Professional Documents
Culture Documents
July 2004
_______________________________________________________________________________________________________________________________________________________________________________________________________________
_
1
Prof. Dr. Christoph Holliger, University of Applied Sciences, Aargau, Klosterzelgstrasse 2,
CH-5210 Windisch, Switzerland. ch.holliger@fh-aargau.ch
2
Dr. Fritz Gassmann, Paul Scherrer Institute PSI, CH-5232 Villigen PSI, Switzerland.
gassmann@psi.ch
2
Table of Contents
4. Wind turbine 9
4.1 Basic considerations for layout 9
4.2 Wind tunnel measurements 10
4.3 Construction 11
4.4 Wing tests 14
4.4.1 Turbulence considerations 14
4.4.2 Mechanical test 15
4.5 Test in Lauffohr location 15
4.5.1 Test site and period 15
4.5.2 Results of measurements 17
4.6 Demo system in Cochabamba (Bolivia) 19
4.6.1 Transport and Preparations of TAUNUS 19
4.6.2 Construction site in Cochabamba 19
4.6.3 Construction of Demo system 19
Literature 39
3
The project began with a feasibility study. The project leaders visited La Paz and
Cochabamba from August 8 to 25, 2000 to find a project partner and to define the
system to be developed. Many contacts were realised with universities, administrative
national or international organisations and local companies. Among others, several
institutes of the Universidad Mayor de San Andres (U.M.S.A.) and the Escuela
Industrial Superior "Pedro Domingo Murillo" in La Paz and the Universidad Privada
Boliviana (UPB) in Cochabamba were visited. Realising that collaboration between
Bolivian universities and the industry is weak, several small and medium private
firms were contacted in La Paz as well as in Cochabamba. The most promising
potential partner was found in Taunus Electromecánica situated in La Paz and El
Alto. This company with over 100 workers produces small series or even single
4
pieces of special metal parts for very different customers in the mining and food
sector.
As interest in solar technology was found with Taunus, two students of UAS
visited the company in February/March 2001 and designed several types of low-cost
solar warm water collectors1 suitable for production in Taunus.
Realising that the interest for solar water heating systems (though being still an
important issue) was smaller than expected, the director of Taunus (Norman Saunero)
was invited to UAS in October 2001 for detailed discussions concerning the project
and the collaboration with the outcome of a strong common interest in wind energy
for pumping ground water for irrigation or domestic use.
Based on this result, a wind turbine was designed at UAS during the year 2002.
Several groups of students designed the following parts and their integration:
• blades on the basis of resin technology2
• blade support unit
• axis and turning mechanism
• storm security system
• tower
• calculation of general layout
During the first semester of 2003, the plans were finished and a prototype
constructed and built in UAS, partly supported by private firms which sponsored the
project.
On July 30, 2003, the system was assembled on a test site in Lauffohr near Brugg
(Switzerland) on a field where an existing hole for pumping ground water was
available (with a ground water table 3.7 m below ground level). It was tested during
one month on this site. The installed pump itself was not part of the project and quite
too small but well suited for demonstration purposes. It was sponsored by SKAT
Foundation (a resource centre promoting exchange of knowledge and experience in
development cooperation, St. Gallen, Switzerland). After disassembling, some parts
(including all 24 blades and the turning mechanism together with the axis) were
transported to La Paz by air-cargo to save time. Most metal parts were produced by
Taunus according to the students' designs.
Between September 17 and 25, 2003, the system was prepared at Taunus
laboratories, transported and then assembled near Cochabamba, Bolivia, on a
agricultural area with a groundwater-table about 5 m below ground level.
3.1 General
The climate of the Altiplano (cf. Fig. 1) is the result of synoptic pressure systems
over the Pacific and over the Amazon basin3: To the west, the Southeast Pacific
Anticyclone (High) produces dry and stable conditions by subsiding air masses; in the
east, in the interior of the continent, a thermal heat Low (Chaco Low) develops during
the summer leading to humid conditions. In addition, the Altiplano is orographically
influenced by its high elevated position (3500-4000 m ASL) between the western and
eastern Andean Cordilleres of Bolivia and finally, the El-Niño-Southern-Oscillation
5
System leads to modifications of the "normal" climate outside El-Niño events. The
investigated period 1999-2001 was about neutral with a tendency towards La-Niña.
