Jmse 12 00335
Jmse 12 00335
Jmse 12 00335
Marine Science
and Engineering
Article
Analysis of Dynamic Characteristics of Rotor Sail Using a 4DOF
Rotor Model and Finite Element Model
Dong Min Kim, Soon Ho Hong, Se Hyeon Jeong and Sun Je Kim *
Department of Naval Architecture & Ocean Engineering, Chungnam National University, Daejeon 34134,
Republic of Korea; kdongmin1@o.cnu.ac.kr (D.M.K.); tnsgh9419@o.cnu.ac.kr (S.H.H.);
sehyeon043@o.cnu.ac.kr (S.H.J.)
* Correspondence: sunjekim@cnu.ac.kr
Abstract: The interest in wind-assisted ship propulsions (WASPs) is increasing to improve fuel
efficiency and to reduce greenhouse gas emissions in ships. A rotor sail, one of the typical WASPs,
can provide auxiliary propulsive force by rotating a cylinder-shaped structure based on the Magnus
effect. However, due to its huge rotating structure, a meticulous evaluation of the influence on the
ship structure and dynamical stability of the rotating structure should be conducted in the design
stage. In this respect, an analysis of the rotating structure for a 30 m height and 3 m diameter rotor sail
was conducted in this study. First, a 4DOF (four-degree-of-freedom) model was derived to simplify
the dynamics of the rotor sail. Using the 4DOF model, natural frequencies for four low-order modes
of the rotor sail were calculated, and frequency responses at support points were predicted. Next, a
comparison and validation with the finite element model of the rotor sail were carried out. For the 1st
and 2nd natural frequencies, a difference of approximately 0.3 Hz was observed between the 4DOF
model and the finite element model, confirming the effectiveness of the 4DOF model for low-order
modes. In analysis with changes in the bearing supporting stiffnesses, it was verified that lower
support bearings have a significant impact on rotor dynamics compared to upper support bearings.
Vibration response at the upper support was also confirmed through frequency response analysis
caused by imbalance at Thom disk and mid-plate. Additionally, when estimating the eccentricity
of the Thom disk as imbalance, a limit of eccentricity error could be set as 24 mm. The presented
modeling procedures and analysis results can be references during early design stage of a novel rotor
Citation: Kim, D.M.; Hong, S.H.; sail structure.
Jeong, S.H.; Kim, S.J. Analysis of
Dynamic Characteristics of Rotor Sail Keywords: wind-assisted ship propulsion; rotor sail; critical speed; harmonic response; rotor analysis
Using a 4DOF Rotor Model and Finite
Element Model. J. Mar. Sci. Eng. 2024,
12, 335. https://doi.org/
10.3390/jmse12020335
1. Introduction
Academic Editor: Joško Parunov As efforts in various industry sectors continue to reduce environmental pollution,
Received: 19 January 2024
interest in eco-friendly vessels also grows in the shipbuilding industry. The proportion of
Revised: 7 February 2024 greenhouse gas emissions from ships has gradually increased from 2.76% in 2012 to 2.89%
Accepted: 13 February 2024 in 2018 [1,2], prompting heightened awareness of environmental concerns. In response,
Published: 16 February 2024 the International Maritime Organization (IMO) is progressively strengthening regulations
on greenhouse gas emissions from ships [3,4]. The regulation of ship emissions primarily
relies on the Energy Efficiency Design Index (EEDI) for new vessels and Energy Efficiency
Existing Ship Index (EEXI) for existing vessels. New vessels can meet the environmental
Copyright: © 2024 by the authors. regulations by utilizing eco-friendly power sources such as electric propulsion, LNG
Licensee MDPI, Basel, Switzerland. (Liquified Natural Gas), and hydrogen propulsion [5,6]. However, existing vessels face the
This article is an open access article challenge of reducing fuel consumption and greenhouse gas emissions while minimizing
distributed under the terms and
structural modifications to their original design.
conditions of the Creative Commons
CCSs (Carbon Capture Systems) and Air Lubrication are considered realistic alter-
Attribution (CC BY) license (https://
natives to reduce fuel consumption and greenhouse gas emissions in existing ships [7–9].
creativecommons.org/licenses/by/
Additionally, a wind-assisted ship propulsion device that can be added onto existing vessels,
4.0/).
called Wind-Assisted Ship Propulsion (WASP), is also being developed [10–12]. Among
the WASP options, the Flettner rotor is considered the most realistic solution [13–15]. The
Flettner rotor, also known as the rotor sail, was first proposed in the 1920s by Anton Flettner
of Germany [16–18]. Although the operating principle has been known for a long time,
it has regained attention recently due to an emphasis on ship efficiency and emission
reduction. The rotor sail produces additional thrust, utilizing the Magnus effect around a
rotating cylinder [19]. When a fluid flows towards a rotating object, a decrease in pressure
occurs as the fluid velocity increases caused by rotation of the rotor. Conversely, at the
points where the wind direction and the rotational linear velocity of the rotor sail are
opposite, an increase in pressure occurs. As a result, additional thrust is generated from
the high-pressure to low-pressure areas, serving as auxiliary propulsion for the ship.
The rotor sail is gaining popularity as it can be installed as an add-on concept on
the ship, requiring minimal structural changes. By implementing a control system that
measures wind direction and intensity to automatically control the rotor sail, it is possible
to improve the energy efficiency of the ship by approximately 5–20% without the need for
additional crew members [20,21]. This trend of applying rotor sails is observed in large
vessels such as VLCCs (very large crude-oil carriers) and bulk carriers [22–25].
In 2015, the M/V Estraden vessel was equipped with rotor sails, which were 18 m in
height and had a diameter of 3 m. Two rotor sails were installed on the ship, generating
approximately 2 MW of propulsion power [26]. Anemoi installed 16 m high and 2 m
diameter rotor sails on a 64,000 DWT (deadweight tonnage) Ultra-Large Bulk Carrier for
operation. Additionally, Norsepower developed a 30 m height and 5 m diameter rotor sail
in 2018, which was installed on the Maersk Pelican vessel [11].
The research on rotor sail up to now has focused on the economic benefits and aerody-
namic performance analysis associated with the utilization of rotor sails. Traut conducted
study about the power-saving impact of a Flettner rotor on selected shipping routes [27].
From this study, we know that a Flettner rotor can contribute the propulsive power from
193 kW to 373 kW per one rotor. And ENERCON conducted an assessment of rotor sail
performance during a voyage from Germany to Portugal, and they reported a 23% fuel
consumption reduction by using the rotor sail [28]. In terms of the performance and design
optimization of rotor sails, research has predominantly focused on aerodynamical analysis.
Notably, Mittal conducted a study evaluating the performance of rotors with respect to
spin ratios using finite element analysis [29]. Additionally, Karabelas focused on research
to reveal the relations between forces at high spin rates with high-Reynolds-number flow
using the 2D numerical computations [30]. Craft assessed the impact and aerodynamic
performance of Thom disks of a rotor sail, utilizing CFD (computational fluid dynamics)
analysis [31]. However, there are several major technical challenges associated with not
only the aerodynamical performance but also dynamic response of WASP. These challenges
include ship stability issues, structural vibrations, constraints on visibility caused by the
rotor sail, cargo handling, control for effective energy efficiency performance, and rud-
der control [17,32]. Additionally, there is a possibility of transmitting additional loads to
the ship structure due to the rotating rotor sail, and conversely, structural and dynamic
characteristics issues may arise from vibrations originating from the engine and other
propulsion systems being transferred to the rotor sail. When retrofitting a rotor sail to an
already-constructed ship, these structural and dynamic characteristic issues can arise [10].
However, studies investigating the dynamic response of a rotor sail and its effect on a ship
structure have not been reported in the published literature yet.
To secure propulsion power utilizing the Magnus effect, rotor sails typically require
large rotating structures with a minimum height of 15 m. Although the rotational speed
is low, the rotational forces from the massive structure can cause structural damage and
the rapid wear of supporting bearings. Therefore, it is necessary to analyze the vibration
response from the rotating rotor sail. This includes evaluating the natural frequencies of
the rotating structure, assessing the possibility of resonance due to excitation sources such
as mass imbalance, and anticipating vibration responses in the rotating structure.
