UNCERTAINTY IN WIND RESOURCE ASSESSMENT - Wind Resource Assessment: A Practical Guide to Developing a Wind Project - page 225

• At some sites, the shear may change with height above the masts in a consistent
way, resulting in a systematic bias in the projected hub-height wind speed.

Each component of the uncertainty for a mast should be combined with the same
component for other masts in either a weighted linear sum or a squared sum depending
on whether the component is correlated or uncorrelated. The weight given to each
mast should be proportional to its influence on the overall array-average wind speed.
One approach to determining the weights is to divide the array into groups of turbines,
each associated with a particular nearby mast. This leads to the following equations:

i = 1 N i σ

i 1 / 2

Uncorrelated σ combined = (15.7)

N T

i = 1 N i σ i
N T

Correlated σ combined = (15.8)

Here, the sums are over M , the number of masts; N i is the number of turbines

associated with mast i ; σ i is the percent uncertainty in the average speed for that group

of turbines; and N T is the total number of turbines in the array. (Where some form of
smooth blending of the predicted speeds from different masts is applied, the uncertainty
equations naturally become more complicated, but the same general principles hold.)
Equation 15.7 implies that the uncorrelated portion of the uncertainty decreases the
more evenly distributed the masts are among the turbines. The effect is demonstrated
in Figure 15-2 for the case of two masts, each having an uncertainty of 10%. If
one mast is the predominant influence for half the turbines and the other mast is the
predominant influence for the other half, the combined uncertainty is reduced to 7.1%.
In fact, assuming the individual uncertainties are the same for all masts (call it

σ 0 ) and
an equal number of turbines are associated with each mast, Equation 15.7 reduces to

σ 0
√ M

σ combined =

(15.9)

This is the familiar equation describing how the uncertainty in the mean value of a
quantity goes down with the square root of the number of independent measurements
of that quantity. At the opposite extreme, if all the turbines are associated with just
one mast i (e.g., because the other masts are much farther from the turbines), then

σ combined = σ i (15.10)

These equations confirm that the uncertainty in the array-average wind speed is
strongly affected not only by the number of wind monitoring masts but also by their
placement within the proposed turbine array. They demonstrate clearly why—as we
have stressed—it is important to place the masts in representative locations throughout
the array.