DATA VALIDATION - Wind Resource Assessment: A Practical Guide to Developing a Wind Project - page 120

Table 9-3. Examples of trend test criteria
Sample parameter a Validation criteria

Wind Speed Average All sensor types

<

1-h Change 5.0 m/s

Temperature Average
1-h Change

5 ◦ C

≤

3-h Change ≤ 1kPa

Differential Temperature
3-h Change Transitions twice above
or below freezing
a All monitoring levels except where noted.
Source: AWS Truepower.

TREND TESTS. These checks are based on the rate of change in a value over
time. Table 9-3 lists sample trend test criteria. The thresholds actually used should be
adjusted as necessary to suit the site conditions. Note that wind direction trends are
not considered because direction can change abruptly during severe weather or frontal
passage events, among other conditions.
The examples of validation criteria in Tables 9-1-9-3 are neither exhaustive nor
do they necessarily apply to all sites. With experience, the analyst will learn which
criteria are most useful in particular conditions.
In addition to these standard tests, two situations usually receive special flags: tower
shadow and icing.

TOWER SHADOW. Tower shadow is identified when two anemometers fail a rela-
tional test and the wind is from a direction in which one of them is downwind of
the mast. The angular width of the zone of tower shadow depends on the geome-
try of the mast but is typically about 30 ◦ on either side of a line directly through
the tower. For example, if the boom points due east from the mast, wind directions
from 240 ◦ to 300 ◦ would be flagged. The shadowed region may be different for
a lattice tower because the boom is typically offset from the center of the tower.
Before applying such a decision rule, it is a good idea to verify the direction of peak
shadow and the width of the shaded zone by plotting the ratio of speeds between
two anemometers at the same height as a function of wind direction (Fig. 9-3). Such
plots can also reveal unexpected influences by the tower and equipment mounted
on it.

ICING. Icing events are usually flagged when the standard deviation recorded by the
direction vanes is zero or near zero, and the temperature is near or below freezing. This
is a conservative approach since direction vanes tend to freeze before anemometers
do. During periods of detectable icing, it is unwise to rely on data from unheated
anemometers even if the speeds are above the offset, since they could be slowed by
moderate ice accumulation.