Research on parameter identification of transmission line galloping model under different degrees of freedom

Abstract

The phenomenon of galloping in transmission lines has long been a topic of significant interest, and establishing an accurate galloping model is highly challenging. Consequently, the sparse identification of nonlinear dynamics (SINDy), Integral SINDy (ISINDy), and Weak SINDy (WSINDy) algorithms are employed to directly identify galloping models from data. A theoretical mechanical model is established for different degrees of freedom (DOFs), using the galloping of iced quad bundle conductors as an illustrative example. Simulated data obtained from the theoretical model are utilized to perform parameter identification of the galloping model using the three algorithms. For the 1-DOF and 2-DOF galloping models, the data is preprocessed using the Hodrick-Prescott (H-P) filter method, and then the parameter identification effects of the three algorithms are compared under different noise levels. The parameter identification effects of the three algorithms for the 3-DOF galloping model are analyzed using noise-free data, while also comparing their robustness across different data lengths. The research findings demonstrate that the WSINDy algorithm exhibits high accuracy and robustness in the parameter identification process of galloping models with varying DOF, surpassing the performance of the SINDy and ISINDy algorithms. The identification models play a crucial role in validating anti-galloping designs and formulating anti-galloping plans in engineering applications.