1 INTRODUCTION

Norway has a very long coastline with many wide

and deep fjords. The increased focus on efficient, safe

and environmentally friendly traffic in coastal areas

has led to a debate on different ways for fjord

crossing. The plan for a ferry-free main road from

Kristiansand to Trondheim along the western part of

Norway is shown in figure 1. The plan has been

prepared by the Norwegian Road Authority. To fulfil

the objective of the plan, eight existing ferry routes

must be removed and replaced by other types of fjord

crossing, such as very long bridges (including floating

types) and subsea tunnels. Based on new cost

estimates for different fjord crossing scenarios, some

of the crossings will continue to be operated with

ferries. However, to fulfil the national goals for coastal

shipping [7] the ferry fleet must be renewed by

introducing low, preferably zero, emission vessels.

For some of the long and exposed routes in the

Manoeuvring Study – Norwegian Double-Ended Ferry

T.E. Berg

, Ø. Selvik

, K. Steinsvik

& D. Leinebø

SINTEF Ocean, Trondheim, Norway

HAV Design, Fosnavaag, Norway

ABSTRACT: The Norwegian coastline has many long fjords where crossings are necessary for transportation of

goods and passengers. In the last decade, the focus on reduced travel time along the main roads in coastal areas

has increased the building of bridges and subsea tunnels. However, at present and in the future many fjord

crossings will depend on ferries. The Norwegian government [7] requires that ferries, like all coastal ships in

Norwegian waters, should be designed for zero or low greenhouse gas (GHG) emissions to meet the national

goal of 50% reduction of GHG from coastal shipping by 2030. As ferry services are regulated by national or local

governmental bodies, all new ferry operations should be performed using zero- or low-emission ferries. Thus,

ferry companies require new and innovative ferry designs with reduced resistance, resulting in reduced

installed propulsion power.

This paper describes work done by the ship designer HAV Design AS (former Havyard Design & Solutions AS

(HDS)) to meet the governmental request for ferries with a low environmental footprint. Work on a double-

ended ferry design is described. In the early design phase manoeuvring performance is not a priority item,

partly due to lack of a simple and reliable manoeuvring performance prediction tool for unconventional ship

designs. It is well known that optimization of resistance can be at the cost of manoeuvring performance. In this

paper, a specific double-ended ferry design will be used as a case. Outcomes of design simulation of

manoeuvring performance are compared to manoeuvring full-scale tests in deep, calm water. Full-scale test

results will later be used to tune a simulation model for a future training simulator for double-ended ferry,

where full-scale manoeuvring tests have been performed, will be used as a test case. This paper shows how the

designer has worked with these two topics in parallel in the final design stage where both experimental and

numerical tools have been used for design verification.

http://www.transnav.eu

the International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 15

Number 1

March 2021

DOI: 10.12716/1001.15.01.05

northern part of Norway, low-emission ferries are the

only solution. Previous tender documents prepared

by the national road authority and regional authorities

responsible for transport services ask for zero or low

emissions fiord crossing ferries. In the process of

evaluating service offers, the two main parameters

have been operational costs and emission footprint. In

these tenders, the cost is weighted 70% and

environmental aspects 30% when comparing offers

from ferry companies, meaning that the concept with

the lowest price will not necessarily win. Lately, the

new tenders from different governmental bodies

require zero emission for fiord crossing ferries. Thus,

the only criteria when comparing offers will be the

total costs (investments and operations) for a given

licence period.

In the competition for new transport contracts, all

ferry companies need new vessels or retrofitting old

vessels with zero/low emission engines. As a result,

national ferry companies have requested new vessel

designs from many Norwegian ship designers. This

paper describes design challenges related to double-

ended ferries, how design tools and model tests were

used to verify a specific ferry design (H-936)

developed by HAV Design for a 120 car capacity

ferry. Sea trials were used to validate manoeuvring

performance of these classes of ferries. Four H-936

ferries have been built and operating on different

routes in Norwegian fjords.

Figure 1. Future ferry free main road (E 39) from

Kristiansand to Trondheim (courtesy Norwegian Road

Authority).

2 DESIGN CONSIDERATIONS

Many of the double-ended car ferries in Norway

operate in sheltered waters, but these waters have

many obstacles, such as islets or underwater reefs,

combined with narrow passages and shallow water

effects close to the ferry quays. Transit routes are often

short, many times shorter than thirty minutes. This

means many docking operations every day, often

between twenty and forty.