J. Mar. Sci. Eng. 2024, 12, 335 3 of 22
J. Mar. Sci. Eng. 2024, 12, 335 cylindrical rotor of the rotor sail is made of composite material, specifically Fiber-Rein- 4 of 22
forced Polymer (FRP), to reduce weight, while the inner tower, drive shaft, and mid-plate
are made of structural steel for marine applications.
Figure
Figure 11 shows
shows aa cross-sectional
cross-sectionalview viewofofthethe rotor
rotor sailsail structure
structure in this
in this study.study. The
The rotor
rotor
sail under investigation is a large rotating structure with a cylindrical rotor with 30 m30
sail under investigation is a large rotating structure with a cylindrical rotor with in
m in total
total height
height and and
3 m 3inmdiameter.
in diameter.The The mid-plate,
mid-plate, whichwhich connects
connects the cylindrical
the cylindrical rotorrotor
and
and the rotor
the rotor shaft,shaft, is located
is located 20 m 20 m above
above the bottom
the bottom of the of the cylindrical
cylindrical rotor.
rotor. The Thebearing
upper upper
bearing set consists
set consists of one
of one roller roller bearing
bearing and abearing,
and a thrust thrust bearing,
with the with
rollerthe rollerpositioned
bearing bearing po- at
sitioned
18.9 m andat 18.9 m andbearing
the thrust the thrust
at a bearing
height ofat18.6
a height
m. The oflower
18.6 m. The lower
bearing bearing 0.21
set is located set ism
located
above the0.21 m above
bottom and the bottom
supports theand supports
radial the radial
load. These load.
detailed These detailed
geometric dimensions geometric
refer to
dimensions refer to the ongoing development of the rotor sail in this
the ongoing development of the rotor sail in this study. The positions of support points study. The positions
and
of support sizes
geometric pointsareand geometric
relevant sizesfrom
not only are relevant
a structuralnotstandpoint
only from but a structural standpoint
also to aerodynamic
but also to aerodynamic
performance. Therefore,performance.
future research Therefore,
shouldfuture
focus onresearch shouldthe
optimizing focus on optimiz-
structure of the
ing thesail,
rotor structure of theboth
addressing rotorstructural
sail, addressing both
reliability andstructural
performancereliability and performance
considerations. In this
considerations. In this study,
study, we concentrate we concentrate
on proposing on proposing
two modeling two modeling
techniques techniques
and an analysis and
method,
an analysis
using method,
a reference usingtoainterpret
model referencedynamic
model toresponses.
interpret dynamic responses.
Figure
Figure1.1.Cross-sectional
Cross-sectionaldiagram
diagramof
ofthe
therotor
rotorsail.
sail.
The
Thecylindrical
cylindricalrotor
rotorand
andThomThomdisk
diskare
arecomposed
composedof ofaa composite
compositematerial
materialwith
withaa
thickness
thickness of 30 mm. The rotating structure of the rotor sail has a total mass of 18,000kg.
of 30 mm. The rotating structure of the rotor sail has a total mass of 18,000 kg.
Generally,
Generally,the
theinner
innertower
towerof ofthe
therotor
rotorsail
sailhas
hasaapartition
partitionwall
wallto tofacilitate
facilitatemaintenance.
maintenance.
However,
However,for
forthe
thesake
sakeofofanalysis
analysissimplification,
simplification,the
thepartition
partitionwall
wallwaswasneglected,
neglected,and
andthe
the
inner
innertower
towerwas
wasmodeled
modeledas asaacylindrical
cylindricalinner
innertower
towerwithwithananidentical
identicalmass
massand
andstiffness.
stiffness.
Detailed
Detailedinformation
informationon onthe
themass
massof ofthe
therotor
rotor sail
sail is
is provided
provided inin Table
Table 1.
1.
The cylindrical rotor operates at a maximum rotational speed of 180 RPM (revolu-
tions per minute), indicating relatively low RPM in terms of rotor dynamics. However,
due to the large diameter of the rotor at 3 m, the linear velocity of the rotor reaches 28.3
m/s, and if vibration occurs due to the high mass, significant vibrational energy can be
transmitted to the structural components of the vessel. Additionally, both the Thom disk
J. Mar. Sci. Eng. 2024, 12, 335 5 of 22
The cylindrical rotor operates at a maximum rotational speed of 180 RPM (revolutions
per minute), indicating relatively low RPM in terms of rotor dynamics. However, due to
the large diameter of the rotor at 3 m, the linear velocity of the rotor reaches 28.3 m/s, and
if vibration occurs due to the high mass, significant vibrational energy can be transmitted
to the structural components of the vessel. Additionally, both the Thom disk and the
cylindrical rotor are made of composite materials, which may introduce mass imbalance
due to manufacturing variations and eccentricity during the assembly process. In this
study, a simplified 4DOF rotor model was implemented to quickly calculate the vibration
response of the rotor sail, allowing for comparison with the vibration levels observed in
the actual system. The model’s validity and the range of its effectiveness were analyzed by
comparing the results with a finite element (FE) model and examining the modeling errors
due to variations in the stiffness at the support points.
Furthermore, stiffness of the upper support point can be considered as the structural
stiffness of the inner tower (k x,s1 , k y,s1 ) and the supporting bearing being connected in
series. Therefore, in Figure 2, it is assumed that there are two stiffnesses arranged in series
at the top. The structural stiffness related to the bending of the inner tower was determined
using the first lateral bending natural frequency ( f b ) obtained through modal analysis
using the finite element model of the inner tower. The relationship equation between the
equivalent stiffness (k eq ) and the natural frequency for a simple cantilever beam was used,
J. Mar. Sci. Eng. 2024, 12, 335 6 of 22
and the symmetric structural stiffness values (k x,s1 , k y,s1 ) are obtained by substituting the
natural frequency [38]. In the equation, m represents the total mass of the inner tower.
s
J. Mar. Sci. Eng. 2024, 12, x FOR PEER REVIEW 1 k eq 6 of 22
fb = 33
(2)
2π 140 m
Figure 2. Schematics
Figure 2. Schematics of
of the
the 4DOF
4DOF rotor
rotor sail
sail model.
model.
Figure 2 shows
The bearing the schematic
stiffness of thedirection
in the radial simplifiedis4DOF rotorconsidered
generally model mentioned above.
to be isotropic.
The bearing stiffness in the radial direction is generally considered to be isotropic.
Therefore, the upper bearing stiffness ( 𝑘 , , 𝑘 , ) and the lower bearing stiffness
Therefore, the upper bearing stiffness (k x,r1 , k y,r1 ) and the lower bearing stiffness (k x,r2 , k y,r2 )
(𝑘 , 𝑘 , ) can be defined symmetrically with respect to the x and y axes as follows:
can, be defined symmetrically with respect to the x and y axes as follows:
, , , ,
= =𝑘 (3)
k x,r1 k,x,s1 , k y,r1
, k y,s1
,
= = k1 (3)
k x,r1 + k x,s1 k y,r1 + k y,s1
𝑘 , =𝑘 , =𝑘 (4)
k x,r2 = k y,r2 = k2 (4)
The distance from the mass center to the upper bearing point is defined as 𝑎, the
distance
The from the mass
distance from center
the massto the lower
center tobearing
the upper point is defined
bearing pointasis𝑏,defined
and theas distance
a, the
from thefrom
distance massthe center mass tocenter
the Thomto the disk of the
lower rotor sail
bearing point is is
defined asas𝑐.b, and the distance
defined
from To
thederive the natural
mass center to thefrequencies
Thom disk of of the
the 4DOF model,
rotor sail we canasestablish
is defined c. the free vibra-
tion To derivecorresponding
models the natural frequencies to each degree of the 4DOF model,[39].
of freedom we canTheestablish
equationsthe of freemotion
vibration
for
models corresponding
the rotational motions along to each thedegree
x and of freedom
y axes [39].
include terms (𝐼of𝜔𝛼,
The equations
coupling 𝐼 𝜔𝛽)for
motion duethe
to
. .
the gyroscopic
rotational effect.
motions along the x and y axes include coupling terms (I p ω α, I p ω β) due to the
gyroscopic effect.
