HomeMy WebLinkAboutPower Predictor1.0 Evaluation RichStromberg 04-2011-APower Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 1
An Evaluation of the Power Predictor 1.0™
A low-cost anemometer, vane and pyranometer used to determine energy potential at a given site
Rich Stromberg
Alaska Energy Authority, April 2011
Acknowledgments: Many thanks to Bruce Wright and Aleutian Pribilof Islands Association for loaning a Power Predictor 1.0™ for this study.
Thanks to James Jensen, Kat Keith, Doug Vaught and Bruce Wright for reviewing the technical content of this report.
Abstract: The Alaska Energy Authority has completed a six-month evaluation of a Power Predictor 1.0™
in near side-by-side testing with a Second Wind anemometer and NRG Symphonie™ datalogger. The
purpose of the evaluation was to determine if this ~$300 product could serve as an alternative to the
Authority’s standard meteorological configuration. The results show that the Power Predictor 1.0™ may
be a viable alternative during the reconnaissance stage of wind energy exploration and site
development and serve as a useful tool to evaluate many potential sites in a fast and economical
fashion. The performance of the unit’s pyranometer was not deemed acceptable, but a future
evaluation of the Power Predictor 2.0™ with upgraded polysilicon PV sensor is warranted.
Figure 1 – Power Predictor 1.0 vane Figure 2 – Datalogger
and anemometer
Background: The Power Predictor 1.0 unit tested is actually the third system used here in the state by
Alaska Energy Authority. The unit tested was graciously loaned to the Authority by Bruce Wright, senior
scientist with Aleutian Pribilof Islands Association. The first unit was installed on a 7-meter portable PVC
tower above the village of Tatitlek in eastern Prince William Sound. This unit failed early due to water in
the datalogger. It is possible that the water-tight seal popped open or was knocked open. No physical
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 2
damage can be seen on the unit to explain the leak. Testing back at AEA indicated no pinholes in the
seal. The second unit was wrapped in plastic and electrical tape with no moisture inside the datalogger.
The third unit tested at the AEA warehouse and the AEA roof had electrical tape wrapped around the
seal with no problems occurring.
The goal in evaluating this unit is to find a low-cost anemometer that can be deployed by one or two
people in the field in locations that are difficult to access with pickups, four-wheelers or other off-road
equipment. Data collection from multiple potential wind sites early in the reconnaissance phase of a
project could allow for better siting of our standard 30/34m NRG Tall Tower™ system. The Authority’s
standard configuration costs $8,160 at the time of publication, weighs 1240 pounds and takes a crew of
5+ to erect utilizing robust anchors and a winch. A portable tower used for wind-site prospecting,
however, could be carried in and erected by two people and cost around $800, including anchors, PVC
pipe, fittings and guy wires.
General product and purchase information can be found at http://www2.powerpredictor.com/
Literature search: All online information that can be found regarding the Power Predictor™ online is
promotional sales information or testimonials. Product specifications and manuals can be found online,
but no independent testing or comparisons with other anemometers are present.
First AEA set up: A tower was installed in the afternoon of Friday, Oct. 24 along the north fence line of
the AEA warehouse at 2601 Commercial Drive in Anchorage. The tower consisted of two 2” ID PVC pipe
sections (10’ long) with 1.5” ID sleeves (2’ long) connected with 3” 5/16 dia. Galvanized bolts, washers
and nuts. (Figure 3) A ½” crescent wrench was used to tighten bolts. These materials were purchased by
the author due to delays in reinstating the Authority’s wind assessment fund. An Air-X tower kit,
including t-base, top collar and guy wires, was not purchased for this reason, but would be part of a
standard test configuration.
Figure 3 – Pipe section detail
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 3
The Power Predictor™ was mounted onto the
top section of PVC pipe with two 3”-long ¼”-
dia. galvanized bolts, plus nuts and washers.
The Second Wind anemometer was mounted
on a galvanized bolt extending from the top of
the PVC pipe. A PVC cap was placed over the
top of the PVC to prevent birds trying to nest
and getting stuck inside. (Figure 4)
The Power Predictor™ datalogger was taped to
the PVC pole around chest height with the
NRG Symphonie™ datalogger sat in a shelter
box at the tower base.
An attempt was made to attach the 20’ tower
to the fence, but the lack of a top collar
anchored to the ground resulted in too much
flex in strong winds and the tower was
reduced to 10’. (Figure 5)
Figure 4 – Top of pole configuration
The purpose of the test was not to assess the wind resource at the
warehouse, but to place the two sensors as close to each other as
possible to determine how closely they track during wind events.
Under the test configuration, the two anemometers were
approximately 1.5 feet apart. The fact that the Second Wind
anemometer is placed higher and extends from the top of the pole,
while the Power Predictor™ is side mounted introduces variation due
to shadowing of the latter from the 2” ID PVC pipe. The Power
Predictor™ cup diameter is approximately 5 inches versus 7.5 inches
diameter for the Second Wind sensor.
