关键词: 灰色系统理论, 神经元预测模型, 矢量运算, 有限元模型修正

Abstract: Based on the multi-scale chirp let sparse signal decomposition (MCSSD), a method for modal parameters identification of time-varying systems is proposed. In the proposed method, the MCSSD method is used to decompose the vibration responses of a multi-degree-of-freedom linear time-varying system into several single-mode vibration responses, at the same time; the corresponding instantaneous frequency of each single-mode vibration response can be obtained. By using the envelope and instantaneous frequency of the single-mode vibration response, the system modal frequency and modal damping ratio can be identified. Compared with the empirical mode decomposition and other time-frequency analysis methods, the proposed method has strong noise immunity and high identification accuracy, further more, the mode confusion problem in decomposing the vibration response can be eliminated. Simulation example of the modal parameters identification of a multi-degree of freedom linear time-varying system shows the effectiveness and accuracy of the proposed method.

Key words: Modal parameters, Multi-scale chirplet, Sparse signal decomposition, Time-varying system