Analyzing the economic benefit of unmanned autonomous ships: An exploratory cost-comparison between an autonomous and a conventional bulk carrier

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Research in Transportation Business & Management journal homepage: www.elsevier.com/locate/rtbm

Analyzing the economic benefit of unmanned autonomous ships: An exploratory cost-comparison between an autonomous and a conventional bulk carrier Lutz Kretschmann⁎, Hans-Christoph Burmeister, Carlos Jahn Fraunhofer Center for Maritime Logistics and Services, Am Schwarzenberg-Campus 4, Building D, 21073 Hamburg, Germany

A R T I C L E I N F O

A B S T R A C T

Keywords: Unmanned autonomous ships Autonomous navigation Cost of shipping Economics of shipping Comparative cost-benefit analysis Economy of unmanned autonomous ships

Unmanned autonomous ships are seen as a key element of a competitive and sustainable European shipping industry in future. But even if the technology to further automate ships will principally be available at some point, this does not imply that autonomous vessels are also the superior choice for the ship owner. In the end the success of autonomous vessels depends on their impact on the profitability of shipping companies. Following a structured approach this paper analyzes the costs of running an autonomous bulker and compares them against a conventional vessel in a cost-benefit analysis. Hereby it provides insights on the (economic) benefit of autonomous vessels for a first-time. Results principally confirm an economic potential. The expected present value of cost of owning and operating the autonomous bulker over a 25-year period is mUSD 4.3 lower than for a conventionally manned ship. Assuming identical cargo carrying capacity this means that the required freight rate of the autonomous bulker which produces a zero net present value is 3.4% lower than the required freight rate of the conventional vessel. This advantageousness is based on one aspect in particular as the paper argues. Besides cost savings associated with reducing crew levels an autonomous ship brings along additional benefits due to changes in ship design.

1. Introduction Shipping is one of the few truly global industries. In principle, competition between all states and nations is possible in maritime international trade due to the freedom of the seas and international regulations. In terms of making international shipping more efficient, several fields of innovation showed a rapid development in the recent past (see Table 1). With the advent of autonomous technology in train and car industries in the last decade, unmanned and autonomous vessels have become of popular interest in latest maritime research and innovation studies as well. The fundament of this kind of vessel lies within the European Waterborne Technologies Platform (Waterborne TP) Implementation Plan from 2011, which defines an autonomous vessel as a ship with “next generation modular control systems and communications technology [that] will enable wireless monitoring and control functions both on and off board. These will include advanced decision support systems to provide a capability to operate ships remotely under semi or fully autonomous control” (Waterborne TP, 2011). Against this background, the intention behind the development of

autonomous vessels is to contribute to all main dimensions of sustainability (e.g. Burmeister, Bruhn, & Rødseth, 2014; Rødseth & Burmeister, 2012): − Economic sustainability by keeping operational expenses low, especially crew-related costs, to facilitate efficient international trade, − Ecological sustainability by enabling new and innovative ways to reduce overall fuel consumption e.g. due to the absence of lifesupport systems on board, and − Social sustainability by increasing safety due to moving trivial operational tasks from fatigue crew to onboard automation and by enabling shore-based and family friendly monitoring jobs for nautical personnel ashore. Of course, individual developments in waterborne transport often relate to more than one field of innovation making a clear attribution difficult. The concept for an autonomous vessel is a good example. Besides a higher automation on board it also covers several aspects closely related to the intelligent ship such as optimized (weather)

⁎

Corresponding author. E-mail addresses: [email protected] (L. Kretschmann), [email protected] (H.-C. Burmeister), [email protected] (C. Jahn). http://dx.doi.org/10.1016/j.rtbm.2017.06.002 Received 21 October 2016; Received in revised form 11 April 2017; Accepted 16 June 2017 2210-5395/ © 2017 Elsevier Ltd. All rights reserved.

Please cite this article as: Kretschmann, L., Research in Transportation Business & Management (2017), http://dx.doi.org/10.1016/j.rtbm.2017.06.002

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

associated with unmanned autonomous ships. The commercial attractiveness for pure retro-fitting of existing vessels might not be given (Rødseth & Burmeister, 2015a). Thus, the paper focuses on the economics of newly designed autonomous ships according to the MUNIN concept. The analysis primarily looks at benefits associated directly with unmanned ships. Benefits associated with ship intelligence, even though part of the MUNIN project to some point, are not considered since they are available for both conventional vessels and unmanned autonomous ships. Further the analysis focusses on costs of owning and operating a ship while revenue is not considered. The analysis in this paper deals primarily with technical and commercial aspects of unmanned autonomous ships. However, it should be noted that unmanned autonomous ships don't appear to pose an unsurmountable substantial obstacle in legal terms and bring along the potential to increase maritime safety. In this context the interested reader is referred to e.g. IMO (2015), Kretschmann et al. (2015a, 2015b), Rødseth and Burmeister (2015a), Safari and Sage (2013), Ringbom, Felix, and Viljanen (2016) and Van Hooydonk (2014), where also effects of unmanned ships on flag state regulations and registrations are discussed. The remainder of this paper is structured as follows. To give an a priori understanding of the subject of investigation, Section 2 outlines the principal concept and operations of the autonomous vessel as developed in MUNIN. Subsequently the methodology of the analysis is introduced in Section 3 and a reference cost model for a conventional bulker is described in Section 4. Section 5 identifies changes in cost of owning and operating an autonomous vessel and gives quantitative estimations of their extent. Finally Section 6 outlines results of the comparative cost-benefit analysis for three scenarios and conclusions are given in Section 7.

Table 1 Selected fields of innovation in waterborne transport. Source: authors. Field of innovation Efficient ship: innovations focusing on efficiency improvements Intelligent ship: innovations that make use of evermore data generated on board in smart applications Autonomous ship: innovations that aim for a higher degree of automation on board

Related to e.g. – – – – – – – – – –

Hull form optimization Energy-saving devices Machinery technology Optimized (weather) routing On-board energy efficiency management Voyage performance management Condition monitoring and management Reduced crew New ship designs Improved safety

routing or on-board energy efficiency management. For a brief review of the history on unmanned and autonomous vessel projects, it is referred to Bertram (2016). Nowadays there are several public-funded research projects in Europe aiming to develop appropriate technology and business concepts that can be mainly attributed to unmanned and autonomous vessels, e.g.: − AAWA - The Advanced Autonomous Waterborne Applications Initiative: A Finish funded project led by Rolls-Royce investigating principle factors and designs enabling autonomous ships (RollsRoyce, n.d). − ReVolt: A Norwegian-funded concept study conducted by DNV-GL about an unmanned and battery powered short-sea container vessel for the Norwegian fjords (DNV GL, 2015). − MUNIN – Maritime Unmanned Navigation through Intelligence in Networks: A European FP7 funded feasibility study about unmanned dry bulk shipping in deep-sea (Fraunhofer CML, n.d.).

