Evaluating methods of wind measurement data
There are several methods of recording wind measurement data and often several methods are applied at the same time.
Rayleigh Method
Evaluates the Rayleigh cumulative probability distribution function. This is a direct variational method, in which the minimum of a functional defined on a normed linear space is approximated by a linear combination of elements from that space. This method yields solutions when an analytical form for the true solution may be intractable. It is applied to calculate likely fluctuations at the site when only the average wind speed of the region is known.
The Rayleigh distribution uses a fixed general formula and reflects regional features only to a certain extent.
Weibull Method
The knowledge of two parameters is required for this method: the scale factor A and the form parameter C. This method describes wind circumstances more exactly than the Raleigh method, as the shape of actual distribution is taken into account. At a value of parameter C = 2, the formula is identical with the Rayleigh distribution.
Classifying measurement data
This procedure is used to structure and minimize the measurement data to an absolute minimum (only the absolute necessities). There are two methods of classification:
- Evaluation of wind speed on two levels
- Scale is divided into segments of constant width.
The frequency is recorded and counted at intervals of one or ten minutes to assess, if the values are within the class limits. Over a period of time a frequency distribution is created to assess the relative distribution, via a division of class values through total number of measurements. The measured data has to be corrected in two steps in order to comply with the requirements and standards (long-term validity at the hub height).
Transformation to hub height
Since wind measurements are usually generated at a lower level than the eventual hub height of the potential wind turbine, a transformation of the data is necessary. This value is usually determined with the roughness length of the site possible for each direction sector (e.g. ground contour). Schedules for roughness length, giving a description of the approximate values of the surroundings, can be applied.
However, a much more reliable and precise method is to collect the wind measurement at two different heights.
Correlation of long-term data
The measurement data should be collected over a period of at least 1 year to ensure that seasonal fluctuations are taken into account. The data of a single year must then be compared with long-term data. Wind speeds can differ largely – up to 20% – from the long-term average.
Long-term correlation data can often be obtained from nearby professional weather stations, airports or from existing wind parks. The measuring standard of this external data may be of a lower standard than that collected from the measuring station.
The most important thing is to ensure reliable continuity of the external data (e.g. data from a weather station that has relocated at some point is not acceptable) and that the site being used for correlation data is comparable.
Wind rose
The unobstructed evaluation of the wind direction is essential to enable the ideal positioning of wind turbines. Information on distribution of wind speeds and the frequency of varying wind directions, as well as the evaluation of the roughness length of the site is required.
To visualise this, one can draw a wind rose diagram based on meteorological observations of wind speed and direction. A circle is divided in 12 - 36 sectors. The radius of the 12 outermost, wide wedges gives the relative frequency of each of the 12 wind directions, i.e. how much percent of the time is the wind blowing from that direction.