Applied Energy 155 (2015) 23–34

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Viability of the application of marine current power generators in the

south Brazilian shelf
Eduardo de Paula Kirinus a,⇑, Wiliam Correa Marques b
a
Instituto de Oceanografia, Universidade Federal do Rio Grande Campus Carreiros, Av. Itália s/n Km 8, CEP 96201-900, Brazil
b
Instituto de Matemática, Estatística e Física, Universidade Federal do Rio Grande Campus Carreiros, Av. Itália s/n Km 8, CEP 96201-900, Brazil

h i g h l i g h t s

 The authors found the best location on the southern Brazilian shelf for the installation of hydrokinetic turbines.
 Two different scenarios using farm of hydrokinetic turbines were simulated.
 The current pattern and energy conversion rates were studied in both scenarios.
 Temporal and spatial variability of the energy conversion were analyzed.
 The results indicate that wake pattern affect nearby turbines impacting the conversion rates.

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

Article history: The continuing growth in the world population has increased the demand and competition for energy,
Received 17 March 2014 such that immense efforts are required to make non-renewable energy sources available. Using marine
Received in revised form 5 May 2015 currents to generate electricity offers a distinct advantage over other renewable energy sources because
Accepted 9 May 2015
of the regular and predictable nature of the resource. Therefore, in addition to promoting the develop-
Available online 14 June 2015
ment of new technologies, global policies for the generation of renewable and clean energy are being
strengthened. Several methods of energy conversion have been developed over the years: the
Keywords:
turbine-based current energy converter, in particular, has demonstrated high energy generation capacity
Coastal current
Wake
and is already in operation. In this study, the three-dimensional model TELEMAC3D was used to investi-
Conversion sites gate the hydrodynamic processes. This model was coupled with an energy conversion module to deter-
TELEMAC3D mine the best energy sites for marine current energy generation in the southern Brazilian shelf. Two
Finite element method viable regions were found in the region of study that exhibited high potential for energy generation from
marine currents; however, the northern region has been found to be a more viable region for the instal-
lation of current converters and can reach an average power of approximately 10 kW per day and inte-
grated values of 3.5 MW per year. The highest levels of power generation were found at 16-day
intervals, showing a high correlation with events associated with the passage of meteorological fronts
along the region of study. In this study, we design a turbine farm with ten helicoidal turbines. Three grids
were used in a one-year simulation of the TELEMAC3D model that was coupled with an energy conver-
sion module. The simulation results were used to identify a suitable region for trial tests on a model tur-
bine farm. For a simulation with physical structures, the northern region site was notable because high
conversion rates were maintained during events of high potential energy. This enhanced electricity gen-
eration occurred because of the intensification of the current field by a physical structure that enhanced
the efficiency of the site. No significant differences in the temporal variability pattern were estimated
between simulations with and without structures; thus, the presence of structures did not change the
temporal energy conversion pattern on the time scales considered in this study. An annual power output
of 59.39 GW h was predicted for the turbine configuration that was chosen for this study, which was
equivalent to 0.22% of the entire energetic consumption of Rio Grande do Sul State in 2010.
Ó 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.
E-mail address: ekirinus@gmail.com (E.P. Kirinus).