The main operation can be divided in four phases:

acceleration – transit – retardation and manoeuvring /

berthing. Very often both retardation and final

manoeuvres overlap. Before entering the quay, waters

are often both narrow and shallow introducing

shallow water and bank effects changing the

manoeuvring performance. In addition to these

effects, the captains have experienced unusual

manoeuvring responses during retardation before

arrival at the ferry port. This has led to development

of special operational procedures on how to use the

fore and aft azipulls in the retardation phase. Adding

the influence of harsh weather conditions increase the

challenges to be coped with by the bridge crew.

When ferry companies compete for governmental

fjord crossing contracts, two factors are weighted in

the selection process – operational costs and emission

footprint. In addition, operational regularity is a

challenging as it is required that operation should be

possible under all weather conditions, except extreme

cases where the weather gets a specific name. As

operational costs are strongly related to energy

consumption (electricity, different types of fuels or

hybrid solutions), the design will focus on resistance

reduction and propulsion efficiency in addition to

selection of engine power sources.

To be able to fulfil governmental requests for

reduced GHG emissions from new ferries, the ferry

companies have requested new vessel designs from

many Norwegian ship designers. Double-ended ferry

designs are often selected to reduce turn-around time

at the ferry quays. As most of the crossings are inside

the fjords, the critical factors for an energy efficient

hull design are:

− Low calm water resistance

− High propulsive efficiency

− Controllability in strong winds near quays

Additional items to be considered are green

water/sea spray on the car deck and directional

stability (to reduce the additional resistance from

using control units).

Figure 2 is a generic illustration of a ship design

process. As can be seen from the “Performance” list,

manoeuvring characteristics is missing. This could

result in new ship designs giving ship masters that is

difficult to handle, especially in confined waters and

ports.

Figure 2 Design process used for a double-ended ferry.

To illustrate design tools used by HAV Design, the

work related to development of their H-936 (120 car)

double-ended ferry is used. The main particulars of

the vessel are listed in Table 1 and a picture of the

vessel in the transit phase is shown in Figure 3. A brief

overview of the tools (experimental and numeric)

used by HAV Design will be described. For

illustrative purposes, the H-936 is used as a case.

Tools applied are listed in Table 2.

Table 1. Main particulars and propulsion units of the HAV

Design H-936 double-ended ferry design.

_______________________________________________

Geometric parameters

_______________________________________________

Length overall Loa 111 m

Length between perpendiculars Lpp 84 m

Breadth B 17,5 m

Max draught Tmax 3,6 m

Test draught T 3,0 m

Block coefficient CB 0,36

Main propulsion 2 x 1200 kW

2 x Rolls-Royce AZP85-EL FF 12.5 TME

_______________________________________________

Figure 3. Double-ended ferry – HAV Design, double-ended

design H-936.

Table 2. Hydrodynamic tools used for designing the H-936

double-ended ferry.

_______________________________________________

Experimental hydrodynamics (model tests)

_______________________________________________

Open water tests, stock and design propeller

Thruster fwd/aft load variations

Resistance tests, 2 waterlines

Self-propulsion tests, 2 waterlines

Trim optimization tests

Seakeeping tests, regular waves

Seakeeping tests, irregular waves - free running

Calm water stop and acceleration tests

_______________________________________________

Numerical hydrodynamics

_______________________________________________

VeSim (seakeeping, manoeuvring and dynamic positioning)

CFD (resistance, manoeuvring, hull optimization)

Stability

Internal weight analysis

Internal route, energy and fuel study

Station keeping analyses for thruster capability

Veres motion code:

- Added resistance indication

- Motion characteristics optimization

CAESES

- Automated CAD/CFD optimization of hull

- Ship speed and powering

- Calm water resistance

- Propeller diameter optimization

- Performance in real seas

Star CCM

Calm water flow analyses optimization

_______________________________________________

3 INITIAL DESIGN VERIFICATION TESTS

Using the given design, HAV Design contracted

SINTEF Ocean to conduct a series of different model

tests and numerical studies (see Table 2) to investigate

design characteristics and to verify the numerical

models used in the early design phase. Experimental

studies were performed in SINTEF Ocean’s Towing

Tank. At this stage studies of manoeuvring

performance were not included. Model scale used for

these tests was 1:12.637.

