SMART COATINGS FOR DRAG REDUCTION IN

YACHTS
T Józsa, I.M. Viola, M Kashtalyan, E Balaras, B Kidd

To cite this version:

T Józsa, I.M. Viola, M Kashtalyan, E Balaras, B Kidd. SMART COATINGS FOR DRAG REDUC-
TION IN YACHTS. INNOVSAIL International Conference on Innovation in High Performance Sailing
Yachts, CVET and Ecole Navale, Jun 2017, Lorient, France. pp.283-288. �hal-01583597�

HAL Id: hal-01583597

https://hal.archives-ouvertes.fr/hal-01583597
Submitted on 7 Sep 2017

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 The Fourth International Conference on Innovation in High Performance Sailing Yachts, Lorient, France

SMART COATINGS FOR DRAG REDUCTION IN YACHTS

T. I. Józsa, Institute for Energy Systems, School of Engineering, University of Edinburgh, UK,
[email protected]
I. M. Viola, Institute for Energy Systems, School of Engineering, University of Edinburgh, UK,
[email protected]
M. Kashtalyan, Centre for Micro- and Nanomechanics, School of Engineering, University of
Aberdeen, UK, [email protected]
E. Balaras, Department of Mechanical and Aerospace Engineering, George Washington Uni-
versity, DC, USA, [email protected]
B. Kidd, AkzoNobel, International Paint Ltd., Felling, UK, [email protected]

In some sailing conditions, the friction drag of a yacht hull can account for more than half of the total
resistance. If the surface of the hull was not rigid but flexible as, for instance, the skin of a dolphin, it
would be possible to decrease the friction drag considerably. Compliant walls can decrease the friction
drag either by delaying the laminar to turbulent transition, or by interacting with a post-transition boundary
layer. In the present work the second of these two mechanisms is explored. We study both actively and
passively controlled surfaces with direct numerical simulations. We consider a canonical channel flow,
where the boundary layer in the half channel represents a thin section of the hull’s boundary layer. The
friction Reynolds number of the boundary layer is Reτ ≈ 180, and we show how the results can be scaled
to higher Reτ . We model the coating as an array of independent tiles, each attached to the hull by a spring
and a damper, and free to move only in the streamwise or the spanwise direction. A drag reduction of 25%
and 3% is achieved with an active and a passive control, respectively, of the proposed surface.

1 INTRODUCTION
2δ
The resistance of sailing crafts can be broken down into fric- D
tion drag and pressure drag. The friction drag is higher than y E
the pressure drag at low sailing speeds, and vice versa [1]. The x
pressure drag can be decreased by shape optimisation, while A F B 4
C z 3 πδ
the friction drag is typically accounted for by minimising the
wetted surface [2, 3, 4, 5, 6]. In fact, decreasing the friction 4πδ
drag for a given wetted surface is challenging.
Choi et al. [7] carried out experiments on a scaled model Figure 1: The computational domain and the coordinate sys-
of Australia II, the yacht that won the 1983 America’s Cup. tem (not in scale).
The experiments showed that a combination of riblets and
polymer additives can reduce the friction drag by as much
as 3.5%. However, the riblets must be designed for a specific tion, or by interacting with a post-transition boundary layer.
flow speed and direction, and outside of these conditions they The second of these strategies is the subject of this work.
lead to a drag increase. Choi et al. [11] and Kang&Choi [12] showed that actively-
Foeth [8] considered air lubrication of the hull. Air is kept controlled wall-normal displacements of a compliant wall can
around the hull by bubble injection, air cavities, or air films. enable drag reduction. This control, known as ‘opposition
This approach has similarities with the use of hydrophobic control’, is based on the flow velocity fluctuations in the near
surfaces, which are known to enable drag reduction [9]. The wall region. It was also proven, both analytically [13] and ex-
air lubrication enables 20% drag reduction. Unfortunately, perimentally [14], that passive compliant coatings can reduce
outside of the optimal flow conditions it can lead to drag in- the friction drag by damping the turbulent fluctuations of a
crease, and the technological realisation is challenging. fully-turbulent boundary layer.
In the present study we consider compliant coatings in- Recent studies, based on Direct Numerical Simulation
spired by the dolphin skin [10]. Compliant walls can decrease (DNS), tried to reproduce these experimental results [15, 16].
the friction either by delaying the laminar to turbulent transi- The proposed passive coatings was deformed in the wall-

