Systems Modelling For Energy-Efficient Shipping

Systems modelling for energy-efficient shipping
Fabian Tillig, Wengang Mao, Jonas W. Ringsberg*

Chalmers University of Technology, Department of Shipping and Marine Technology, Division
of Marine Technology, SE-412 96 Gothenburg, SWEDEN
*
E-mail: Jonas.Ringsberg@chalmers.se
Summary
There is a wide selection of methods and models that can be used to predict and monitor energy utilisation
in ships. There is, however, no single model/method that can be applied generally to a vessel to increase
its energy efficiency. The shipping industry faces the challenge of reducing fuel consumption and air
emissions. There is a need to understand how much energy is needed and used by the entire energy system
of a ship but with a resolution that can analyse this usage at a subsystem or component level. This report
presents a state-of-the-art investigation of published models and methods within the research area of ship
energy efficiency. Emphasis is placed on the existing models and methods for energy systems modelling
(prediction, monitoring and improvements) and their applicability, strengths and weaknesses. The report
also presents a review of green ship energy concepts such as wind power as auxiliary propulsion in ships,
together with ship routing optimisation. The outcome of the study highlights two important issues. First,
there is a need to develop a generic holistic model that would be applicable for ship energy efficiency
analysis and simulations. The current state-of-the-art shows that parts of such a model exist, but they need
to be combined to interact on a common basis. Second, the potential of wind power to act as auxiliary
propulsion and ship routing optimisation can significantly reduce energy consumption and improve the
energy efficiency of ships. Several technical concepts have been evaluated and tested, and wingsails are a
solution that, together with ship routing algorithms, can significantly lower fuel consumption.
Contents
Summary ........................................................................................................................................................................ 1
1. Introduction .............................................................................................................................................................. 2
2. General description of a ship as a system model ...................................................................................................... 3
2.1 Division into components ................................................................................................................................ 4
2.2 Connection points and challenges .................................................................................................................... 7
3. Monitoring of energy consumption/ship performance monitoring ........................................................................... 8
4. Tools and models for energy efficiency prediction ................................................................................................ 11
4.1 Holistic models .............................................................................................................................................. 11
4.2 Subsystem models .......................................................................................................................................... 12
4.3 Commercial tools and software...................................................................................................................... 13
5. Energy-saving measures and wind power as auxiliary propulsion ......................................................................... 13
5.1 Measures and their barriers to improve energy efficiency in shipping .......................................................... 14
5.2 Wind power as auxiliary propulsion in ships ................................................................................................. 17
5.3 Feasibility study for wind assist propulsions ................................................................................................. 23
6. Ship routing systems for energy-efficient shipping ................................................................................................ 25
6.1 Basic capabilities of current routing systems ................................................................................................. 26
6.2 Optimisation algorithm in routing systems .................................................................................................... 28
6.3 Future routing concepts .................................................................................................................................. 31
7. Concluding remarks ............................................................................................................................................... 33
8. References .............................................................................................................................................................. 33
1 Report No. 15:153

ISSN: 1652-9189
1. Introduction
More than 90% of the total transport of goods in the world currently uses vessels, which are
unfortunately a significant source of emissions. These emissions account together for
approximately 3% of global CO2 emissions, 1-3% of SOx and more than 10% of NOx. The
emissions are expected to increase further by between 150 and 200% before 2050 due to
increased transport volumes. The European Commission’s strategic document “A strategy to
reduce atmospheric emissions from seagoing ships” underlines the importance of reducing
emissions from ships to improve human health and the environment. Additionally, the EU and
the Swedish national climate goals have the following objectives for the transport sector in 2020:
• a 20 percent improvement in energy efficiency,

• a 40 percent reduction in greenhouse gas emissions, and
• the use of at least 10% renewable energy.
To achieve these goals, sustainable sea transport with minimal environmental impact is required.
At the same time, because of the rapid increase in the price of oil, the fuel for certain types of
ships represents approximately 70% of the total transport cost. An increased awareness of the
environmental impact, the global economic crisis and high oil prices are the driving forces for the
development of more energy-efficient maritime transport with lower fuel consumption. This has
been particularly evident over the last decade, as several national and international industrial and
research projects have targeted energy-efficient shipping. These projects have studied, analysed
and evaluated conceptual solutions and proposals for measures that could lead to more energy-
efficient maritime transportation. A common denominator for the projects is that they are often
focused on and sub-optimise a solution or an area of the ship, hereafter referred to as subsystems,
without considering the fact that each subsystem must interact with the ship and its resulting total
energy system as a whole. This sub-optimisation makes it difficult to apply results of these
projects to other ship types and operating conditions than those used in the original research
projects. However, shipping needs general methods and models to implement the right measures
that can actually lead to credible energy saving measures. Consequently, there is a lack of a
holistic overview, methodology and generic model.
The increased research interest in “green shipping” in the maritime industry is also reflected in
the form of innovative ship design, operation optimisation, and smarter fuel-efficient engines.
Additionally, various green concepts using wind have been investigated within the Swedish
shipping industry, such as wingsails, kite-sails, Flettner rotors and wind turbines. A few years ago,
Marinvest AB and the Department of Shipping and Marine Technology at the Chalmers
University of Technology investigated the possibility of using wingsails. They found that, as for
other concepts, a major obstacle to the practical implementation was the lack of an integrated
system for optimisation and accurate estimation of energy savings and impact on the ship (e.g.,
stability conditions, fatigue). The lack of such a system entails high risk in both the economic
investment and the ship’s structural safety. Systems perspective along with reliability-based
analysis and optimisations can be performed in a better way.
2 Report No. 15:153

ISSN: 1652-9189
This report summarises the results from a state-of-the-art investigation of the application of and
research into ship energy efficiency, focusing on models and methods for energy systems
modelling and wind on ships as an alternative green concept. It also includes methods for ship
route planning and route optimisation as part of energy efficiency. The report is divided into
several chapters. Chapter 2 presents a general description of a ship as a system model, how it can
be divided into subsystems or components, and where the connection points are between
subsystems and challenges that follow from that. In Chapter 3, the results from a literature review
and state-of-the-art monitoring of energy consumption/ship performance monitoring are
presented. Different approaches are identified, and their advantages and limitations are discussed.
Groups primarily working in Europe on this topic have been identified and are presented in
connection to each approach. Chapter 4 continues to present and discuss existing tools and
numerical models for energy efficiency prediction. As part of this chapter, in-house developed
codes and commercial software packages are presented. Chapters 5 to 6 emphasise different
solutions and techniques for wind propulsion. In addition to technical solutions for wind
propulsion, routing tools and models that can maximise the benefit of using sails are presented
and discussed. Some examples of ship routing commercial software are presented. Finally, in
Chapter 7, concluding remarks from the study are outlined.
2. General description of a ship as a system model

Systems modelling is a technique that is used to visualise, analyse and potentially optimise
complex systems in a broad range of applications. Energy system models are widely used to
analyse the energy consumption and efficiency of energy systems in buildings or entire countries.
These models can provide deep insight into the processes and interactions between different parts
of a system or subsystems. The main difficulty with systems modelling is finding the right level
of abstraction to be able to capture the nature of the system while also describing it by means of
mathematical equations.
A system is a combination of equipment, machinery, external influences and processes that

together fulfil a mission, and can be decomposed in specific functions. A subsystem is a complex
group of machinery, equipment or processes that are used together to fulfil one certain function
of the full system. Merchant ships at sea can be considered such energy systems, with the general
mission to transport goods. This overall mission can be decomposed into specific functions
(Mermeris et al. 2011):
• platform functions (carrying cargo, safety),

• hotel functions (HVAC, accommodation),
• general support functions (cooling, fresh water supply, fuel supply), and
• mission specific functions (cargo handling).
In Grimmelius (2003), system models of physical systems are categorised using four dimensions:
• model level, i.e., the level of detail in the description of the process,
• model time domain, i.e., the time domain in which the model is developed,
• application time domain, i.e., the time domain in which the model is applied, and
• model data, i.e., the amount and character of required data to produce useful output.
3 Report No. 15:153

ISSN: 1652-9189
In the first dimension, model level, three approaches can be identified: black box, grey box and
white box models (Grimmelius 2003). Though the last is fully based on physical principles and
might result in complex differential equations and slower models in terms of computational time,
a black box model describes a pure input/output relationship. A grey box model is a combination
of transfer functions similar to the black box and white box methods. A considerably deeper
understanding of the natural process is required if a white box model is to be applied, but the
simulation will at the same time deliver more valuable results and a deeper insight into the
process (Grimmelius 2003).
Model time domains can be divided into steady-state and time-dependent domains, and the latter
can be further divided into time domain behaviour or frequency domain behaviour. Steady state
models are the least expensive in terms of computational effort but are unable to represent
dynamic processes where a time dependent model is needed. Of great importance for the required
performance of the model is the application time. If real-time use is required, the computational
effort must be minimised and well specified. If real-time use is not required, only practical
limitations should be kept in mind. For example, if the model should be used for a huge number
of simulations within an optimisation loop, the computational time should not exceed a specified
maximum for practical reasons. Finally, the available or required input to the model is of
importance for model development. In Grimmelius (2003), four categories are specified:
• models requiring limited measurements,

• models requiring limited system parameters,
• models requiring mainly parameters, and
• models requiring mainly extensive system parameters.
When applied to a ship’s energy system model, several approaches must be considered. A
simulation model of an engine or engine component might require a time-based simulation,
whereas a prediction of the ship’s propulsive power requirement can be performed with a steady
state simulation if only one operational point is of interest. For the model input, the requirements
are different for each subsystem. An engine simulation for example requires a considerable
amount of system parameters but can be completed with only a few measurements, whereas the
propulsive power prediction requires much more measurements of the ship’s condition and the
weather. Altogether, a holistic ship energy model will fall into several categories and thus be
much more complex than a subsystem model. The complete system must therefore be divided
into its components to define the structure of the entire system model.
2.1 Division into components

A general energy system can be divided into three main components (Baldi 2013, Osses et al.
2014): supply, conversion and consumption/demand. Additionally, a component for waste can be
inserted, which becomes useful, particularly with the focus of a ship’s waste heat recovery and
cooling systems. An overview of a generic ship energy system is given in Figure 1.
4 Report No. 15:153

ISSN: 1652-9189
Figure 1: Overview of a ship energy system.
On the supply side, usually wind is not part of the efficiency evaluation because it is difficult to
measure how much of the available wind energy is actually used by sails. Additionally, the
energy available from wind is considered available without costs other than those for the sails and
equipment. Thus, the efficiency can be evaluated using the difference between the theoretical
energy available in the fuel on board and the energy that was produced from this fuel. In shipping,
efficiency is often considered the quotient of used energy (fuel) and transported cargo, which
makes the ship’s light weight part of the energy system. Using this formula, a lighter ship with
the same cargo capacity will certainly have a better efficiency, assuming that the hull form and
ship systems are of the same quality.
The components demanding electrical power or steam are rather easy to capture because they
physically exist in the engine, cargo or accommodation compartments of the ship. For an existing
ship, the energy consumption of these components can be measured, and those measurements can
be used for model development. A generic model of these components, conversely, requires
many system parameters and assumptions or special models for environmental parameters.
The total demand for propulsion is also measured rather easily for an existing ship because the
propeller shaft power can be measured by means of revolutions and torque. Simulating and
predicting the propulsion power demand is, however, complicated. In fact, the propulsion of a
ship is a complete and complex subsystem. A ship’s calm water resistance can be divided into
viscous form, frictional and wave making components. Additionally, added resistance in waves,
due to wind, due to steering and due to added roughness will appear under service conditions,
according to Kristensen and Lützen (2012).
Figure 2 provides an overview of the components of a ship as an energy system, including the
links to the environment. The energy input and output is illustrated for each component. Of
course, not all of the components exist in every ship’s energy system.
5 Report No. 15:153

ISSN: 1652-9189
Figure 2: Overview of system components and influences to the energy system.
6 Report No. 15:153