The outcomes of these tests were compared to

numerical predictions from the tools described in the

previous section. Figure 4 compares predicted power

(on the electric motor) and measured values for sea

trial with four sister vessels. The figure shows that

predictions based on the model test overestimates the

measured from the four sister vessels of the H-936

design for speeds above 11 knots. For the first three

delivered vessels (ships 1 - 3), trial trip results have

been corrected using the ISO standard [4]. This will

also be done for the fourth ship (yard trials not

analysed yet). The figure includes computational fluid

dynamics (CFD) calculated resistance. As can be seen,

these results compare well with sea trial data.

Figure 4. Comparison between sea-trial data on sister

vessels and model test results.

4 DESIGN VERIFICATION OF MANOEUVRING

PERFORMANCE BASED ON MODEL TESTS

For an MSc thesis [5] HAV Design and NTNU

(Norwegian University of Science and Technology)

collaborated to fund the building of a new hull model

and to run additional model tests in SINTEF Ocean's

Towing Tank. For Planar Motion Mechanism (PMM)

and oblique towing tests the model scale was 1:15.33

giving an approximately 8 m long mode (somewhat

smaller than the model used for resistance and

propulsion tests). Figure 5 shows the model in the

Hexapod system.

The focus of these tests was to investigate the

quality of numerical tools for simulation of

manoeuvring performance. In principle, two methods

were used to generate input data to SINTEF Ocean's

six degree for freedom (6 DOF) time-domain vessel

simulation tool VeSim [9]. The first one is purely

numerical using the Hullvisc program to generate the

hydrodynamic input file for VeSim [6]. It is based on

linear slender body theory and a cross-flow drag

formulation for non-linear damping forces. The

second one is based on experimental data from model

scale oblique towing and PMM tests. In the thesis,

Leinebø compared outcomes from these sets of input

data. A mathematical three degree of freedom (3 DOF)

solver for the non-linear coupled surge, sway and yaw

equations (SIMAN), was also used with only

numerical input data. The simulations were

performed with the aft propulsion unit working and

the fore azipull turned off.

Figure 5. HAV Design H-936 connected to seakeeping

carriage for PMM tests.

Challenges were experienced using an azipull

propulsion model, which were solved by utilising a

simplified propulsion and rudder model. They were

tuned to and compared against a working model in

SIMAN (previous time-domain manoeuvring

simulation tool used by SINTEF Ocean). This model

showed good agreement with different standard

manoeuvring tests [3] with rudder/azipull angles up

to 20-25 degrees (with the same numerical hull input).

Figure 6 shows the results of a complete spiral

simulation using both time-domain simulation tools.

In figure 7 there is good agreement for the full-scale

sea trial and predicted turning circle manoeuvre when

VeSim used a 20- and 25-degrees rudder angle and

SIMAN used a 35 degrees azipull angle.

Figure 6 shows some challenges when using the

numerical HullVisc tool (based on slenderbody and

strip theory) to create the input field for VeSim. This

is mainly due to the design of double-ended ferries

where the combination of main dimensions is outside

the parameter range for regression formulas used

when developing the present numerical tool. The

numerical method experienced some unstable

solutions because of this. A new numerical tool based

on CFD are under development at SINTEF Ocean, and

it is presently tested for a smaller HAV Design (50 car)

double-ended ferry. Until CFD results are validated

against model test results, the simpler and faster

HullVisc tool will be applied in the early design

phase.

Figure 6. Complete spiral manoeuvre simulations of H-936.

Figure 7. Turning circle manoeuvre. Simulations vs full-

scale of H-936 (identical SB and PS simulations)

Two different ways of using experimental results

were applied. Both are based on mapping force

measurements from the experiments to common time

series for surge and sway forces and yaw moment. A

modified least square method was then used to

estimate hull force coefficients in a 3 degree of

freedom model developed by Ross [8] integrated into

the six degree of freedom model used in VeSim. In the

first one, only PMM test data was used and in the

second the combined oblique towing and PMM tests

were used. Including the oblique towing force

measurements gave a large difference in some of the

hydrodynamic coefficients (also for added mass

coefficients). This showed how different set of

coefficients in the motion equations could be

developed from model tests. Even if the values of the

individual coefficient in the simulation varied

significantly, the outcomes from VeSim simulations

were nearly the same for IMO's turning circle test [3].

But this will give challenges for manoeuvres outside

the measured areas.

