Content uploaded by Ammar Amory
Author content
All content in this area was uploaded by Ammar Amory on Jan 14, 2019
Content may be subject to copyright.
Content uploaded by Ammar Amory
Author content
All content in this area was uploaded by Ammar Amory on Jan 13, 2019
Content may be subject to copyright.
Modelling and CFD Simulation of a Micro
Autonomous Underwater Vehicle SEMBIO
Ammar Amory and Erik Maehle
University of Luebeck - Institute of Computer Engineering
Luebeck, Germany
Email: amory, maehle@iti.uni-luebeck.de
Abstract—The design of Autonomous Underwater Vehicles
(AUV) is very complex due to high demands of the marine
environment such as depth range, endurance, payload, and
energy efficiency. The hull design and its hydrodynamic resis-
tance is one of the most important factors directly affecting
power requirements and the maneuverability of the vehicle. The
minimizing of the hydrodynamic resistance of AUVs leads to
high energy efficiency. This paper presents and focuses on the
hydrodynamic analysis and fluid flow simulation of the SEMBIO
micro AUV using computational fluid dynamic (CFD). The SEM-
BIO design was mainly inspired by the hydrodynamic features
of a guitarfish. To complete the processes of study, SolidWorks
software was employed to establish a three-dimensional (3D)
model for the CFD simulation, ANSYS Fluent software was used
to analyze the SEMBIO model to get comparative data about
hydrodynamic properties. Finally, a comparison was carried out
between SEMBIO AUV and the more traditional MONSUN AUV
design. The results illustrated that hydrodynamic properties and
resistance forces of SEMBIO surpassed those of MONSUN.
I. INTRODUCTION
In recent years, the interest of using autonomous underwater
vehicles (AUV) in a wide range of applications and researches
of marine environment has been growing [1]. The design and
construction of micro AUVs is a very promising solution
to meet the different requirements of an AUV that aims to
be employed in swarm applications, it is small and easy to
handle as well as low cost, in addition, micro AUVs allow
the exploration of confined spaces. Several parameters of the
AUV design remain fixed such as drag, which is related to the
underwater fluid flow. The hull shape of the robot is designed
to be appropriate to the application requirements. Typically,
most common AUVs use a torpedo shape and streamline
hull that is used in high speeds [2]. An other type is non
torpedo shape that is used mostly in remotely operated vehicles
(ROVs) and for some special AUVs, which are usually used
for shorter duration missions or the inspection of other large
bodies underwater such as icebergs or dams [3].
The advantages of a streamline hull motivated a lot of
researchers to study the hydrodynamic characteristics of un-
derwater robots. The researchers have worked to evaluate op-
timum designs of AUVs in order to improve the performance
on water resistance. Some works have considered designing
different mechanical structures that will provide optimal hy-
drodynamic characteristics of a hull shape to minimize the
drag force [4]. Moreover, the streamlined hull is intended to
improve the energy efficiency of the vehicle by reducing the
Fig. 1. The SEMBIO micro AUV is small and streamline-shaped and is fully
actuated with six thrusters that allow forward and vertical movement.
drag. In addition, a number of works have developed a special
shape such as fish-like, dish shape, or spherical underwater
robots to achieve the adequate hull form design. In this regard,
an important tool in the recent years, Computational Fluid
Dynamics (CFD), plays an important role in AUV design [5].
The CFD analysis has been used extensively in underwater
vehicle design, since it can analyze the flow field around the
hull and helps to understand the complex physics of the flow
phenomena. It provides the designer with a visualization of the
flow field, which is difficult and requires a long time to obtain
by field experiments. Furthermore, CFD simulation supports
the designer in testing and proves a different model of hull
shape form a previously produced first prototype model for
tests.
This paper presents the streamlined design of the pressure
hull, which constitutes the main vehicle body in which every
component and electronic device is enclosed, and the 3D draw-
ings created with the CAD software SolidWorks. Furthermore,
the paper is focused on minimizing the hydrodynamic resis-
tance of our SEMBIO AUV to obtain high energy efficiency
which leads to less installed power and longer endurance. It
was achieved by designing a streamline-shaped hull that was
inspired by the hydrodynamic features of a guitarfish.
