close

Вход

Забыли?

вход по аккаунту

?

1.4038166

код для вставкиСкачать
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
Aaron S. Alexander
py
ed
ite
d
Trapped Cylindrical Flow with Multiple
Inlets for Savonius Vertical Axis Wind
Turbines
Co
Assistant Professor, Member of ASME
ot
Department of Engineering Technology
Oklahoma State University
tN
378 Cordell South
sc
rip
Stillwater, OK 74074
Email: aaron.s.alexander@okstate.edu
Ma
nu
Arvind Santhanakrishnan
Assistant Professor
Department of Mechanical and Aerospace Engineering
Ac
ce
pt
ed
Oklahoma State University
218 Engineering North
Stillwater, OK 74074
Email: askrish@okstate.edu
http://www.appliedfluidslab.org/
ABSTRACT
Savonius VAWTs typically suffer from low efficiency due to detrimental drag production
during one half of the rotational cycle. The present study examines a stator assembly created
1
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017
ASME
withbythe
objective of trapping cylindrical flow for application in a Savonius vertical axis wind
turbine (VAWT). While stator assemblies have been studied in situ around Savonius rotors in
the past, they have never been isolated from the rotor to determine the physics of the flow field,
raising the likelihood that a moving rotor could cover up deficiencies attributable to the stator
design. The flow field created by a stator assembly, sans rotor, is studied computationally using
ed
ite
d
3D numerical simulations in the commercial computational fluid dynamics (CFD) package StarCCM+. Examination of the velocity and pressure contours at the central stator plane show
that the maximum induced velocity exceeded the free stream velocity by 65%. However, flow
py
is not sufficiently trapped in the stator assembly, with excess leakage occurring between the
stator blades due to adverse pressure gradients and momentum loss from induced vorticity. A
Co
parametric study was conducted on the effect of the number of stator blades with simulations
ot
conducted with 6, 12, and 24 blades. Reducing the blade number resulted in a reduction in the
tN
cohesiveness of the internal swirling flow structure and increased the leakage of flow through the
stator. Two unique energy loss mechanisms have been identified with both caused by adverse
sc
rip
pressure gradients induced by the stator.
Ma
nu
Keywords: vertical axis wind turbine; stator; CFD; Savonius wind turbine.
Introduction
Unlike rotors used on horizontal axis wind turbines (HAWTs), the rotors used on Savonius vertical
ed
axis wind turbines (VAWTs) will cycle through all possible angles of attack during each full rotation.
This means that a Savonius VAWT will, at times in its cycle, be moving in opposition to the wind
pt
direction which leads to the creation of detrimental drag. As the Savonius design is a drag-based VAWT,
ce
this means that the rotor will create only detrimental drag during half of the cycle. Even a Darrieus
Ac
lift-based VAWT will have substantial parts of the cycle where the only force created will be due to
detrimental drag [1–3]. In order to avoid this effect and to optimize the wind’s angle of incidence,
researchers have attempted to install various types of stators around VAWT rotors.
The simplest solution that has been examined is to block off the Savonius rotor from the wind
during the part of the cycle when it is moving upstream. Kotb and Aldoss [4] computationally studied
how the flow field was modified by the imposition of such a blockage. Shaghnessy and Probert [5]
2
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017 by ASME
conducted
experimental tests with a V-shaped blockage and found it was possible to increase power
output by 20%. Using optimization procedures on the wall blockage and blade shape, Mohamed et
al. [6] further increased the power output by 38.9% above the standard Savonius design. Additionally,
Zhang et al. [7] have used an optimization algorithm to determine the proper placement of downstream
Savonius turbines, so that re-use of the upstream turbine wake results in increased performance in the
ed
ite
d
downstream turbines. Craig et al. [8] conducted a similar analysis for Darrieus VAWTS demonstrating
that an arrays of turbines can be oriented such that the incoming streamwise velocity can be increased
by 56%.
py
As an extension to the simple blockage, several studies have examined actively ducting the flow
from away from the upstream moving side onto the power producing side. Sabzevari [9] conducted
Co
wind tunnel studies of ducted flow around a Savonius rotor where the duct channeled flow from a large
ot
area onto the blades. This resulted in a doubling of rotor speed albeit with the trade-off of losing the
tN
omnidirectional characteristic of VAWTs. Sivasegaram [10] followed up Sabzevari’s study by examining the impact of individual components of the duct on power output. He concluded that a power
sc
rip
amplification factor of 1.5 would be possible with the ducted design. In an analogous design, Fukutomi
et al. [11] generated similar results with a maximum power amplification factor of 1.7.
