Page 110 - Fluid Power Engineering

88 Chapter Six

lowest height anemometer starts increasing because of higher turbu-
lence caused by surface roughness and changes in contour. Therefore,
there is a balance between increasing the distance between top and
lowest height anemometers to get a better estimate of shear and de-
creasing the accuracy the lowest height anemometer. As a guideline,
the lowest anemometer height may be set to 30 m for a 60 m met-tower.
The overall error in estimation of wind speed in a measurement
campaign can be computed using the independence assumption,
which assumes that each source of uncertainty is independent of their
other. Under this assumption, the overall error is:

N

ε = ε 2 (6-4)
i
i=1
where ε is the overall error, ε i is the error because of each component
of uncertainty, and N is the number of components of uncertainty.

Example of Error Estimate
Consider NRG #40 calibrated anemometers installed at 60, 40, and
30 m in a complex terrain with maximum flow angle of 6 . What is
◦
the uncertainty associated with estimating wind speed at 80 m?
The purpose of this illustrative example is to expose the reader to
the lack of complete data to compute uncertainties. Several assump-
tions based on judgment are made below regarding the source and
amount of uncertainties:

Anemometer accuracy in complex terrain. According to

Table 6-2, uncertainty with 3 of inflow angle is 2.4% and up
◦
◦
to 15 is 8.3%. As an approximation, a linear relationship will
yield an uncertainty of 3.9%.
Calibration, dynamic overspeeding, and inflow angle are in-

cluded in the anemometer accuracy above. Tower shadow,
boom, and mounting effects are not included and are inde-
pendent of anemometer accuracy. Since information about
predominant wind direction is not available, the nature of bias
(positive or negative) because of tower shadow is not known.
Errors of 1.5% because of tower shadow and 0.5% because of
boom and mounting effects will be assumed.
Error because of shear can be estimated by examining Figs.

6-10 and 6-11. Although complete data is not available, the
error is likely to be in the range of 4.7 and 2.4%. A 4.7% error
is for a case with 50-m tower and Class 2 anemometer; 2.4% is
with 60-m tower and Class 1.5 anemometer; this example is
of a 60-m tower and Class 2.4 anemometer. In this example,