.5 Snap load characteristics variation due to winch speed in H s = 4 m. . . . . . . . . . 55

C.1 ROV slender element properties in SIMO. . . . . . . . . . . . . . . . . . . . . . . . . . xvii

xiii
Nomenclature

Abbreviations

AHRS Attitude and heading reference system

DAF Dynamic amplification factor
DNV Det Norske Veritas
DPS Degrees per second
IMU Inertial measurement unit
JONSWAP Joint North Sea Wave Observation Project
LARS Launch and recovery system
OROV Observation class remotely operated vehicle
OS Offshore standard
ROV Remotely operated vehicle
RP Recommended practice
SIMO Simulation of marine operations
SWL Safe working load
TMS Tether management system
WROV Work class remotely operated vehicle
XKF Xsens Kalman Filter

Roman symbols

Symbol Description
A Effective umbilical area
A 33 Added mass in heave
A∞
33 High-frequency limit added mass in heave
ac t Crane tip acceleration
G
A Inertial reference frame accelerations

xv
AL Local reference frame accelerations
Am Total added mass
MP
A Accelerations at point of measurement
Ap Projected area
As Slamming area
A SP Accelerations at specified point
à w Water line area
aw Water particle acceleration
B1 Linear damping
b1 Linear damping coefficient
B2 Quadratic damping
b2 Quadratic damping coefficient
C Snap velocity correction factor
Ca Added mass coefficient
CD Drag coefficient
Cs Slamming coefficient
D Characteristic length
d Depth of penetration
E Modulus of elasticity
FB Buoyancy force
FD Drag force
F h yd Hydrodynamic force
F l i ne Lifting line force
FM Mass force
F sl am Slamming impact force
F snap Snap force
F st at i c Static force
F w,sl end er Wave force on slender element
F t ot al Total force
Fρ Varying buoyancy force
g Acceleration of gravity
H Wave height
Hs Significant wave height
h Submergence relative to water surface
I Mass moment of inertia
K System stiffness

xvi
k Wave number
KC Keulegan-Carpenter number
K0 Connection flexibility
L Length of lifting line
LIB Rotation matrix
M Mass of lifted object
m Mass per unit length
Re Reynolds number
MP
r Distance from measurement point to specified point
SJ JONSWAP wave spectrum
T Tension
Tp Wave spectrum peak period
Tz Zero-up-crossing period
T0 Natural period
V Volume of displaced water
vc Winch speed
vc t Crane tip velocity
vf f Free fall velocity
vr Relative velocity
vs Slamming velocity
v snap Snap velocity
vw Water particle velocity
V0 Total submerged volume
Xn Amplitude of oscillation
zc t Motion of crane tip

Greek symbols

Symbol Description
β Wave direction
δV Change in volume displacement
δj JONSWAP period parameter
γ JONSWAP peakedness parameter
φ Roll about x-axis
Φ0 Velocity potential
Φζ Wave component phase angle

xvii
ηc t Vertical crane tip motion amplitude
η̈ 3 Vertical acceleration of object
Λ Angular acceleration
λ Wave length
θ Pitch about y-axis
ψ Yaw about z-axis
ρ Density of salt water
σ JONSWAP width parameter
ω Wave angular frequency
ω Angular velocity
ωp JONSWAP angular peak period
ζa Wave amplitude

xviii
Chapter 1

Introduction

1.1 Background

At the semi-submersible rig Snorre B, the work class remotely operated vehicle (WROV) is
essential to perform drilling and production activities. Examples are operation of valves,
observation during drilling and template intervention. Because of this, the weather criterion
for the ROV should not be lower than criteria for the production and drilling activities.

Currently, the weather criterion for launching the WROV is based on simplified calculations
from DNV’s Recommended Practice (Det Norske Veritas, 2014b). The limit for lunch and
recovery is set at four meters significant wave height, which leads to restrictions for other
subsea operations dependent on the ROV. Due to assumptions and uncertainties in these
calculations, the resulting forces may be overestimated. Better methods for determining the
dimensioning effects for the launching system can reduce the safety factors in the calculations.
Thus, operation of the system in larger waves may be allowed. This requires a better
understanding of how the dimensioning forces occurs during launch and recovery of the
ROV.

1.2 Objectives

The objectives for the work is to present the background for the current weather criterion and
investigate limiting parameters. A preliminary model of the system for numerical analysis in
SIMO has been made previously for this purpose. In addition, acceleration measurements on

1
the real operation will be used to investigate the real behaviour of the ROV during launch and
recovery. This will also contribute in the validation of the SIMO model.

Based on time domain simulations, forces acting on the ROV system during the lifting
operation through the splash zone can be evaluated more accurate. This model can be used to
investigate the current weather criterion and critical effects for increasing the weather window.
Further, the effect of modifications on the lifting system can be evaluated.

The main objectives of the project can be summarized as:

1. Describe the ROV launch and recovery phases and the lifting system at Snorre B with
focus on limiting factors for the weather criterion.

2. Present the most important force contributions acting during launch and recovery of the
ROV based on the simplified calculation method from Det Norske Veritas (2014b),
numerical analysis in SIMO and full scale motion measurements.

3. Identify important parameters in procedures and the ROV launch and recovery system to
extend the weather window, and investigate the effect of changing these by using time
domain analysis in SIMO.

1.3 Literature review

The problem of determining forces on structures in waves has been discussed thoroughly in
the literature. Most of the theory used in this work was developed during the 20th century, and
the linear wave representation is based on theories introduced in the 1800s. The method for
calculating the wave forces are based on the Morison equation, which was introduced in
Morison et al. (1950). This formulation proposes that the forces acting on a section of a pile due
to wave motion can be divided into drag and inertia forces, represented by the use of drag and
inertia coefficients. It requires that the body is small compared to the wave length resulting in
an uniform incident flow in the vicinity of the body (Sarpkaya and Isaacson, 1981).

To use the Morison equation on subsea modules, the force coefficients are found from
empirical data based on the parameters characterizing the flow. During the 1980s, Marintek
conducted a research program concerning marine operations. As a part of this, a series of
model tests were performed to determine hydrodynamic coefficients for different subsea
structures and modules in both steady and oscillatory flow. The tests are described in Øritsland
and Lehn (1987) and a summary of the coefficients are presented in Øritsland (1989). An

2
alternative way of presenting the drag term from the Morison equation was used to provide a
linear and quadratic damping term. This description was found to give the most convenient
representation both for a partly and fully submerged body (Øritsland and Lehn, 1987). The
results are presented with corresponding Reynolds (Re) number and Keulegan-Carpenter (KC)
number, where the KC number represents the amplitude of the fluid motion relative to the
body size and the Re number represents the ratio between the inertial and viscous forces.
However, three-dimensional complex structures makes it difficult to determine the force
coefficients due to factors such as interaction effects. Uncertainties due to input parameters
makes it necessary to properly verify results from time domain simulation programs.

At the University of Victoria, Canada, motions of a deep sea remotely operated vehicle system
were logged and used to develop a continuous one-dimensional model of the system
behaviour. In their case, operation depths of more than 1000 meters were the concern, unlike
this work where the wave zone is the main focus. However, the tether length during deep sea
operations gave resonance with the vessel motion leading to slack tether and snap loads. The
motion measurements are described in Driscoll et al. (2000b), while the model development is
presented in Driscoll et al. (2000a). Further, the model was used to develop and optimize a
passive heave compensator to reduce the risk of slack tether and decrease the tension
variation, as described in Driscoll et al. (2000c) and Driscoll et al. respectively. By tuning the
stiffness and damping, slack umbilical were avoided even in extreme wave conditions.

1.4 Structure of the report

In Chapter 2, the remotely operated vehicle (ROV) and launch and recovery system (LARS) are
presented together with the different launch and recovery phases. The basic theory related to
the operation is presented in Chapter 3. Also the theory behind the simplified analysis and
time domain simulation software SIMO is referred to. Further, strategies for the full scale tests,
SIMO modelling and input values used in the calculations are discussed in Chapter 4.

In Chapter 5, the results are presented and discussed. This includes sensitivity analysis,
parameter variation and comparison between the different methods. The conclusion and
suggestions for further work are presented in Chapter 6.

3
Chapter 2

Problem description

The dimensioning factors during launch and recovery of the ROV will be influenced by many
aspects. This includes the semi-submersible platform, launch and recovery system and the
properties of the ROV system. Also the procedure used during launch and recovery of the ROV
is important. In this section, both the equipment and the phases during the launch and
recovery are presented.

2.1 Equipment

The semi-submersible platform Snorre B is equipped with two ROVs, one work class ROV and
one observation class ROV. An A-frame is used to launch the observation class ROV from the
platform side, while the work class ROV is launched through a deck opening. Unlike the work
class ROV, the observation class ROV does not have manipulators and payload capacity. This is
why many of the subsea tasks conducted at Snorre B requires use of the work class ROV. Thus,
only the launch and recovery of the work class ROV is analysed.

2.1.1 Snorre B

Snorre B is an integrated production, drilling and quarters semi-submersible unit located at

the Tampen area in the Norwegian North Sea. The hull is symmetrical about both the roll and
pitch axis, and consists of four columns connected by four pontoons. Figure 2.1 illustrates the
top-side of the platform. A SIMO model of Snorre B has been provided by Statoil.

5
 Figure 2.1: Snorre B, view from north-west. Photo by Harald Pettersen, Statoil.

2.1.2 ROV system

The main components of the work class ROV system is a tether management system (TMS), the
ROV and a tool skid. A steel armoured umbilical connected to the TMS is used to lift the
system. In addition, the umbilical transfers power and signal between the surface and the TMS.
During launch and recovery, the TMS is connected to the top of the ROV. This TMS
arrangement is called a top-hat, as opposed to a garage where the TMS surrounds the ROV. At
the operational depth, the ROV is disconnected from the top-hat. Power and signal are
transferred between the TMS and the ROV by a smaller cable called tether. The tool skid is fixed
to the underside of the ROV. In Figure 2.2, the ROV and the tool skid is placed on the deck
hatches. The lower part of the TMS is seen on top of the ROV.

Normally, the ROV and tool skid is designed to have neutral weight submerged. This is done to
preserve the manoeuvrability during operations. On the contrary, the weight of the submerged
TMS will give a downward acting force. This leads to a low centre of buoyancy compared to the
centre of gravity when the ROV and TMS are connected. If the centre of gravity is located above
the centre of buoyancy, loss of tension in the umbilical will lead to a rotational motion of the
submerged ROV system.

6
 Figure 2.2: The WROV system as seen from behind.

Snorre B is equipped with an old work class ROV. This has made it difficult to obtain precise
properties of the system. Most of the properties are found in the calculations done as basis for
the current weather criterion (JMC Engineering, 2006a) and the load test document for the
launch and recovery system (Saipem LTD, 2012). In Table 2.1, a summary of the data is
presented.

Table 2.1: Properties of the ROV system.

Length ROV [m] 3.2
Width ROV [m] 1.5
Mass ROV [kg] 3000
Mass TMS [kg] 2300
Mass tool skid [kg] 1100

2.1.3 Launch and recovery system

The work class ROV is launched through a deck opening using a winch. From the winch, the
umbilical is passed through a spooling device and over a sheave to the bullet at the top-hat.
The sheave is placed approximately 18.5 meters north and 6.5 meters west relative to the centre
of the platform. The clearance between the sheave and deck hatches is 5 meters, while the
distance from the deck opening to the water surface is approximately 22 meters. A photo of the
sheave system is shown in Figure 2.4.

7
A hydraulic winch with maximum speed of 0.8 m/s is used for the lifting operation. The safe
working load (SWL) for the launch and recovery system is 9000 kg at 4 meters significant wave
height (COSALT Offshore Norge AS, 2012). In Table 2.2, a summary of the properties are
presented.

Table 2.2: Properties of the launch and recovery system.

Max winch speed [m/s] 0.8
Umbilical effective stiffness (EA) [kN] 38 985
Umbilical diameter [mm] 42

Due to a small deck opening, guidewires are used to reduce the ROV motions through the deck
opening. This is done by connecting two guide wires to a guide frame (cursor) which is
connected to the TMS. A hydraulic system in the guide frame can be used to rotate the ROV
system. This is normally done before the ROV is lifted through the deck opening during
recovery.

Figure 2.3: Drawing of the sheave suspension in front of the winch (JMC Engineering, 2006b).

The sheave is supported by a frame connected to a suspension system. All the data about the
suspension system are taken from JMC Engineering (2006b), where structural analysis and
dynamic behaviour analysis are presented. The system is illustrated in Figure 2.3. It consists of
two passive pneumatic hydraulic cylinders preloaded by two gas accumulators. The cylinders
are fastened to the girder supporting the main deck. In addition, two shock absorbers are used
to reduce impact forces from the sheave frame due to sudden slack umbilical. The suspension

8
is preloaded with 87.4 kN , corresponding to the weight of the ROV system including the guide
frame. Further, the stiffness is linearised and based on the assumption of an adiabatic process
due to fast response of the system (JMC Engineering, 2006b). This gives a suspension stiffness
of 196 kN /m. Both stiffness and damping properties of the system is based on assumptions.
This must be considered if the sheave suspension is used to further change the weather
criterion.

2.2 Phases during launch and recovery

To better understand the launch and recovery system and procedure, the operation will be
divided into different phases based on forces acting on the ROV and the physical behaviour. In
addition, this will be used to identify the most critical phase for determining the weather
criterion.

2.2.1 Lift off

The lift off phase covers lifting the ROV system off deck, moving the hatches covering the deck
opening and lowering the ROV system through the deck opening. As mentioned in the
equipment presentation, a guidewire system has been installed to lead the ROV system
through the deck opening. In addition to restrict motions of the ROV, the cursor will relieve the
umbilical tension while rotating the ROV before it is lifted through the deck opening during
retrieval.

The cursor is lifted by the same umbilical as the ROV system. This will introduce an extra mass
of 2600 kg. However, the dynamic amplification is low compared to other phases because of
limited platform motion in the relevant weather conditions. Therefore, this phase is not
considered critical in terms of the umbilical tension and weather criterion.

Increasing the mass of the lifted object will introduce more wear in the umbilical. This is
especially relevant because of large concentrated loads over the sheave and at the spooling
arrangement into the winch. The concentrated loads due to bending of the umbilical will be
present independent of the guide frame. However, in case of modifications to the launching
system, the design and need for a guiding system through the deck opening has to be
considered. The current system gives a low clearance between the sheave and cursor, as shown
in Figure 2.4. This will restrict modifications of the sheave to reduce the concentrated loads on
the umbilical.

9
 Figure 2.4: The blue cursor frame on top of the TMS illustrating clearance to the sheave.

2.2.2 Lifting below deck

In this phase, the lifting between the deck opening and water surface is covered. After the ROV
system is lowered through the deck opening, the cursor frame will be disconnected and act as a
pivot point for the pendulum motion. This is illustrated in Figure 2.5.

The distance from the mean water level to the deck opening is approximately 22 meters. This
distance combined with limited damping makes the operation vulnerable for pendulum
motion. Both platform motions and wind forces can excite the motion. This will depend on the
natural period of the pendulum motion in air. A simplified assumption of the pendulum period
q
is T0 = 2π Lg , where L is length of hoisting line and g is acceleration of gravity. With a
umbilical length between 5 and 20 meters, the natural period increases from 4.5 to 9.0 seconds.
However, since the natural period depends on length of the hoisting line, it is unlikely to
coincide with the period of the excitation force over a long period. Assuming that the clearance
to the columns is sufficient, the most critical effect from horizontal motion will be during the
water impact. The water surface impact will be discussed for the next phase.

Also the vertical natural period is checkedqto determine the characteristic behaviour in the
M
vertical direction. This is done by T0 = 2π K
, where M is the mass term and K is the stiffness

10
 Figure 2.5: Drawing of the ROV below the deck opening. Adapted from Bruset (2014)

term in the equation of motion. Assuming only umbilical stiffness contribution with a length of
20 meters, K = ELA = 1950 kN /m. This gives T0 = 0.34 s. Compared to the first order wave force
periods, this indicates a stiffness dominated system with a quasi-static behaviour. Including
the sheave suspension system reduces the stiffness to 178 kN /m. The natural period is
increased to 1.2 seconds, which also gives a low frequency ratio.

2.2.3 Lifting through the wave zone

This phase covers the impact with the water surface and the transit through the wave zone.
Normally, the winch speed is reduced before the water impact to reduce the slamming force.
Once the ROV has penetrated the water, it is desirable to transit the wave force area as quickly
as possible. The intention is to reduce the time of wave force interaction.

