International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Contents lists available at ScienceDirect

International Journal of Naval Architecture and Ocean Engineering

journal homepage: http://www.journals.elsevier.com/
international-journal-of-naval-architecture-and-ocean-engineering/

Potential energy savings of air lubrication technology on merchant

ships
Young-Rong Kim*, Sverre Steen
Department of Marine Technology, Norwegian University of Science and Technology (NTNU), Trondheim, Norway

a r t i c l e i n f o a b s t r a c t

Article history: As the reduction of greenhouse gas emissions has become an important issue, measures and devices to
Received 24 January 2023 reduce energy consumption are in increasing demand. In this study, the potential energy saving due to
Received in revised form the application of air lubrication technology in merchant ships is analyzed. We propose a simplified
12 April 2023
empirical model, covering three different air lubrication technologies, based on the experimental results
Accepted 15 April 2023
and assumptions taken in the existing studies. The bottom surface area covered with air is important for
Available online 1 May 2023
the efficiency of the air lubrication system, according to the sensitivity analysis. From the global fleet
analysis, net-percentage power saving varies according to the operational profile as well as the tech-
Keywords:
Air lubrication system
nology. Net-percentage power savings of 2e5% from air bubble, 8e14% from air layer, and 16e22% from
Drag reduction air cavity technology were obtained assuming calm-water conditions. The methodology can be adopted
Energy saving in early design stage and fleet-wide analyses of various energy-saving measures.
Fleet-level analysis © 2023 Society of Naval Architects of Korea. Production and hosting by Elsevier B.V. This is an open
Ship transportation access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction wetted surface area of the vessel through lubrication between the
water flow and the hull surface.
Along with the acceleration of global warming, the international Mitsubishi Heavy Industry first installed its air lubrication system
community is paying keen attention to greenhouse gas emissions. (MALS) on a newly built ship and showed up to 12% net energy
The International Maritime Organization (IMO) released an initial savings in a sea trial of a module carrier (Mizokami et al., 2010). As
strategy, which aims to reduce greenhouse gas emissions by at least the nominal thickness of the air layer that forms on the bottom of the
50% in shipping by 2050 compared to 2008 levels and reduces hull increases, it has also been confirmed that net energy savings
emissions in stages as soon as possible, at Marine Environment increase as well. In a subsequent study, Kawabuchi et al. (2011)
Protection Committee (MEPC 72) in 2018 (IMO, 2018). In addition, analyzed the distribution of air bubbles on the hull surface and its
IMO adopted EEXI (Energy Efficiency eXisting Ship Index) and CII effect on propeller performance using CFD. Silverstream developed
(Carbon Intensity Indicator) as direct and short-term measures for an air carpet technology that covers the entire bottom of the ship by
this (IMO, 2021). In response to this international trend, various injecting micro bubbles from air release units, and it was confirmed
types of energy-saving devices and measures have been studied that a net energy reduction of about 4% could be achieved from
and applied to ships (Bouman et al., 2017; Zhang et al., 2021). actual operations of 40 k DWT tanker (Silberschmidt et al., 2016). In
Air Lubrication Systems (ALS) have been proposed as a prom- Lee et al. (2017), they observed results from model tests, sea trials,
ising energy-saving technology that can effectively lower fuel and in-service data from two ships fitted with air-lubrication sys-
consumption and greenhouse gas emissions from ships. Frictional tems (SAVER) made by Samsung Heavy Industry (SHI). As a result, in
resistance generally accounts for the largest portion of a ship's total the case of a heavy cargo carrier, power savings of 8.8% were esti-
resistance, especially in slow-moving ships, where it can account mated in the sea trial results, and on the basis of long-term trip data,
for more than 80% of the total. Air supplied from ALS to the bottom power savings of roughly 4e5% were recorded for a LNG carrier. In
of the hull can reduce frictional resistance by reducing the effective the meantime, Damen group unveiled the Air Chamber Energy
Saving (ACES) system, which uses a chamber-shaped design on the
bottom of the hull to create a cavity where air is supplied to prevent
water from coming into contact with the lower hull surface (Pavlov
* Corresponding author.
et al., 2020). According to several investigations (ABS, 2019; Gebraad
E-mail address: [email protected] (Y.-R. Kim).
Peer review under responsibility of The Society of Naval Architects of Korea. et al., 2021), there have been about 50 ships with air lubrication

https://doi.org/10.1016/j.ijnaoe.2023.100530
2092-6782/© 2023 Society of Naval Architects of Korea. Production and hosting by Elsevier B.V. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

systems installed by 2021, including some test cases, and interest in stress. At the same time, the effective viscosity is reduced due to an
the technology continues to increase considering recent orders. increase in void fraction, which consequently serves to suppress the
Many previous studies have looked at the applicability and turbulence of the flow and reduce skin friction (Park and Lee, 2018;
performance of air-lubrication systems for certain ship cases based ABS, 2019). As the injected air flux increases from this state, a
on model tests, sea trials, or CFD. On the other hand, some studies transition occurs in which the air bubbles and the air layer coalesce
have suggested a simplified method for estimating energy savings with each other in the gas-liquid mixture. When sufficient air is
in the air lubrication system. Ma €kiharju and Ceccio (2011), and injected into the near wall region of the turbulent boundary layer,
Ma €kiharju et al. (2012) established a method of calculating the the air is aggregated with each other to form a continuous air layer
energy economy of the air lubrication system using experimental separating the hull surface from the water flow. It was found that
data and presented the energy-saving results based on the such a developed air layer, so-called ALDR can significantly reduce
assumption of various situations for the U.S. Great Lakes vessel. frictional resistance compared to bubbly flow (Ceccio, 2010; Elbing
Comer et al. (2019) applied a similar method to perform an analysis et al., 2013). PCDR reduces frictional resistance by injecting air into
on route-based fuel and emission reduction of the three ships a recess or cavity at the bottom of the hull to separate the lower part
installed with ALS. of the hull from water (Lay et al., 2010). A typical hull design for
In this study, a simplified model, capable of applying different PCDR consists of a slightly downward sloping closure downstream
types of ALS considering various ship types and profiles, is pre- from the starting wall of the cavity into which air is injected, which
sented. Through the suggested model, this study aims to assess the forms a partial cavity to trap the air. This drag reduction effect by
impact of ALS on different ship types and obtain knowledge to help the cavity air layer is associated with the design of the bottom
reduce maritime emissions. The suggested model is intended for cavity and the continuous injection of air to maintain a stable air
early-design estimations, fleet-wide studies, and similar applica- layer (Wu and Ou, 2019).
tions where quick calculations requiring limited input are desired. To analyze the impact of an air lubrication system on the speed-
The underlying idea for the simplified method is that the layer of air power performance of a ship and to get insight into the optimal hull
produced by the air lubrication system reduces frictional resistance design and arrangement of the air lubrication system, several
by covering a portion of the hull's surface area with air, or air methods such as a model test, full-scale measurements, and CFD
bubbles. In order to enable the evaluation of the effectiveness of air computation are typically used. Fig. 1 presents the net-percentage
lubrication technology for ships with different design characteris- power savings of each air lubrication system collected from these
tics and operating profiles, the aforementioned simplified empirical studies as a horizontal box plot, along with a scatter plot of the
approach is adopted. The model has been established based on collected data. Several studies have used various metrics such as fuel
previously published experimental results and various information consumption, gas emission, drag reduction, and power saving, but
found in open literature, and throughout this paper, the calculation since net-percentage power savings defined in Eq. (1) will be used as
process and basic assumptions are explained. It includes three a performance metric of ALS throughout this study, only the results
types of air lubrication: air bubble, air layer, and air cavity. that can estimate such value are presented here, and the detailed
In Chapter 2, the general concepts and different types of air sources are presented in Table A.1 in appendix A. Although each
lubrication systems are presented, and relevant studies used to experiment was classified and listed by ALS type, some experiments
develop the simplified method are also introduced. In the following may correspond to a transitional region depending on the injected
chapter, the background of the various formulas employed in the airflow. According to the 25%e75% quartile ranges corresponding to
model to calculate the potential savings of an air lubricating system both ends of the box, BDR indicates a net-percentage power saving of
is discussed. Additionally, by comparing the results of the model about 3e6%, ALDR of 4e12%, and PCDR of 16e22%. It is clear that
with those of other studies, the overall properties and performance there are scatters in any type of air lubrication because the effec-
are demonstrated. Chapter 4 presents the research outcomes based tiveness of power savings is highly dependent on the operational
on the model that has been established. Here, parametric studies are profiles of the ship, the details of the air lubrication arrangement,
performed on the main factors of an air lubrication system, and and the experimental setup.
potential power savings for the global fleet in calm water and a
specific vessel in the real sea are evaluated. The last chapter presents
the conclusions obtained from the study and proposes future works.

2. Background

Frictional resistance often predominates among the resistance

components that a ship encounters when moving through water, and
it is heavily influenced by the wetted surface area, operating speed,
and viscosity of the fluid. Basically, the main principle of air lubrication
technology is to reduce frictional resistance by reducing direct contact
with water, that is to reduce the wetted surface area by releasing air
bubbles and covering some part of the bottom surface area of the hull.
Air lubrication can be classified into three different techniques; the air
bubble concept that injects micro air bubbles at the bottom of the hull,
the air film concept covers the bottom surface with a continuous air
layer through increased air flux, and the air cavity concept fills the
recessed area beneath the hull with air (Foeth et al., 2009; ABS, 2019).
In the rest of the text, air bubble concept will be referred to as BDR
(Bubble drag redction), air film as ALDR (Air layer drag reduction), and
air cavity as PCDR (Partial cavity drag reduction).
BDR reduces the local density by injecting numerous micro- Fig. 1. Potential net-percentage power savings achieved from previous studies. The
bubbles into the boundary layer, thereby reducing the Reynolds detailed data sources are listed in Table A.1.

