A step excitation optimization method for intensive modal identification of cable net structures

摘要

Compared with the direct force measurement of cables, dynamic modal testing is more economic and applicable in engineering for monitoring the stiffness change of cable net structures caused by pretension deviation. Normally, it is sufficient to detect the variation of structural key stiffness by testing a few modes (target modes) that are, however, likely located in the region of dense modes. To effectively evaluate the stiffness of an existing cable net structure, the identification accuracy of these target modes should be improved when using conventional time-domain methods, e.g., the eigensystem realization algorithm (ERA). Structural free vibration can be induced by a step excitation that is generated by simultaneously releasing loads hung on the cable net. In this paper, the modal energy in the structural free vibration response is expounded to be determined by the work done by the step excitation loads in the direction of this mode shape. Therefore, the identification accuracy of a target mode can be improved by optimizing the layout and magnitude of the suspended loads to increase its energy proportion in the structural free vibration response and suppress those of its adjacent modes. For a certain set of degrees of freedom (DOFs) that loads are allowed to be suspended, a method is put forward to construct a step excitation load that can guarantee that the suppressed modes do not have an energy contribution to its induced free vibration. Referring to the effective independence (EI) method, an iterative algorithm is developed to select a given number of DOEs for the suspension of the step excitation loads that can ensure the target mode has a relatively large energy proportion. With the solution of generalized Rayleigh entropy, the magnitude of the loads suspended from the selected DOEs is further optimized to achieve the maximum energy proportion of the target mode. An illustrative saddle-shaped cable net roof is employed to validate the accuracy and validity of the proposed method by comparing the modal identification results under the optimized step excitation, conventional impact excitation and random excitation, respectively. (C) 2019 Elsevier Ltd. All rights reserved.