2. MUNIN concept of the unmanned autonomous vessel The use case investigated in MUNIN is a partly unmanned dry bulk carrier. In the principal operational concept of MUNIN, the vessel is fully unmanned during deep sea voyages, which is operated by autonomous onboard navigation and lookout systems while being monitored from a shore side control center (see also Fig. 1). In the taxonomy of MUNIN, the above described vessel is part of the class of Unmanned Maritime System, as it is a “maritime system that operates full or part time without humans in direct control” (Rødseth, Tjora, & Burmeister, 2014). Only during port approach and berthing an onboard control team is on the vessel and directly operates it from the bridge. During the main deep-sea leg, an Autonomous Navigation System acts as the officer-ofthe-watch with regards to operative decision-making, while the lookout

While the first project is basically about technical design according to the project description, the latter two also include research on the commercial feasibility of the developed concepts. This represents a key issue in the context of industrial autonomous system design, as one of the first questions asked is going be “Can the [vehicle] do the job, and if so, at a lower cost?” (Stokey et al., 1999). In other words: unless an innovation is commercially viable it will not find its way into practice. Surprisingly, so far little attention has been given to the economy of unmanned autonomous ships (Bertram, 2016). Against this background the paper intends to open up the discussion on economic issues

Fig. 1. MUNIN vision. Source: Burmeister, Bruhn et al., 2014.

2

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

− Operating costs are all expenses to keep the ship in an operational status incl. periodic maintenance. For the ship owner these costs are fixed (unless the ship is laid up) and independent of a particular voyage the ship may or may not be trading in. − Voyage costs, on the contrary, are variable costs directly associated with a particular deployment of the vessel and include port call cost and cost for loading and unloading cargo. Thus, they only occur if a ship enters a particular voyage and can be directly attributed to that voyage. − Capital costs are all expenses related to the purchase of a ship and associated costs (e.g. interest).

task is performed by an Advanced Sensor Module (see e.g. Burmeister, Rødseth, Porathe, & Bruhn, 2014; Burmeister, Bruhn et al., 2014). Notwithstanding the autonomous technology onboard, the shore control center is normally also connected to the vessel. In contrast to today's operation, the humans in the shore control center are not executing a direct remote control, but only monitoring a proper execution of the vessel's autonomous systems according to the predefined voyage plan (Rødseth, Kvamstad, Porathe, & Burmeister, 2013). For more details about the operation and system architecture, it is referred to Rødseth and Burmeister (2015a) and Burmeister, Rødseth et al. (2014). Due to the MUNIN concept's inherent cooperation of unmanned and manned operations, focusing on dry bulk voyages is very promising with regards to commercial feasibility, as those trades typically consist of long voyages with only one loading and discharging port, while at the same time the cargo doesn't need regular intervention while underway (Rødseth & Burmeister, 2012). Further, efforts for transferring a bulker - a rather simple type of vessel - to autonomous operation are presumed to be comparatively low. Taking into account the above, the analysis of cost of running a ship in this paper primarily addresses changes in crew placement and technical system architecture.

Following a structured approach the analysis calculates the costs of running an autonomous bulker and compares them against a conventional vessel. The methodology of this comparative cost-benefit analysis is as follows: − A cost model for a conventional manned bulker - representing the as-is-processes and technical systems - is developed based on typical cost figures for operating a bulk carrier under a characteristic yearly operating profile. − Changes in operating, voyage and capital costs for an autonomous bulker are identified. Appropriate methods are applied to come up with a quantitative estimation of the extent costs can be expected to change. − Taking into consideration the identified cost changes a cost model for the autonomous bulker is developed. − Both models are used to calculate the expected present value (EPV) of cost of owning and operating a conventional and an autonomous bulker over their operational lifetime.

3. Methodology of the analysis Unmanned autonomous vessels have to compete with conventional manned ships and thus in general the same principal economics apply. In order to assess the economics of autonomous ships a slightly adapted version of the shipping cash-flow model defined by Stopford (2009) is used. It describes how revenue is generated by a ship and after costs are deducted creates free cash flow which is used to cover taxes, pay dividends and generate a profit for the ship owner (see Fig. 2). In regards to ship revenue, it is fair to assume that the freight rate is set exogenous and predominantly influenced by commodity price (see e.g. Corbett, Wang, & Winebrake, 2009; Hummels, Lugovskyy, & Skiba, 2007) while productivity and cargo capacity can be considered to be equal for autonomous and manned bulkers (Kretschmann et al., 2015b). Accordingly the first main assumption underlying the analysis is that both an autonomous and a conventional bulker have the same potential to generate revenue over their operating life. Of course a certain risk element is associated with the operation of autonomous ships which might have a financial impact. Quite likely the risk profile of autonomous ships will be different from conventional ships. However, it will not necessarily be larger as risks can be reduced to an acceptable level with proper technical solutions. Furthermore, on autonomous vessels it will be possible to control the factor causing more than two third of maritime accidents - human error - much better (Sanquist, 1992). Nonetheless a key element of developing the MUNIN concept has been to ensure that the unmanned ship systems can autonomously sail on an intercontinental voyage at least as safe and efficient as manned ships (Rødseth & Burmeister, 2015a). Accordingly the second main assumption underlying the analysis is that unmanned vessels and manned vessel maintain an equal safety level. Considering the above the main focus of the paper are the costs of owning and operating a ship. The following cost categories are considered (Neylan, 2011; Stopford, 2009):

In this context the EPV of costs is defined as: N

EPV of costs =

1

In order to take time value of money into account the discount rate is set at 8% which is slightly below typical weighted average cost of capital in the shipping industry (IMO, 2011). The operational lifetime is the second important parameter. Demolition ages vary over time. They are primarily influenced by vessel age, technical developments, regulatory changes, prices of scrap metal, current market state, and market expectations (Stopford, 2009). Here an operational lifetime of 25 years is chosen for both vessels respectively which is in line with demolition ages currently seen in dry bulk markets (Sand, 2016). Besides the EPV of costs the required freight rate (RFR) will be used as another measure to compare both vessels. The RFR is that freight rate which produces a zero net present value over the assumed life of the ship (Watson, 1997): N

RFR =

Operating Costs

PW(cost of owning and operating vessel) cargo tonnage

Dividing the EPV of costs of the unmanned vessel by the EPV of costs of the conventional vessel shows how much lower the RFR of the autonomous bulker which produces a zero NPV over the assumed life would be under the (previously discussed) assumption that both vessels transport an equal amount of cargo. This measure shall be referred to as the relative RFR. Using the described methodology results for three different scenarios are calculated and discussed:

Free cash flow

Capital Costs

∑ 1

Ship Revenue Depends on: 1. Cargo capacity 2. Productivity 3. Freight rates

∑ PW(cost of owning and operating vessel)

− Scenario A: only the effects of a reduced crew on board and new shore/port services are analyzed. − Scenario B: corresponds to Scenario A and additionally considers improved fuel efficiency of the autonomous vessel. − Scenario C: corresponds to Scenario B but the autonomous vessel

Voyage Costs

Fig. 2. Shipping cash-flow model. Source: Stopford, 2009.

3

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

voyage costs. Fuel prices used in this analysis are based on the forecast of the crude oil price published by the IEA World Energy Outlook. In the latest version from 2015 a price of USD 80 per barrel is predicted for 2020 (Birol, 2015). Crude oil price is converted into HFO and MDO prices based on the average price ratio between them over the period from 2005 to 2014 (see Fig. 3). Over the considered period HFO was trading at about 70% of the price of crude oil while MDO had cost about 115% of the price of crude oil. Applying this ratio to the predicted price of crude oil (USD 80 per barrel) corresponds to a HFO fuel price of USD 422 per tonne and a MDO fuel price of USD 693 per tonne. It should be mentioned, that there is no certainty that the price ratio seen in the past will continue in future. E.g. global Sulphur limits could make HFO more expensive compared to crude oil. However, a more detailed forecast of marine fuel price is out of scope here.