2
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

3. Methodology added resistance due to wind, and RWave is added resistance in

waves.
3.1. Modeling of air lubrication technology For the estimation of the calm water resistance of a ship, it can
be estimated from various methods listed in Table 1 to suit the
The purpose of the method developed in this study is to evaluate dimensions and operating range of each ship. To use wind resis-
the potential energy saving of air lubrication systems on merchant tance coefficients taking into account different ship types and
ships and to obtain practical knowledge through the obtained re- windage area above the waterline, wind tunnel test results from
sults. It is anticipated that the applicability of air lubrication tech- Blendermann (1996), Fujiwara (2006), and ISO (2015) were gath-
nology will vary because ships operating globally have diverse ered. Moreover, the Combined Method, by Kim et al. (2022b) is
design characteristics and operating profiles. Since it is an analysis of used in the model to compute the added resistance in arbitrary
a wide variety of general-purpose levels, it is necessary to develop a wave headings using a few basic inputs. In order to obtain the
model that can simulate the overall trend in energy savings using the propulsive efficiency for various ships, this study uses the simpli-
fundamental ship information. Therefore, the energy economic fied method (Kristensen and Lützen, 2012), which can obtain a
calculation approach used in Ma €kiharju et al. (2012) was adopted in quick estimate from Wageningen B-series (Oosterveld and Van
this study as the performance evaluation method of ALS, and Oossanen, 1975) using a limited input value, and the methods
required volumetric fluxes of gas for the air-lubrication were esti- found in Birk (2019). In Nagamatsu et al. (2002)'s full-scale exper-
mated based on experimental data obtained from large cavitation iment, a bubble injector was dedicatedly designed to prevent the
tunnel in Elbing et al. (2008) and Makiharju et al. (2010). In addition, decrease in propeller efficiency due to the inflow of air bubbles into
a number of assumptions and simplifications were made regarding the propeller. However, according to later studies (Kawakita et al.,
the application and composition of the air lubrication system based 2011; Kawabuchi et al., 2011; Jang et al., 2014), the loss of propul-
on the findings of earlier studies that were published. sive efficiency before and after starting the air lubrication system
was less than about 1%, demonstrating that air bubbles may not
have much of an impact on a propeller. Based on this fact and for
3.1.1. Energy saving by air lubrication system
the simplicity of the model, this study neglect to include any
The energy savings by the air lubrication system are determined
change in propulsive efficiency caused by air bubbles.
by the reduction of the power required to overcome the frictional
In Section 5, the resistance of different commercial ships in the
drag on the lubricated surface and the power consumed to inject
global fleet is estimated. The resistance is calculated using the well-
gas into the bottom surface. Here, the performance index of the air
established empirical methods listed in Table 1. A method for se-
lubrication system uses the percentage of net power saving to total
lection of the best empirical methods for each ship has been
brake power, that is, net-percentage power saving, as stated in Eq.
established (Kramel et al., 2021). In the fleetwide calculation in this
(1). The total resistance and overall efficiency can be used to esti-
work, an updated version is applied, where some additional
mate the total brake power, as shown in Eq. (2).
empirical methods are included, as listed in Table 1. The main
Psave  Pcons Pnet feature of this resistance calculation method is that it requires few
PN ½% ¼  100 ¼  100 (1) input parameters. If more detailed information is available, more
PB PB
advanced resistance prediction methods can be applied.

RT  V
PB ¼ (2)
hT 3.1.2. Power saving by the air lubrication
The air lubricating device serves to reduce the resistance of the
where PN is net-percentage power saving, Psave is power saved by air area covered with air on the bottom surface among the frictional
lubrication system; it can be replaced by Psave,wc if there is an in- resistance generated from the fluid surrounding the hull during
fluence of weather, Pcons is power consumed by air compressor, Pnet ship operation. Therefore, the power that can be saved from the air
is net power saving by air lubrication system, PB is total brake po- lubrication can be calculated by simply taking into account the
wer, RT is total resistance in real sea conditions, hT is overall effi- power due to the frictional resistance generated by the wetted
ciency, and V is ship speed. surface area, the proportion of the air covered area to the total
The total resistance in real sea conditions can be simply wetted surface, and the drag reduction achieved by the air covered
expressed as Eq. (3), and it is considered as the sum of the calm area, as shown in Eq. (4).
water resistance and the additional resistance caused by wind and
waves. Here, it is possible to estimate each of the resistance com- Aa
Psave ¼ PF DR (4)
ponents that make up the ship's total resistance using established Aw
empirical methods, and the methods used in this study are listed in
Table 1. where PF is the power required to overcome the frictional drag, DR is
frictional drag reduction fraction due to air lubrication, Aa is air
RT ¼ RCalm þ RWind þ RWave (3) covered area at the bottom surface, and Aw is wetted surface area.
As in Eq. (5), the power needed to overcome the frictional drag
where RCalm is total resistance in calm water conditions, RWind is can be obtained by multiplying the total brake power by the

Table 1
Estimation of resistance components using empirical methods used in the study.

Component Method

RCalm Holtrop-Mennen (Holtrop and Mennen, 1982), Hollenbach (Hollenbach, 1998), Guldhammer (Guldhammer and Harvald, 1974; Kristensen and Lützen, 2012),
Oortmerssen (Van Oortmerssen, 1971; Helmore, 2008)
RWind Blendermann (Blendermann, 1996), Fujiwara (Fujiwara, 2006), STAJIP (ISO, 2015)
RWave Combined Method (Kim et al., 2022b)
hT Kristensen (Oosterveld and Van Oossanen, 1975; Kristensen and Lützen, 2012), Birk (Birk, 2019)

3
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

proportion of frictional drag to total drag. Here, the frictional drag

coefficient of the flat plate can be calculated using the ITTC 1957
friction line (ITTC, 1978) from Eq. (7), and accordingly, the frictional
resistance is obtained from Eq. (6). As indicated in Eq. (3), the total
resistance can be determined from empirical methods.

RF
PF ¼ PB  (5)
RT

1
RF ¼ rw Aw CF V 2 (6)
2

0:075
CF ¼ (7)
ðlog 10 Rn  2Þ2

where rw is water density, RF is frictional resistance, CF is frictional

coefficient, and Rn is Reynolds number.
According to Silberschmidt et al. (2016), the estimated Fig. 3. The ratio of bottom area to wetted surface area according to Cb of a ship at
appendage drag of the air release units attached to the bottom various draught.

surface of LNG carriers or cruise ships was less than 0.5% of the
total. In this study, the influence of appendages for all ALS types is 8
ignored for simplicity, and in the case of PCDR, it is assumed that >
> 0:8227Cb  0:201 T=Td ¼ 1:0
<
there is a newly built ship with proper design alterations for the Ab 0:8449Cb  0:1927 T=Td ¼ 0:9
¼ (9)
cavity form. By ignoring appendage drag for the ALS, it is in fact Aw > > 0:871Cb  0:1834 T=Td ¼ 0:8
:
assumed that great care has been taken to design the ALS in a 0:8942Cb  0:1698 T=Td ¼ 0:7
careful way.
Among the wetted surface areas under the waterline of a ship, where Ab is bottom surface area of a ship, Cb is block coefficient, T is
the air covered area, which can be expected to reduce frictional sailing draught, and Td is design draught.
resistance by the air lubrication system, is expressed in the form of Meanwhile, the arrangement of the air release device and the
Aa/Aw as shown in the following Eq. (8) to facilitate calculation in hull bottom design of the specific ship may affect the air covered
this study. Here, wetted surface area, bottom area, and air covered area. Kim et al. (2021) found that the reduction rate of frictional
area are defined as shown in Fig. 2. The area that can be covered resistance gradually increased as the air injection holes were placed
with air bubbles or layers increases as the flat bottom surface of the wider in the width direction, and Park and Lee (2018) reported that
hull increases. This implies that the potential energy saving from it was more effective to inject air distributedly in multiple locations
the air lubrication grows. than in a single injection location. In this regard, it is important to
The bottom area of a ship can be estimated from the particular appropriately arrange the injectors to increase the covering area of
hull shape of the ship, but as it is nearly impossible to get air at the bottom of the hull as much as possible. Based on the result
comprehensive hull shape data for ships at the fleet level, this study in Wu and Ou (2019), 0.84 was used as a ratio of air covered area to
proposed regression equations to estimate the bottom area of a ship the bottom area (Aa/Ab) in this study. However, this value can be
(refer to Eq. (9)). They have been developed based on 22 ships with changed as needed depending on each ship's ALS configurations.
various hull shapes (refer to Table A.2 in appendix A), and presented Fig. 4 depicts the boundaries for three drag reduction regions
the ratio of bottom surface area to wetted surface area (Ab/Aw) ac- based on flow rate in the air bubble injection experiment on the flat
cording to draught ratio (T/Td) and block coefficient (Cb) as illus- plate as given in Elbing et al. (2008). According to the flow rate, I
trated in Fig. 3. Here, Cb is based on the design draught of the ship. It represents the BDR region, II the transitional region between the BDR
is obvious that the Ab/Aw rises as Cb increases, and the Ab/Aw of the and ALDR, and III the ALDR region. The transitional gas injection rate
ballast condition is higher than that of the laden condition. For bulk for BDR (qtrans) and the critical gas injection rate for ALDR (qcrit) are
carrier and tankers with normally blunt hull shapes, the flat bottom indicated by the vertical lines in the figure. In region I, when air is
area tends to be bigger, whereas, for container ships and ferries injected into the bottom surface of the hull, the flow of air-liquid
with typically slender hull shapes, it tends to be smaller. mixture predominates at the turbulent boundary. As the air flux
gradually increases, some air bubbles are combined to form a partial
air layer, and a transition in which the mixed flow and the air layer
Aa A Aa
¼ b (8) coexist occurs (qtrans). Here, in region II, as the air flux increases, the
Aw Aw Ab
frictional drag reduction starts to rapidly increase from 20% to 80%.
When the air flux exceeds the critical value (qcrit), a continuous air
layer is completely developed, and the drag reduction is 80% or more.
The drag reduction values shown in Fig. 4 were measured at a
location of 6.05 m in the streamwise direction from the air injector
at the bottom of the plate, which is approximately half the length of
the entire plate. In fact, as the bubbles move toward the down-
stream direction, the sizes of the bubbles change due to the co-
alescences and splits or the bubbles escape from the near wall
boundary layer, thereby reducing the drag reduction effect
Fig. 2. Bottom-up view showing air covered area, bottom area, and wetted surface area (Kodama et al., 2005; Elbing et al., 2008; Verschoof et al., 2016). In
of a ship. some studies, endplates were installed along the entire length to
4
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

