2nd IEEE ENERGYCON Conference & Exhibition, 2012 / Advances in Energy Conversion Symp

A BUILDING-INTEGRATED EOLIC SYSTEM

FOR THE EXPLOITATION OF WIND ENERGY IN URBAN AREAS

S. Carcangiu, A. Montisci

University of Cagliari, Department of Electrical and Electronic Engineering,

Piazza d’Armi – 09123 Cagliari (Italy)
ABSTRACT interested to exploit wind energy in urban areas, both to
integrate the energy produced by the sun, and to increase the
Renewable energy exploitation in the urbanized areas has continuity of production, so that the energy to store or to
been receiving an increased attention over recent years exchange with the electric network can be reduced.
because the proximity to electricity users makes it possible In this paper an innovative system is presented that
to obtain better global efficiency and to save territory. In aims to solve some drawbacks of wind generation (such as
this work a new wind energy system that uses the roof of the the encumbrance of the blades, noise, cut-in and cut-off
building in combination with a centripetal turbine is velocities of the wind, presence of turbulence, etc…) in
presented. This “Aeolian roof” could be considered as an order to make possible its extensive use in urban areas. The
amplifier of the wind speed that should allow the wind generation system substitutes the roof of the building
turbine to work under both lower and higher velocities than carrying out the same covering function. Fig. 1 shows this
the conventional aerogenerators and therefore to increase its “Aeolian roof” that consists of an external static structure,
delivered power output. Considering the local wind climate the stator, whose task is conveying the flow towards the
of the area of interest, in this work, firstly, a Computational centre of the roof, where a vertical axis centripetal turbine
Fluid Dynamics (CFD) model is used to simulate the flow transforms the wind energy in mechanical energy.
into the eolic system. Then, a laboratory-scale prototype has Such a layout of the system entails several advantages.
been made in order to test the potentialities of the proposed Firstly, the visual impact and safety problems of the turbine
system for fulfillment of energy demands. are definitively avoided, because it is hidden inside the
structure. Secondly, the wind direction is not a problem,
Index Terms— Wind energy, Aeolian Roof, Integrated because the vertical axis of the rotor allows it to capture the
Wind Turbines, Buildings, Computational Fluid Dynamics wind no matter which direction it comes from. Thirdly, the
stator allows one to intercept a wide section of flow, even if
1. INTRODUCTION the dimension of the turbine is very small, depending on the
ratio between the inlet and the outlet cross section of the
As it is well known, wind and sun are the most promising stator. This allows one to exploit a wider range of wind
renewable energy sources, both for the concentration and velocities with respect to common wind generators, because
availability [1]. Typically, the wind energy is preferred in of a lower cut-in and unlimited cut-off velocities.
extra urban areas, because of the better ratio between In this work, in according with the wind characteristics
produced energy and used land with respect to other in the built environment, a laboratory-scale prototype of this
renewable energy sources. Therefore, the integration of new system, representing “Aeolian roof”, firstly modeled
wind energy conversion systems into buildings is a small with a CFD tool, is realized and monitored in order to
but growing trend, and it has a great scope in generating evaluate its energetic performances. This prototype is
electricity from the wind. In fact, in Europe there is positioned in the roof of a building sited at Cagliari in the
increasing interest from home owners and businesses to south of Sardinia, where the speed and the direction of the
install small wind turbines on the rooftop of houses, wind have been monitored during the last year by a weather
buildings of school, hospitals, commercial buildings, or station.
even on tall buildings [2]. However the main reasons for this The present paper is organized as follows. In Section 2
small trend of progress appear to be the urban constraints, the wind characteristics of the site where the prototype is
the high turbulence triggered by neighboring buildings, the positioned are reported and the layout of the system is
difficulty to re-orienting the generator on the basis of the described. In Section 3 the tridimensional numerical model
wind direction, and so on [3]. Furthermore, wind patterns in of the system and the equations describing the physical
the urban environment are complex, because the air passes phenomenon are reported. In Section 4 the experimental
over, around, and between the buildings, and creates high setup is described and the obtained results are discussed.
turbulence [4]. In spite of these problems, one could be Some conclusions will end the paper.

978-1-4673-1454-1/12/$31.00 © 2012 IEEE 172

 Fig. 2 Wind rose shows the prevailing wind direction and speed.

Fig. 1. Example of the building-integrated eolic system studied.

