Structural Continuous Modal Identification by Video Measurement

In order to realize the identification of spatially high-density mode shapes, a method of continuous modal identification based on video measurement is proposed. Firstly, a vision modal analysis method based on optical flow is introduced to analyze enough regions, and the structural frequencies, damping ratios and initial displacements are obtained. Secondly, singular value decomposition is applied to the gradient matrix of reference image to get the principal direction vectors, and the initial displacements are projected to the principal directions to get the principal displacements. Finally, the total displacements are recovered and corrected by the structural continuity condition, and the structural continuous mode is obtained. An arch test is used to verify the method. The identification frequencies and damping ratios are compared with the results of laser Doppler vibrometer. The identification errors of the first five frequencies are less than 0.15%. The identification errors of damping ratios are less than 0.07% except the third order. Compared with finite element method, the modal assurance criterion values of the first five orders are all greater than 0.95. The proposed method only analyses the selected areas, which results in low computational load, and the displacements of regions with insufficient gradient information can be recovered and corrected to obtain the structural continuous mode. © 2025 Nanjing University of Aeronautics an Astronautics. All rights reserved.