Table 2 Ship particulars of the reference panamax bulker. Source: authors. Ship particular

Value

Length over all Breadth Design draught Service speed Displacement Main engine

230 m 32 m 14,5 m 15,5 kn 90,600 t 10,230 kW

uses Marine Diesel Oil (MDO) as main fuel while the conventional bulker continues to use Heavy Fuel Oil (HFO) as main fuel. Additionally a sensitivity analysis is carried out to determine the impact of important input variables on the relative RFR.

4.2.2. Fuel costs – main engine Besides fuel price the ships fuel consumption, its operational profile and the type of fuel are required to calculate yearly fuel costs. Fuel consumption is determined for a standard motor type of vessels comparable to the reference vessel (MAN Diesel & Turbo, 2014a). With an operating point at 85% Maximum Continuous Rating (MCR) - at service speed and loaded - and including a reasonable mark-up to correct for test bench conditions a fuel consumption of 182.5 g/kWh is calculated (MAN Diesel & Turbo, 2014b). For the vessel sailing at service speed under loaded conditions this corresponds to a fuel consumption of 38 tons per day which is in line with fuel consumption values found in literature (see e.g. Neylan, 2011). In ballast condition the fuel consumption is somewhat reduced (26 tons per day). The yearly operational profile was already introduced in Section 4. In sea passage the vessel uses HFO and in the state “ship maneuvering” MDO. This way, developments towards stricter environmental standards which require high grade fuel - particularly close to shore in Sulphur Emission Control Areas - are taken into account. The IMO MARPOL Convention recognizes several of these areas – currently in the Baltic Sea, North Sea, North American, and United States Caribbean Sea area – to minimize airborne emissions from ships primarily close to shore (IMO, 2017). Vessels operating within a Sulphur Emission Control Area are required to use marine fuels with a maximum content of 0.1% Sulphur (in general a requirement MDO fulfils but HFO does not), or implement appropriate abatement technologies on board (Ballini & Bozzo, 2015). All things considered, fuel consumption of the main engine accounts for 67% of voyage cost in the reference case.

4. Reference cost model for the conventional bulker The reference vessel for the analysis is a panamax bulk carrier. Its bridge is conventionally manned at all times while the engine room is usually only manned during daytime. Maintenance work is carried out continuously by the crew. The main engine is fueled with HFO and drives a single fixed propeller. Auxiliary engines run on MDO. Main ship particulars of the reference vessel are summarized in Table 2. In order to determine the cost of running the ship - both for the conventional and the autonomous bulker - an operating profile is required. Data for a collection of bulk carriers with similar specifications as the reference vessel was used to define the yearly operational profile (see Table 3). “In sea passage” state is the path between pilot points. Two-thirds of days in sea passage are assumed to be in loaded condition while the remaining one-third is in ballast condition. “Ship at berth or waiting” is ship not moving. “Ship maneuvering” is a collection of all other states. 4.1. Operating costs Operating costs are different for every ship depending on, amongst others, company policy, flag, ship type, and age. In this analysis statistics on average operating costs for a large number of panamax bulkers forms the basis for defining the reference structure of operating costs. Table 4 gives an overview of the considered costs. Even though periodic maintenance is usually not included in operating costs it is shown in the table as well. 4.2. Voyage costs

4.2.3. Fuel costs – auxiliary engines For reasons of comparability both the autonomous and the conventional bulker are assumed to use diesel generator sets for auxiliary energy supply. Associated fuel costs are calculated via the specific fuel consumption of the diesel generators, the electric energy consumption on board, and the fuel type used to run the generators. The specific fuel consumption is determined for the standard type generator set proposed in the MUNIN general technical system redesign (Wehner, Schmidt, & Fentzahn, 2013). For an operating point at 85% MCR and including a reasonable mark-up to correct for test bench conditions a specific fuel consumption of 213.1 g/kWh is calculated (MAN Diesel & Turbo, 2014c). A distinction between different operating points is not made. Based on data collected by the California Air Resources Board the average auxiliary to propulsion ratio for bulk carriers is 0.22 (Lindhjem & Browning, 2007). Thus, the power of the reference bulker's auxiliary engines is assumed to be 2271 kW. Auxiliary engine load factors are used to come up with electric energy consumption on board for different operational states (see Table 5). The diesel generators use MDO. Based on these considerations, the fuel consumption for auxiliaries

Considered voyage costs are fuel costs of the main engine and the auxiliary engines as well as port call costs. In order to calculate fuel cost the fuel price has to be determined first. 4.2.1. Fuel price Fuel costs alone can represent between 50% and 70% of the total cost of owning and operating a ship. Accordingly changes in fuel prices have a huge impact on the costs of operating a vessel and future fuel price developments are the number one uncertainty when determining Table 3 Yearly operational profile. Source: authors based on data from Atlason and Jónsson (2014). State

Days per year

Ship at berth/waiting Ship maneuvering Ship in sea passage

120 days 29 days 216 days

4

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

Table 4 Operating costs of reference vessel. Source: authors based on data from Neylan (2011), Greiner (2011), E.R. Schiffahrt (2015). Cost type

Includes

Costs per year on average

Change over lifetime

Share of operating costs

Crew costs Stores & consumables Maintenance & repair Insurance General costs Periodic maintenance

Direct and indirect costs (e.g. wage, travel, victualling, etc.) General stores (e.g. deck, cabin, engine) and lubricants Routine basic work on-board & more complex work during port calls Hull & Machinery, Protection & Indemnity Expenses associated with administration & management Dry dockings

USD USD USD USD USD USD

Slight increase Slight increase Slight increase Slight decrease Constant Large increase

45% 14% 13% 15% 13% –

USD per tonne

945,000 289,000 268,000 313,000 269,000 420,000

m USD 100

1.200

Capesize Panamax

1.000

80

Handysize

800

60 600

40 400

MDO Crude Oil

200

20

HFO

0

0 2005

2007

2009

2011

2013

2002

2004

2006

2008

2010

2012

Fig. 3. Crude oil and marine fuel price development, 2005 to 2014. Source: authors based on data from MTNZ, EIA, Insee.

Fig. 4. Bulker new building prices, 2002 to 2013. Source: authors based on data from EquityGate Advisors (2013).

Table 5 Electric energy consumption for different operational states. Source: Lindhjem & Browning, 2007.

vessel. They include the new building price, cost of financing and a payment received upon the sale of the vessel. For the ship owner – besides own funds invested up front –they typically come in the form of regular payments of interest and redemption. However, in this analysis capital costs shall be treated as a one-time payment at the commissioning of the ship. Thus, capital costs can be understood as the discounted value of all payments associated with the purchase (and sale) of the vessel. To estimate the capital costs for the reference vessel new building prices of panamax bulk carriers are referred to. Fig. 4 shows the development of new building prices for capesize, panamax and handysize bulk carriers between 2002 and 2013. The average new building price for a panamax bulker during that period was mUSD 34. This amount is chosen as capital costs for the conventional bulker in the analysis. While it can be argued that capital costs of mUSD 34 are rather high it is reasonable here since additional costs, such as cost of financing, are not considered separately.