This paper is structured as follows. Section II describes
the 3D modelling for CFD analysis of SEMBIO. Section III
describes the preprocessing steps for CFD analysis to generate
the mesh model of the robot. Section IV characterizes the
main hydrodynamic parameters of SEMBIO and analyses its
motions in detail based on CFD, in addition, a comparison
between the hydrodynamic analysis of SEMBIO and the
torpedo shaped AUV MONSUN is made. Finally, Section V
and VI summarizes our conclusions and future work.
II. 3D MODELLING FOR CFD ANA LYSIS OF SEMBIO
The SEMBIO AUV is a small and an inexpensive AUV for
monitoring and inspection tasks in coastal and inland aquatic
areas (see Figure 1). SEMBIO is intended to be hand-sized,
and is focused towards being developed for AUV swarms with
a swarm capability that is characterized by cost-effectiveness,
robustness, and energy-efficiency. For the latter, SEMBIO is
e.g. provided with a solar and energy management system
(SEMS) to increase the mission time [6].
In this study, the selection of the guitarfish as a bio-inspired
model is justified on the following grounds. The guitarfish has
a highly effective forward propulsion thanks to the geometric
shape of its front part, particularly the shovel-shaped nose,
as well as a low-drag profile, as its body and fins have a
streamlined shape (see Figure 2) [7]. Its size also makes the
guitarfish a suitable inspiration, which is usually between 75
cm (similar to the size of a micro AUV of <1 m) and
170 cm long. The guitarfish differs from other species of
fish in that it does not need to move its body as it swims,
owing to the flatness and large size of its body (its swimming
thus resembles an underwater flight); this makes it a good
model for the design of a surface that can carry various
instruments such as sensors (i.e., payload) for exploring the
aquatic environment.
The 3D modelling of SEMBIO from the first drawing to
the end-outcome was achieved with the help of the SolidWorks
software, which is useful, as it allows alterations to be brought
to the model. The development of the 3D model was based
on a subjective ”resemblance” to capture the main features
and measurements of the guitarfish morphology. The SEMBIO
model was given the same dimensions as a guitarfish of small
size at 70 ×27 ×10 cm. The final version was achieved
after extensive effort, and is illustrated in Figure 3. This AUV
robot assembly consists of the vehicle body, upper cover,
front enclosure, and six thrusters (four vertically and two
horizontally) as propulsion system. More detailed information
concerning the architecture of SEMBIO’s power system, main
controller, depth control, and software can be found in [6],
[8].
Fig. 2. The frontal part of the shovelnose guitarfish.
Fig. 3. Representation of the SEMBIO model in 3D based on SolidWorks
Fig. 4. The MONSUN AUV designed for swarm applications such as
environmental monitoring tasks
III. CFD ANA LYSI S OF SEMBIO A ND MONSUN
A. Preprocessing and Geometry
Improvement and optimization of the 3D shape model were
achieved via CFD simulations, which were useful for the
assessment of the hydrodynamic properties of the proposed
vehicle and also provided guidelines for enhancing its design.
Meanwhile, the investigation and analysis of MONSUN have
also been significantly beneficial for the SEMBIO design in
terms of ensuring that design errors and ambiguities were
avoided. MONSUN as shown in Figure 4 is another swarm-
capable micro AUV which is developed at our Institute [9].
The full-scale geometrical 3D SEMBIO model was im-
ported into ANSYS Fluent in the SolidWorks format. The
chosen CFD solver was ANSYS Fluent to afford a central
role of ANSYS in the AUV design, enabling the execution
and visualization of recent CFD results. In ANSYS Fluent,
SEMBIO was fixed in a rectangular domain, which was built
and added for a water flow field. The dimensions of the pool
were 1 m×1 m×2 m to ensure that there was sufficient space
for a dynamic fluid flow (see Figure 5 and 6). MONSUN was
also examined in the same context for comparative purposes.
