Advanced damage identification of jacket structure based on data-driven variational nonlinear chirp mode decomposition and radial basis function neural network

This paper proposes a new structural damage identification method based on data-driven variational nonlinear chirp mode decomposition (DDVNCMD) and radial basis function (RBF) neural network. Firstly, a novel time-frequency analysis methodology, termed DDVNCMD, is put forward for intrinsic mode function (IMF) decomposition. Next, an RBF neural network is successively constructed, where the extracted IMF components serve as the input and the structural stiffness coefficients are taken as the output. Then, damage localization and quantification are carried out through the trained RBF neural network. Finally, the proposed method is validated by a jacket structure under three environmental loads and three damage scenarios. Results show accurate damage localization and quantification, with maximum absolute errors ranging from 2.516% to 3.279% across all cases, which indicates that the proposed method can effectively locate and quantify damage of jacket structure and exhibits potential for practical structural health monitoring applications.