Northfield Schools Wind Monitoring Data Summary

The following data analysis and summary was prepared by RENew Northfield Board Member, and electrical engineer, Chris Ludewig.

Background:

Wind Monitoring for the Northfield School District began on November 8, 2001, with the installation of a temporary 40-meter high meteorological tower by the Minnesota State Departmet of Commerce Energy Office on the site of the future Northfield Middle School. Data was collected from this site for what ultimately became a full year, and compared with data from the same time period collected from a long term wind monitoring site near Nerstrand which was considered to have adequate wind for a turbine.

Analysis of the first several months of data from the Middle School Site showed a lower-than-hoped-for wind speed, which was averaging only 81 % of the wind speed from the same time period from the Nerstrand site (at a 50-meter comparison height). In that same time period, a WaSP computer analysis of the Northfield area was done by the State Energy Office, resulting in a higher resolution relative wind power map specific to the Northfield vicinity. This map suggested better potential locations for wind turbines in nearby rural areas surrounding Northfield.

As a result, a recommendation was made to the School Board in April of 2002 to further investigate these potential “off-campus” sites. A Wind Energy Task Force was subsequently formed at the direction of Superintendent Terry Tofte, and began meeting at the end of May, 2002. Carleton College also joined the discussions at that time expressing an interest in partnering with the School District in its efforts.

After a more detailed investigation–taking into account the potential wind resource available, proximity to existing 3-phase power lines, and the willingness of the landowners to host the installation of a temporary met tower–two new rural sites were chosen for further wind monitoring. The State Energy Office tower was relocated in November of 2002 to the Carlyle Aldrich farm in the service territory of Great River Energy, and a recently purchased RENew Northfield tower was erected on a ridge on the Paul Lindenfelser farm in the service territory of Xcel Energy. A summary of the results of 5-plus months of monitoring on these two sites follows below.

Site Details

Relative Wind Power

The four sites mentioned above, Middle School, Lindenfelser, Aldrich, and Nerstrand, are shown located on the Relative Wind Power map and the accompanying Topographic map.

The Aldrich Site is located about three miles SSE of Northfield at an elevation of approximately 1,150 feet on a broad ridge along the north of the Prairie Creek area.

The site is approximately 6.25 miles WNW of the Nerstrand reference site and is approximately 1.5 miles from an existing Steele Waseca Coop. substation. This site was monitored with a 40-meter tall tower supplied by the State Energy Office. The tower was equipped with anemometers at 10 meters, 30 meters, and 40 meters and had wind vanes at 30 meters and 40 meter heights. Data was taken as 10 minute averages, was collected by the state, and provided for this analysis in an Xcel format. This site was monitored for a total of 156 days from November 6, 2002, to April 10, 2003.

The Lindenfelser site is located about two miles east of Northfield at an elevation of 1,050 feet on a somewhat narrower ridge just to the south of State Hwy 19. This site is approximately 8.25 miles NNW of the Nerstrand reference site, and is located about 1.5 miles east of the end of an existing Xcel Energy 3 phase distribution line. This site was monitored with a 40-meter tall tower supplied by RENew Northfield. The tower was equipped with anemometers at 20 meters, 30 meters, and 40 meters, and had wind vanes at 30 meter and 40 meter heights. Data was taken as10 minute averages, collected via memory card and imported into Xcel for this analysis. This site was monitored for a total of 155 days from November 9, 2002, to April 12, 2003. The Nerstrand reference site is located on a communications tower approximately 1.5 miles NE of Nerstrand at an elevation of 1,180 feet. This site has been continually monitoredsince June,1995, by the State Energy Office. The tower is equipped with anemometers at 30 meters, 50 meters, and 70 meters and wind vanes at 30-meters and 70-meter heights. Data is collected as one hour averages and is maintained in a database by the State Energy Office.

The Middle School Site is located on the southern edge of Northfield at an elevation of 1,000 feet. It was monitored by the same 40-meter tower supplied by the Energy Office for the Aldrich site. The site was monitored from November, 2001, to November, 2002.

Data Collection and Analysis

Data from the sites has been collected monthly and compared hour by hour to determine the speed at the two temporary sites relative to the Nerstrand reference site. All analysis has been done using an Xcel spreadsheet.

The effect of tower shading has been taken into effect for all sites. The Nerstrand site has two anemometers at all heights and shading effects are removed by selecting the appropriate anemometer depending on the wind direction. On the temporary towers where only one anemometer was available at some of the measurement heights data is filtered out when it is determined that shading is affecting the measurement.

In order to be able to compare wind speeds from the temporary sites to those from Nerstrand at the 50- and 70-meter heights (which are higher than the temporary towers), the wind speeds from the temporary towers have been extrapolated to heights of 50 meters and 70 meters using the wind shear formula. The wind shear coefficient used in the formula is first calculated using the wind speed measurements from the lowest and highest anemometer on each tower. The measured data from the 40-meter height at the top of the temporary towers is used as the base number in the equation and extrapolated to the taller heights using the wind shear coefficient previously calculated. Wind shear is not calculated for wind speeds less than seven mph due to the high variability in the wind averages at those low speeds.

Wind power density, a measure of how much energy is available in the wind, has also been calculated. The power in the wind is related to the air density and the cube of the velocity of the wind. Air density is determined using the measured temperatures at the sites and the altitude of the sites.

The formula used to calculate wind power density comes from the MN State Energy Office WRAP report.

In order to allow direct comparison to the Nerstrand data, the Aldrich and Lindenfelser data is converted to one-hour averages to match the Nerstrand data format.