Ma
nu
In order to restore omnidirectionality to the VAWT design, Sivapalan and Sivasegaram [12, 13] used
an eight-bladed rotor with external vanes numbering up to eight flat blades at a variety of angles. They
found that they could increase power output by up to 60% for the three-vaned stator.
ed
Altan et al. [14, 15] used a curtain design similar to Sivapalan and Sivasegaram to augment the
flow around a Savonius rotor. Using a double-bladed Savonius rotor, they placed two curtains in the
pt
upstream-only position. At higher wind speeds, performance actually dropped indicating a possible
ce
increase of back pressure at higher speeds. At moderate speeds, they reported an increase of power
Ac
coefficient from 0.16 to 0.35 with the addition of the shroud. Pope et al. [16–18] examined the performance of what they termed a “Zephyr wind turbine”. This turbine had a stator with tabs on the back
of each stator blade (reverse winglets). Their modeling demonstrated that the stator tabs provided a
detrimental performance on power output, likely due to induced vortices in the flow.
All of the above research involved channeling the flow along the power producing path and then
allowing it to leave the stators on the downstream side. The current research hypothesizes that forcing
3
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
the2017
flowbytoASME
swirl inside a stator, by minimizing exit on the downwind side and forcing it to exit above the
stator, will result in an increase in harnessed energy. This study examines the flow field created by such
a stator. In the long term, the results of the current study will be used to design and evaluate stators for
use in standard Darrieus and Savonius VAWTs. A similar design has been proposed by Korprasertsak
et al. [19–21]. They used CFD analyses of the rotors to determine that the stators could increase the
ed
ite
d
rotational velocity of the Savonius rotor by up to 53%.
One class of atypical VAWTs, termed tornado-type, has also implemented a similar system [22–24].
These turbines utilized the low pressure vortex created by the trapped flow to pull in air from below
py
through an axial fan turbine positioned vertically. While the concept of tornado-type wind turbines
Co
is promising, the actual performance is quite low in comparison to the footprint [25]. Hsu [25] and
Volk [26], among others, made the argument that the tornado-type turbines should measure Cp based
ot
on area of the internal turbine fan rather than the more typical frontal area of the turbine. This allows
tN
them to report Cp much higher than the Betz limit, but disregards the fact that for the same footprint,
sc
rip
conventional turbines would harvest more energy. In fact, when frontal area is taken into account the
power coefficient will drop under 0.25 for even the best reported results. This is in comparison to
measured power coefficients of conventional VAWT turbines in excess of 0.4 [27]. Another atypical
Ma
nu
VAWT design combines Darrieus and Savonius rotors into the same system to lower the starting speed
and increase the power output [28].
The current study examines whether it is possible to induce the same rotational flow exhibited by the
ed
tornado-type turbine while still harvesting wind over the entire frontal area. As such, the flow field in
pt
the posited stator seeks to be to be similar to the results reported by de Farias Neto, Legentihomme, and
ce
Legrand [29], in which the flow tangentially enters a cylinder with an initially strong angular component.
Ac
As the flow travels around the cylinder, it gradually acquires an axial component. The axial component
then grows to dominate the velocity field.
In order to understand how the flow will react to becoming trapped inside the stator, a CFD model
was used to examine the flow structures within the stator, as well as effects of varying the number of
stator blades. Being a first study of this stator design, the rotor was not included in the design of the
experimental and computational models.
