Long term vibration data analysis from wind turbine -statistical vs energy based features

Abstract

Abstract. Wind turbines are operating in varying conditions. Therefore, the recorded signal is highly nonstationary. The typical approach for damage detection in long term data is based on the energy and spectral analysis. This method, suffer for several drawbacks, especially for the signals with high contamination. Thus, the alternative approach is the application of statistical parameters that may indicate the damage. In order to indicate the frequency band corresponding to the damage the proper statistics are used. In this paper, the well know spectral kurtosis and another statistic are applied to the long-term vibration data from wind turbine. Their performance is compared with the standard energy based methods. It is showed that selectors are able to track the damage development and distinguish between healthy and unhealthy case.

Application of Machine Learning Tools for Long-Term Diagnostic Feature Data Segmentation

“…Just to mention several practical implementations, one may focus on [6] where the procedure of load-dependent features processing with application to wind turbine bearings was proposed. Another example proposed by Ignasiak et al [7] discussed a comparative study on statistical and energy-based parameters analysis in a long-term context, also for a wind turbine. Wodecki et al [8] analyzed the overheating problem for heavy-duty machines operating in the underground mine.…”