Loss of tension in the umbilical is a condition that might occur in this phase. If the combined
slamming impact force, buoyancy force and vertical wave forces is larger than the weight of the
ROV system, tension in the umbilical is lost. The result will be a snap load in the umbilical as
the tension is restored. In addition to limit the weather criterion due to the safe working load of
the lifting cable, snap loads can reduce the life of the umbilical due to damage on the structural
armour and internal conductors (Driscoll et al., 2000c).

For simplified calculations, it is common to divide this phase into several stages and determine
if snap loads will occur for each phase. However, the wave zone represents a complicated

11
environment which makes it difficult to consider assumptions around the forces without
detailed analysis. Based on the description of the other phases, the wave zone is assumed to be
the dimensioning phase for the weather criterion. A more detailed theoretical description of
the wave zone lifting is presented in Chapter 3.

2.2.4 Operating below the wave zone

The ROV is further lowered to the operating depth of 350 meters. Even if the wave forces
decreases exponentially, the platform motion will still influence the dynamic behaviour. As the
umbilical gets longer, the vertical stiffness is reduced and weight of the system increases.
Combined, this may lead to larger weight of the system and resonance with the sheave vertical
motion.
q
M +A 33
As previously, the vertical natural period is taken as T0 = 2π K . A 33 is the vertical added
mass and will increase the natural period found in air. Assuming an umbilical length of 350
meters and A 33 = M , including the sheave stiffness gives a natural period of 2.7 seconds. This is
regarded as outside the resonance domain for the first order wave forces. In addition, it is a
conservative estimate. However, the natural period have to be checked if further reduction of
the sheave stiffness is carried out.

12
Chapter 3

Theory of the analysis

As previously discussed, lifting through the wave zone is regarded as the most critical phase
during launch and recovery of the ROV. In addition, the forces on the lifted object are difficult
to predict. Oscillating flow because of waves, water impact effects and varying degree of
submergence due to sheave motion and incoming waves are examples of factors influencing
the total force.

In this section, the basic theory relevant to lifting through the wave zone is presented. It is
important to understand assumptions and limitations in the calculations when evaluating the
results. In addition, the calculation methods and limitations of DNV’s simplified method and
the time domain simulations in SIMO are discussed.

3.1 Wave zone analysis

3.1.1 Environment

The environment will include wind, waves and current. However, it is assumed that first order
wave forces will dominate the object behaviour and platform motions in the wave zone. As the
focus will be on forces and effects during the wave zone lifting, only waves are included in the
environment modelling.

In this case, the waves are based on linear wave potential theory. This means that the
undisturbed incoming wave field can be expressed in terms of the velocity potential Φ0 (SIMO
Project Team, 2013). Derivation of the velocity potential for a regular wave at infinite water
depth can be found in Faltinsen (1990). Water particle velocity and acceleration can be found

13
from the velocity potential. In linear theory, it is assumed that the velocity potential and fluid
velocity are constant between the mean free surface level and the actual free surface level. This
is illustrated by Figure 2.2 in Faltinsen (1990).

For deep water waves we have (SIMO Project Team, 2013):

ζa g kz
Φ0 = e cos(ωt − kxcosβ − k y si nβ + Φζ ) (3.1)
ω
2
where ζa is the wave amplitude, g is the acceleration of gravity, k = ωg is the wave number, β is
the direction of wave propagation and Φζ is the wave component phase angle.

The dimensions of the ROV in horizontal plane is assumed small compared to the wave length.
This assumption can be evaluated in relation to the wave breaking limit. The limit is set from
the maximum wave steepness, and gives λ = 7H for deep water waves (Det Norske Veritas,
2014b). H is the wave height and λ is the wave length.

A real sea state will be irregular and have random height, length, shape and speed of
propagation (Det Norske Veritas, 2014b). This can be represented by a wave spectrum. A wave
spectrum represents the energy distribution of the sea state. The main parameters for most
wave spectra are significant wave height HS and spectral peak period T p . In this case, the
JONSWAP spectrum is used to simulate wind sea. It describes wind sea conditions for the most
p
severe sea states, and is assumed to be a reasonable model for 3.6 < T p / H s < 5 (Det Norske
Veritas, 2014b).

The JONSWAP spectrum is defined by a non-dimensional peak shape parameter, γ, a spectral

width parameter, σ, as the following relation:

( ωω −1)2
5 ωp 5 − 5 ( ωp )4 exp(−
p
)
S J (ω) = H s2 T p ( ) e 4 ω (1 − 0, 287 l nγ)γ 2σ2 (3.2)
32π ω


0, 07 ω ≤ ωp 2π
σ= where ωp =
0, 09 ω < ω Tp
p

The relationship between the spectrum peak period, T p , and the zero-up-crossing wave
period, T z , is T z = T p (0.6673 + 0.05037 γ − 0.006230γ2 + 0.0003341γ3 ) (Det Norske Veritas,
2014b). Usage of the JONSWAP spectrum in SIMO is described in Section 4.2.4.

14
3.1.2 Motions and forces in the wave zone

The equation of vertical motion is derived from the second law of motion, mass multiplied by
P
the acceleration equals the sum of external forces, M a = F ext . For a small object compared to
the wave length, the external forces will mainly consist of buoyancy, inertia, wave excitation,
slamming, and drag forces during launch (Det Norske Veritas, 2014b). Based on the description
in Det Norske Veritas (2014b), the different hydrodynamic forces will be presented in this
section, before the resulting equation of motion is expressed.

Buoyancy

The buoyancy force equals the weight of displaced water, and the centre of buoyancy equals
the centre of the volume displacement. Leaking air that was trapped in the submerged object
may cause a decreasing buoyancy force after the object is completely submerged. This must be
considered based on perforation of the object when the dimensioning forces are determined.
The time dependant buoyancy force is expressed by:

F B (t ) = ρg V (t ) (3.3)

where ρ is the mass density of water, g is the acceleration of gravity and V (t ) is the displaced
volume of water.

The static weight can then be expressed by:

F st at i c = M · g − F B (t ) (3.4)

where M is the structural mass.

Inertia

The inertia force due to motion of the object is given by the mass of the object, M , and added
mass in vertical direction due to vertical acceleration, A 33 , multiplied with the vertical
acceleration of the object, η̈ 3 :
F I = −(M + A 33 )η̈ 3 (3.5)

15
Wave excitation

Assuming a small object compared to the wave length, the wave excitation force can be found
by:
F w,i = (ρV + A i i ) · a w,i + F D,i (3.6)

where the first term represents the inertia force due to water particle acceleration, a w , and F D,i
is the viscous drag force due to the relative velocity on the object in direction i . In the inertia
term, A i i is the added mass term in direction i due to motion in the same direction and V is the
submerged volume of the object. This is for a fully submerged object. If the object is partly
submerged, a hydrostatic force due to the wave elevation is added (Det Norske Veritas,
2014b).

Viscous drag

The viscous drag force in oscillating flow can be found from the drag term in the Morison
equation:
1
F D = ρC D A p |v r |v r (3.7)
2
where C D is drag coefficient in oscillatory flow, A p is projected area normal to the flow
direction and v r is relative velocity. Depending on the relative velocity between the object and
fluid, the viscous drag force can act as either damping or excitation of the system (Det Norske
Veritas, 2014b).

Model tests have shown that the drag coefficient in oscillatory flow is dependent on the
Keulegan-Carpenter (KC) number in addition to the Reynold’s number (Øritsland and Lehn,
1987). Decay test have also shown that a linear term may contribute significantly to the
damping in oscillatory flow. Further, Øritsland and Lehn (1987) states that by expressing the
viscous drag with a linear and quadratic term, the KC dependency is reduced. This gives the
following expression:
F D = B 1 v r + B 2 v r |v r | (3.8)

where B 1 and B 2 is the linear and quadratic damping term respectively. This expression is
further discussed in Section 3.1.3.

16
Slamming

The slamming force can be expressed as the rate of change of fluid momentum (Faltinsen,
1990):
d d A∞
33 2
F sl am (t ) = (A ∞ v s ) = v (3.9)
d t 33 dh s
where A ∞
33 is the high frequency added mass in heave, v s is the slamming impact velocity and h
is the submergence relative to the water surface.

Equation of vertical motion and line force

If these forces are combined, the equation of vertical motion for a lowered object is obtained
(Det Norske Veritas, 2014b):

1 2
d A∞
33
(M +A 33 )η̈ 3 = B 33 (v 3 −η̇ 3 )+B 33 (v w −η̇ 3 )|(v w −η˙3 )|+(ρV +A 33 )a w + (ζ̇−η̇ 3 )2 +ρg V (t )−M g +F l i ne (t )
dh
(3.10)

where η(t ) is vertical object motion, v w is the water particle vertical velocity, v c is the winch
speed and F l i ne (t ) is force in the hoisting line.

The time dependant force in the hoisting line can found by using the cable stiffness, F l i ne (t ) =
M g −ρg V (t )+ K (z ct − η). Then K is the cable stiffness multiplied by the difference in crane tip
motion, z c t , and motion of the lifted object, η. If the winch speed is constant and the crane tip
motions are neglected, the lifting line force can be expressed as (Det Norske Veritas,
2014b).

d A∞
33
F l i ne (t ) = M g −ρg V (t )−(ρV + A 33 )a w − (ζ̇−v c )2 −B 1 (v w −v c )−B 2 (v w −v c )|(v w −v c )| (3.11)
dh

3.1.3 Hydrodynamic properties

A difficult part of the analysis in the wave zone is to determine hydrodynamic properties. Free
surface interaction and oscillating flow combined with a complex structure will introduce
complicating effects. In this case, drag and added mass forces are calculated using
dimensionless coefficients.

As mentioned in Section 3.1.2, the drag force can be split into linear and quadratic damping
terms. The physical representation of the linear and quadratic terms can be skin friction and
form drag respectively (Det Norske Veritas, 2014b). According to Øritsland and Lehn (1987),

17
this will make it possible to use constant damping coefficients in oscillating flow. This
assumption is only valid for a limited domain of the Keulegan–Carpenter number. K C < 10 is
mentioned as a valid domain based on results from model tests. It is defined as K C = 2π · XDn ,
where X n is the amplitude of oscillation and D is the characteristic length. Due to a complex
structure and uncertain motion, both X n and D are difficult to determine. However, Øritsland
and Lehn (1987) states that the range is assumed to be valid for most conditions related to the
lifting operations that is covered, which is similar to this case.

The linear damping coefficient b 1 and quadratic damping coefficient b 2 are defined as
(Øritsland and Lehn, 1987):

3π2
p
D/2g
b1 = B 1 · (3.12)
2ρ A p D
B2
b2 = (3.13)
0.5ρ A p
where B 1 and B 2 are the linear and quadratic damping terms respectively, D is the
characteristic body length and A p is the projected area in the velocity direction.

The added mass coefficient is defined as (Øritsland, 1989):

Am
Ca = (3.14)
ρV0

where A m is total added mass and V0 is total submerged volume. As shown in Figure 6-11 in
Det Norske Veritas (2014a), the added mass depends strongly on submergence relative to the
free surface. In addition, the slamming impact force is defined as the added mass variation
with depth. Because of a high slamming impact velocity compared to the water particle
velocity is assumed, the vertical added mass is taken as its high frequency limit (Det Norske
Veritas, 2014a).

The coefficients used in the SIMO model are discussed in Section 4.2.3. Further, sensitivity
analysis have been conducted to check validity and importance. The results are presented in
Chapter 5.

3.2 Simplified analysis method

The simplified analysis described in Det Norske Veritas (2014b) will for many cases provide a
good estimate of the forces involved in a lifting operation. The current weather criterion for

18
ROV launch and recovery at Snorre B is based on this method (JMC Engineering, 2006a). A
fundamental property of the method is that it should overestimate the forces. This provides
safety margins, but may lead to a unnecessary strict weather criterion. In addition, the results
will depend on the method and assumptions used to determine the input parameters.

3.2.1 Main assumptions

In Det Norske Veritas (2014b), three main assumptions are listed for the method:

• Small object in the horizontal plane compared to the wave length.

• The vertical motion of the object follows the vertical motion of the crane tip.

• The dominating loads on the object are caused by the vertical relative motion between
object and water so that other modes of motion can be disregarded.

A small body can be based on the relation 5 · D < λ, where D is the horizontal extent of the
object and λ is the wave length (Det Norske Veritas, 2014b). This should hold for critical wave
heights with the length of the ROV equal to 3.2 meter compared to the wave breaking limit
7 · H < λ for deep water.

The vertical motion of the object equal to the crane tip motion requires no amplification due to
vertical resonance. In the wave zone, the following requirement is set by DNV for resonance
not to occur (Det Norske Veritas, 2014b):

T p > 1.6T0 (3.15)

q
M +A 33 +θmL
where T p is the peak wave period, T0 = 2π K
is resonance period of the hoisting
system, M is mass of the object in air, A 33 is heave added mass of object, K is stiffness of the
hoisting system, m is mass per unit length of lifting line, L is length of lifting line and θ is an
adjustment factor. Based on the natural periods found in Section 2.2, this assumption is
valid.

The last assumption is questionable for wave heights in the region of the weather criterion.
Large waves may introduce other significant force contributions.

To further simplify the calculations, wave kinematics

qindependent of the wave period can be
HS
used. This will introduce the requirement T Z ≥ 10.6 g
, where T Z is the zero-up-crossing
period and HS is the significant wave height. According to Det Norske Veritas (2014b), this
wave period domain will be valid for most sea state conditions.

19
3.2.2 Force calculations

In the simplified method the different force contributions are calculated separately, put
together and then compared to the accept criteria. The hydrodynamic forces are varying
buoyancy, F ρ , hydrodynamic mass, F M , slamming impact, F sl am , and hydrodynamic drag, F D .
Each of the force components will vary in time. This makes it difficult to determine the time
instant with the largest total hydrodynamic force. The simplified method introduce the
following relation: q
F h yd = (F D + F sl am )2 + (F M − F ρ )2 (3.16)

The total hydrodynamic force F h yd is added to the static force F st at i c to find the tension in the
lifting cable.
F st at i c = M g − ρg V (3.17)

where M is structural mass and V is volume displacement.

The vertical water particle velocity and acceleration are found from the velocity potential Φ0 ,
2
v w = δΦ
δt
0
and a w = δδtΦ20 . Wave kinematic equations independent of the wave period then
gives:
− 0.35d
πg HS e
p
v w = 0.30 H s (3.18)
− 0.35d
a w = 0.10πg e H s (3.19)

where v w is the vertical water particle velocity, a w is the vertical water particle acceleration and
d is the distance from water plane to centre of gravity of submerged part of the object. These
equations are based on maximum wave height Hmax for a 30 minutes long operation, where it
is assumed a wave amplitude of ζa = 0.9H s .

Mass force

The characteristic mass force consists of the combined inertia force due to crane tip
acceleration and water particle acceleration. The object can be divided into items, i . Based on
this, the force contributions can be calculated separately and be included dependant on the
launch phase. q
F Mi = [(M i + A 33i )a ct ]2 + [(ρVi + A 33i )a w ]2 (3.20)

where M i is mass of object item in air, A 33i is heave added mass of object item, Vi is volume of
displaced water of object item relative to still water level and a ct is the crane tip
acceleration.

20
Slamming impact force

When the object penetrates the water surface, a slamming impact force will occur. This is taken
as:
F sl am = 0.5ρC s A s v s2 (3.21)

where C s is the slamming coefficient,

q A s is horizontal plane area of object subjected to
2 2
slamming force and v s = v c + v ct + vw , where v c is the winch speed.

Drag force

The characteristic drag force on an item is defined:

F Di = 0.5ρC D A pi v r2 (3.22)

where C D is drag coefficient in oscillatory flow of submerged part of the item, A pi is projected
q
2 2
vertical area of submerged part and the relative velocity is defined as v r = v c + v ct + vw ,
where v c is the winch speed.

Varying buoyancy force

Waves will introduce a varying buoyancy force:

F ρ = ρ δV g (3.23)
q
where δV = Aew ζ2a + η2ct is the change in volume of displaced water from still water surface to
instantaneous water surface, Aew is the mean object area in the instantaneous water surface
and η c t characteristic single amplitude vertical motion of the crane tip.