 n1 !
m_ g n P2 n
Pcomp ¼ P 1 (11)
hc r1 1 n  1 P1

where Pcomp is the power needed to compress a specified quantity

of gas, he is the efficiency of electrical motor (he ¼ 0.9), hc is effi-
ciency of an air compressor (hc ¼ 0.6), m_ g is the mass flow rate of air
necessary to maintain the given volume flow rate of air on the
bottom surface, r1 is the initial density of the air where it is com-
pressed, P1 is the atmospheric pressure, P2 is the air delivery
pressure from the compressor, and n is the polytropic index, chosen
as the value valid for adiabatic processes (n ¼ 1.4).
The pressure (P2) required by the compressor to deliver air to the
bottom of the hull to achieve air lubrication, consisting of static
pressure and dynamic pressure of the bottom air inlet of the hull, and
pressure loss due to the piping as shown in Eqs. (13)-(14). In general,
since the hull is deeply submerged in water, the influence of static
pressure contributes the most to the compressor power. The amount
of pressure loss caused by piping losses varies on a number of factors,
including the piping length, roughness, and the relevant design of
the air lubrication system. Some of the existing articles calculated
frictional pressure loss and minor loss from a moody chart assuming
Fig. 4. Drag reduction of three regions according to the gas injection rate measured
a certain pipe surface roughness (Ma €kiharju et al., 2012; Comer et al.,
from the model tests on the flat-plate. I, II, and III represent BDR region, transition
region between BDR and ALDR, and ALDR region, respectively. The figure is adapted 2019), while others (Ceccio and Ma €kiharju, 2012; Jang et al., 2014;
from Elbing et al. (2008) with minor modifications. Gallardo Martínez, 2016) used a range of 1e1.5 atm for the pressure
drop due to piping losses. As a cautious estimate for the pressure
drop caused by pipe, 1.5 atm was used in this study.
trap the air bubble in the bottom to achieve an effect (Kawashima
et al., 2007; Hoang et al., 2009). As such, it is very important to r1 P3
maintain the continuity of air bubbles at the bottom of the hull. It is m_ g ¼ qw (12)
P1
assumed that the generated air layer persists along the bottom of
the hull with the bubbles evenly distributed and maintaining the
P2 ¼ P3 þ DPloss (13)
level of drag reduction at the certain air flux measured in the
experiment. These experimental results were used as a criterion for
designing the ALS model in this study, and thus a drag reduction of 1
P3 ¼ rw gT þ rw V 2 (14)
20% in the transition region of BDR and 80% in the critical region of 2
ALDR were assumed. Since the cavity closure has a drag reduction
of more than 95% once it has been completed, as per Lay et al. where q is the volumetric gas flux per unit span, w is the width of
(2010)'s analysis, a conservative 95% is used for the PCDR here. In air covered area, P3 is the pressure under the hull, DPloss is pressure
other words, the relevant fixed drag reduction value and the air flux drop due to piping losses, and g is gravitational acceleration.
necessary under specified circumstances for each type of ALS are For the estimation of the volume flux of gas required to achieve a
employed as indicated in Fig. 4. However, in model tests or real given air lubrication type beneath the hull, the experimental data of
ships injected with different air flow rates, it is anticipated that a Elbing et al. (2008) and Makiharju et al. (2010) conducted in the
slightly varying drag reduction may be attained. large cavitation tank was used. Fig. 5 shows volumetric air flux per
span (q) for each air lubrication type according to the flow rate
measured from the experiments. In Elbing et al. (2008), air flux was
3.1.3. Power consumption by air compressor
converted to an air layer nominal thickness, and drag reduction
An air compressor or blower must be used to send air to the
according to thickness was used in their studies. In this model, as
outlet nozzles at the bottom of the hull in order to form and
shown in Fig. 5, gas flux according to flow speed was used as
maintain an air bubble layer beneath the ship's surface. The power
reference data for calculating compressor power for each ALS type.
used by the compressor varies according to the pressure and vol-
In addition, the gas fluxes of the air lubrication system estimated
ume of air transferred, which has a significant impact on the real
from the sea trials of the bulk carrier and module carrier are also
gain of an air lubrication system. According to Buckingham and
displayed (Hoang et al., 2009; Mizokami et al., 2010). According to
Pearson (2019), using compressor manufacturer data may be
M€akiharju et al. (2012), these investigations hypothesize that an air
more accurate in estimating compressor power consumption, but
layer or transitional region formed on the hull's bottom because the
in this paper, keeping simplicity and versatility prioritized, the
reduction in friction drag was lowered by 20e40%, which is partly
power needed to compress the gas at a specified mass flow rate is
compatible with the outcomes of Elbing et al. (2008), and
determined using the polytropic process (M€akiharju et al., 2012), as
Makiharju et al. (2010).
shown in Eqs. (10)-(11). The expansion or compression process
The air flux required to achieve the air layer or air cavity grows
including heat transfer is approximately described by the poly-
proportionally as the flow rate rises. In this study, BDR used qtrans in
tropic process equation (Nag, 2013).
the transition region between BDR and ALDR with 20% drag
Pcomp reduction as in Fig. 4. The required air flux in the corresponding
Pcons ¼ (10) state is much less than the air flux for maintaining the ALDR at the
he
smooth and rough plates. In the case of ALDR, the result was ob-
tained in a state in which the qcrit, that is, the continuous air layer is
5
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Fig. 5. The volumetric air flux per unit span required for different air lubrication techniques according to the flow speed. The figure is adapted from M€akiharju et al. (2012) with
minor modifications.

fully developed. It can be confirmed that the rough surface requires at the transition gas injection rate of BDR, and the gas flux neces-
additional gas flux to achieve the same extent of frictional drag sary to accomplish ALDR and PCDR was then applied as demon-
reduction as on the smooth surface. In the case of PCDR, it is strated in the study of Ma€kiharju et al. (2012). Here, the gas flux of
separated into the gas flow necessary to establish or maintain the ALDR was selected using the regression equation of the rough plate
cavity, and the required airflow of them is significantly different assuming that the surface of the hull would be somewhat rough
(Ma€kiharju et al., 2013). Since the gas flux needed to maintain the rather than completely smooth. Instead of using the establishment
cavity is less than half that needed to generate it and is almost gas flux of PCDR, maintenance was employed since it was assumed
identical to the gas flux needed for the BDR, hence the power used that keeping a cavity using an air lubricating system during the
by the compressor for the PCDR is actually very little. voyage would be more common. In this model, it is assumed that
According to Makiharju et al. (2010), in the flat plate experi- the air lubrication system is automatically controlled, such as the
ment, the air cavity did not easily reach the beach (end of the volume gas flux according to the flow speed shown in Fig. 5, and
closure) at a lower flow speed, and the flow in the cavitation tunnel that the system automatically shuts off if the compressor consumes
fluctuated with overshooting the beach at a higher speed. There- more power than the saving power or if it is outside the operating
fore, the corresponding study used a range of flow speeds showing speed range.
a stable flow rate change while the closure area can be completely The curve fits for volumetric air flux per unit span (q) for BDR at
filled with air, and this limited range is also applied in this study as the transitional gas injection rate are as follows:
shown in Fig. 5. Similar experimental results can be found in other
studies. More power was required to maintain the air cavity than qBDR ¼ 0:008V  0:0405 5:5 < V < 12:5 (15)
not lubricating in some low-speed conditions, and if the flow is too The curve fits for volumetric air flux per unit span for ALDR on a
low, the water might re-attach too close to the cavity step (Pavlov rough surface, and the curve fits for maintaining PCDR are
et al., 2020). In addition, according to Butterworth et al. (2015), expressed:
the efficiency of the air cavity decreases as the speed increases. If it
is too high, the bubbles may escape from the side of the cavity, qALDR ¼ 0:00126V 2  0:00755V þ 0:0391 5:5 < V < 12:5
resulting in negative net savings. However, the effect of reducing
(16)
drag on the air layer and the leakage of air from the bottom of the
hull is greatly influenced by the design characteristics of the bottom
cavity (Slyozkin et al., 2014; Butterworth et al., 2015). If the air flow qPCDR ¼ 0:00701V 2  0:0866V þ 0:277 5:5 < V < 7:5 (17)
rate is optimized and the hull is properly designed for the PCDR,
Table 2 shows the specifications and power of the compressor
considering the ship's operating characteristics, the air lubrication
used for air lubrication systems reported in several studies. Here,
system may also be functional in various speed ranges (Pavlov et al.,
main dimensions without accurate information from the references
2020). From the review of previous work summarized in the pre-
are obtained through a similar ship or simple estimation method
ceding discussion, it might be concluded that air cavity is an
(Kim et al., 2022a). By substituting the given information into Eqs.
immature type of ALS, requiring further research. However, we still
(10)-(13), the compressor powers of the various ships listed in the
chose to include it in this study.
table are estimated, and they are compared in Fig. 6. As the oper-
To determine the amount of air required for each ALS, the model
ating conditions and information of the vessel and ALS compressor
uses regression equations based on experimental data on the flat
shown in the table do not exactly match the setting used in this
plate as shown in Eqs. (15)-(17). At sufficiently high Reynolds
model, it may be a rather rough estimate. However, as shown in the
numbers, the dependence of gas flux on Reynolds number can be
figure, estimates are quite well correlated with the reported data.
weakened (Lay et al., 2010), and in experiments with different size
scales of comparable shape, the normalized gas flux for air lubri-
cation may be in the same range (Makiharju et al., 2010; Ma €kiharju 3.1.4. Weather correction for the efficiency of air lubrication system
et al., 2012). Thus, this study used the results of these model scales Depending on the environment the vessel is operating in, the ALS
to estimate the actual gas flow rate on a full-scale ship. A curve performance may change. According to the sea-trial results for a
fitting equation was developed based on the experimental results cruise ship from Foreship (Pavlov et al., 2020), a relatively small
6
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Table 2
Compressor specifications for air lubrication of ships from reported data. The asterisk symbol in the table represents the estimated value.

