Enhanced semi-automated Bayesian modal identification approach for Ultra-high voltage transmission towers

Abstract

The accuracy and efficiency of the modal identification for Ultra-high voltage (UHV) transmission towers under harsh operational environments is seriously affected by environmental noises. In general, the initial natural frequency drowned in the noises is evaluated by the experience of engineers before performing a Bayesian modal identification approach. It will lead to potential errors and limitations to the application of the Bayesian modal identification approach. In this study, a coherent clustering enhanced semi-automated Bayesian modal identification method is proposed and then implemented in the operational modal analysis of an instrumented UHV transmission tower under ambient excitation. The coherent clustering is adopted to obtain the quantitative value of the initial natural frequency, in place of the empirical evaluation. Then, the Fast Bayesian Fast Fourier Transform (FBFFT) method could accurately identify the most probable values (MPV) and the corresponding posterior uncertainty of modal parameters. Moreover, the first ten modes of the UHV transmission tower are identified, and the results of modal parameters are compared to the stochastic subspace identification (SSI) method. The results show that the coherent clustering enhanced semi-automated Bayesian method could accurately identify modal parameters with posterior uncertainty, even for modes weakly excited. The proposed method could provide a reliable basis for the modal updating, damage detection, and dynamic reliability assessment of UHV transmission towers.