Snap force

In case of slack umbilical, a snap force will occur. The expression can be derived from
conservation of energy assuming kinetic energy is transformed into potential spring energy,
and is taken as:

p
F snap = v snap K (M + A 33 ) (3.24)

21
K is stiffness of the hoisting system, M is mass of the object in air and A 33 is heave added mass
of the object. The snap velocity is v snap = v f f +C v r where v f f is free fall velocity, v r is relative
vertical velocity as defined for the drag force and C is a correction factor given as C = 0 for
vf f
v f f > 0.7v r , C = cos[π( vr
− 0.2)] for 0.2v r < v f f < 0.7v r and C = 1 for v f f < 0.2v r .

The free fall velocity is calculated from the drag force:

s
2F st at i c
vf f = (3.25)
ρ ApCD

where A p is projected area of submerged part of the object in vertical direction and C D is the
drag coefficient of submerged part of the object in vertical direction. The static force is given as
F st at i c = M g − ρV g , where V is the volume of submerged part of the object and M will include
entrapped water in the object.

3.2.3 Accept criteria

The capacity checks of the lifted structure and lifting equipment will depend on occurrence of
slack in the lifting cable. Det Norske Veritas (2014b) recommends a slack criterion of
F h yd ≤ 0.9 · F st at i c−mi n . Should slack occur, a snap force will be present as tension in the
umbilical is restored. This lead to a large amplification factor and should be avoided. Two
accept criteria are mentioned for the simplified method in Det Norske Veritas (2014b), with and
without snap loads. The characteristic force without snap load is taken as:

F t ot al = F st at i c + F h yd (3.26)

and with snap load:

F t ot al = F st at i c + F snap (3.27)

where F h yd is Equation 3.16 and F snap is Equation 3.24. Lifting standards, e.g. DNV-OS-H205,
applies dynamic amplification factors (DAF) for capacity checks. For offshore lifting, Det
F t ot al
Norske Veritas (2014b) gives the following equation: D AF conv = Mg
, where M is the mass of
object in air and F t ot al is the largest force during the lifting operation.

22
3.3 SIMO time domain analysis

Simulation of Marine Operations (SIMO) is a computer program where time domain

simulation is applied to solve equations of motion. This section is based on (SIMO Project
Team, 2013), and will give a brief overview of the most important parts relevant to the model
used in this analysis. For further information, the SIMO theory manual (SIMO Project Team,
2013) should be consulted.

SIMA has been used as the graphical interface of SIMO to make the model, run analysis,
visualize the simulations and post-process the results.

3.3.1 Coordinate systems

To model the system and interpret the results, it is necessary to obtain an overview of the
coordinate systems. SIMO uses multiple right-handed cartesian coordinate systems where
counter clockwise rotations are positive. As a reference, a global earth fixed coordinate system
is used. That is positioned with the xy-plane in the calm water surface and an upward pointing
z-axis. Local body fixed coordinate systems are used for bodies and will follow the local body
motion. The local system is used to provide coordinates for body elements. In addition, a body
related coordinate system which only follows the horizontal motion of floating bodies
exist.

3.3.2 Environment

The ocean environment can be simulated by wind, waves and current in SIMO. Only waves are
simulated in this project.

Linear wave potential theory is used to simulate waves, as described in section 3.1.1. Both
regular waves and irregular random waves can be modelled. A number of different wave
spectra enables simulation of wind sea, swell sea or a combination. In addition, a user
specified numerically defined wave spectrum can be applied.

23
3.3.3 Force models

Body types

The type of body used will determine which force and motion models that are used in the
simulation. In SIMO Project Team (2013), four different body types are presented.

• Type 1: Large volume body with 6 degrees of freedom where the total motion is simulated
in time domain.

• Type 2: Large volume body with 6 degrees of freedom where the motions are separated
into frequency domain and time domain.

• Type 3: Small volume body with 3 degrees of freedom (translations) where the forces are
determined by position dependent hydrodynamic coefficients.

• Type 4: Same as type 1, but with fixed or prescribed body position.

If the main body is simulated as body type 1, distributed element force can be used. There are
two ways of modelling distributed element force, either long slender elements or concentrated
fixed elements. By using distributed elements, small body theory is used to calculate the forces.
The forces are then transferred to the main body.

In theory, slender elements may be used to model all properties of the body. The only
requirement for the main body is structural mass. However, this may be set to zero for the main
body and defined in the distributed elements. By defining structural mass in the distributed
elements, mass moment of inertia will be calculated based on the element locations.

Three coordinate systems are used in the body modelling, a global system, a body fixed system
and a local coordinate system for the distributed elements. They are all orthogonal and right
handed (SIMO Project Team, 2013). The hydrodynamic coefficients of slender elements are
defined in the local element coordinate system.

Slender elements

The hydrodynamic forces on the slender element are based on the Morison equation. In
addition, buoyancy forces F ρ and slamming forces F sl am are modelled.

Both the gravity F = −mg d s and buoyancy force F ρ = ρV g d S acts in the global z-direction.

24
The wave forces are represented by:

F w,sl end er = (ρV + ρV C a ) a w + B 2 v r |v r | + B 1 v r (3.28)

V is submerged volume per unit length, a w is water particle acceleration components in local
strip coordinate system and v r is the relative velocity. Based on the theory presented
previously, the first term represents the Froude-Krylov and diffraction force, the second term is
the Morison equation drag term and the third term is linear drag (SIMO Project Team,
2013).

The slamming force is expressed as:

δA m δh
F sl am = − · vr (3.29)
δh δt

where h is the distance between instantaneous surface elevation and strip origin in global
z-direction (SIMO Project Team, 2013).

Coupling elements

To model the lifting cable, a simple wire coupling can be used. It is modelled as a linear spring,
∆L = KT . In this case, T is the umbilical tension, K is the effective axial stiffness and ∆L is the
umbilical elongation. The possibility to include crane boom stiffness contribution, K 0 , is added
to the stiffness input as:

1 L 1
= + (3.30)
K E A K0

The umbilical properties are given as the unstretched length, L, modulus of elasticity, E , and
effective cross section area, A.

In addition, a tensioner may be given for the simple wire coupling (SIMO Project Team, 2013).
This will simulate a passive pneumatic hydraulic cylinder. The supplied pressure will hold a
mean tension in the cylinder. The input parameters consist of pretension, maximum rate of
change in pretension, stiffness at the specified pretension and stroke length.

25
Solution methods

The calculation of water particle motion and forces due to waves are done using cosine series.
That implies summation of the harmonic components in time domain. Pre-generated
harmonic components by Fast Fourier transform will be a faster calculation method. However,
that method will not take into account prescribed change in position using a winch. If the
initial position of the ROV is in air, wave forces at the surface will be applied when the body is
submerged regardless of the depth.

For the motion calculations, three integration methods can be used, Modified Euler, 3rd order
Runge Kutta or Newmark predictor-corrector method (SIMO Project Team, 2013). The results
in this report have been calculated using Runge Kutta.

26
Chapter 4

Procedure

In this chapter, calculation methods and input for the simplified method and the time domain
analysis are discussed. Further, the set-up and procedure for the acceleration measurements
are described.

4.1 DNV Simplified method

As mentioned previously, the current weather criterion is based on the simplified calculation
method from DNV-RP-H103 (Det Norske Veritas, 2014b). The procedure is to determine the
most critical phase and find the largest tension in the lifting cable. In Section 3.2, the
background for the method is given.

4.1.1 Calculation method

The method of calculation will be dependant on the load case analysed. This will determine
which hydrodynamic forces that are influencing the system. Thus, the results will be affected
by the chosen load case, and the assumptions made for which forces that contributes. The
assumptions are based on the calculations for the current weather criterion. Based on that, a
better understanding of the background for the current wave height limit can be
obtained.

The calculations can be divided into stages. In the first part, the hydrodynamic forces are
calculated and compared to the static weight of the object. If the hydrodynamic forces exceeds
the static force by a certain factor, slack umbilical leading to a snap load will occur. The next

27
part will then be to calculate the snap force in the umbilical. After this, the last part will be to
determine the largest tension in the umbilical. That is done by summing the static and
dynamic forces, where the dynamic force will be the largest of the hydrodynamic or snap
force.

4.1.2 Input data

For simplicity, the calculations are done independent of the wave period. This means that
significant wave height H s is the only weather parameter in the calculations. In Table 4.1, the
environment and motions of the LARS are given. These values are based on JMC Engineering
(2006a), where v ct is velocity of the sheave, a ct is acceleration of the sheave and v c is assumed
hoisting velocity during water impact.

Table 4.1: Wave height and sheave motion.

H s [m] v ct [m/s] a ct [m/s 2 ] v c [m/s]
4 0.5 0.46 0.17

Also the properties of the umbilical are based on JMC Engineering (2006a). The length is set to
30m, which will be the approximate length during the water surface impact. Table 4.2 presents
the properties.

Table 4.2: Properties of the umbilical.

E A[N ] L[m] K [N /m]
39 · 106 30 1.3 · 106

The hydrodynamic properties are chosen specifically for the load case analysed. In this
analysis, that corresponds to right after the water surface impact. The vertical projected area
A p is taken as the full horizontal plane area of the ROV, the volume displacement V is 0.5 m 3
and d is assumed to be small.

Table 4.3: ROV properties at the chosen load case.

M [kg ] V [m 3 ] A p [m 2 ] d [m]
6400 0.5 4.8 0.2

The hydrodynamic forces are calculated based on force coefficients. Det Norske Veritas (2014b)
provides recommendations for the values. These may in some cases be unnecessary

28
conservative due to limited information about the system. The slamming coefficient C s should
not be lower than 5 for structures other than smooth circular cylinders and the drag coefficient
C D should be higher than 2.5 in oscillatory flow unless model tests or CFD analysis has been
conducted. The chosen coefficients are presented in Table 4.4.

Table 4.4: Hydrodynamic coefficients.

C D [−] C a [−] C s [−]
2.5 0.8 5

4.1.3 Calculations

The calculations are based on the formulas presented in Section 3.2. In Table 4.5, water particle
velocity v w and acceleration a w , static weight F st at ,vertical added mass A 33 and the relative
velocity v r are given. The slamming velocity v sl am is taken as the relative velocity.

Table 4.5: Static force, vertical added mass and calculated velocities and accelerations.
F St at [kN ] A 33 [kg ] v w [m/s] a w [m/s 2 ] v r [m/s] v sl am [m/s]
56.775 410 3.27 3.03 3.48 3.48

The hydrodynamic forces and total force are presented in Chapter 5.

4.2 SIMO Modelling

Where determining a conservative weather criteria is the main objective of the simplified
method, the intension of time domain analysis is to model the physical behavior during launch
and recovery correctly. In this section, the modelling methods and simulation procedures in
SIMO are presented.

4.2.1 Platform motions

Two methods have been tested to model the platform motions, a full body type 1 model and a
simple body type 2 model using motion transfer functions as input. Statoil has provided a
SIMO model of Snorre B to simulate the motions in the sheave point. This also included the
motion transfer functions.

29
In order to use the motion transfer functions, the platform had to be modelled as body type 2.
However, this lead to large transient motions of the ROV during the simulated operation. These
motions were not damped out properly while the ROV was in air. The full body type 1 model
did not cause the same problems. Thus, it was decided to use the body type 1 model. Instead,
the motion transfer functions were used to validate the resulting platform motions.

The origin of the platform local coordinate system is placed at the calm water surface in the
middle of the platform. At the simulation start-up, this corresponds to the global coordinate
system.

4.2.2 Launch and recovery system

The model of the launch and recovery system represents the connection between the ROV and
the platform. For the dynamic properties, correct motion of the sheave and stiffness in the
lifting system are most important.

A body point connected to the platform is placed at the approximate winch position. SIMO
enables winch simulation at a body point, where hoisting speed, acceleration and operation
time can be given for several intervals. Guide points are fixed to the platform body to simulate
the sheave and cursor. This will provide the correct cable length from the winch, which will
influence the cable stiffness. In addition, the pendulum pivot point in the deck opening is
simulated by the lowest guide point representing the cursor. The body point positions relative
to the platform body coordinate system are given in table 4.6.

Table 4.6: Position of LARS components relative to the body coordinate system of Snorre B.
Coordinate x y z
Winch [m] 6.5 15 26.5
Sheave [m] 6.5 18.5 27
Cursor [m] 6.5 18.5 22

The umbilical is modelled as a simple wire coupling between the winch and a body point fixed
to the ROV system body. Initial length, connection flexibility, material damping and line
stiffness are given as properties. The stiffness in the umbilical, elasticity modulus multiplied by
cross section area, is assumed the same as used for the simplified calculations (JMC
Engineering, 2006a). Based on the coupling force description in the theory manual (SIMO
Project Team, 2013), material damping is chosen as 2 % of the stiffness. As the cable stiffness is
assumed to provide the majority of the system stiffness, the connection flexibility is set
negligibly low. The input values are presented in table 4.7

30
 Table 4.7: Simple wire coupling input.
Initial length [m] Connection flex. [m/N] Stiffness (Ea) [N] Material damping [Ns]
10 10−10 39 · 106 780 · 103

The sheave suspension will influence the stiffness and damping in the system. Not including it
in SIMO will introduce a simplification compared to the real system. On the other hand, this
will make it easier to troubleshoot the model, conduct sensitivity analysis and evaluate the
results from the simplified calculations.

To check how the sheave suspension affects the results, it has been modelled in two ways. The
first is to insert the suspension stiffness , 1/K 0 = 1/196000 m/N , as the connection flexibility in
the simple wire coupling input. An alternative way is to use the tensioner described in Section
3.3.3. Based on the assumptions in JMC Engineering (2006b), the input parameters are chosen
as presented in Table 4.8.

Table 4.8: Tensioner input.

Pretension [N] Max. rate of change in pretension [N/s] Stiffness [N/m] Stroke length [m]
87 400 8.6 · 105 196 000 0.7

4.2.3 ROV system

The ROV system is modelled as a body type 1 assigned with structural mass and mass moment
of inertia. To model the buoyancy and hydrodynamic forces, slender elements have been
attached to the main body. The model is visualized in Figure 4.1. The black sphere represents
centre of gravity for the ROV system.

Main body

To be able to build the model using slender elements, the main body has to be type 1 (SIMO
Project Team, 2013). The only input requirement for this body is structural mass, which can be
set to zero. In this case, structural mass, centre of gravity and mass moment of inertia is
defined for the main body. By doing this instead of defining structural mass for the slender
elements, a better control over the mass properties is obtained.

Information about the mass moment of inertia has not been obtained. Thus, it is assumed that
the mass of the ROV and tool skid is homogeneous distributed in a cuboid. By doing this, the

31
 Figure 4.1: Visualization of the ROV and LARS model.

1
mass moment of inertia can be found from I = 12 M (a 2 + b 2 ), where M is total mass and a and
b are lengths out from the rotation axis (Irgens, 1999). The input values are found in Table
4.9.

Table 4.9: Mass properties of the ROV system.

Mass [kg ] I xx [kg · m 2 ] I y y [kg · m 2 ] I zz [kg · m 2 ]
6400 6000 3000 6000

Slender elements

Slender elements are used to simulate buoyancy and hydrodynamic properties. This is done by
distributing the elements to match the geometry of the ROV system. Volume, linear and
quadric drag coefficients and added mass coefficient is then defined for each element. Both
the distribution and the sum of the properties should match the real object. Depth dependent
volume and force coefficients are defined for the elements with length in the horizontal
plane.

32
In Figure 4.1, the distribution of the slender elements are presented. The ROV system consists
of the ROV, TMS and tool skid, as shown in Figure 2.2. This is modelled by rectangular frames
placed low, mid and high in the horizontal plane. Vertical corner elements are used to simulate
the forces between the horizontal planes. All these slender elements are defined with
buoyancy, drag coefficients and added mass coefficients. In addition, two slender elements are
placed high in the ROV frame. Only buoyancy is modelled for these. The intention is to make
an element distribution that models the real hydrodynamic properties by using as few
elements as possible.

Hydrodynamic properties

The method to determine hydrodynamic properties can introduce a significant error source. In
this case, the hydrodynamic properties are based on model test data from Øritsland (1989).
Several different structures are tested to determine force coefficients.

Two types of structures has been chosen from Øritsland (1989), a buoyant type body to
represent the ROV with tool skid and a working tool body to represent the TMS. Total linear and
quadratic drag force and added mass are calculated for the ROV and the TMS in x, y and z
direction. Then the hydrodynamic coefficients of the slender elements in SIMO are adapted to
provide the total calculated properties. In Table 4.10, the results of the calculations from
Øritsland (1989) is presented. The SIMO slender elements input is adapted to match the
calculated values, and is presented in Appendix C. The local coordinate system of the ROV is
the same as shown in Figure 4.4.