Table 3 compares numerical and experimental

input data used in VeSim with the average overshoot

angles from full-scale trials. The table presents the

differences between the full-scale average overshoot

angles and the predicted overshoot angles using

different VeSim input. Negative values represent

lower overshoot angles in the simulations compared

with full-scale results, and positive values vice versa.

The experimental input shows the improvement of the

overestimation of the vessel’s course-keeping and

course-changing ability in the numerical input. The

RMS values from the sea trials varied between 1-2

degrees, resulting in a high percentage deviation at

low overshoot angles. The 5°/5° zig-zag manoeuvre

was requested by the master during sea trials in May

2019 (see section 5).

Guidance notes from NAUT(AW) class notation

states [1]: The characteristic parameters should be within

15% of the parameters obtained from the full-scale trials. If

deviation exceeds this figure, the whole full-scale trial

program should be completed.

Comparing overshoot angle predictions from

VeSim with the measured sea trial data, it is seen that

some results deviate more than 15%.

Table 3. Zig-Zag manoeuvre test, numerical (HullVisc) and

experimental input (PMM) data in VeSim compared with

full-scale results of H-936.

_______________________________________________

Zig-Zag Manouvre Overshoot angles differences

Simulations 1st (Deg) 2nd (Deg) 3rd (Deg)

_______________________________________________

Numerical input 5/5 + 0.7 - 1.5 - 0.6

10/10 - 1.9 - 5.0 - 4.7

20/20 - 4.9 - 2.8 - 5.6

Experimental 5/5 + 1.3 + 0.3 + 1.3

Input 10/10 - 0.1 - 1.8 - 1.7

20/20 - 1.2 + 0.6 + 2.3

_______________________________________________

When applying the present simplified numerical

tool HullVisc to generate input data to VeSim, the

result is large discrepancies between simulated

turning circle characteristics to the ones coming from

experimental input data. This is an expected result as

the HullVisc tool used for a double-ended ferry is

outside of the validity range of correlation factors

used in HullVisc as they are adapted to conventional

merchant ships. The hydrodynamic coefficients in the

manoeuvring equations coming from experimental

PMM results give better prediction of IMO standard

manoeuvres. In the early design phase, there is a need

for a numerical tool for generating input, for

unconventional ship designs such as the double-

ended ferry, to time-domain manoeuvring simulation

tools. As mentioned earlier, there is an ongoing

activity to develop a specific CFD code for better

estimation of hydrodynamic coefficients for the

manoeuvring equations for unconventional ship

designs.

Both numerical and experimental input in the

turning circle simulations showed a much higher peak

in rate of turn in simulations, with this a quicker

established drift angle. The simulations also gave too

low a period between overshoot angles, because of

higher peak values in rate of turn. Further

investigation showed that the simulation did not

capture the full effect of the fore azipull.

5 VALIDATION OF MANOEUVRING

PERFORMANCE – IMO STANDARD

MANOEUVRES

For documentation of the manoeuvring performance

of the H-936 design, the ferry company Fjord1 made

the double-ended ferry MF Suløy available for sea

trials in May 2019. The tests were performed in

Vartdalsfjorden where the water depth is 300 m,

figure 8. Weather conditions were excellent, with no

appreciable waves and a very low wind speed (2.8

m/s, direction 166°). Tidal current during the test

period was not measured. Two sources were used for

data collection. The main data source was the vessel's

Integrated Automation System (IAS), where among

position, heading, thruster settings, rpm and thruster

angle were measured. In addition, SINTEF Ocean a

dual Global Positioning System (GPS) for measuring

position and heading. All signals were recorded

synchronously as time series with a sampling rate of 1

Hz.

Figure 8. Test area for manoeuvring test with car ferry "MF

Suløy".

An overview of the main test program is given in

Table 4. The first part of the test program included

IMO standard tests. Outcomes of these tests were

used in the preparation of the vessel's Manoeuvring

Booklet [2]. The second part was manoeuvres

specified by the master on the ferry. Finally, some

IMO standard tests were repeated with a small

forward trim (obtained by positioning heavy trucks at

the bow).

Table 4. Manoeuvring tests from MF "Suløy" test campaign

May 2019. Test speed 13 knots.