http://dx.doi.org/10.1016/j.apenergy.2015.05.028
0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
24 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34
1. Introduction
The continuing growth of the world population has increased
the demand and competition for energy, such that immense efforts
are required to make non-renewable energy sources available.
Therefore, in addition to promoting the development of new tech-
nologies, global policies for the generation of renewable and clean
energy are being strengthened.
The oceans are an important and infinite source of renewable
energy [1]. Such energy can be obtained from waves, tidal oscilla-
tions or tidal currents, thermal energy, osmotic gradient and ocean
currents. Two ways to obtain energy from the currents are through
potential energy (variations in the sea level) and kinetic energy
(ocean currents and their water masses). In general, the technique
is defined as marine hydrokinetic energy extraction. According to
[2], the energy from river and estuary flow, tidal currents and other
artificial water channels is considered a viable source of renewable
energy.
The Canadian Hydrology Center (CHC) conducted a survey of
the available marine energy along the Canadian coastal region
[3]. This study demonstrated that the average electrical power at
190 study points would constitute approximately 63% of
Canadian energy demand. A similar study was carried out by [4]
for the United States, where the authors identified regions with
the highest potential for hydropower.
Defne et al. [5] investigated the energy potential of waves and
tidal currents along the southeastern coast of the United States
and identified a power conversion that ranged between 1.0 and
3.0 MW/year. Additionally, the Electric Power Research Institute
(EPRI) documented 16 TWh/year in Alaska and 0.6 TWh/year in
Puget Sound [6–11].
Studies and effective energy conversion from the astronomical
and meteorological tides are in advanced stage of development.
Several countries as: United States [4,5,12,13], Canada [3], United
Kingdom [14,15], Ireland [16], Scotland [17], among others have
an inventory about the potential of their hydrological resources.
Instead of this high potential, on the structural point of view, the sys-
tems for marine energy conversion still have high production costs.
In this sense, the development of modern and efficient convert-
ing structures, with lower costs and environmental friendly, like
the triple helicoidal turbines have increased the applicability for
the marine energy conversion around the world [18]. Khan et al.
[2] analyzed the existing turbines and suggested the dominance
of the axial turbines with commercial capacity. On the other hand,
the choice of the turbine used for each local depends on economi-
cal, political, technical and environmental factors [19–23].
Technology for energy conversion through the oceanic currents
still under development with only few models connected to the
electric grid [14]. The large scale turbines installed nowadays are
the SeaGen and the SeaFlow turbines [24,25] (Marine Current
Turbines), Tidal Stream Turbine (Hammerfest Strom AS, Norway)
and the scale model from the Open Centre Turbine (Open-Hydro
Ltd., Ireland).
Some countries already have some prototypes installed. Since
2003, a prototype was successfully installed 1 km off Foreland
Point, near Devon (UK), at 30 m depth. This prototype has a simple
rotor of 15 m diameter and generates 300 kW of electric power
with currents of 2.3 m/s on bidirectional way. After 2006, the
responsible company received permission to install a turbine with
double rotor capable to produces 1 MW of electric power in
Strangford Lough, near the North Ireland coast, in order to convert
energy and confirm the commercial purpose of this technology
[26].
Another example is the Kobold, that consists in a turbine with
vertical axis under development since 1995. An important
characteristic of this turbine is that the rotation is independent
from the direction of the currents. After 2002, one prototype was
installed at the Messina strait, 200 m off the Italian coast. This tur-
bine still in operation connected to the electric grid and prelimi-
nary tests indicated that the turbine produces 25 kW with
current velocities of 1.8 m/s [27].
In Brazil, approximately 80% of the population lives within
200 km of the coastline [28]. However, there is no mapping of
the coastal zones regarding the energetic potential viable for con-
version using hydrokinetic turbines. A recent study have showed
two spots of high power availability off the shores of the Rio
Grande do Sul state, that can generate 3.5 MW of power during a
year [29]. Marques et al. [23] studied the effect of hydrodynamic
and morphodynamic processes on the installation of six hydroki-
netic turbines that attained 5 GW of power annually.
Tavora et al. [30], applied Geographic Information System (GIS)
analysis in order to evaluate the social, economic and environmen-
tal strains in the Rio Grande do Sul state regarding the usage of
marine current turbines. These authors verified that in the north-
ern SBS there is a spot of high power availability which is also
the less impacted region for this purpose.
The Southern Brazilian Shelf (SBS) is located between 28°S and
35°S (Fig. 1), is continentally bounded by Rio Grande do Sul state
and has a slightly rugged shoreline that is oriented Northeast –
Southwest. The bathymetry of this region is quite soft, with a high
slope and a shelf break near the 180 m isobath [31]. The region of
study is located near the Brazil–Malvinas confluence, which is
known for its high spatial and temporal variability [32,33] and
the convergence of several water bodies [34,35].
Thus, the Southwest Atlantic Ocean is one of the most dynamic
regions of the global ocean [36,37] and is characterized by large
thermohaline contrasts and intense mesoscale activity [38]. The
high seasonality of the wind fields [39,40] is dominated by north-
east (NE) winds during the summer and southwest (SW) winds
during the winter that drive the coastal circulation through the
SW and NE, respectively [23,35,41,42]. These winds can be
enhanced by El Niño Southern Oscillation (ENSO) events [39].
The recent annual energy report of the Rio Grande do Sul state
[43] briefly mentioned the use of marine currents as a potential
energy source for power harvesting that could easily enhance the
Brazilian matrix of energy. In this context, the aim of this study is
to study the potential of using energy converters (turbine-type)
along the SBS by applying a three-dimensional model of ocean circu-
lation coupled with an energy model to evaluate the energy conver-
sion and local circulation pattern of a turbine farm.
2. Methodology
2.1. Hydrodynamic numerical model
The hydrodynamic processes patterns along the southern
Brazilian shelf were studied using the three-dimensional
finite-element open-source TELEMAC3D model.1 In the
TELEMAC3D model, the Navier–Stokes equations are solved by con-
sidering local variations in the free surface of the fluid, while
neglecting density variations in the mass conservation equation,
using the hydrostatic pressure and Boussinesq approximations.
In the model, finite element techniques are used to solve the
hydrodynamic equations [44], and the sigma coordinate system
is used to follow the surface and lower boundaries for vertical dis-
cretization [45]. The Multidimensional Upwind Residual
Distribution (MURD) [44] method is applied for the advection of
three-dimensional variables under TELEMAC3D. On the other
1
http://www.opentelemac.org.
 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34
Fig. 1. Region of study: southern Brazilian shelf with a maximum depth of approximately 3.500 m.
hand, the boundary conditions in TELEMAC3D are applied accord-
ing with the advective and diffusive methods used.
The advection of velocities and the diffusion of velocities and
tracers are solved using characteristic methods. According to
[44], TELEMAC3D uses this method and prescribes the boundary
conditions using a radiative condition. The advection of tracers
are solved using the distributive method MURD + PSI and, accord-
ing to [44], this method guarantee conservation and monotonicity.
In this way, the prescription of tracers at the boundaries is applied
using imposition by the Dirichlet boundary condition. A detailed
description regarding the mathematical and numerical methods
used by TELEMAC3D model can be found in [44].
A time step of 90 s and a Coriolis coefficient of
7:70  105 rad s1 (at 32°S) were used in all of the simulations.
The horizontal turbulence process was performed using the
Smagorinsky model. This turbulent model closure is generally used