The sum of all these influences results in a distinct seasonal pattern with the following
main characteristics:
• In austral summer (December-March), more than 80% of the annual
precipitation of 350-400 mm is delivered due to east winds bringing moist air
from the Chaco Low and due to strong convection over the high and intensively
irradiated (due to its elevation) plain.
• In austral winter (July-August), conditions are dry associated with strong
westerly flow over the entire region.
It is important to realize that over the Altiplano both, wind and precipitation regimes,
are linked to a continental scale and therefore show a seasonal cycle dominating over
local effects. This means especially for winds, that they will show roughly the same
characteristics over large portions of the Altiplano. However, winds have to be
expected being modified by local orographical details (channeling effects, effect of
shallow basins).
Fig.1: Orography of the Altiplano and the eastern Andean Cordillera of Bolivia
(horizontal: Western Longitude in deg., vertical: Southern Latitude in deg.,
gray-scale: Altitude ASL in m). The meteo station indicated as La Paz lies on
the Altiplano above the city. The orography is smoothed due to the 1km x 1km
grid and so, heights of mountain peaks are underestimated. Data from NOAA-
GLOBE (Global Land One-kilometer Base Elevation), May 1999.
6
3.2 Altiplano
The wind data for La Paz and Oruro could be purchased from NOAA4. All data
are hourly averages, i.e. gusts clearly exceed the given values for wind speeds.
La Paz: The meteo station lies in a flat region near the north-east border of the
Altiplano, where it opens towards the valley in which La Paz is situated. The
dominating orography is the eastern Andean Cordillera of Bolivia parallel to about
135 degrees. During only 15% of the time, the wind speed is below 1.5 m/s and the
overall average is 3.5 m/s. The recorded maximum was 23 m/s (average over 1 hour!).
The most frequent wind speed is around 2 m/s (during 27% of the time). During the
remaining 58% of the time, wind speeds are over 2.5 m/s and therefore suitable for
wind generators. The wind speeds are regularly distributed over the whole year:
Monthly averages are all between 3.1 and 3.8 m/s.
The wind regime in El Alto is produced by a continental scale process and
therefore can be assumed as being typical for large portions of the Altiplano. Winds
blow regularly the whole year over and wind directions are channeled by the
orography. This leads to two clearly defined main wind directions separated by 180
degrees. These characteristics make wind generators a successful means of producing
energy on the Altiplano.
Oruro: The meteo station Oruro is situated 10 km south-east of the town near the
lake Uru Uru (approx. 150 km2) at the foot of Cordillera Azanques, where the
direction of the eastern Andean Cordillera of Bolivia changes from about 135 to 180
degrees. Several nearby ridges can produce channeling effects and in the shallow
basin, "cold air lakes" can develop leading to stagnant situations (cf. Fig. 1). An
important characteristic of the wind regime in Oruro is the extremely high percentage
of time without wind: 57% of all measured values were recorded as wind speed 0.
However, as soon as there is wind, the speed is relatively high: After discarding all
hours with wind zero, the average wind speed of the remaining 43% of time is 5.4 m/s
and the component of wind speed into the direction 300 deg. is still 3.7 m/s. The
maximum wind speed recorded is nearly 31 m/s (average over 1 hour!) occurring
during one afternoon. The best winds are found in the afternoon (2-8 pm): During
67% of time, average is 5.8 m/s and 300 deg.-component is 4 m/s (i.e. only 33% zero
wind).
Oruro shows disadvantages for wind energy because of the high percentage of
time with zero winds (57%). In addition, prevalent wind directions are not separated
by 180 deg. and clear directions can be found only in summer. The advantage is, that
winds, if present, are frequently rather strong (> 4 m/s). Wind-driven pumps would
work especially in summer afternoons during more than 50% of the time (analysis
excludes 8 pm - 7 am because of lack of data!).
For a detailed description of wind regimes in La Paz and Oruro, see a special
report5.
7
3.3 Cochabamba
To estimate wind climatology, data from a meteo-station at the periphery south-
east of the centre of the town Cochabamba (CBBA) were analyzed6 (cf. Fig. 2). The
climate of CBBA results from its exceptional orographical location characterized by
surrounding high mountain chains. It is influenced by the same synoptic pressure
systems as mentioned above. They cause a distinct seasonal pattern with the following
main characteristics:
• In austral summer (November-March), about 80% of the annual precipitation of
500 mm is delivered due to east winds bringing moist air from the Chaco Low.