Ship type L [m]  B [m]  T [m] Speed [knots] Air flow [m3/min] Compressor power [kW] References

Bulk carrier 230  43  6.6e12.8 14 150e250 500e840 Mizokami et al. (2013)

Module carrier 153  38  4.5 13.25 40.5e94.5 72e211 Mizokami et al. (2010)
Tanker 168  32  10.6 11e14 e 150e230 Silberschmidt et al. (2016)
Container 350  51*  15.5* 24 200e550 680e1900 Mizokami et al. (2013)
Container 321*  45.6  14.75 19 133 600 Borusevich et al. (2017)
Ferry 105  17.9  6.3 14 26e110 13e60 Nagamatsu et al. (2002)
Passenger ship 240  32.2*  7.8* 17 100e200 230e460 Mizokami et al. (2013)

Table 3
Weather correction factor for the efficiency of air lubrication system according to the
sea-state.

Sea-state Max sig.wave height [m] Correction factor [-]

1 0.1 1
2 0.5 1
3 1.25 1
4 2.5 1
5 4 0.8
6 6 0.6
7 9 0.1
8 e 0

Psave;wc ¼ Psave  Cwc (18)

where Psave,wc is saved power by air lubrication system after

weather effect correction, Cwc is weather correction factor.

3.2. Comparison with previous studies

Fig. 6. Comparison of estimated compressor power from the model and the reported
This section examines the validity of the model by comparing the
compressor power in Table 2.
results obtained from several experiments such as model tests, CFD,
and full-scale measurements with the estimates of the model pro-
tendency of net power saving due to air bubbles according to ALS on- posed in Section 3.1. Fig. 7(a) shows the results of CFD analysis for the
off was seen for Beaufort scale 6 and above compared to less than 320 m cruise ship from Foreship (Pavlov et al., 2020). Here, the ship's
Beaufort scale 4. According to the model test of a tanker by speed changed from 14 knots to 22 knots, and net-percentage power
Borusevich et al. (2017), as the sea-state (ss) increased, cavity savings for four different air flow rates were shown. According to
instability due to waves occurred in PCDR. As a result, the power CFD calculations, the power saved by air bubbles tends to drop
saving efficiency of the air cavity system dropped by 20% in ss5, 40% constantly as ship speed increases at relatively low flow rates of
in ss6, and 90% in ss7, and was hardly functional above. The extreme 2.3e7.3 kg/s, however, at 10 kg/s, the saving rises until 19 knots and
pitch motion of the ship in the rough sea is found to have the po- then declines. BDR estimations from the proposed model reflect a
tential to seriously disturb the air layer on the bottom of the hull. As a trend where savings gradually decrease as speed increases, while the
result, the drag reduction from the air layer system is reduced, and ALDR estimates gradually grow with speed and then gradually fall
more air is needed to keep the air lubrication at its calm water level. beyond 18 knots. Although the figures do not exactly match, it seems
It has been challenging to evaluate the effect of waves on the air to capture the saving trend of ALS according to the speed and flow
lubrication system because the majority of ALS research has used rate of the ship. Compared to the CFD results, the volatility of net-
model experiments in towing tanks or sea trials in relatively calm percentage power saving with speed seems small, but comparing
water conditions. As a result, this study used the findings of the model tests and full-scale measurements in Fig. 7(b), it can be
Borusevich et al. (2016) to roughly represent the effects of weather seen that the volatility may not be so large.
in the model. As indicated in Eq. (18), in order to estimate the po- The net-percentage power savings achieved by the air lubrica-
wer saving reflecting the weather effect, the power savings esti- tion system during the laden voyage of a LNG ship are shown in
mated by Eq. (4) is multiplied by the correction factor as shown in Fig. 7(b). In the Model test, about 5e6% saving is attainable, and the
Table 3. Although it is cautiously expected that the air cavity system optimal performance can be achieved near the ship's service speed,
will have a greater loss due to ship motion than air bubble or air which is similar to the trend of ALDR predictions from the proposed
layer on the bottom surface, the coefficient is identically given to all model. In this experiment, two rows of air injectors in the forward
ALS kinds and ship types. Nevertheless, further tests and full-scale and one in the aft were arranged at the bottom of the ship model,
observations are required to fully understand how weather affects and the nominal thickness of the air layer was formed 2, 3, and
ALS performance in relation to the sea state and ship design. A 5 mm, respectively, which is presumed to correspond to transi-
comparison of ALS performance according to the application of tional air layer drag reduction. The results show net-percentage
actual sea conditions and weather correction factors is further power savings of roughly 4e5% from the ship's real operations,
discussed in Section 4.4. and they tend to decline slightly as speed increases, which is similar
to the BDR predictions made by our model.
7
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Fig. 7. Comparison of estimates of the proposed model with previous experimental results (a) Cruise ship (Pavlov et al., 2020), (b) LNG carrier (Lee et al., 2017), (c) Container ship A
(Borusevich et al., 2017) and Container ship B (Pavlov et al., 2020). The figures are adapted from the referenced papers with minor modifications.

Fig. 7(c) shows the results of PCDR analysis of two container accordance with various ship operating and design conditions.
ships. As shown in the figure, it was not possible to collect the Here, a supramax-class general cargo ship was selected for the case
power consumption of the compressor according to speed from the study and had the dimensions shown in Table 4. The settings for
relevant research (Borusevich et al., 2017; Pavlov et al., 2020), they each simulation case were specified as shown in Table 5, and the
were compared in terms of power saving by ALS, without cor- findings thereof are depicted in Fig. 8; the upper graph displays the
recting for compressor power. Looking at the predicted results, it net-percentage power saving, while the lower graph displays the
can be seen that it is in quite good agreement with the results of net power saving. To investigate the influence of block coefficient in
two ships within the computable area. Case 4, it is assumed that the hull shape design has been modified,
The BDR estimates in this model tended to be somewhat smaller i.e., that the air lubrication system has been installed on different
than the results of other studies, apparently because the BDR was ships (85e100% Cb).
calculated based on the transitional gas injection rate. Some of the In Fig. 8(a), the net power saving increases for all ALS types as ship
studies related to the air bubble system may have progressed beyond speed rises. While, the net-percentage power saving gradually de-
the area where only air bubbles exist to the transitional region where clines in BDR, and for ALDR and PCDR, it is gradually increased to a
air bubbles and air layers coexist through the adjustment of air flux. certain speed and then decreased. In principle, it is advantageous for
Overall, the ALDR has a tendency to overestimate, which is probably ships to operate at a low service speed because the frictional resis-
because this model assumes that a completely continuous air layer tance is predominant at low speeds, while at high speeds, the wave-
has been formed. In the case of PCDR, it tends to be almost similar to making resistance contributes more to the total resistance. However,
the experimental results within the application range. Fig. 7 shows since the air flux required by the compressor varies for each ALS type
that the prediction method presented in Section 3.1 gives fairly good as illustrated in Fig. 5, the speed conditions at which the maximum
estimates, seen in the light of the simplicity of the model and the saving could be achieved from the ALS could be slightly different.
complexity of the physics it attempts to represent. The comparison in The lower and upper limits of the net-percentage power saving
Fig. 7 is also a reminder that our model is not intended to be an are depicted in Fig. 8(b), which only changes the ratio of the air
accurate representation and not a replacement for model tests or covered area from 60% to 100% under the same conditions as in
detailed numerical studies. Case 1. Due to the hull structure and arrangement of air release
Due to the various assumptions and the settings of ALS in the units, it is practically impossible to completely cover the bottom
model, these comparison results were not accurately matched. This region with air; nonetheless, the range is assumed to be 60e100% to
model assumes that the air lubrication system is automatically examine the effect of the air covered area. Depending on the ALS
controlled to maintain the drag reduction level at a specific air flow type, this difference in the air covered area may lead to a significant
rate measured in the experiment. Moreover, an air layer is gener- difference in savings of 3e10%. In other words, it is clear that the
ated and maintained along the bottom of the hull without loss of configurations of the ALS installation, which decide how much air
bubbles, and the reduction in propulsive efficiency due to bubbles can cover the bottom region, can have a significant impact on
is ignored. In particular, during the actual operation of the ship, the performance in addition to the ship's flat bottom area.
drag reduction effect of the air layer may have been influenced by
several uncertain factors such as the ship's motion, environments,
and ship-specific conditions. As a result, it is difficult to replicate
Table 4
the precise conditions of tests undertaken in prior studies, hence Ship basic information.
this can only provide an approximation of the model's validity.
General cargo

4. Results Length [m] 194

Breadth [m] 32
Design draught [m] 12.6
4.1. Parametric study Block coefficient [-] 0.79
Wetted surface area at design draught [m2] 9370
The energy saving trend of the air lubrication technology was Deadweight tonnage [ton] 50700
investigated using a parametric study on changes in ship speed, air Maximum continuous rating [kW] 10780
Service speed [knots] 15.5
covered area, loading condition, and block coefficients in
8
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Table 5
Parametric study of the air lubrication systems for the target ship. Sea-states are assumed to be calm water condition.