for maritime domains with large-scale eddy phenomena, and the
mixing coefficient is calculated by considering the size of the mesh
elements and the velocity field [46].
A mixing length model for buoyant jets was implemented to
evaluate the vertical turbulence processes and to better represent
the stratification. The vertical diffusion coefficients were calculated
by incorporating density effects into the model via a damping fac-
tor that depended on the Richardson number.
2.2. Energy module
Converters can be used to harness the power of ocean currents.
Similar to the technology of wind converters, a submerged rotor is
forced to rotate by the fluid surrounding it. Khan et al. [2] recently
evaluated the equipment available to capture hydrokinetic energy
and found that 76 pieces of equipment, including turbines, are
either in operation or in the early stages of research.
The hydrodynamic simulations used in this study were per-
formed using the TELEMAC3D model. Energy conversion from the
currents into electrical power was calculated using the energy
module [23]. This module uses the turbine standard equation
(Eq. (1)) to calculate the electric power in watts (W) from the inci-
dent flow velocity.
25
The principle of energy conservation is used during each time
step of the hydrodynamic model (Fig. 2) to calculate the current
velocity, which is then transferred to the energy conversion mod-
ule. The module converts a portion of the energy from the currents
into power through the electric power equation. In the energy con-
version module, the current velocity is updated to maintain the
energy balance of the TELEMAC3D model.
Möller et al. [35] and [23,42,47], among others, showed that the
SBS region exhibits a multidirectional and highly dynamic circula-
tory pattern that is strongly influenced by the passage of meteoro-
logical fronts. Thus, in this study, we used the Gorlov converter
[48], which can capture the energy of multidirectional currents.
In this way, this turbine is more appropriate for environments
with high temporal and spatial variability on circulation. In addi-
tion, on the numerical point of view, this turbine is more easily
implemented because of its applicability considering multidirec-
tional currents.
The helicoidal turbine of the Gorlov has a rectangular sectional
area (h  D) and its power coefficient (g) is equal to 0.35 [48]. The
power gained from a helicoidal converter is determined by Eq. (1),
where: P represents the electric power in watts (W), q is the water
density and v is the current velocity. Table (1) shows the turbine
technical parameters that were used in the energy conversion
module.
1
PðWÞ ¼ gqðh  DÞv 3 ð1Þ
2
2.3. Simulated scenarios
The energy conversion potential and the effect of the installa-
tion of energy converters on the natural hydrodynamic processes
of the SBS were investigated by performing three simulations over
365 days: the physical parameters applied are given in the next
section. The simulated period ranged from January 01st to
December 31st of a climatological year.
One simulation was conducted using only the hydrodynamic
processes. After selecting the area of interest, a mesh was created
for ten turbines (Fig. 3a). The farm grid direction (Fig. 3b) was
26 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34
Fig. 2. Fluxogram of interactions between TELEMAC3D and energy module
(adapted from [23]).
idealized as being parallel to the coast, with 200 m between each
turbine in the x- and y-directions. Given the computational limita-
tions, the optimal shape for a turbine had 4 nodes (Fig. 3c), with
10 m between each node.
The orientation of the farm grid was defined parallel to the
coast in accordance with the residual pattern of the currents over
the adjacent continental shelf. The distance of 200 m between each
turbine on x and y directions were used in order to minimize the
wake effect, which can impact negatively the flow velocity [49–
57].
This distance varies with the incident current intensity, accord-
ing to [56,50], the lateral wake can impact the flow far from 3 tur-
bine diameters, while the frontal wake can reach 6 to 10 diameters
of distance [57]. The distance between each turbine was chosen
accordingly to these previous references.
The conversion model interacted with the turbine and acquired
the velocity of the node (which is shown as a red bullet at Fig. 3c);
this velocity was converted into power and the lost kinetic energy
was released to the turbine node (which is represented by yellow
bullets).
This scenario was improved by implementing an energy sink for
conversion in two simulations using the farm grid (with the afore-
mentioned interactions). One simulation was performed without
the conversion structures, and another simulation was performed
using the three-dimensional physical structure of the turbines
(Fig. 3d), where the depth of the turbine nodes was changed in
the FORTRAN source code.
2.4. Initial and boundary conditions
In this study, climatology data were used to impose the initial
and boundary conditions to investigate the power generation char-
acteristics and the processes occurring on a turbine farm. This data
was created using the long time series data base from the Brazilian
Table 1
Turbine technical parameters.
Start-in speed 0.2 m/s
Cut-in speed 1.5 m/s
Efficiency coefficient 0.35
Nominal power 170 kW
Turbine height 14 m
Turbine ray 10 m
National Water Agency (ANA), the Ocean Circulation and Climate
Advanced Modeling Project (OCCAM) and Reanalysis (National
Oceanic Atmospheric Administration – NOAA).
The climatological changes were incorporated using a monthly
mean for a temporal data discharge series from January of 1940 to
December 2006. The OCCAM data from 1990 to 2004 were pro-
cessed to obtain the velocity components, temperature, salinity
and sea surface height. The wind and air temperature fields were
collected from Reanalysis from 1948 to 2012 [58].
The water body boundaries used in this study were the Guaíba
River, the Camaquã River and the São Gonçalo channel (Fig. 4)
based on the discharge time series obtained from the Brazilian
National Water Agency2 (ANA) website. The discharge of the São
Gonçalo channel was assumed to be constant at 760 m3/s due to
the absence of the observed data series.
The oceanic boundary was forced by the astronomical tides,
water levels, current velocity, salinity and temperature fields
(Fig. 4). The temporal and spatial variability in the winds and air
temperature were prescribed along the surface boundary. The air
temperature data along the ocean surface were also used to incor-
porate heat exchange with the atmosphere into the model
calculations.
The numerical model was initialized from a rest state with an
initial elevation of 0.50 m, which is the approximate average tide
in the region [59]. The amplitude and phase data were also pre-
scribed along the oceanic border using the Grenoble Model
FES95.2 (Finite Element Solution v. 95.6).
2.5. Calibration and validation
Monteiro et al. [60], and [61–63] calibrated and validated a
two-dimensional model for the Patos Lagoon estuary.
Subsequently, Marques et al. [23,41,42,47] performed a set of sim-
ulations to calibrate and validate a three-dimensional numerical
model for an area consisting of the Patos Lagoon and the adjacent
coastal region.
The results of these calibration and validation tests indicated
that the TELEMAC3D model could be used to investigate the SBS
with an acceptable degree of accuracy. The aforementioned studies
yielded values for many physical parameters (such as the wind
influence coefficient, the friction coefficient and the turbulence
models) that were used in the present study (Table 2).
3. Results and discussion
The hydrodynamic conditions for this region are characterized
by the convergence of different water masses. The average velocity
of the current (Fig. 5a) was calculated for this region, and a maxi-
mum mean value of 0.4 m/s was found for the two highlighted
regions.
This mean value exhibited some variability, where the highest
standard deviations around the mean velocities were found for
the regions with the highest mean velocities (Fig. 5b). This result
suggests that although these regions are appropriate sites for
energy converters, periods of low power generation may be
encountered because the velocity deviations were closer to the
average values in these regions.
In a previous study, [29] concluded that the southern region is
less viable for installing marine turbine currents in the SBS than
the northern region, which has emerged as a significant potential
power producer that can attain mean values of 10 kW over a day.
[30] achieved the same conclusion as [29] using GIS analysis.
2
http://www.ana.gov.br.
 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34 27