An initial inspection showed that instantaneous readings on the two
systems varied in gusty conditions, but steady state looked fairly
close. Both loggers will record 10-minute data, their clocks synched
up. The recorded time period, however, did not match up with the
NRG recording at :00, :10, :20, etc. and the Power Predictor™
recording at :06, :16, :26, etc. It is not clear how to correct this offset. Figure 5 – 20’ pole assembly
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 4
Sensor specs:
Power Predictor™ 3-cup pulse anemometer with data logger
• Wind speeds accurate to +/- 3% in independently verified tests
• 10 second recording interval, average taken every 10 minutes
• Wind vane for measuring prevailing wind direction and turbulence (requires due south
orientation)
• Solar sensor for measuring solar irradiation (also requires due south orientation)
• Self-contained waterproof data logger engineered for low power operation
• LCD screen provides live wind speed and solar indication
• 512 MB SD memory card included with USB adaptor
• Power Predictor Web application works through your browser - works with PC or Mac
• Requires USB 1.0 or 2.0 port on your computer
• Minimum 30 days data required before Power Report generated
• PP3 9V battery supplied
• 5m cable (6m extension available)
• Rugged UV resistant ABS plastic
• RoHS compliant and fully CE certified
http://www.bettergeneration.com/power-predictor/introducing-the-power-predictor-
anemometer.html
Second Wind C3™ Conical cups measure 51 mm (2 inches) in diameter
• Rotor diameter is 190 mm (7.5 inches)
• Standard AC output, frequency proportional to cup rotational speed
• Shielded AC pickup coil, 4100 turns of #41 wire
• Four-pole Indox 1 magnet rotates with the cup assembly
• Fully hardened beryllium-copper shaft running in self-lubricating modified Teflon bearings,
with protective boot to make the system dirt and water resistant
• Rated bearing PV (pressure-velocity) factor is 20,000
– At 15 mph PV is approx. 500.
– At 100 mph PV is approx. 2,000.
• Rotor assembly moment of inertia = 68 x 10–6 S-ft2 (or 92.2 x 10–6 kg-m2)
• Distance constant = 10 feet (3.0 meters)
• Transfer Function: m/s = (Hz x 0.766) + 0.324 *miles per hour = (Hz x 1.714) + 0.725+
• Accuracy: within 0.1 m/s (0.2 mph) for the range 5 m/s to 25 m/s (11 mph to 55 mph). This
equates to max 2% error.
http://www.secondwind.com/Sensors/C3-anemometer.html
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 5
Results from AEA Warehouse Site: After one month, data for the two sensors was compared. Even
though there had been some severe wind events in Anchorage, the highest 10-minute average seen
during the month was 6.5 meters per second. Prevailing winds are likely blocked by nearby buildings and
vegetation even though the immediate several hundred feet around the site is clear. At the end of the
first month, it was noticed that the anemometer channel on the NRG datalogger did not have a .350
offset value as is apparent in the presence of values less than .4 m/s. (Figure 6)
Figure 6 – Trend Chart from AEA Warehouse Site
This offset value was corrected on Oct. 25. The data trend shows several key observations.
Null readings on the NRG/Second Wind sensor default to 0.4m/s while the Power Predictor™
will read zero unless the wind speed is at least 0.4 m/s at which point the anemometer begins to
spin.
Both sensors move with each other – each wind peak can be detected by the Second Wind and
the Power Predictor™ anemometers. The correlation coefficient (Excel™ CORREL procedure) is
0.929 for all data pairs and .952 for all data after the NRG offset value was corrected.
With one low-speed spike as an exception, the Second Wind anemometer always peaks higher
than the Power Predictor by ~ 0.5 m/s.
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 6
A high-wind event was analyzed to get better visual resolution on how well the two sensors track each
other. (Figure 7) The two sensors appear to track very closely.
Figure 7 – “High” wind event – AEA warehouse location
Probability plots for the two datasets show the Second Wind anemometer to be consistently higher
reading than the Power Predictor. The smoothness of the curves depends on which data column was
sorted. (Figures 8 and 9)
Figure 8 – Prob plot NRG sort Figure 9 – Prob plot PwrPred sort
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 7
To test the theory that the 2”-dia. PVC pipe is blocking some of the wind, radial graphs were printed.
The Power Predictor™ has to be mounted on the south side of the pole to provide accurate wind
direction values and solar irradiance values. Thus, one would expect a 2” pipe approximately one foot
from the anemometer to interfere with wind readings from the north and possibly other directions. A
radial graph of average wind speeds by direction show very low readings coming from the north (Figure
10) and possible interaction with winds out of the south due to a pressure increase on the windward
side of the PVC pipe.
Figure 10 – Radial graph of average wind speed by wind direction – Warehouse site
A similar radial graph, but plotting the maximum observed wind speed for a compass direction exhibits
an even stronger effect of lower wind readings coming from the north (pole blocking) and south (pole
pressure buildup) sides of the test configuration. (Figure 11)
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 8
Figure 11 – Radial graph of maximum wind speed by wind direction – Warehouse site
Because no higher wind events were seen at the warehouse site, the test tower was moved to the roof
of the Alaska Energy Authority building located at 813 W. Northern Lights Blvd in Anchorage in early
December. This new site also allowed for correlation of solar irradiance data since a Li-Cor LI-200SZ
pyranometer had been collecting data on the AEA roof since late February 2010.