𝑚𝑥 + 𝑐 (𝑥 − 𝛼𝑎) + 𝑐 (𝑥 + 𝛼𝑏) + 𝑘 (𝑥 − 𝛼𝑎) + 𝑘 (𝑥 + 𝛼𝑏) = 0 (5)
.. . . . .
m x + c1 x − αa + c2 x + αb + k1 ( x − αa) + k2 ( x + αb) = 0 (5)
𝑚𝑦 + 𝑐 𝑦 + 𝛽𝑎 + 𝑐 𝑦 − 𝛽𝑏 + 𝑘 (𝑦 + 𝛽𝑎) + 𝑘 (𝑦 − 𝛽𝑏) = 0 (6)
..
. . . .
my + c1 y + βa + c2 y − βb + k1 (y + βa) + k2 (y − βb) = 0 (6)
𝐼 𝛽 + 𝑐 𝑦 + 𝛽𝑎 𝑎 − 𝑐 𝑦 − 𝛽𝑏 𝑏 − 𝑘 (𝑦 − 𝛽𝑏)𝑏 + 𝑘 (𝑦 + 𝛽𝑎)𝑎 + 𝐼 𝛼𝜔 = 0 (7)
.. . . . . .
IT β + c1 y + βa a − c2 y − βb b − k2 (y − βb)b + k1 (y + βa) a + IP αω = 0 (7)
𝐼 𝛼 − 𝑐 (𝑥 − 𝛼𝑎)𝑎 + 𝑐 (𝑥 + 𝛼𝑏)𝑏 − 𝑘 (𝑥 − 𝛼𝑎)𝑎 + 𝑘 (𝑥 + 𝛼𝑏)𝑏 − 𝐼 𝛽𝜔 = 0 (8)
.. . . . . .
IT α −
In this c1 x𝑚− represents
case, αa a + c2 xthe + αbtotal b −mass − αa
k1 ( x of + k2 (sail,
the) arotor x + αb )b − IP βω
including the=cylindrical
0 (8)
rotor, Thom disk, and mid-plate. 𝐼 denotes the transverse (in the X and Y directions)
moment of inertia of the rotating body, while 𝐼 represents the axial moment of inertia.
𝑐 , 𝑐 are the damping values at each bearing support point (see Table 2), and 𝜔 signifies
the rotational speed of the cylindrical rotor. The mass, moment of inertia, and geometry
information used in the 4DOF model are summarized in Table 2, which was obtained from
J. Mar. Sci. Eng. 2024, 12, 335 7 of 22
In this case, m represents the total mass of the rotor sail, including the cylindrical
rotor, Thom disk, and mid-plate. IT denotes the transverse (in the X and Y directions)
moment of inertia of the rotating body, while IP represents the axial moment of inertia.
c1 , c2 are the damping values at each bearing support point (see Table 2), and ω signifies
the rotational speed of the cylindrical rotor. The mass, moment of inertia, and geometry
information used in the 4DOF model are summarized in Table 2, which was obtained from
the three-dimensional geometry model of the rotor sail shown in Figure 3.
J. Mar. Sci. Eng. 2024, 12, x FOR PEER REVIEW 7 of 22
Symbols
Table 2. Modeling Parameters
parameter for the 4DOF rotor sail model. Values
m Mass of rotor 18,009 kg
Symbols Parameters Values
𝑚a Distance
Mass of rotor from CG (center of gravity) to the upper 18,009
0.5672 mkg
supporting point
Distance from CG (center of gravity) to the upper sup-
𝑎b Distance 0.5672mm
porting pointfrom CG to the lower supporting point 17.9728
𝑏c Distance
Distance fromfromCGCGto to thelower
the Thomsupporting
disk point 11.8207
17.9728mm
𝑐d Distance from
Distance fromCGCGto to
the
theThom disk
mid-plate 11.8207
1.8172 m m
𝑑 IP Distance
Momentfrom CG to along
of inertia the mid-plate
the rotor axis 1.8172 m
81,243.84 kgm2
𝐼I Moment
Momentof inertia along
of inertia the rotor
for lateral axis of the rotor
direction 81,243.84
940,097 kgmkgm2 2
T
𝐼 Moment of inertia for lateral direction of the rotor 940,097 kgm 2
kr11 Stiffness of the upper roller bearing 2.5 × 107 N/mm
𝑘 Stiffness of the upper roller bearing 2.5 × 107 N/mm
kr12 Stiffness of the upper thrust bearing 1.0 × 107 7N/mm
𝑘 Stiffness of the upper thrust bearing 1.0 × 10 N/mm
4
𝑘 kr2 Stiffness at the lower supporting
Stiffness at the lower supporting point point 2.0××10
2.0 104N/mm
N/mm
𝑐 c1 Damping
Damping coefficient
coefficient ofof uppersupporting
upper supporting point
point 1.0 103 3Ns/m
1.0××10 Ns/m
𝑐 c2 Damping
Dampingcoefficient ofoflower
coefficient lowersupporting point
supporting point 1.0××10
1.0 3
10 Ns/m
3 Ns/m
Figure 3.
Figure Three-dimensional geometry
3. Three-dimensional geometry model
model of
of the
the rotor
rotor sail.
The previous
The previous set
set of
ofequations
equations of
ofmotion
motioncan
canbe
berepresented
representedin
inmatrix
matrixform
formas
asfollows:
follows:
𝑚 0 0 0 𝑥 𝑐 +𝑐 0 0 𝑐 𝑏−𝑐 𝑎 𝑥
⎡ 0 𝑐 +𝑐 𝑐 𝑎−𝑐 𝑏 0 ⎤
0 𝑚 0 0 𝑦 𝑦
+⎢ ⎥ +
0 0 𝐼 0 𝛽 ⎢ 0 𝑐 𝑎−𝑐 𝑏 𝑐 𝑎 +𝑐 𝑏 𝐼 𝜔 ⎥ 𝛽
0 0 0 𝐼 𝛼 ⎣𝑐 𝑏 − 𝑐 𝑎 0 −𝐼 𝜔 𝑐 𝑎 +𝑐 𝑏 ⎦ 𝛼
𝑘 +𝑘 0 0 𝑘 𝑏−𝑘 𝑎 (9)
𝑥 0
⎡ 0 𝑘 +𝑘 𝑘 𝑎−𝑘 𝑏 0 ⎤
⎢ ⎥ 𝑦 0
=
⎢ 0 𝑘 𝑎−𝑘 𝑏 𝑘 𝑎 +𝑘 𝑏 0 ⎥ 𝛽 0
⎣𝑘 𝑏 − 𝑘 𝑎 0 0 𝑘 𝑎 +𝑘 𝑏 ⎦ 𝛼 0
Assuming a harmonic response for the displacement and angular responses, we can
express the free vibrational model of the rotor sail in a simple matrix multiplication form
J. Mar. Sci. Eng. 2024, 12, 335 8 of 22
.. .
x c2 b − c1 a x
m 0 0 0 ..
c1 + c2 0 0 .
0 m 0 0 y 0 c 1 + c 2 c 1 a − c 2 b 0 y
.. + . +
0 0 IT 0 β 0 c1 a − c2 b c1 a2 + c2 b2 IP ω β
.. .
0 0 0 IT α c2 b − c1 a 0 − IP ω c a2 + c b2 α
1 2 (9)
k2 b − k1 a
k1 + k2 0 0 x 0
0 k 1 + k 2 k 1 a − k 2 b 0 y 0
=
0 k 1 a − k 2 b k 1 a2 + k 2 b2 0 β 0
k2 b − k1 a 0 0 k 1 a2 + k 2 b2 α 0
Assuming a harmonic response for the displacement and angular responses, we can
express the free vibrational model of the rotor sail in a simple matrix multiplication form
as follows:
x = a1 est , y = a2 est , β = a3 est , α = a4 est (10)
h i
[ M ] s2 + [ C + G ] s + [ K ] { X } = {0} (11)
where [ M ] represents the inertia matrix, which consists of the mass and rotational inertia
components from the previous equation; [C + G ] represents the damping matrix, including
the gyroscopic term; and [K ] represents the stiffness matrix composed of the support
stiffness values.
