Enhancing Instantaneous Angular Speed Estimation with an Adaptive Multi-Order Probabilistic Approach

Abstract

The Multi-Order Probabilistic Approach (MOPA) is a well established method utilised for the estimation of instantaneous angular speed (IAS) based on vibration signals. However, it has some shortcomings and challenges, such as the selection of its input parameters and those for the generation of the time-frequency representation. This work delves deeper into the selection of the frequency bandwidth that defines the region in which the spectra will be used to obtain the IAS. More specifically, a variable range is proposed, to reduce the interference of neighbouring harmonics. Furthermore, the optimal window size for each signal segment is defined using sparsity indicators that evaluate the information content. The variable region algorithm is tested individually and in combination with the adaptive window short-time Fourier transform, using simulated and real data from offshore wind turbines. The results show that the proposed method is much more robust and at least as accurate as the standard method in the IAS estimation. Furthermore, this study demonstrates that the variable MOPA can be deployed to new datasets with a minimum of prior knowledge, as there are fewer parameters to be selected by the user, reducing thus the overall time and effort for accurate IAS assessment.

Keywords: MOPA, IAS, signal processing, monitoring