Environmental Engineering Reference
In-Depth Information
15.1 MEASUREMENT UNCERTAINTY
This is the uncertainty in the free-stream wind speed as measured by the anemometers
after data validation and adjustments. It reflects not just the uncertainty in the sensitiv-
ity of the instruments when operating under ideal wind-tunnel conditions but also their
performance in the field, where they may be subject to turbulent and off-horizontal
winds, the possible effects of the tower and other obstacles on the observed speeds, and
problems such as icing that may be missed in the validation. There may be additional
uncertainties associated with certain anemometers and anemometer types, including
those resulting from manufacturing or design flaws or damage incurred in the field.
1
The uncertainty associated with anemometer response under ideal conditions (called
the
sensor response uncertainty
) is typically estimated to be from 1.0% to 1.5% for
a single anemometer. Considering that they are used for power curve testing and
thus represent the de facto standard for estimating turbine output, class I sensors
may be assumed to have a somewhat lower uncertainty than other sensors. The other
components of the measurement accuracy vary depending on the circumstances. High
turbulence, significant vertical winds, short booms or other factors contributing to
tower effects on the free-stream speed, can all lead to greater uncertainty. The general
range of estimates for a single anemometer mounted in accordance with the guidelines
presented in this topic, assuming good data quality, is around 1.5-2.5%.
The measurement error can be reduced by averaging the data from two sensors
mounted in different directions at the same height on the mast. For those direction
sectors where neither sensor is in the direct shadow of the tower, it is possible to
reduce the measurement uncertainty by a factor of up to the square root of two
(1.414), implying a combined uncertainty range as low as 1.1-1.8%. This strategy
also reduces the risk of systematic error introduced by tower effects. For instance, with
two anemometers mounted at right angles to each other, when one is directly upwind
of the tower, the other points to the side. The first sees a decrease, and the other an
increase, in the free-stream speed. As a result, the average of the two measurements
is nearly unbiased.
The benefit of averaging is lessened when there are significant biases affecting both
sensors, however. Possible examples include the effects of turbulence and vertical
winds, which are likely to be similar for both sensors if they are the same model.
Where it is suspected these effects might be substantial, only partial credit for the
averaging should be taken.
15.2 HISTORICAL WIND RESOURCE
This uncertainty addresses how well the site data (after any MCP adjustments) may
represent the historical norm. It is related to the amount of on-site data, the interannual
1
A significant real-world example of a manufacturing or design flaw is the problem of “dry friction whip,”
a vibratory mode experienced by a portion of NRG Maximum 40 anemometers manufactured between May
2006 and December 2008. The problem typically reduces average wind speeds by up to several percent
and causes unusually wide scatter compared to a problem-free reference anemometer (1, 2).
Search WWH ::
Custom Search