2. WIND RESOURCE ANALYSIS AND SYSTEM

DESCRIPTION

When contemplating the embedding of wind power

generation into a building design, the first consideration
must be the local wind climate of the area [5]. As mentioned
above, the prototype developed in this study is placed on the
roof of a building of Cagliari (latitude 39N, longitude 9E) in
the south of Sardinia. During the last year a weather station
Fig. 3 Weibull pdf of the distribution of the wind speed during the
has monitored the wind trend in order to establish its main last year in the roof of a building of Cagliari.
directions and speed.
Wind Energy often is expressed in the form of a Wind month or 1 year. It has been used commonly both in wind
Rose with reference to a Weibull Distribution. In Fig. 2 the speed and wind energy analysis.
graphical representation of the distribution of wind speeds In Fig. 3 the Weibull pdf histogram for the average
and directions is shown. This wind rose indicates that most 10 m wind speed distribution in correspondence of the
of time the wind blows from the north and north-west and weather station, is shown. This plot has been obtained
that these winds are also the strongest and therefore considering respectively k = 1.7 and λ = 4.2 as Weibull
potentially those which will generate the most part of the parameters, that have been obtained using the MATLAB
averaged energy. However, the contribution of wind Statistical Toolbox considering the collected data from the
blowing from the other directions is not negligible. weather station.
The wind speed (and indirectly the power) is described Considering this wind trend, a prototype of the eolic
by the two parameters Weibull distribution [6] that is an system has been developed with four inlet sections in order
asymmetrical distribution whose probability density to favour the more frequent directions and speeds of the
function (pdf) is given by: wind.
k ⎛U ⎞
k −1
⎡ ⎛ U ⎞k ⎤ The prototype is composed by a fixed part, the stator,
f (U ) = ⎜ ⎟ exp⎢− ⎜ ⎟ ⎥ (1)
λ⎝λ ⎠ with square shape and area equal to 1 m2. The moving part,
⎢⎣ ⎝ λ ⎠ ⎥⎦ the impeller, is a vertical axis centripetal turbine, with 7
where k is the shape parameter, λ the scale parameter and U blades. The turbine is a commercial product (see Fig. 9), so
the wind speed. The two parameters Weibull distribution is that in designing the diameter of the turbine the choice is
accepted as the best among the different wind speed limited to the catalogue values. In this case, the diameter of
distribution models used in literature, because one can the impeller is equal to 12 cm and the inlet section of its
adjust the parameters to suit a period of time, usually 1 ducts is equal to 5x1 cm2. The stator blades have a

173
 Network
deficit of power has to be managed by a set of lead-acid
batteries.
Control Unit
Not represented in the diagram of Fig. 4 is a data
AC DC
acquisition system that records all variables, i.e. voltages,
PM ~ = currents and wind speed.
generator
DC AC
Wind kWh
= ~
turbine
Rectifier Inverter 3. THE CFD MODEL

The tri-dimensional system, shown in Fig. 1, has been

Battery
bank
modelled and the commercial software Comsol
Multiphysics® 4.2a [7], which implements the CFD
analysis, has been used to solve the model equations.
Fig. 4 A schematic of the energy conversion system.
A 3D analysis of the model has been performed, taking
helicoidal shape, so that a great ratio between the two into account a wind flow around and inside the building
terminal sections of the duct can be achieved. model described by the Navier-Stokes equations [8] that
The rate of convergence of stator duct is a crucial have been solved for the velocity field and the pressure.
design parameter. In fact, a great ratio between the two Considering an incompressible flow and a stationary study,
terminal cross sections implies that a wide tube of flux is the nonlinear system of equations is the following:
captured, but it is processed by a small turbine, in this way
obtaining advantages both in terms of cut-in and cut-off
((
ρu ⋅ ∇u - ∇ ⋅ μ ∇u + (∇u )
T
)) + ∇ ⋅ p = F (2)
velocities of the wind. In fact, the friction and the inertia of ∇⋅u = 0
the impeller are smaller, and then the turbine can get to where μ denotes the dynamic viscosity, u the velocity
rotate even for very small wind velocities. In turn, in the vector, ρ the density of the fluid, p the pressure and F is a
presence of strong wind, the centrifugal strength is strongly body force term.
dependent on the diameter of the impeller, so that a small The modelled fluid is air with viscosity 10−5 [N·s/m2]
turbine can carry higher rotating velocities. and density 1.2 kg/m3. The first equation is the momentum
On the other hand, part of the wind hinting the building balance, and the second one is the equation of continuity for
bypasses it and the ratio between the bypassing flow and the incompressible fluids.
one evolving in the system depends on the reciprocal fluid The following boundary conditions are applied in the
dynamic impedance of two alternative paths. For this computational domain:
reason, the designer has to find a crossover point between
the two conflicting exigencies of having a small turbine and • the inlet and the outlet conditions are set on the
capturing a great quantity of energy. opposite planes YZ;
The outlet flow of the turbine has vertical direction and • assuming the distribution of wind speed during the last
it escapes through a divergent duct, at the end of which the year provided by the monitoring of the weather station
exhaust flow intersects the bypassing flow. The divergent (Fig. 3), the inlet velocity vector is set normal to the
duct allows one to increase the velocity of the air in the
turbine, and the depression on the top of the building
contributes to reduce the impedance of the internal path of
the air. The impedance of the bypassing flow can be
increased by means of mouldings in the facade of the
building, but these further possibilities to improve the
performance have not been explored in this work.
Regarding the mechanical to electrical power
conversion in the proposed system, this can be achieved
through the electrical energy conversion system shown in
Fig. 4. The wind energy is converted to electricity by a
Permanent Magnet (PM) generator. The electrical energy is
produced at variable frequencies. First, it is conditioned for
24-volts dc voltage by a control rectifier (Fig. 4). The
control rectifier is equipped with the necessary control
circuitry to protect the batteries against overcharging. The
energy has to be supplied to the energy network after being
transformed to 230-volts ac voltage by a bi-directional,
microprocessor-controlled inverter/charger; any surplus, or
Fig. 5 Geometry and pressure map of the 3D model.