Case Vs [knots] T [m] Cb [-] Aa/Ab [-]

1 10.5e15.5 (0.7e1.0Vs) 12.6 0.79 0.84

2 10.5e15.5 (0.7e1.0Vs) 12.6 0.79 0.6e1.0
3 15.5 8.8e12.6 (0.7e1.0Td) 0.79 0.84
4 15.5 12.6 0.68e0.79 (0.85e1.0Cb) 0.84

Fig. 8. Results of parametric study for the air lubrication systems: (a) speed, (b) ratio of air covered area to the bottom surface area, (c) draught, and (d) block coefficient.

9
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Case 3 demonstrates that the net saving of the air lubrication result of the preliminary analysis, the interaction between speed,
decreases as the draught increases. This is because the energy draught, and the ratio of air covered area to the bottom surface area
consumed by the compressor to supply air to the bottom of the ship was not significant, there was little difference between the total
is increasing with increasing draught. Additionally, the underwater index and the first index. Thus, only the first-order Sobol indices are
area increases along with the draught, increasing the hydrody- displayed in the figure. Within the parameter range defined in this
namic drag forces, which has the effect of increasing the overall research, the ratio of air covered area to the bottom surface area is
required power. the most influential as it accounts for about 46e85% of the total
Fig. 8(d) shows the parametric study results of the block co- variance. In BDR, speed is a relatively more important parameter
efficients. The bottom area of the ship generally tends to widen as than ALDR and PCDR, which is about 35e42% of the total. The
the block coefficient increases, thus even if it is assumed that the change in the Sobol index of various parameters according to block
same percentage of air is covering the hull bottom, it can be seen coefficient is not significant, but the influence of draught is rela-
that the amount of frictional resistance can be decreased. That is, tively greater in ships with a large block coefficient.
the efficiency of air lubrication system is high at a high block co-
efficient as can be observed from the figure.
4.3. Case study of global fleet in calm water condition

4.2. Global sensitivity analysis On about 48,710 global fleet registered in the Seaweb database,
performance analysis in calm sea conditions according to the
The Sobol method, a global sensitivity analysis method, was installation of the air lubrication system was carried out. Here, five
used to determine how each parameter affected the model's ship categories of bulk carrier, chemical tanker, general cargo,
output. The Sobol sensitivity index can be used to quantify each container ship, and oil tanker were examined, and ship type and
parameter's contribution to the variance of the model output. A low size were categorized in accordance with the IMO's fourth green-
Sobol index indicates that the variation of the output caused by a house gas study (IMO, 2020), as indicated in Table 6. The draught
change in the corresponding parameter is relatively small (Homma ratio for each ship type was used as given in Table 7 in order to
and Saltelli, 1996; Saltelli and Annoni, 2010). The first-order index is assume the laden and ballast voyage of the ship operation (Olmer
measuring the direct effect of each parameter on the variance of the et al., 2017). Here, ballast-only voyages are uncommon for
model. It can be expressed as Eq. (19), and it means an expected container ships unlike other ship types, thus the average draught
decrease in the variance of the model when Xi is fixed. The total ratio is applied for all voyage types. According to the study's
index, which includes both the first-order index, as well as the specified bin size, Fig. 10 depicts the composition of each type of
sensitivity due to the interaction between that parameter and all ship, and the distributions of ship parameters employed in this case
other parameters, can be expressed as Eq. (20). The larger the dif- study are shown in Fig. 11.
ference between the first and total index, the greater the effect of Fig. 12 displays the comparison of the global fleet's potential
sensitivity on variance due to the interaction. net-percentage power savings by type of air lubrication. The blue
and red boxes represent the case study assuming that there is no
Var½EðYjXi Þ environmental force when navigating at service speed under
Si ¼ (19)
VarðYÞ ballast and laden conditions.
The saving of the air lubrication system during a ballast voyage
Var½EðYjXi Þ is higher than that of a laden voyage, as can be demonstrated in this
STi ¼ (20)
VarðYÞ plot. Container ships deliver the same results because their draught
is assumed to be the same. Inspecting the overall results for each
where, E(Y|Xi) stands for the predicted output value when Xi is ship type, bulk carriers and tankers with flat bottom shapes that
fixed, and Xi denotes all uncertain parameters except Xi. can hold more air bubbles in the hull bottom, i.e., generally asso-
Here, as shown in Fig. 9, global sensitivity was examined within ciated with high block coefficients, are advantageous. Referring to
the range (Vs: 10.5e15.5, T: 8.9e12.6, Aa/Ab: 0.6e1.0) that was taken Fig. 11, it can be seen that a ship with a relatively low operating
into account in the previous section. Within this range, it was speed compared to the size of the ship, that is, the Froude number,
assumed that the parameters were distributed uniformly. For each rather than the operating speed, has a more direct effect on power
sensitivity analysis group, different block coefficients were used of saving. The potential net-percentage power savings of BDR, ALDR,
0.68, 0.74, and 0.79 to support the assumption that the air lubri- and PCDR are 2e5%, 8e14%, and 16e22%, respectively, when the
cation system was installed on the different ships, respectively. As a results of the entire fleet of 25e75% quantile in ballast and laden

Fig. 9. Sensitivity analysis with Sobol's indices for each parameter according to different block coefficient. The bars in the figure represent the first order Sobol index using 10,000
random samples from the Monte Carlo method.

10
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Table 6 voyages are taken into account.

Vessel type and categories. A more detailed comparison according to the bin size for each
Ship type Bin size Capacity Unit ship type is displayed in the following Fig. 13. The PCDR results of
Bulk carrier 1 0-9999 DWT
some container ships are not provided here, which were not
2 10,000e34,999 calculated because the service speeds of the vessels belonging to
3 35,000e59,999 the corresponding bin sizes were outside the operating range of the
4 60,000e99,999 PCDR. It is clear that even within the same ship type, performance
5 100,000e199,999
can vary significantly depending on the operation profile and hull
6 200,000-
Chemical tanker 1 0-4999 DWT characteristics. Additionally, it can be shown that overall savings
2 5000e9999 tend to rise as bin size grows. This is thought to be the case since the
3 10,000e19,999 proportion of frictional resistance in total resistance increases with
4 20,000e39,999
ship size. On the other hand, the net-percentage power savings
5 40,000-
Container ship 1 0e999 TEU
tend to no longer increase but rather slightly decrease in the case of
2 1000e1999 a large tanker such as bins 7 and 8 because the air compressor's
3 2000e2999 power consumption rises due to the hull design with deep draught.
4 3000e4999 Based on these results, coastal barges with flat bottom hull shapes
5 5000e7999
that typically operate at low draught and low speeds are believed to
6 8000e11,999
7 12,000e14,999 be an ideal ship type, although not investigated in this study.
8 14,500e19,999 The results shown here are obtained under the assumption of
9 20,000- calm water conditions and can be somewhat optimistic due to the
General cargo 1 0-4999 DWT various assumptions and simplifications. In order to obtain the cor-
2 5000e9999
3 10,000e19,999
responding amount of drag reduction on an actual ship, an appro-
4 20,000e30,000 priate design must follow. However, the estimated results in Fig. 12
Oil tanker 1 0-4999 DWT are fairly consistent and are distributed in a similar range to the
2 5000e9999 previous research as in Fig. 1, despite the fact that a ship-to-ship
3 10,000e19,999
comparison between them cannot be done due to different experi-
4 20,000e59,999
5 60,000e79,999 mental setups, such as ship speed, loading conditions, and air
6 80,000e119,999 lubrication system.
7 120,000e199,999 Considering the actual use of air lubrication technology, the
8 200,000- following characteristics can be considered from the above analysis
results. For ships that are already in operation, the retrofit is
comparatively easy with BDR and ALDR and a certain degree of drag
Table 7 reduction can be expected. Furthermore, it is investigated that the
Average draught ratio according to the voyage type of different ship types. Draught
impact on the ship's maneuvering and sea-keeping capabilities is
ratio is defined as the ratio of actual draught to design draught.
not significant (Thill et al., 2005; Foeth et al., 2009; Gallardo
Ship type Ballast voyage Laden voyage Martínez, 2016). However, according to the hull shape and oper-
Bulk carrier 0.58 0.91 ating profile, it is required to assess the actual gain between savings
Chemical tanker 0.66 0.88 by air lubrication and consumption by the air compressor. Mean-
General cargo 0.65 0.89
while, PCDR is expected to be effective when the ship maintains an
Oil tanker 0.60 0.89
Container ship 0.82
air lubrication system throughout actual operations due to the
relatively small air flux required to keep the air cavity. According to
Zverkhovskyi and De Jong (2020), a recent tank test study on PCDR,
it may be particularly suitable for hybrid and electric ships because
it can significantly reduce energy consumption, helping to reduce
the ship's capacity required for energy storage or power generation
systems. However, initial capital costs may be high due to the need
for a hull design specifically suited for PCDR. Additionally, the
system can be functional in a limited range, and there is still some
ambiguity regarding air cavity loss caused by ship motion.
Furthermore, since most studies on PCDR are based on model ex-
periments, further studies on a large scale are needed to determine
effective and feasible operational range.