In the free vibration model, the natural frequencies associated with the rotational
characteristics can be obtained as the imaginary parts of the solutions to the characteristic
equation, which can be expressed as follows:
h i
det [ M ]s2 + [C + G ]s + [K ] = 0 (12)
(∆m1 + ∆m2 )ω 2 r
(∆m1 + ∆m2 )ω 2 re− π2 i iωt
{ F} =
∆m dω 2 r − ∆m cω 2 r e− π2 i e
(13)
2 1
− ∆m2 dω 2 r − ∆m1 cω 2 r
When incorporating the imbalance force into the equation as an external excitation
term, it can be represented in matrix form as follows:
n .. o n.o
[ M] X + [C + G ] X + [K ]{ X } = { F } (14)
To perform numerical integration for calculating the response, new state variables Q
can be defined as follows: ( )
X
Q= . (15)
X
J. Mar. Sci. Eng. 2024, 12, 335 9 of 22
. I 0 0 0 I
Q= + Q (16)
0 [ M ] −1 F −[K ] −[C + G ]
.
After obtaining Q, the time response can be calculated by numerically integrating
it over time steps. To calculate the amplitude of the frequency response, the following
transfer function matrix with the force vector can be used:
h i −1
{X} = −[ M]ω 2 + [C + G ]iω + [K ] { F} (17)
In this study, the natural frequencies and frequency response results of the 4DOF
model were compared with the FEM analysis results to evaluate the limitations and validity
of the analysis.
was also performed to calculate the harmonic response due to imbalance mass.
The finite element analysis model of the rotor sail can be divided into the rotor part
and inner
depicted tower,5,
in Figure aswhere
shownthe in Figure
bottom4.surface
The boundary
is fixed to conditions of the inner
the foundation of thetower are as
ground.
Thedepicted in Figure
top surface 5, wheretothe
is subjected bottom
a point surface
mass is fixed
of 5000 to the foundation
kg, representing of the ground.
the weight of equip-The
top(e.g.,
ment surface is subjected
electric to a point
drive, gearbox, mass
etc.) of 5000
required forkg,
therepresenting
operation ofthetheweight of equipment
rotor sail. The top
(e.g.,
of the electric
inner tower drive,
and gearbox,
the drive etc.)
shaftrequired for thebyoperation
are supported of thewhile
two bearings, rotorthe
sail.bottom
The top
of of
the cylindrical rotor and the inner tower are connected through one bearing using a body-the
the inner tower and the drive shaft are supported by two bearings, while the bottom of
cylindrical
body connection rotorcondition.
and the inner
The tower
bearingarestiffness
connected through
in the radialone bearingisusing
direction a body-body
considered to
connection
reflect condition.
the support The bearing stiffness in the radial direction is considered to reflect
stiffness.
the support stiffness.
Figure
Figure 4. Model
4. Model of the
of the rotor
rotor sail.sail.
J. Mar. Sci. Eng. 2024, 12, 335 10 of 22
Figure 4. Model of the rotor sail.
In the
In the modal
modal analysis,
analysis, the the
thinthin structure
structure of the
of the ThomThom
diskdisk
cancan generate
generate multiple
multiple local
local
modes.
modes of However,
interest in these
this structural
study. local
Therefore, modes
the of
Thom the disk
disk was are
setnot
as relevant
modes. However, these structural local modes of the disk are not relevant to the rotorto
a rigid to the
structurerotor
modes the
prevent of interest in this
occurrence study.modes.
of local Therefore, the Thom disk
The cylindrical was
rotor, set as a rigid
mid-plate, structure
and drive shaftto
prevent the occurrence of local modes. The cylindrical rotor, mid-plate, and
were fully bonded to each other. The boundary conditions for the finite element analysis drive shaft
were
can be fully bonded
seen in Figureto6.each other. The boundary conditions for the finite element analysis
can be seen in Figure 6.
Figure
Figure6.6.Boundary
Boundaryand
andload
loadconditions
conditionsfor
forthe
therotor
rotormodel.
model.
InInthe
themodal
modalanalysis,
analysis,the
thenatural
naturalfrequencies
frequencieswere
werecomputed
computedfor
fora atotal
totalofofseven
seven
rotationalspeeds
rotational speeds(0(0RPM,
RPM,3030RPM,
RPM,6060RPM,
RPM,9090RPM,
RPM,120
120RPM,
RPM,150
150RPM,
RPM,180 180RPM).
RPM).ForFor
the harmonic response analysis considering imbalance, rotating forces were applied as
imbalance conditions at the positions of the Thom disk and the mid-plate. With various
magnitudes and phases of the imbalance, force analyses were repeatedly conducted. To
validate the reliability of the finite element model, a convergence test was conducted by
varying the number of elements. As shown in Figure 7, it was confirmed that the value of
J. Mar. Sci. Eng. 2024, 12, 335 11 of 22
the harmonic response analysis considering imbalance, rotating forces were applied as
imbalance conditions at the positions of the Thom disk and the mid-plate. With various
magnitudes and phases of the imbalance, force analyses were repeatedly conducted. To
validate the reliability of the finite element model, a convergence test was conducted
by varying the number of elements. As shown in Figure 7, it was confirmed that the
value of the second natural frequency converged to a certain value with an increase in the
number of elements. To enhance the efficiency of the analysis, the number of elements at
which the value saturated was applied to the model. The inner tower was modeled with
260,000 nodes and 130,000 elements. The rotor model was implemented with 70,000 nodes
J. J.Mar.
Mar.Sci.
Sci.Eng.
Eng.2024,
2024,12,
12,x xFOR
FORPEER
PEERREVIEW
REVIEW 1111ofof2222
and 50,000 elements. The overall model of the rotor sail, including the inner tower and
other components, consisted of 330,000 nodes and 180,000 elements, as shown in Figure 8.
7.95
7.9 Finalmodel
model
Final
7.85
7.8
7.75
7.7
0 0.5 1 1.5 2 2.5
Number of elements 10 5
Figure 7.Convergence
Figure7.7.
Figure Convergence
Convergence test
test
testforfor
for finite
finite element
element
finite model
model
element ofofrotor
of rotor
model rotor sail.
sail. sail.
Figure
Figure 8.Finite
Figure8.8. Finiteelement
Finite element
element model
model
model ofrotor
rotor
ofofrotor sail.
sail.
sail.
5. Modal Analysis
5.5.Modal
ModalAnalysis
Analysis
Natural Frequencies
NaturalFrequencies
Natural Frequencies
To understand the rotational vibration characteristics of the rotor sail, it is necessary to
Tounderstand
To
examine understand
the system’sthe therotational
rotational
eigenvalues, vibration
orvibration
natural characteristics
characteristics
frequencies, ofofthe
in the free therotor
rotorsail,
vibration sail,itDue
state. itisisnecessary
necessary
to
to
theexamine
gyroscopicthe system’s
effect, the eigenvalues,
natural or
frequencies natural
of the frequencies,
rotor change within
thethe free vibration
rotational
to examine the system’s eigenvalues, or natural frequencies, in the free vibration state. speed. state.