Table 4.10: Hydrodynamic properties of the ROV system.

TMS ROV Total
x y z x y z x y z
M a [kg ] 769 513 769 2306 1538 2306 3075 2050 3075
B 1 [N s/m] 169 169 127 462 282 451 631 451 578
B 2 [N s 2 /m 2 ] 1538 1538 1153 4369 3075 4920 5907 4613 6073

For each structure that was model tested, a solidity ratio describing projected area, a fullness
factor describing volume and the weight in air related to the submerged weight are given.
These parameters are compared to the ROV and TMS. Based on this, the most similar modelled
structure is chosen. The drag and added mass coefficients for oscillating flow have been
determined by decay tests in low frequency, wave frequency and high frequency. Results from
the wave frequency test is used for the linear and quadratic drag coefficient.
Keulegan–Carpenter number K C < 6 and Reynold number Rn < 4.35 · 104 are assumed

33
(Øritsland and Lehn, 1987). Added mass coefficient is taken from the high frequency decay
test. This is to model the slamming force due to high speed water entry, which depends on the
change in added mass relative to submergence (Det Norske Veritas, 2014a). The formulas used
to calculate the total force is given in Section 3.1.3.

The input values for the slender elements and properties of the model are given in Appendix C.
In addition to the hydrodynamic properties, the total displacement is found to be 4215 kg and
centre of buoyancy is 1.61 m higher than the lowest slender element frame, z = 0.11 m in the
ROV body fixed coordinate system.

To model the varying forces when the ROV system is partly submerged, depth dependant
coefficients are defined in SIMO. For the buoyancy and drag, they are assumed to vary linearly
with the submergence. The added mass depth dependency is based on Figure 6-11 in
DNV-RP-C205 (Det Norske Veritas, 2014a). In Table 4.11, the depth dependency is presented
for the relative vertical position based on the element radius, R.

Table 4.11: Depth dependency of the horizontal slender element properties.

Vert. pos z/R[−] C V [−] C B 1 [−] C B 2 [−] C M a [−]
1 0 0 0 0
0 0.5 0.5 0.5 0.5
-1 1 1 1 0.65
-3 1 1 1 1

4.2.4 Sea state representation

The waves are modelled as a two parameter JONSWAP spectrum with significant wave height,
H s , and peak period, T p , as input. The peakedness parameter is taken as
p
γ = exp[3.484 (1 − 0.1975 δ J T p4 /H s2 )] and δ J = 0.036 − 0.0056 T p / H s (SIMO Project Team,
p p
2013). This is valid for 3.6 H s < T p < 5 H s . Outside this domain the peakedness parameter is
p p
γ = 1.0 for T p ≥ 5 H s and γ = 5.0 for T p ≤ 3.6 H s . Based on this, the zero-up-crossing wave
period can be calculated from the relationship in Section 3.1.1.

According to Det Norske Veritas (2014b), a lower limit for the zero-up-crossing period should
p
be T z = 8.9 · H s /g , which gives T z = 5.7 s in H s = 4 m and T z = 6.4 s in H s = 5 m. In addition, a
maximum period of T z = 13 s is mentioned. However, only wave periods up to T z = 10.7 s are
used. The wave periods run in the SIMO analysis are presented in Table 4.12.

34
 Table 4.12: Zero-up-crossing period for the JONSWAP spectrum.
Hs = 4 m Hs = 5 m
T p [s] γ T z [s] T p [s] γ T z [s]
7 5.0 5.6 8 5.0 6.4
8 3.0 6.2 9 2.9 6.9
9 1.6 6.6 10 1.6 7.3
10 1.0 7.1 11 1.1 7.9
11 1.0 7.8 12 1.0 8.5
12 1.0 8.5 13 1.0 9.3
13 1.0 9.3 14 1.0 10.0
14 1.0 10.0 15 1.0 10.7

4.2.5 Verification of SIMO model

To ensure that the variables are used correctly in the SIMO model, several tests have been
conducted. The intention is to prevent errors like wrong input units and misunderstanding the
coordinate systems. In addition, the numerical solver properties have been checked to provide
stable simulations and correct results.

Model properties

To ensure correct input of the model parameters, simulations without waves has been run. The
umbilical tension will indicate mass and buoyancy of the ROV, effects from the winch dynamics
and depth dependent hydrodynamic coefficient.

In Figure 4.2, the tension during a launch in still water is presented. This graph illustrates the
mass in air, buoyancy variation during water penetration, winch acceleration, hydrodynamic
drag force and how the tension variations are damped. Before the winch is started, the tension
corresponds to the mass of 6400 kg. The winch acceleration gives a lower tension
corresponding to a reduction in gravitational acceleration. When the ROV penetrates the water
surface, buoyancy and slamming impact force reduces the tension. Deceleration of the winch
increases the tension after 190 seconds, before the tension is stabilized at 21.4 kN . This
corresponds to the buoyancy force of 41.4 kN .

The drag coefficient will influence the tension when the ROV is submerged due to the winch
speed. In this run, a velocity of 0.8 m/s leads to a vertical drag force of 3.4 kN . Using a
projected area of 4.8 m 2 in Formula 3.7 gives a quadratic drag coefficient of 2.16. This is based
on the assumption that the quadratic drag force is the only contribution. The input for the drag

35
coefficient is intended to model drag force in oscillating flow, which is larger than the steady
flow coefficient without surface interaction (Øritsland and Lehn, 1987). Thus, this force is
probably too high in a simulation without waves. However, it is in the same region as the drag
coefficient proposed for the simplified method in Det Norske Veritas (2014b).

Depth dependant coefficients are confirmed during the water penetration. The tension
reduction occur over a time span due to buoyancy, drag and slamming. Individual
contributions from drag and buoyancy are difficult to distinguish. However, the slamming
impact contribution is dependent on change in added mass. Thus, a small increase in tension
is observed after each element with vertical added mass coefficient becomes fully submerged.
The time between the first impact and the first tension rise is approximately 0.3 seconds. By
calculating the time it takes to apply 65 % of the added mass coefficient with 0.8 m/s winch
speed, 0.26 seconds is obtained. The last 35 % is added over the next 0.26 seconds. This
confirms that the tension increase is due to reduction in the slamming impact force.

Sensitivity analysis and further discussion on the hydrodynamic properties are presented in
Chapter 5.

Umbilical tension
60

50
Tension [kN]

40

30

20

10

60 80 100 120 140 160 180

Time [s]

Figure 4.2: Time domain analysis without waves.

The full body type 1 Snorre B model provided by Statoil has been used to simulate the platform
motions in SIMO. This model also includes first order motion transfer functions that can be
used to model the platform motions as body type 2. Instead of doing that, the transfer
functions have been used to check the motions obtained from the body type 1 model in regular
waves. The results indicates similar motions. However, a certain time is required to stabilize
the motions. It is assumed that the same results are valid for irregular waves. Because of this,
the winch is started at 160 seconds in simulations with irregular waves to make sure that the
platform motions are stabilized.

36
Numerical solving properties

The numerical properties and procedures will also influence the results. In this case, time
series are generated by summation of harmonic components in time domain. This will ensure
that the slender elements are subjected to wave kinematics corresponding to the actual
position, as mentioned previously. Thus, this method is the only applicable for lifting operation
simulations with vertical position change in the wave zone. Further, a sufficiently low time step
must be used in the numerical integration.

The time step is checked by ensuring that tension peaks during snap loads are captured. A
simulation with irregular waves where three subsequent snap loads occur was checked. This
was done by reducing the time step until the maximum tension stabilized for the three spikes.
Since 0.005 and 0.001 seconds gave the same results, a time step of 0.005 seconds was used for
further analyses. In addition to the integration time step, also the storage interval was checked.
After the tension spikes had stabilized, the storage interval was increased until the peaks
changed. A integration time step of 0.005 seconds combined with a storage interval of 0.01
seconds is concluded to be sufficient.

4.3 Motion measurements

To investigate the actual behaviour of the ROV system during launch and recovery, an inertial
measurement unit (IMU) was fitted to the ROV. In addition to look at the physical behaviour,
the results can to some extent be used to validate analytical results. This will represent a simple
and low cost full scale test of the ROV motions.

4.3.1 Equipment and installation

The IMU consists of an accelerometer, gyroscope and magnetometer, which together works as
a three-dimensional compass and provides accelerations and rotation rates that can be
transformed into an inertial reference system. Both the IMU and the camera house used as
protection were borrowed from NTNU. The IMU was of type Xsens MTi-28 with device ID
01301955, and sensor specifications are summarized in Table 4.13. More information about the
IMU can be found in Xsens Technologies B.V. (2010).

Before the equipment was sent to the platform, the camera house had to be modified. Better
internal IMU fastening was mounted and the opening mechanism was replaced. In addition,

37
 Table 4.13: IMU sensor specifications.
MTi-28A53G35 Acceleration Rate of turn Magnetic field
Units m/s 2 d eg /s mG auss
Scale [units] 50 300 +/- 750
Bandwidth [Hz] 30 40 10
A/D Resolution [bits] 16 16 16
Dimentions 3 3 3

the connector on the camera house was replaced at the platform by the ROV operators, who
also conducted the measurements. The connector change was done to ensue compatibility
with the RS-232 communication and power supply through the ROV.

Several aspects had to be considered when mounting the IMU to the ROV. This includes
connection options on the ROV and limiting the noise in the measured data. Acceleration and
angular velocity will be affected by vibrations and magnetic disturbance will induce magnetic
deviation on the 3D compass. Based on this, the IMU was mounted to the frame in front of the
buoyancy element, as shown in Figure 4.3.

Figure 4.3: Position of IMU as seen in front of the ROV. Photos received form the ROV operators.

The location of the IMU provides distance to thrusters and electric motors. However, it also
gives a distance to the centre of rotation for the ROV system. This is important to consider
when comparing accelerations from the IMU with the results from the time domain

38
simulations. The IMU is positioned 1.67 m in front of the tether termination on the ROV and
1.70 m over the bottom of the tool skid. In SIMO, the ROV local coordinate system is placed in
the middle of the horizontal slender elements frame, and 1.5 m over the lowest frame. As
indicated by Figure 4.4, the horizontal distance in y-direction will have greatest influence on
the difference between measurements and simulations.

4.3.2 Test procedure

The IMU manufacturers software, MT Manager, was used to change the IMU settings and log
sensor data. By starting the record mode, accelerations, angular velocities and orientation are
stored in a binary log file. This file can be opened in MT Manager to export the data to an ASCII
file.

During launch, the record mode was started before the ROV was lifted from deck and ended
when the ROV was situated below the wave zone. The same time span was covered during
recovery. Based on recommendations from Xsens Technologies B.V. (2010), the IMU was
powered up a while before the recording was started to stabilize the build in filters.

Wave data has been obtained from the weather monitoring system used at Snorre B, MIROS. It
not possible to export time series of the wave data. Thus, the weather information has only
been used to get an indication of environment at the operation time. Plots of wave height and
wave periods were retrieved from Larsen (2015).

4.3.3 Post processing measurements

After extracting the data from the binary log files, Matlab was used to post process the results.
This is divided into three parts:

• Transforming the accelerations into an inertial reference system.

• Further calculations to evaluate the measurements.

• Presentation of the results.

The accelerations are obtained by measuring force per unit mass. This means that the
gravitational acceleration will be part of the measurements. Further, the direction of the
accelerations will be given in the local IMU body fixed coordinate system. By transforming the
measurements into an inertial reference frame, the global accelerations are obtained. Thus, the
gravitational acceleration will only be part of the vertical acceleration. This also gives the

39
opportunity to compare measured accelerations with global accelerations retrieved from the
SIMO analysis. Figure 4.4 illustrates the origin and coordinate system of the ROV in SIMO and
the IMU in an inertial reference frame.

ROV
x x
IMU

y z z
y

Figure 4.4: Origin of the coordinate systems from the IMU and the ROV as modelled in SIMO

The accelerations in the inertial reference frame, AG = [AGx AGy AGz ]T , are calculated by
multiplying the local reference frame accelerations, A L = [A Lx A Ly A Lz ]T by a rotation matrix, L I B
(Driscoll et al., 2000b).

AG = L I B A L (4.1)

 
cos ψ · cos θ cos ψ · sin θ · sin φ − sin ψ · cos φ cos ψ · sin θ · cos φ + sin ψ · sin φ
 
LIB =  sin ψ · cos θ sin ψ · sin θ · sin φ + cos ψ · cos φ sin ψ · sin θ · cos φ − cos ψ · sin φ (4.2)
 
 
− sin θ cos θ · sin φ cos φ · cos θ

The orientation angles are defined such that φ is roll about x, θ is pitch about y and ψ is yaw
about z, and obtained by using an attitude and heading reference system filter (AHRS). A build
in Kalman filter, XKF-3, can be used to obtain the orientation output from the IMU (Xsens
Technologies B.V., 2010). However, the transformed accelerations indicate a significant drift off
in the horizontal plane. Thus, two open source AHRS filters were tested. One is developed by
Sebastian Madgwick, while the other is Madgwick’s implementation of Mayhony’s AHRS
algorithm. The Matlab code for the filter developed by Madgewick was obtained from x-IO
Technologies (2012), while the implementation of Mayhony’s filter was taken from Madgwick
(2013).

40
Sebastian Madgwick’s filter combines integration of angular velocities from the gyroscope,
gravitational acceleration from the accelerometer and magnetic field measured by the
magnetometer to obtain the orientation. By doing this, drift off due to integration of angular
velocity and error due to magnetic deviation is corrected. The procedure is further described in
Madgwick et al. (5550). The other filter was based on Robert Mahony’s work, but did not involve
data from the magnetometer. Assuming both the platform and the ROV equipment influences
the magnetic field measurement differently over time, excluding the magnetometer data might
give more accurate results. In addition, the magnetometer measurements is most important to
provide a reference for the yaw orientation, which is of less importance in this work.

Positioning the IMU in front of the ROV pitch axis will introduce transient accelerations as the
ROV rotates. To remove these, an attempt to transform the acceleration measurements to the
centre of the ROV was made. Formula 4.3 was used to transform the accelerations in the point
of measurement, A MP , to the accelerations at a specific point, A SP , where the specified point is
given by a vector r M P (Driscoll et al., 2000b).

A SP = A M P − L I B · (Λ × r M P + ω × (ω × r M P )) (4.3)

The transformation requires the angular velocity, ω, and angular acceleration, Λ. However, the
measurements of angular velocity includes noise in the same frequency range as measured
spikes at in the wave zone. Unfiltered angular velocities gave vertical accelerations up to 60
m/s 2 , while filtered measurements gave unrealistic change around the acceleration spikes.
Thus, the position compensated accelerations are considered unreliable and have not been
included.

An attempt to obtain velocity and position by integrating the acceleration measurements was
also made. To avoid drift off due to noise, the time series was high pass filtered. However, the
filter will introduce uncertainties and remove position change due to the winch speed. This
makes it difficult to obtain reliable results. Instead, integrating the wave zone accelerations
over a limited time span is used to get an idea of the velocity change over a period in time.

41
4.3.4 Uncertainties and validation of measurements

The uncertainties in the measurements mainly consists of sensor measurement error, post
processing error and interpretation error due to lack of other references.

Measurements from the sensors in the IMU have predictable noise and error, as documented
in the test and calibration certificate found in Appendix D. However, external factors and post
processing may introduce significant error sources. In this test, calculating the orientation
used to transpose the acceleration measurements into an inertial reference frame will
introduce an error source.

To check the results when transforming the accelerations into the inertial reference system, the
two open source AHRS filters and the built in Kalman filter has been compared. As mentioned,
the Euler angles from the IMU gives steady non zero horizontal acceleration, which indicates
that the gravitational component has not been removed. Both of the open source AHRS filters
uses an error correction factor to account for measurement uncertainties like sensor noise,
calibration errors and sensor axis non orthogonality as mentioned in Madgwick et al. (5550).
The factor has been chosen such that the steady horizontal accelerations are removed. This
implies that the mean horizontal acceleration is close to zero in the time series. Further, the
filter giving acceleration trends most similar to the acceleration obtained from the build in
Kalman filter was used. In this case, that was the algorithm based on Mayhony’s work. The
results were compared to the angular velocities and accelerations for the local body fixed
coordinate system as a reference.

