A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Autoregressive spectrum-guided variational mode decomposition for time-varying modal identification under nonstationary conditions
Highlights The modal identification of nonstatinary and time-varying conditions are investigated by an improved variational mode decomposition method. Determining the initial center frequency that is near the real center frequency of the mode is effective to improve the performance of variational mode decomposition. Balancing factor is another crucial parameter to decompose intrinsic mode functions accurately for time-varying system and nonstarionary condition. Numerical studies and study of practical structures show the effectiveness of the proposed method for modal identification of time-varying cases and nonstarionary conditions.
Abstract Modal identification is an important research subject in structural health monitoring. It is a mature subject for modal identification of time-invariant and stationary conditions. Because time-varying properties are valuable for damage detection and time-varying system identification, and because the environmental loads do not always satisfy the assumption of stationarity, it is necessary to develop accurate identification methods for time-varying structures, structures under nonstationary excitations, or the coupling of the two conditions. The variational mode decomposition (VMD) method, which adaptively decomposes the signal into narrow-band intrinsic modes, shows the potential for time-varying and nonstationary conditions. However, parameter selection greatly affects the accuracy of mode decomposition. In this study, a new modal identification method based on an autoregressive spectrum-guided variational mode decomposition method is proposed for the application of time-varying and nonstationary conditions. First, the determination of the initial center frequencies for VMD is proposed using an autoregressive spectrum-guided method to greatly improve the accuracy. Second, a method to select the optimal balancing factor is developed to improve the decomposition accuracy by controlling the proper bandwidth. Subsequently, an instantaneous frequency identification method is presented by detecting the ridge of the time–frequency distribution. Finally, studies of numerical and actual structures demonstrate the capacity of the proposed method for practical applications of time-varying cases and nonstationary conditions.
Autoregressive spectrum-guided variational mode decomposition for time-varying modal identification under nonstationary conditions
Highlights The modal identification of nonstatinary and time-varying conditions are investigated by an improved variational mode decomposition method. Determining the initial center frequency that is near the real center frequency of the mode is effective to improve the performance of variational mode decomposition. Balancing factor is another crucial parameter to decompose intrinsic mode functions accurately for time-varying system and nonstarionary condition. Numerical studies and study of practical structures show the effectiveness of the proposed method for modal identification of time-varying cases and nonstarionary conditions.
Abstract Modal identification is an important research subject in structural health monitoring. It is a mature subject for modal identification of time-invariant and stationary conditions. Because time-varying properties are valuable for damage detection and time-varying system identification, and because the environmental loads do not always satisfy the assumption of stationarity, it is necessary to develop accurate identification methods for time-varying structures, structures under nonstationary excitations, or the coupling of the two conditions. The variational mode decomposition (VMD) method, which adaptively decomposes the signal into narrow-band intrinsic modes, shows the potential for time-varying and nonstationary conditions. However, parameter selection greatly affects the accuracy of mode decomposition. In this study, a new modal identification method based on an autoregressive spectrum-guided variational mode decomposition method is proposed for the application of time-varying and nonstationary conditions. First, the determination of the initial center frequencies for VMD is proposed using an autoregressive spectrum-guided method to greatly improve the accuracy. Second, a method to select the optimal balancing factor is developed to improve the decomposition accuracy by controlling the proper bandwidth. Subsequently, an instantaneous frequency identification method is presented by detecting the ridge of the time–frequency distribution. Finally, studies of numerical and actual structures demonstrate the capacity of the proposed method for practical applications of time-varying cases and nonstationary conditions.
Autoregressive spectrum-guided variational mode decomposition for time-varying modal identification under nonstationary conditions
Yao, Xiao-Jun (author) / Yi, Ting-Hua (author) / Qu, Chun-Xu (author)
Engineering Structures ; 251
2021-11-01
Article (Journal)
Electronic Resource
English