INNOV'SAIL 2017
283
 The Fourth International Conference on Innovation in High Performance Sailing Yachts, Lorient, France

normal direction by the instantaneous variations of the fluid rigid wall, no-slip boundary conditions are prescribed on the
pressure. The drag reduction was not confirmed by these sim- C and D surfaces. As shown in the following two Sections
ulations. The pressure-driven deformations could not mimic (2.3 and 2.4), the no-slip boundary conditions are modified to
the active wall-normal opposition control because the wall study the effect of different active and passive coatings.
pressure is not correlated with the near-wall velocities. In- A time dependent streamwise pressure gradient is applied
deed, this was also anticipated by the pioneering study of Choi to keep the bulk Reynolds number Re = ub δ/ν = 2800 con-
et al. [11]. stant in the channel. The friction Reynolds number of the rigid
In this paper we consider in-plane deformations that are wall simulation is Reτ = uτ δ/ν ≈ 180. As showed in Sec-
driven by the wall shear stresses. The wall shear stresses are tion 3.4, the results can be scaled to higher Reτ based on the
directly correlated with the velocity field near the wall. The correlations between the coating dynamics and the turbulent
goal of our study is to identify beneficial coating dynamics, statistics. In the rigid wall simulation, the Reynolds number
and quantify the corresponding drag reduction. based on the centre line velocity (uc ) is Rec = uc δ/ν ≈ 3300.
The centre line velocity represents the free-stream velocity of
2 METHOD the water flow around the yacht. While ub is the same for ev-
ery simulation, uτ and uc depends on the resulting flow field.
We investigate turbulent channel flows with DNS. These high- The flow between the no-slip wall boundary C and the mid-
fidelity computational fluid dynamics simulations allow the dle of the channel represents the boundary layer around the
same confidence level as experimental measurements. yacht hull. The periodic boundary conditions on the A-B
Symbols with the superscript plus (+ ) indicate quantities surfaces allow to consider only a short streamwise section of
that are made dimensionless with the kinematic viscosity of the hull boundary layer without modelling its spatial devel-
the fluid ν, the friction velocity uτ and the density ρ. Any opment from the bow. Thus, the thickening of the boundary
other quantity is made dimensionless with the channel half layer within the considered section is neglected. On the upper
height δ, the bulk velocity ub and the density ρ. Angle brack- half of the channel, instead of a free-stream flow, a boundary
ets h i enclose time-averaged quantities, and the prime super- layer symmetric in-the-average is modelled. For the purpose
script (0 ) indicates the zero-mean fluctuating component of a of the present investigation, this has a negligible effect on the
time-varying quantity. The dot superscript (˙) stands for the boundary layer statistics.
temporal derivative.
Simulations were run on the UK national supercomputer
ARCHER. The wall time of each simulation was around 14 2.3 ACTIVE CONTROL
hours using 288 cores from 12 Intel E5-2697 v2 Ivy Bridge To identify beneficial in-plane wall deformations and the cor-
processors. responding maximum friction drag reduction, we consider
actively-controlled in-plane wall displacements following the
2.1 SOLVER approach of Choi et al. [11]. The compressible coating is
modelled by an array of independent tiles, where each tile cor-
We solve numerically the governing equations of Newtonian
responds to one grid cell. The velocity of each tile is set by
fluids for incompressible flow. The flow domain in a chan-
the wall boundary condition. For the reference rigid simula-
nel between infinite plates is discretised with a Cartesian grid
tion, the tile velocity is zero and we apply a no-slip boundary
(Figure 1). A fractional step method [17] is used for the time
condition. We then consider one of the two following con-
advancement, with a second-order implicit Crank-Nicolson
trol strategies, where the local coating velocity is actively set
scheme for the convective and the viscous terms in the wall-
based on the nearby flow velocity.
normal direction, and a third-order low-storage Runge-Kutta
scheme for the other terms. The spatial derivatives are dis-
cretised with a second-order central finite difference scheme Active u0 -control
on a staggered grid. The pressure Poisson equation is solved
directly using fast Fourier transforms in the two periodic di- The streamwise deformation velocity of the wall ξ˙ at time step
rections. For a more detailed description of the numerical n is equal in direction and in magnitude to the streamwise ve-
scheme and the implementation, we refer to the work of locity fluctuation u0 from the previous time step at a distance
Balaras [18]. After the flow has become fully developed, the yc from the wall:
time step ∆t+ ≈ 0.115 is kept constant. The presented statis- ξ˙n = u0 |n−1
yc . (1)
tics are the results of averaging over a period t+ ≈ 23000.