Fig. 3. Turbine region for area of study: (A) numerical grid with high degree of refine for the region of interest; (B) converter farm with 10 helicoidal turbines represented
using 2 arrays parallel to the coast; (C) schematic showing interactions between energy conversion module and turbines; and (D) conversion structures represented using a
three dimensional shape similar to a monopile; site depth is approximately 18 m.

The rotational pattern was studied to understand the tendency

for maintaining turbulence in each scenario. The average behavior
of the power generation for both sites was analyzed considering
the residual velocity field that was associated with the mean field
power converted (which is shown using power isolines in Fig. 6).
The average power converted was higher than 1.4 kW (Fig. 6a
and b) for some turbines. The simulation without structures
(Fig. 6a) showed a higher mean power for turbines 1, 2, 3 and 6
than the other turbines (considering the North–West turbine to
be the first turbine and the South–East to be the tenth turbine,
see Fig. 7). The simulation with structures (Fig. 6b) resulted in
enhanced power generation for turbines 7, 8 and 10 along with
the aforementioned turbines.
The same South–West circulation pattern was observed for
both simulations, which could be explained by the dominance of
the North quadrant winds during the period considered. In addi-
tion, there was a slight intensification of the residual velocity in
the simulation with structures (Fig. 6b), where the residual veloc-
ity vectors were enhanced between and around the turbine arrays.
We observed a shadow zone in the circulation pattern behind
the turbines, which could be related to the wake generated by
Fig. 4. Finite element mesh: liquid and surface boundary conditions for
TELEMAC3D model are highlighted. the adjacent turbines. This effect means that the flow incident to
a boundary (in this case a turbine) should generate small eddies
Therefore, only the northern region was investigated in the present that move downstream more slowly than the upstream incident
study. velocity and return to balance the momentum in time [64,17].
In the simulation without structures (Fig. 6a), this wake effect
3.1. Current pattern and energy conversion was purely hydrodynamic, and the variations in the velocity fields
resulted from the conversion of energy, which caused perturba-
The most efficient scenario for a turbine farm (with or without tions to the local vorticity pattern. However, in the simulation with
the presence of the structures) was determined by performing a structures (Fig. 6b), this process was caused by (1) modifications to
spatial variability analysis considering the temporal variations for the hydrodynamic processes from the energy sink and (2) the
the simulation. In this analysis, the capacity of each turbine was effect of the shape of the turbine body. These factors resulted in
used to convert the current energy into power according to the a reduction of the velocity and were also observed by [50] for
hydrodynamic pattern. experimental trials in a flume tank.
The superficial vorticity pattern for each scenario is shown in
Fig. 8. The vectors represent the surface residual velocity, and the
Table 2 colors represent the extents of the turbine wake: the same residual
Physical parameters and numerical treatment used on simulation.
pattern can be observed in both parts of the figure, and the primary
Time step 90 s difference between the two parts was caused by the turbine wake.
Coriolis Coefficient 7:70  105 N m1 s1 For the scenario without structures (Fig. 8a), more intense curls
Wind influence coefficient 1:0  106 (both cyclonic and anti-cyclonic) were observed than for the sce-
Horizontal turbulence Smagorinsky model nario with vertical structures (Fig. 8b).
Vertical turbulence Mixing length (Jet) model
This result agreed with those from the previous residual analy-
Bottom friction Manning law
Scheme for diffusion Implicit sis, showing that using the conversion module alone (Fig. 8a) pro-
Scheme for advection N-scheme for tidal flats moted the generation of stronger curls, which increased the
Solvers for diffusion Conjugate gradient superficial turbulence. This rotation effect decreased the intensity
Preconditioning of matrix Diagonal
of the superficial flow, i.e., the turbine behind the wake path was
28 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34

Fig. 5. (A) Average current velocity (m/s) and (B) its standard deviation over the entire simulation period; the southern region is enclosed in the red-dashed area and the
northern region is enclosed in the black-dashed area. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 6. Residual velocities and mean power (kW) generated for turbines, shown as isolines for (A) site without structures and (B) site with structures.

affected negatively by this strong curl, which decreased the turbine
efficiency.
Although the same pattern was observed for the scenario with
structures (Fig. 8b), the curl was smoothed, indicating that the
fluxes surrounding the turbines were more homogeneously dis-
tributed, resulting in less rotation than for the case without struc-
tures. The superficial current remained well orientated, which
reduced the velocity reduction from eddy and curl effects. In this
case, the entire turbine farm received a slightly stronger current,
resulting in higher conversion rates compared to the case without
structures.
The bottom layer curl (Fig. 9) exhibited the same pattern as was
observed for the surface layer: in the absence of structures
(Fig. 9a), stronger curls were created that could deplete the veloc-
ity and dissipate energy, whereas the presence of the vertical struc-
tures (Fig. 9b) generated weaker curls, thus retaining the current
energy for energy conversion. Given these results, only the sce-
nario with vertical structures is analyzed in the rest of the paper.
3.2. Temporal variability analysis
The temporality of the energy conversion for the entire turbine
farm was analyzed using the wavelet method described by [65,66]
with a time series (Fig. 10a) for the electric power of ten convert-
ers. In the wavelet method, a global wavelet power spectrum was
 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34 29

Fig. 7. (A) Location of vertical cross section, which begins at 0 m (above turbines 1 and 6) and ends at 1.800 m (below turbines 5 and 10); cross sections A and C are on left and
right sides, respectively, of turbine farm; cross section B is in the middle of the farm; turbines are numbered as shown.

