In this paper, a demodulation technique based on narrowband post-processing is proposed as a wind turbine tower resonance-monitoring tool.
This method provides resonance estimation in time domain, and a confidence interval is given per estimation. The algorithm deployment and its capability of circumventing the modulation effect in the dynamics involved in the tower acceleration response make this method an appropriate candidate for post-processing, especially in automated routines.
In this work, the impact of the main parameters needed to adjust this algorithm is analyzed. The benefits related to its robustness and computational requirements will be stressed and compared with other demodulation techniques. This will be demonstrated with either modulated synthetic signals or real tower turbine acceleration registers. Copyright © 2013 John Wiley & Sons, Ltd.