Duetotothe
Due thegyroscopic
gyroscopiceffect,
effect,the
thenatural
naturalfrequencies
frequenciesofofthe
therotor
rotorchange
changewith
withthe
therotational
rotational
speed. Therefore, the Campbell diagram, which plots the natural frequencies
speed. Therefore, the Campbell diagram, which plots the natural frequencies against the against the
rotationalspeed,
rotational speed,can
canbebeused
usedtotoassess
assessthethecharacteristics
characteristicsofofthetherotor.
rotor.InInthis
thisstudy,
study,toto
validate the approach of the simplified 4DOF model, the Campbell diagrams
validate the approach of the simplified 4DOF model, the Campbell diagrams of the 4DOF of the 4DOF
model and the finite element model were compared. Figure 9 shows the results of the
J. Mar. Sci. Eng. 2024, 12, 335 12 of 22
Therefore, the Campbell diagram, which plots the natural frequencies against the rotational
speed, can be used to assess the characteristics of the rotor. In this study, to validate the
approach of the simplified 4DOF model, the Campbell diagrams of the 4DOF model and
the finite element model were compared. Figure 9 shows the results of the modal analysis
using the finite element model. The major modes of rotational vibration in the rotor sail
are bending modes with respect to the support points, as observed in Figure 9. In the
first vibration mode shape of the finite element analysis result, the deflection is primarily
caused by horizontal displacement at the upper support point rather than bending of the 12 of 2
J. Mar. Sci. Eng. 2024, 12, x FOR PEER REVIEW
cylindrical rotor itself. This mode shape corresponds to the first mode natural frequency in
the Campbell diagram.
Figure
Figure 9. 9. Finiteelement
Finite element modal
modal analysis
analysisresults.
results.
The solid lines and markers shown in Figure 10 represent the natural frequencies
of The solidcalculated
the rotor lines andusing
markers
the 4DOFshown in Figure
model 10 rotational
at various representspeeds.
the natural frequencies o
The analysis
theresults
rotor using
calculated
the 4DOFusing the 4DOF
model show amodel
total ofatfour
various
natural rotational speeds.
frequencies, Thetwo
including analysis re
forward whirl natural frequencies (4DOF fn2, 4DOF fn4 in Figure 10)
sults using the 4DOF model show a total of four natural frequencies, including two for and two backward
whirl
ward natural
whirl frequencies
natural (4DOF (4DOF
frequencies fn1, 4DOF fn2,fn34DOF
in Figure
fn4 10). In rotor
in Figure dynamics,
10) and twothe backwar
natural frequencies of forward whirl and backward whirl are influenced by the gyroscopic
whirl natural frequencies (4DOF fn1, 4DOF fn3 in Figure 10). In rotor dynamics, the na
effect of the rotor, causing variations with respect to rotor RPM. For forward whirl, the
ural frequencies
natural frequency ofincreases
forwardwithwhirl and backward
an increase whirl
in RPM, while forare influenced
backward by natural
whirl, the the gyroscopi
effect of the rotor, causing variations with respect to rotor RPM. For
frequency decreases with an increase in RPM. However, in the case of the 1st mode shape forward whirl, th
natural
(4DOFfrequency
fn1 and fn2)increases with
in Figure 10, thean increase
rotor exhibitsin RPM, while
a cylindrical for backward
motion whirl,
centered around thethe natu
supports, and the gyroscopic effect has no impact on the variation of
ral frequency decreases with an increase in RPM. However, in the case of the 1st modthe natural frequency.
However,
shape (4DOF in fn1
the second
and fn2)mode
in shape
Figure(4DOF fn3 rotor
10, the and fn4), which aexhibits
exhibits a conical
cylindrical mode centere
motion
centered around the supports, it can be confirmed that there is a difference in the natural
around the supports, and the gyroscopic effect has no impact on the variation of the na
frequencies of forward whirl and backward whirl due to the gyroscopic effect.
ural frequency. However, in the second mode shape (4DOF fn3 and fn4), which exhibits
conical mode centered around the supports, it can be confirmed that there is a differenc
in the natural frequencies of forward whirl and backward whirl due to the gyroscopi
effect.
The dashed lines and markers in Figure 10 represent the natural frequencies obtaine
from the finite element analysis results, which only display the four major natural fre
quencies related to the global lateral vibration modes of the rotor. When comparing th
results of the finite element analysis with the 4DOF model, the natural frequencies of th
lower mode, specifically the cylindrical mode, show similar values around 4 Hz. Since th
cylindrical mode is not primarily influenced by the geometric bending of the rotor, it doe
not vary significantly with the rotational speed, resulting in similar results between th
4DOF model and the finite element model. However, for the 3rd and 4th modes, the 4DO
model no longer produces the same results. This can be attributed to the natural frequen
24, 12, x FOR PEER REVIEW 13 of 22
J. Mar. Sci. Eng. 2024, 12, 335 13 of 22
ings and five stiffness values for the lower bearing. The resulting first and 1000 second natural
Top Thrust Bearing 1 10 100 1500
frequencies with respect to the rotational speed can be observed in Figures 11 and 12.
(reference)
Figure 11 illustrates the variation in0.05
Bottom Bearing
natural frequencies
0.1
with
0.5
changes in2 the stiffness3
(reference)
of the upper bearings. It can be observed that the variations in the stiffness of the upper
bearings have little effect on the natural frequency values. This result is consistent for both
the 4DOF model and the Thisfinite
includes one case
element of hard
model, support and
indicating thethree cases ofin
similarity soft support.
trends Therefore, the
between
analysis was conducted for five different stiffness values for each of the two upper bearings
the two models. This can be attributed to the fact that the stiffness of the upper support
bearings does not significantly affect the natural frequencies as it operates in series with
the bending stiffness of the inner tower connected to it. Therefore, variations in the stiff-
ness of the bearings, which are relatively higher compared to the inner tower, do not have
a significant impact on the natural frequencies.
Figure 12 demonstrates the changes in natural frequencies with variations in the stiff-
sumption,
sumption,leading leadingto tosimilar
similarnatural
naturalfrequencies
frequenciesbetween
betweenthe thetwo
twoanalyses.
analyses.Therefore,
Therefore,in in
order
order to to approximate
approximate the the second
second mode
mode eveneven with
with the
the 4DOF
4DOF model,
model, the
the stiffness
stiffness ofof the
the
lower roller bearing needs to be set to a value less than 1/4 of the reference
lower roller bearing needs to be set to a value less than 1/4 of the reference stiffness. How- stiffness. How-
ever,
ever,considering
consideringthat thatthe
theoperating
operatingrange
rangeof ofthe
therotor
rotorsail
sailmainly
mainlyfalls
fallswithin
withinthe theopera-
opera-
J. Mar. Sci. Eng. 2024, 12, 335 14 of 22
tional
tionalspeed
speedrangerangewhere
wherethe thenatural
naturalfrequency
frequencyof ofthe
thefirst
firstmode
modeisisnot
notreached
reached(below
(below33
Hz),
Hz),ititisisexpected
expectedthatthathigher
higherfrequency
frequencymodesmodeswillwillhave
haveaalimited
limitedimpact
impacton onthethedynamic
dynamic
response of the
response of the rotor rotor sail. This will
sail.stiffness
and five This will be confirmed
be confirmed
values through
for the lowerthrough the results
theThe
bearing. results of the
of the
resulting harmonic
harmonic
first fre-
and secondfre-natural
quency
quencyresponse
responseanalysis
analysisin
frequencies the
thefollowing
inwith respect section.
to the
following rotational speed can be observed in Figures 11 and 12.
section.
Figure
Figure11.
11.Natural
Naturalfrequencies
frequencies
Figure of
ofthe
the
11. Natural rotor
rotorsail with
sailof
frequencies with respect
sailto
respect
the rotor toupper
upper
with support
support
respect stiffness.
stiffness.
to upper support stiffness.
Figure
Figure12.
12.Natural
Naturalfrequencies
frequencies
Figure of
ofthe
the
12. Natural rotor
rotorsail with
sailof
frequencies with respect
sailto
respect
the rotor tolower
lower
with support
support
respect stiffness.
stiffness.
to lower support stiffness.