2.2 BOUNDARY CONDITIONS Active w0 -control

The computational box is 4πδ×2δ×4πδ/3 in the streamwise, The spanwise deformation velocity of the wall ζ̇ at time step
wall-normal and spanwise direction, respectively. The corre- n is opposite in direction and equal in magnitude with the
sponding grid has a number of grid points nx × ny × nz = spanwise velocity fluctuation w0 from the previous time step
290 × 251 × 290. at a distance yc from the wall:
Periodic boundary conditions are used on the domain faces
A-B and F -E (Figure 1). For the baseline simulation of a ζ̇ n = −w0 |n−1
yc . (2)

INNOV'SAIL 2017
284
 The Fourth International Conference on Innovation in High Performance Sailing Yachts, Lorient, France

2.4 PASSIVE CONTROL 3.0

u′+
RMS < u′+v ′+ >
2.5 ′+
vRMS
We model an anisotropic compressible coating as a series of Moser et al. [19]
′+
tiles, which can overlap to each other. Each tile is independent 2.0 wRMS
by the adjacent tiles. The tile has a mass m and it is attached to 1.5
the hull by a spring with a constant k, and a damper with a vis-
cous damping coefficient c. The tiles are free to move only in 1.0
the streamwise direction. The overlapping tiles are a discrete 0.5
model of a continuous compressible coating. Overlapped tiles 0.0
indicate a region where the coating is compressed, and sepa-
rated tiles indicate a region where the coating is stretched. −0.5
Each tile corresponds to one grid cell on the wall and has −1.0
0 20 40 60 80 100 120 140 160 180
a surface area ∆x∆y. At every time step, for each tile we y+
solve the streamwise wall deformation ξ, the velocity ξ˙ and
the acceleration ξ¨ due to the force F = τx ∆x∆y, where τx
Figure 2: Validation of the turbulent statistics of the rigid wall
is the streamwise wall shear stress. Therefore the boundary
simulation against literature data.
condition of the passive coating can be written as

mξ¨n + cξ˙n + kξ n = F n−1 . (3) found that the peak drag reduction with the u0 - and the w0 -
control were around 10% and 30%, respectively. They found
A passive spanwise control has not been considered. In
an optimal control distance of yc+ ≈ 10 for both controls. We
fact, in the passive control, the wall velocity is driven by the
believe that the differences in the control efficiency are proba-
wall shear stress, which has the same direction as the near-
bly caused by the detailed of the implementation, for instance
wall velocity. Once the wall has reached the initial deforma-
by the phase lag of one time step between the measured signal
tion due to the mean flow, the wall velocity fluctuation has the
and the actuation (see equations 1 and 2).
same direction as the near-wall velocity fluctuation. This sug-
gests that spanwise passive control cannot mimic the active
25
w0 −control, where the wall velocity fluctuation has opposite
direction as the near-wall velocity fluctuation (cf. equation 2).
20

3 RESULTS AND DISCUSSION

15
u′-control
DR [%]

The drag reduction (DR) is used to quantify the control effi-

ciency and it is defined as w′-control
10

∆pcontrolled − ∆prigid
DR = , (4) 5
∆prigid
where ∆p denotes the time-averaged pressure drop through 0
the channel. 0 2 4 6 8 10 12 14
yc+