Further verification of the transformed accelerations is difficult due to lack of other references
for the ROV motion. This is also the case when the results are interpreted. Examples are the
time of the first water impact, background for acceleration spikes and the degree of
submergence of the ROV system. However, the combined results from the accelerometer and
gyroscope compared to the knowledge about the launch and recovery procedure will reduce
the risk of misinterpretation. This makes it possible to discuss theories about the behaviour of
the ROV in the wave zone based on the motion measurements.

42
Chapter 5

Results and discussion

The results from simplified calculations, motion measurements and time domain analysis are
presented in this chapter. In addition, comparison of the results and discussions around the
consequences of the findings are carried out.

5.1 Simplified method

DNV recommends to divide the water penetration into four load cases between the first water
impact and fully submerged (Det Norske Veritas, 2014b). However, to distinguish between the
cases will be more difficult with a small volume object in large waves. In this case, each of the
hydrodynamic force coefficients are calculated and checked by the snap load criterion given in
Det Norske Veritas (2014b), F h yd < 0.9 · F st at i c . The hydrodynamic load components are given
in Table 5.1.

Table 5.1: Static and hydrodynamic loads from simplified calculations.

F st at [kN ] F sl am [kN ] F M [kN ] F D [kN ]
56.8 149.1 4.2 74.5

The total hydrodynamic force is found from Formula 3.16. Not all of the force contributions are
in phase and will contribute at the same time. However, both the slamming impact force and
viscous drag force will individually lead to slack umbilical in the used load case according to
the snap load criterion. Based on this, it can be concluded that snap load will occur for the
respective load case.

43
In case of snap load occurrence, a more accurate calculation of the lifting operation is
recommended by Det Norske Veritas (2014b). However, a simplified method to determine the
snap load is provided. The results are presented in Table 5.2.

Table 5.2: Snap load calculation.

v f f [m/s] C [−] v snap [m/s] F snap [kN ] F snap + F st at i c [kN ]
3.07 0 3.07 288.3 346.0

The capacity check should be based on the largest total force, as mentioned in Section 3.2.3. In
the simplified method, this is either F t ot al = F st at + F snap or F t ot al = F st at + F h yd . To calculate
the total hydrodynamic load, it is assumed that the slamming impact force do not have full
effect at the same time as the hydrodynamic drag, mass and varying buoyancy force. Thus, the
largest total force will be due to the snap load:

F snap + F st at i c 346.0
D AF snap = = = 5.51 [−] (5.1)
M ·g 62.8

The dynamic amplification factor of 5.51 is governed by the static weight, the free fall velocity
when the umbilical tension is restored, the mass and the cable stiffness. Since the velocity is
multiplied by the square root of the product of stiffness and mass, uncertainty in free fall
velocity has the largest potential for error in the force.

The result obtained in JMC Engineering (2006a) concludes with D AF snap = 6.5, which is even
higher. This is mainly due to calculating the dynamic amplification factor by
F st at i c +F snap
D AF = F st at i c , using C D = 2.0 and a structural mass of M = 7000 kg in their calculations. It
should be noted that the current launch and recovery system is not dimensioned for a dynamic
amplification factor of 6.5. By doing analysis of the sheave suspension arrangement, the DAF
has been reduced to 3.0 (JMC Engineering, 2006b).

Based on the simplified calculations, the slamming force is the single largest contribution for
snap loads to occur. The drag force is about half in size, but still large enough to cause slack
umbilical. Motion measurements and SIMO analysis will be used to further investigate the
validity of this.

44
5.2 Motion measurements

During the motion measurement test, six launches and two recoveries were logged in total. In
Table 5.3, the loggings are presented with wave data taken from the MIROS system at Snorre B.
Due to the short time period available and downtime on the ROV system, only this limited
amount loggings were performed. Thus, generalisations from the results must be carefully
handled. However, both launch and recovery of the ROV have been covered in four meter
significant wave height. This gives the opportunity to evaluate examples of behaviour in
weather conditions used around the current weather criterion. Plots of the most interesting
period from each logging are found in Appendix A.

Table 5.3: Received log files with corresponding wave condition.

Local time Phase H s [m] T z [s] Sampling [Hz] Length [s]
2015-04-02 19h58 Launch 3.7 7.9 100 270.7
2015-04-04 14h19 Launch 1.5 8.3 100 230.7
2015-04-05 03h47 Launch 1.7 8.3 100 7969.8
2015-04-09 22h36 Launch 3.0 6.9 100 4744.4
2015-04-12 20h32 Recovery 4.0 7.5 200 421.9
2015-04-13 01h47 Launch 3.5 7.1 200 360.4
2015-04-14 00h15 Recovery 3.0 6.5 200 591.4
2015-04-14 02h50 Launch 4.0 7.0 200 213.9

The launch phases discussed in Section 2.2 are recognised from the IMU data. Before lift-off
from deck, vertical accelerations varying with the wave period are observed. This corresponds
to the motion of the platform. Accelerations with a higher frequency dominates the
measurements after lift-off and until the guide frame lands on the platform deck. The impact
between the guide frame and deck gives large vertical acceleration spikes. After disconnecting
the guide frame, a steady yaw velocity is observed. In addition, the gyroscope indicates local
pitch and roll oscillations combined with a pendulum motion with lower frequency. The first
water impact is identified by larger horizontal accelerations and angular velocity changes with
lower frequency than in air.

The data have been evaluated based on accelerations, orientation and rotation rates only. This
is sufficient information to identify phases as lift off from deck and the first water impact. On
the other side, details concerning effects in the wave zone are easily misunderstood. Especially
acceleration spikes are difficult to assess. The fact that accelerations and orientation are
obtained through filters and transformations must also be kept in mind. This is discussed in
Section 4.3.4. Position and coordinate system of the IMU are presented in Section 4.3.3.

45
5.2.1 Above the water surface

The most relevant measurement phase is the lowering from the deck and through the wave
zone. Spikes in the vertical acceleration identifies the landing of the guide frame on the deck.
For both recoveries and two of the launches, these represents the highest vertical accelerations
in the time series. However, the duration is less than 0.1 seconds, which indicates that the
acceleration spikes do not influence the tension in the umbilical much.

After disconnecting the ROV from the guide frame, a steady angular velocity in yaw is observed
for most measurements. In addition, oscillations with an amplitude below 0.1 degrees per
second and a period between one and two seconds are present for all three angular velocities.
Because of the short period and low amplitude, this is assumed to represent local pitch, roll
and yaw motion of the ROV.

The two recovery measurements indicates pendulum motion in addition to local rotations of
the ROV. This is based on angular velocities with a period decreasing from around seven
seconds toward zero as the umbilical length is reduced. Reduction of the period as the cable
length, L, is reduced coincides with the simplified pendulum natural period formula
p
T = 2π L/g . The angular velocity amplitudes increases as the period decreases. Also
oscillations that are damped out short after the ROV is assumed to be out of the water are
shown. That can be local pitch and roll motions due to exit forces from the water. The recovery
measurements are found in Figure A.5 and A.7.

5.2.2 Wave zone

The first water impact during launch is identified by a distinctive change from the pattern
described above the water surface. Larger horizontal accelerations and higher angular
velocities with longer period of oscillation are observed. This phase is shown in Figure 5.1,
taken from the launch 14th of April at 02:50.

It is difficult to estimate the position of the ROV related to the water surface. The ROV might get
hit by a wave top followed by a period without water contact. In addition, the winch speed is
unknown, and will vary for each measurement logging. When the accelerations stabilize with
longer periods and the high frequency oscillations in the angular velocities disappears, the ROV
is assumed to be fully submerged. Assuming maximum wave height of
1.8 · H s = 1.8 · 4 m = 7.2 m and winch speed of 0.4 m/s, 18 seconds transit time from wave top to
wave trough is obtained. Thus, the ROV may exit the water several seconds after the first water

46
 Accelerometer (Global reference system)
X
2 Y
Z

1
Accelerations [m/s 2 ]

-1

-2

-3

125 130 135 140 145 150 155 160 165

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
125 130 135 140 145 150 155 160 165
Time [s]

Figure 5.1: Wave zone accelerations during launch 2015-04-14 at 02:50

impact. This is a possible explanation for the variations in the period of the horizontal
acceleration between 125 and 145 seconds in Figure 5.1. The motions stabilize from around
145 seconds.

Also an increase in vertical acceleration is measured in this phase. Considering that the IMU is
mounted in the forward part of the ROV, the gyroscope data indicates that the cause is pitch
motion causing local acceleration at the front of the ROV. Roll motion does not have the same
effect since the position is along the mid axis of the length direction. An example is shown by
the angular velocity peak at 142 seconds, which does not seem to influence the vertical
acceleration much.

5.2.3 Acceleration spikes

Large vertical positive acceleration measurements with a duration shorter than 0.5 seconds are
observed in the wave zone during launch. This does not necessarily translate into large tension
increase in the umbilical. In this section, possible causes of the spikes and the consequences
are discussed.

Table 5.4 shows the largest accelerations measured in the wave zone during launch. In
addition, an approximate time duration of the impulse is included. This shows that most
measurements have vertical acceleration spikes of duration less than 0.05 seconds. Based on

47
the short duration, it is uncertain if these spikes indicate large vertical acceleration of the
whole ROV system or only local accelerations of the camera house or fastening beam. The fact
that the camera house is covered by the fastening beam in front, the buoyancy element in the
back, but is unprotected from the bottom, supports that the camera house suffers from direct
water impact from below.

Table 5.4: Maximum positive acceleration in wave zone with approximate duration.
Logging Acceleration [m/s 2 ] Duration [s]
2015-04-02 19:58 8.3 0.04
2015-04-04 14:19 4.0 0.05
2015-04-05 03:47 2.9 0.04
2015-04-09 22:36 2.5 0.06
2015-04-13 01:47 9.7 0.22
2015-04-14 02:50 13.3 0.01

However, the acceleration spike shown in Figure 5.2 stands out from the others. Between 126.0
and 127.575 seconds the acceleration is negative without sudden changes. After this, a short
negative acceleration spike combined with a positive pitch velocity spike appears. This
happens before the large vertical acceleration measurement of 9.7 m/s 2 with a duration of 0.22
seconds is observed. Four similar oscillations over 1.05 seconds with amplitudes between 1.15
and 1.35 m/s 2 follow after the spike, giving a mean period of 0.26 seconds.

Accelerometer (Global reference system)

X
8 Y
Z

6
Accelerations [m/s 2 ]

-2

124 126 128 130 132 134

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
124 126 128 130 132 134
Time [s]

Figure 5.2: Acceleration spike during launch 2015-04-13 at 01:47

48
By calculating the area over the unfiltered signal, the negative acceleration gives a vertical
velocity change of 1.47 m/s before the spike. Doing the same for the spike indicates a positive
vertical velocity change of 0.83 m/s. One theory is that the negative acceleration occurs with a
slack umbilical when the ROV is partly or fully submerged. When the umbilical tension is
restored, a pitch correction is caused together with the positive acceleration spike. The
following oscillations indicates low damping, which can be caused by the ROV being fully or
partially out of the water.

A simple estimate of the q

natural period assuming the cable as the main stiffness contribution,
(M +A 33 )
E A/L, indicates T0 = 2 π K
= 0.36 s. The cable input E A is taken from the SIMO input
and the length is assumed 20 meters. Since the system is assumed to be out of the water,
vertical added mass term is not taken into account. Thus, the mass term is taken as
M = 6400 kg . If the ROV is fully or partially submerged, added mass will increase the natural
period. The calculated natural period is 38 % longer than the measured period. Including the
effect of the snap load damper will lower the stiffness. Thus, the natural period will become
longer. If the acceleration spike and oscillations are due to a snap load, this indicates that the
current sheave damper system has limited effect on the stiffness during snap loads.

After the wave zone transit, the winch is stopped to change operating place from outside on
deck to the control container. This is shown in Figure 5.3 to illustrate a controlled velocity
change of the ROV. The area corresponds to a velocity change of 0.88 m/s, which is higher than
the maximum winch speed of 0.8m/s. Probable causes are positive heave acceleration of the
platform or measurement post processing inaccuracies.

Also in this case, vertical acceleration oscillation is observed. However, the period is longer and
decreasing amplitude indicates higher damping. Three oscillations are observed during 3.95
seconds, giving an approximate period of 1.32 seconds. Using a cable length of 120 meters
based on full winch speed after the acceleration spike, the natural period is found to be 1.07
seconds neglecting the sheave suspension stiffness. Vertical added mass of 3000 kg is included
in this case. Higher measured period coincides with lower stiffness due to the sheave damper
system. In JMC Engineering (2006b) the stiffness of the suspension is assumed to be 196.0
kN /m, leading to a 37.6 % reduction with a combined stiffness of 122.2 kN /m. The natural
period calculation then gives 1.74 seconds. This is 32 % longer than the measured period,
which is between the two calculated periods. Uncertainties in cable length and stiffness, the
suspension stiffness and added mass will influence the result.

49
 Accelerometer (Global reference system)
1.2 X
Y
1 Z

0.8
Accelerations [m/s 2 ]

0.6

0.4

0.2

-0.2

-0.4

-0.6
244 246 248 250 252 254 256 258
Angular velocities [dps]

Gyroscope
0.05
X
Y
0 Z

-0.05
244 246 248 250 252 254 256 258
Time [s]

Figure 5.3: Winch stop below wave zone during launch 2015-04-13 at 01:47.

5.3 Time domain analysis

In this section, results from the SIMO model are presented and discussed. Sensitivities when
changing input parameters are first checked to determine relationships and dependencies
between the input and results. Further, the model is used to obtain better understanding of
critical factors for the ROV motion and umbilical tension.

5.3.1 Sensitivity analysis

Sensitivity analysis have been conducted to illustrate how the different properties of the model
influences the results. This will give a better understanding of the model behaviour and how
inaccurate input can influence the results. Examples are the ratio between drag and inertia
forces and relationship changes with different coefficients.

50
 Hydrodynamic coefficients

Ten launches with winch speed 0.8 m/s were run with different wave seed using a two
parameter JONSWAP wave spectrum with T p = 8 s and H s = 4 m. All ten runs were checked
with different coefficients, but only the run with highest umbilical tension is presented here.
For the critical run with the initial input parameters, zero tension occurred twice with following
snap loads.

Added mass will influence the inertia forces and the slamming impact force. In Figure 5.4,
umbilical tension for three runs with the same environment, but changing added mass
coefficient, are shown. The slamming impact force on the lowest frame of slender elements is
evident before the buoyancy element gets submerged. Increasing the added mass coefficient
by 50 % has limited impact on the tension during the slamming phase. On the other side,
removing the added mass will reduce the tension by 5 - 10 kN . For SIMO’s slamming model,
this corresponds to removing the slamming contribution. The largest change in tension occurs
during the re-tensioning after the first period of slack umbilical. Fifty percent higher added
mass increases the tension from 44 kN to 51 kN , while eliminating it causes a snap load of only
24 kN .

Added mass dependency

70
Base case (1.0 Am)
High added mass (1.5 Am)
60 No added mass

50
Tension [kN]

40

30

20

10

0
185 190 195 200 205 210
Time [s]

Figure 5.4: Sensitivity analysis of added mass coefficient on umbilical tension.

51
In Figure 5.5, tension variation due to change in quadratic drag coefficient is presented. Fifty
percent increase in the quadratic drag coefficient gave an additional period of no umbilical
tension leading to a snap load. In addition, the first Snap load increased from 45kN to 105kN ,
an increase of 133%. On the other hand, reducing it by 50 % eliminated all snap loads due to
slack umbilical.

Drag dependency
Base case (1 Cd )
100
High drag (1.5 C d )
Low drag (0.5 C d )
90

80

70
Tension [kN]

60

50

40

30

20

10

0
170 180 190 200 210 220 230 240
Time [s]

Figure 5.5: Sensitivity analysis of drag coefficient on umbilical tension.

Based on these results, both the added mass and drag coefficients seems to be important for
the umbilical tension. Small changes in the added mass coefficient do not affect the umbilical
tension much. However, the reduction in tension when eliminating added mass proves the
importance of including it. The added mass term does not seem to influence the occurrence of
slack umbilical. On the contrary, change in the quadratic drag coefficient will influence both
occurrence of zero umbilical tension and the size of the following snap load. Consequently,
uncertainties in quadratic drag coefficient are more critical than for the added mass
coefficient. This is further discussed in Section 5.3.6.