6.6.Harmonic
HarmonicResponse
Response Figure 11 illustrates the variation in natural frequencies with changes in the stiffness
of the upper bearings. It can be observed that the variations in the stiffness of the upper
Typically,
Typically,thethe mass
massimbalance
bearings imbalance of
ofrotating
rotating
have little effect
bodies
bodies
on the natural
isiscontrolled
controlled
frequency
through
values. through
balancing
This resultbalancing
op-
op-
is consistent for both
erations
erationsbased
basedon the
onthethe permissible
permissible
4DOF model and residual
the finiteunbalance
residual unbalance amount
element model,amount defined
defined
indicating in
inISO
ISO[39].
the similarity [39]. How-
How-
in trends between
ever,
ever, rotor
rotor sails,
sails, being
the twolarge
being large rotating
models. This can
rotating structures composed
be attributed
structures to the factof
composed composite
that
of materials
the stiffness
composite of the upper
materials andsupport
and
bearings does not significantly affect the natural frequencies as it operates in series with the
bending stiffness of the inner tower connected to it. Therefore, variations in the stiffness
J. Mar. Sci. Eng. 2024, 12, 335 15 of 22
of the bearings, which are relatively higher compared to the inner tower, do not have a
significant impact on the natural frequencies.
Figure 12 demonstrates the changes in natural frequencies with variations in the
stiffness of the lower bearing. As the stiffness of the lower bearing gradually decreases
and becomes softer, the results of the 4DOF model analysis and the finite element analysis
become more similar. This occurs when the stiffness of the lower bearing falls below a
certain threshold, resulting in the dominance of the natural frequencies and mode shapes
influenced by the bearing stiffness rather than the bending stiffness of the cylindrical
rotor. In this case, the analysis results of the 4DOF model approximate the rigid shaft
assumption, leading to similar natural frequencies between the two analyses. Therefore, in
order to approximate the second mode even with the 4DOF model, the stiffness of the lower
roller bearing needs to be set to a value less than 1/4 of the reference stiffness. However,
considering that the operating range of the rotor sail mainly falls within the operational
speed range where the natural frequency of the first mode is not reached (below 3 Hz), it is
expected that higher frequency modes will have a limited impact on the dynamic response
of the rotor sail. This will be confirmed through the results of the harmonic frequency
response analysis in the following section.
6. Harmonic Response
Typically, the mass imbalance of rotating bodies is controlled through balancing oper-
ations based on the permissible residual unbalance amount defined in ISO [39]. However,
rotor sails, being large rotating structures composed of composite materials and heights of
up to 30 m, pose limitations on controlling residual unbalance using conventional balancing
equipment. Therefore, the mass imbalance of the rotor sail is likely to induce vibration in
the rotating body. In this study, the permissible residual unbalance amount for the rotor
sail was calculated based on the G6.3 standard used in applications such as pumps and
electric motors. This calculation was then applied to the rotor sail to evaluate its harmonic
response at various excitation frequencies. The permissible residual unbalance amount
varies depending on the rotor’s mass and rotational speed. For a maximum rotational
speed of 180 RPM and a total rotor weight of 18,009 kg, it can be calculated as 6.018 kgm.
There are two main locations where mass imbalance can occur in the rotor sail: Firstly,
at the topmost Thom disk, which has a wide plate-like structure relative to the cylindrical
rotor. If the rotation axis and Thom disk are offset during installation, it can result in mass
imbalance. Secondly, at the mid-plate, which is composed of ship structural steel and has
a higher mass concentration compared to the cylindrical rotor. In this study, harmonic
response analyses were conducted for three cases where imbalance mass could exist in the
Thom disk and mid-plate. Figure 13 illustrates these three mass imbalance conditions. Case
1 represents imbalance mass in the Thom disk, while case 2 represents imbalance mass in
the mid-plate. Since it is not feasible to perform a balancing check for the entire rotor sail,
which is a large rotating structure, static balancing was only conducted, but the possibility
of dynamic imbalance still exists. Therefore, case 3 considers the scenario where the Thom
disk and mid-plate have simultaneous imbalances of the opposite phase, representing a
dynamic imbalance condition. The analysis was performed using both the 4DOF model
and finite element analysis for all three conditions, and the results were compared and
analyzed. Figure 14 presents the mode shapes for each case as obtained from the finite
element analysis.
ducted, but the possibility of dynamic imbalance still exists. Therefore, case 3 considers
the scenario where the Thom disk and mid-plate have simultaneous imbalances of the
opposite phase, representing a dynamic imbalance condition. The analysis was performed
using both the 4DOF model and finite element analysis for all three conditions, and the
J. Mar. Sci. Eng. 2024, 12, 335 results were compared and analyzed. Figure 14 presents the mode shapes for each case as22
16 of
obtained from the finite element analysis.
Figure
Figure 14.14. Harmonic
Harmonic response
response finite
finite element
element analysis
analysis result
result of of rotor
rotor sail.
sail.
Figure
Figure 15 15 shows
shows thethe frequency
frequency response
response (harmonic
(harmonic response)
response) forfor
thethe three
three imbalance
imbalance
cases, specifically the vibration velocity results at the topmost section of the
cases, specifically the vibration velocity results at the topmost section of the inner tower, inner tower,
which corresponds to the support region of the upper bearing. The results
which corresponds to the support region of the upper bearing. The results shown as solid shown as solid
blue lines and circular markers represent the calculations from the 4DOF model, while thethe
blue lines and circular markers represent the calculations from the 4DOF model, while
redred triangular
triangular markers
markers represent
represent thethe results
results from
from thethe
FEMFEM analysis.
analysis. Considering
Considering that
that thethe
maximum rotation speed of the rotor sail is 180 RPM (3 Hz), the analysis focused on thethe
maximum rotation speed of the rotor sail is 180 RPM (3 Hz), the analysis focused on
harmonic response up to 3 Hz, as it represents the maximum excitation frequency due to
harmonic response up to 3 Hz, as it represents the maximum excitation frequency due to
mass imbalance.
mass imbalance.
From the results shown in Figure 15, it can be observed that the analysis results from
the 4DOF model and FEM exhibit relatively similar trends for each imbalance case. Since
the first natural frequency occurs around 4 Hz, the overall trend in the analysis results up
to 3 Hz shows a monotonic increase. However, it can be noted that Case 1 exhibits larger
deviations compared to Cases 2 and 3. This is due to the mass imbalance imposed solely
on the Thom disk, as shown in case 1 of Figure 14, which leads to bending of the rotating
part of the rotor sail. The FEM results consider this bending effect, resulting in higher
lateral displacements and velocities compared to the assumption of a rigid body rotation
in the 4DOF model. In Case 2, where mass imbalance occurs in the mid-plate, which is
closer to the center of gravity, the bending effect decreases, resulting in similar harmonic
J. Mar. Sci. Eng. 2024, 12, x FOR PEER REVIEW 17 of 2
J. Mar. Sci. Eng. 2024, 12, 335 17 of 22
Figure15.
Figure 15.Harmonic
Harmonicresponses
responses
of of three
three imbalance
imbalance cases
cases (Top:
(Top: Case
Case 1, Middle:
1, Middle: CaseCase 2, Bottom: Cas
2, Bottom:
3).
Case 3).
From
The the results velocity
vibration shown inresults
Figure 15,
foritthe
cancombined
be observed thatimbalance
mass the analysisconditions
results fromoccurrin
the 4DOF model and FEM exhibit relatively similar trends for each imbalance
in the Thom disk and mid-plate are shown in Figure 16. The mass imbalances case. Since
in the Thom
the first natural frequency occurs around 4 Hz, the overall trend in the analysis results up
disk and mid-plate are depicted in an out-of-phase condition, and the vibration velocit
to 3 Hz shows a monotonic increase. However, it can be noted that Case 1 exhibits larger
is plotted. The maximum vibration velocity occurs when the mass imbalance is concen
trated in the mid-plate.