• In austral winter (May-August), conditions are dry associated with prevailing
westerly flow over the entire region.
providing a coupling of the valley air to the supraregional flow, causing maximum
wind speeds around sunset.
4. Wind turbine
4.1 Basic considerations for layout
Many wind turbines with different characteristics were developed for a variety of
purposes. The first step in the layout process for a new wind turbine consists therefore
of ensuring an optimal match between the physical properties of a turbine system to
be chosen and its application7. According to the aim being e.g. to produce electricity
or to drive a piston water pump, the choice of the wind turbine will be radically
different. With this choice, limits are set for the efficiency of the resulting system,
which can be quantified by the power coefficient normally symbolized by Cp, defined
as the maximum available power PM divided by the "ideal power" PB according to
Bernoulli's law:
PB = 12 r A u 3
PM
Cp =
PB
where r is the air density, A the area swept over by the blades and u the undisturbed
windspeed measured far away from the wind turbine. Fig. 5 shows power coefficients
† for different types of wind turbines.
Fig. 5: Cp is the power coefficient as defined above and tsr is the tip velocity to
windspeed ratio defined as the speed of the blade-tips divided by the wind
speed (measured far from the wind turbine). The indicated Betz limit is the
result of a experimentally validated engineering approximation (see:
http://windturbine-analysis.com/intro-betz.htm). The result is equivalent to a
transfer of 2/3 of the wind power to the wind turbine. A high speed propeller
has 1 to 3 blades and reaches efficiencies up to 45%. The chosen system
resembles most the American multiblade with a maximum efficiency of 30%.
10
To drive a water pump, a continuous flow of water even during periods with low wind
speeds is more important than high efficiency. As the turbine should be directly
mechanically coupled to a piston pump, a high torque at zero rotations per minute
(RPM) is essential to start the system as soon as enough wind is present. Because this
torque at 0 RPM is approximately proportional to the number of blades, a wind
turbine with 24 blades has been chosen, mostly resembling the American multiblade
type in Fig. 5. A start at about 2 m/s even with low air density as in El Alto (0.82
kg/m3) is possible with this design. As soon as the turbine begins to rotate, the torque
increases and the turbine is further accelerated until it finds its steady state frequency
(see also 4.7).
20
15
10
0
0 2 4 6 8 10 12 14 16 18 20
Rotations per Minute
v (m/s) =
Power measured PBernoulli =
Pmeasured=
45 Eta=
11
The maximum power according to Bernoulli's law for a wind turbine with a diameter
of 0.5 m for a wind speed of 4 m/s would be 7.54 W. As power scales with the square
of the length scale, ideal max. power for the 1:1 turbine would be 271 W, leading to a
power coefficient of 14% (calculated with 38 W, see maximum at 8 RPM in Fig. 6) at
a tip velocity to windspeed ratio (tsr) of 0.31. This result is in good agreement with
the American Multiblade characteristics given in Fig. 5.
4.3 Construction
This subchapter illustrates some characteristic data of the system and documents
components with photos and short explanations.
Fig. 8: The 24-blade turbine seen from the back hand side.
towards the foreground side. This is part of the safety system turning the turbine out
of the wind at high wind speeds. On the right hand side, perpendicular to the axis, the
0.8 m2 storm plate will be attached (see Fig. 10 for details of the safety system). On
the other side, a rod will be fixed to support the 20 kg weight which drives the system
into the operational position during normal wind speeds.
Fig. 10: Details of the safety system. During strong winds, the 0.8 m2 plate at
the left hand side (partly covered by the tree) turns the turbine by 90 degrees
(the storm plate then touches the rod supporting the main tail which remains
in wind direction all the time) lifting up the 20 kg weight seen in the
foreground. This weight opens the system back into the operational state as
soon as winds recede (see also Fig. 13).