174
 u = 5 m/s

Fig. 6 Velocity map of the 3D model. Fig. 8 Velocity profile of the minimal section of the 3D duct.

in the duct owing to the pressure differential between the

duct inlet at the front façade and the outlet at the roof.
Fig. 6 and Fig. 7 show the magnitude of the wind
velocity of the solution of the 3D model and 2D section
respectively. In Fig. 8 the velocity profile along an
intermediate line of cross-section is shown. As it can be
seen, the stator makes it possible to increase the wind
velocity up to 8 m/s, in this way concentrating the power of
the wind in a limited area.
The available power P of an air flow crossing the
section A (measured in m2) is equal to:
1
2 ∫A
P= ρ ⋅U 3dA (3)

where P is the power, ρ is the density of the air and U is the

velocity of the wind.
Fig. 7 Velocity map of a bidimensional section of the model. As the power is proportional to the cube of velocity, the
reduction of flow due to a decrease of the minimal section
inlet surface and equal to u ⋅ n = 5 [m / s ] ; can be partially compensated by the velocity increase. By
integrating the power in the minimal section of the duct, we
• the Outlet boundary condition sets the constant
obtain the power available at the inlet of the turbine. Such
pressure to atmospheric value of p0 = 1 [Bar];
value can be compared with the power of the wind crossing
• the No-slip boundary condition, that eliminates all an equal section in a point far from the building, so that the
components of the velocity vector ( u = 0 ), is set on the velocity is uniformly equal to 5 m/s. The ratio between the
ground and the walls of the building (gray surfaces in two values of power gives us the coefficient of power Cp
Fig. 5); due to the use of the statoric structure. In the analyzed
• a Symmetry condition, applied to all the remaining problem the obtained value of Cp is equal to 1.8.
surfaces, states that the zero normal velocity and the
normal gradients of all variables are equal to zero. 4. INITIAL EXPERIMENTS

In Fig. 5 the 3D model of the system and the pressure 4.1. Experimental set-up
contours are reported. The overall number of elements is
240 660, and a nonlinear solver has been used, with 178 448 In order to test the model, it was decided to conduct a series
degrees of freedom. As can be noted in Fig. 5 the wind of preliminary experiments. Taking into account the system
approaching the building generates a pressure distribution modelled by the CFD tool, a laboratory-scale prototype (up
on the front façade. Therefore high wind speeds are induced to 600 mm high, 1000 mm wide and 1000 mm deep) was

175
 3

2.5

Voltage [V]
1.5

0.5

0
0 5 10 15 20 25 30
Days

Fig. 10 Voltage monitored by the wireless multimeter.