4
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017 by ASME
Methods
Model Design
The CFD model was created to a scale that would allow for eventual wind tunnel comparison and
was dimensioned as shown in Fig. 1. The support plates and inner cylinder were all 1/4” thick and
the blades were 0.06” thick. Twenty-four blades were used for the baseline model while additional
ed
ite
d
simulations were run with twelve and six blades to study the effectiveness of larger blade spacing. Each
blade had a chord length of 1.75” and made a 56◦ angle with the stator outer diameter. The inner hub
py
had a diameter of 5”.
Co
Computational Fluid Dynamics (CFD)
The numerical simulations used the commercial CFD code Star-CCM+ [30]. The utilized physics
ot
model was a three-dimensional, SST (Menter) k-ω RANS with unsteady, incompressible, turbulent,
tN
segregated flow (see Table 1 for full details). A polyhedral mesh was used adjacent to the stator to
allow for smooth size transitions. Further away from the stator, the mesh was changed to hexahedral
sc
rip
cells to take advantage of the more efficient cell placement. The near-wall effects were simulated using
Star-CCM+’s “all-y+” model in which turbulent boundary layer theory is used to model the velocity
Ma
nu
distribution in the cell nearest the wall when y+ is greater than thirty, but when y+ is less than one the
boundary layer is directly simulated. If y+ falls between one and thirty then a hybrid method is utilized
to determine the boundary layer profile. In the current simulation, the layer near the walls of the stator
assembly was maintained at y+ less than one, but the walls of the wind tunnel were maintained at y+
ed
slightly more than thirty. This arrangement allowed for maximum fidelity of the boundary layer in the
ce
walls.
pt
stator while also still accurately modeling the development of the boundary layer at the wind tunnel
Ac
The total mesh count was 18 million cells with an average base size in the stator region of 2.7mm
which was reduced to 0.9mm near the stator blade tips. In the hexahedral section the mesh base size
was increased to 9mm. As can be seen in Fig. 2, the areas of high churn were treated with a reduced
cell sizes.
The cross section of the flow field was set up as a virtual wind tunnel with a cross section of 1m x 1m
and a no-slip boundary condition at all four walls. A total physical time of 0.5 seconds was simulated,
5
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
requiring about 64,800 processor hours. The average time step of 3.0 x 10-5 seconds for the 33.5 m/s
case was controlled so that a mean Courant number of 0.8 was maintained. The flow field was initialized
with the output from previous results using coarser grids such that the initial solution was already mostly
converged. Free stream wind speeds of 20 m/s, 25 m/s, 30 m/s, and 33.5 m/s were studied. The system
ed
ite
d
boundary conditions and settings are summarized in Table 1.
Numerical Uncertainty
Initial simulations were conducted at coarse resolutions on the 33.5 m/s velocity case to establish
py
a baseline flow field. The mesh was gradually increased in resolution until the engineering parameters
of interest (mass flow into the stator, mass flow out of the stator sides, mass flow out of the stator
Co
top, and velocity distribution at 90◦ ) remained constant between simulations. Each refinement in this
ot
process resulted in an increase in the number of cells by 10-30%. A grid dependence study was then
tN
completed whereby the base cell sizes were reduced by another 10%, resulting in an approximately
38% increase in the number of cells to 22 million. Both the base simulation and the grid dependence
sc
rip
simulation converged to a stator entry mass flow rate of 0.75 kg/s with 0.465 kg/s exiting the stator
sides and 0.285 kg/s exiting out of the stator top. Additionally, each simulation resulted in all residuals
progressing below 10−5 . Figure 2b shows the fidelity in velocity distribution at 90◦ between the different
Ma
nu
cell counts. The grid dependence study showed that the 18 million cells used in the base design were
sufficient for resolving the flow characteristics of interest.
pt
Baseline Flow Field
ed
Results
ce
Contours of the tangential velocities inside the stator are shown in Fig. 3. These figures demonstrate
Ac
that the stator array trapped the flow and created a high tangential velocity component with local velocities that exceed the incoming free stream velocity by up to 64.5% (see Table 2). The general shape
of the flow patterns remained constant over the simulation time period, but the size and position of the
structures were found to vary with time.