Aux. engine load factor Energy consumption

Theoretical maximum

Ship in sea passage

Ship maneuvering

Ship at berth/ waiting

1

0.17

0.45

0.22

2271 kW

386 kW

1022 kW

500 kW

in sea passage is about two tons MDO per day. All things considered, auxiliary engine fuel consumption accounts for 10% of voyage costs in the reference case. 4.2.4. Port call costs Port call costs include different fees and charges associated with services the vessel receives in port. The level of port call costs is primarily determined by the pricing policy of the port. Charging practices vary considerably from one port to another. Accordingly, it is difficult to come up with a general estimation of port call costs without specifying a set of voyages and thus ports of call. Typical port call costs found in literature are USD 147,000 for a panamax bulker trip from Australia to Europe and a range of USD 35,000 to USD 40,000 per call for bulk carriers respectively (DA-Desk, n.d.; Stopford, 2009). USD 100,000 is chosen as the cost per port call in this analysis. With an average voyage length of 14 days the vessel has 15 port calls per year. This means that port call costs account for 23% of voyage costs in the reference case.

4.4. Overview of reference bulker costs Fig. 5 shows all costs incurred over the lifetime of the reference 20% Reference 10 Vessel

16

53%

46

26%

7

16

34

129

mUSD Operating costs (crew)

4.3. Capital costs

Voyage costs (fuel aux. engine)

Operating costs (other)

Voyage costs (port call)

Voyage costs (fuel main engine)

Capital costs

Fig. 5. EPV of costs over lifetime for reference bulk carrier. Source: authors.

Capital costs are all expenses associated with the purchase of the 5

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

5.1.3. Maintenance crews in port On the autonomous vessel a boarding crew is responsible for maintaining the propulsion plant, auxiliary plants, supply systems, electrical and automation systems, etc. It is assumed that a crew consisting of nine engineers and technicians carries out the work during the time the ship is at berth or waiting (120 out of 365 days) (see Kretschmann et al. (2015b) for details). Associated costs are estimated to be around USD 135,000 per vessel per year (plus 15% for miscellaneous costs and profit).

Table 6 Considered cost changes. Source: authors. Operating costs

Voyage costs

Capital costs

Crew wages (−) Crew related costs (−) Shore control center (+) Maintenance crews (+)

Air resistance (−) Light ship weight (−) Hotel system (−) Boarding crew for port calls (+)

Deckhouse (−) Hotel system (−) Redundant technical systems (+) Autonomous ship technology (+)

5.2. Voyage costs Minus (−) represents a reduction of costs; plus (+) an increase.

5.2.1. Reduced fuel consumption Several aspects which could contribute to a higher fuel efficiency of autonomous vessels are discussed by experts in the field (see e.g. Jokioinen, 2016). This analysis considers reduced air resistance, lower light ship weight and the hotel system on board. All effects taken together fuel consumption of the autonomous vessel is reduced by approximately 6%. Compared to other estimations of potential fuel savings for autonomous ships – e.g. 12 to 15% (Arnsdorf, 2014) – this is a rather conservative estimate. Again, it shall be stressed that only effects directly related to an autonomous ship are considered here while possible ship intelligence effects are excluded. However, such ship intelligence effects may be included in suggested fuel savings elsewhere.

bulker discounted to the time of commissioning of the ship. Total EPV of costs is mUSD 129. Voyage costs account for the largest part (53%) followed by capital costs (26%). Operating costs make up the remaining 20%. 5. Considered cost changes for the autonomous bulker An obvious cost saving potential of autonomous vessels is reducing operating cost related to the crew on board. On the other side e.g. additional technology that enables the autonomous functioning of the vessel and staff in a shore control center which monitors the autonomous voyage increase costs. Table 6 summarizes all cost changes considered in this analysis.

5.2.2. Air resistance Autonomous ships no longer need to support a crew living on board and are not bound to minimum sight restrictions from the bridge. This makes new ship designs possible which do not include a deckshouse structure as it is found on conventional vessels today. Removing the deckshouse reduces air resistance and thus improves fuel efficiency. In calm conditions air resistance is basically a function of the ships speed and the surface area exposed to the wind above the waterline. It typically represents about 2% of the total resistance of a vessel but potentially much more in head winds (MAN Diesel & Turbo, 2011). Without a deckshouse structure the surface area exposed to wind is reduced. Accordingly propulsive power and fuel consumption are lower. To estimate the effect, the frontal air resistance with and without the deckshouse is calculated. Wind resistance coefficients are adopted from Blendermann (1996). cd = 0.68 (specified for a tanker) is used for the reference vessel and cd = 0.45 (specified for a car carrier with a closed fore section) for the autonomous vessel. Wind area frontal for the reference panamax bulk carrier is approximately 422 m2 in design condition and 617 m2 in ballast condition. Wind area frontal of the deckshouse alone is calculated to be 313 m2 following the MüllerKöster method (Bertram & Schneekluth, 1998). All things considered, the required propulsion power without deckshouse is 1% lower (at service speed in loaded condition, no additional true wind). Reductions in fuel consumption are assumed to be proportional.

5.1. Operating costs 5.1.1. Crew wages and crew related costs A crew is no longer required on board the autonomous bulker. This results in average savings of USD 945,000 per year for crew wages and related costs (see previous section). Additional cost reductions are expected since the crew is no longer living on board for extended periods of time. This includes costs for general stores (e.g. medical, cabin, safety equipment), maintenance of e.g. life rafts and spares for the hotel system. Together this amounts to an additional estimated cost reduction of USD 67,000 per year on average. 5.1.2. Shore control center The autonomous vessel requires certain shore and port based service which conventional ships do not require. In particular this includes the shore control center and maintenance crews that conduct necessary repairs while the ship is in port. The MUNIN project developed an organizational layout of a shore control center which monitors 90 vessels at a time (Porathe, Costa, & Tjora, 2014). The concept is the starting point for estimating personnel, operating and investment costs. A detailed calculation of the cost of running the shore control center was carried out as part of the MUNIN project (Kretschmann et al., 2015b). According to the report, overall personal costs of the shore control center are mUSD 10.4 per year which corresponds to USD 116.000 per vessel. 15% is added to this figure for miscellaneous costs and profit of the company offering the service. Besides personnel costs there are several investment and operating costs associated with setting up and running the shore control center. A detailed study was done as part of the MUNIN project. It considers onetime costs for equipment (e.g. situations rooms, software, hardware, office equipment) and annual costs in terms of rent and operational costs (e.g. power supply, software, training costs) (Kretschmann et al., 2015b). All things considered, investment costs add up to mUSD 2.1 and operating costs per year are around USD 875,000. Since several ships are monitored, expenses per vessel are reduced accordingly. 15% is added for miscellaneous costs and profit of the operating company in this analysis.