Fig. 8. Superficial residual velocities and rotational pattern for (A) site without structures and (B) site with structures; arrow shows scale of 0.1 m/s.

used to demonstrate the occurrence of energy conversion events
on over time.
Analyzing the local power spectrum (Fig. 10b) showed that pos-
itive correlations (i.e., the red contours) were enhanced for veloci-
ties above 0.5 m/s, which increased the energy conversion. The
figure also shows that the physical processes that maintained the
operation of the turbine farm were controlled by two primary
groups of temporal scales. The first group of approximately 6 days
dominated the period because of forcing from cyclic changes in the
direction of the wind pattern.
The second group of temporal scales consisted of correlations
over periods above 16 days, suggesting that the physical processes
shorter than 20 days were the primary mechanisms that controlled
the electric energy conversion along the inner continental shelf.
The global spectrum of energy (Fig. 10c) confirmed this hypothesis,
indicating the occurrence of processes with time scales longer than
16 days (which may have been extended for some transients) at a
95% confidence level throughout the study period.
This pattern was similar to that obtained by [23] in terms of the
occurrence of the processes and the cycles of occurrence. Thus, the
pattern was strongly influenced by the passage of frontal meteoro-
logical systems that generated further changes in the wind direc-
tion and the intensity of currents.
3.3. Spatial variability analysis
The significance of each turbine in the farm structure was eval-
uated by correlating the incident current velocity with the power
for the turbine arrays. The efficiency of each turbine was studied
on different time scales using two-dimensional cross-spectral
wavelet analysis, where the dominant processes at the turbine
sites were considered to occur on temporal scales longer than
1 day and shorter than 30 days.
The significance of each turbine and its variance in time is
shown by the strongest correlation (i.e., the red contours) for the
Morlet wavelet. The values to the right of the tendency line show
the turbines with the optimal locations at a 95% confidence level.
For the simulation with structures, a cross-spectral analysis
(Fig. 11a) showed an intensified power conversion for several tur-
bines during the primary conversion events. The temporal series
30 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34

Fig. 9. Bottom residual velocities and rotational pattern for (A) site without structures and (B) site with structures; arrow shows scale of 0.05 m/s.

Fig. 10. Integrated current velocity and electric power time series used for cross-wavelet analysis and (B) local and (C) global wavelet power spectrum of the time series using
Morlet wavelet; thick contour lines enclose regions above 95% confidence for a red-noise process with a lag-1 coefficient of 0.95; cross-hatched regions indicate cone of
influence where edge effects become important. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

for the mean variance (Fig. 11b) indicated, at a 95% confidence
level, that four events of high power conversion sustained the high
power levels and were also intensified in this scenario.
Variance values above 35 kW (Fig. 11b) were observed during
the extreme event in October, which enhanced the highest power
capacity for this scenario. The mean variance of each turbine
(Fig. 11c) indicated that a higher correlation between the variables
resulted in an average power of approximately 7 kW. Moreover,
this result shows that turbines 1, 2, 3, 6 and 7 were the most effi-
cient turbines on the farm, considering the energy conversion only.
The different turbine efficiencies were attributed to the differ-
ent turbine locations on the farm and the effect of the incident cur-
rent on the wake patterns.
3.4. High energy conversion events
The velocity time series (Fig. 10a) and the variance time series
(Fig. 11b) showed four events of high energy conversion during
the entire year: two primary events were chosen to analyze the
pattern of occurrence of these events and the corresponding effect
on the vertical fluxes surrounding the turbine farm.
The May event and the event at the end of September were cho-
sen, which will be referred to as events I and II, respectively, in this
section. For both events, time series for the wind intensity and
direction (Fig. 12a and b) were taken to identify the mechanism
controlling each event. The vertical cross sections were also ana-
lyzed for these periods.
Event I was characterized by south wind intensities reaching up
to 10 m/s (Fig. 12a), whereas event II was characterized by strong
north winds with intensities up to 15 m/s (Fig. 12b). Regardless of
the difference in the wind direction, these high-intensity winds
were the primary mechanism of these high energy conversion
events. This result was obtained because of the dominance of
southwest and northeast winds in the region, and because
high-intensity transients were strongly correlated with the pas-
sage of frontal meteorological systems. Thus, event I produced a
60 kW conversion for the first turbine, and event II produced a con-
version of almost 120 kW (see Fig. 12c and d, respectively).
 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34 31

Fig. 11. (A) Cross-spectral analysis between power (kW) and current velocity (m/s) for events on time scales longer than 1 day and shorter than 30 days; thick contour lines
enclose turbines for a red-noise process with a lag-1 coefficient of 0.27 above a 95% confidence level; T1-T10 denote the turbine numbers, as defined in Fig. 12; (B) mean
variance for the period of study, where values beneath the tendency line correspond to a 95% confidence level; and (C) temporal series of spatial mean variance of each
turbine. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 12. Wind intensity during high energy conversion events: (A) event I with southern winds and (B) event II with north winds; residual currents and conversion rates (kW)
during (C) event I and (D) event II.
32 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34