As mentioned above, part of the objective for the present study was to see if it would be possible to
induce swirling flow similar to the tangentially injected flow modeled in the de Farias Neto et al. study
6
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017The
by ASME
[29].
velocity vectors in Fig. 3 show that, while swirling flow was established, the flow still had a
significant amount of leakage from the internal cylinder on the downstream side instead of out the top
of the stator. In fact, only 1/3 of the flow exited from above the stator. Figure 4 shows the bypass rates
as a function of Reynolds number based on free stream velocity and the external diameter of the stator.
ed
ite
d
This quantity was defined as:
ṁout
∗ 100%,
ṁin
(1)
py
where ṁin and ṁout represent mass flow entering and exiting the stator array respectively. Figure 1
Co
shows the boundaries where mass flow at the entry and exit were calculated. Note that flow moving
inwards towards the center of the cylinder is measured as mass flow entry and flow crossing the boundary
ot
outwards from the center of the cylinder is measured as mass flow exit. For a stator configuration with 24
tN
blades, the bypass rate between the four Reynolds number cases averaged 63% with a standard deviation
sc
rip
of just one percentage point demonstrating that little variation occurs with change in Reynolds number.
To show that bypass rate is dependent upon the gap between the blades, bypass rates for twelve and six
ed
Ma
nu
blades have been included on Fig. 4 as well.
pt
Tangential Velocity Profiles
ce
Examining the tangential flow profiles over time shows how the flow varies in the stator. Figure 5
Ac
plots the tangential velocity profiles at the 0◦ , 90◦ , 180◦ , and 270◦ positions from zero to three seconds
with 0.05s intervals. The 0◦ position is defined at the upstream position, and the angle increases with the
direction of flow (counter-clockwise), as shown in Fig. 1. These results show that the flow field at the
0◦ , 180◦ , and 270◦ positions are not steady and vary in time. This variation in time explains the relative
difference in the velocity increases shown in Table 2. The maximum velocities found in the flow field
are going to fluctuate with the small changes in unsteady flow structures.
7
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017 by Coefficient
ASME
Pressure
Contours
Non-dimensional pressure coefficients were evaluated using the following equation:
pstatic
1
2
2 ρV∞
(2)
ed
ite
d
Cp =
where pstatic is the static pressure, ρ is the fluid density, and V∞ is the free stream velocity. Pressure
coefficient contours are shown in Fig. 6 for the 20 m/s and 33.5 m/s free stream velocities. The expected
reduction in pressure in areas of high velocity are present as well as a high pressure field upstream due
py
to the induced pressure drop of the stator. Along the external sides of the stator (at 90◦ and 270◦ ), lower
Co
pressure areas were created due to high velocity and wake recirculation.
ot
Effects of Vertical Position and Blade Number
tN
Figure 7 demonstrates the velocity contours at heights 5cm above and below the centerline and
sc
rip
at the centerline. The general shape of the flow field remains the same with only minor differences
between them. The vorticity variation with height (Fig. 8) is similar although the vorticity reveals some
interesting characteristics of the flow. The inside tips of all the blades generated significant vorticity.
Ma
nu
On the incoming side, these streams of high vorticity join together so that by the time the flow reaches
the 90◦ position the entire cross-section is experiencing vorticity well above the free stream levels. The
downstream blades demonstrate the leakage effect as the small streams of high vorticity are created as
ed
the flow passes the back side of each blade.
Pressure and velocity contours for the 6-blade and 12-blade systems are shown in Fig. 9. The
pt
contours demonstrate the increased bypass and reduction in tangential flow with the fewer number of
ce
blades. Yet, the increase in local velocity from the inner edge of the blades is starting to become apparent
Ac
in the 12 blade arrangement.
Conclusions
Removing the influence of the rotor from the stator system gives a chance to study effects that may
still exist but not be apparent in the full system. The physics of primary interest is the presence of the
three main sources of adverse pressure gradients shown in Fig. 6. The first (APG1) is located upstream
8
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
by ASME
of 2017
the stator
(at 0◦ outside of the stator blades) where impingement upon the blades creates a high
pressure due to stagnation points. The second source of adverse pressure gradient(s) (APG2) is actually
a multiple occurrence as they are located on the inside of the downstream stator blades (135◦ to 270◦ ).
