Abstract
Frequency methods such as frequency spectrum analysis, frequency spike detection, demodulation, envelope spectrum method have been widely used for condition monitoring of engineering structural systems. Different from the conventional frequency methods, the transmissibility function (TF) represents the relationship between different system output responses such as, e.g. vibration and acoustic emission sensor measurements. This paper introduces a simple and effective generalized transmissibility damage indicator (GTDI) for TF based condition monitoring. Unlike the conventional transmissibility damage indicator (TDI), the new GTDI can improve the detection sensitivity, reduces noise effects and avoid dynamic loadings effects. This is achieved by combining multiple groups of data to obtain more accurate transmissibility analysis, exploiting all the available TFs, and using multiple references. This has two advantages. First, it does not require any other priori knowledge about the system responses. Therefore the method can be used for the condition monitoring of a wide range of components or systems. Further, the method can be easily implemented using Fast Fourier transform (FFT) or power spectra density (PSD) methods and therefore is computationally efficient. These make
the method very suitable for implementing online real-time condition monitoring. The method is investigated by simulation studies and then applied to analyze the vibration data of the main bearing of operating wind turbines, producing very promising results.
the method very suitable for implementing online real-time condition monitoring. The method is investigated by simulation studies and then applied to analyze the vibration data of the main bearing of operating wind turbines, producing very promising results.
Original language | English |
---|---|
Pages (from-to) | 6347-6359 |
Journal | IEEE Transactions on Industrial Electronics |
Volume | 63 |
Issue number | 10 |
Early online date | 14 Jun 2016 |
DOIs | |
Publication status | Published - Oct 2016 |