ISSN: 1652-9189
2.2 Connection points and challenges
In a ship as an energy system, three main connection points can be identified: the propeller-hull
interaction, the propeller-main engine interaction and the response of the ship in environmental
conditions. The latter can be further divided into the response of the different components. The
resistance of the ship is primarily dependent on the wind, waves and water temperature. Further
components such as the HVAC system or the cargo handling (in particular heating and cooling)
are dependent on the air temperature and wind chill effects. Additional, secondary connections
can be identified, such as the response of the generators (especially in connection with shaft
generators), waste heat recovery systems and boilers on the main engine load.
The propeller-hull interaction is often referred to as the propulsive efficiency (Kristensen and
Lützen 2012). It is defined as the difference between the effective power (resistance times speed)
and the power that can be measured at the propeller shaft. For single propeller ships, the
propulsive efficiency is approximately 0.6-0.7, and twin skeg ships can have a propulsive
efficiency up to 0.85. To understand the reason for these losses, the total efficiency can be
divided into six components (Dyne 1995) as shown in Figure 2 and listed below:
• axial losses,
• rotational losses,
• axial gains,
• rotational gains,
• losses due to infinite number of blades of the propeller, and
• frictional losses.
To define these components, one must analyse the wake pattern of the hull with a working
propeller just before and behind the propeller (Dyne 1995). Axial and rotational gains appear if
the propeller fills in a wake or recovers a rotation that exists in front of the propeller. Axial and
rotational losses occur when rotations or axial velocities that are higher than the ship’s speed
appear in the slipstream of the propeller. Theoretically, a ship can have a propulsive efficiency
higher than one, assuming that the propeller can fill the entire wake and recover the existing
rotation to be zero in the slipstream. In practice, however, this is not the case. Energy-saving
devices (ESDs) often work by creating a rotation in front of the propeller or by neutralising the
rotation in the slipstream to increase rotational gain or decrease rotational loss. Wake equalising
ducts aim to decrease the axial losses (accelerating ducts) or increase the axial gains (decelerating
ducts); see Dyne (1995). ESDs or specially designed hull forms that should increase the wake in
the propeller plane cause a higher propeller load and might lead to higher rotational losses.
Unfortunately, the wake pattern of a hull with a working propeller is not easy to predict. To run a
full analysis using the above theory, full viscous CFD computations with a working propeller
have to be performed.
In model tests or theoretic predictions, the propulsive efficiency is determined by using either
approximations (in theoretic predictions) or tests without a working propeller to determine the
resistance of the hull. Later, tests with a working propeller can be used to measure propeller
power. According to this method, the propulsive efficiency is divided into the propeller open
water efficiency, the relative rotational efficiency (the difference between the open water
efficiency and the efficiency in behind conditions) and the hull efficiency (ITTC 1999). A clear
7 Report No. 15:153

ISSN: 1652-9189
explanation of which losses and gains are captured for efficiency in this method can
unfortunately not be given. Because the latter method is used in model test predictions, many
statistical data are available. Thus, this method is often used for theoretical predictions.
In conclusion, the propeller-hull interaction is challenging because the propulsive efficiency

depends not only on the hull form and propeller design but also on the operational condition, i.e.,
the propeller load. The propeller load is often defined using the advance ratio J = va/nD, where va
is the advance speed of the propeller, D is the diameter and n is the revolution (rpm). For a fixed
pitch propeller (FPP), the torque and thrust are directly coupled to the rpm. Additionally, the
velocity of advance is not equal to the ship speed due to hull form, propeller design and propeller
load specific efficient wake (we) of the hull.
For calm water conditions and clean hull and propeller (FPP), the advance ratio of the propeller is
almost constant over a wide speed range. Due to added resistance caused by wind, waves or
fouling, the advance ratio can be considerably reduced because the speed of advance is close to
constant but the revolutions of the propeller must be increased due to the higher demand of thrust
and torque. Additionally, propeller fouling causes higher torque values at the same rpm and
similar thrust. These points together lead not only to a higher propeller load but also to
difficulties concerning the engine-propeller relation because the engine has upper limits for
deliverable torque at a certain rpm and for the rpm itself. Thus, an involuntary speed reduction
caused by one of these effects must be accounted for when modelling the energy system. In
conclusion, finding the equilibrium of the propulsion system of the ship, which consists of
resistance, propulsive efficiency, propeller and engine characteristics, is a major challenge.
3. Monitoring of energy consumption/ship performance monitoring

Ship performance monitoring describes the process of measuring the ship’s actual energy
consumption and predicting the theoretical consumption considering the environmental
conditions. Tools, equipment and methods for monitoring the performance of ships have existed
for a long time, as reported by Crane (1982) and Tosi et al. (1982). However, the wide
application of such systems was first driven by the rising fuel costs and the declining market
situation during the past decade. The purpose of such analysis and measurements is monitoring of
the vessel condition, particularly hull and propeller fouling. It is also important for the control of
the ship’s operational state, which offers the opportunity of operational optimisation, especially
with regard to speed and draft/trim (Hansen 2011). In addition to performance monitoring with
regard to the ship’s resistance and propulsion performance, monitoring the ship’s energy system,
especially the main engine, is possible and was developed for steam plants in the 1980s (Crane
1982).
Monitoring of the energy consumption of the ship can be performed using mass flow meters
measuring the fuel consumption of the main and auxiliary engines. Due to earlier discussed
environmental influences on the ship’s hydrodynamic resistance and power consumption, such
measurements will not give any indication on how well the ship is operated or the conditions
under which the systems, the hull and the propeller exist. To analyse the performance of the ship,
a power prediction model must be used, as discussed and presented by Hansen (2011) and
Andersen et al. (2005). For this approach, multiple measurements are needed, including the ship
speed, the fuel consumption, the loading condition of the ship (draft and trim) and environmental
8 Report No. 15:153

ISSN: 1652-9189
conditions regarding wind and waves. Using the measured environmental data and loading
conditions, a prediction of the fuel consumption can be performed using the theoretical model.
The predicted consumption can then be compared with the actual consumption, and the
performance of the ship can be evaluated (Hansen 2011).
Difficulties can be identified at two points (Hansen 2011). The first point is data logging and
acquisition, and the second is the prediction of added resistance due to waves and loading
conditions. In Hansen (2011), manual data sampling proved to introduce many uncertainties and
possible errors, and Caprino et al. (1993) discussed the automatic logging of available parameters.
Automatic data logging can often be performed using the ship’s existing systems (Hansen 2011)
and customised time steps. The latter provides the opportunity to adjust the amount of data by
determining which of the measured parameters are actually time-dependent. These time-
dependent parameters can then be sampled with shorter intervals. However, only data that can be
measured on board can be collected using the ship’s systems. One main problem thus remains,
which is the measuring of wave height and frequency. In general, five options are available:
1. Manual estimation of the wave height by deck officers. This is most likely the most uncertain
option.
2. Estimation of the wave height using the wind speed, geographical position and statistical wave
data. This option requires detailed analysis of the geographical position and weather in close
areas and might only work well for a certain amount of time.
3. Estimation of the wave height and period using on board measurements of the motions of the
ship. This option was used in Hansen (2011).
4. Measurement of the wave height using on board wave radars. This option will certainly give
adequate results but requires additional equipment on board the ships.
5. Measurement of the wave height using radar pictures from the ship’s radar system. A 3D
Fourier analysis of the picture from a conventional radar systems was developed by Young et
al. (1985), but an application of such a system could not be found.
Once the data are collected, the second problem is the prediction of the added resistance or the
prediction of the baseline performance of the new hull in the present condition. The resistance of
a ship in service can be divided by the calm water resistance at a certain loading condition for
which resistance data can be obtained from model tests. A theoretical prediction or computation
is available, and numerous added resistances due to the ship’s condition and the environment are
listed below (ISO 2002):
• calm water resistance at even keel and known loading condition,

• additional resistance due to trim and loading condition,
• additional resistance due to salinity, density and water temperature,
• additional resistance due to steering,
• additional resistance due to wind,
• additional resistance due to waves, and
• additional resistance due to shallow water.
Additionally, effects from current and drift must be considered if the ship’s speed is not measured
as pure longitudinal speed through the water. In the ISO guidelines (ISO 2002) and the ITTC
9 Report No. 15:153

ISSN: 1652-9189
proceedings (ITTC 2002), the necessary measurement data and methods to evaluate the above-
listed effects are presented. Most of the effects are captured by analytical formulas, except for the
additional resistance due to the loading condition and trim as well as the additional resistance due
to waves.
The additional resistance due to draft and trim is highly dependent on the hull form, particularly
the bow region, and does not have linear relationship with draft or trim. Especially for ships with
a large bulbous bow, the residual (wave making) resistance is influenced by the draft at the
forward perpendicular, as shown in Larsen et al. (2012), where a variation of the propulsive
efficiency was found in addition to the variation in resistance over draft and trim of a large cargo
vessel. The non-linear behaviour of the power variation was shown in Liu et al. (2011). It has
been suggested that viscous CFD computations or model tests should be performed to capture
these effects.
Added resistance due to waves and ship motion is difficult to evaluate with a high level of
accuracy. Early strip theory methods were developed in the 1970s (Boese 1970) and can
primarily be used for early design evaluation when quick responses are needed. In Journée (1992),
it was found that the results were satisfactory for conventional, slender mono hull ships. However,
due to the nature of the strip theory, the results would not be satisfactory for ships with large
bulbous bows, very full block ships or if the wave encounter frequency is very high (Pérez 2006).
More satisfactory results can be obtained using 3-dimensional, non-linear methods, as discussed
by Fang et al. (2006), including Maruo’s far field method, as shown by Liu et al. (2011) and
proposed in ISO (2002). The implementation and validation of a nonlinear boundary element
method for ship motions is discussed by Kjellberg (2013). The method presented by Liu et al.
(2011) was able to provide a satisfactory evaluation of the added resistance for different wave
lengths for both, slender and full ships. Additionally, the reported computational time was very
short, which offers the opportunity to use such a method to compute a large number of cases.
This would be necessary if applied for the performance evaluation of ships in seaway.
Aside from computational methods, model tests in regular waves will still give the most reliable
results and are widely used to validate the computational methods. However, extensive model
test results, including trim variation tests, were available for the study undertaken by Hansen
(2011). The resulting performance indices or estimated roughness of the hull showed a
considerable large spread over time. This might be due to the rather simple approach that was
used to estimate the added resistance due to waves, which is an empirical formula based on the
wave height, ship’s breadth, water line length and the block coefficient (ITTC 2005, Hansen
2011). Thus, no influence of the hull form, especially of the forebody form, was analysed. This
spread shows the very high importance of an accurate estimation of the added resistance due to
seaway.
The result of performance analyses is often shown as a “slip”, a “performance index” or the
theoretical roughness of the hull at a certain time (Hansen 2011). The slip or the performance
index is the difference in predicted power and measured power, whereas the hull roughness is
obtained by assuming that the difference in the predicted and measured power is caused purely by
added frictional resistance.
10 Report No. 15:153

ISSN: 1652-9189
4. Tools and models for energy efficiency prediction
Numerous tools and methods for energy and performance prediction of ships are available both
commercially and academically. These will be divided into two groups, i.e., holistic and
subsystem models, and discussed in the following chapters. Commercially available tools and
software will be mentioned in Chapter 4.3.
4.1 Holistic models

As of today, no commercial tools are available for the holistic prediction of the energy
consumption of a ship. Holistic means that all parts, including the propulsion, resistance, engine,
hoteling, cargo handling and environmental influences, are taken into account. The mandatory
components of such a holistic model were discussed above and can be found in Figure 2.
Furthermore, there have been two holistic models reported from different universities:
• the entire ship model and ship impact model reported from University College London
(Calleya 2014, Calleya et al. 2014), and
• the dynamic energy model from University of Strathclyde (Mermeris et al. 2011).
The first model from University College London is based on a method to apply rather small
changes to the baseline ship from which the energy systems and behaviour is known. The method
is divided into two levels of detail: the ship impact model and the whole ship model (Calleya et al.
2014). Whereas the ship impact model provides a high level view of the ship and its impact on
the fleet, the whole ship model provides a more detailed view on the changes of the single ship
itself. The model is built using Paramarine and Matlab (Calleya et al. 2014). The hydrodynamic
properties of the models ships are obtained using the Holtrop-Mennen method (Holtrop et al.
1982). Due to the nature of this method, the ship model cannot give any information about the
engine load, propeller-engine or propeller-hull interaction. Thus, this model is more suitable in an
early stage of the design process when the level of details is low compared with other holistic or
subsystem models.
The dynamic energy model from the University of Strathclyde is assembled from numerous sub
system models using Matlab/Simulink (Mermeris et al. 2011). As the name indicates, this model
is time dependent and can thus provide information regarding changing energy consumption over
time. It is not clearly described by Mermeris et al. (2011) whether the model is reversible, i.e., if
the user can, at the same time, provide the ship’s speed as input to the propulsion system and the
next time the main engine power. Mermeris et al. (2011) also mentioned that the resistance and
propeller-hull interaction were not yet integrated. According to Marzi (2014) the integration of
the hydrodynamic components is done by means of CFD computation and response surface
technologies for the hull resistance and an artificial neural network (ANN) for propeller
properties. One drawback of using Simulink for such models can be attributed to the fact that
Simulink models information flow and not a real physical model. Thus, the model becomes more
complicated and more difficult to work with. Overall, this model seems more flexible and
detailed than the first mentioned model, but it is not clear how it can be used when limited
information is available, such as with early ship design.