4.4. Case study of a target ship in real sea condition

In the previous section, due to a large amount of calculation of

the global fleets, results were obtained by assuming calm water
conditions, i.e., external environmental factors are ignored, but in
this section, the performance of ALS according to the weather effect
is analyzed. For ease of calculation, it is assumed that a ship oper-
ates a fixed trade pattern at service speed annually, and three
Fig. 10. Percentage of vessels by bin size obtained from the sea-web database. different scenarios are compared: a calm sea condition, a real sea
condition, and the real sea condition where weather adjustment
factors are applied.
The general cargo ship in Table 4 was used, and a scenario was
11
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Fig. 11. Distribution of ship parameters relevant to ALS according to the ship type used in the study.

Fig. 12. Comparison of potential net power saving of global fleet by air lubrication type.

assumed in which the ship sailed the route between Rotterdam, scenarios, respectively. The ship typically experiences headwinds at
Netherlands (NLRTM), and New York, USA (USNYC), at a constant the start of the voyage outside the Strait of Gibraltar, and in the
service speed of 15.5 knots, as indicated in Fig. 14. NLRTM-USNYC is North Atlantic Ocean, net-percentage power savings tend to decline
assumed as a laden voyage and in the opposite direction is ballast dramatically as a result of rather strong external environments,
voyage, and 50 waypoints were uniformly defined throughout the since the total resistance increases significantly, while the frictional
route. The histograms of the apparent wind speed, apparent wind resistance reduction by ALS is not influenced. In addition, seasonal
angle, significant wave height, and relative wave angle that the ship variations show that from June to August, there is an average saving
may experience while operating are presented in Fig. 15, which was of about 8%, while from December to February, the average saving
created using meteorological data from the appropriate route from in some areas drops to less than 4%. It is evident that there is a
the ECMWF reanalysis weather hindcast data for 2020 (In the significant variation in the performance efficiency of air lubrication
figure, 0 represents headwind and head wave). On average, the depending on the region and season. In Fig. 16(b), the difference in
ship encounters more headwinds from NLRTM to USNYC and net-percentage power saving according to the actual weather
following winds from USNYC to NLRTM during the voyage. conditions at sea can be confirmed. In calm sea conditions, the
Fig. 16 represents an example of ALDR in the ship's laden voyage savings of the ship is constant annually at all waypoints, but when
among the results of annual energy-saving simulations, and considering the weather profiles of the real sea, the net-percentage
Fig. 16(a) and Fig. 16(b) show seasonal changes and three different power saving changes geographically and seasonally, and there is a

12
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Fig. 13. Comparison of potential net power saving of global fleet in bin categories according to ship type: (a) bulk dry, (b): chemical, (c): container, (d): general cargo, (e) oil.

Fig. 14. (a) Yearly mean wind speed, (b) Yearly mean significant wave height at North Atlantic Ocean in 2020. Arrows in the figures represent the mean direction of the wind and
waves. Orange circle is route from NLRTM to USNYC and red diamond is from USNYC to NLRTM.

difference of about 30% on average. When the weather adjustment the ballast voyage and the laden voyage is also shown in Fig. 17,
factor proposed in this study is additionally applied, it tends to which is expected to be largely due to the meteorological charac-
decrease by about 5% more than that. Thus, the weather correction teristics of such a specified route, where the annual weather con-
factor itself is of minor importance compared to the increase of dition is much more severe in the laden voyage, as shown Fig. 15.
resistance and related power consumption due to wind and waves.
Fig. 17 shows the results of three scenarios for all ALS types and 5. Conclusions
voyage types at once. It can be seen that the overall energy-saving
effect decreases by roughly 15e35% compared to the calm water Most of the existing studies related to ALS have introduced ship-
conditions, taking into account the actual weather environment specific approaches such as model tests, CFD, and full-scale mea-
and weather correction effect. The difference in efficiency between surements, and few simplified models can be applied to various
13
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

Fig. 15. Histograms of the actual weather conditions that ship encounters: (a) apparent wind speed, (b) apparent wind angle, (c) significant wave height, (d) relative wave angle.

Fig. 16. Net power savings of a ship using ALDR at laden voyage (NLRTM-USNYC): (a) seasonal changes, (b) weather influences.

vessels with different operating profiles and evaluate energy-saving in relation to the configuration of ALS. The ratio of air covered area
trends. This study presents an easy-to-use tool that can be applied at to the bottom surface area could explain 46e85% of the total vari-
the global fleet level in order to assess the energy saving potential of ation of net-percentage power savings within the given parameter
air lubrication systems of different configurations. The tool might range in this study, according to a sensitivity analysis utilizing the
also be useful for early-design considerations of installation of air Sobol index. As higher draught increases the energy needed to
lubrication systems. The overall theoretical background underpin- overcome hydrodynamic drag forces and increases the energy of
ning the modeling, the estimation methods of several factors, and the compressor used to supply air, the savings are larger in ballast
the assumptions were addressed. On the basis of the established than in laden conditions.
model, parametric and sensitivity analyses were carried out, and According to the case study conducted under the assumption
insights on the variables influencing ALS performance were pro- that the entire global fleet would be equipped with ALS, the
vided. Additionally, potential energy reduction trends for each ship possible net-percentage power savings would be BDRs of 2e5%,
type were examined, and changes in performance according to the ALDRs of 8e14%, and PCDRs of 16e22%. The level of savings iden-
location and seasonal effects were discovered for specific waters. tified in the fleet-wide study agrees fairly well with the level of
Through the parametric study, despite air lubrication systems savings of the various studies of individual ships found in the
being more efficient for ships operating at low speeds, the para- literature. Overall, bulk carriers and tanker with blunt hulls and
metric study indicates that the ideal operating conditions may vary moderate running speeds showed larger savings than container
slightly depending on the type of air lubrication. The efficiency of ships with slender hulls and high operating speeds. The operation
ALS increases as the block coefficient rises because high block profile and hull features, however, can significantly affect perfor-
usually comes with large flat bottom area. Not only this, but it is mance even within the same ship class. Considering the actual
also important how much air can cover the bottom area of the ship weather environment and weather correction effect, it can be

14
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

results. In fact, careful hull and system design suited to those systems
will be needed to achieve the same level of power reduction as the
results achieved from this work. The model would benefit from an
improved model of the effect of waves and ship motions on the ALS,
as well as the effect of ALS on the propulsive efficiency. Further study
of airflow at large scales and high Reynolds numbers, as well as
closer examination using full-scale measurements from ALS-
equipped ships, are necessary in order to close the gap between
model-scale and full-scale results and to more accurately capture the
impact of the air layer below the hull. In addition, BDR and ALDR can
be interconverted depending on the gas flow rate, and theoretically,
there is an optimal gas flow rate that can provide maximum energy
savings. This may vary depending on the design characteristics and
operational profile of the vessel and need be constantly controlled
during actual ALS operation. In a future study, it is planned to analyze
the energy saving and related emission reduction potential from the
optimized operation of air lubrication, taking into account the actual
operational pattern and environment, including the wave correction
on the air lubrication effectiveness.

Declaration of competing interest

Fig. 17. Comparison of average annual net power saving of ALS according to the
The authors declare that they have no known competing
application of weather effects.
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
observed that the effect decreases by about 15e35% compared to
the calm water conditions. Acknowledgements
The approach taken in this study can be applied to provide
preliminary performance estimates when considering the instal- This study is part of the research projects CLIMMS e Climate
lation of an air lubricating system during the ship's initial design change mitigation in the maritime sector (Research Council of
phase, and an evaluation of the anticipated performance of the Norway (RCN) project number 294771).
global fleet can also be taken into consideration. In the area of in-
ternational shipping, this will help to emphasize the potential of air Appendix A
lubrication technology to reduce emissions.
Nonetheless, it should be noted that some of the assumptions and
simplifications of the model might lead to somewhat optimistic

Table A.1
Relevant studies on the energy saving of air lubrication systems used in Fig. 1.

Type of ALS Method References

BDR Model test Kodama et al. (2005); Lee et al. (2017)

Sea-trial Latorre et al. (2002); Nagamatsu et al. (2002); Hoang et al. (2009);
Kumagai et al. (2015); Silberschmidt et al. (2016); Lee et al. (2017);
Mizokami and Kuroiwa (2019); Pavlov et al. (2020); Silverstream (2022)
CFD Pavlov et al. (2020)
ALDR Model test Jang et al. (2014); Lee et al. (2017); Sindagi et al. (2022)
Sea-trial Kodama et al. (2008); Hoang et al. (2009); Mizokami et al. (2010);
Lee et al. (2017); Pavlov et al. (2020)
CFD Fotopoulos and Margaris (2020)
PCDR Model test Butterworth et al. (2015); Borusevich et al. (2016);
Borusevich et al. (2017); Pavlov et al. (2020)
Sea-trial Aronietis et al. (2011); Borusevich et al. (2016); Pavlov et al. (2020)

Table A.2
Dimension of ships used for the regression equations in Eq. (9)

Ship type Lpp [m] B [m] Td [m] Cb [-] References

Tanker 161e323 28e60 9e21 0.72e0.85 Pinkster (1980); Bunnik (1999); Larsson et al. (2013);
Park et al. (2019a); Hinostroza et al. (2019); Kim et al. (2019)
General cargo 60e194 15e32 3.2e12.6 0.56e0.80 Gupta et al. (2019); Gerritsma and Beukelman (1972);
Kracht (1984); Alamsyah et al. (2018)
Bulk carrier 192 36 11.2 0.85 Lee et al. (2019)
Container 119e355 19e51 6e14.5 0.58e0.76 El-Moctar et al. (2012); Van (1997); Simonsen et al. (2013);
Park et al. (2019b); Reguram et al. (2016); Liu et al. (2015)
Ro-Ro/Ferry 158e178 21e32 6.1e8.2 0.54e0.56 Tsujimoto et al. (2009); Surendran et al. (2005)
Total 60e355 15e60 3.2e21 0.55e0.85

15
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

References nonlinear models. Reliab. Eng. Syst. Saf. 52, 1e17.

