Discrete analytical mode decomposition with automatic bisecting frequency selection for structural dynamic response analysis and modal identification

Abstract

Analytical mode decomposition (AMD) is one of dynamic response decomposition methods. However, the discretization of the dynamic responses and the selection of the bisecting frequency may cause the failure of the decomposition. In this study, the discrete analytical mode decomposition (DAMD) is investigated. In addition, an automatic bisecting frequency selection algorithm based on the auto-regressive power spectrum is proposed to effectively extract modal responses from the measured dynamic response. The efficiency of the proposed method is verified by a simulated dynamic signal with amplitude decays. Moreover, a 36-story frame structure with closely spaced modes subjected to a Gaussian white noise is simulated as a numerical example, and the dynamic responses of the frame structure are obtained based on the numerical simulation. The individual modal responses of the frame structure considering noise effect are effectively extracted by using the proposed method. Finally, the dynamic responses of a steel-concrete composite girder bridge are measured in the field under operational conditions. It is demonstrated that the measured dynamic responses are effectively decomposed into the modal responses. Based on the extracted modal responses, the modal parameters including natural frequencies, damping ratios and modal shapes of the 36-story frame structure and the tested bridge are further successfully identified accordingly.