ISSN: 1652-9189
4.2 Subsystem models
A subsystem model is a model of only one part of the entire ship. This could be the main engine,
the hull resistance or a combination of some parts. The different types of subsystem models
existing are discussed in the following items.
Models for resistance and propulsion

One can classify most performance analysis and monitoring tools and models as subsystems for
resistance and propulsion. This is true if the main engine parameters are not taken into account,
with the exception of the main engine power. Most of the common commercial tools are models
of this type. These tools are discussed in Chapter 4.3. In addition, numerous power prediction
methods are available that can be considered models for resistance and propulsion. Methods for
the calm water resistance are described by Kristensen et al. (2012) and Holtrop et al. (1982). A
simple method for the added resistance in seaway is described in ITTC (2005), but both should
be classified as more theoretical/empirical methods than as energy models.
Main engine models

There are numerous main engine models found in the literature, and most of them use slightly
different approaches and are more or less flexible. Most of the more detailed models are focused
on either 2- or 4-stroke engines. An example for a theoretical engine model is reported by Osses
et al. (2014). In such a model, the shaft load is simply set by an assumption and no hydrodynamic
properties or loads from shaft generators are taken into account. A different approach of using
CFD to model the diesel engine is reported by Jin et al. (2013). Such a model can give accurate
results for certain operational points but may be too time-consuming to be included in a holistic
energy model.
Main engine and propulsion models

These models capture the main engine performance coupled to the hydrodynamic properties of
the hull and are therefore very well suited for more advanced performance analysis, as discussed
and presented by Hansen (2011). In Schulten (2005), a coupled main engine and propulsion
model is used to predict the power consumption and engine behaviour of a ship during
manoeuvring. The models used at the University of Athens (Papalambrou et al. 2012) and
Newcastle University (Tian et al. 2012) follow the same strategy of considering only the main
engine and the hydrodynamics. The approach to leave out all other on-board systems apart from
the main engine seems suitable if the main interest is the performance of the ship in terms of
speed or manoeuvring but lacks information if the total fuel consumption or emission is of
interest.
On board energy systems models

The models in this category focus on all on board energy systems, such as main and auxiliary
engines, boilers, hydraulics, tank heating, HVAC, fans and pumps, but neglect the
hydrodynamics of the ship’s hull. Normally, the propeller load is set as constant or roughly
estimated using approximation methods such as in Holtrop et al. (1982). Such a model is
discussed by Baldi (2013) and is available in some commercial tools as discussed in Chapter 4.3.

ISSN: 1652-9189
4.3 Commercial tools and software
Numerous commercial systems are available for performance monitoring and analysis, as
summarised in Table 1. Although the input and the results of such systems are rather similar
(except for the Kongsberg system, which is more engine focused), the methods are rather
different. In general, there are some black box systems such as the Marorka system, and some
white/grey box systems such as the Marin system. All of these are only focused on parts of the
whole, complex energy system “ship”, and most of them on the hull resistance and main engine
performance.
From the literature, the most holistic system for the on-board energy systems seems to be the
DNV Cossmos (DNV 2012), but it does not include the ship’s hydrodynamics.
Table 1: Examples of companies selling commercial software for performance monitoring.

Company name Website
NAPA www.napa.fi
Marorka www.marorka.com
Marin www.marin.nl
Kongsberg www.kongsberg.com
Insatech www.insatechmarine.com
Seaweb www.sea-web.com
Amarcon www.amarcon.com
Weather Routing Inc. www.wriwx.com
Kyma www.kyma.no
5. Energy-saving measures and wind power as auxiliary propulsion

In current shipping market, the bunker price has increased more than 6 times since 2000. It is
likely to further increase to 1000 USD/ton over the service life of today’s newly built ships.
Improving energy efficiency to reduce fuel consumption is becoming one of the most important
topics. In addition to the increasing pressure on high bunker prices, its main driving forces also
result from stricter regulations on energy performance and air emissions from ships, and
ambitious goals to strength market positions, trade flexibly, branding and attract external
stakeholders, who put more weight on energy efficiency to make their investment decisions. It is
known that shipping is a substantial emitter, accounting for approximately 3% of the total global
CO2, 1-3% of SOx and more than 10% of NOx (Buhaug et al. 2009). Emissions from shipping are
expected to increase 150-200% by 2050 due to the growth of the maritime transport sector (EC
2011). The effect of air emissions of NOx and SOx from shipping on environmental and health
impacts has also been realised to a large extent (Corbett et al. 2007, Eyring et al. 2010). The
International Maritime Organization (IMO) has established regulations to reduce fuel
consumption and air emission (EEDI, EEOI, SEEMP, ECA); see IMO (2012, 2013a, 2013b). The
inclusion of emission tax in fuel price is also likely to be established in future shipping market
(DNV 2011). According to the future fuel cost scenarios by Eide et al. (2011) and DNV GL
(2014) for the reduction of CO2 and SOx emissions, the added premiums may push the price to
2000 USD/ton by 2040.
The increased awareness of environmental impact, global economic crisis and high oil price acts
as the catalyst to stricter regulation requirements (DNV GL 2014) to push shipping toward more

ISSN: 1652-9189
energy-efficient transport. Currently, research interests have been focused on the maritime
industry to seek energy-efficient measures in terms of innovative ship design, optimal operation,
smarter propulsion, and efficient machinery. In the following Chapter 5.1, general pros/cons of
those realised measures are addressed, as are the barriers that hinder their practical
implementations. In particular, this report focuses on the status of green shipping concepts of
utilising wind power as extra propulsion, as summarised in Chapter 5.2. Furthermore, Chapter 5.3
narrows down shipping concepts, research progress and technical matureness with respect to the
implementation of energy saving devices to provide wind assist propulsion for ships.
5.1 Measures and their barriers to improve energy efficiency in shipping

Different stakeholders and policy makers in the maritime industry have identified a vast range of
measures, which aim at reducing the fuel consumption and air emissions in shipping. A summary
of these measures and their potential savings are listed in Table 2. These are based on extensive
investigation of various industry projects and research activities, e.g., ABS (2013), Buhaug et al.
(2009), DNV GL (2014), Faber et al. (2011) and Wärtsilä (2010). It should be noted that the
percentages of potential energy savings by implementing these measures can differ significantly
depending on the analysis source. The colours in the table are to give a basic reference and
relative savings potential in comparison with different measures.
In Table 2, the ability to implement each measure is categorised into new building and existing
vessels. Due to significant differences in the specific characteristics of each ship type, some
measures may be limited to certain groups of ships. For example, the utilisation of lightweight
material for ship construction may be only beneficial to ships with larger superstructures but have
limited application for conventional ships such as bulk carriers and tankers (Faber et al. 2009).
Furthermore, some measures are mutually exclusive. A simple summation of all measures will
greatly overestimate the energy-efficiency improvement potentials of employing multiple
measures for a given ship, as noted by Johnson et al. (2014), Kesicki and Strachan (2011), and
Kesicki and Ekin (2012).
As shown in Table 2, a large energy savings potential is available through the implementation of
a combination of different measures for both newly built and existing ships, according to the
extensive market survey carried out by DNV GL (2014). Shipping companies tend to apply well-
known and mature measures that require little investment and have easily accessed benefits such
as slow-steaming, weather routing and optimised voyage planning. The biggest incentive of
implementing energy efficient measures is to be in compliance with regulation requirements
rather than market-focused goals, such as strengthening their market position and economic
profits. Most such measures rely on service providers’ or manufacturers’ data and analysis, in
which there is often a lack of transparency leading to significant uncertainties of possible energy
savings. One of the most significant challenges in undertaking these tasks is the lack of data. To
develop robust and accurate models, detailed ship cost effectiveness data and
technical/operational data are needed.
In addition to non-transparent data and models for energy-efficient assessment, other barriers
hinder the implementation of energy-saving measures in the shipping market. These were
categorised by Jafarzadeh et al. (2014) and were discovered based on an extensive market survey
in Norway and international research activities by Cagno et al. (2013), Faber et al. (2009, 2011)

ISSN: 1652-9189
and Fleiter et al. (2011). In general, these barriers can be grouped into five levels: information
barriers, economic barriers, (intra and inter) organisational barriers, technical barriers and policy
barriers.
Table 2: Available energy-saving measures and concepts with potential savings and main
barriers for their practical implementation.
Fuel saving
Concept categorizes Detailed measures Barries
Potentials *
Main dimensions optimisation, e.g. slender and
Policy and regulation
larger ships, and less ballast water volumn
Ship energy system optimisation based on actural
planned operaitonal profiles, fuel price, trade Information and organiation
Hull form routes, loading conditions etc.
optimisation and Optimisation of stern bigle , rudder and propeller
New Information, economic,
propeller for inflow considerations (e.g. better skeg design,
organizational and technical
building configuration better propeller balde design)
vessels Optimisation of hull shape and bulbous bow Information and technical
Better design with consideration of added
Technical and economic
resistance due to wind and waves in open sea
Influence from IMO EEDI on ship design Policy and organization
High strength steel Technical
light weight
construction Composite material Technical and economic
Waste heat recovery system
Alternative fuel for power
Machineary Economic, organisation and
Hybrid auxiliary power generation
system information
Optimum heating and cooling systems
Adaptive pump and power management systems
Propulsion improving devices Technical and information
Energy saving Skin friction reduction Technical and economic
devices Wind power and other renewable devices for Technical, economic and
Existing auxiliary propulsion organization
vessels Ship speed reduction Organisation
Optimised ship maintenance schedule, hull and
Information and economic
propeller cleaning
Ship operation Autopilot adjustment Technical
manamgement Intelligent engine adjustment according to weather
Technical and economic
and optimisation and loading conditions
Trim, ballast and rudder control optimisation, air
Technical and information
lubriation etc.
Weather routing and voyage optimisation Information and technical
* The potentials of energy savings can differ significantly from individual ships, even by implementing the same
measure. The percentages here, rather, give an average savings potential indication based on general ship types.
The potentials are divided into the following four levels:
Fuel saving up to 2% 5% 10% ≥20%
Marker
For most newly built vessels, current ship designs provided by shipyards have already been
optimised to have the best operation performance. New changes and arrangements to a specific
ship are often negotiated with shipyards possible with a high construction cost. However, the
energy savings due to the installation of these measures on new building ships will eventually
benefit most ship chartering companies, which are not responsible for the payment of the extra
installations. There obviously exist some conflicts of interest between different stakeholders of
the shipping industry. This is referred to as inter-organisational barriers for the implementation of
energy efficient measures. Even within the same company, upper management may have
different motivations from technical advisors to update their conventional designs to more

ISSN: 1652-9189
energy-efficient ship concepts. This is often associated with risk of the actual energy gains
through the new implementations. In some cases, the organisational barriers are also combined or
due to the technical maturity of certain measures and methodologies. For example, it is well
recognised that a ship’s optimisation design should be carried out for its actual operational
conditions and future service weather environments, whereas construction contracts between
shipyards and ship owners normally refer to calm water conditions with a specific service speed.
All of the subsystems are therefore arranged and optimised based on requirements for the ship
operated in calm waters, whereas ships rarely sail under such conditions. The reason behind this
is that the mature information and technology for such conditions are not available to judge the
success of the ship construction project. Another more obvious example is the utilisation of high
tensile steel (HTS) for ship construction. This could help to reduce a ship’s plate thickness, thus
leading to a significant total weight deduction. However, the technical problems associated with
the use of HTS such as fatigue and welding issues also limit its wide usage in ship construction.
The same problem is also faced for the use of composite materials, which are also strongly
limited due to cost and technical immaturity.
For currently existing ships, many measures are available in the market aimed at increasing
energy efficiency and reducing air emissions. The implementation of these measures can be
categorised into machinery systems, energy-saving devices and ship operation management and
optimisations, as shown in Table 2. The services of implementing these measures within
machinery systems are often provided by marine engine companies. The main barriers for such
implementations are often related to economic concerns from the ship owners’ perspective.
Organisational barriers occur when the shipping company has different priorities than the
measures’ implementation. Information barriers are essential for the application of such measures
because the data and analysis provided by engine companies often lack transparency. The
consequence is that larger percentages of energy savings are claims but not actually observed in
the market. These information barriers can also be found in retrofitting energy saving devices,
such as installing a duct to harmonise and stabilise the flow distribution entering into the
propeller to increase propulsion efficiency or the re-configuration of propellers. The lack of
information can also be induced by the lack of technical solutions to prove the efficiency of
energy saving devices. Most often, these energy saving devices need large amounts of
investments for their implementations. The technical maturity of these measures should be further
developed to ensure the confidence to estimate their investment and payback time. One example
is the utilisation of wind assist propulsion for ships. Although large potentials can be expected
from such measures, their arrangement on board ships, installations, maintenance and impacts on
a ship’s overall safety, manoeuvrability, and optimum utilisation of wind forces actually require a
complex system with different levels of models and technologies to support the entire concept. In
addition to retrofitting or updating subsystems in the existing ships, optimal ship operation
management, such as slow steaming and optimisation voyage plans, are cheaper, straightforward,
efficient and highly implemented in the shipping industry. Due to the advancement of
information and communication technology, the increased reliability of weather forecast provided
by metrological institutes and large established database on long-term marine weather
environments, higher energy savings from an optimal ship operation can be realised through the
integration of an voyage optimisation system with some energy saving measures, such as wind
assist devices and propulsion retrofitting etc.