5.2.3. Light ship weight Innovative autonomous ship designs without deckshouse will be characterized by a reduced light ship weight. This in turn has a positive impact on fuel consumption. For a panamax tanker the American Bureau of Shipping indicates a 0.34% reduction of fuel consumption for a 1% change in steel weight due to possible adjustments of the block coefficient while deadweight is maintained constant (ABS, 2013). The light ship weight of the reference bulker is approximately 12,000 t (see e.g. Mikelis, 2012). Details of a ship's light weight by individual ship sections are rarely published. Thus, in order to come up with a reasonable estimation of the reduced light ship weight of an autonomous bulker, approximation methods are applied (Bertram & Schneekluth, 1998). Steel weight of superstructure and deckhouse is calculated to be 430 t. Further, equipment and outfitting of the living quarters (e.g. walls insulation, sanitary installations, 6

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

do not require and certain costs for the ship owner are associated. MUNIN has developed several ideas how boarding crews might be dispatched in future. Options include e.g. a transfer of crew by helicopter or transfer vessel or a remote control of the vessel from a pilot boat (Porathe, 2013). However, these concepts for crew transfer and manning during approaching and berthing are not mature enough to come up with sensible cost estimations. Instead it is assumed that port call costs are 20% higher for the autonomous vessel which corresponds to an additional USD 20,000 per call. To put this into context: a small helicopter deployed in crew transfer for offshore wind parks in Europe can costs from USD 1300 to USD 5000 per hour (Jahn & Münsterberg, 2015). Besides planned assistance by boarding crews for approaching and berthing, unforeseen events during the journey of the vessel would require other forms of assistance. One example is deep sea towing due to a failure of the main engine. However, cases of unscheduled assistance are not discussed in depth here. They are expected to be rather rare events which in the long run will be covered by insurance or P & Ilike instruments.

Table 7 Electrical power balance for a container vessel. Source: authors based on Mau (1984). Consumer

Auxiliary systems for propulsion service Auxiliary systems for ship operation Heating ventilation air conditioning Galley and laundry Deck machinery Ventilation cargo space Lighting Other auxiliary systems Total connected load

Nominal power Total

At sea

At sea without crew

Assumed reduction

1168 kW

403.9 kW

403.9 kW

–

142.8 kW

76.6 kW

76.6 kW

–

374.3 kW

309.3 kW

0

100%

178.6 kW 609.5 kW 49.6 kW 91 kW 42.2 kW 2656 kW

138.4 kW 137.5 kW 43.5 kW 81 kW 37 kW 1227.2 kW

0 137.5 kW 43.5 kW 40.5 kW 37 kW 739 kW

100% – – 50% – 40%

kitchens, and furniture) are considered. As a rule the weight of the accommodation area is proportional to the volume of the deckshouse. Here 70 kg/m3 is assumed for a deckshouse volume of 4690 m3. Accordingly weight of equipment and outfitting in the accommodation and living quarters is 328 t. Weight of additional miscellaneous systems (partly below the main deck) is expected to be another 152 t. This includes e.g. life rafts plus mountings, waste water treatment systems, air conditioning, fresh water tanks as well as associated auxiliary systems. All things considered, this corresponds to a reduction of the light ship weight by 7.6% compared to the reference vessel. Applying the factor between light ship weight and fuel consumption introduced earlier it results in a reduction of the fuel consumption by 2.6%.

5.3. Capital costs Capital costs from the perspective of the ship owner include all expenses associated with the purchase of the vessel. Besides cost of financing they are primarily determined by the new building price of the vessel which in turn is linked to the production costs at the shipyard (plus a profit margin). Since the new building price is influenced by market forces (and thus difficult to determine) production costs are a better indicator to estimate capital costs. Two considerations are essential to identify how production costs for an autonomous vessel differ from a conventional bulker. On the one hand, systems compulsory on a conventional ship are no longer required on an autonomous ship (e.g. deckhouse and hotel system). On the other hand, the autonomous vessel requires new autonomous ship technology and, most likely, some systems on board will need to be redundant to ensure that the vessel is at least as safe as a manned ship.

5.2.4. Hotel system If the crew is no longer living on board the autonomous vessel for extended periods of time there is no need to have a fully equipped hotel system. Removing the hotel system also means that the electric power consumption is reduced. In order to identify what part of total electric power demand on board is associated with the hotel system an electrical power balance of a container ship is shown in Table 7. In the last column consumers presumed to be directly related to the crew living on board are subtracted from nominal power at sea. This way the total connected load for the vessel decreases by 40% (from 1227 kW to 739 kW). It is unlikely that this reduction of total connected load (in kW) is directly applicable to a bulk carrier. However it is reasonable to assume that the relative reduction of electrical power consumption for the autonomous bulker is in the same order of magnitude. Accordingly electric energy consumption of the autonomous bulker in sea passage is expected to be 60% of the electric energy consumption calculated for the conventional bulker previously. This corresponds to a reduction by 154 kW. For the other operational statuses the same reduction is assumed. All things considered, auxiliary engine fuel consumption can be reduced by about 33%.

5.3.1. Deckhouse and hotel system Material and production costs of the autonomous vessel will be lower if it does not have a deckhouse and hotel system as they are found on today's ships. Several methods are discussed in literature to estimate the construction costs of a ship (e.g. Caprace & Rigo, 2009; Ross, 2004). A top-down cost estimation approach, typically used in a rather early stage of vessel design, would principally lead to the desired result. Unfortunately, no complete ship design for the MUNIN project's autonomous vessel is available. This limits the applicability of top-down cost estimation methods. Instead, and in order to give an indication how capital costs of the autonomous ship will likely differ from a conventional bulker, typical values of the distribution of building costs for different technological or weight groups are referred to (see e.g. Papanikolaou, 2014; Shetelig, 2013). Costs of the hotel and accommodation section represent approximately 5% of the total costs. In case a hotel and accommodation section is no longer necessary (or only a significantly reduced version to support boarding crews for short periods of time) the production costs are reduced accordingly. The deckshouse – even though it has a higher complexity than the main hull which increases production costs (Lamb, 2004) – is only responsible for a small part of the hull weight. Removing it might reduce overall production costs by another couple of percent.

5.2.5. Boarding crew for port calls The initial idea of the MUNIN project envisioned a boarding crew from a local departure and approach service to control the vessel between berth and pilot point. Thus, approaching and berthing is still executed by a conventional crew on board. In case a route includes a canal passage a boarding crew controls the vessel as well. According to the MUNIN concept, it is the owner's obligation to arrange the local crew. However, in the long run a port or canal authority might be willing to offer such a service (similar to the canal helmsmen in Kiel Canal). Of course boarding crews are a service which conventional vessels

5.3.2. Autonomous ship technology and redundant technical systems For safe operation the autonomous vessel has to be equipped with specific autonomous ship technology (e.g. Autonomous Navigation System, Advanced Sensor Module) and have a certain degree of redundancy (e.g. communication, electrical system, propulsion). Both increases the production costs of the vessel and accordingly capital 7

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

scenario. This corresponds to a relative RFR of 99.6%. Accordingly the RFR of the autonomous bulker which produces a zero NPV is 0.4% lower than the RFR of the conventional vessel.

costs. First concepts for autonomous ship technology have been developed as part of the MUNIN project. Due to the innovative nature of this technology and the very early stage of development, coming up with detailed calculations of their cost is hardly possible at this point. Based on a good first estimate they might add 5%, or mUSD 1.7, to the costs of the vessel. The picture is somewhat similar regarding additional cost for redundant technical systems on board. A detailed estimation would need to be based on a full ship design which is not available. However, looking at twin screw ship designs allows drawing some conclusions. The main argument for twin screw tankers is that they would reduce machinery failures – an important cause of oil spillage – due to fully redundant engine rooms. Large twin screw tankers are estimated to cost approximately 5 to 8% more than equivalent single screw ships (Devanney, 2008). Concluding by analogy and being somewhat conservative, the redundancy of technical systems on board the autonomous bulker can be expected to increase cost by 10%. All things considered, production costs of an autonomous vessel are probably higher than for a conventional bulker but not significantly. However, throughout the MUNIN project some professionals have expressed the expectation that production costs for unmanned autonomous ships might even be lower than for a conventional manned vessel. Notwithstanding this, capital costs of the autonomous vessel are defined as 110% of the conventional bulker's capital costs in the base case of the analysis.