IMO, 2018. Strategy on reduction of ghg emissions from ships. MEPC 304, 72.
IMO, 2020. Fourth imo ghg study: reduction of ghg emissions from ships. MEPC 75,
ABS, 2019. Air Lubrication Technology. American Bureau of Shipping, Spring, TX,
7e15.
USA.
IMO, 2021. G. uidelines on the method of calculation of the attained energy effi-
Alamsyah, M.A., Hakim, M.L., Utama, I., 2018. Study of shear and pressure flow on
ciency existing ship index (eexi). MEPC 76/15/Add.2/Annex 7.
the variation of ship hull shapes as one of the biofouling growth factors. In:
ISO, E., 2015. Sistema de Gest~ ao da Qualidade, 9001: 2015.
Proceedings of the 3rd International Conference on Marine Technology. SCI-
ITTC, 1978. Report of performance committee. In: Proceedings 15th ITTC,
TEPRESS-Science and Technology Publications, pp. 97e105.
pp. 389e392.
Aronietis, R., Crozet, Y., Ferrari, C., Frouws, K., Grootbod, H., Guihery, L., Kapros, S.,
Jang, J., Choi, S.H., Ahn, S.M., Kim, B., Seo, J.S., 2014. Experimental investigation of
Laroche, F., Lloyd, M., Rouboumtsos, A., et al., 2011. Innosutra Project Deliverable
frictional resistance reduction with air layer on the hull bottom of a ship. Int. J.
D4 and D5-Topical Assessment of Innovative Successes and Not-Yet-Successes.
Nav. Archit. Ocean Eng. 6, 363e379.
Birk, L., 2019. Fundamentals of Ship Hydrodynamics: Fluid Mechanics, Ship Resis-
Kawabuchi, M., Kawakita, C., Mizokami, S., Higasa, S., Kodan, Y., Takano, S., 2011. Cfd
tance and Propulsion. John Wiley & Sons.
predictions of bubbly flow around an energy-saving ship with mitsubishi air
Blendermann, W., 1996. Wind Loading of Ships Collected Data from Wind Tests
lubrication system. Mitsubishi Heavy Industries Technical Review 48, 53e57.
Tunnel in Uniform Flow. Institut fur Schiffbau der Universitat Hamburg.
Kawakita, C., Takano, S., Kodan, Y., Mizokam, S., 2011. Experimental investigation of
Borusevich, V., Poustoshny, A., Sverchkov, A., Trincas, G., 2017. Future Outlook of
the behavior of injected air on the ship bottom and its influence on propeller.
Artificial Cavity Application for Reducing Hydrodynamic Resistance of
J. Jpn. Soc. Nav. Archit. Ocean Eng. 12, 43e50.
Containerships.
Kawashima, H., Kodama, Y., Hinatsu, M., Hori, T., Makino, M., Ohnawa, M.,
Borusevich, V., Pustoshny, A., Sverchkov, A., Trincas, G., 2016. Impact of air cavity
Takeshi, H., Sakoda, M., Kawashima, H., Matsuno, F., 2007. A research project on
technology on ship drag reduction: experience from research studies. In: 10th
application of air bubble injection to a full scale ship for drag reduction. In:
Symposium on High-Performance Marine Vehicles-HIPER 16. TUHH,
Fluids Engineering Division Summer Meeting, pp. 265e274.
pp. 94e107.
Kim, H.T., Kim, H.T., Kim, H.J., Kim, J.J., 2021. Study on the evaluation of frictional
Bouman, E.A., Lindstad, E., Rialland, A.I., Strømman, A.H., 2017. State-of-the-art
drag reduction by air lubrication and the arrangement of air injection parts for a
technologies, measures, and potential for reducing ghg emissions from
liquefied natural gas carrier. Journal of the Society of Naval Architects of Korea
shippingea review. Transport. Res. Transport Environ. 52, 408e421.
58, 144e157.
Buckingham, J., Pearson, D., 2019. Modelling Alternative Propulsion Technologies
Kim, T., Yoo, S., Oh, S., Kim, H.J., Lee, D., Kim, B., 2019. Numerical and experimental
for Merchant Vessels. RINA, Power & Propulsion Alternatives for Ships, 23rd
study on the estimation of added resistance of an lng carrier in waves. Int. J.
January.
Offshore Polar Eng. 29, 24e32.
Bunnik, T.H., 1999. Seakeeping Calculations for Ships, Taking into Account the Non-
Kim, Y., Steen, S., Muri, H., 2022a. A novel method for estimating missing values in
linear Steady Waves.
ship principal data. Ocean. Eng. 251, 110979.
Butterworth, J., Atlar, M., Shi, W., 2015. Experimental analysis of an air cavity
Kim, Y.R., Esmailian, E., Steen, S., 2022b. A meta-model for added resistance in
concept applied on a ship hull to improve the hull resistance. Ocean. Eng. 110,
waves. Ocean. Eng. 266, 112749.
2e10.
Kodama, Y., Takahashi, T., Makino, M., Hori, T., Ueda, T., Kawamura, N., Shibata, M.,
Ceccio, S.L., 2010. Friction drag reduction of external flows with bubble and gas
Kato, H., Inoue, T., Suzuki, T., et al., 2005. Practical application of microbubbles
injection. Annu. Rev. Fluid Mech. 42, 183e203.
to shipsdlarge scale model experiments and a new full scale experiment. In:
Ceccio, S.L., M€akiharju, S., 2012. Air Lubrication Drag Reduction on Great Lakes
Proceedings of the 6th International Symposium on Smart Control of
Ships. Great Lakes Maritime Research Institute.
Turbulence.
Comer, B., Chen, C., Stolz, D., Rutherford, D., 2019. Rotors and bubbles: route-based
Kodama, Y., Hinatsu, M., Hori, T., Kawashima, H., Takeshi, H., Makino, M.,
assessment of innovative technologies to reduce ship fuel consumption and
Ohnawa, M., Sanada, Y., Murai, Y., Ohta, S., 2008. A full-scale air lubrication
emissions. J. ICCT working pa per 11, 6.
€kiharju, S., Wiggins, A., Perlin, M., Dowling, D.R., Ceccio, S.L., 2013. experiment using a large cement carrier for energy saving (result and analysis).
Elbing, B.R., Ma
In: Proc. Japan Soc. Naval Architects and Ocean Engineers Conference,
On the scaling of air layer drag reduction. J. Fluid Mech. 717, 484e513.
pp. 163e166.
Elbing, B.R., Winkel, E.S., Lay, K.A., Ceccio, S.L., Dowling, D.R., Perlin, M., 2008.
Kracht, A., 1984. Einfluss des Bugwulstes auf den Leistungsbedarf eines Schiffes im
Bubble-induced skin-friction drag reduction and the abrupt transition to air-
Seegang. Forschungszentrum des Dt. Schiffbaus.
layer drag reduction. J. Fluid Mech. 612, 201e236.
Kramel, D., Muri, H., Kim, Y., Lonka, R., Nielsen, J.B., Ringvold, A.L., Bouman, E.A.,
El-Moctar, O., Shigunov, V., Zorn, T., 2012. Duisburg test case: post-panamax
Steen, S., Strømman, A.H., 2021. Global shipping emissions from a well-to-wake
container ship for benchmarking. Ship Technol. Res. 59, 50e64.
perspective: the mariteam model. Environ. Sci. Technol. 55, 15040e15050.
Foeth, E., Eggers, R., van der Hout, I., Quadvlieg, F., 2009. Reduction of Frictional
Kristensen, H.O., Lützen, M., 2012. Prediction of resistance and propulsion power of
Resistance by Air Bubble Lubrication. SNAME Maritime Convention, OnePetro.
ships. Clean Shipping Currents 1, 1e52.
Fotopoulos, A.G., Margaris, D.P., 2020. Computational analysis of air lubrication
Kumagai, I., Takahashi, Y., Murai, Y., 2015. Power-saving device for air bubble gen-
system for commercial shipping and impacts on fuel consumption. Computa-
eration using a hydrofoil to reduce ship drag: theory, experiments, and appli-
tion 8, 38.
cation to ships. Ocean. Eng. 95, 183e194.
Fujiwara, T., 2006. A new estimation method of wind forces and moments acting on
Larsson, L., Stern, F., Visonneau, M., 2013. Cfd in ship hydrodynamicsdresults of the
ships on the basis of physical components models. J. Jpn. Soc. Nav. Archit. Ocean
gothenburg 2010 workshop. In: MARINE 2011, IV International Conference on
Eng. 2, 243e255.
Computational Methods in Marine Engineering. Springer, pp. 237e259.
Gallardo Martínez, A., 2016. Investigation of the Air-Lubrication Effect on Friction
Latorre, R., Miller, A., Philips, R., 2002. Microbubble resistance reduction for high-
Resistance.
speed craft. SNAME Transactions 110, 259e277.
Gebraad, J., Quispel, M., Karaarslan, S., Lehne, M., Rafael, R., Eppich, M., Janssens, G.,
Lay, K.A., Yakushiji, R., Makiharju, S., Perlin, M., Ceccio, S.L., 2010. Partial cavity drag
Smidt, H., Barcanescu, M., 2021. Structuring towards zero emission waterborne
reduction at high Reynolds numbers. J. Ship Res. 54, 109e119.
transport, d2.1. state-of-play of decarbonisation of waterborne transport
Lee, C.M., Yu, J.W., Choi, J.E., Lee, I., 2019. Effect of bow hull forms on the resistance
“technology application atlas. https://www.waterborne.eu/projects/
performance in calm water and waves for 66k dwt bulk carrier. Int. J. Nav.
coordination-projects/steerer/results.
Archit. Ocean Eng. 11, 723e735.
Gerritsma, J., Beukelman, W., 1972. Analysis of the resistance increase in waves of a
Lee, J., Kim, J., Jang, J., McStay, P., Raptakis, G., Fitzpatrick, P., 2017. Full Scale Ap-
fast cargo ship. Int. Shipbuild. Prog. 19, 285e293.
plications of Air Lubrication for Reduction of Ship Frictional Resistance. SNAME
Guldhammer, H., Harvald, S.A., 1974. Ship resistance-effect of form and principal
Maritime Convention, OnePetro.
dimensions.(revised). In: Danmark, Danmarks Tekniske Hojskole, Kademisk
Liu, W., Suzuki, K., Shibanuma, K., 2015. Nonlinear dynamic response and structural
Forlag, vol. 8. Danish Technical Press, St. kannikestrade. DK 1169 Copenhagen.
evaluation of container ship in large freak waves. J. Offshore Mech. Arctic Eng.
Gupta, P., Steen, S., Rasheed, A., 2019. Big data analytics as a tool to monitor hy-
137.
drodynamic performance of a ship. In: International Conference on Offshore
M€akiharju, S., Ceccio, S.L., 2011. Air Lubrication Drag Reduction on Great Lakes
Mechanics and Arctic Engineering. American Society of Mechanical Engineers.
Ships. University of Michigan.
V07AT06A059.
Makiharju, S., Elbing, B., Wiggins, A., Dowling, D., Perlin, M., Ceccio, S., 2010. Per-
Helmore, P., 2008. Update on van oortmerssen's resistance prediction. In: Pacific
turbed partial cavity drag reduction at high Reynolds numbers. In: Proc. 28th
2008 International Maritime Conference.
Symp. On Naval Hydrodynamics.
Hinostroza, M., Xu, H., Soares, C.G., 2019. Manoeuvring test for a self-running ship
M€akiharju, S.A., Elbing, B.R., Wiggins, A., Schinasi, S., Vanden-Broeck, J.M., Perlin, M.,
model in various water depth conditions. In: Sustainable Development and
Dowling, D.R., Ceccio, S.L., 2013. On the scaling of air entrainment from a
Innovations in Marine Technologies. CRC Press, pp. 187e196.
ventilated partial cavity. J. Fluid Mech. 732, 47e76.
Hoang, C., Toda, Y., Sanada, Y., 2009. Full scale experiment for frictional resistance
M€akiharju, S.A., Perlin, M., Ceccio, S.L., 2012. On the energy economics of air
reduction using air lubrication method. In: The Nineteenth International
lubrication drag reduction. Int. J. Nav. Archit. Ocean Eng. 4, 412e422.
Offshore and Polar Engineering Conference. OnePetro.
Mizokami, S., Kawakado, M., Kawano, M., Hasegawa, T., Hirakawa, I., 2013. Imple-
Hollenbach, K.U., 1998. Estimating resistance and propulsion for single-screw and
mentation of ship energy-saving operations with mitsubishi air lubrication
twin-screw ships-ship technology research 45 (1998). Schiffstechnik 45, 72.
system. Mitsubishi Heavy Industries Technical Review 50, 44e49.
Holtrop, J., Mennen, G., 1982. An approximate power prediction method. Int.
Mizokami, S., Kawakita, C., Kodan, Y., Takano, S., Higasa, S., Shigenaga, R., 2010.
Shipbuild. Prog. 29, 166e170.
Experimental study of air lubrication method and verification of effects on
Homma, T., Saltelli, A., 1996. Importance measures in global sensitivity analysis of
actual hull by means of sea trial. Mitsubishi Heavy Industries Technical Review