3.1 VALIDATION OF THE REFERENCE CASE

Figure 3: Drag reduction of the active controls as a function
The reference simulation with a rigid wall is validated against of the control distance.
the DNS simulation of Moser et al. [19]. Figure 2 shows the
Root Mean Square (RMS) of the streamwise velocity fluc-
tuations (u0 ), the wall normal velocity (v 0 ) and the spanwise 3.3 PASSIVE CONTROL
velocity (w0 ) across the boundary layer. All velocity statistics
show a good agreement with Moser et al. Excellent agree- The simulations of a wide range of coating parameters m,
ment is also found for the Reynolds stress hu0 vi. c and k suggest that the maximum drag reduction is around
3%. Further research is currently ongoing to identify the op-
timal coefficients and the corresponding drag reduction. In
3.2 ACTIVE CONTROL
the following, we analyse the flow field achieved with a set
We search for the optimal control distance yc that enables the of parameters (m = 15∆x∆y/Reb , c = ∆x∆y/Reb and
two active controls described by equations 1 and 2. Figure 3 k = ∆x∆y/Reb ) that results in a drag reductions of 2.87%.
shows the drag reduction as a function of yc . The u0 -control The deformation of the coating has only a small influence
results in a peak drag reduction around 8% at yc+ = 8. The on the pressure gradient, while it causes well distinguish-
w0 -control is more efficient than the u0 -control, with a peak able changes in the turbulent statistics. Figure 4 shows the
drag reduction around 25% at yc+ = 12. These results are first component of the Reynolds stress tensor hu0 u0 i the RMS
in qualitative agreement with those of Choi et al. [11], who of the wall-normal vorticity fluctuations ωy0 and of the span-

INNOV'SAIL 2017
285
 The Fourth International Conference on Innovation in High Performance Sailing Yachts, Lorient, France

wise vorticity fluctuations ωz0 . Since the wall of the compli- 4π/3

ant coating is moving in the streamwise direction, at the wall

hu0 u0 i 6= 0. This results in a higher peak of hu0 u0 i in the buffer
2π/3

z
layer compared to the rigid wall. The spanwise vorticity fluc-
tuation ωz0 = ∂v 0 /∂x − ∂u0 /∂y is dominated by the second
term, −∂u0 /∂y. As the coating moves in the streamwise di- 0
0 π 2π 3π 4π
rection, the wall-normal gradient of the streamwise velocity x
fluctuations (∂u0 /∂y) drops down, and results in a decreased
spanwise vorticity. −0.250 −0.125 0.000 0.125 0.250
ξ˙

8 ′
0.4 Figure 6: Instantaneous deformation velocity of the coating.
RMS(ωz+) < u′+u′+ >
7

6 0.3 3.4 SCALING AND RESULTS

RMS(ωy+), RMS(ωz+)
′
5
< u′+u′+ >

Our results show that the tile spacing and dynamics should
4 0.2 be correlate with those of the streaks. The streaks scale with
the wall variables ν/uτ and uτ . The spanwise spacing of the
′

3
RMS(ωy+)
′
streaks is roughly 100 ν/uτ . The streamwise spacing varies
2 0.1 from 1000 ν/uτ to 10 000 ν/uτ [21]. To estimate the required
rigid
1 tile spacing for a full scale yacht under realistic sailing con-
coating
ditions, we assume that the deformations and the deformation
0 0.0
0 10 20 30 40 velocities of the coating scale with those of the streaks.
y+ Using a 1/7th power law [22], we can estimate the friction
Reynolds number (Reτ ) along different positions of the hull,
Figure 4: Velocity and vorticity statistics in the first quarter of for different sailing speeds. The viscous length scale ν/uτ
the boundary layer. decreases with the sailing speed, and increases with the dis-
tance from the bow. Recalling that the present simulations
have been performed at Reτ ≈ 180, our results are represen-
The high and low momentum regions of the streamwise ve- tatives of the boundary layer at 0.2 m and 0.02 m from the
locity, also known as buffer-layer streaks, are one of the most bow for a sailing speed of 1 m/s and 5 m/s, respectively. Let
distinctive flow features of turbulent boundary layer flows scale our results to the boundary layer at 10 m from the bow
[20]. These high and low momentum regions leave a foot- of a yacht sailing at 5 m/s. In these conditions, the boundary
print on the wall shear stress at the wall. The streamwise layer thickness is ca. 0.12 m, the friction Reynolds number is
high and low shear regions are visualised in Figure 5 by the Reτ = 2 · 104 and the viscous length scale is 6 · 10−6 m. The
contours of the wall-normal gradient of the streamwise ve- spanwise tile spacing should be lower than the streak spacing,
locity fluctuation at the wall. Since the wall deformation ve- which is 6 · 10−4 m. Scaling the deformation velocity of the
locity tries to follow the shear originating from the streaks, wall with the friction velocity, we find that tiles within a span-
the wall velocity also shows a similar pattern than the wall wise distance of 0.5 millimetre would move in opposite direc-
shear stress (Figure 6). The wall velocity results in higher tions with an average streamwise velocity of 0.3 m/s. Scaling
production of wall-normal vorticity fluctuation near the wall: the deformation based on the viscous length scale, leads to a
ωy0 = ∂u0 /∂z − ∂w0 /∂x. maximum streamwise deformation of 5 · 10−3 m.
Assuming that the drag reduction is the same at higher
Reynolds numbers, we can estimate the speed gain enabled
4π/3
by a 3% drag reduction. Let consider, for instance, an Amer-
ica’s Cup Class yacht. This class was used in the America’s
2π/3
Cup from 1992 to 2007. At the typical upwind speed of 5 m/s,
z