Depth dependent coefficients

To simulate vertical variations of the hydrodynamic properties for the horizontal slender
elements, depth dependency has been added to the volume and force coefficients. This is
particularly important for the slamming force due to the slamming model in SIMO. In Figure

52
5.6, the vertical distance of active coefficients has been changed 50 %. The influence is largest
for the tension spike around the first water impact and for the snap load during re-tensioning
of the umbilical. However, the influence is small compared to the sensitivity analysis of the
coefficient values. Thus, which of the coefficients that are responsible for the deviation is not
further investigated.

Depth dependence
70 Initial depth dependence (1.0 d)
Small depth dependence (0.5 d)
Large depth dependence (1.5 d)
60

50
Tension [kN]

40

30

20

10

0
182 183 184 185 186 187 188 189 190
Time [s]

Figure 5.6: Depth dependent coefficients sensitivity for the lower element frame.

Horizontal forces

To check the importance of the horizontal hydrodynamic coefficients, a ROV model without
horizontal added mass and drag was run. In Figure 5.7, the results from 30 wave seeds in
H s = 4 m and T p = 8 s are presented. The results indicates that the vertical hydrodynamic
forces only will increase the umbilical tension. Thus, the risk of zero tension is non-existing.
This indicates that the ROV moves horizontally to the position where the wave forces do not
work against the gravitational force. Consequently, also the horizontal hydrodynamic
coefficients are important to obtain correct umbilical tension.

53
 Umbilical tension without horizontal added mass and drag
70

60

50
Tension [kN]

40

30

20

10

0
82 84 86 88 90 92 94 96
Time [s]

Figure 5.7: Umbilical tension in H s = 4 m and T p = 8 s without horizontal drag and added mass
coefficients.

5.3.2 Winch speed

The winch speed is an important parameter during launch and recovery. In addition to control
the time that the ROV is situated in the splash zone, also the slamming forces and the steady
vertical drag forces will be influenced.

Based on the sensitivity analysis of the added mass coefficient, the slamming force is assumed
to have a limited effect. Drag forces will reduce the umbilical tension when the winch speed is
increased due to a larger relative velocity. Thus, the probability for slack umbilical will rise. On
the contrary, the ROV will transit the wave zone faster with a higher winch speed. This way the
ROV is exposed to wave forces for a shorter time span, and unfavourable wave kinetics leading
to slack umbilical can be avoided

To check which effects that are most important, two winch speeds have been run with 60 wave
seeds in H s = 4 m and T p = 8 s. In addition, 30 wave seeds were run for each speed with
T p = 7 s. The results are found in Table 5.5. Due to different winch speed, the ROV is subjected
to different wave kinetics at the same wave seed. That is the reason for running the large
number of simulations. Slack in the umbilical occurs in 10 of the 60 runs with winch speed
0.4 m/s and 15 of the 60 runs with winch speed 0.8 m/s. Also the periods of zero tension are in
general longer for the highest winch speed. However, the largest umbilical tension occurs for
the lowest winch speed. Several factors will influence this. A shorter period of slack combined
with lower winch speed means that the ROV will travel a shorter distance. Since the wave
kinematics decreases with depth, the wave forces will be higher when the snap load occurs.
Further, lower winch speed will reduce the vertical drag force due to a steady velocity. A winch
speed of 0.8 m/s gives a steady drag force of 3.4 kN , while 0.4 m/s reduces it to 0.85 kN .

54
 Table 5.5: Snap load characteristics variation due to winch speed in H s = 4 m.
Spectrum peak period T p [s] 7 8
Winch speed [m/s] 0.4 0.8 0.4 0.8
No of periods with slack [-] 3 12 10 15
Mean period of slack [s] 1.53 1.28 0.74 0.97
Max tension [kN] 133.1 132.6 112.5 111.7

Based on this, it seems like the winch speed will influence the probability for slack umbilical,
while other factor are more important for the maximum snap load tension. This will be further
investigated in Section 5.3.6.

5.3.3 Horizontal translation

The clearance between the ROV and the platform hull can cause problems during launch and
recovery. Even though the deck opening is positioned midway between the platform hull
columns, the horizontal distance between the pontoon side and the sheave position is only
about 6 meters (Bruset, 2014). To check the risk of impact, the horizontal ROV position relative
to the sheave has been plotted in Figure 5.8. The red lines in the top plot indicates the
approximate position of the pontoon wall relative to the sheave position, while the blue lines
represent the ROV position for 30 wave seeds in the direction of incoming waves.

As observed in Figure 5.8, the centre of the ROV passes the 6 meter mark once in H s = 4 m and
T p = 7 s. However, the pontoon position is normally 12.5 meters below the mean water surface
(Bruset, 2014). When considering where the largest translations occur, impact between the
ROV and the pontoon does not occur in this case. Taking into account the ROV’s horizontal
extent, the latest critical position happens 195 seconds out in the time series when the ROV
buoyancy element is 7.2 meters below the mean water surface.

To investigate if this represents the worst case in H s = 4 m, a longer wave period and lower
winch speed have been checked. In addition, the case with H s = 4 m and T p = 7 s has been run
using 30 additional wave seeds to check if the first runs provides a good representation. The
latest critical position occurs with the buoyancy element 7.3 meters below the mean water
surface, which coincides well with 7.2 meters. On the contrary, a longer wave period gives
smaller translation. Setting T p = 8 s with H s = 4 m gives the latest critical position at 5 meters
below the mean surface. Further, reducing the winch speed from 0.8 to 0.4 m/s with T p = 8 s
decreases the depth to one meter below the mean water surface. This indicates that impact is
most likely with a short wave period combined with high winch speed.

55
 ROV translation
8
Pontoon position [m]
6
Horizontal position [m]

-2

-4

-6
175 180 185 190 195 200 205
Vertical position [m]

20
15 Mean water surface
10
5
0
-5
-10
-15
-20
175 180 185 190 195 200 205
Time [s]

Figure 5.8: ROV horizontal and vertical position during 30 wave seeds in H s = 4 m and T p = 7 s.

Also the motion of the pontoon has to be taken into consideration. If the ROV motion is 180
degrees out of phase with the pontoon motion, both the critical depth and horizontal
translation will be reduced for the ROV. To check which motions that can be expected, the
platform translations from SIMO has been transferred to a point at z = −10 m and y = 24.5 m.
In Figure 5.9, the pontoon position in H s = 5 m and T p = 9 s relative to the deck opening
position is presented. The horizontal motion is limited to an amplitude of 0.7 m, while the
vertical motion amplitude is maximum 0.3 m. Based on this, the motion of the pontoon will
only have a limited effect on the risk of impact between the ROV and the pontoon.

Considering that the real position of the pontoon is more than 10 meters below the mean water
surface, the SIMO analysis indicates that impact between the ROV and the pontoon does not
occur in H s = 4 m.

56
 Pontoon position relative to deck opening
7.4
7.2
Horizontal position [m]

7
6.8
6.6
6.4
6.2
6
5.8
5.6
5.4
5.2
5
4.8
150 200 250 300
Vertical position [m]

-29.8

-29.9

-30

-30.1

-30.2
150 200 250 300
Time [s]

Figure 5.9: Pontoon horizontal and vertical position relative to the deck opening position during
30 wave seeds in H s = 5 m and T p = 9 s.

5.3.4 ROV recovery phase

Most of the analysis are based on the ROV launch phase. That is assumed to be most vulnerable
to slack umbilical leading to snap loads. In the recovery phase, the winch speed will increase
the tension due to drag forces, and slamming on the ROV are less probable. However, the ROV
is still exposed to wave forces and water exit effects. Thus, also recovery has been simulated.
Higher mass due to entrapped water has not been included as the ROV system is assumed to
have a high degree of perforation.

Figure 5.10 shows the umbilical tension for 30 wave seeds during recovery of the ROV. In
contrast to launching the ROV in the same sea state, the umbilical tension is never near zero.
The lowest tension obtained is 14.5 kN , corresponding to 23 % of the static weight in air. This
indicates that the launching operation is more critical than the recovery operation in terms of
slack umbilical. In addition, the tension in the umbilical does not exceed the transient tension
variations in air observed during the launch simulations, despite using the maximum winch
speed of 0.8 m/s.

57
 Umbilical tension during recovery

70

60

50
Tension [kN]

40

30

20

10

0
60 80 100 120 140 160 180
Time [s]

Figure 5.10: Umbilical tension during recovery for 30 wave seeds in H s = 4 m and T p = 8 s.

5.3.5 Weather condition

To investigate how the weather condition influences the results, 30 wave seeds have been run
using a JONSWAP wave spectrum for different significant wave heights and peak periods. The
sea state properties are presented in Table 4.12. Umbilical tensions from each run are
presented in Appendix B. Two aspects are particularly interesting, the maximum umbilical
tension and if slack in the umbilical occurs. That is indicated by the maximum and minimum
tension for each run in the plots.

The current weather criterion is set at a significant wave height of four meters regardless of the
wave period. Therefore, four meters significant wave height has been checked with a spectrum
peak period varying between 7 and 14 seconds. In addition, five meters significant wave height
has been used to investigate consequences of a higher wave limit. Due to a long simulation run
time, the number of conditions has been limited. Change in the direction of incoming waves
has not been checked. One reason is that the platform heave transfer function changes little
due to the symmetrical hull. In addition, wave shadow effects due to interaction with the
columns are not modelled. That is assumed to affect the forces more than change in the
platform motion. Thus, changing the wave direction by 45 degrees is considered to introduce
more uncertainties in the results.

58
The general observation is that the maximum umbilical tension increases with lower wave
period and increasing wave height. This is as expected due to larger water particle velocities
and accelerations with increasing wave amplitude and frequency. Since the wave period is
increased until slack umbilical is avoided, snap loads occurs for all sea states except the two
longest periods. However, the snap loads are significantly higher for lower wave periods. Most
of the slack umbilical incidents do not cause higher tension than the weight of the ROV system
in air. This is due to the high buoyancy of the ROV and tool skid, which will increase the risk of
slack umbilical and reduce the total force during snap loads when the ROV is submerged.

Impact between the ROV and the pontoon can also be critical. This was discussed in Section
5.3.3, but mainly for H s = 4 m. To check the influence from wave height, the translation at
H s = 4 m and T p = 7 s has been compared with the result from H s = 5 m and T p = 8 s. In both
simulations, a winch speed of 0.8 m/s was used. By assuming a critical horizontal translation
of 4.4 meters, the latest occurrence was found to be with a buoyancy element submergence of
6.1 meters. That is less than 7.2 meters found for H s = 4 m, and supports the theory that
pontoon impact is less important for the weather criterion.

5.3.6 Snap loads in SIMO

The time domain analysis indicates that the largest umbilical tension occur due to snap loads
after a period of slack umbilical. When the period without tension increases, the size of the
tension spike gets larger. Thus, it is important to investigate which effects cause zero tension in
the umbilical.

Wave forces

To check how the wave forces influence the largest umbilical tension, vertical wave kinematics
have been plotted against the tension in Figure 5.11. Particle velocity and acceleration have
been taken at the middle of the slender element simulating the buoyancy element in the length
direction. Before the element is submerged, the data represents surface wave kinematics.

The umbilical tension reaches zero when the declining vertical particle acceleration crosses the
increasing vertical particle velocity. Since the drag forces are dependant to the velocity
component while the inertia forces are proportional to the acceleration, this indicates that the
drag coefficient is most important for occurrence of slack umbilical. When the velocity reaches
zero, re-tensioning in the umbilical is observed. This happens as the acceleration is at the
lower extreme value, and will induce the highest downwards wave induced inertia force. Thus,

59
 Wave kinematics during slack umbilical
10 Particle velocity [m/s]
Particle acceleration [m/s 2 ]
Wave elevation [m]
Tension [N 10 -4 ]
8

-2

185 190 195 200 205 210

Time [s]

Figure 5.11: Vertical wave kinematics compared to umbilical tension.

inertia forces influence the snap load as indicated during the sensitivity analysis. Also the drag
force will start acting downwards. However, the downwards velocity will be much smaller than
the acceleration during the snap load.

Based on this, it seems like drag forces are most important for occurrence and length of slack
umbilical. The inertia term will act maximum downwards during the re-tensioning of the
umbilical. In addition, the inertia force will influence the time of slack umbilical and snap
velocity. This is because of the negative vertical water particle acceleration most of the period
of slack umbilical. Assuming that the tension is restored when the particle velocity reaches
zero, the maximum snap load increase found in the drag coefficient sensitivity analysis will be
limited by the period of positive particle velocity.

Quantifying the snap load

The simplified calculation of the snap load, F snap , can be derived from conservation of energy.
Kinetic energy is assumed to be transformed into potential energy as spring energy. This gives
p
F snap = v snap · K · (M + A 33 ), where v snap is the velocity absorbed, K is the system stiffness
and M + A 33 corresponds to the mass term including vertical added mass when submerged.
However, this implies that other external forces do not contribute by external work during the
snap load.

60
To check which parameters that are important for the umbilical tension, the snap load in
Figure 5.12 has been evaluated. Following the same wave kinetics pattern as in Figure 5.11, a
snap load occurs at maximum wave elevation. The ROV velocity plot indicates a downward
vertical velocity of 1.415m/s before the snap load. Since the winch speed is 0.8 m/s and the
sheave velocity is 0.171 m/s, the snap velocity is taken as v snap = 1.415 m/s− 0.800 m/s+
0.171 m/s = 0.786 m/s. Using a cable length of L = 34.5 m and a mass term of
M + A 33 = 6400 kg + 3000 kg = 9400 kg gives a snap force of F snap = 81.0kN . To obtain the total
force, the weight is added in F t ot al = (M − ρV ) · g + F snap . In this case M = 6400 kg and
ρV = 4215 kg , giving F t ot al = 102.5 kN . The tension plot shows 111.7 kN , 8.9 % higher than
calculated. This indicates that additional forces will influence the ROV during the snap load,
but that the assumption can give an indication of the total force.

Vertical motion of ROV during snap load

10

-5 ROV position [m]

ROV acceleration [m/s 2 ]
Tension [0.1 kN]
-10 Wave elevation [m]

155 160 165 170 175 180 185 190 195 200

0.5
0
-0.5
ROV velocity [m/s]
-1
Sheave velocity [m/s]
-1.5
155 160 165 170 175 180 185 190 195 200
Time [s]

Figure 5.12: ROV vertical motion during snap load in H s = 4 m and T p = 8 s.

The calculation gave a lower total force than what was obtained in SIMO. This indicates that
other external forces contributes as work. Based on the maximum downward water particle
acceleration, the wave induced inertia forces are assumed to influence on the umbilical tension
during snap loads. In addition, the inertia force contribution to the snap velocity during slack
umbilical is included in the calculation. This is done by taking the ROV velocity from the SIMO
results. Contribution from the inertia force is confirmed by the sensitivity analysis of the added
mass coefficient presented in Figure 5.4. Although, whether the contribution was due to ROV
velocity change or forces acting during the force impulse was not established.

61
For a snap load occurring with a fully submerged ROV, the simple formula provides an
indication of the total force during snap loads. However, the calculation is based on known
steady drag force, added mass and stiffness of the system. These properties are easy to
determine from the input values, but are subject to uncertainties in the real system. In
addition, the sheave and winch speed are taken into account. Based on this, a snap velocity
taken as the free fall velocity without umbilical tension, v f f , corrected for crane tip velocity,
v c t , and winch speed, v c seems to be a good approximation. This will require a good
approximation of the free fall velocity of the ROV exposed to wave forces.

5.3.7 Stiffness and damping in the launching system

The launch and recovery system influences both the stiffness and damping in the problem. For
most of the simulations, the effect from the sheave stiffness is left out of the model by assuming
large stiffness. Then the stiffness and damping properties of the umbilical will be the
dominating effects. To check the validity of this assumption, three different ways of modelling
the stiffness and damping in the launching system have been checked. This will also give an
indication of the effect by using a sheave suspension system.

In addition to assuming large sheave stiffness, both adding a tensioner at the winch and
reducing the sheave stiffness for the connection flexibility in the simple wire coupling input
have been checked. The input was based on assumptions around the current sheave
suspension system. This represents course simplifications of the system, but will give an
indication of how the snap loads are affected by reducing the stiffness in the system.

Based on the assumptions in JMC Engineering (2006b), the stiffness of the sheave suspension is
set to 196 kN /m. Using the same umbilical characteristics with a length of 30 meters, the
effective axial stiffness is reduced from 1299.5 kN /m to 170.3 kN /m. In Figure 5.13, the effect
by including the sheave stiffness is illustrated by the yellow line compared to only model the
umbilical stiffness for the blue line. The tensioner will represent a passive pneumatic hydraulic
cylinder (SIMO Project Team, 2013), and the effect from using the tensioner is presented as the
red line. By using the tensioner, better representation of the combined stiffness and damping
properties can be obtained.