ISSN: 1652-9189
5.2 Wind power as auxiliary propulsion in ships
The EU Commission’s Communication on a “Strategy to reduce atmospheric emissions from
seagoing ships” indicated the importance of reducing these emissions from ships for the
improvement of human health and environment. The EU and Swedish national climate goal in
the transport sector includes 20% improvement in energy efficiency, 40% reduction of
greenhouse gas emissions and a minimum of 10% renewable energy by the year 2020. It also
requires sustainable maritime transport with minimal environmental impact. Further motivated by
the high bunker fuel cost, many research activities have focused on the concepts of wind energy
utilisation as auxiliary propulsion in ships, which was extensively investigated in the 1980s due
to the 1979 oil crisis (Letcher 1982, Schenzle 1985). In Figure 3, some of the most commonly
discussed wind assist ship concepts are presented. Different refinements of these concepts and
their combinations were also proposed. The state-of-the-art research and feasibility of these wind
assistance ship concepts during 1980s (1979-1985) were extensively reviewed by, e.g., Bergeson
and Greenwald (1985) and Nance (1985).
Figure 3: Different green shipping concepts using wind power as auxiliary propulsion.
The Flettner rotor ship concept uses the Magnus effect (Magnus 1853) to generate a propulsive
force from the airﬂow around rotating cylinders. As shown in Figure 4, it is assumed that a ship is
heading to the west and that the wind is blowing to the north. To generate propulsion force to
push the ship toward the west, the cylinder must be rotated clockwise. The rotating cylinder will
cause a fractional drag force on the surrounding airflows, causing the speed of the airflow in front
of the cylinder to be greater than that behind. Based on the Bernoulli equation, this can cause
lower air pressure in front of the cylinder and generate a lift force on the cylinder to push the ship
forward.

ISSN: 1652-9189
Figure 4: A sketch of the Magnus effect for a Flettner rotor ship with a vertical rotating cylinder.
The first system to prove that the Magnus effect concept could provide auxiliary propulsion to
ships was designed by the German engineer Anton Flettner (Flettner 1925). The demonstration
project was conducted on a schooner (the Buckau) that was outfitted with cylindrically shaped,
rotating masts that were 15 meters tall and 3 meters in diameter. It crossed the Baltic Sea and
North Sea from Gdansk to Scotland in February 1925. Over the past century, the Flettner rotor
ship concept never took hold in the commercial shipping market due to its non-competitive
capabilities in comparison with steam and diesel engines as ship propulsions. However, the
concepts were briefly revisited in the 1980s by maritime academia and research institutions
driven by the concerns of the oil crisis (Nance 1985, Bergeson and Greenwald 1985). Flettner’s
cylindrical rotors were further revised by Thom (1934) with discs distributed along the cylinder.
The performance of cylinder rotors with and without discs for ship auxiliary propulsion was
experimentally studied by Clayton (1985). However, the distributed disc cylinder concept was
abandoned by the inventor Thom because the power required to spin the rotor with large discs
became too expensive.
Currently, due to high bunker prices and correspondingly increased market pressure toward a
more energy efficient merchant shipping industry and environmental concerns through the
international and national regulations, Flettner’s concept is now under the spotlight of the entire
maritime community as an important alternative to using wind power as auxiliary propulsion to
ships. The research activities on Flettner’s concepts mainly focus on the computation of lift and
drag forces generated from the cylinder rotor because a high Reynolds numbers airflow makes
the reliable computation a bit challenging (e.g., Aoki and Ito 2001, Karabelas et al. 2012, Craft et
al. 2014). In the industry market, dozens of patents have been granted to support and utilise
Flettner rotor concept related technologies and subsystems during the past few years.
Due to the large weight and high central gravity of the rotors, a ship’s structural integrity and the
vessel’s hydrostatic and seakeeping capability should be carefully planned and analysed to ensure
its safety. After 10 years of technical development, the first modern vessel, i.e., the 10,000 dwt
cargo vessel E-Ship 1 (ENERCON 2013), was completed and launched in 2010. It was equipped
with four Flettner rotors that were 27 m height and 4 m in diameter, demonstrating the design
feasibility of the technology. As shown in Figure 4, Flettner rotors are controlled through a single
parameter, i.e., the rotational speed.
The concept of utilising wind force from towing kites to provide propulsion forces to ships has
been systematically developed since the 1980s (Duckworth 1985, Wellicome and Wilkinson

ISSN: 1652-9189
1984). Similar to other wind assist ship concepts, the incentive for the development of towing
kite propulsions is originated from financial reasons and environmental concerns of bunker fuel
consumption in shipping. Schlaak et al. (2009) investigated the worldwide shipping market for
the implementation of wind towing kites as auxiliary ship propulsions. Based on similar results
from field trials with a kite into a study, they concluded that kite systems could be retrofitted to
approximately 60,000 of a total of 90,000 ships. Significant fuel saving potential has been
presented for a multi-purpose freighter, estimated between 1% and 21% for a ship speed of 15
knots and between 4% and 36% for a ship speed of 13 knots.
Before the practical installation of kite systems in ships, two technical problems must be solved.
The first is a reliable prediction of mechanical performance of auxiliary wind propulsion using
kites. The second is the kite control/operation system, as well as the effect of ship structure and
stability due to installation and optimum ship route plan system, which is required for all wind
assist propulsion concepts. Mechanical models for kite performance prediction have been
researched and developed by e.g., Argatov et al. (2009), Lloyd (1980), Naaijen et al. (2006),
Wellicome and Wilkinson (1984), and Williams (2006). The proposed zero mass theory
(Wellicome and Wilkinson 1984) has even been validated by published experimental data. The
theory is based on the assumption that the aerodynamic forces are larger than the kite mass,
allowing for acceleration effects to be neglected. Therefore, the entire kite system can be
considered to be in an equilibrium flight state. Dadd et al. (2010) also used the zero mass kite
manoeuvring theory to predict kite line tension and other parameters connected with a kite’s
performance. Good agreement has been found between the computed results and a real test
dataset recorded using a purpose-specific kite dynamometer. In addition, the kite operation
system for auxiliary ship propulsion has been investigated. For example, Grosan and Dinu (2010)
studied the influence of a ship’s stability and manoeuvrability with zero forward speed under
certain wind conditions, Wrage (2007) patented his technical achievement for the kite launch and
recovery systems, and Erhard and Struch (2012) presented their research efforts on an autopilot
kite control. These technical developments have enabled kite systems to be installed
commercially for trans-oceanic voyages. These have become the most popular wind assist ship
propulsion concept in the current shipping market due to its technical maturity. The industry
partner SkySails GmbH (Skysail 2014) is the market and technology leader in the field of marine
towing kite systems. SkySails kites are key technology for capturing the vast potential of high-
altitude winds, and SkySails is the first company in the world that has succeeded in developing
towing-kite technology into an industrial application. Figure 5 presents a prototype of the
SkySails towing kite implemented in an oceangoing commercial vessel. The latest product
generation developed by the company can replace up to 2 MW of the main engine's propulsion
power. Another successful market player on the kite assist ship propulsion uses a system of
different size and shape of kites developed by KiteShip (Kiteship 2014), which has a successful
track record of using kites to provide propulsion force to yachts. The company has also built the
world’s largest sailing kite in 1997 and recently applied their systems to power ocean going
tankers with auxiliary energy.

ISSN: 1652-9189
Figure 5: (Left) A SkySails towing kite implemented in oceangoing commercial vessel, and
(right) another kite propulsion concept developed by Kiteship.
Another more challenging concept is to install one or several windmills on board a ship. This
concept can use not only provide propulsion forces but also the electrical power generated by
windmills. Windmill propulsion generally provides high propulsive force only when the ship
speed is less than approximately half the wind speed (Blackford 1985). This finding implies that
wind turbine propulsion is more preferable for slow steaming ships or sailing in high wind
conditions. However, high wind conditions may induce extra challenges to a ship’s structural
safety and its stability in open sea operations. Ship motions, when operated in the open sea, can
significantly decrease the efficiency of the power generation from on board windmills.
Furthermore, to use the electrical power generated by the retrofitted windmills, a large investment
must be made to revise the marine engine system in ships. Maintenance costs may become too
expensive during the operating period of windmills due to their working conditions under moist
and salty sea environments. Hence, no merchant ships have been equipped with windmills to
provide auxiliary ship propulsions (Bøckmann and Steen 2011).
One century of the history of the maritime industry included merchant ships using hybrid
propulsions between steam engine and traditional sails. Eventually, by the early 1900s, shipping
companies migrated their ships to steam ships. Gradually, every transoceanic sailing-ship
company went out of business because of the high energy efficiency of a marine engine
combined with a modern propeller design and the low cost of energy resources. Furthermore,
merchant ships with a steam/marine engine were capable of sailing across oceans faster than the
wind. They could also be operated in a more regular and flexible manner. Currently, due to the
high bunker price and environmental concerns related to bunker burning, hybrid ships using sails
as auxiliary propulsions have attracted much attention among different stakeholders in the
maritime transport sector. Sails on masts are often categorised as traditional soft sails and
wingsails, which are airfoil-like structures similar to airplane wings.
Traditional soft sails are still used for primary propulsion to many commercial fishing crafts in
developing countries. These sails are also used on larger trading vessels in trade areas, including
along South Pacific routes, with a long distance and high fuel costs, as noted by Bose (2008),
Teeter and Cleary (2014). Soft sails are often designed as fore-and-aft sails or square sails, as
described by Angvik (2009), Nance (1985), and Silvanius (2009). The largest advantage of using
soft sails on merchant ships is their relatively low cost installation and maintenance. A major
drawback is the lack of auto operation/control of the soft sails. These sails must either rely on
extra crew members to pilot sails or use very complex and expensive automated system for the
launch, recovery and operation of the sails. Furthermore, the fuel savings using soft sails are

ISSN: 1652-9189
usually marginal, and handling is difficult (Schenzle 1985). The technical requirements for the
practical applications of soft sails for auxiliary ship propulsion have been discussed at several
conferences (e.g., ICSACFV 1983, WASP 1985, and WINDTECH 1985).
Due to the growing theoretical understanding of aerodynamics and advancements in wingsail

design and construction in the aircraft industry, Abbott and von Doenhoff (1959), rigid airfoil-
like sails (wingsails) have become more attractive alternatives for providing wind propulsions to
ships. The appearance of wingsails looks similar, whereas the details can differ (Atkins 1994):
• plain aerofoil section,

• aerofoil with plain flap,
• aerofoil with single slotted flap,
• aerofoil with double-slotted flap, and
• aerofoil with variable “nose” camber.
These details can increase the wingsails’ maximum lift coefficient and lift-to-drag ratio. In
addition, different materials for wingsail construction, size and arrangement of wingsails on the
ship, and the pilot/control/operation method can be quite different from those wingsail systems.
Figure 6 lists a few wingsail design proposals from the early 1980s; see Nance (1985). The most
successfully implemented concept is the folding wingsail developed by the Japanese company
NKK and fitted first to the Shin-Aitoku-Maru, a 1400 DWT tanker (Fujiwara et al. 2003,
Watanabe et al. 1982). After the successful installation of the system on board the Shin-Aitoku-
Maru tanker, it was also retrofitted to approximately 10 ships of different sizes, ranging from 600
to 31,000 tons, built in six different shipyards, and operated by six different companies in Japan.
This system has been demonstrated to possess a potential fuel savings of up to 20-30%.
Figure 6: Typical wingsail proposals developed in the early 1980s (Nance 1985).