the friction drag is approximately half of the resistance [6]. In

these conditions, a 3% change in the friction drag would result
0 in a speed increase of almost 1%.
0 π 2π 3π 4π
x

−6 −3 0 3 6
4 CONCLUSIONS
∂u′/∂y
We investigated the potential hydrodynamic drag reduction of
Figure 5: Instantaneous distribution of the wall-normal gradi- compliant coatings. We performed direct numerical simula-
ent of the streamwise velocity fluctuation at the wall. tions of a fully turbulent channel flow at friction Reynolds
number Reτ ≈ 180 and we scaled the results to higher
Reynolds number conditions. We modelled the hull surface

INNOV'SAIL 2017
286
 The Fourth International Conference on Innovation in High Performance Sailing Yachts, Lorient, France

as an array of tiles. Each tile is smaller than a squared vis- [6] I. M. Viola, R. G. J. Flay, and R. Ponzini, “CFD anal-
cous length and can move only in the streamwise or spanwise ysis of the hydrodynamic performance of two candi-
direction, depending on the control. date Americas Cup AC33 hulls,” Transactions of the
We found that, if the in-plane displacement of each tile is Royal Institution of Naval Architects Part B: Interna-
actively prescribed based on the flow velocity fluctuations at tional Journal of Small Craft Technology, vol. 154, no. 1,
y + ≈ 10, than the friction drag would decrease by up to 25%. pp. 1–12, 2012.
We also considered a compliant coating that was passively
deformed by the streamwise wall shear stress. Each tile is at- [7] K.-S. Choi, G. E. Gadd, H. H. Pearcey, A. M. Savill, and
tached to the hull by a spring and a damper, and free to move S. Svensson, “Tests of drag-reducing polymer coated on
only in the streamwise direction. This passive anisotropic a riblet surface,” Applied Scientific Research, vol. 46,
coating enabled a drag reduction of ca. 3%. no. 3, pp. 209–216, 1989.
The drag reduction is the result of the interaction of the [8] E. J. Foeth, “Decreasing frictional resistance by air lu-
coating with the buffer-layer streaks. Noting that the scala- brication,” in 20th International HISWA Symposium on
bility of these flow features with the Reynolds number is well Yacht Design and Yacht Construction, Amsterdam, The
established, it is possible to scale the present results to higher Netherlands, 2008.
Reynolds numbers. As an example, we scaled the results for a
hull section at 10 m from the bow and a sailing speed of 5 m/s. [9] J. P. Rothstein, “Slip on superhydrophobic surfaces,”
We found that two points of the passive coating, that are less Annual Review of Fluid Mechanics, vol. 42, pp. 89–109,
than half millimetre apart in the streamwise direction, would 2010.
move in opposite directions with an average velocity of 0.3 [10] M. Gad-el Hak, “Compliant coatings for drag reduc-
m/s. Each point would move by maximum 5 · 10−3 m. tion,” Progress in Aerospace Sciences, vol. 38, no. 1,
For an America’s Cup Class monohull, this passive coating pp. 77–99, 2002.
