When opening WindRose.xls(m)file, Excel displays a security warning, asking whether or not to enable WindRose's macros. The software will run in both cases, but functionality is reduced with macros disabled.
WindRose runs OK in both 32bit & 64bit versions of Windows. However, it fails to install in the 64bit versions of MS-Office (because of this Microsoft bug) and a workaround is currently investigated. Note also, that Microsoft recommends the 32bit Office versions for most users and this version is installed by default, even on 64-bit Windows operating systems.
WindRose accepts a linear correction (y=a·x+b) for all channels (Table 3, Input Sheet). If no values are given there, it is assumed that a=1.0 and b=0.
Now, assume that (a0, b0) are the wrong coefficients with whom the data files were created, instead of the correct ones (a1, b1). Then, the following correction coefficients (a2, b2) should be entered in the Table 3 of the Input sheet, in order to produce correct results:
a2 = a1 / a0 and b2 = b1 - b0 · (a1 / a0)
In case you have multiple anemometers but not all of them have wrong coefficients, then when setting the correction coefficients, remember to set a=1.0 and b=0.0 for the rest ones. Example: if you have 3 anemometers and the correct coefficients for the 2nd one are: (1.003, -0.38) then you should write: 1.0;1.003;1.0 (multiplier) and 0.0;-0.38;0.0 (offset).
4. When setting correction coefficients and the 1st value is negative, then Excel does not accept the cell value.
Assume you have 2 vanes and their "north" should be corrected by -15° and 8°. When setting the correction offset you should put either an apostrophe (') in front of the; '-15;8 (denoting to Excel that this is a text cell) or format the cell as a text, before writing -15;8.
There is an undocumented feature that solve this. If your data have a time step <10min (1sec, 30sec, 1min,...), then WindRose produces automatically a 10minAve.txt file at the end of a run. However, note the following:
1. It is suggested to process data chunks (1-2 months) as WindRose allocates huge memory buffers, and you risk to obtain either an "Out of memory" message or an Excel crash.
2. For 1sec data most probably you will not have a Usdv column. In that case, just set as Usdv column the same column as U, because WindRose needs at least 5 colums to run (U,Dir,Usdv,Time, Date). The created 10minAve.txt file will contain the correct 10min based Usdv values.
When calculating turbulence intensities, it is assumed that the average wind speed does not vary in time. However, sometimes (e.g.: rapid weather changes) there are periods of some hours that the average wind speed systematically increases or decreases and during these periods significantly higher turbulence intensities are calculated. Increased TIs might affect the wind turbine class for the specific site. De-trending is a process where this trend is subtracted using a simple empirical model, based on the work of K. Hansen etal and Damanscke etal.
A nice graphical example of such a case which needs to be detrended is given in K. Hansen etal
Here, when analyzing period P3, the (linear) trend is first subtracted, in order to deduce meaningful turbulence intensity. When raw data (1sec) are not available, then the authors propose a simple empirical model based on the 10min averaged data (the common wind data case).