ISSN: 1652-9189
According to Cooke Associates (2014), the Walker Wingsail (Walker 1985) was a more
aerodynamically efficient wingsail that was developed and applied to the M.V. Ashington
recreational vessel during the period between 1986 and 1988. Another wingsail proposal, the
Boatek wingsail, is believed to have an even higher efficiency and larger lift coefficient. The
Walker wingsail has been compared with the Boatek wingsail through a numerical simulation of
two-dimensional section aerodynamic characteristics. According to the analysis by Cooke
Associate (2014), the Boatek design exhibits better aerodynamic characteristics, most likely due
to the use of cambered, as opposed to symmetrical, wing elements. Cambered designs, however,
must be rotated each time the boat changes tack. Figure 7 presents two examples of the Walker
and Boatek wingsails. Although the economic benefits and technical development of the two
wingsail proposals are not proven by their practical installation on merchant vessels, these
designs are applied extensively on recreational vessels. For their application on merchant ships,
robust and advanced sail-assist/operation systems are usually required to avoid the need for extra
crew sources to handle the wingsails on board. Both the development of sail-assist systems and
optimisation of wingsails requires good understanding of the mechanical behaviour/response of
wingsails under windy conditions. Insight can be gleaned from the aircraft industry for both
numerical analysis and experimental tests; see Atkins (1994), Bergeson and Greenwald (1985),
Fujiwara et al. (2003), Marie and Courteille (2014), Satchwell (1985), Scherer (1974), and
Sheldahl and Klimas (1981). However, the special characteristics of wingsail applications on
ships, such as the mechanical interaction between aerodynamic on wingsails and hydrodynamic
response under a ship’s open sea operation, and the low and irregular wind speed encountered by
wingsails also bring further challenges to sail assist ships; see Firestein (1985), Prandtl and
Tietjens (1957), and Spaans (1985). It is well known that today’s computational capability
enables advanced CFD computations of fluid dynamics for understanding the physical
phenomena of wind assist ship performances. However, the complete coupling of aerodynamic
and hydrodynamic analysis still requires too much computational effort. In particular, wind assist
operation systems often need robust models to describe such a ship’s aerodynamic and
hydrodynamic performance via a fast and efficient approach. More challenges in addition to the
technical development of these concepts are described in the following chapter.
Figure 7: Modern aerodynamically efficient wingsail designs:

(left) Walker wingsail and (right) Boatek wingsail.

ISSN: 1652-9189
5.3 Feasibility study for wind assist propulsions
In the previous chapter, potential barriers that hinder the practical applications of different
measures for energy efficient shipping were briefly presented. Although the maritime industry
has accumulated long periods of sailing experiences using wind power as the main propulsion, its
application on today’s ships to provide auxiliary propulsions is facing new challenges with
respect to economic investment, design and operation of the devices, as well as operational
viabilities. Based on a literature review of the technical development of different wind assist ship
concepts, Table 3 summarises the feasibilities of implementing these concepts to identify further
research and development efforts to enable practical applications. All of these economic and
technical barriers can be overcome by engineering intelligently before marketing.
Table 3: Feasibility study with respect to the potential application of wind assist devices
Viability of economic investment

In general, to generate a large wind force to provide auxiliary propulsion to merchant ships, wind
assist devices should also be appropriately large. Thus, the investment required to design,
construct and install is usually not a small cost. In particular, these devices should be designed
separately for individual ships to achieve a higher efficiency and energy gains in comparison with
its investment for real installations. Furthermore, due to severe working environments (high
moisture and salinity) of these wind assist devices in the open sea, one also needs to account for
long-term maintenance costs during their operation. For ship owners, the preferred payback time
of their investment is 5 to 10 years. However, usually the current payback time is often longer
than that. For example, according to O’Rourke (2006), to install the wind device of a skysail, the
cost of retrofitting a cargo ship with a row of masts and strengthening the ship to dissipate stress
due to the sail was estimated to be more than 10 million USD. This would take approximately 15
years to recoup their costs through fuel savings. In addition, a high level of maintenance costs is
usually associated with the initial installation investments of such expensive devices. The
maintenance sometimes can double the payback time. It should be noted that the fuel savings cost
is also strong related to future market oil price and future fuel cost scenarios for reduction of
shipping emissions Eide et al. (2011) and Faber et al. (2011). This may shorten payback time and
make the wind assist concepts more attractive to the shipping market.
Furthermore, large uncertainties are included in the estimation of fuel savings by implementing
these devices. One important source leading to this uncertainty is a lack of reliable
methods/models for predicting the performance (aerodynamic and hydrodynamic coupling
analysis) of these devices under real operating conditions, e.g., WASP (1985) and WINDTECH

ISSN: 1652-9189
(1985). Another reason is that the estimation of fuel savings should rely on the encountered sea
environments (e.g., trade routes, wind, wave) of a ship’s future sailing service. Fortunately, large
data sources are freely available to provide the long-term statistics of sea environments along
specific ship routes. For example, Clauss et al. (2007) and Traut et al. (2014) used weather data
from the European Centre for Medium-Range Weather Forecast (ECMWF) and the UK Met
office’s Unified Model to access wind statistics to estimate potential energy savings of different
wind assist devices along trade routes.
Moreover, the fuel savings by implementing these devices could be further maximised through
the optimised planning of ship routes for each individual ship voyage. Currently, most ocean-
crossing vessels are instrumented with a routing plan system, where a routing optimisation tool is
used to assist ship operation in a more optimal way based on weather forecast information. The
routing tool has the potential to save up to 5% of fuel cost and reduce structural fatigue damage
by 50%, Mao et al. (2012). The routing optimisation system is also one of the key factors in
maximising the use of wind propulsion in practical shipping. Spaans (1985) investigated different
optimisation algorithms for the utilisation of wind assist devices to reduce the energy cost of
shipping. However, there is no routing tool that can account for lift and drag wind forces for
routing optimisation.
Design and installation challenges

The work efficiency of the wind assist devices is of greatest concern for their practical
installation and practical applications. The aerodynamic performances of these devices were
extensively studied via wind tunnel tests (see Argatov 2006, Bergeson and Greenwald 1985,
Blackford 1985, Clayton 1985, Craft et al. 2012, Flettner 1925, and Satchwell 1985) and CFD
analysis. This was possible due to the rapid increase in computational capability (see Atkins 1994,
Cooke Association 2014, Craft et al. 2010, and Johnson 2011). However, the reliability of those
established models to describe the operational performances of these devices obviously requires
further development. Furthermore, some devices have a high weight and centre of gravity, which
may affect a ship’s seakeeping behaviour in addition to its high building cost. Wind devices are
often installed on ships through large masts. In unfavourable winds, large masts can only create
drag force, which causes ships to heel, sometimes dangerously. Masts and their pivoting sails
take up valuable cargo space on the deck. Loading and unloading is more expensive because the
cranes must work around the masts, Bose (2008). Tall masts or wingsails exceeding 100 meters
in height must often use collapsible equipment to avoid operation problems when meeting waters
with bridges. These collapsible masts will significantly increase the building cost.
Furthermore, due to the large forces generated by wind devices, the ships’ structural components
should be strengthened to enhance ship structural safety. For example, following the success of
wingsail implementation on the tanker Shin-Aitoku-Maru, another vessel installed a similar
wingsail on board that failed during high wind in a Japanese harbour. The high forces, partly
from the wingsail, broke the moorings of the vessel, which was blown across the harbour. The
parts seriously damaged other vessels in the harbour as well as the harbour itself. The accident
discredited wingsail safety and indicates a need for further technical development of the wind
assist device applications.

ISSN: 1652-9189
Operational challenges leading to energy efficiency (fuel saving)
In addition to the design and construction of wind assist devices, both the operation of a ship
fitted with these devices and the device handling/control requires robust performance prediction
models to describe the ship and wind devices’ response in real ship operations. Numerically
modelling complex ship motions in real sea environments has always been one of the greatest
challenges for hydrodynamic analysis in the maritime industry. The development of wind assist
ship performance models requires even more sophisticated aerodynamics and hydromechanics
coupling analysis.
A sailing ship sails with a constant heel and leeway angle to make use of wind propulsions to
power the ship, whereas current merchant ships are mainly powered by marine engines connected
with well-designed propellers. The installation of auxiliary wind propulsion devices will add at
least two extra wind forces to ships, depending on the number and arrangements of the devices.
Wind forces, which should be sufficiently large to ensure the device working efficiency (note that
large wind assist devices may block a ship’s operation view on the bridge, which is quite
essential in particular in port operations) will significantly affect a ship’s response behaviour.
This includes effects on the ship’s resistance, propeller efficiency, ship manoeuvrability and ship
motions. Furthermore, the drag and lift forces generated from wind assist devices have to interact
with the thrust force provided by a ship’s propellers. The ship’s sailing course might also be
affected by such interactions. To achieve the highest work efficiency of these devices, an
optimisation procedure should be developed to guide the operation of the ship, referred to as
routing optimisation (optimum voyage plan) systems. In this case, routing optimisation is a key
factor in helping captains to operate their ships and maximise the utilisation of wind propulsion
through the iterative consideration of wind device performance and the ship’s main power
management. This helps achieve the expected time of arrival with minimum fuel cost. In the
following chapter, the characteristics of routing systems available in the current shipping market
will be summarised to give a state-of-the-art description of such services. Further development
requirements and directions to account for the wind assist device in future routing systems will
also be discussed.
6. Ship routing systems for energy-efficient shipping

Ship routing systems, also known as weather routing or optimisation voyage plans depending on
their functionalities, are the discipline of producing the most favourable route with respect to
expected time of arrival (ETA), passing waypoints, ship speed, heading and engine power
profiles for a given voyage trade based on encountered weather forecast information and a ship’s
operational performance. The primary use for these systems is ocean crossing transits, but they
are also used for coastal navigation. In the current routing system market, the most interesting
objective of ship routing systems is to provide a service to minimise fuel costs during sailing
while arriving on time, considering the safety of crew, ship, and cargo, as well as air emissions in
certain control areas.
To further the understanding of the structure of a ship routing system, it can be broken down into
comprehendible categories and sub-components. Figure 8 is intended to provide a general
overview of this structure. The basic idea of ship routing is to provide/suggest a ship route
(including ship course, speed, heading and engine power), as illustrated in a digital chart with
plotted forecasted weather information. At its early stage, a system should be able to account for

ISSN: 1652-9189
the expected time of arrival and safety issues, which can be a ship’s motions/stability in the open
sea, structural safety under extreme sea conditions, cargo safety and crew/passenger comfort.
Figure 8: Overall structure of a weather routing system.
6.1 Basic capabilities of current routing systems

A ship routing system is designed to provide the service of optimising and planning a ship’s
sailing course/route with respect to the earliest possible time of safe arrival or expected time of
safe arrival (ETA) based on weather forecast information and hindcast statistics such as air
pressure, waves, wind, and current. The safety factors considered in the service include a ship’s
stability, seakeeping, structural integrity, cargo protection, and passenger comfort. The
functionalities of these systems are summarised as follows, based on 6 companies providing
routing services in the worldwide market as shown in Table 4.
Table 4: Typical routing systems in the worldwide shipping market.

No. of vessels
Company Website
installed
Voyage
4000 http://weathernews.com
planning
BonVoyage 3500+ http://www.appliedweather.com
SPOS 2500+ http://www.meteogroup-maritime.com
Seaware unknown http://www.seaware.se
OCTOPUS unknown http://www.amarcon.com
SeaRoutes unknown http://www.searoutes.gr

ISSN: 1652-9189
To optimise a ship’s sailing route to minimise fuel cost in operation and enhance safety, the basis
of a routing system is the access to reliable weather forecast information. There exists a wide
range scatter of weather forecasting provided by different institutes; hence, some systems also
use satellite and hindcast weather statistics to calibrate their accepted weather forecast
information from metrological institutes. Normally, a routing system receives updated weather
forecast information from meteorological institutes to ships every 6 to 24 hours. Depending on
the trade regions, weather information can be chosen with different resolutions. The weather
information is primarily used to plan an optimised ship route for a sailing voyage. In addition, it
can be plotted in a sea chart for captains to visualise the weather change, which may help them
adjust their planned routing schedule. In the planning process of a routing schedule, the
information contained in the sea chart such as water depth and surrounding marine traffic is also
necessary. In general, the most important weather forecast information received from
metrological institutes for routing plan can include the following:
• pressure contours,
• surface wind,
• temperature and moisture,
• wave and swell, and
• current and tides.
Furthermore, some routing systems also provide wave and wind statistics based on historical data
for the locations along the planned course. The primary target of a routing service is to provide
ships with a reliable ship sailing schedule to achieve the ETA. The expected time of arrival can
be to plan a ship’s route with respect to the earliest possible time of safe arrivals or required time
of safe arrival, depending on the transport contract. It will not be distinguished in this report and
will be denoted as ETA hereafter. To have a good plan for the ship with an expected time of
arrival, routing systems can implement the following information for optimisation:
• requirements from ship operation company or ship owners,

• departure and arrival port database, and
• interface with other transportation.
Though ship safety issues have been casually dismissed by the current maritime field because the
technical issues seemed to be advanced for many years, the routing systems can implement
technical solutions to provide ships with safety-related services. These would include, for
example, ship structural ultimate strength in large storms (wave/wind loads), structural fatigue
damage, risks of collision and ground along the route, and ship stability problem. The stability
may also be related to seakeeping properties, i.e., ship motions and accelerations of
roll/pitch/heave for different loading and weather conditions. These technical issues should be
considered separately for each individual ship. However, some routing systems may focus on
more general problems connected with the weather forecast information and just set a threshold
value of the weather (wind/wave) with respect to the ship safety resistance.