B. Mesh Generation
A very important factor in hydrodynamics analysis is the
flow field mesh. This is because the amount of mesh deter-
mines how well the hydrodynamics analysis performs and how
complex the computation is. As such, prior to commencing the
analysis, mesh smoothing was undertaken and the 3D model
was simplified prior to meshing. For instance, to make the
mesh simpler and to improve its quality, elements such as
motors and bolts were disregarded. As illustrated in Figure 7,
ANSYS Fluent with CFD-Mesh was employed to carry out
Fig. 5. Computerized representation of the flow domain associated with
SEMBIO
Fig. 6. Computerized representation of the flow domain associated with
MONSUN
the meshing procedure. Prior to initiating the CFD simulation,
it was necessary to define the face spacing that dictated the
surface area of the mesh that came into contact with the
surface of SEMBIO. The dimensions of face spacing had
to be sufficiently small to capture the SEMBIO geometry.
Furthermore, on the SEMBIO surface, the number of inflation
layers increased. The same work was suitably applicable to
MONSUN, with the results being presented in Figure 8.
Fig. 7. The SEMBIO mesh represented in 3D
C. Setup and Solving
The commercial CFD analysis code (CFD solver) was em-
ployed to model and solve the fluid flow surrounding the SEM-
BIO on the basis of ANSYS Fluent. For the purpose of these
calculations, the flow field and water pressure surrounding the
SEMBIO body were determined based on the motion mod-
eling of the fluid with the incompressible, isothermal RANS
(Reynolds-averaged NavierStokes) equations. These equations
are employed in the majority of studies due to the fact that,
Fig. 8. The MONSUN mesh represented in 3D
by comparison to the potential flow theory, they provide a
better treatment of viscous effects. The equations comprise a
standard solution of the ensemble-averaged, steady-state, and
3D Navier-Stokes equations characterizing the properties of
the flow, like velocity, pressure, temperature, and density. In
the present case, to enable the closure of the RANS system of
equations by ANSYS Fluent, the k−ωturbulence model was
selected, with kand ωrespectively denoting the turbulence
kinetic energy and the viscous dissipation rate. This model
helps to anticipate how the turbulence and flow behave around
the hull, and its selection for the purposes of the project was
justified, as it is a widely employed turbulence model for
engineering simulations that is robust and can be applied to
different flows. Furthermore, the computation time was cut by
using the symmetry condition. Table I presents an overview of
the requirements for the pre-processing stage, including mesh
generation.
TABLE I
PRINCIPAL CONDITIONS EMPLOYED IN THE NUMERICAL COMPUTATION
OF SEMBIO
Object Name Mesh
Water Domain size 1m×1m×2m
Turbulence model k−ωmodel
Reynolds number 22.47 ×105
AUV Dimensions 0.7m×0.1m×0.27m
Total no. of elements (nodes) 1,819,692 (334,157)
IV. RES ULT S OF CFD ANALYSIS AND THE COMPARISON
OF SEMBIO A ND MONSUN
The hull design was improved many times, with the final
results being based on driving in forward direction at a speed
of 1 m/s. ANSYS Fluent was employed to perform every step
of the vehicle’s hull analysis. Hence, CFD post-processing
from ANSYS Fluent supplied comprehensive data related to
the velocity and pressure distribution surrounding both SEM-
BIO and MONSUN. The streamline of the SEMBIO shape
and the movement and contact of the fluid flow within the
domain are presented in Figure 9. For comparison purposes,
Figure 10 presents the fluid flow of MONSUN.
As is shown in Figure 11, the pressure distribution surround-
ing SEMBIO is homogeneous, apart from a few tiny points.
The interaction between the fluid and SEMBIO is the reason
for the occurrence of maximal pressure, denoted by the red
Fig. 9. Illustration of the fluid flow during the CFD analysis of SEMBIO
Fig. 10. Illustration of the fluid flow during the CFD analysis of MONSUN
color, at the tip of the nose of the vehicle and on the side
edge of each thruster ring. Such areas of maximal pressure
are so few that they are almost insignificant. In the other parts
of the SEMBIO surface, the pressure is not as high and is
more or less uniform due to the smooth flow. Thus, SEMBIO
demonstrates a good hydrodynamic performance thanks to the
smooth streamline and balanced pressure distribution.
