Wind Characteristics - Wind Energy: Renewable Energy and the Environment - page 45

In general, whatever the wind speed is at any point in time, over the next hour the behavior ought
to be similar. This is called persistence: v ( t t 0 ) ~ v ( t 0 ), where t is variable. However, a histogram
does not give a time sequence of data, nor does a wind speed-duration curve tell the length of calm
periods. As more wind turbines are installed, wind farm operators and utilities will be interested in
predicting wind speeds, average variation by season and time of day, duration of low wind speeds,
and values for the next 1 to 36 h.

3.10 VARIATIONS IN WIND POWER POTENTIAL
Since the motion of the atmosphere varies on a scale from seconds to years, wind power and
wind energy will also vary on the same time scale. The annual average wind power (6 m height)
for Amarillo, Texas, was 220 watts/m 2 for the period 1962-1977 [6]; however, the variation from
one year to the next can be quite large. A minimum of 2 years of data are needed to obtain an
estimate for the annual wind power potential, and 5 years of data are needed to obtain a mean
value within 6% of the long-term mean. Most people assume that if you have 2-3 years of data,
then that will suffice, along with longer-term regional data for comparison, to determine the wind
power potential. The annual wind power potential (Figure 3.14) for White Deer and Dalhart,
Texas, shows the correlation between sites, which are 140 km apart in the same region. Data
were sampled at 1 Hz and averaged over 1 h. Therefore, for a region where long-term base data
are available for comparison, 1-2 years of data would suffice for determining the wind power
potential at a specific location.
The seasonal variation for most of the United States is high wind speeds in the spring, with low
wind speeds in the summer ( Figure 3.15 ) . Notice the standard deviations at 10 and 50 m are com-
parable, and the average value for both is 0.6 m/s. Also, the standard deviation of the wind speed
by month is close to the same as the standard deviation of wind speeds for an individual month
(744 data points). The most notable exception to general seasonal variation is the mountain passes
in California between the coast and inland deserts. The windy season corresponds to heating of
the deserts in the summer, where the hot air rises and is replaced with cooler air flowing in from
the ocean.
There are also variations with the movement of synoptic weather patterns, which is represented
by a 4- to 5-day variation. The diurnal (daily) variation is due to heating during the day. These fre-
quency representations ( Figure 3.16) are common to many locations [7]. The peak at 0.01 cycle/h
corresponds to a period of 100 h, which is the 4- to 5-day variation, and the peak near 0.1 cycle/h
corresponds to the diurnal variation.

600

White deer Dalhart

500

400

300

96 97 98 99 00 01 02 03 04 05 06

FIGURE 3.14 Annual variation of wind power potential at 50 m for White Deer and Dalhart, Texas.