ISSN: 1652-9189
The analysis of ship safety can be carried out via different approaches:
• Performance models based on experimental tests and numerical simulations.

• Development of robust models to describe a ship’s motions and wave-induced loads applied
on the ship’s hull to gain a quick understanding of the ship’s seakeeping behaviour and
structural integrity response for different operational conditions during practical sailings.
• Hull monitoring system.
• Installing sensors along the hull to measure the six degrees of motions of the ship. The strain
signal is also occasionally recorded to investigate/check ship structural integrity.
• Post-voyage analysis.
• Analysing a ship’s response characteristics after the planned voyage is complete. This will
help the engineer/inspector understand more about the ship’s safety status. The insight
gained from the post-analysis can also help further optimise the ship’s future routes.
The most popular topic in the current maritime market is, most likely, energy-efficient shipping,
which can help the industry reduce fuel consumption and air emissions. In addition, the shipping
performance and efficiency may include the following criteria:
• minimum fuel consumption,

• best cargo protection,
• less ship speed loss, and
• passenger comfort such as avoiding large motions and whipping phenomena.
6.2 Optimisation algorithm in routing systems

The core element in the development of a routing system is to choose a powerful optimisation
algorithm, which has a crucial effect on the computation time, solution and quality of routing
plans. In the maritime community, the most common algorithms are Dynamic Programming,
Dijkstra Algorithm, and the genetic-evolutionary algorithm. In a study by Larsson and Simonsen
(2014), the so-called DIRECT (DIviding RECTangles) algorithm (Finkel 2003), which was
developed based on Lipschitz optimisation (Jones et al. 1993), was also implemented in a routing
system. The DIRECT algorithm illustrates how the algorithm moves toward an optimum. It is
simple in nature and requires only adjustment of few settings. The algorithm searches locally and
globally simultaneously, and intensification of a local search is not restrictive for the global,
which is contrary to many other algorithms. A state-of-the-art analysis of selected algorithms is
presented as follows.
3D Dynamic Programming
The Bellman principle of optimality forms the basis of dynamic programming, stating that “an
optimal scheme has the built in feature, that no matter what the initial circumstance and
selections are, the remaining decisions must comprise an optimal scheme regarding that
circumstance”. The more conventional 2D Dynamic Programming method (Chen et al. 1998)
may be used to compute an optimum route by determining headings. However, it is restricted to
determining the route using constant speed or power for the entirety of the route. The 3D
Dynamic Programming model (3DDP), as noted by Avgouleas (2008), Wei and Zhou (2012),
allows for the optimisation of the sailing profile through speed and power control.

ISSN: 1652-9189
Shao et al. (2006) described the development of a forward dynamic programming method for
weather routing. The method provides a route based on both course and power to allow for the
ship to slow down or speed up relative to its average speed throughout the journey. The 3DDP
method, (3D Dynamic Programming) shows significant fuel savings compared with the more
common 2D Dynamic Programming method. The method presented does not account for the
uncertainty of weather but only handles weather forecast data directly. The method was measured
against three different Genetic Algorithms (GAs) by Shao et al. (2006). The 3DDP method
showed improved performance compared with all three GA methods. The 3DDP method is
dependent on a pre-processed grid.
Pros & Cons: Because the method relies on a predefined grid, impassable areas may be easily
handled by skipping grid points. The pre-processed grid may be created relatively simply by
spreading grid points out perpendicularly at intervals following the great circle route. A
secondary advantage of the method lies in a by-product resulting from the nature of the solution
space, which produces optimal routes for different arrival times, thus allowing the user an active
selection. Though this method shows the advantage of easily handling impassable areas, finding
the shortest route around or between islands requires significant grid resolution. This makes the
accuracy of the result highly dependent on the grid resolution, which directly couples to the
required computation time.
Dijkstra
The algorithm works based on two principles. The first principle forms the basis for dynamic
programming (Sniedovich 2006): “A sub-route of a shortest route is itself a shortest route”. The
second principle is that “With a given shortest distance, x, between points A and C, a path going
from point A to C through a third point B will always be of a distance greater than or equal to x”.
From the starting point, all connecting lines to neighbouring points are simulated and assigned
cost values. This value may be distance or fuel consumption or a weighted combination. From
the line with the lowest end point value, possible lines to the next neighbouring points are
simulated and assigned costs. When a line reaches a point that has already been tested, the second
principle stated above is used. If the end point cost for the newly tested route is smaller than the
original, the new cost is assigned to that point, and the original route dismissed as part of an
optimum solution. Following this routine, routes will evolve toward the end point, and one
optimum route remains. An example of such a route optimisation can be observed in Figure 9.
The bold lines mark the routes with the smallest values, and the one stretching from point “Start”
to point “Destination” is the optimum route.
Figure 9: Illustration of the Dijkstra algorithm.

ISSN: 1652-9189
Mannarini et al. (2013) described the development of a conceptual model relying on a
modification of the Dijkstra algorithm (Dijkstra 1959). The method only accounts for wave
height and direction, along with safety restrictions based on these parameters. The described
method does not handle voluntary speed reductions. Padhy et al. (2008) presented another
modification to the Dijkstra algorithm with the capability of handling voluntary speed reduction.
The method uses weather conditions as influences on the obtained speed and may handle
impassable areas through assigned weight of edges.
Pros & Cons: The Dijkstra algorithm will always find the most optimal route from the given grid.
For cases of constant speed and handling of ship motions through reduction curves only, the
algorithm is very fast. However, possible routes are restricted to move along nodes of the grid
and are thus highly grid-dependent. In particular, for the 3DDP method, this method will have
high grid dependency. Modifying the algorithm for handling dynamic weather, which is essential
for varying speed, may be complex. Arrival time and minimum fuel consumption must be
handled simultaneously, as shown by Zeszyty (2012).
Genetic Algorithm
The genetic algorithm replicates nature’s evolution with principles of survival of the fittest. It
does so through an initial population of solutions created through random sampling routes with
advantages in fuel consumption, ETA, safety, or a combination of these. The routes are evaluated
and given a fitness value based on fuel consumption, ETA, safety or predetermined
combinations. A selection operator determines the routes containing the best fitness values,
which may then be crossed over, mutated or handled through specialised operators to form a new
solution population.
Often cited is the work of Hinnenthal (2008), who provided an elaborate suggestion of the use of
a multi-objective genetic algorithm. Ship responses and constraints are modelled using ship
response operators. In this proposed method, the objectives of the optimisation are both fuel
consumption and arrival time, optimising both route and velocity profiles.
Pros & Cons: GAs are generally implemented fairly easily, requiring minimum tuning and
modification from the base form. The algorithm may be stopped at any time and an optimal
solution produced based on a Pareto front and ranking. Through mutation, crossover and other
specialised operators, the algorithm is not likely to become stuck in local minima given proper
tuning. From the Pareto fronts, the operator has the possibility, through ranking parameters, of
selecting a most suitable solution with regard to ETA, fuel consumption, and safety. Conversely,
when using this algorithm, arrival time and minimum fuel consumption must be handled
simultaneously. The algorithm does not ensure an optimum solution but will instead give an
approximation for it, which improves with the number of iterations. The GA is not very fast and
is highly dependent on the population size of the initial generation of solutions. The selection of
first generations may be very influential on the solution because these couple directly to the
degree of mutations and crossovers. The parameters must be tuned and are likely to be case-
specific for optimal computing.
The conclusion is that the less advanced algorithms leaning toward brute force are phased out in
favour of more advanced algorithms such as versions of 3D Dynamic Programming and Genetic
Algorithms. These algorithms perform better when the optimisations are multi-objective and

ISSN: 1652-9189
when computation time is limited. Algorithms with the capability of handling black box problems
are favoured because the aspects of the computations and totality of the modelling are too
complex to be based on derivatives. Both the dimensionality of the solutions and dynamics of the
solution space limit the number of applicable algorithms. Increased insight into system models
and handling of the routing as a total voyage system are likely to lead the way to even more
advanced algorithms being implemented.
6.3 Future routing concepts

The state-of-the-art and future development concepts of routing systems are summarised in
Figure 10 based on three main activities. The first activity was mapping of the market leaders
within weather routing software. Based on the literature, the selected software providers were
contacted and were asked to fill out a questionnaire. The results of the study are presented. The
second activity was a study of current and previous research to map the evolution of weather
routing. The third activity was participation in network groups and workshops on weather routing
to gain insight into the practical use of weather routing and its most current research and
development. The general findings are illustrated in Figure 10. Features and functions are divided
into three main categories depending on whether the functionality is mature and included in
current software, less mature and in the process of implementation or are only present in research
and possibly part of future implementation. The top weather routing software has the possibility
of routing with varying speed or power profiles, enabling ships to speed up or slow down to
avoid constraining weather. To prevent involuntary increases to resistance, many components are
considered when routing with the leading software, including: wind, waves, currents, fouling, and
engine maintenance. Basic capabilities of a routing system were described in previous chapters.
The following text focuses on upcoming and future concepts, which may be included in routing
systems.
Figure 10: General overview of state-of-the-art and future routing concepts.

ISSN: 1652-9189
In the current maritime community, the main driving force is to minimise fuel consumption for
shipping. Constantly increasing computing power enables response based routing, which in turn
will increase safety for ships, cargo and on board personnel. Response-based routing will most
likely be coupled with on board monitoring systems to achieve even more precise models. The
added resistance in seaway is of great interest to ship routing services. Speed reductions are
handled in Shao et al. (2006) through the reduction equations presented by Kwon (2008). The
equations are a simplification that leads to reduction percentages. The most advanced response-
based routing system will be able to use models established based on model test data, on-board
hull monitoring data and numerical simulations, which can achieve the highest level of precision
to accurately describe a ship’s response under random sea way. Furthermore, it is possible to
integrate the ship response model with other energy-efficient measures, such as wind assist
devices, through the coupling of aerodynamic and hydrodynamic considerations. Using a reliable
optimisation algorithm, the routing optimisation system could serve as a platform for
implementations and make the best use of these energy-efficient measures to reduce fuel
consumption and air emission in a more systematic way.
Another important development area is connected to the uncertainty and spread of weather
forecasts. There is a possibility of risk-based routing. The difference between the paths of routes
is another possible constraint. A route ahead of a storm would often be considered a higher risk
route than a route behind the same storm. Weather uncertainty handling is one of the areas of
research that have the greatest potential for improved routing. An obvious element is the
discussed timespan of forecasts. Further, accounting for the uncertainties related to the forecasted
weather may lead to very different routes. A discussed solution is weighing in probable
forecasted weather from ensemble and super ensemble forecasts. Few software or research
projects have yet to use this possibility. Some examples of the use of ensemble forecasts are
presented by Hinnenthal (2008) and Skoglund et al. (2012).
Other focus areas in the routing domain are related to the new SECA regulations and piracy
protection, which are pressing matters for ship owners and operators. As of January 1, 2015,
SECA areas (Sulphur Emission Controlled Areas) came into force. These areas result in higher
operational costs for larger ships operating in these areas. Current weather routing providers
handle the SECA areas differently. Some providers only display the areas without considering
their effects into the optimisation, others account for the effects of possibly switching fuel or
requiring scrubbers, and still others add them as no-go areas. The providers that are approached
but have not yet implemented the SECA areas are to implement the areas in coming releases.
Piracy is a current and highly prioritised issue within shipping. Piracy services are seldom
directly included in routing software but are often provided as a separate or add-on service.
Piracy is also connected to risk-based routing, which is further discussed below. Fatigue routing
is a discussed topic. Research has shown that container vessels in the North Atlantic trade may
increase their fatigue life by 50% by implementing fatigue routing (Mao et al. 2012). Fatigue
routing does not present a direct coupling to immediate costs and has therefore not been
implemented yet. The inclusion of fatigue into multi-objective routing is at present a means of
taking routing one step further.