Fig. 13. Vertical cross sections during event I; colored surface shows transversal
Fig. 14. Vertical cross sections during event II; colored surface shows transversal
velocity for each time step, and vectors show direction of resultant between vertical
velocity for each time step and vectors represent direction of resultant between
and transversal velocities. (For interpretation of the references to color in this figure
vertical and transversal velocities. (For interpretation of the references to color in
legend, the reader is referred to the web version of this article.)
this figure legend, the reader is referred to the web version of this article.)
 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34
The residual currents of these events showed that the currents
were strongly influenced by the wind direction. Marques et al. [47]
showed that the wind setup was the primary mechanism that con-
trolled the mesoscale dynamics of the SBS. As the wind changed
direction and increased in intensity (Fig. 12a and b), coastal cur-
rents were driven in the same direction as the wind, promoting
high energy conversion for the turbines.
Three cross sections (A, B and C) were taken in the
northeast-southwest direction with the starting point at the 0-m
mark (Fig. 7) and the ending point at 1.800 meters. The colored
surface shows the transversal velocity for each time step, and the
vectors indicate the direction of the resultant of the vertical and
transversal velocities.
The fluid motion followed the wind setup for both events. The
cross section analysis for event I (Fig. 13) showed that the power
of the southern turbines was enhanced because of their location
relative to the incident current (i.e., the northeast wind-driven cur-
rent, which is shown in Fig. 12c).
The cross section C (Fig. 13c) shows the bottom current encoun-
tering the conversion structures (from 400 to 1200 m). Before this
interaction, the currents intensified because of the wind setup and
then, following energy conversion, suffered a velocity loss of 35%,
which was the same turbine efficiency that was designed for the
study (see Table 1).
For event II (Fig. 14), we observed that the wind setup con-
trolled the current, directing the entire water column to the south-
west (Fig. 12d). A slight reduction in the velocity was only
observed for cross section A at the middle of the bottom layer at
the south of the farm (from 1200 to 1800 m) (Fig. 14a). This effect
corresponded to a reflex of the wake pattern that was generated by
the turbine farm at the end of the wake, resulting in decreased
velocities and energy dissipation. This effect was smoothed for
cross sections B and C.
The wind intensity was the primary mechanism for both events.
In addition, the farm configuration ensured that one turbine was
always highly affected by the farm wake. The primary direction
of the current prevented the final turbine from reaching the max-
imum efficiency, i.e., the 1st turbine for event I and the 5th turbine
for event II.
4. Conclusions
The southern Brazilian shelf is highly dynamic and constantly
influenced by cycling winds (North–East/South–West) and
wind-driven coastal currents. These conditions make the usage of
marine current turbines for energy generation viable, for which
the use of helicoidal turbines is recommended. The results of this
study led to the following conclusions.
A comparison of the numerical simulations with and without
physical conversion structures demonstrated that these structures
did not affect the temporal pattern of the energy conversion. The
structures had a positive effect in intensifying the velocity field
surrounding the turbine farm and increasing the energy
conversion.
However, the energy converters appropriated a portion of the
kinetic energy from the coastal currents, thereby generating diver-
gence and convergence zones in accordance with the dominant
direction of the currents. The changes in the hydrostatic balance
generated instabilities in the fluid motion, partly because of the
conversion energy and partly because of the presence of the struc-
tures that generated the wake behind the turbines. This effect
decreased the energy conversion of the next turbine, while also
intensified the surrounding velocity fields.
Turbulent processes decreased the intensity of the superficial
flow, i.e., the turbine behind the wake path was negatively affected
33
by the strong curl, which decreased the turbine efficiency.
Structures smoothed this effect, such that the fluxes surrounding
the turbines were more homogeneously distributed, which
decreased the rotation and the energy dissipation of the curl.
The wind setup with intensities up to 10 m/s was the control-
ling mechanism for high energy conversion events. The energy
conversion associated with the configuration of the turbine farm
decreased the velocities at the north of the farm for event I, and
the change in the velocities in event II. During both events, slight
differences in the vertical flow structure were observed, demon-
strating that during a high wind intensity event, regardless of the
wind direction, the entire water column remained almost steady,
thereby increasing the energy conversion for all of the turbines
except those that were negatively affected by the turbine wake
effect.
The turbine farm showed a large capacity for converting current
energy for four primary energetic events. This turbine farm
reached an annual energy capacity of 59.39 GW h (16.5 MW) using
ten turbines, which was equivalent to 0.22% of the entire energetic
consumption of Rio Grande do Sul State in 2010.
In future studies, we shall improve the farm configuration and
perform a trial test to implement a diffuser structure near the tur-
bines to intensify the incident current fields.
Acknowledgements
The authors are grateful to the Agência Nacional do Petróleo –
ANP and Petrobras for fellowships from the Programa de
Recursos Humanos (PRH-27) that provided bursaries, the
Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul
(FAPERGS) for sponsoring this research under contract 1799123