The third source (APG3) is located between the stator blades and cylinder on the upstream side where
the flow streams from between multiple stator blades converge.
ed
ite
d
For APG1, the velocity vector field in Fig. 3 shows that this pressure gradient results in some flow
being diverted around the outside of the stator assembly. While this does not meet our objective of
channeling all of the entering flow through the stator, the presence of a similar pressure gradient and
py
flow diversion also exists in standalone wind turbines [31–33] so the reduction in energy harvested due
Co
to this adverse pressure gradient would be small. Care should still be taken in arbitrarily increasing the
number of blades as the upstream high pressure effect will increase as the stator become more of a blunt
ot
body.
tN
The high pressure of the multiple APG2 sources forces flow to the right, toward lower pressure
external air, and out of the stator envelope. This effect is in conjunction with the large magnitude of
sc
rip
vorticity generated within the inside of the previous blade. The high vorticity region removes momentum
from the tangential flow, making the flow more susceptible to being directed in an alternative direction.
Ma
nu
The high pressure of the stagnation point then forces flow to the right, toward lower pressure external
air, and out of the stator envelope. In a system with a rotor, these leaks are likely to be amplified by the
forward motion of the rotor pushing on air ahead of the blades. Thus, the blade on the Savonius rotor
to be used in conjunction with a stator assembly should be designed so as to push the air away from the
ed
outside and back toward the center of rotation.
pt
Finally, since the velocity vectors of Fig. 3 show vectors impinging on the cylinder, it is probable that
ce
the high pressure of APG3 is due to a stagnation effect. Yet, the high pressure that spreads out away from
Ac
the apparent stagnation point is highly asymmetric with low pressure above and high pressure below.
The high pressure is in a region of higher velocity flow which would normally represent lower pressures.
Thus, it is likely that the whole high pressure zone is a combination of the stagnation point along with
the convergence of multiple flow streams coming from between stator blades (315◦ to 45◦ ). As a result
of this high pressure region, the flow from the upper left quadrant (approximately 315◦ ) is diverted into
clockwise (negative) tangential flow. That clockwise flow meets the original counter-clockwise (posi9
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017
by ASMEflow at about the 225◦ location forcing both flows out of the stator assembly through
tive)
tangential
adjoining stator vanes. While the high pressure created by the stagnation point would mostly be ameliorated with the addition of a moving rotor, the high pressure due to the converging streams would most
likely remain. Of the three pressure gradients, this is the most serious as the diverted clockwise flow
will provide pure resistance to the movement of a rotor in the assembly. One possible way to minimize
ed
ite
d
this effect would be to introduce angled slots just on a small portion of the blades near the outside of
the assembly as shown in Fig. 10. When the flow is parallel to the blade the slots will provide minimal
hindrance to the flow, but for the flow going around the outside of the blades the slots would direct the
py
flow onto the curved inner portion of the adjacent blade. Not only will this direct more flow into the
assembly, but the impingement of flow on the blade will create high pressure stagnation points. The
Co
high pressure should help to keep flow from exiting to the sides of the stator assembly which will hinder
tN
ot
backwards circulation inside the assembly.
Acknowledgments
sc
rip
This work was completed utilizing the High Performance Computing Center facilities of Oklahoma
State University at Stillwater. The authors would like to thank Geoffrey Donnell for his assistance with
Ma
nu
the work.
References
ed
[1] Strickland, J. H., Webster, B., and Nguyen, T., 1979, “A vortex model of the Darrieus turbine: an
analytical and experimental study,” ASME J. Fluids Eng., 101(4), pp. 500–505.
pt
[2] Gosselin, R., Dumas, G., and Boudreau, M., 2013, “Parametric study of H-Darrieus vertical-axis
ce
turbines using uRANS simulations,” In 21st Annual Conference of the CFD Society of Canada,
Ac
Sherbrooke, Canada, pp. 6–9.
[3] Ikoma, T., Masuda, K., Maeda, H., and Sasanuma, T., 2007, “A basic study on characteristics
of torque and hydrodynamic force of darrieus water turbines,” In Conference Proceedings of
JASNAOE, Vol. 4, pp. 51–54.