ISSN: 1652-9189
7. Concluding remarks
The current report summarises the outcome of the project “Examination of concepts, available
models and methods useful for prediction, monitoring and improvement of ship energy
efficiency”, funded by the Swedish Energy Agency, Contract No. 39422-1. The purpose was to
deliver a state-of-the-art study that compiles and examines the models and methods available for
the prediction and monitoring of the vessel’s energy use, with the aim of working toward
improved energy efficiency. It also includes an inventory of green concepts that utilise wind
power and a special section on wingsails on ships and the route planning investigated.
The outcome of the study highlights at least two important issues. First, there is a need to develop
a generic holistic model that is applicable for the energy efficiency analysis and simulations for
ships. Current state-of-the-art technology shows that parts of such a model exist, but they need to
be combined to interact. Second, using wind power as auxiliary propulsion and ship routing
optimisation can significantly reduce energy consumption and improve the energy efficiency of
ships. Several technical concepts have been evaluated and tested, and wingsails are a solution that,
together with ship routing algorithms, can lower fuel consumption significantly. In addition to
these issues, a discussion on the scopes of different models, methods and software is presented;
their limitations, strengths, weaknesses and flexibility to be applied or extended to the new
extended use are also discussed. It is the authors’ opinion that the means for future energy
efficient shipping is to build a generic holistic energy systems model of a vessel that can be
applied to demonstrate and develop concepts in a virtual but yet “realistic” environment. This
model can be used efficiently to enhance the performance characteristics of existing vessels by
using the model together with the results from ship monitoring and performance measurements.
Acknowledgements
The project was funded by the Swedish Energy Agency, contract no. 39422-1, “Examination of
concepts, available models and methods useful for prediction, monitoring and improvement of
ship energy efficiency”.
8. References
Abbott, I. H. and von Doenhoff, A. E. (1959). Theory of Wing Sections. Dover Publications.
ABS (2013). Ship energy efficiency measures advisory. Technical report, Houston, USA.
Andersen, P., Borrod, A.-S., Blanchot, H. (2005). Evaluation of the Service Performance of
Ships, Marine Technology, Vol. 42 (4), pp. 177-183
Angvik, I. (2009). Wind Assisted Ship Propulsion, master thesis, NTNU, Norway.
Aoki, K. and Ito, T. (2001). Flow characteristics around a rotating cylinder, Proc. School of
Engineering Tokai University, Vol. 26, pp.29-34.
Argatov I, Rautakorpi P, Silvennoinen R. (2009). Estimation of the mechanical energy output of
the kite wind generator. Renewable Energy, Vol. 34 (6), pp.1525-1532.
Atkins, D.W. (1994). The CFD assisted design and experimental testing of a wingsail with high
lift device. PhD dissertation, University of Salford, Salford, UK.

ISSN: 1652-9189
Avgouleas, K. (2008). Optimal ship routing. Master's thesis, Massachusetts Institute of
Technology, Cambridge, USA.
Baldi, F. (2013). Improving ship energy efficiency through a systems perspective, Licentiate
thesis, Chalmers Technical University, Technical Report No. 13:147
Bergeson, L. and Greenwald, C.K. (1985). Sail assist developments, 1979-1985, Journal of wind
engineering and industrial aerodynamics, Vol. 19 (1-3), pp.145-114.
Blackford, B. L. (1985). Optimal Blade Design for Windmill Boats and Vehicles, Journal of Ship
Research, Vol. 29 (2), pp.139 – 149.
Boese, P. (1970), Eine Einfache Methode zur Berechnung der Widerstandserhöhung eines
Schiffes in Seegang, Institut für Schiffbau der Universität Hamburg, Bericht 258
Bose, N. (2008). Marine powering prediction and propulsors, SNAME,USA.
Buhaug, Ø., Corbett, J.J., Endresen, Ø., Eyring, V., Faber, J., Hanayama, S., Lee, D.S., Lee, D.,
Lindstad, H., Markowska, A.Z., Mjelde, A., Nelissen, D., Nilsen, J., Pålsson, C., Winebrake,
J.J., Wu, W. and Yoshida, K. (2009). Prevention of air pollution from ships – second IMO
GHG study. International maritime organization, London, UK.
Bøckmann, E. and Steen, S. (2011). Wind Turbine Propulsion of Ships, Second International
Symposium on Marine Propulsors smp’11, Hamburg, Germany.
Calleya, J. (2014). Ship Design Decision Support for a Carbon Dioxide Constrained Future, PhD
Thesis, University College London
Calleya, J., Pawling, R., Greig, A. (2014). Ship Design and Emissions Abatement Options
considering EEDI, Influence of EEDI on Ship Design, RINA, 24-25 September 2014, London,
UK
Caprino, G., Della Loggia, B. (1993). Recent developments on the integrated ship performance
monitoring systems, Proceedings of the 20th ITTC, The On-board Monitoring session
Chen, H. Cardone, V. and Lacey, P. (1998). Use of operation support information technology to
increase ship safety and efficiency, SNAME transactions, Vol.106, pp.105-127.
Clauss, G.F., Siekmann, H., Tampier, B.G. (2007). Simulation of the operation of wind-assisted
cargo ships, In: Hauptversammlung der Schiffbautechnischen Gesellshaft, Berlin, Germany.
Clayton B.R. (1985). BWEA initiative on wind assisted ship propulsion (WASP), Journal of
wind engineering and industrial aerodynamics, Vol. 19 (1-3), pp.251-276.
Cooke Associates, (2014). Available at: www.cookeassociates.com
Corbett, J. J., Wang, C., Winebrake, J. J., Green, E. (2007). Allocation and Forecasting of Global
Ship Emissions; Clean Air Task Force and Friends of the Earth International, Boston, USA.
Craft, T.J., Iacovides, H., Johnson, N. and Launder, B.E. (2014). Back to the future: Flettner–
Thom rotors for maritime propulsion? Turbulence, Heat and Mass Transfer 7.
Crane, H. E. (1982). Ship efficiency analyzer, Google Patents, US4334425, available at
www.google.co.in/patents/US4334425

ISSN: 1652-9189
Dadd, G.M. Hudson, D.A. and Shenoi, R.A. (2010). Comparison of two kite force models with
experiment J. Aircr., Vol.47 (1), pp. 212–224.
Dalsoren, S., Eide, M., Endresen, O, Mjelde, A., Gravir, G, Isaksen, I. (2009). Update on
emissions and environmental impacts from the international fleet of ships: the contribution
from major ship types and ports, Atmospheric Chemistry and Physics, Vol. 9, pp. 2171-2194.
Dijkstra, E.W. (1959). A note on two problems in connexion with graphs, Numerische
Mathematik, Vol.1, pp.269–271.
DNV (2011) Greener shipping in North America. 2011.
http://viewer.zmags.com/publication/4ae38aec#/4ae38aec/1.
DNV (2012). Managing the complexity if ship machinery systems, DNV Research and
Innovation, Position Paper 11-2012
DNV GL (2014). Energy management study 2014. Technical report, Oslo, Norway.
Duckworth, R. (1985). The application of elevated sails (kites) for fuel saving auxiliary
propulsion of commercial vessels. J Wind Eng and Ind Aerodynam, Vol. 20 (1-3), pp.297–
315.
Dyne, G. (1995). The Principles of Propulsion Optimization, Transactions RINA, No. 137, pp.
189-208
Eide, M.S., Longva, T., Hoffmann, P., Endresen, Ø., Dalsøren, S.B. (2011). Future cost scenarios
for reduction of ship CO2 emissions, Maritime Policy Management, Vol. 38 (1), pp.11-37.
ENERCON (2013). Rotor sail ship “E-Ship 1” saves up to 25% fuel. Available at:
http://www.enercon.de/en-en/2224.htm.
Erhard, M., Struch, H. (2012). Control of towing kites for seagoing vessels, Cornell University e-
prints, (arXiv:1202.3641v1), pp.1–10.
European Commission. (2011). Commission Staff Working Paper - Pollutant emission reduction
from maritime transport and the sustainable waterborne transport toolbox. Brussels, Belgium.
Faber, J., Markowska, A.Z., Nelissen, D., Davidson, M., Eyring, V., Cionni, I. et al. (2009).
Technical support for European action to reducing greenhouse gas emissions from
international maritime transport. Delft, The Netherland.
Faber, J., Wang, H., Nelissen, D., Russell, B., Amand, D.S. (2011) Marginal abatement costs and
cost effectiveness of energy-efficiency measures, SNAME report, USA.
Fang, M.-C., Chen, G.-R. (2006). On the nonlinear hydrodynamic forces for a ship advancing in
waves, Ocean Engineering, Vol. 33 (14), pp. 2119-2134
Finkel, D.E. (2003). DIRECT Optimization Algorithm User Guide, Center for Research in
Scientific Computation, North Carolina State University, USA.
Firestein, G. (1985). Dynamic response of a wingsail mast to unsteady aerodynamic loads
including ship motion effects, Journal of wind engineering and industrial aerodynamics,
Vol.20 (1-3), pp.23-38.

ISSN: 1652-9189
Fleiter, T., Worrell, E., Eichhammer, W. (2011). Barriers to energy efficiency in industrial
bottom-up energy demand models e a review, Renew Sustain Energy Rev, Vol.15 (6),
pp.3099-3111.
Flettner, A. (1925). The Flettner rotor ship, Engineering.
Fujiwara, T., Hirata, K., Ueno, M., Nimura, T. (2003). On aerodynamic characteristics of a
hybrid-sail with square soft sail, Proceedings of 13th International Offshore and Polar
Engineering Conference, Hawaii, USA.
Grimmelius, H.T. (2003). Simulation Models in Marine Engineering: From Training to Concept
Exploration, Compit, Hamburg
Grosan, N., Dinu, D. (2010). Considerations regarding kite towed ships’ maneuvering,
Proceedings of the 3rd International Conference on Maritime and Naval Science and
Engineering.
Hansen, S. V. (2011). Performance Monitoring of Ships, PhD Thesis, Technical University of
Denmark, Department of Mechanical Engineering
Hinnenthal, J. (2008). Robust Pareto Optimum Routing of Ships utilizing Deterministic and
Ensemble Weather Forecasts, PhD thesis, Technischen Universitat Berlin, Germany.
Holtrop, J., Mennen, G. (1982). An Approximate Power Prediction Method, International
Shipbuilding Progress 29 (335), pp. 166-170
ICSACFV. (1983). Proceedings of the International Conference on Sail-Assisted Commercial
Fishing Vessels, Florida Sea Grant College and the Society of Naval Architects and Marine
Engineers, Tarpon Springs.
International Maritime Organization (IMO) (2012). EEDI e rational, safe and effective, Available
at: http://www.imo.org/MediaCentre/HotTopics/GHG/Pages/EEDI.aspx.
International Maritime Organization (IMO) (2013a). Greenhouse gas emissions. IMO. Available
at:
http://www.imo.org/OurWork/Environment/PollutionPrevention/AirPollution/Pages/GHGEmi
ssions.aspx.
International Maritime Organization (IMO) (2013b). Prevention of air pollution from ships. IMO.
Available at:
http://www.imo.org/OurWork/Environment/PollutionPrevention/AirPollution/Pages/Air-
Pollution.aspx.
ISO (2002). Ships and marine technology – Guidelines for the assessment of speed and power
performance by analysis of speed trial data, ISO 15016
ITTC (1999). 1978 ITTC Performance Prediction Method, Recommended Procedures,
Performance, Propulsion
ITTC (2002). The Specialist Committee on Speed and Powering Trials, Final Report and
Recommendations to the 23rd ITTC
ITTC (2005). Full Scale Measurements, Speed and Power Trials, Analysis of Speed/Power Trial
Data, ITTC recommended Procedures and Guidelines