6.2. Scenario B: reduced crew and increased fuel efficiency The second scenario, ceteris paribus, considers the effect of improved fuel efficiency which is expected for the autonomous bulker. Fig. 7 includes a short scenario description and shows the impact of different effects on the EPV of costs. Over a 25-year period the EPV of costs of the autonomous bulker is mUSD 4.3 lower than the EPV of costs of the reference vessel in this scenario. This corresponds to a relative RFR of 96.6%. Accordingly the RFR of the autonomous bulker which produces a zero NPV is 3.4% lower than the RFR of the conventional vessel.

6.3. Scenario C: reduced crew, increased fuel efficiency and high grade fuel Autonomous operation using HFO as main fuel would be technically challenging. Thus a robust solution for an autonomous vessel is to use high grade fuels such as MDO instead of HFO. For this reason the technical redesign of the MUNIN Autonomous Engine Room was built around the assumption that the autonomous vessel uses MDO as main fuel (Wehner et al., 2013). At the same time, this has a large impact on voyage costs due to the higher price of MDO. The third scenario reflects the recommendations from the technical redesign to use MDO as a main fuel instead of HFO. The conventional bulker on the other hand continues to use HFO as main fuel. Fig. 7 includes a short scenario description and shows the impact of different effects on the EPV of costs. Over a 25-year period the EPV of costs of the autonomous bulker is mUSD 19.2 higher than the EPV of costs of the reference vessel in this scenario. To put this into context: the price difference between MDO and HFO would need to reduce to about 12% to justify an investment in the autonomous bulker (in terms of a breakeven of the EPV of costs). With regards to freight rates the relative RFR in this scenario is 114.8%. Accordingly the RFR of the autonomous bulker which produces a zero NPV is 14.8% higher than the RFR of the conventional vessel.

6. Comparative cost-benefit analysis In this section the results of three different scenarios are presented which compare the autonomous vessel against the reference bulker. They make use of the cost models described before to calculate the EPV of costs over lifetime. Additionally a relative RFR is determined for each scenario. Fig. 6 gives a first overview of the respective results. 6.1. Scenario A: reduced crew The first scenario considers the effects of a reduced crew on board and additional costs for new shore/port services (identified changes in fuel consumption are not included). Thus, it helps to understand whether cost savings associated with removing the crew on board are sufficient to offset increases in new building costs and additional costs associated with new shore/port services. Fig. 7 includes a short scenario description and shows the impact of different effects on the EPV of costs. Over a 25-year period the EPV of costs of the autonomous bulker is mUSD 0.5 lower than the EPV of costs of the reference vessel in this 20% Reference 10 Vessel

53%

16

6.4. Sensitivity analysis A sensitivity analysis is carried out to determine the impact of different input variables on the relative RFR. Results are shown in a spider diagram in Fig. 8. Base case and starting point for the sensitivity analysis is the second scenario with a relative RFR of 96.6%. The following main input variables are changed on a percentage basis for the autonomous vessel only: fuel consumption, new building price, port call cost, and costs of new shore/port services. Fuel price and voyage length are changed equally for both vessels. The respective effect on the relative RFR is illustrated in the diagram. A reading below 100% indicates that the autonomous bulker is favorable in terms of lower cost over lifetime. The chart shows that the fuel consumption of the autonomous vessel has the largest impact on the relative RFR. Reducing the fuel consumption of the autonomous vessel by 10% would lower the relative RFR by 3.6 percentage points. The second largest impact has the new building price of the autonomous vessel. Here a 10% reduction lowers the relative RFR by 2.9 percentage points. Port call costs and costs of shore and port services have a somewhat lesser impact. Since fuel price and voyage length affect the EPV of costs of both vessels the effect on the relative RFR is smaller. Furthermore, the effect is reversed. A reduction in fuel prices increases the relative RFR, so does a reduction of the voyage length. Thus, autonomous vessels are more favorable the higher the fuel price and the longer the voyage between ports.

26% 7

46

16

129

34 3

S c enar io A

15

46

Scenario B

15

45

7

19

12 8

37 3

4

19

37

124 3

Scenario C 15

68

4

19

37

1 48 mUSD

Operating costs (crew)

Voyage costs (port call)

Operating costs (other)

Capital costs

Voyage costs (fuel main engine)

New shore/port services

Voyage costs (fuel aux. engine)

Fig. 6. EPV of costs over lifetime for reference and autonomous vessel. Source: authors.

8

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

110% of conv. Bulker

Main fuel type

HFO both vessels

Considered effects

Reduced crew New shore/port services

EPV costs, conv. bulker

128.7 mUSD

EPV costs, auton. bulker

128.2 mUSD

Relative RFR

99.6%

+3,4

Due to crew and -10,4 related expenses Due to new shore +3,3 & port services Due to additional port call expenses

+3,2

Due to better fuel efficiency

0,0

Overall

Scenario B: Main assumptions

126,2

121,2

123,7

-0,5 133,7

New building costs

Due to higher newbuilding price

131,2

Fuel price based on USD 80 per barrel of crude oil

Scenario A: Changes of the autonomous bulker’s EPV of costs:

128,7

Scenario A: Main assumptions

EPV [inmUSD]

Scenario B: Changes of the autonomous bulker’s EPV of costs:

Fuel price based on USD 80 per barrel of crude oil

EPV costs, conv. bulker

128.7 mUSD

EPV costs, auton. bulker

124.3 mUSD

Relative RFR

96.6%

Scenario C: Main assumptions

Due to additional port call expenses

+3,2

Due to better fuel efficiency

-3,8

Overall

-4,3 133,7

Reduced crew New shore/port services Better fuel efficiency

131,2

Considered effects

Due to new shore +3,3 & port services

128,7

HFO both vessels

+3,4

Due to crew and -10,4 related expenses

126,2

Main fuel type

Due to higher newbuilding price

123,7

110% of conv. Bulker

121,2

New building costs

EPV [in mUSD]

Scenario C: Changes of the autonomous bulker’s EPV of costs:

Fuel price based on USD 80 per barrel of crude oil

128.7 mUSD

EPV costs, auton. bulker

147.8 mUSD

Relative RFR

114.8%

+3,2

Due to fuel eff. & MDOasmainfuel

+19,7

Overall

-19,2 148,7

EPV costs, conv. bulker

Due to additional port call expenses

143,7

Reduced crew New shore/port services Better fuel efficiency

138,7

Considered effects

Due to new shore +3,3 & port services

133,7

Conv. bulker: HFO Auton. bulker: MDO

Due to crew and -10,4 related expenses

128,7

Main fuel type

+3,4

123,7

110% of conv. Bulker

118,7

New building costs

Due to higher newbuilding price

EPV [in mUSD]

Fig. 7. Main assumptions and EPV over lifetime in considered scenarios. Source: authors.