and the Conselho Nacional de Desenvolvimento Científico e
Tecnológico (CNPq) for contracts 456292/2013-6 and
305885/2013-8.
We would also like to acknowledge the Brazilian Navy for pro-
viding detailed bathymetric data for the coastal area and the
Brazilian National Water Agency and NOAA for supplying the river
discharge and wind data sets, respectively. Although data were
obtained from governmental databases, the views presented in this
study do not represent the views of the government.
References
[1] Cruz JMBP, Sarmento AJNA. Sea state characterization of the test site of an
offshore wave energy plant. Ocean Eng 2007(24):763–75.
[2] Khan MJ, Bhuyan G, Iqbal MT, Quaicoe JE. Hydrokinetic energy conversion
systems and assessment of horizontal and vertical axis turbines for river and
tidal applications: a technology status review. Appl Energy
2009;86(10):1823–35. http://dx.doi.org/10.1016/j.apenergy.2009.02.017.
[3] Cornett A. Inventory of Canadas marine renewable energy resources. Tech.
rep., Ottawa: Canadian Hydraulics Centre; 2006.
[4] Hall DG, Reeves KS, Brizzee J, Lee RD, Carroll GR, Sommers GL. Feasibility
assessment of the water energy resources of the United States for new low
power and small hydro classes of hydroelectric plants. Tech. rep., U.S.
Department of Energy, Office of Energy Efficiency and Renewable Energy,
Wind and Hydropower Technologies Program, Idaho Operations Office; 2006.
http://www1.eere.energy.gov/water/pdfs/doewater-11263.pdf.
[5] Defne Z. Multi-criteria assessment of wave and tidal power along the Atlantic
coast of the southeastern USA. Ph.D. thesis, Georgia Institute of Technology;
2010.
[6] EPRI. Tidal power in North America environmental and permitting issues. Tech.
rep., Devine Tarbell & Associates Inc., Electric Power Research Institutes. EPRI-
TP-007-NA; 2006.
[7] EPRI. Tidal InStream energy resource assessment for Southeast Alaska. EPRI-
TP-003-AK; 2006.
[8] EPRI. Maine Tidal InStream Energy Conversion (TISEC): survey and
characterization of potential project sites. EPRI-TP-003-ME; 2006.
[9] EPRI. Massachusetts Tidal In-Stream Energy Conversion (TISEC): survey and
characterization of potential project sites. EPRI-TP-003-MA; 2006.
[10] EPRI. New Brunswick Tidal In-Stream Energy Conversion (TISEC): survey and
characterization of potential project sites. EPRI-TP-003-NB; 2006.
[11] EPRI. Nova Scotia Tidal In-Stream Energy Conversion (TISEC): survey and
characterization of potential project sites. EPRI-TP-003-NS; 2006.
34 E.P. Kirinus, W.C. Marques / Applied Energy 155 (2015) 23–34
[12] Defne Z, Haas KA, Fritz HM. GIS based multi-criteria assessment of tidal stream
power potential: a case study. Renew Sustain Energy Rev
2011;15(5):2310–21. http://dx.doi.org/10.1016/j.rser.2011.02.005.
[13] Defne Z, Haas KA, Fritz HM, Jiang L, French SP, Shi X, et al. National
geodatabase of tidal stream power resource in USA. Renew Sustain Energy Rev
2012;16(5):3326–38. http://dx.doi.org/10.1016/j.rser.2012.02.061.
[14] Gross R. Technologies and innovation for system change in the UK: status,
prospects and system requirements of some leading renewable energy
options. Energy Policy 2004;17:1905–19.
[15] James V. Marine renewable energy: a global review of the extent of marine
renewable energy developments, the Developing Technologies and Possible
Conservation Implications for Cetaceans. Wiltshire; 2013.
[16] Rourke FO, Boyle F, Reynolds A. Marine current energy devices: current status
and possible future applications in Ireland. Renew Sustain Energy Rev
2010;14:1026–36. http://dx.doi.org/10.1016/j.rser.2009.11.012.
[17] Hardisty J. The analysis of tidal stream power. The University of Hull,
Kingston-upon-Hull, UK: John Wiley & Sons, Ltd; 2009.
[18] Gorlov A. Tidal energy. Academic Press; 2001. p. 2955–60. http://dx.doi.org/
10.1006/rwos.2001.0032.
[19] Khan MJ, Iqbal MT, Quaicoe JE. River current energy conversion systems:
progress, prospects and challenges. Renew Sustain Energy Rev
2008;12:2177–93.
[20] Grabbe M, Lalander E, Lundin S, Leijon M. A review of the tidal current energy
resource in Norway. Renew Sustain Energy Rev 2009;13:1898–909.
[21] Guney MS, Kaygusuz K. Hydrokinetic energy conversion systems: a technology
status review. Renew Sustain Energy Rev 2010;14:2996–3004. http://
dx.doi.org/10.1016/j.rser.2010.06.016.
[22] Vennell R. Estimating the power potential of tidal currents and the impact of
power extraction on flow speeds. Renew Energy 2011;36(12):3558–65. http://
dx.doi.org/10.1016/j.renene.2011.05.011.
[23] Marques WC, Fernandes EHL, Malcherek A, Rocha LAO. Energy converting
structures in the southern Brazilian shelf: energy conversion and its influence
on the hydrodynamic and morphodynamic processes. J Earth Sci Geotech Eng
2011;1(1):61–85.
[24] Douglas CA, Harrison GP, Chick JP. Life cycle assessment of the Seagen marine
current turbine. Proc Inst Mech Eng Part M: J of Eng Marit Environ
2008;222:1–12. http://dx.doi.org/10.1243/14750902JEME94.
[25] Fraenkel PL. Marine current turbines: pioneering the development of marine
kinetic energy converters. Proc IMechE Part A: J Power Energy 2007:159–69.
[26] M.C.T. MCT. SeaGen. <http://www.seageneration.co.uk/>; 2006.
[27] Matacena E. Kobold. <www.pontediarchimede.com>; 2002.
[28] IBGE. Censo 2010; 2010.
[29] Kirinus EP, Marques WC, Stringari CE. Viabilidade de conversão da energia de
correntes marinhas na Plataforma Continental Sul do Brasil. Vetor
2012;22:83–103.
[30] Pereira JTB, Kirinus EdP, Marques WC, Gandra TBR, de Barros GP, Matzenauer
HB. Exploring the project potential of marine current turbines: a case study in
the southern Brazilian shelf region. Int J Geosci 2014;5(13):1547–60. http://
dx.doi.org/10.4236/ijg.2014.513126.
[31] Zembruscki S. Geomorfologia da margem continental sul brasileira e das
bacias oceânicas adjacentes. Tech. rep., Projeto REMAC, Rio de Janeiro; 1979.
[32] Podestá GP. Utilización de datos satelitarios en investigaciones oceanográficas
y pesqueras en el Océano Atlántico Sudoccidental. In: Boschi EE (Ed.), El mar
argentino y sus recursos pesqueros. Mar del Plata, Argentina; 1997. p. 195–
222.