For those interested in more detail on wind monitoring techniques and data analysis, including the equations used to calculate wind shear, extrapolate wind speeds, and calculate wind power density, see the 2002 Wrap Report from the MN State Energy Office here >>

An excellent wind energy educational website, created by the Danish Wind Industry, can be found at:www.windpower.org

Wind Monitoring Data Results
Here are the measured averages for the sites monitored
from Nov. 9, 2002, to April 10, 2003:

Wind Speeds (mph)

Table: Measured averages for wind measurement sites.

*red values are calculated based on
wind shear

Wind Shear Coefficient

Table: Wind shear coefficient.

Wind Energy Density (watts per square meter)

Table: Wind energy density.

Figure 3: Hour by hour
wind speed comparison for December showing significant correlation between
sites.

Graph: Hour by hour wind speed comparison.

Figure 4: Wind speed distribution
plot. Note that a turbine would produce power for all speeds greater than
3 meters per second.

Graph: Wind speed distribution plot.

Figure 5: Wind Rose (directional distribution
of wind). Note the similar shape of the pattern, although it is rotated
somewhat from one site to the next.

Graph: Wind rose.

Figure
6: Plot of the ratio of the hourly average windspeed from the Lindenfelser
site versus the Nerstrand site. It shows the relative speed of the wind
at the Lindenfelser site averaging out to approximately 106 % of the Nerstrand
speed after only about ten days of monitoring. This is very close to the
final value that was eventually attained at the end of the entire 5-plus
months of monitoring. This indicates that a site can be characterized relative
to the Nerstrand site within a fairly short time period.

Graph: Plot of ratio of the hourly average windspeed from the Lindenfelser site versus the Nerstrand site.

Final Site Selection

Based on the data collected, it is apparent that the Lindenfelser site has significantly better wind. In addition, Xcel Energy is required to pay slightly better rates for the energy produced than Great River Energy and the power grid interconnection has proven to be somewhat easier and less expensive in the Lindenfelser area. For all of these reasons, the ridge that the Lindenfelser monitoring site was located on has been determined to be the best location for the potential School District and Carleton turbine projects. A wind lease option has been signed with one of the landowners in the area and a preliminary siting for the turbine locations has been completed (as shown in Figure 2 above).

Turbine Energy Production Estimates

For the purposes of estimating actual wind turbine power
production, the seven-plus years of historical wind data from the Nerstrand
site will be used, adjusted in an appropriate manner that is based on the
measurements from the Lindenfelser site relative to the Nerstrand site.
Because wind power density is a better indicator of turbine performance
than average wind speed, the 70 meter Nerstrand data will be adjusted upward
by the amount needed to result in 13% more energy in the wind, as calculated
for the Lindenfelser site. This is accomplished by increasing each hourly
average wind speed value by 4.2% prior to calculating turbine output. In
addition, these modified 70-meter wind speed averages have been adjusted
to obtain a value for each of the specific hub heights of the various turbines
modeled.

In order to estimate turbine energy production, regression analysis has
been used to curve fit a 9th order polynomial to the wind turbine power
curves. Once the equations are determined, an hour-by-hour turbine production
estimate can be calculated for each turbine from the adjusted Nerstrand
wind data. From this data it is then possible to estimate average yearly
energy generation and turbine capacity factors. In addition, the time varying
nature of the energy produced by a potential turbine can be observed.

Bids were received from three turbine manufacturers for a total of seven
different turbines. Production estimates for each are shown below. It should
be noted that these are estimates only. There is some range of uncertainty
in these numbers that is unavoidable. For purposes of the financial pro
forma analysis the developer has reduced these estimates by 20% to take
into account any uncertainties in the wind data and analysis, as well as
other factors that can result in less than 100% turbine availability.

Based on a combination of total project cost and power output estimates,
the NEG Micon NM82-1650 is the turbine that has been chosen from the bidding
process.

Turbine Production Estimates

Table: Turbine production estimates.


Variations In Revenue:

Figure 7 below is a plot depicting how gross
revenue from a turbine might change over time due to the variability of
the wind. The plot was created using seven years of actual data from the
Nerstrand site.

On a monthly basis over that 7 year period, revenue would have varied
from a low of 45 %, to a high of 152 % of nominal.

On a quarterly basis revenue would have varied from a low of 70 % to a
high of 127 %.
The seasonal minimum numbers typically occur in the summer, and the maximums
in the winter.

On a year by year basis, the revenue would have varied from a low
of 86 %, and a high of 109 % of nominal.

Figure 7: How gross revenue might change
over time due to wind variability. Create from seven years of actual data
from the Nerstrand site.

Graph: How gross revenue might change over time due to wind variability.

Conclusions

Reasonable correlation is seen in the wind patterns between
the temporary monitoring sites and the Nerstrand site. Therefore, the
seven-plus years of historical data from the Nerstrand site is valuable
for estimating Turbine output.

· The Lindenfelser site has the best wind resource, with slightly
more energy in the wind than at Nerstrand, and significantly more energy
in the wind than at the Aldrich site. The Lindenfelser ridge is also more
desirable because Xcel will pay a little more per kWh, and the grid interconnect
is likely to be easier and less expensive than at the Aldrich site. The
actual turbine sites are to be located about ¾ mile to the SW of
the Lindenfelser site on the same ridge and results are expected to be
similar.

· The NEG Micon NM82-1650 1.65MW turbine is the overall winner
in the bidding process based on its higher production numbers in this
wind resource relative to the total project cost for this turbine.

· The financial pro forma uses energy production numbers 20
% lower than what has been calculated to provide margin for error. This
should be adequate to account for any measurement errors or deviations
from predictions in the wind shear calculations. No formal analysis of
the possible range of error in the analysis has been done, however,

· Turbine gross revenue will likely vary by +/- 15 % on a yearly
basis based on the historical record from the Nerstrand site.