7. Conclusion

reference vessel. First, a cost model for a conventional bulker - representing the as-is-processes and technical systems – is developed. Next, changes in operating, voyage and capital costs for an autonomous vessel are identified and estimated quantitatively. This allows

This paper analyzes the costs of owning and operating an autonomous bulk carrier and compares them to a conventionally manned 9

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

104%

Fuel Consumption* New Builidng Cost* Port Call Cost* Shore and Port Services* Fuel Price** Voyage Length**

102%

Relative RFR

100% 98%

* Altered only fora utonomous vessel ** Altered for both vessels

96% 94% 92% 90% 88%

-20%

-10%

Base Case

+10%

+20%

Value relative to base case (Scenario B) Fig. 8. Spider diagram – impact of input variables on relative RFR. Source: authors.

results are still associated with a certain degree of uncertainty – mainly due to the early stage the development towards autonomous vessels is in. Several unknowns remain e.g. related to the need for extensive sensors on unmanned ships and the level of functional redundancy. Against this background the paper intends to open up the discussion on economic issues associated with unmanned ships. It does not aim to fully explore all dimensions of the questions at hand. Some important fields for further research are discussed below. Currently the MUNIN concept study on an autonomous bulker is the only publicly available source which provides comparatively detailed technical information. Other applications for unmanned ships are either mere thought experiments or still in a (very) early stage of development. Hopefully, more autonomous ship designs and concepts on a sufficiently detailed level will be developed over the coming years. As these become available, they should be the basis for further investigations into the economic benefit of autonomous ships. Most importantly, detailed designs would allow more precise estimations of fuel efficiency gains and changes in new building cost. This is crucial - as the sensitivity analyses revealed - because fuel consumption and capital cost of the autonomous vessel have the largest impact on the relative RFR. Besides that, continued research is also necessary on maintenance strategies, the concrete design of the shore control center and port call procedures of the autonomous vessel. The analysis in this paper only covered one particular type of vessel – an unmanned bulk carrier as developed in the MUNIN project. However, the paper does not argue that a bulk carrier is necessarily the best application for unmanned autonomous ships. By now many new ideas have emerged such as e.g. autonomous inland shipping or partly unmanned bridge watches. Other ship types might be promising as well. A first qualitative analysis can be found in e.g. Rødseth and Burmeister (2015b), where the applicability of unmanned systems on deep-sea container vessel and offshore supply vessels are discussed. The same applies for different forms of assisted operation as an intermediate step towards total autonomy (for examples see e.g. Burmeister, Rødseth et al. (2014)). As these ideas become more mature, it would be worthwhile to carry out a structured cost-benefit analysis to get a sense concerning the economic effects for different ship types and trades, compare results and identify the most advantageous applications of

developing a cost model for an autonomous bulker. Both models are used to calculate the EPV of costs over a period of 25 years. Results are compared in three different scenarios in order to identify the potential impact of unmanned autonomous ships on the profitability of shipping companies. The first scenario analyzes the effects of removing the crew on board. The EPV of costs for both vessels is almost on the same level in this case. Thus, reducing crew costs only will most likely not be enough to justify total autonomy. However, innovative autonomous ship designs can be expected to be more fuel efficient than conventional ships. This effect is considered in the second scenario. Here the EPV of costs of the autonomous bulker is mUSD 4.3 lower than the reference bulkers EPV of costs. This means that the required freight rate of the autonomous bulker which produces a zero net present value is 3.4% lower than the required freight rate of the conventional vessel. Accordingly autonomous ships have the potential to increase the profitability of shipping companies - within the assumptions underlying this analysis. Additional economic advantages, besides the ones considered here, are likely because it is easier to realize benefits associated with ‘the intelligent ship’ on autonomously acting vessels. Moreover, improved fuel efficiency would also contribute to the goal of reducing emissions from shipping. The third scenario, however, highlights an important limitation. Autonomous operation using HFO as main fuel would be technically challenging. Thus, the autonomous vessel uses MDO as main fuel instead of HFO in this scenario. The result is a substantial increase of voyage costs. As a consequence the EPV of costs of the autonomous vessel is mUSD 19.1 higher than for the conventional bulker which still uses HFO. Thus, if going for a completely unmanned vessel requires switching to MDO as main fuel, and as long as vessels are not limited to shortsea trades in Sulphur Emission Control Areas, it is unlikely that the unmanned vessel would be financially viable. In future, however, tighter environmental regulations may diminish the difference in price between marine fuel types and thus reduce the disadvantage of using MDO. Based on the findings discussed in this paper it is safe to argue that autonomous ships can have a positive impact on the profitability of shipping companies. However, it needs to be emphasized that the 10

Research in Transportation Business & Management xxx (xxxx) xxx–xxx

L. Kretschmann et al.

Hummels, D., Lugovskyy, V., & Skiba, A. (2007). The trade reducing effects of market power in international shipping. Journal of Development Economics, 89(1), 84–97. IMO (2011). Reduction of GHG emissions from ships. Marginal abatement costs and cost effectiveness of energy-efficiency measures. MEPC 62/INF.7. IMO (2015). The IMO regulatory framework and its application to Marine Autonomous Systems, MSC 95/INF.20. IMO (2017). IMO Sulphur oxides (SOx) – Regulation 14. Date of access: 21/06/2017 http://www.imo.org/en/OurWork/Environment/PollutionPrevention/AirPollution/ Pages/Sulphur-oxides-(SOx)-%E2%80%93-Regulation-14.aspx. Insee International prices of imported raw materials - Heavy fuel oil (Rotterdam). Date of access: 09/02/2016 http://www.insee.fr. Jahn, C., & Münsterberg, T. (2015). Offshore-Windenergie – Kostensenkung durch Logistiksimulation. In M. Rabe, & U. Clausen (Eds.), Simulation in Production and Logistics 2015. Stuttgart, Germany: Fraunhofer IRB Verlag. Jokioinen, E. (2016). Introduction. Rolls-Royce. Remote and autonomous ships the next steps. Kretschmann, L., et al. (2015a). MUNIN deliverable 9.2: Qualitative assessment. MUNIN report. Kretschmann, L., et al. (2015b). MUNIN deliverable 9.3: Quantitative assessment. MUNIN report. Lamb, T. (2004). Ship design and construction volume II. Jersey City, USA: The Society of Naval Architects and Marine Engineers. Lindhjem, C. E., & Browning, L. (2007). LADCO 2005 commercial marine emissions. LADCO reports. MAN Diesel & Turbo (2011). Basic principles of ship propulsion. MAN Diesel & Turbo (2014a). Propulsion trends in bulk carriers: Two-stroke engines. MAN Diesel & Turbo (2014b). Marine engine: IMO Tier II and Tier III programme 2015. MAN Diesel & Turbo (2014c). L28/32H: Project guide - Marine, four-stroke GenSet compliant with IMO Tier II. Mau, G. (1984). Handbuch Dieselmotoren im Kraftwerks- und Schiffsbetrieb. Wiesbaden: Vieweg & Teubner. Mikelis, N. (2012). The Hong Kong convention and its guidelines for safe and green ship recycling. MIMA lecture. MTNZ, Ministry of Transport New Zealand Transport-related price indices. Date of access: 09/02/2016 http://www.transport.govt.nz. Neylan, P. (2011). Ship operating cost 2011–2012 annual review and forecast. London: Drewry. Papanikolaou, A. (2014). Ship design: Methodologies of preliminary design. Heidelberg: Springer. Porathe, T. (2013). MUNIN deliverable 7.1: Error and human intervention report in the context of autonomous vessels. MUNIN report. Porathe, T., Costa, N., & Tjora, Å. (2014). MUNIN deliverable 7.4: Organizational lay-out of SOC. MUNIN report. Ringbom, H., Felix, C., & Viljanen, M. (2016). Legalities. Rolls-Royce. Remote and autonomous ships the next steps. Rødseth, Ø., & Burmeister, H.-C. (2012). Developments towards the unmanned ship. Proceedings of International Symposium Information Ships - ISIS 2012, Hamburg, Germany, August. Rødseth, Ø., & Burmeister, H.-C. (2015a). Risk assessment for an unmanned merchant ship. TransNav, the International Journal on Marine Navigation and Safety of Sea Transportation, 9(3), 357–364. Rødseth, Ø. J., & Burmeister, H.-C. (2015b). MUNIN deliverable 10.2: New ship designs for autonomous vessels. MUNIN report. Rødseth, Ø., Kvamstad, B., Porathe, T., & Burmeister, H.-C. (2013). Communication architecture for an unmanned merchant ship. Proceedings of OCEANS 2013, MTS/IEEE, Bergen, Norway, June (pp. 1–9). . Rødseth, Ø., Tjora, Å., & Burmeister, H.-C. (2014). MUNIN deliverable 4.7: Draft architecture standard. MUNIN report. Rolls-Royce, Rolls-Royce to lead autonomous ship research project Date of access: 09/02/ 2016 http://www.rolls-royce.com. Ross, J. M. (2004). A practical approach for ship construction cost estimating. Proceedings of the 3rd International COMPIT, Siguënza, Spain, May (pp. 98–110). . Safari, F., & Sage, B. (2013). MUNIN deliverable 7.2: Legal and liability analysis for remote controlled vessels. MUNIN report. Sand, P. (2016). Demolition age drops as the dry bulk market enters another challenging year, Bimco news, last updated: 05 February 2016. Sanquist, T. F. (1992). Human factors in maritime applications: A new opportunity for multi-modal transportation research. Proceedings of the Human Factors 36th Annual Meeting-1992: 1123–1127. Sage: Publications. Schröder-Hinrichs, J.-U., et al. (2015). Introducing the concept of resilience into maritime safety, MARISA conference. Shetelig, H. (2013). Shipbuilding cost estimation parametric approach. Norwegian University of Science and Technology. Stokey, R., et al. (1999). AUV bloopers or why murphy must have been an optimist: A practical look at achieving mission level reliability in an autonomous underwater vehicle. Proceedings of the 11th International Symposium on Unmanned Untethered Submersible Technology, Durham, New Hampshire, August (pp. 32–40). . Stopford, M. (2009). Maritime economics. London: Routledge. Van Hooydonk, E. (2014). The law of unmanned merchant shipping. The Journal of International Maritime Law (2014) 20 JIML. Waterborne TP (2011). Waterborne implementation plan: Issue May 2011. Waterborne strategic research agenda. Watson, D. G. M. (1997). Practical ship design. Elsevier ocean engineering book series. vol. 1. Oxford, UK: Elsevier Science Ltd. Wehner, K., Schmidt, H., & Fentzahn, E. (2013). MUNIN deliverable 6.2: Specification concept of the general technical system redesign, MUNIN report.