Results from AEA Main Office Site: After four months, data for the two sensors was compared again.
This time there had been four periods where the 10-minute average peaked out around 7.5 meters per
second. Due to the placement of the test tower, there was interaction with part of the building roof and
the Second Wind/NRG instruments would have been less impacted by the roof than the Power
Predictor™. As was seen before, both units track well with high-wind events, but the Power Predictor™
never reads as high for wind speed. (Figure 12)
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 9
Figure 12 – Trend Chart from AEA Main Office Roof Site
The correlation coefficient (Excel™ CORREL procedure) is 0.9002 for all data pairs and the delta between
the average of each data set is .255 meters per second with the Second Wind/NRG configuration having
the higher average. Median delta is only .1 meters per second.
As before, probability plots for the two datasets show the Second Wind anemometer to have
consistently higher readings than the Power Predictor™. The smoothness of the curves depends on
which data column was sorted. (Figure 13) This time, the delta between the two moving-average curves
is higher – 1 meter per second of higher. Radial graphs of the average and maximum speeds (Figures 14
and 15) indicate additional shadowing from the nearby roof on the east side of the test configuration.
The Power Predictor’s lower position on the test pole would make it more susceptible to interference
from this sloping roof. After seeing this impact, it is recommended that the AEA office roof site not be
used in future studies. If higher wind regimes are needed, the author has a cabin east of Chickaloon with
some exposed hills. Data could be collected each weekend. Also, technical reviewer Bruce Wright has
property north of Palmer that is being used to test and compare two Power Predictors.
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 10
Figure 13 – Probability plot (sorted by NRG value)
Figure 14 - Radial graph of average wind speed by wind direction – AEA roof site
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 11
Figure 15 - Radial graph of maximum wind speed by wind direction – AEA roof site
Comparison of solar irradiance sensors: With the test pole moved to the AEA roof, a near side-by-side
comparison could be made between the Power Predictor 1.0™ solar sensor which has a vertical, south-
facing placement and the LiCor-200 pyranometer that sits at a latitude tilt, south-facing. The Power
Predictor 1.0™ sensor was not blocked by any building components or by any part of the test
configuration. This pole was anchored within 50 feet of AEA’s pyranometer, which has been collecting
solar irradiance data since late Feb. 2010.
Detailed specs for the Power Predictor 1.0™ solar sensor are not available. Specs for the Li-Cor LI-200SZ
are in Figure 16. This model is one of the standards used by the National Renewable Energy Laboratory
to measure solar irradiance.
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 12
Figure 16 – Pyranometer specifications
Results from Power Predictor 1.0 solar sensor are severely inadequate. The trend graph in Figure 17
clearly shows that the Power Predictor™ is binning to discrete values while the Li-Cor 200 outputs
continuous data values. The 255-point moving average trend lines appear to move together and the
correlation coefficient (Excel™ CORREL procedure) is a surprising .8567. The actual values output from
the solar sensor, however, do not allow for any detailed analysis. The average solar irradiance for the
study period from 12/3/10 to 4/6/11 is 92.50 watts per square meter for the Li-200 and 60.89 watts per
square meter for the Power Predictor 1.0™. The Power Predictor1.0™ is also limited by its fixed vertical
mount.
It should be noted that the Power Predictor 2.0™ has incorporated a completely different style of solar
sensor – a photovoltaic cell mounted at a near horizontal orientation.
Power Predictor 1.0 Evaluation Alaska Energy Authority P a g e | 13
Figure 17 – Pyranometer vs. solar sensor comparison
Conclusions/Recommendations: Data from the evaluation was reviewed by Douglas Vaught,
professional engineer with V3 Energy, LLC, Bruce Wright senior scientist with Aleutian Pribilof Islands
Association and James Jensen, wind program manager with Alaska Energy Authority. The consensus was
that the Power Predictor™ could be used for low-cost reconnaissance work throughout Alaska to
identify sites for possible wind energy development. Based on this study, we would assume the Power
Predictor™ underestimates the average wind speed by 0.5 meters per second.
Additional wind data collection with a 34-meter NRG Tall Tower™ would be required before any large-
scale development were to take place, but the Power Predictor™ data may suffice in locations where
small turbines (< 10-kilowatt rated capacity) were to be installed. The Power Predictor™ could also be
used in a fleet of as many as five sensors in the vicinity of a proposed wind farm to identify the best
location to set up a large meteorological tower for longer-term data collection.
Follow-on studies are recommended to assess the solar photovoltaic sensor in the 2.0 model as well as
to compare multiple units side-by-side to verify data consistency from one unit to the next.
In the future, the AEA office roof will only be used to compare solar irradiance sensors. Further wind
studies will be conducted at AEA’s warehouse site or potential rural locations in the Matanuska Valley.