Figure 17 presents the vibration velocity at the upper bearing support for the cas
where the mass imbalances in the Thom disk and mid-plate have the same phase (in-phas
condition) in the combined mass imbalance condition. The maximum vibration velocit
J. Mar. Sci. Eng. 2024, 12, 335 18 of 22
deviations compared to Cases 2 and 3. This is due to the mass imbalance imposed solely on
the Thom disk, as shown in case 1 of Figure 14, which leads to bending of the rotating part
of the rotor sail. The FEM results consider this bending effect, resulting in higher lateral
displacements and velocities compared to the assumption of a rigid body rotation in the
4DOF model. In Case 2, where mass imbalance occurs in the mid-plate, which is closer to
the center of gravity, the bending effect decreases, resulting in similar harmonic response
J. Mar. Sci. Eng. 2024, 12, x FOR PEERwithin
REVIEW the desired range of 3 Hz. 18 of 22
The vibration velocity results for the combined mass imbalance conditions occurring
in the Thom disk and mid-plate are shown in Figure 16. The mass imbalances in the Thom
disk and mid-plate are depicted in an out-of-phase condition, and the vibration velocity is
reaching approximately
plotted. The 10.6 mm/s.
maximum vibration This
velocity maximum
occurs when thevibration velocity
mass imbalance value can be evalu-
is concentrated
ated
in theby comparing it with the allowable vibration level at the support of the rotor sail.
mid-plate.
J. Mar. Sci. Eng. 2024, 12, x FOR PEER REVIEW 18 of 2
reaching approximately 10.6 mm/s. This maximum vibration velocity value can be evalu
ated by comparing it with the allowable vibration level at the support of the rotor sail.
Figure 16. Harmonic responses due to combined dynamic imbalances (out-of-phase condition).
Figure 16. Harmonic responses due to combined dynamic imbalances (out-of-phase condition).
Figure 17 presents the vibration velocity at the upper bearing support for the case
where the mass imbalances in the Thom disk and mid-plate have the same phase (in-
phase condition) in the combined mass imbalance condition. The maximum vibration
velocity occurs when the mass imbalances in the Thom disk and mid-plate are in the same
phase, reaching approximately 10.6 mm/s. This maximum vibration velocity value can be
Figure 16.by
evaluated Harmonic responses
comparing dueallowable
it with the to combined dynamic
vibration levelimbalances (out-of-phase
at the support of the rotorcondition).
sail.
Figure 17. Harmonic responses due to combined dynamic imbalances (in phase condition).
The vibration model of the rotor sail is linear, so the response under the condition of
combined mass imbalance is essentially the linear combination of the responses under
each individual mass imbalance condition. Therefore, we can expect the vibration velocity
Figure 17. Harmonic responses due to combined dynamic imbalances (in phase condition).
at the upper
Figure supportresponses
17. Harmonic to be predicted using the
due to combined following
dynamic simplified
imbalances linearcondition).
(in phase equation:
𝑦 = 0.237 ∙ ∆𝑚1 + 0.825 ∙ ∆𝑚2 (18)
The vibration model of the rotor sail is linear, so the response under the condition o
combined mass
In a rotor imbalance
sail, is vibrations
significant essentiallycan
theoccur
lineardue
combination
to the massofimbalance
the responses
of theunde
mid-
each individual
plate. mass
However, for imbalance
rotational condition.
structures Therefore,
made we can
of structural expect
steel, the vibration of
the management velocity
mass
J. Mar. Sci. Eng. 2024, 12, 335 19 of 22
The vibration model of the rotor sail is linear, so the response under the condition of
combined mass imbalance is essentially the linear combination of the responses under each
individual mass imbalance condition. Therefore, we can expect the vibration velocity at the
upper support to be predicted using the following simplified linear equation:
In a rotor sail, significant vibrations can occur due to the mass imbalance of the mid-
plate. However, for rotational structures made of structural steel, the management of mass
imbalance can be achieved using a balancing machine. On the other hand, in the case of
the Thom disk, which is made of composite material and assembled with the cylindrical
J. Mar. Sci. Eng. 2024, 12, x FOR PEER REVIEW
rotor, there can be eccentricity (offset) errors between the rotational axis and the Thom disk 19 of 2
(Figure 18). If we limit the vibration velocity to 6.3 mm/s, we can deduce from the previous
linear equation that a mass imbalance of up to 26.58 kg is permissible for the Thom disk.
Considering the eccentricity error as the eccentricity from the rotational axis to the total
model
massderived in this
of the Thom disk,study
with aallows
weightfor a quick
of 1096.9 kg,estimation of the
we can provide the response to rotor
guideline for the vibra
tions, making
maximum it useful
allowable for providing
eccentricity design
error to 24 criteria
mm. Thus, and vibration
the rotor guidelines during
model the desig
derived
in this study allows for a quick estimation of the response to rotor vibrations,
phase. Furthermore, by implementing this model in a real-time simulation environmen making it
useful for providing design criteria and guidelines during the design phase. Furthermore,
it can be applied to evaluate the structural integrity of the rotor sail by monitoring th
by implementing this model in a real-time simulation environment, it can be applied to
signal deviation between the measured signals in an actual rotor sail and the rotor sa
evaluate the structural integrity of the rotor sail by monitoring the signal deviation between
model.
the measured signals in an actual rotor sail and the rotor sail model.
Figure 18. Mass eccentricity of the Thom disk due to installation concentricity error.
Figure 18. Mass eccentricity of the Thom disk due to installation concentricity error.
7. Conclusions
In this study, two methods for analyzing the dynamic response of a rotor sail were
7. Conclusions
discussed. Firstly, the 4DOF simplified rotor sail model was derived to obtain low-order
In this study, two methods for analyzing the dynamic response of a rotor sail wer
natural frequencies and to assess dynamic responses quickly. And then, finite element
discussed. Firstly,
analysis was the 4DOF
introduced for thesimplified rotor sail
rotor sail to validate themodel was derived
effectiveness to simplified
of the 4DOF obtain low-orde
natural frequencies
rotor sail model. By and to assess
comparing dynamicapproaches,
two modeling responses we quickly. And then,analysis
could summarize finite elemen
resultswas
analysis as follows:
introduced for the rotor sail to validate the effectiveness of the 4DOF simpl
fied- rotor
Thesail
4DOFmodel.
modelBycan
comparing two modeling
estimate low-order approaches,
four natural we could
frequencies summarize
including two ana
forward whirl
ysis results as follows:natural frequencies;
- The 4DOF model can quickly estimate the first natural frequency for cylindrical mode
- The 4DOF
with aboutmodel
0.3 Hzcan estimate
difference withlow-order
FEA results;four natural frequencies including two for
- ward
The whirl natural
4DOF rotor sail frequencies;
model is valid for dynamic response analysis in designs where no
- The 4DOF
critical model
speed can i.e.,
occurs, quickly
belowestimate the first
the first natural natural frequency for cylindrical mod
frequency;
with about 0.3 Hz difference with FEA results;
- The 4DOF rotor sail model is valid for dynamic response analysis in designs wher
no critical speed occurs, i.e., below the first natural frequency;
- A bottom supporting bearing significantly effects on the dynamic responses of th
J. Mar. Sci. Eng. 2024, 12, 335 20 of 22
Appendix A
Table A1. Nomenclature in the manuscript.
References
1. Faber, S.; Hanayama, S.; Zhang, S. Fourth IMO GHG Study 2020-Final Report; International Maritime Organisation (IMO): London,
UK, 2020.
2. Wang, X.; Shipurkar, U.; Haseltalab, A.; Polinder, H.; Claeys, F.; Negenborn, R.R. Sizing and control of a hybrid ship propulsion
system using multi-objective double-layer optimization. IEEE Access 2021, 9, 72587–72601. [CrossRef]
3. The United Nations. Unsustainable Development Goals. Goal 14: Conserve and Sustainably Use the Oceans, Seas and Marine
Resources. 2020. Available online: https://www.globalgoals.org/goals/14-life-below-water/ (accessed on 7 February 2024).
4. The International Maritime Organization. Initial IMO Strategy on Reduction of GHG Emission from Ships. Annex 11. Resolution
MEPC.04.2018;304(72). Available online: https://unfccc.int/news/initial-imo-strategy-for-reducing-shipping-emissions-set-
for-adoption?gad_source=1&gclid=EAIaIQobChMIhueWv-KshAMVHNXCBB2HeQtOEAAYASAAEgIa5_D_BwE (accessed on 7
February 2024).
5. Castles, G.; Bendre, A. Economic benefits of hybrid drive propulsion for naval ships. In Proceedings of the 2009 IEEE Electric
Ship Technologies Symposium, Baltimore, MD, USA, 20–22 April 2009; pp. 515–520.
