Contributions for enhanced tracking of (onshore) wind turbines modal parameters

Abstract

The structural performance of wind turbines is highly dependent on their dynamic behaviour, meaning the monitoring of their structural health should take their dynamic response into account, which can be pursued through the continuous application of Operational Modal Analysis (OMA) to time-series recorded on the structure and subsequent tracking of modal properties. However, it is impractical to manually analyse long periods of data, being therefore essential to automatize OMA.Furthermore, it is necessary to link the results obtained in each dataset. This is usually based on the comparison between estimated modal properties and a trusted set of references. However, the identification of said set of references is normally performed manually, taking time, and producing a result that may be biased. Additionally, in the case of wind turbines, due to the diversity of their operating condition, several sets of references have to be defined, making this process even more cumbersome and subjective.In this sense, this work takes advantage from a continuous monitoring program of an onshore wind turbine, to propose and validate a procedure for the automatic determination of a set of reference modal properties and their subsequent use in modal tracking, which is then validated with results from 18 months of continuously recorded data. Additionally, a complementary very expedite procedure for a preliminary evaluation of the wind turbine dominant frequencies, and detection of eventual problematic resonances, is presented and evaluated using alternative sensors (accelerations, strains and rotations) and data from 3 different turbine models.