16
Y.-R. Kim and S. Steen International Journal of Naval Architecture and Ocean Engineering 15 (2023) 100530

47, 41e47. Conference of Shipping in Changing Climate, pp. 10e11. Newcastle, UK.
Mizokami, S., Kuroiwa, R., 2019. Installation of Air Lubrication System for Ro-Pax Silverstream, 2022. Silverstream® system performance. URL: https://www.
Ferry and Verification of its Effect in Actual Seas Based on Onboard Measure- silverstream-tech.com/what-is-air-lubrication/.
ment Data, vol. 29. Japan Society of Naval Architects and Ocean Engineers, Simonsen, C.D., Otzen, J.F., Joncquez, S., Stern, F., 2013. Efd and cfd for kcs heaving
pp. 1e9. and pitching in regular head waves. J. Mar. Sci. Technol. 18, 435e459.
Nag, P., 2013. Engineering Thermodynamics. Tata McGraw-Hill Education. Sindagi, S., Vijayakumar, R., Saxena, B.K., 2022. Experimental investigation on ship's
Nagamatsu, T., Kodama, Y., Kakugawa, A., Takai, M., Murakami, K., Ishikawa, S., model in carrying out energy economics of bdr/als methodology. Ships Offshore
Kamiirisa, H., Ogiwara, S., Yoshida, Y., Suzuki, T., et al., 2002. A full-scale Struct. 17, 1437e1446.
experiment on microbubbles for skin friction reduction using” seiun maru” part Slyozkin, A., Atlar, M., Sampson, R., Seo, K.C., 2014. An experimental investigation
2: the full-scale experiment. J. Soc. Nav. Archit. Jpn. 15e28, 2002. into the hydrodynamic drag reduction of a flat plate using air-fed cavities.
Olmer, N., Comer, B., Roy, B., Mao, X., Rutherford, D., 2017. Greenhouse Gas Emis- Ocean Eng. 76, 105e120.
sions from Global Shipping, 2013e2015 Detailed Methodology. International Surendran, S., Lee, S., Reddy, J.V.R., Lee, G., 2005. Non-linear roll dynamics of a ro-ro
Council on Clean Transportation, Washington, DC, USA, pp. 1e38. ship in waves. Ocean. Eng. 32, 1818e1828.
Park, D.M., Lee, J.H., Jung, Y.W., Lee, J., Kim, Y., Gerhardt, F., 2019a. Experimental and Thill, C., Toxopeus, S., van Walree, F., 2005. Project energy-saving air-lubricated
numerical studies on added resistance of ship in oblique sea conditions. Ocean. ships (pels). In: Proceedings of the 2nd International Symposium on Seawater
Eng. 186, 106070. Drag Reduction, pp. 1e16.
Park, D.M., Lee, J.H., Lee, J., Kim, B.S., Kim, B.S., Yang, K.K., Kim, Y., Lee, Y.G., Kim, T., Tsujimoto, M., Kuroda, M., Shibata, K., Sogihara, N., Takagi, K., 2009. On a calculation
Yang, J.H., et al., 2019b. Comparative study on added resistance of a container of decrease of ship speed in actual seas. J. Jpn. Soc. Nav. Archit. Ocean Eng. 9,
ship in waves. In: The 29th International Ocean and Polar Engineering Con- 79e85.
ference. OnePetro. Van, S., 1997. Measurement of flows around a 3600teu container ship model. In:
Oosterveld, M.W.C., Van Oossanen, P., 1975. Further computer-analyzed data of the Annual Autumn Meeting. SNAK, Seoul, Korea, 1997.
wageningen b-screw series. Int. Shipbuild. Prog. 22, 251e262. Van Oortmerssen, G., 1971. A power prediction method and its application to small
Park, S.H., Lee, I., 2018. Optimization of drag reduction effect of air lubrication for a ships. ISP 18.
tanker model. Int. J. Nav. Archit. Ocean Eng. 10, 427e438. Verschoof, R.A., Van Der Veen, R.C., Sun, C., Lohse, D., 2016. Bubble drag reduction
Pavlov, G.A., Yun, L., Bliault, A., He, S.L., 2020. Air Lubricated and Air Cavity Ships. requires large bubbles. Phys. Rev. Lett. 117, 104502.
Springer. Wu, H., Ou, Y.p., 2019. Experimental study of air layer drag reduction with bottom
Pinkster, J.A., 1980. Low Frequency Second Order Wave Exciting Forces on Floating cavity for a bulk carrier ship model. China Ocean Eng. 33, 554e562.
Structures. Zhang, X., Bao, Z., Ge, Y.E., 2021. Investigating the determinants of shipowners'
Reguram, B.R., Surendran, S., Lee, S.K., 2016. Application of fin system to reduce emission abatement solutions for newbuilding vessels. Transport. Res. Trans-
pitch motion. Int. J. Nav. Archit. Ocean Eng. 8, 409e421. port Environ. 99, 102989.
Saltelli, A., Annoni, P., 2010. How to avoid a perfunctory sensitivity analysis. Environ. Zverkhovskyi, O., De Jong, J., 2020. Damen air cavity system of sustainable pas-
Model. Software 25, 1508e1517. senger ships. In: Proceedings of the Intern. Conf.«Sustainable and Safe Pas-
Silberschmidt, N., Tasker, D., Pappas, T., Johannesson, J., 2016. Silverstream system- senger Ships».eAthens (Greece). Royal Institution of Naval Architects & Hellenic
air lubrication performance verification and design development. In: Institute of Marine Technology, pp. 65e70.

17