[4] Kotb, M., and Aldoss, T., 1991, “Flowfield around a partially-blocked Savonius rotor,” Applied
Energy, 38(2), pp. 117–132.
10
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)[5]
2017
by ASME
Shaughnessy,
B., and Probert, S., 1992, “Partially-blocked Savonius rotor,” Applied Energy, 43(4),
pp. 239–249.
[6] Mohamed, M., Janiga, G., Pap, E., and Thévenin, D., 2011, “Optimal blade shape of a modified savonius turbine using an obstacle shielding the returning blade,” Energy Conversion and
Management, 52(1), pp. 236–242.
ed
ite
d
[7] Zhang, B., Song, B., Mao, Z., and Tian, W., 2017, “A novel wake energy reuse method to optimize
the layout for savonius-type vertical axis wind turbines,” Energy, 121, pp. 341–355.
[8] Craig, A. E., Dabiri, J. O., and Koseff, J. R., 2016, “Flow kinematics in variable-height rotating
py
cylinder arrays,” ASME J. Fluids Eng., 138(11), p. 111203.
[9] Sabzevari, A., 1977, “Performance characteristics of concentrator-augmented Savonius wind ro-
Co
tors,” Wind Engineering, 1, pp. 198–206.
ot
[10] Sivasegaram, S., 1979, “Concentration augmentation of power in a Savonius-type wind rotor,”
tN
Wind Engineering, 3, pp. 52–61.
[11] Fukutomi, J., Shigemitsu, T., and Daito, H., 2011, “Study on performance and flow condition of a
sc
rip
cross-flow wind turbine with a symmetrical casing,” ASME J. Fluids Eng., 133(5), p. 051101.
[12] Sivapalan, S., and Sivasegaram, S., 1982, “Power augmentation in a Savonius-type wind-turbine
by using a single air-deflecting vane,” Regional Journal of Energy, Heat and Mass Transfer, 4,
Ma
nu
pp. 187–193.
[13] Sivasegaram, S., 1986, “Power augmentation in wind rotors - A review,” Wind Engineering, 10,
pp. 163–179.
ed
[14] Altan, B. D., Atilgan, M., and Ozdamar, A., 2008, “An experimental study on improvement of
ce
pp. 1673 – 1678.
pt
a Savonius rotor performance with curtaining,” Experimental Thermal and Fluid Science, 32(8),
Ac
[15] Altan, B. D., and Atilgan, M., 2010, “The use of a curtain design to increase the performance level
of a Savonius wind rotors,” Renewable Energy, 35(4), pp. 821–829.
[16] Rowe, J., 2004, “Vertical axis wind turbine,”, May 25 US Patent 6,740,989.
[17] Pope, K., Rodrigues, V., Doyle, R., Tsopelas, A., Gravelsins, R., Naterer, G. F., and Tsang, E.,
2010, “Effects of stator vanes on power coefficients of a zephyr vertical axis wind turbine,” Renewable Energy, 35(5), pp. 1043–1051.
11
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017
by ASME
[18]
Pope,
K., Dincer, I., and Naterer, G., 2010, “Energy and exergy efficiency comparison of horizontal
and vertical axis wind turbines,” Renewable Energy, 35(9), Sept., pp. 2102–2113.
[19] Korprasertsak, N., Korprasertsak, N., and Leephakpreeda, T., 2014, “Cfd modeling and design of
wind boosters for low speed vertical axis wind turbines,” Advanced Materials Research, 1016,
pp. 554–558.
ed
ite
d
[20] Korprasertsak, N., and Leephakpreeda, T., 2015, “Optimal design of wind boosters for low speed
vertical axis wind turbines,” Applied Mechanics and Materials, 798, pp. 195–199.
[21] Korprasertsak, N., and Leephakpreeda, T., 2016, “Analysis and optimal design of wind boosters
py
for vertical axis wind turbines at low wind speed,” Journal of Wind Engineering and Industrial
Aerodynamics, 159, pp. 9–18.