The pressure distribution surrounding MONSUN is shown
in Figure 12. It can be seen that maximal pressure occurs over
the whole tip of the nose, while minimal pressure, denoted by
Fig. 11. The distribution of pressure on the surface of the hull of SEMBIO
Fig. 12. The distribution of pressure on the surface of the hull of MONSUN
the blue color, occurs at the edge of the windward side of each
fin. A compression dysfunction is generated by the pressure
difference on the fins.
As presented in Figure 13, the velocity associated with
SEMBIO was measured over the entire cross-section of the
hull. It decreased slightly at the tip of the nose and subse-
quently increased and was maintained to the end of the hull,
where it decreased again at the starting point of the tail curve.
For comparative purposes, the Figure 14 illustrates the velocity
results of MONSUN.
Fig. 13. The distribution of velocity around the SEMBIO hull at a speed of
1 m/s
According to the velocity results, by comparison to MON-
SUN, the velocity of SEMBIO around the hull was greater
(1.164 vs 1.172). Figure 14 shows the flow behind each fin,
which was not present in SEMBIO, the latter generating a
lower proportion of drag due to its higher velocity. An identical
pattern of an increasing velocity across the bow curvature and
the curve of the tail section as well as a reduction in velocity
at the bow tip and tail end was exhibited by every type of
hull. The vehicle noses differed significantly between the two
models, the spot being greater on MONSUN compared to
SEMBIO and most likely causing a rise in drag.
The shape and speed of a body moving underwater deter-
mines the drag forces acting on that body. The drag force
Fig. 14. The distribution of velocity around the MONSUN hull at a speed
of 1 m/s
(water resistance) Fdcan be simulated and determined by
ANSYS Fluent based on a given velocity for movement in
the horizontal direction. More specifically, the amount of
force needed by a vehicle to attain a particular speed can be
determined by ANSYS Fluent. The following equation permits
the calculation of the drag coefficient Cd, which represents
a key hydrodynamic feature and is an important factor to
measure the drag force of the interactions between a vehicle
and fluid, and thus determines the hydrodynamic parameters
of the vehicle.
Cd=Fd
1
2ρv2A
where, the density of the liquid is represented by ρ, the
velocity of the body is denoted by v, the reference area (the
cross-sectional area) is denoted by A.
The value obtained for the drag coefficient associated with
the horizontal movement was 0.12. This value suggests that
SEMBIO has a more streamlined shape than MONSUN, as
the latter’s Cdwas 0.48. ANSYS Fluent was also used to
determine Fdassociated with movement in a horizontal direc-
tion, which in turn was needed to obtain the thrust force FT
for SEMBIO (Table II). In addition to theoretical calculations,
the determination of FTalso was done by empirical work
on the SEMBIO hull. Thus the empirical work involved the
submersion of SEMBIO at a depth of 0.5 m in a swimming
pool and a simple and approximate method was employed
to measure the force required to thrust SEMBIO in a forward
direction, with subsequent measurement of the speed. Table III
provides the values of the drag force yielded by these exper-
iments. As can be seen, there was no significant difference
between CFD and empirical results. However, there were two
factors that may have impacted the results: the fact that the
empirical method employed was simple and possibly prone to
measurement errors and waves in the swimming pool.
According to the CFD results, to reach a speed of
1 m/s, a minimum of 5.44 N had to be generated by
the horizontal thruster. In the case of MONSUN, a
drag force value of 7.06 N was indicated by ANSYS
Fluent as being necessary to move the vehicle in a
horizontal direction at a speed of 1 m/s. However, it is
not very reasonable to compare the two vehicles from this
perspective, since they differ in terms of cross-section
(SEMBIO = 0.01685 m2; MONSUN =
0.015531 m2) and 3D surface area
(SEMBIO = 0.34 m2; MONSUN = 0.29 m2). In other
words, in comparison to MONSUN, SEMBIO had a volume
that was 1.2 times larger, yet the results still turned out in
favor of SEMBIO. Furthermore, MONSUN was transferred
from real life into SolidWorks, and the real MONSUN was
closely approximated in the drawings. It was drawn on the
basis of a Myring form [10] that does not match exactly with
reality, i.e. the drawn MONSUN is not an accurate reflection
of the real MONSUN. In spite of these considerations,
however, the comparative analysis still highlighted that the
SEMBIO design is more advantageous and superior compared
to MONSUN.