would enable a speed increase of about 1%. Ongoing research
aims to identify a passive coating, with different mechanical [11] H. Choi, P. Moin, and J. Kim, “Active turbulence control
properties, that would enable higher drag reduction. for drag reduction in wall-bounded flows,” Journal of
Fluid Mechanics, vol. 262, pp. 75–110, 1994.
5 ACKNOWLEDGEMENT [12] S. Kang and H. Choi, “Active wall motions for skin-
friction drag reduction,” Physics of Fluids, vol. 12,
This work received funding from AkzoNobel’s Marine Coat-
no. 12, pp. 3301–3304, 2000.
ings business (International Paint Ltd.), the Energy Technol-
ogy Partnership [ETP106]. This work used the ARCHER UK [13] P. W. Carpenter and P. J. Morris, “The effect of
National Supercomputing Service (http://www.archer.ac.uk). anisotropic wall compliance on boundary-layer stability
and transition,” Journal of Fluid Mechanics, vol. 218,
REFERENCES pp. 171–223, 1990.
[14] T. Lee, M. Fisher, and W. H. Schwarz, “Investigation
[1] L. Larsson and R. Eliasson, Principles of yacht design.
of the stable interaction of a passive compliant surface
Intl Marine Pub, 1995.
with a turbulent boundary layer,” Journal of Fluid Me-
[2] M. Caponnetto, A. Castelli, P. Dupont, B. Bonjour, P. L. chanics, vol. 257, p. 373, 1993.
Mathey, S. Sanchi, and M. L. Sawley, “Sailing yacht de-
[15] T. Endo and R. Himeno, “Direct numerical simulation
sign using advanced numerical flow techniques,” in 14th
of turbulent flow over a compliant surface,” Journal of
Chesapeake Sailing Yacht Symposium, Annapolis, MD,
Turbulence, vol. 3, 2002.
USA, 1999.
[16] S. Xu, D. Rempfer, and J. Lumley, “Turbulence over a
[3] L. Imas, G. Buley, B. Baker, and B. Ward, “CFD-
compliant surface: numerical simulation and analysis,”
based hydrodynamic analysis of high performance rac-
Journal of Fluid Mechanics, vol. 478, pp. 11–34, 2003.
ing yachts,” in 3rd High Performance Yacht Design Con-
ference, Auckland, New Zealand, 2008. [17] J. Van Kan, “A second-order accurate pressure-
correction scheme for viscous incompressible flow,”
[4] R. Azcueta and S. R., “6 degrees of freedom CFD ap-
SIAM Journal on Scientific and Statistical Computing,
plied to the design of an IMOCA Open 60,” in The 2nd
vol. 7, no. 3, pp. 870–891, 1986.
International Conference on Innovation in High Perfor-
mance Sailing Yachts, Lorient, France, 2010. [18] E. Balaras, “Modeling complex boundaries using an
external force field on fixed Cartesian grids in large-
[5] P. Groenenboom, B. Cartwright, P. de Luca, eddy simulations,” Computers & Fluids, vol. 33, no. 3,
A. Kamoulakos, D. McGuckin, and L. Olivari, pp. 375–404, 2004.
“Simulation based design for high performance com-
posite sailing boats,” in The Second International [19] R. D. Moser, J. Kim, and N. Mansour, “Direct numerical
Conference on Innovation in High Performance Sailing simulation of turbulent channel flow up to Reτ = 590,”
Yachts, Lorient, France, 2010. Phisics of Fluids, vol. 11, no. 4, 1999.