Co
[22] Yen, J. T., 1975, “Tornado-type wind energy system,” In Energy 10; Annual Intersociety Energy
ot
Conversion and Engineering Conference, Vol. 1, pp. 987–994.
tN
[23] Yen, J. T., 1978, “Tornado-Type Wind Turbine,” US Patent 4,070,131.
[24] Hsu, C., 1984, “Tornado type wind turbines,”, June 5 US Patent 4,452,562.
sc
rip
[25] Hsu, C., and Minachi, A., 1990, “Performance tests of tornado-type wind turbine models,” Journal
of Propulsion and Power, 6(2), pp. 181–185.
[26] Volk, T., 1982, “Performance of tornado wind energy conversion systems,” Journal of Energy,
Ma
nu
6(5), pp. 348–350.
[27] Eriksson, S., Bernhoff, H., and Leijon, M., 2008, “Evaluation of different turbine concepts for
wind power,” Renewable and Sustainable Energy Reviews, 12(5), pp. 1419–1434.
ed
[28] Rassoulinejad-Mousavi, S., Jamil, M., and Layeghi, M., 2013, “Experimental study of a combined
ce
211.
pt
three bucket h-rotor with savonius wind turbine,” World Applied Sciences Journal, 28(2), pp. 205–
Ac
[29] de Farias Neto, S., Legentilhomme, P., and Legrand, J., 1998, “Finite-element simulation of laminar swirling decaying flow induced by means of a tangential inlet in an annulus,” Computer
Methods in Applied Mechanics and Engineering, 165(1-4), pp. 189 – 213.
[30] CD-adapco, 2016, Star-CCM+.
[31] Sanderse, B., and Koren, B., 2009, “Energy preservation in the numerical calculation of wind
turbine wakes,” Euromech Colloquium 508 on Wind Turbine Wakes, p. 16.
12
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017
by ASME A., Murai, Y., Tasaka, Y., and Takeda, Y., 2011, “Interactive flow field around two
[32]
Shigetomi,
savonius turbines,” Renewable Energy, 36(2), pp. 536–545.
[33] McLaren, K., Tullis, S., and Ziada, S., 2012, “Computational fluid dynamics simulation of the
aerodynamics of a high solidity, small-scale vertical axis wind turbine,” Wind Energy, 15(3),
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Co
py
ed
ite
d
pp. 349–361.
13
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
2017
ASME
List
of by
Tables
1
Simulation boundary conditions and settings. . . . . . . . . . . . . . . . . . . . . . .
2
Comparison of free stream velocity (V∞ ) to maximum internal flow tangential velocity
15
(Vθ max ). Re is calculated based on external stator diameter and the free stream velocity.
ed
ite
d
The percent increase is calculated by (Vθ max −V∞ )/V∞ . . . . . . . . . . . . . . . . . .
List of Figures
Schematic of the stator model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
(a.) Mesh distribution around the stator assembly (b.) Result of the grid sensitivity study
17
py
1
16
Co
between the low (15 million cells), mid (18 million cells), and high (22 million cells)
cell count cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Contours of normalized tangential velocity inside the stator at the central plane with
ot
3
18
tN
velocity vectors overlaid on (d.) as representative of the shape of the flow field. Flow is
from left to right and is normalized by free stream velocity (V∞ ). Re = (a.) 278,311, (b.)
sc
rip
347,888, (c.) 417,466 (d.) 466,171. Reynolds number (Re) is defined based on external
. . . . . . . . . . . . . . . . . .
19
4
Bypass rate vs. Re. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
5
CFD time evolution results of the tangential velocity inside the stator at the indicated
Ma
nu
diameter of the stator and free stream velocity (V∞ ).
polar positions for V∞ = 33.5 m/s and Rmax = 105mm (central plane of stator). . . . . .
Pressure coefficient (equation 2) contours for (a.) 20 m/s and (b.) 33.5 m/s (mid-height
ed
6
21
of the stator). Adverse pressure gradients (APG) 1 and 2 are shown in (a.) while APG3
Depiction of the variation of the vorticity contour (V∞ = 33.5 m/s) with height for 5cm
below the centerline, the centerline, and 5cm above the centerline (left to right). . . . .