_______________________________________________

Test type Number Test parameters Comment

of tests

_______________________________________________

Turning circles 7 Azipull angle 35° One 720°

Zig-Zag 11 5°/5°, 10°/10°,

20°/20°

Direct spiral 4

Reversed spiral 9 5°/min, 10°/min, Investigating

15°/min width of

hysteresis loop

Stopping 21 Varying engine

control modes,

manual control

_______________________________________________

Figure 9 is a picture of the vessel performing an

IMO zig-zag test. As can be seen, the sea in the test

area is calm. Some initial validation studies using the

full-scale tests have been described in section 4. Figure

7 (in section 4) compares turning circle paths and

ways to handle the deviations between measurements

and predictions by either changing the control angel

of the azipull unit or replace the unit by a

conventional rudder/propeller system. It was found

that the applied generic azipull model gave too high

control forces for high angles. Simulations were thus

also run using different azipull angles – the best

turning circle results were obtained using a 25°

control angle to port and 20° to starboard. Also, for

the overshoot angles, there are some differences

between measured and predicted values. Improved

outcomes from VeSim are obtained by tuning the

hydrodynamic coefficients and especially control

system parameters using the deviations between the

initial VeSim predictions and full-scale

measurements.

There are two main reasons for the difference

between sea trial manoeuvres and VeSim predictions.

The first one is due to limitations introduced by the

Hexapod system used by SINTEF Ocean for PMM

tests. For highly manoeuvrable vessels, such as

double-ended ferries with azipull systems, it is not

possible to obtain the high drift angles and yaw rates

that are measured during turning circles and other

tests applying large control unit angles. The second

one comes from the existing model of the hull and

azipull unit interaction in VeSim. CFD is presently

used to study these interactions and new models for

interaction effects will be developed for double-ended

ferry designs.

Figure 9. Zig-zag test seen stern-wise from the bridge of MF

Suløy.

The results of the manoeuvring sea trials have been

used by SINTEF Ocean and HAV Design to tune hull

and propulsion force models for the VeSim simulation

tool. The goal of this work has been to reduce the

differences between sea trial measurements and

VeSim predictions. The reliability of VeSim with

numerical input for new hull forms needs to be

improved so it can be used as one of the design tools

by HAV Design in their work to develop new zero-

/low-emission double-ended ferries with high

manoeuvring performance. Based on tuning

experience, it is concluded that approvements are

needed with respect to:

− Four quadrant asipull model

− Interaction effects between hull and asipull

Two alternative ways of improving VeSim input

for early design phase studies are presently

investigated. One is to extend the present HullVisc

tool to unconventional ship designs by using 3D

potential flow theory. The other one is to generate

input data from a CFD based PMM.

6 SOME COMMENTS ON OPERATIONAL

EXPERIENCE

The IMO standard tests give information for the ship

designer more than for ship captains. Despite double

ended ferries showing a low degree of directional

instability, based on direct and reverse spiral tests,

and qualifying well within IMO’s criteria on zig-zag

tests, there is no guarantee that the ferry is steerable in

critical and typical manoeuvring situations, especially

during the retardation phase. One specific effect that

is purely related to the hull and propulsion units is

the behaviour during retardation. Some of the ferry

captains initially reported unusual effects related to

control of the vessel during retardation manoeuvres.

Analysing the reports, it was concluded that lack of

experience with the actual control system (azipulls

fore and aft) caused some of these effects. The higher

the initial speed when initiating retardation, the more

unstable the vessel will be in the retardation phase.

Such an operational characteristic increases the stress

level for the captains during the final phase of a

voyage. To improve this behaviour, a set of new of

test manoeuvres have been suggested by the captains

to identify the best operational procedure for

operating both of the azipulls during retardation and

docking. The outcome of such tests, and a specific

operational guideline for these part of the operation

under varying environmental conditions, should be

documented in the vessel's Manoeuvring Booklet.

Based on discussions with double-ended ferry

masters, they are asking for better documentation of

low-speed manoeuvring performance, especially for

harsh weather conditions. This should also be part of

the vessel specific Manoeuvring Booklet. In addition

to these tests, which will reveal behaviour during sea

trials, it is even more important to detect

manoeuvring challenges during the design process

and hence modify design to overcome this. Hence, a

methodology for developing a reliable simulation

model that capture the behaviour during retardation

should be specified. The essence is to conduct

systematic CFD studies/ model tests to generate

hydrodynamic coefficient input to the time domain

simulation model so it can be used as an early design

tool. This tool (for instance a simplified version of

VeSim) could then be used by designers to investigate

manoeuvring performance, both IMO standard

manoeuvres and ship specific low speed manoeuvres

requested by captains.