ISSN: 1652-9189
Jafarzadeh, S., Ellingsen, H., Utne, I.B. (2012). Emission reduction in the Norwegian fishing
fleet: towards LNG? Second International Symposium on Fishing Vessel Energy Efficiency
(E-Fishing). Vigo, Spain.
Jin, W., Vassalos, D. (2013). Energy Modeling of Large 2-Stroke Marine Diesel Engine using
CFD, Low Carbon Shipping Conference, London
Johnson, H., Johansson, M., Andersson, K. (2014). Barriers to improving energy efficiency in
short sea shipping: an action research case study, J Clean Prod, Vol.66, pp.317-327.
Johnson, N.D. (2011). High-Reynolds-number flow past a rotating cylinder with and without
Thom disc. MSc-by-Research dissertation, Faculty of Engineering & Physical Sciences, The
University of Manchester, UK.
Jones, D.R. Perttunen, C.D. Stuckman, B.E. (1993). Lipschitzian optimization without the
lipschitz constant, Journal of Optimization Theory and Application, Vol. 79 (1), pp.157-181.
Journée, J. M. J. (1992). Quick Strip Theory Calculations in Ship Design, PRADS´92, Conf. on
Paractical Design of Ships and Mobile Structures, Volume I, Newcastle upon Tyne, UK
Karabelas, S.J., Koumroglou, B.C., Argyropoulos, C.D., Markatos, N.C. (2012). High Reynolds
number turbulent flow past a rotating cylinder. Applied mathematical modelling, Vol.36,
pp.379-398.
Kesicki, F., Ekins, P. (2012). Marginal abatement cost curves: a call for caution, Climate Policy,
Vol.12 (2), pp.219-236
Kesicki, F., Strachan, N. (2011). Marginal abatement cost (MAC) curves: confronting theory and
practice, Environmental Science & Policy, Vol.14 (8), pp.1195-1204
KiteShip (2014). Available: http://www.kiteship.com.
Kjellberg, M. (2013). Fully Nonlinear Unsteady Three-Dimensional Boundary Element Method
for Ship Motions in Waves, PhD Thesis, Chalmers Technical University Göteborg
Kristensen, H., O. and Lützen, M. (2012). Prediction of Resistance and Propulsion Power of
Ships, Technical University of Denmark/ University of Southern Denmark, Project no. 2010-
56, Report No. 04, Emissionsbeslutningsstottesystem
Kwon, Y.J. (2008). Speed loss due to added resistance in wind and waves. The Naval Architect -
RINA, UK.
Larsen, N. L., Simonsen, C. D., Nielsen, C. K., Holm, C. R. (2012). Understanding the physics of
trim, 9th annual Green Ship Technology (GST) conference, Copenhagen
Letcher, J.S. (1982). Optimal performance of ships under combined power and sails, Journal of
Ship Research, Vol.26.
Liu, S., Papanikolaou, A., Zaraphonitis, G. (2011). Prediction of added resistance of ships in
waves, Ocean Engineering, Vol. 38 (4), pp. 641-650
Lloyd, M. (1980). Crosswind kite power, J Energy, Vol. 4 (3), pp.106-111.

ISSN: 1652-9189
Magnus (1853). Über die Abweichung der Geschosse, und: Über eine abfallende Erscheinung bei
rotierenden Körpern (On the deviation of projectiles, and: On a sinking phenomenon among
rotating bodies), Annalen der Physik, Vol. 164 (1), pp.1-29.
Mannarini, G., Coppini, G., Oddo, P., Pinardi, N. (2013). A prototype of ship routing decision
support system for an operational oceanographic service, The International Journal on Marine
Navigation and Safety of Sea Transportation, Vol.7 (1), pp.53-59.
Mao, W. Ringsberg, J.W. and Rychlik, I. (2012). What is the potential of using ship fatigue
routing in terms of fatigue life extension? In Proceedings of the 22nd International Offshore
and Polar Engineering Conference Rhodes, Greece.
Marie, S., Courteille, E. (2014). Sail-assisted motor vessels weather routing using a fuzzy logic
model. Journal of Marine Science and Technology, Vol. 19 (3), pp.265-279.
Marzi, J. (2014). Targeted Advanced Research for Global Efficiency of Transportation Shipping
– Final Public Report, M40 (36), Targeted Advanced Reasearch for Global Efficiency of
Transportation Shipping
Mermeris, D., Vassalos, D., Dodworth, K., Sfakianakis, D. and Mermiris, G. (2011). Dynamic
energy modeling – a new approach to energy efficiency and cost effectiveness in shipping
operations, Low Carbon Shipping Conference, London
Naaijen, P., Koster, V., Dallinga, R. (2006). On the power savings by an auxiliary kite propulsion
system. Int Shipbuild Progress, Vol. 53 (4), pp.255–279.
Nance, C.T. (1985). Outlook for wind assistance, Journal of Wind Engineering and Industrial
Aerodynamics, Vol. 19 (1-3), pp.1-17.
O’Rourke, R. (2006). Navy Ship Propulsion Technologies: Options for Reducing Oil Use-
Background for Congress. CRS report for congress.
Osses, J. R. P. and Bucknall, R. W. G. (2014). A Marine Engineering Plant Model to Evaluate the
Efficiency and CO2 Emmisions from Crude Oil Carriers, Shipping In Changing Climates:
Provisioning The Future Converence, Liverpool
Padhy, C. P. Sen, D., Bhaskaran, P.K. (2008). Application of wave model for weather routing of
ships in the north indian ocean, Natural Hazards, Vol.44, pp.373-385.
Papalambrou, G., Glaros, S., Angelou, E., Kyrtatos, N. (2012) Integrated Simulation Model of
Ship Hull, Propeller and Propulsion Engine for Performance Prediction and Fuel Consumption
Reduction, ELINT
Pérez Arribas, F. (2006). Some methods to obtain the added resistance of a ship advancing
waves, Ocean Engineering, Volume 34 (7), pp. 946-955
Prandtl, L., Tietjens, O.G. (1957). Applied Hydro- and Aero-Mechanics. Dover Publications,
New York, USA.
Satchwell, C.J. (1985). Aerodynamic design of marine aerofoils, Journal of wind engineering and
industrial aerodynamics, Vol. 20 (1-3), pp.1-22.

ISSN: 1652-9189
Schenzle, P. (1985). Estimation of wind assistance potential, in Proceedings of the International
Symposium on Windship Technology, WINDTECH’85, University of Southampton,
Southampton.
Scherer, J.O. (1974). Aerodynamics of high performance wingsails, Marine technology, Vol.11
(3).
Schlaak, M., Kreutzer, R., Elsner, R. (2009). Simulating possible savings of the skysails- system
on international merchant ship fleets, Int J Marit Eng, Vol.151 (4), pp.25–37.
Schulten, P.J.M. (2005). The interaction between diesel engines, ship and propellers during
manoeuvring, PhD thesis, TR diss 4446, Delft University of Technology
Shao, W., Zhou, P., Thong, S.K. (2006). Development of a novel forward dynamic programming
method for weather routing, Journal of Marine Science and Technology, Vol.11 (1), pp.239-
251.
Sheldahl, R. E., Klimas, P. C. (1981). Aerodynamic Characteristics of Seven Symmetrical Airfoil
Sections Through 180-Degree Angle of Attack for Use in Aerodynamic Analysis of Vertical
Axis Wind Turbines. Technical Report SAND80-2114, Sandia National Laboratories.
Silvanius, M. (2009). Wind assisted propulsion for pure car and truck carriers, master thesis,
KTH Center for Naval Architecture, Sweden.
Skoglund, L. Kuttenkeuler, J. and Rosen, A. (2012). A new method for robust route optimization
in ensemble weather forecasts, KTH Center for Naval Architecture, Sweden.
SkySails (2014). Available at: http://www.skysails.de.
Sniedovich, M. (2006). Dijkstra's algorithm revisited: the dynamic programming connexion,
Control and Cybernetics, Vol. 35 (3), pp.600-620.
Spaans, J.A. (1985). Windship routeing. Journal of wind engineering and industrial
aerodynamics, Vol. 19 (1-3), pp.215-250.
Thom, A. (1934). Effects of discs on the air forces on a rotating cylinder, Aero. Res. Counc.
R&M 1623.
Tian, G., Juando, A., Heslop, J., Solla, M., Roskilly, T., Sarasquete, A. Rodiguez, J. (2012). A
holistic ship model for variable speed generation system on a RoRo vessel, Low Carbon
Shipping Conference, London
Tosi, F., De Gaetano, S., Verde, E. (1982). Propeller and Hull Efficiency Monitoring system,
ISSOA ’82
Traut, M., Gilbert, P., Walsh, C., Bows, A., Filippone, A., Stansby, P., Wood, R. (2014).
Propulsive power contribution of a kite and a flettner rotor on selected shipping routes,
Applied Energy Vol.113, pp.362-372
Walker, J.G. (1985). A high performance automatic wingsail auxiliary propulsion system for
commercial ships, Journal of Wind Engineering and Industrial Aerodynamics, Vol. 20 (1-3),
pp.83-96.

ISSN: 1652-9189
WASP. (1985) Wind assisted ship propulsion – a development strategy. British Wind Energy
Association, Report of the BWEA Wind Assisted Ships Committee.
Watanabe, T., Endo, Y., Nakanishi, K., Takeda, K. (1982). Sail-equipped motor ship Shin Aitoku
Maru and studies on larger ship, in Proceedings of the 12th AIAA Symposium on
Aero/Hydronautics of Sailing, American Institute of Aeronautics and Astronautics/Society of
Naval Architects and Marine Engineers, pp. 238–249, Los Angeles, USA.
Weber, L. (1997). Some reflections on barriers to the efficient use of energy. Energy Policy, Vol.
25 (10), pp.833-835.
Wei, S., Zhou, P. (2012) Development of a 3d dynamic programming method for weather
routing. TransNav, Vol. 6 (1), pp.79-85.
Wellicome, J., Wilkinson, S. (1984). Ship propulsive kites: an initial study, University of
Southampton, UK.
Williams P. (2006). Optimal wind power extraction with a tethered kite. AIAA Guidance,
Navigation, and Control Conference and Exhibit, AIAA, Keystone, Colorado, USA.
WINDTECH (1985). International Symposium on Windship Technology, University of
Southampton, Southampton, UK.
Wrage, S. (2007). Positioning device for a free-flying kite-type wind-attacked element in a wind-
powered watercraft, U.S. Patent No. 20007/0272141 A1.
Wärtsilä (2010). Boost energy efficiency. Technical report.
Young, I. R., Rosenthal, W., Ziemer, F. (1985). A Three-Dimensional Analysis of Marine Radar
Images for the Determination of Ocean Wave Directionality and Surface Currents, Journal of
Geophysical Research, Vol. 90 (C1), pp. 1049-1059
Zeszyty, N. (2012). A comparison between dijkstra algorithm and simplified ant colony
optimizations in navigation. Scientific Journals, Maritime University of Szczecin, Vol. 29
(101), pp.25-29.

ISSN: 1652-9189

Systems Modelling For Energy-Efficient Shipping

Uploaded by

Copyright:

Available Formats

Systems Modelling For Energy-Efficient Shipping

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Systems Modelling For Energy-Efficient Shipping

Uploaded by

Copyright:

Available Formats

Systems modelling for energy-efficient shipping

Fabian Tillig, Wengang Mao, Jonas W. Ringsberg*

1 Report No. 15:153

• a 20 percent improvement in energy efficiency,

2 Report No. 15:153

2. General description of a ship as a system model

A system is a combination of equipment, machinery, external influences and processes that

• platform functions (carrying cargo, safety),

3 Report No. 15:153

• models requiring limited measurements,

2.1 Division into components

4 Report No. 15:153

5 Report No. 15:153

6 Report No. 15:153

7 Report No. 15:153

In conclusion, the propeller-hull interaction is challenging because the propulsive efficiency

3. Monitoring of energy consumption/ship performance monitoring

8 Report No. 15:153

• calm water resistance at even keel and known loading condition,

9 Report No. 15:153

10 Report No. 15:153

4.1 Holistic models

11 Report No. 15:153

Models for resistance and propulsion

Main engine models

Main engine and propulsion models

On board energy systems models

12 Report No. 15:153

Table 1: Examples of companies selling commercial software for performance monitoring.

5. Energy-saving measures and wind power as auxiliary propulsion

13 Report No. 15:153

5.1 Measures and their barriers to improve energy efficiency in shipping

14 Report No. 15:153

15 Report No. 15:153

16 Report No. 15:153

17 Report No. 15:153

18 Report No. 15:153

19 Report No. 15:153

20 Report No. 15:153

Due to the growing theoretical understanding of aerodynamics and advancements in wingsail

• plain aerofoil section,

21 Report No. 15:153

Figure 7: Modern aerodynamically efficient wingsail designs:

22 Report No. 15:153

Viability of economic investment

23 Report No. 15:153

Design and installation challenges

24 Report No. 15:153

6. Ship routing systems for energy-efficient shipping

25 Report No. 15:153

Figure 8: Overall structure of a weather routing system.

6.1 Basic capabilities of current routing systems

Table 4: Typical routing systems in the worldwide shipping market.

BonVoyage 3500+ http://www.appliedweather.com

SPOS 2500+ http://www.meteogroup-maritime.com

Seaware unknown http://www.seaware.se

OCTOPUS unknown http://www.amarcon.com

SeaRoutes unknown http://www.searoutes.gr

26 Report No. 15:153