9
10
23
Ac
for 5cm below the centerline, the centerline, and 5cm above the centerline (left to right).
8
22
Depiction of the variation of normalized velocity contour (V∞ = 33.5 m/s) with height
ce
7
pt
is shown in (b.). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
24
Pressure coefficient (equation 2) and normalized velocity (V∞ = 33.5 m/s) contours for
(a.) 6 and (b.) 12 blades (mid-height of the stator). . . . . . . . . . . . . . . . . . . . .
25
Slot concept shown on a single blade. . . . . . . . . . . . . . . . . . . . . . . . . . .
26
14
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
py
– Simulation boundary conditions and settings.
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Co
Table 1
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
15
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
Re
V∞ (m/s) Vθ max (m/s) Increase
278,311
20
31.1
56%
347,888
25
40.7
62.8%
417,466
30
47.7
59%
466,171
33.5
55.1
64.5%
– Comparison of free stream velocity (V∞ ) to maximum internal flow tangential velocity (Vθ max ). Re is calculated based on external stator diameter and the
free stream velocity. The percent increase is calculated by (Vθ max −V∞ )/V∞ .
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Co
py
ed
ite
d
Table 2
16
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
– Schematic of the stator model.
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Fig. 1
Co
py
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
17
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
sc
rip
tN
ot
Co
py
(a).
(b).
– (a.) Mesh distribution around the stator assembly (b.) Result of the grid
sensitivity study between the low (15 million cells), mid (18 million cells), and
high (22 million cells) cell count cases.
Ac
ce
pt
ed
Ma
nu
Fig. 2
18
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
Co
py
ed
ite
d
a.
Ma
nu
sc
rip
tN
ot
b.
d.
Ac
ce
pt
ed
c.
– Contours of normalized tangential velocity inside the stator at the central
plane with velocity vectors overlaid on (d.) as representative of the shape of the
flow field. Flow is from left to right and is normalized by free stream velocity (V∞ ).
Re = (a.) 278,311, (b.) 347,888, (c.) 417,466 (d.) 466,171. Reynolds number (Re)
is defined based on external diameter of the stator and free stream velocity (V∞ ).
Fig. 3
19
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
– Bypass rate vs. Re.
Ac
ce
pt
ed
Ma
nu
Fig. 4
sc
rip
tN
ot
Co
py
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
20
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
ot
Co
py
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
– CFD time evolution results of the tangential velocity inside the stator at the
indicated polar positions for V∞ = 33.5 m/s and Rmax = 105mm (central plane of
stator).
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
Fig. 5
21
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
b.
tN
ot
Co
py
a.
– Pressure coefficient (equation 2) contours for (a.) 20 m/s and (b.) 33.5 m/s
(mid-height of the stator). Adverse pressure gradients (APG) 1 and 2 are shown in
(a.) while APG3 is shown in (b.).
Ac
ce
pt
ed
Ma
nu
sc
rip
Fig. 6
22
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
– Depiction of the variation of normalized velocity contour (V∞ = 33.5 m/s)
with height for 5cm below the centerline, the centerline, and 5cm above the centerline (left to right).
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Co
py
Fig. 7
23
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
– Depiction of the variation of the vorticity contour (V∞ = 33.5 m/s) with
height for 5cm below the centerline, the centerline, and 5cm above the centerline
(left to right).
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Co
py
Fig. 8
24
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c)
a. 2017 by ASME
Co
py
ed
ite
d
b.
– Pressure coefficient (equation 2) and normalized velocity (V∞ = 33.5 m/s)
contours for (a.) 6 and (b.) 12 blades (mid-height of the stator).
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Fig. 9
25
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
– Slot concept shown on a single blade.
Ac
ce
pt
ed
Ma
nu
sc
rip
tN
ot
Co
py
Fig. 10
ed
ite
d
Journal of Fluids Engineering. Received February 25, 2017;
Accepted manuscript posted October 10, 2017. doi:10.1115/1.4038166
Copyright (c) 2017 by ASME
26
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Документ
Категория
Без категории
Просмотров
0
Размер файла
18 817 Кб
Теги
4038166
1/--страниц
Пожаловаться на содержимое документа