The alternative modelling do not influence the time duration of slack umbilical. In contrast,
the resulting snap load is reduced to 75.7 % with the tensioner and to 62.0 % with reduced
sheave stiffness compared to only including umbilical stiffness. Further, the decreased stiffness
gives a longer period of oscillation following the snap load. A longer natural period will reduce

62
 Sheave suspension sensitivity
120
Cable stiffness
100 Tensioner
Reduced stiffness
Tension [kN]

80

60

40

20

182 184 186 188 190 192 194 196

15
Cable stiffness
Acceleration [m/s 2]

Tensioner
10 Reduced stiffness

-5
182 184 186 188 190 192 194 196
Time [s]

Figure 5.13: Tension and vertical acceleration variation due to change in stiffness.

the ratio between the length of the force impulse and the natural period of the system. This will
decrease the dynamic amplification of the displacement for the following oscillations, as
illustrated by Figure 3.5 in Larsen (2012). Assuming a triangular shaped impulse lasting 1/6 of
the natural period, the dynamic amplification factor is around 0.5 for the following free
oscillation. The example corresponds to an impulse duration of 0.25 seconds combined with a
natural period of 1.5 seconds. In comparison, an impulse duration ratio of 1.0 will cause a
dynamic amplification factor of around 1.5. Thus, lower stiffness can prevent additional snap
loads caused by the first oscillation after the snap load impulse.

It is clear that decreasing the stiffness can be beneficial in terms of reducing maximum
umbilical tension. However, the stiffness reduction due to the current passive suspension
system will be limited by the piston stroke length. If the stiffness is set too low, the piston will
hit the end stops with high velocity. This will cause large forces in the suspension system and
the reduced stiffness effect disappears. The suspension stiffness has to be adjusted so that the
stroke length is utilized, but the piston does not hit the end stops. The natural period must also
be taken into consideration. Based on the current stiffness, the natural period is around 1.5
seconds in the wave zone and approximately 2.3 seconds at the operation depth of 350 meters.
Further decreasing the stiffness may increase the natural period into the wave band, causing
resonance in vertical motion at the operation depth. This can make it difficult to dock the ROV
onto the TMS and cause snap loads in the umbilical at operation depth.

63
Due to uncertainties in the input parameters and improvement potential, the effects of
different sheave stiffness and damping modelling has not been further investigated at this
stage. Both the value of the suspension stiffness and the linear behaviour found from JMC
Engineering (2006b) should be verified. In addition, the improvement potential of a modified
system will be dependant on space limitations and the chosen solution.

5.4 Comparison of results

Results and discussions of the measurements, SIMO analysis and simplified calculations have
now been presented. In this section, relevant results from each method will be compared to
give an indication of the representation of the real ROV launch and recovery operation.

When comparing the results, it is important have the simplifications in the SIMO model and
uncertainties in the input parameters in mind. Examples are neglected wind forces and sheave
suspension stiffness in the SIMO model and only a JONSWAP wave spectrum using linear wave
theory is used to model the irregular waves. The real environment and equipment will be more
complex. Also the measurements includes uncertainties due to transformation, filtering and
lack of references.

5.4.1 Steady accelerations

In Figure 5.1, measured accelerations in the wave zone are presented. This represents a launch
in approximately H s = 4 m and T z = 7 s, based on the obtained MIROS data Larsen (2015). The
horizontal accelerations are found to be between −3.5 m/s 2 and 2.5 m/s 2 . In addition, the
period indicates that the ROV enters the water more than once. To check if the SIMO model
can reproduce these accelerations, four plots of the ROV acceleration in global y-direction are
presented in Figure 5.14. The vertical position of the buoyancy element relative to the wave
elevation is used to give an indication of the submergence of the ROV. Different wave seed have
been used to obtain the data in H s = 4 m and T p = 8 s with winch speed 0.4 m/s.

The results from the current SIMO model confirms that horizontal accelerations in the order of
3 m/s 2 are possible. The upper right plot shows the largest positive acceleration obtained after
running 30 different wave seeds. After 195 seconds, the acceleration increases from −2 m/s 2 to
3.5 m/s 2 . The sudden decrease in relative vertical position indicates impact with a wave top.
Variations in vertical relative position is assumed to be mainly due to the wave elevation. Rise

64
in horizontal acceleration combined with a local position decrease is also seen in the other
plots.

Global ROV y-acceleration in the splash zone

10
6 8
y-acceleration [m/s 2 ] y-acceleration [m/s 2 ]
4 Vertical position [m] 6 Vertical position [m]
4
2 2
0 0
-2
-2
-4
-4 -6
-6 -8
-10
195 200 205 210 215 220 195 200 205 210 215 220

10 10
8 y-acceleration [m/s 2 ] 8 y-acceleration [m/s 2 ]
6 Vertical position [m] 6 Vertical position [m]
4 4
2 2
0 0
-2 -2
-4 -4
-6 -6
-8 -8
-10 -10
195 200 205 210 215 220 195 200 205 210 215 220
Time [s] Time [s]

Figure 5.14: ROV acceleration in global y-direction versus position relative to wave elevation in
H s = 4 m and T p = 8 s with winch speed 0.4 m/s.

Both the measured accelerations and the results from SIMO shows large variation in the
horizontal accelerations. In addition, the SIMO results includes large transient accelerations
before the winch is started, which will influence the ROV accelerations until the first water
contact. This makes direct comparison difficult. However, the results have given an indication
of the magnitude of the accelerations due to water impact and wave forces. The analysis results
shows that horizontal accelerations up to 3.5 m/s 2 occurs with the current SIMO model in a
similar sea state. The largest horizontal accelerations happens when a wave top is hit, as
indicated by the upper right plot in Figure 5.14. Based on the motion measurements, the SIMO
model seems to provide realistic horizontal accelerations in the wave zone.

5.4.2 Vertical acceleration spikes

Due to lack of references, it is difficult to verify theories related to the motion measurements.
That is particularly critical for the suspected snap load occurrence presented in Figure 5.2. To
further investigate possible causes, the acceleration measurements are compared to vertical
ROV accelerations during a snap load occurrence in SIMO.

65
In Figure 5.15, the largest snap load occurrence during launch in H s = 4 m and T p = 8 s with
winch speed 0.4 m/s is presented. The top plot shows global vertical velocity and acceleration
of the ROV related to the umbilical tension, while the bottom plot indicates the corresponding
vertical wave properties. In addition, the ROV buoyancy element vertical position is presented
to indicate the degree of submergence.

Snap load with winch velocity 0.4 m/s

Tension [0.1*kN]
10 ROV vertical velocity [m/s]
ROV vertical acceleration [m/s2 ]
8

-2
204 206 208 210 212 214 216 218 220

4
2 ROV vertical position [m]
Wave elevation [m]
0 Particle vertical velocity [m/s]
-2 Particle vertical acceleration [m/s2 ]
-4
-6
204 206 208 210 212 214 216 218 220
Time [s]

Figure 5.15: ROV motion and wave kinetics in H s = 4 m and T p = 8 s with winch speed 0.4 m/s.

The vertical ROV accelerations from SIMO shows a similar pattern as the measurement in
Figure 5.2. First there is a negative acceleration lasting approximately 1.5 seconds. In the
analysis this accelerations has a slightly shorter duration, but ends at a higher negative
acceleration. Both cases ends by an acceleration spike of around 9.5 m/s 2 followed by
oscillations. The lower plot in Figure 5.15 shows that slack in the umbilical starts when the ROV
is partially submerged, but the snap load happens when the ROV is fully submerged under the
wave top. Damping in water explains the underdamped behaviour for the following
oscillations. The first three oscillations in the measured vertical acceleration indicates an
undamped behaviour, suggesting that the ROV is out of the water. In addition, the vertical
acceleration decreases before the spike in the measurement. Assuming slack umbilical,
reduction in static weight as parts of the ROV system exits the water can be an explanation.
Also decreasing downward water particle acceleration before the wave exit can be a cause. This
requires the relative contribution from the inertia forces to be important.

66
Based on the comparison, the analysis from SIMO indicates that the measured spike can be
due to re-tension in the umbilical. Also the snap velocity of 0.79 m/s found in Figure 5.12
coincides well with the approximated velocity change of 0.83 m/s during the acceleration spike
in Figure 5.2. However, limited references makes it difficult to conclude. In addition, the
comparison shows that the vertical position of the ROV relative to the water surface may be
different during the acceleration spikes. If the measured spike is re-tensioning in the cable, the
vertical position relative to the water surface will influence the snap load due to different mass
term, buoyancy force and exit forces in the water surface.

5.4.3 Umbilical tension

The simplified calculations gave larger snap loads compared to the SIMO analysis. A free fall
velocity of 3.07 m/s is the main reason. That was based on the drag force expression, and was
not dependant on the duration of slack umbilical. Neither the SIMO analysis nor the
measurements proves the occurrence of such large vertical velocities. This can indicate that the
procedure for finding the snap velocity is over conservative. However, a certain safety factor is
required due to uncertainties in the analysis input. Also the limited amount of measurements
combined with uncertainties in the post processing must be taken into consideration. One
example is that pitch motions of the ROV can influence the vertical acceleration
measurements. Considering the gyroscope data will limit that uncertainty.

The largest snap loads from each set of wave seeds in SIMO seems to be a result of the same
occurrence. The ROV enters the water, the vertical water particle velocity causes slack
umbilical and the snap load occurs as the vertical water particle velocity reaches zero and the
particle acceleration is maximum downwards below the wave top. This is not confirmed by the
assumed snap load in the measurement in Figure 5.2. If the ROV is partially out of the water,
exit forces and lower buoyancy forces will increase the umbilical tension as mentioned. In
addition, added mass may still contribute in the mass term for determining the snap load.
Consequently, verification of the snap loads occurring in the SIMO analysis is not achieved
with the current data available. On the contrary, the vertical acceleration pattern during slack
umbilical and the snap load in SIMO shows similarities with the measured acceleration in
Figure 5.2.

67
5.5 Weather criterion

Measurements, time domain simulations and simplified calculations confirms the complexity
and variability of lifting operations in the wave zone. This means that the weather criterion
must be based on a conservative estimation of a close to real procedure.

Large variability in the forces within each sea state makes it difficult to evaluate the weather
criterion. However, the acceleration measurements and the SIMO simulations gives an
indication of the most critical factors. The results shows that lowering the ROV through the
wave zone is the dimensioning phase. That is based on horizontal and vertical accelerations in
the measurements and large umbilical tension obtained in SIMO. Lower measured
accelerations and no slack umbilical occurrence in the simulations confirms the assumption
that recovery is less critical. Further, parameter variations in SIMO indicates that snap loads
following from slack umbilical provides the dimensioning forces for the weather criterion. In
addition, pontoon impact has not occurred in the performed simulations.

The current ROV launch and recovery system is dimensioned for snap loads in the umbilical.
That is based on the simplified calculations, where the dimensioning force is reduced by a
sheave suspension system. The simulations in SIMO without the sheave suspension has not
resulted in higher tension than the safe working load for the umbilical in H s = 4 m and
H s = 5 m. However, even if the safe load is not exceeded, other effects might cause limitations
for the weather criterion. That can be damage to the inner core of electrical and optical
conductors in the umbilical, which is especially relevant for tension impulses (Driscoll et al.,
2000c). At Snorre B, the combination of snap loads and concentrated umbilical loads over the
sheave wheels and winch spooling device can be critical. In addition, uncertainties in the
SIMO input parameters requires a conservative estimate of the forces. The sensitivity analysis
proved that changes in the drag coefficients are particularly critical. A 50 % increase in the
quadratic drag coefficient more than doubled the umbilical tension during a snap load.

The modelling of a sheave suspension system in SIMO reduced both the snap load and the
following tension oscillations. This indicates that a proper sheave suspension arrangement is
the best solution to reduce umbilical tension. However, including this in the weather criterion
evaluation requires detailed knowledge about the properties of the suspension system. When
designing an effective snap load reduction system, information about duration and size of the
impulses will also be necessary. That can be obtained using reliable measurements combined
with statistics from time domain simulations.

68
Chapter 6

Conclusions and further Work

6.1 Conclusions

Based on the description of the launch and recovery phases and procedure, the wave zone was
assumed to be the most critical when determining the weather criterion. Thus, this phase has
been further investigated using simplified calculations, time domain simulations in SIMO and
motion measurements of the ROV. The results confirms that large forces occurs in the wave
zone. In addition, variability in the simulations and measurements indicates the complexity of
the wave zone transit.

The largest umbilical tensions obtained in the time domain simulations were a consequence of
slack umbilical leading to snap loads. Thus, this seems to be the most critical effect during
launch. Impact between the ROV and the pontoon has not occurred in the SIMO simulations
performed, and is assumed to be of less concern. By reducing the winch speed from 0.8 to 0.4
m/s, the number of slack umbilical occurrences decrease. However, the maximum umbilical
tension was of the same size, with an increase of less than 1 % for two sea states. Simulations of
the recovery phase did not lead to slack umbilical, which indicates that the launch phase is
more critical.

Occurrence of slack umbilical coincides with the results from simplified method calculations.
However, the large slamming impact force was not obtained in the time domain simulations.
This indicates that a slamming coefficient of 5 is conservative. The simulations also indicates
that the simplified method overestimates the snap velocity. This resulted in a dynamic
amplification of 5.5 compared to maximum 2.2 from SIMO in H s = 4 m. Further, the
acceleration pattern during snap loads in SIMO was compared to the motion measurements.

69
One occurrence with similar vertical acceleration shape was found. This was based on the
velocity change during the acceleration spike and the pattern prior to the spike. The
comparison can indicate that slack umbilical occurred. However, lack of references for the ROV
position and umbilical tension during the measurement makes it difficult to conclude.

The SIMO analysis indicates that the drag force due to vertical water particle velocity is the
dominating effect for slack umbilical to occur. This means that the drag coefficients used in
SIMO is especially important for the dimensioning forces. Increasing the drag coefficients by
50 % increased one snap load by 133 %. However, the wave kinetics during slack umbilical
indicates that the maximum snap load will be limited by the period of positive vertical water
particle velocity. Wave induced inertia forces also contributes to the snap load, but were less
sensitive to change in the force coefficient. Despite uncertainties in the input parameters,
measured horizontal ROV accelerations were of similar magnitude as the accelerations
obtained in SIMO. This suggests that the model provides an indication of the real ROV
behaviour in the wave zone.

Based on the critical factors, reducing the launch and recovery system stiffness by
implementing an effective sheave suspension is assumed to be a good measure to reduce the
umbilical tension. However, the effect will depend on the characteristics. Implementing the
current sheave suspension in SIMO by changing the stiffness and using a tensioner reduced a
snap load to 62 % and 76 % of the original force respectively. This was compared to assuming
high stiffness in the sheave. Uncertainties in the properties of the current suspension system
can result in simulating lower forces than the real system loads. In addition, the measurement
with the suspected slack umbilical indicates a limited effect of the current system. Thus, more
detailed analysis will require verified properties of the current suspension system or of an
alternative new system. This should be obtained before extending the weather window based
on a sheave suspension system.

70
6.2 Recommendations for Further Work

In this work, the main focus has been to determine the critical factors for the weather criterion
and measures that can influence these. This has been based on simulations and measurements
where assumptions and uncertainties influences the results.

The interpretation of motion measurements requires position references to reduce the

uncertainties. Measures like video of the behaviour during water entry or data from the other
ROV sensors synchronized in time, can be used to provide the position relative to the water
surface. More detailed full scale measurements can also include better environmental
description. An example can be time series of the wave elevation.

Further, including motion measurements of the sheave frame can be used to investigate the
sheave suspension system behaviour. Based on the measurements, the effect of the suspension
on stiffness and damping can be evaluated. This may be used to investigate whether the end
stops in the suspension cylinders are hit. In addition, accelerations of the ROV can be
compared to the sheave motions. Based on the comparison, a better indication of the
umbilical tension during acceleration spikes may be obtained. If the weather criterion should
be changed based on the sheave suspension, the influence on stiffness and damping needs to
be verified. Reliable measurements of the umbilical tension will provide a good basis to
evaluate the weather criterion, and can be used together with motion measurements as
described in Driscoll et al. (2000b).