6. Sulligoi, G.; Castellan, S.; Aizza, M.; Bosisch, D.; Piva, L.; Lipardi, G. Active front-end for shaft power generation and voltage
control in FREMM frigates integrated power system: Modelling and validation. In Proceedings of the 21st International
Symposium on Power Electronics, Electrical Drives, Automation and Motion, Sorrento, Italy, 20–22 June 2012; pp. 452–457.
7. Tadros, M.; Ventura, M.; Soares, C.G. Review of current regulations, available technologies, and future trends in the green
shipping industry. Ocean. Eng. 2023, 280, 114670. [CrossRef]
8. Vidovic, T.; Simunovic, J.; Radica, G.; Penga, Z. Systematic overview of newly available technologies in the green maritime sector.
Energies 2023, 16, 641. [CrossRef]
9. Aakko-Saksa, P.T.; Lehtoranta, K.; Kuittinen, N.; Järvinen, A.; Jalkanen, J.P.; Johnson, K.; Jung, H.; Ntziachristos, L.; Gagné, S.;
Takahashi, C.; et al. Reduction in greenhouse gas and other emissions from ship engines: Current trends and future options. Prog.
Energy Combust. Sci. 2023, 94, 101055. [CrossRef]
10. Lindstad, E.; Stokke, T.; Alteskjær, A.; Borgen, H.; Sandaas, I. Ship of the future—A slender dry-bulker with wind assisted
propulsion. Marit. Transp. Res. 2022, 3, 100055. [CrossRef]
11. Wang, Y.; Zhang, X.; Lin, S.; Qiang, Z.; Hao, J.; Qiu, Y. Analysis on the Development of Wind-assisted Ship Propulsion Technology
and Contribution to Emission Reduction. IOP Conf. Ser. Earth Environ. Sci. 2022, 966, 012012. [CrossRef]
12. Petković, M.; Zubčić, M.; Krčum, M.; Pavić, I. Wind Assisted Ship Propulsion Technologies-Can they Help in Emissions Reduction?
Nase More 2021, 68, 102–109.
13. Tillig, F.; Ringsberg, J.W. Design, operation and analysis of wind-assisted cargo ships. Ocean Eng. 2020, 211, 107603. [CrossRef]
14. Ammar, N.R.; Seddiek, I.S. Wind assisted propulsion system onboard ships: Case study Flettner rotors. Ships Offshore Struct. 2021,
17, 1616–1627. [CrossRef]
J. Mar. Sci. Eng. 2024, 12, 335 22 of 22
15. Lu, R.; Ringsberg, J.W. Ship energy performance study of three wind-assisted ship propulsion technologies including a parametric
study of the Flettner rotor technology. Ships Offshore Struct. 2019, 15, 249–258. [CrossRef]
16. Bergeson, L.; Greenwald, C.K. Greenwald, Sail assist developments 1979–1985. J. Wind. Eng. Ind. Aerodyn. 1985, 19, 45–114.
[CrossRef]
17. Pearson, D.R. The use of flettner rotors in efficient ship design. In Proceedings of the Influence of EEDI on Ship Design Conference,
London, UK, 24–25 September 2014.
18. De Marco, A.; Mancini, S.; Pensa, C. Preliminary analysis for marine application of flettner rotors. In Proceedings of the 2nd
International Symposium on Naval Architecture and Maritime, Istanbul, Turkey, 23–24 October 2014.
19. Borg, J. Magnus Effect: An Overview of Its Past and Future Practical Applications; AD-A165 902; NAVSEA: Washington, DC, USA,
1986; Volume 1–2.
20. Kuuskoski, J. Norsepower rotor sail solution. In Proceedings of the International Conference on Maritime Energy Management,
Malmo, Sweden, 24–25 January 2017.
21. Chou, T.; Kosmas, V.; Acciaro, M.; Renken, K. A Comeback of Wind Power in Shipping: An Economic and Operational Review on
the Wind-Assisted Ship Propulsion Technology. Sustainability 2021, 13, 1880. [CrossRef]
22. Sun, W.; Tang, S.; Liu, X.; Zhou, S.; Wei, J. An Improved Ship Weather Routing Framework for CII Reduction Accounting for
Wind-Assisted Rotors. J. Mar. Sci. Eng. 2022, 10, 1979. [CrossRef]
23. Ma, R.; Wang, Z.; Wang, K.; Zhao, H.; Jiang, B.; Liu, Y.; Xing, H.; Huang, L. Evaluation Method for Energy Saving of Sail-Assisted
Ship Based on Wind Resource Analysis of Typical Route. J. Mar. Sci. Eng. 2023, 11, 789. [CrossRef]
24. Seddiek, I.S.; Ammar, N.R. Harnessing wind energy on merchant ships: Case study Flettner rotors onboard bulk carriers. Environ.
Sci. Pollut. Res. 2021, 28, 32695–32707. [CrossRef]
25. Zhang, P.; Lozano, J.; Wang, Y. Using Flettner Rotors and Parafoil as alternative propulsion systems for bulk carriers. J. Clean.
Prod. 2021, 317, 128418. [CrossRef]
26. Norsepower—Your Provider of Auxiliary Wind Propulsion. Available online: http://www.nor-sepower.com/ (accessed on 7
February 2024).
27. Traut, M.; Gilbert, P.; Walsh, C.; Bows, A.; Filippone, A.; Stansby, P.; Wood, R. Propulsive power contribution of a kite and a
Flettner rotor on selected shipping routes. Appl. Energy 2014, 113, 362–372. [CrossRef]
28. Schmidt, A.E.E. Ship 1 A Wind-Hybrid Commecial Cargo Ship. In Proceedings of the 4th Conference on Ship Efficiency, Hamburg,
Germany, 23–24 September 2013.
29. Mittal, S.; Kumar, B. Flow past a rotating cylinder. J. Fluid Mech. 2003, 476, 303–334. [CrossRef]
30. Karabelas, S.J.; Koumroglou, B.C.; Argyropoulos, C.D.; Markatos, N.C. High Reynolds number turbulent flow past a rotating
cylinder. Appl. Math. Model. 2012, 36, 379–398. [CrossRef]
31. Craft, T.J.; Iacovides, H.; Johnson, N.; Launder, B.E. Launder, Back to the future: Flettner-Thom rotors for maritime propulsion?
In Proceedings of the Seventh International Symposium on Turbulence Heat and Mass Transfer, Palermo, Italy, 24–27 September
2012; Volume 7.
32. Pan, P.; Sun, Y.; Yuan, C.; Yan, X.; Tang, X. Research progress on ship power systems integrated with new energy sources: A
review. Renew. Sustain. Energy Rev. 2021, 144, 111048. [CrossRef]
33. Lloyd’s Register. Guidance Notes for Flettner Rotor Approval; Lloyd’s Register Group Limited: London, UK, 2015.
34. Vance, J.M. Rotordynamics of Turbomachinery; Wiley: Hoboken, NJ, USA, 1991.
35. Prandtl, L. The Magnus effect and wind-powered ships. Naturwissenschaften 1925, 13, 1787–1806.
36. Thom, A. Effects of Discs on the Air Forces on a Rotating Cylinder; Reports & Memoranda 1623; Aerospace Research Council: Ottawa,
ON, Canada, 1934.
37. De Marco, A.; Mancini, S.; Pensa, C.; Calise, G.; De Luca, F. Flettner Rotor Concept for Marine Applications: A Systematic Study.
Int. J. Rotating Mach. 2016, 2016, 3458750. [CrossRef]
38. Lei, X.; Wang, Y.; Wang, X.; Lin, G.; Shi, S. Revisited on the Free Vibration of a Cantilever Beam with an Asymmetrically Attached
Tip Mass. Math. Probl. Eng. 2021, 2021, 8024879. [CrossRef]
39. ISO 21940-11:2016; Mechanical Vibration-Rotor Balancing-Part 11: Procedures and Tolerances for Rotors with Rigid Behaviour.
Switzerland, 2016. Available online: https://www.iso.org/standard/54074.html (accessed on 7 February 2024).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.