INNOV'SAIL 2017
287
 The Fourth International Conference on Innovation in High Performance Sailing Yachts, Lorient, France

[20] S. J. Kline, W. C. Reynolds, F. A. Schraub, and P. W. Prof. Maria Kashtalyan is Director of the Centre for
Runstadler, “The structure of turbulent boundary lay- Micro-and Nanomechanics, and Professor at the School of
ers,” Journal of Fluid Mechanics, vol. 30, no. 04, Engineering of the University of Aberdeen. She authored
pp. 741–773, 1967. more than 160 publications in the areas of solid mechanics
and modelling deformation and fracture of anisotropic
[21] J. Jiménez and R. D. Moser, “What are we learning and heterogeneous materials, including 60 papers in peer-
from simulating wall turbulence?,” Philosophical Trans- reviewed international scientific journals, 5 chapters in
actions of the Royal Society of London A: Mathematical, major reference books on composites, and 40 papers in
Physical and Engineering Sciences, vol. 365, no. 1852, peer-reviewed conference proceedings. Three of her papers
pp. 715–732, 2007. were included into Top 25 most read articles of the European
[22] H. Schlichting and K. Gersten, Boundary-layer theory. Journal of Mechanics A/Solids (July September 2011),
Springer, 1960. International Journal of Solids and Structures (April June
2007) and European Journal of Mechanics A/Solids (July
2004 March 2005).
6 AUTHORS BIOGRAPHY

Mr Tamás István Józsa is a PhD student at the University Prof. Elias Balaras is a Professor at the Department
of Edinburgh, working under the supervision of Dr Viola. of Mechanical and Aerospace Engineering at the George
He earned a mechanical engineering bachelor’s degree at Washington University. His current research program aims at
the Budapest University of Technology and Economics the development of robust numerical techniques for parallel,
(BUTE) and two master’s degrees in mechanical engineering large-scale simulations of multiscale, multiphysics problems
modelling and in computational fluid dynamics at BUTE in physical and biological systems. Emphasis is given at
and Cranfield University. His current PhD project aims to large-eddy and direct numerical simulations, fluid-structure
investigate the skin friction reduction potential of compliant interactions and biological fluid dynamics. He has been
coatings for marine applications. The supervisory team the recipient of several awards including the Marie-Curie
includes Prof. Kashtalyan, Dr Kidd and Prof. Balaras. Mr fellowship from the European Commission in 1994 and the
Józsa has been the recipient of numerous awards, including CAREER award from the National Science Foundation in
the Pattantyús-Pálffy Prize for the best bachelor’s thesis 2003. Dr. Balaras has published over 100 papers in refereed
with a hydrodynamics topic at BUTE; a travel grant from journals and conference proceedings, and served as reviewer
International Paint Ltd.; an EPSRC ARCHER RAP award; for numerous journals and government programs related
the best poster in the Marine Energy Theme at the 2016 ETP to fluid mechanics, biological flows, high performance
Annual Conference; and the Principals Go Abroad Fund. computing and turbulence. His research has been featured in
He has been a Visiting Scholar at the George Washington several media outlets, including the New York Times and the
University for 2 months in 2015 and 2 months in 2017. Sunday Times.

Dr Ignazio Maria Viola is Senior Lecturer at the Institute Dr Barry Kidds is a Section Manager at AkzoNobels Ma-
for Energy Systems of the School of Engineering, University rine Coatings business (International Paint Ltd.) based in the
of Edinburgh and a Fellow of the Royal Institution of Naval UK. He graduated from Heriot-Watt University with an Hon-
Architects. He is Editor-in-Chief of the Journal of Sailing ors degree in Applied Chemistry followed by a PhD from The
Technology (SNAME), and Member of the Editorial Board University of Manchester, Institute of Science and Technol-
of Ocean Engineering (Elsevier) and the International Jour- ogy. Barry has over 20 years of experience in the research
nal of Small Craft Technology (RINA). His background is in and development of coatings technologies with current focus
yacht sail aerodynamics, and his current research focuses on on fouling control coatings. Barry has led several technical
the control of vortical flow structures by means of flexible sur- projects with the aim of developing new products as well as
faces. He wrote more than 100 scientific publications and was managing the development of AkzoNobels marine coatings
awarded two RINA Medals of Distinctions and a RINA Medal R+D laboratory in Singapore. Over the last 6 years Barry has
of Exceptional Merit. managed the Hydrodynamics Team which focusses on devel-
oping a greater understanding of how fouling control selection
affects ship efficiency. AkzoNobel create everyday essentials
to make people’s lives more liveable and inspiring. Team Ak-
zoNobel is the first to enter the forthcoming edition of Volvo
Ocean Race.

INNOV'SAIL 2017
288