TABLE II
DRAG F ORC E DATA OBTAI NE D FRO M CFD RES ULTS O F SEMBIO
Drag force-CFD results 1.95 3.41 4.48 5.44 6.52
Speed (m/s) 0.25 0.5 0.75 1.0 1.25
TABLE III
DRAG F ORC ES O BTAI NED F ROM E XP ERI ME NTAL R ESU LTS OF SEMBIO
Drag force exp. results 1.0 1.5 2.0 2.5 3.0 4.9 5.8
Speed (m/s) 0.20 0.32 0.41 0.52 0.60 0.72 1.0
V. CONCLUSION
This paper presents the 3D modelling of the SEMBIO AUV
that was implemented by SolidWorks, and a computational
fluid dynamic model that was built and conducted using the
ANSYS Fluent software. As a result of the CFD analysis,
pressure and velocity distributions over the AUV bodies were
obtained. CFD simulations also provided information on the
water flow around the AUVs, and drag forces were obtained as
well. The CFD results indicated that the streamlined SEMBIO
hull has a smaller drag coefficient than the torpedo shaped
MONSUN hull, which leads SEMBIO to consume less energy
or showing a higher energy efficiency.
VI. FUTURE WO RK
For future work, we should search for more accurate and
realistic numerical CFD simulation model of SEMBIO when it
is swimming in different types of motions like turning, diving
or any other motion, also in different turbulence models. In
addition, we will conduct the same study at different flow
speeds, e.g. 0.5,2.5and 4m/s for a broader comparison.
REFERENCES
[1] W. H. Wang, X. Q. Chen, A. Marburg, J. G. Chase, and C. E. Hann,
“A low-cost unmanned underwater vehicle prototype for shallow water
tasks,” in 2008 IEEE/ASME International Conference on Mechatronic
and Embedded Systems and Applications, Institute of Electrical and
Electronics Engineers (IEEE), Oct 2008.
[2] R. Panish and M. Taylor, “Achieving high navigation accuracy using
inertial navigation systems in autonomous underwater vehicles,” in
OCEANS 2011 IEEE - Spain, pp. 1–7, June 2011.
[3] B. W. Hobson, A. D. Sherman, and P. R. McGill, “Imaging and sam-
pling beneath free-drifting icebergs with a remotely operated vehicle,”
Deep Sea Research Part II: Topical Studies in Oceanography, vol. 58,
no. 1112, pp. 1311 – 1317, 2011.
[4] A. Alvarez, V. Bertram, and L. Gualdesi, “Hull hydrodynamic optimiza-
tion of autonomous underwater vehicles operating at snorkeling depth,”
Ocean Engineering, vol. 36, no. 1, pp. 105–112, 2009.
[5] K. Vasudev, R. Sharma, and S. Bhattacharyya, “A multi-criteria decision
based robust optimization model for design of AUVs,” in Proceedings
of the 7th International Conference on Advanced Computing and Com-
munication Technologies-India. ICACM (India)-2013. Inderscience, UK,
ISBN, vol. 1152615797, pp. 7–12, 2013.
[6] A. Amory and E. Maehle, “SEMBIO - a small energy-efficient swarm
AUV,” in OCEANS 2016 MTS/IEEE Monterey, pp. 1–7, IEEE, 2016.
[7] F. Fish and G. Lauder, “Passive and active flow control by swimming
fishes and mammals,” Annu. Rev. Fluid Mech., vol. 38, pp. 193–224,
2006.
[8] U. Behrje, A. Amory, B. Meyer, and E. Maehle, “System identification
and sliding mode depth control of the micro AUV SEMBIO,” in 50th
International Symposium on Robotics (ISR 2018), (Munich, Germany),
June 2018.
[9] B. Meyer, C. Renner, and E. Maehle, “Versatile sensor and communica-
tion expansion set for the autonomous underwater vehicle MONSUN,”
in Advances in Cooperative Robotics, pp. 250–257, World Scientific,
2017.
[10] D. Myring, “A theoretical study of body drag in subcritical axisymmetric
flow,” The Aeronautical Quarterly, vol. 27, no. 3, pp. 186–194, 1976.