7 VESSEL SPECIFIC MANOEUVRING

HANDBOOK

As described prior, the results from service speed

turning circles, zig-zags and stopping tests were used

to produce the Wheelhouse Poster and the Pilot Card.

Additional IMO manoeuvres such as spiral tests and

manoeuvres specified by the master (especially low-

speed tests) were used by HAV Design in the

development of the ship specific Manoeuvring

Booklet, see figure 10. As mentioned in the previous

section, vessel characteristics during the retardation

phase should be found here.

Figure 10. Front page of Manoeuvring Booklet for double-

ended ferry MF Suløy.

8 CONCLUSIONS

From the MSc study, it has been shown that present

numerical models to predict coefficients in

manoeuvring equations need to be improved. The

position of the separation point, used in the

calculation of hydrodynamic derivatives, deviates

significantly from values used for traditional

displacement vessels. The use of regression type

coefficients is not recommended for double-ended

ferries, partly due to the fore and aft symmetry of the

hull. Based on the comparison of simulated standard

manoeuvres and full-scale measurements, it is

concluded that more work is needed to understand

the influence of the forward azipull in all phases of

the vessel operation. Modifications of VeSim to

include this influence are presently investigated.

Application of least square methods to identify

linear and non-linear force coefficients from captive

model test must be used carefully. Motion parameters

in low-speed manoeuvres may be outside the speed

and acceleration domains used in the tests. More sea

trial data should be obtained for further validation

studies of the manoeuvring models for double-ended

ferries.

An early design phase tool for investigating

manoeuvring performance of double-ended ferries is

under development. Using sea trial data from a set of

double-ended ferries (50, 80 and 120 cars) developed

by HAV Design, SINTEF Ocean works to improve

methods (3D slender-body theory with empirical

corrections and CFD PMM) for creating the

hydrodynamic input files to the time domain VeSim

simulation tool.

IMO standard manoeuvring tests are of little value

for ship captains on double-ended ferries. They are

requesting more information on low-speed

manoeuvring performance which should be

documented in the vessel's Manoeuvring Booklet.

Addition to these tests, which will reveal

behaviour during sea trials, it is even more important

to detect manoeuvring challenges during the design

process and hence modify design to overcome this.

Hence, a methodology for developing a reliable

simulation model that capture the behaviour during

retardation should be specified. The essence is to

conduct systematic CFD studies/ model tests to build

a time domain simulation model to be used as an

early design tool. Such a tool would be used to predict

outcomes of IMO standard manoeuvres, low speed

manoeuvres requested by ferry captains, and special

off design performance characteristics of a specific

ship.

ACKNOWLEDGEMENT

The authors acknowledge the support from Fjord1 which

made it possible to use MF Suløy for an extensive full-scale

manoeuvring test program in May 2019. We also thank the

bridge crew on the ferry for their enthusiasm during the

tests and their proposals for low-speed tests whose results

are included in the Manoeuvring Booklet. Finally, we

appreciate the support from Norwegian Electric Systems

(NES) personnel in their assistance to link the vessel's data

logging system to SINTEF's sea trial instrumentation

package.

REFERENCES

1. DNV GL: Guidance notes NAUT AW class,

https://rules.dnvgl.com/docs/pdf/DNVGL/RU-

SHIP/2015-10/DNVGL-RU-SHIP-Pt6Ch3.pdf, last

accessed 2021/02/24.

2. International Maritime Organisation: Provision and

display of manoeuvring information on board ships.

IMO Resolution A.601(15). , London, UK (1987).

3. International Maritime Organisation: Resolution

MSC.137(76) – Standards for Ship Manoeuvrability,

MSC 76/23/Add.1. , London, UK (2002).

4. ISO 15016:2015: Ships and marine technology —

Guidelines for the assessment of speed and power

performance by analysis of speed trial data. , Geneva

(2015).

5. Leinebø, D.: Prediction of Maneuverability of Double-

Ended Ferry. NTNU (2020).

6. Martinussen, K., Ringen, E.: Manoeuvring Prediction

During Design Stage. Presented at the International

Workshop on Ship Manoeuvrability at the Hamburg

Ship Model Basin , HSVA (2000).

7. Norwegian Government: The Government’s action plan

for green shipping. , Oslo (2019).

8. Ross, A.: Nonlinear Manoeuvring Models for Ships – A

Lagrangian Approach. NTNU (2008).

9. SINTEF Ocean: VeSim,

https://www.sintef.no/en/software/vesim/, last accessed

2021/02/24.