If more reliable measurements are obtained, the model of the ROV system in SIMO can be
improved. Especially the drag coefficient has proved to be important for occurrence of slack
umbilical in the wave zone and the resulting forces. Thus, the force coefficients should be
properly verified. Alternative methods for determining the coefficients can also be carried out.
This can be model tests or numerical calculations. Further, different methods to model the
ROV can be compared. The current model has been made based on assumptions about mass
and buoyancy distribution, which introduces additional uncertainties.

An improved model verified by measurements can provide statistics to determine a new

weather criterion. Additional parameters in the criterion can also be considered, for example
the wave period or platform motion. Variation of the wave period in the SIMO simulations
indicates variability within each significant wave height, as illustrated in Appendix B. Reliable
data of all the criterion parameters must then be available for the decision making prior to the
operation. A clear limit for the accepted loads in the umbilical and launch and recovery system
will also be required when determining the criterion. In addition to the safe working load limit,

71
other factors like bending radius of the umbilical can affect the accept criterion. High
concentrated loads will influence wear and damage to the internal components of the
umbilical, which increases maintenance costs and down time of the ROV system. This is
especially important if snap loads in the umbilical are accepted. Based on this, the benefits of
extending the weather criterion should also be considered against higher maintenance
costs.

Further work can be summarized in the following points:

• Further verify the dimensioning effects found from the SIMO simulations by using
reliable measurements combined with alternative modelling and force coefficient
determination for the analysis.

• Establish more accurate properties of the current sheave suspension system by tests,
measurements and observations.

• Determine the potential of a modified sheave suspension system in terms of space

limitations at Snorre B and forces in the launch and recovery system.

• Combine measurements and simulations to establish a new weather criterion.

72
Bibliography

Bruset, O. (2014). Snorre B info – ROV Operations.

COSALT Offshore Norge AS (2012). Certificate No. AHLY43GRU for test and examination of
lifting equipment and appliances. Report.

Det Norske Veritas (2014a). DNV-RP-C205 Environmental Conditions and Environmental

Loads.

Det Norske Veritas (2014b). DNV-RP-H103 Modelling and Analysis of Marine Operations.

Driscoll, F. R., Buckham, B., and Nahon, M. Numerical optimization of a cage-mounted passive
heave compensation system. In Oceans 2000, volume 2, pages 1121–1127. IEEE.

Driscoll, F. R., Lueck, R. G., and Nahon, M. (2000a). The motion of a deep-sea remotely
operated vehicle system part 2: Analytical model. Ocean Engineering, 27(1):57–76.

Driscoll, F. R., Lueck, R. G., and Nahon, M. (2000b). The motion of a deep-sea remotely
operated vehicle system. Part 1: Motion observations. Ocean Engineering, 27(1):29–56.

Driscoll, F. R., Nahon, M., and Lueck, R. G. (2000c). A comparison of ship-mounted and
cage-mounted passive heave compensation systems. Journal of Offshore Mechanics and
Arctic Engineering, 122(3):214–221.

Faltinsen, O. M. (1990). Sea loads on ships and offshore structures. Cambridge University Press,
Cambridge.

Irgens, F. (1999). Formelsamling mekanikk: statikk, fasthetslære, dynamikk, fluidmekanikk.

Tapir, Trondheim. 1. utg. 1987 3. utg.

JMC Engineering (2006a). Dynamic amplification for ROV sheave. Report S6-KC-NCE-1000.

JMC Engineering (2006b). Structural calculation of the ROV-sheave suspension at Snorre-B.

Report S6-KC-NCE-1001.

73
Larsen, C. M. (2012). Marin dynamikk: kompendium for bruk i faget TMR 4182 Marin
dynamikk ved Institutt for marin teknikk, Fakultet for ingeniørvitenskap og teknologi, NTNU,
volume UK-2012-09. Marinteknisk senter., Trondheim. Omslagstittel: TMR4182 Marin
dynamikk.

Larsen, K. (2015). MIROS Data - Snorre B.

Madgwick, S. (2013). Oscillatory-Motion-Tracking-With-x-IMU.

http://github.com/xioTechnologies/Oscillatory-Motion-Tracking-With-x-IMU. Online;
accessed 14-April-2015.

Madgwick, S. O. H., Harrison, A. J. L., and Vaidyanathan, R. (22275550). Estimation of imu and
marg orientation using a gradient descent algorithm. In Rehab Week Zurich 2011 - 2011 IEEE
International Conference on Rehabilitation Robotics, ICORR 2011.

Morison, J. R., Johnson, J. W., and Schaaf, S. A. (1950). The force exerted by surface waves on
piles. Journal of Petroleum Technology, 2(5):149–154.

Øritsland, O. (1989). A summary of subsea module hydrodynamic data, volume 511110.05.

Norsk marinteknisk forskningsinstitutt, Trondheim. Arkivnr. MT51 89-0045.

Øritsland, O. and Lehn, E. (1987). Hydrodynamic forces on subsea modules during lifting
operations, volume 511003.02. Norsk marinteknisk forskningsinstitutt, Trondheim. Arkivnr
MT51 87-0194.

Saipem LTD (2012). MRV09 Load Test. Report WI-SUKL-ASST-947-E.

Sarpkaya, T. and Isaacson, M. (1981). Mechanics of wave forces on offshore structures. Van
Nostrand Reinhold, New York.

SIMO Project Team (2013). SIMO - Theory Manual Version 4.0 rev. 3. Report 516412.00.03.

x-IO Technologies (2012). Open source IMU and AHRS algorithms.

http://www.x-io.co.uk/open-source-imu-and-ahrs-algorithms/. Online; accessed
14-April-2015.

Xsens Technologies B.V. (2010). MTi and MTx User Manual and Technical Documentation.
Report MT0100P.

74
Appendix A

Motion measurements

In this appendix, accelerations and angular velocities from the IMU data loggings are
presented. To remove the gravitational acceleration, the accelerations have been transferred
into a global reference system. However, neither the gyroscope nor accelerometer data have
been subjected to further filtering to avoid removing relevant details in the measurements.
Only the parts showing launches or recoveries are included in the plots.

Accelerometer (Global reference system)

8 X
6 Y
Z
4
Accelerations [m/s 2 ]

2
0
-2
-4
-6
-8
-10

20 40 60 80 100 120 140 160 180 200 220

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
20 40 60 80 100 120 140 160 180 200 220
Time [s]

Figure A.1: Launch from 2015-04-02 at 19:58

i
 Accelerometer (Global reference system)
4
X
Y
3 Z

2
Accelerations [m/s 2 ]

-1

-2

-3

20 40 60 80 100 120 140 160 180 200

Angular velocities [dps]

Gyroscope
0.2
X
Y
0 Z

-0.2
20 40 60 80 100 120 140 160 180 200
Time [s]

Figure A.2: Launch from 2015-04-04 at 14:19

Accelerometer (Global reference system)

6
X
Y
4 Z
Accelerations [m/s 2 ]

-2

-4

20 40 60 80 100 120 140 160 180 200 220

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
20 40 60 80 100 120 140 160 180 200 220
Time [s]

Figure A.3: Launch from 2015-04-05 at 03:47

ii
 Accelerometer (Global reference system)
4 X
Y
3 Z
Accelerations [m/s 2 ]

-1

-2

-3

150 200 250 300 350

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
150 200 250 300 350
Time [s]

Figure A.4: Launch from 2015-04-09 at 22:36

Accelerometer (Global reference system)

6 X
Y
4
Z
2
Accelerations [m/s 2 ]

-2

-4

-6

-8

-10
220 230 240 250 260 270 280 290 300 310 320
Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
220 230 240 250 260 270 280 290 300 310 320
Time [s]

Figure A.5: Recovery from 2015-04-12 at 20:32

iii
 Accelerometer (Global reference system)
X
8 Y
Z
6
Accelerations [m/s 2 ]

-2

-4

50 100 150 200 250 300 350

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
50 100 150 200 250 300 350
Time [s]

Figure A.6: Launch from 2015-04-13 at 01:47

Accelerometer (Global reference system)

6 X
Y
4 Z
Accelerations [m/s 2 ]

-2

-4

-6

150 160 170 180 190 200 210 220 230 240 250
Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
150 160 170 180 190 200 210 220 230 240 250
Time [s]

Figure A.7: Recovery from 2015-04-14 at 00:15

iv
 Accelerometer (Global reference system)
X
10 Y
Z
Accelerations [m/s 2 ]

-5

-10

40 60 80 100 120 140 160 180 200

Angular velocities [dps]

Gyroscope
0.5
X
Y
0 Z

-0.5
40 60 80 100 120 140 160 180 200
Time [s]

Figure A.8: Launch from 2015-04-14 02:50

v
Appendix B

SIMO analysis

In this appendix tension plots extracted from SIMO are presented. For each sea state, 30
different wave seeds are run. The winch is started 160 seconds after the simulation start, and
the velocity is set to 0.8 m/s. All simulations presented are without the sheave suspension
modelling. Total simulation time was 300 seconds, but only the most critical phase is
presented. A two parameter JONSWAP spectrum was used, and γ together with the
relationship between T z and T p are found in Table 4.12.

Hs = 4 m and Tp = 7 s

120

100
Tension [kN]

80

60

40

20

0
175 180 185 190 195 200 205 210
Time [s]
120 Maximum tension [kN]
Tension [kN]

100 Minimum tension [kN]

80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.1: Umbilical tension in H s = 4 m and T p = 7 s.

vii
 Hs = 4 m and Tp = 8 s

100

80
Tension [kN]

60

40

20

0
175 180 185 190 195 200 205 210
Time [s]
120
Maximum tension [kN]
Tension [kN]

100
Minimum tension [kN]
80
60
40
20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.2: Umbilical tension in H s = 4 m and T p = 8 s.

Hs = 4 m and Tp = 9 s
100

80
Tension [kN]

60

40

20

0
175 180 185 190 195 200 205 210
Time [s]
100
Tension [kN]

50 Maximum tension [kN]

Minimum tension [kN]

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.3: Umbilical tension in H s = 4 m and T p = 9 s.

viii
 Hs = 4 m and Tp = 10 s
70

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
80
Tension [kN]

60
Maximum tension [kN]
40
Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.4: Umbilical tension in H s = 4 m and T p = 10 s.

Hs = 4 m and Tp = 11 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]

20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.5: Umbilical tension in H s = 4 m and T p = 11 s.

ix
 Hs = 4 m and Tp = 12 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
80
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]

20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.6: Umbilical tension in H s = 4 m and T p = 12 s.

Hs = 4 m and Tp = 13 s
60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.7: Umbilical tension in H s = 4 m and T p = 13 s.

x
 Hs = 4 m and Tp = 14 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.8: Umbilical tension in H s = 4 m and T p = 14 s.

Hs = 5 m and Tp = 8 s
150
Tension [kN]

100

50

0
175 180 185 190 195 200 205 210
Time [s]
160 Maximum tension [kN]
Tension [kN]

140 Minimum tension [kN]

120
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.9: Umbilical tension in H s = 5 m and T p = 8 s.

xi
 Hs = 5 m and Tp = 9 s

120

100
Tension [kN]

80

60

40

20

0
175 180 185 190 195 200 205 210
Time [s]
140
120 Maximum tension [kN]
Tension [kN]

100 Minimum tension [kN]

80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.10: Umbilical tension in H s = 5 m and T p = 9 s.

Hs = 5 m and Tp = 10 s

80
Tension [kN]

60

40

20

0
175 180 185 190 195 200 205 210
Time [s]
80
Tension [kN]

60
Maximum tension [kN]
40
Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.11: Umbilical tension in H s = 5 m and T p = 10 s.

xii
 Hs = 5 m and Tp = 11 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.12: Umbilical tension in H s = 5 m and T p = 11 s.

Hs = 5 m and Tp = 12 s

70

60
Tension [kN]

50

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
80
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.13: Umbilical tension in H s = 5 m and T p = 12 s.

xiii
 Hs = 5 m and Tp = 13 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
80
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.14: Umbilical tension in H s = 5 m and T p = 13 s.

Hs = 5 m and Tp = 14 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
80
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.15: Umbilical tension in H s = 5 m and T p = 14 s.

xiv
 Hs = 5 m and Tp = 15 s

60

50
Tension [kN]

40

30

20

10

0
175 180 185 190 195 200 205 210
Time [s]
Tension [kN]

60
Maximum tension [kN]
40 Minimum tension [kN]
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number [-]

Figure B.16: Umbilical tension in H s = 5 m and T p = 15 s.

xv
Appendix C

Slender elements properties

The properties of the slender elements used to build the ROV model in SIMO are presented in
Table C.1. Linear drag input, B 1 , are set to 10 % of the quadratic drag input based on the
calculated values in Table 4.10, and are not included in the table. In Figure 4.4, the coordinate
system for the directions is shown. Input refers to the values used in SIMO, while total is
calculated to be compared to the calculations.

xvi
 Element name Length [m] Radius Displ. CoB [m] Am input [kg /m] Total A m [kg ] B 2 input [N s 2 /m 3 ] Total B 2 [N s 2 /m 2 ]
2
y x z R [m] A [m ] V [m 3 ] z y x z y x z y x z y x z
buoy_ROV_1 3,2 0 0 0,37 0,430 1,376 2,064 0 0 0 0 0 0 0 0 0 0 0 0
buoy_ROV_2 3,2 0 0 0,37 0,430 1,376 2,064 0 0 0 0 0 0 0 0 0 0 0 0
lFrame_ROV_1 3,2 0 0 0,103 0,033 0,107 0,000 0 95 120 0 304 384 0 190 245 0 608 784
lFrame_ROV_2 3,2 0 0 0,103 0,033 0,107 0,000 0 95 120 0 304 384 0 190 245 0 608 784
lFrame_ROV_3 0 1,5 0 0,103 0,033 0,050 0,000 80 0 120 120 0 180 180 0 245 270 0 367,5
lFrame_ROV_4 0 1,5 0 0,103 0,033 0,050 0,000 80 0 120 120 0 180 180 0 245 270 0 367,5
tFrame_ROV_1 3,2 0 0 0,103 0,033 0,107 0,213 0 95 120 0 304 384 0 190 245 0 608 784
tFrame_ROV_2 3,2 0 0 0,103 0,033 0,107 0,213 0 95 120 0 304 384 0 190 245 0 608 784
tFrame_ROV_3 0 1,5 0 0,103 0,033 0,050 0,100 80 0 120 120 0 180 180 0 245 270 0 367,5
tFrame_ROV_4 0 1,5 0 0,103 0,033 0,050 0,100 80 0 120 120 0 180 180 0 245 270 0 367,5
mFrame_ROV_1 0 0 2 0,103 0,033 0,067 0,067 80 95 0 160 190 0 180 190 0 360 380 0

xvii
mFrame_ROV_2 0 0 2 0,103 0,033 0,067 0,067 80 95 0 160 190 0 180 190 0 360 380 0
mFrame_ROV_3 0 0 2 0,103 0,033 0,067 0,067 80 95 0 160 190 0 180 190 0 360 380 0
mFrame_ROV_4 0 0 2 0,103 0,033 0,067 0,067 80 95 0 160 190 0 180 190 0 360 380 0
mFrame_TMS_1 0 0 2 0,103 0,033 0,067 0,200 80 95 0 160 190 0 180 190 0 360 380 0
mFrame_TMS_2 0 0 2 0,103 0,033 0,067 0,200 80 95 0 160 190 0 180 190 0 360 380 0
mFrame_TMS_3 0 0 2 0,103 0,033 0,067 0,200 80 95 0 160 190 0 180 190 0 360 380 0
mFrame_TMS_4 0 0 2 0,103 0,033 0,067 0,200 80 95 0 160 190 0 180 190 0 360 380 0
tFrame_TMS_1 1,5 0 0 0,103 0,033 0,050 0,200 0 95 120 0 142,5 180 0 190 245 0 285 367,5
tFrame_TMS_2 1,5 0 0 0,103 0,033 0,050 0,200 0 95 120 0 142,5 180 0 190 245 0 285 367,5
tFrame_TMS_3 0 1,5 0 0,103 0,033 0,050 0,200 80 0 120 120 0 180 180 0 245 270 0 367,5
Table C.1: ROV slender element properties in SIMO.

tFrame_TMS_4 0 1,5 0 0,103 0,033 0,050 0,200 80 0 120 120 0 180 180 0 245 270 0 367,5
Total 4,112 1.61 2000 3021 2976 4500 6042 6076
Appendix D

IMU Calibration certificate

xviii
xix