[33] Souza RB, Robinson IS. Lagrangian and satellite observations of the Brazilian
Coastal Current. Cont Shelf Res 2004;24:241–62.
[34] Castro BM, Lorenzzetti JA, Silveira ICA, Miranda LB. Estrutura termohalina e
circulação na região entre o Cabo de São Tomé (RJ) e o Chuí (RS). In: Rossi CL,
Wongtschowsky DB (Eds.), O ambiente oceanográfico da plataforma
continental e do talude na região sudeste-sul do Brasil. Edusp, São Paulo;
2006, Ch. 1. p. 11–20.
[35] Möller OO, Piola AR, Freitas AC, Campos EJ. The effects of river discharge and
seasonal winds on the shelf off southeastern South America. Cont Shelf Res
2008;28(13):1607–24. http://dx.doi.org/10.1016/j.csr.2008.03.012.
[36] Chelton DB, Schlax MG, Witter DL, Richmann JG. GEOSAT altimeter
observations of the surface circulation of the Southern Ocean. J Geophys Res
1990;95:877–903.
[37] Piola AR, Matano RP. The South Atlantic Western Boundary Currents Brazil/
Falkland (Malvinas) Currents; 2001.
[38] Gordon AL. Brazil–Malvinas confluence – 1984. Deep-Sea Res
1989;36:359–84.
[39] Piola AR, Matano RP, Palma ED, Möller OO, Campos EJD. The influence of the
Plata River discharge on the western South Atlantic shelf. Geophys Res Lett
2005;32:L01603. http://dx.doi.org/10.1029/2004GL021638.
[40] Braga MF, Krusche N. Padrão de ventos em Rio Grande, RS, no período de 1992
a 1995. Atlântica 2000;22:27–40.
[41] Marques WC, Fernandes EHL, Monteiro IO, Möller OO. Numerical modeling of
the Patos Lagoon coastal plume, Brazil. Cont Shelf Res 2009;29:556–71.
[42] Marques WC, Fernandes EHL, Moller OO. Straining and advection
contributions to the mixing process of the Patos Lagoon coastal plume,
Brazil. J Geophys Res 2010;115(C6):C06019. http://dx.doi.org/10.1029/
2009JC005653. <http://www.agu.org/pubs/crossref/2010/2009JC005653.
shtml>.
[43] Capeletto GJ, De Moura GHZ, Balanço Energético do Rio Grande do Sul 2010:
ano base 2009; 2010.
[44] Hervouet JM. Free surface flows: modelling with the finite element methods;
2007
[45] Hervouet JM, Van Haren L. Recent advances in numerical methods for fluid
flows. In: Floodplain processes. Hoboken, Chichester: John Wiley & Sons; 1996.
p. 183–214.
[46] Smagorinsky J. General circulation experiments with the primitive equation, I.
The basic experiment. Weather Rev 1963;91:99–164.
[47] Marques WC, Fernandes EHL, Moraes BC, Möller OO, Malcherek A. Dynamics of
the Patos Lagoon coastal plume and its contribution to the deposition pattern
of the southern Brazilian inner shelf. J Geophys Res 2010;115(C10):C10045.
http://dx.doi.org/10.1029/2010JC006190.
[48] Gorlov BA. Helical Turb Fish Saf 2010:1–14.
[49] Ji B, Luo X, Wua Y, Peng X, Xu H. Partially-averaged NavierStokes method with
modified ke model for cavitating flow around a marine propeller in a non-
uniform wake. Int J Heat Mass Transfer 2012;55:6582–8. http://dx.doi.org/
10.1016/j.ijheatmasstransfer.2012.06.065.
[50] Myers LE, Bahaj AS. An experimental investigation simulating flow effects in
first generation marine current energy converter arrays. Renew Energy
2012;37(1):28–36. http://dx.doi.org/10.1016/j.renene.2011.03.043.
[51] Akwensivie F. In the wake of a marine current turbine. Master thesis,
University of Strathclyde; 2010. p. 111.
[52] Larsen GC. A simple wake calculation procedure. Denmark: Riso National
Laboratory; 1988.
[53] Li Y, Çalisal SM. Numerical analysis of the characteristics of vertical axis tidal
current turbines. Renew Energy 2009;35(2):435–42. http://dx.doi.org/
10.1016/j.renene.2009.05.024.
[54] Maganga F, Germain G, King J, Pinon G, Rivoalen E. Experimental
characterisation of flow effects on marine current turbine behaviour and on
its wake properties. Renew Power Gener 2010;4(6):498–509. http://
dx.doi.org/10.1049/iet-rpg.2009.0205.
[55] Li Y, Calisal SM. Three-dimensional effects and arm effects on modeling a
vertical axis tidal current turbine. Renew Energy 2010;35(10):2325–34. http://
dx.doi.org/10.1016/j.renene.2010.03.002. <http://linkinghub.elsevier.com/
retrieve/pii/S0960148110001047>.
[56] Myers L, Bahaj AS. Simulated electrical power potential harnessed by marine
current turbine arrays in the Alderney Race. Renew Energy 2005;30:1713–31.
http://dx.doi.org/10.1016/j.renene.2005.02.008.
[57] Giles J, Myers L, Bahaj A, Shelmerdine B. The dowstream wake response of
marine current energy converters operating in shallow tidal flows. In: World
renewable energy congress. Linköping, Sweden: Marine and Ocean
Technology; 2011. p. 2270–7.
[58] Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M,
Saha S, White G, Woollen J, Zhu Y, Chelliah M, Ebisuzaki W, Higgins W,
Janowiak J, Mo KC, Ropelewski C, Wang J, Leetmaa A, Reynolds R, Jenne R,
Joseph D. The NCEP/NCAR 40-Year Reanalysis project. Tech. rep., Bulletin of the
American Meteorological Society; 1996.
[59] Möller OO, Castaing P, Salomon JC, Lazure P. The influence of local and non-
local forcing effects on the subtidal circulation of Patos Lagoon. Estuaries
2001;24:297–311.
[60] Monteiro IO, Pearsom M, Möller OO, Fernandes EHL. Hidrodinâmica do Saco da
Mangueira: mecanismos que controlam as trocas com o estuário da Lagoa dos
Patos. Atlântica 2006;27:8–101.
[61] Fernandes EHL. Modelagem do Transporte de Elementos Particulados e
Dissolvidos entre a Região Estuarina da Lagoa dos Patos e o Oceano
Atlântico. Ph.D. thesis, Universidade Federal de Rio Grande; 1997.
[62] Fernandes EHL, Mario-Tapia I, Dyer KR, Mller OO. The attenuation of tidal and
subtidal oscillations in the Patos Lagoon estuary. Ocean Dyn 2004;54(3–
4):348–59. http://dx.doi.org/10.1007/s10236-004-0090-y.
[63] Fernandes EHL a, Monteiro IO, Möller OO. On the dynamics of Mangueira
BayPatos Lagoon (Brazil). J Coastal Res 2007;10047:97–107. http://dx.doi.org/
10.2112/1551-5036-47.sp1.97. <http://www.bioone.org/doi/abs/10.2112/
1551-5036-47.sp1.97>.
[64] Kundu PK, Cohen IM. Fluid mechanics. 2nd ed. Orlando, USA: Academic Press;
2002.
[65] Torrence C, Compo GP. A practical guide to wavelet analysis. Tech. rep.,
Bulletin of the American Meteorological Society; 1997.
[66] Morettin PA. Ondas e ondaletas: da análise de Fourier à análise de ondaletas.
Edusp, São Paulo; 2004.