70
Fig. 9 Vertical axis centripetal turbine

Wind energy output (Wh/day)

60
mounted on a metallic structure.
50
constructed in fiberglass.
The closed impeller, sited at the centre of the fiberglass 40

structure, is a vertical axis centripetal turbine, with 7 blades 30

(see Fig. 9) mounted on a metallic structure supported by a
ball bearing. A brushless DC generator with electronic 20

commutation using permanent magnets offering speeds up 10

to 4000 rpm [9] is connected to the axis of the impeller (see
Fig. 4). Finally, a wireless multimeter is connected to the 0
0 5 10 15 20 25 30
Days
generator in order to remotely collect the measured data.
Fig. 11 Daily wind energy output obtained.
4.2. Experimental results

In the last month, the data of voltage generated by the registered, the system has produced a correspondent
prototype, at standard conditions of temperature and quantity of energy, without the needed of put the system in
atmospheric pressure, have been collected and analyzed. cut-off configuration. In fact it has been possible to exploit
The data are in time-series format, where the resolution of all the incoming energy because of the little centrifugal
the series is 1 min. Each day then contains a number N of forces that made sustainable the strong wind.
data points equal to 1440. In Fig. 10 the time diagram of the
voltage is shown. 5. CONCLUSIONS
The daily energy E produced by this laboratory-scale
prototype in a time period T (one day, in this preliminary The integration of wind energy converters into buildings is a
work), is calculated by the following discrete relation, challenging problem which is still largely unsolved. A wind
considering the monitored instantaneous power P: generation system that permits to exploit the wind energy in
N built up areas has been presented in this paper. In the
E = ∑ Pi ⋅ Δti (4) proposed eolic generation system the turbine is in-shore,
i =1
because it is contained inside a flow conveyor which
performs at the same time as stator of the turbine and as roof
where Δti is the sampling time (60 sec). of the building. This allows one to employ turbines which
The daily energy (eq. 4) has been divided by 3600 in cross-section is small with respect to the elaborated flow
order to covert it in Wh. In Fig 11 the daily wind energy tube, in this way saving installation costs and increasing the
output is reported. As can be noted, in all the monitored functioning range of admitted wind velocities. As is used to
days the system has produced at least a small quantity of do in urban areas, a vertical axis turbine is adopted, so that
energy, excepted the 14th and 15th when the system was out the wind direction does not affect the efficiency of the
of order for maintenance. This in accordance with the fact system, nor any reorienting apparatus is necessary.
that the cut-in velocity of the system is very low due to the The CFD models and the initial experiments on a
small inertia of the impeller and to the multiplication effect prototype have shown the potentialities of this new system
of the statoric structure. Furthermore, in the two days when offering many possibilities to increase its efficiency and the
a velocity much above the average value has been quantity of produced energy.

176
 6. ACKNOWLEDGEMENT

This work is supported by the operating program of Regione [4] D. Ayhan, Ş. Sağlam, “A technical review of building-mounted
Sardegna (European Social Fund 2007–2013), L.R.7/2007, wind power systems and a sample simulation model”, Renewable
and Sustainable Energy Reviews, vol. 16 (1), pp. 1040–1049, 2012.
“Promotion of scientific research and technological
innovation in Sardinia”. [5] A. Di Piazza, M.C. Di Piazza, A. Ragusa and G. Vitale,
“Statistical processing of wind speed data for energy forecast and
7. REFERENCES planning”, Proc. InternationalConference on Renewable Energy
and Power Quality 2010 (ICREPQ10), pp. 23-25, March 2010.
[1] J. F. Manwell, J. G. McGowan, A. L. Rogers, Wind Energy,
Theory, Design and Applications, Contract NAS2-11665, Muadyne [6] J.V Seguro, T.W. Lambert, “Modern estimation of the
Report 83-2-3, John Wiley and Sons, pp. 154-162, 2006. parameters of the Weibull wind speed distribution for wind energy
analysis”, J Wind Eng. Ind. Aerodyn, vol. 85(1): pp.75-84, 2000.
[2] J. Cace, E. ter Horst, K. Syngellakis, et al, “Urban wind
turbines: Guidelines for small wind turbines in the built [7] http://www.comsol.com
environment”, Wineur Intelligent Energy Europe, pp. 1-41, 2007.
[8] R. Temam, Navier-Stokes equations, theory and numerical
[3] A.S. Bahaj, L. Myers, P.A.B. James, “Urban energy generation: analysis, AMS-Chelsea Series, AMS, Providence, 2001.
influence of micro-wind turbine output on electricity consumption
in buildings”, Energy and Buildings, vol. 39 (2), pp. 154–165, [9] http://uk.rs-online.com/web/p/dc-motors/5366052/
2007.

177