Application of diagnostic algorithms for wind turbines

Source: Diagnostyka, Numer 2(50) / 2009, Pages 7-11
Author: Tomasz Barszcz (EC Systems)

Summary:

The paper presents the choice of diagnostic algorithms for condition monitoring of wind turbines. The device under monitoring is the power train, i.e. a main bearing, gears and a generator. The object also contains several roller bearings. The typical mechanical structure and characteristic frequencies are shown. The next chapter presents the sensor location and choice of sampling parameters. Additionally the monitoring should include a few key process parameters (output power, wind speed and rotational speed). Quick changes of the operational point are an important feature of wind turbines and should be addressed. There are several diagnostic methods which could be used for the monitoring of a wind turbine. A few of them should be selected for the proper detection and identification of the most important faults. The main part of the paper presents malfunctions and diagnostic algorithms which should be used to detect those malfunctions. To avoid the influence of varying operational point, some additional preprocessing should take place. Apart from relatively simple methods mentioned above, there is a number of more advanced methods. The example of such an algorithm based on the spectral kurtosis is also presented.

Keywords: turbine diagnostic methods, proper faults detection, diagnostic algorithm.

Download the full article for free »