13
c) Turning mechanism
Fig. 11 shows some of the 12 plastic rolls which enable the rotation into the wind
direction. This solution was chosen to avoid expensive bearings. A disadvantage is
the considerable torque (caused by friction) needed to turn the turntable, which with
axle, turbine and windvane weighs approx. 350 kg.
d) Tower
For strong average winds of 140 km/h (the max. 1 hour average of wind speed
found in the Oruro data record is 110 km/h), the turbulence spectrum was simulated
for both cases (using D = 100 m for the case with an obstacle). It follows from the
above relations that wake turbulence (b) is much stronger for high wind situations and
so is the important condition for layout leading to transient forces of 12'400 N on the
wind turbine5. This value was used for the layout of the tower.
15
Fig. 13: Wind turbine on the test site in Lauffohr in August 2003. The metal plate at
the right hand side is part of the storm safety system. During winds above
20!m/s, the drag excerted on this plate exceeds the weight of the cylinder
seen at the left hand side and the turbine is turned out of the wind. The
pumping rod transmitting force to the pump situated below ground water
level is clearly visible. The anemometer at about half the height of the tower
and the electronics box below it was used to measure performance of the
turbine (the farmer's pumping equipment at the lower right hand corner is not
part of the test).
17
Fig. 14: Pump frequency in rotations per minute (RPM) and wind speed (m/s) for
measurement 4 from August 23, 2003 14:50-18:50. During the first 7
minutes, the turbine was directed perpendicular to wind direction and
therefore did not rotate. As soon as wind speed exceeded 3 m/s, the system
turned into the wind and began to work. At the end of the record, the turbine
is turning with about 4 RPM during an average wind speed slightly below 1
m/s, demonstrating the very low friction losses.
19
Cochabamba
Road to
Oruro-La Paz
Fig. 17: Cochabamba is situated in a large basin surrounded by high montains. Due to
this geographical pecularity, inversion layers formed through the night
prevent winds until the afternoon. Rather strong winds are regularly observed
between 3 and 10 pm (see Fig. 3).
21
5.3 Recommendations
With the aim to produce and successfully sell such wind-driven pumping systems
in Bolivia, a rigorous cost estimate for a somewhat reduced system (e.g. cables
instead of balancing profiles for the tower, blade thickness reduced to 70%) is of
central importance. Furthermore, a three-blade system should be purchased and
installed on the same test site. A cost-efficiency comparison between the two systems
should answer the question related to the general type of the turbine. Finally, a market
analysis must answer the question of the size of the production run and which system
of what size could be marketed in which region in Bolivia and/or country in Latin
America.
22
APPENDIX
Construction Plans
A1) Support Unit for Blades
A1.0) Overview
here to be
bent upwards
Wind
Direction
24
A1.3) U-Profiles
Literature
1
P. Willi & P.P. de Tommaso, Nachhaltige Energietechnologie-Entwicklung in
Bolivien, Thermische Solarenergie zur Warmwasserbereitung, Schlussbericht,
28.3.2001.
2
D. Kohli et al., Projekt Windradflügel für Bolivien, FHA, 30.10.2002.
3
M. Vuille: Atmospheric circulation over the Bolivian Altiplano during dry and wet
periods and extreme phases of the Southern Oscillation, Int. J. Climatol. 19,
1999, 1579-1600.
4
National Climatic Data Center (NCDC, Asheville, USA). The list of all available
stations and years can be found at:
http://lwf.ncdc.noaa.gov/oa/climate/surfaceinventories.html
5
F. Gassmann & K. Schöllhorn, F&E-Projekt nachhaltige Energietechnologie,
Zusammenarbeit FHA - Bolivien, Technical Report: Meteorological Situation
on the Bolivian Altiplano, Wind and Turbulence, FHA, 25.10.2002.
6
F. Gassmann, F&E-Projekt nachhaltige Energietechnologie, Zusammenarbeit FHA -
Bolivien, Technical Report: Meteorological Situation in Cochabamba (Bolivia),
Wind frequency, FHA, 12.11.2002.
7
P. Fraenkel et al., Windpumps – A guide for developing workers, Intermediate
Technology Publications, Stockholm Environment Institute, 1993.
8
C. Bopp, Der Wind weht, das Wasser fliesst, Aargauer Zeitung, Dienstag, 5. 8.
2003, S. 11.
9
K. Ruch, Windrad hat den Test bestanden, Aargauer Zeitung, Freitag, 29. 8. 2003,
Regionalteil Brugg-Windisch.