autonomous ships. In this context, it would also be interesting to broaden the assessment perspective to put a stronger emphasis on macroeconomic and societal aspects as well as the risk element associated with unmanned vessels. A main assumption of the analysis presented in this paper is that an autonomous and a conventional vessel maintain an equal safety level. Notwithstanding, some new accident types are probably unavoidable in connection with autonomous vessels. One might compare this to the introduction of new technologies to shipping in the past. Radar as an aid to navigation led to a new accident type, the “radar assisted collision”. Similar developments were seen with the introduction of AIS and ECDIS (Schröder-Hinrichs et al., 2015). Anyhow, radar has no doubt had a positive impact on safety and reduced maritime accidents at large (Allianz, 2012). Right now it is too early to conclude whether autonomous vessels will have a similar impact. However, first quantitative and semi-quantitative analysis of relevant incident types point towards slightly lower risk (Kretschmann et al., 2015b). Future research should build on these results and aim to get a clearer understanding of the risks associated with a higher degree of autonomy in shipping. In case a different risk profile is confirmed for autonomous ships it should, of course, be taken into account when analyzing their economic benefit as well. Funding The research leading to these results has received funding from the European Union's Seventh Framework Programme under the agreement SCP2-GA-2012-314286. References ABS (2013). American Bureau of Shipping. Ship energy efficiency measures. Status and guidance. ABS ship energy efficiency measures advisory. Allianz (2012). Safety and shipping 1912–2012. Allianz Global Corporate & Specialty AG. Arnsdorf, I. (2014). Rolls-Royce drone ships challenge $375 billion industry: Freight. Bloomberg businessweek, 25 February. Atlason, G. F., & Jónsson, Þ. S. (2014). MUNIN deliverable 6.8: Constant engine efficiency concept, MUNIN report. Ballini, F., & Bozzo, R. (2015). Air pollution from ships in ports: The socio-economic benefit of cold-ironing technology. Research in Transportation Business & Management, (17), 92–98. Bertram, V. (2016). Unmanned & autonomous shipping – A technology review. 10th symposium on high-performance marine vehicles (pp. 10–24). . Bertram, V., & Schneekluth, H. (1998). Ship design for efficiency and economy. Oxford, UK: Butterworth-Heinemann. Birol, F. (2015). World energy outlook 2015. Paris: International Energy Agency. Blendermann, W. (1996). Wind loading of ships - collected data from wind tunnel tests in uniform flow. Technical report, TUHH. Burmeister, H.-C., Bruhn, W. C., & Rødseth, Ø. J. (2014). Can unmanned ships improve navigational safety? Proceedings of the transport research arena 2014, Paris, France. Burmeister, H.-C., Rødseth, Ø. J., Porathe, T. E., & Bruhn, W. C. (2014). Autonomous unmanned merchant vessel and its contribution towards the e-navigation implementation: The MUNIN perspective. International Journal of e-Navigation and Maritime Economy, 1(1), 1–14. Caprace, J.-D., & Rigo, P. (2009). Multi-criteria decision support for cost assessment techniques in shipbuilding industry. Proceedings of the 8th International COMPIT, Budapest, Hungary, May (pp. 6–21). . Corbett, J. J., Wang, H., & Winebrake, J. J. (2009). The effectiveness and costs of speed reductions on emissions from international shipping. Transportation Research Part D, 14(8), 593–598. DA-Desk Lauritzen Bulkers, a venerable shipping company. Date of access: 09/02/2016 http://www.da-desk.com/customers/case-studies/lauritzen-bulkers-standing-testtime. Devanney, J. (2008). The argument for twin screw tankers. WMU Journal of Maritime Affairs, 7(1), 353–380. DNV GL (2015). DVL GL homepage. Date of access: 09/02/2016 https://www.dnvgl. com/technology-innovation/revolt/index.html. E.R. Schiffahrt (2015). Going easy on the budget. Ship & shore, 15(2), 6–7. EIA, Europe Brent Spot Price FOB Date of access: 09/02/2016 http://www.eia.gov/dnav/ pet/hist/LeafHandler.ashx?n=pet&s=rbrte&f=m. EquityGate Advisors (2013). Bulk shipping market update. Date of Access: 15/02/2016 http://www.equitygate.de. Fraunhofer CML, MUNIN result, Date of access: 09/02/2016 http://www.unmannedship.org/munin/about/munin-results-2/. Greiner, R. (2011). Ship operating costs: Current and future trends. Presentation at propeller club 2011, Piraeus, Greece.

11