The optimization and autonomous identification of modal parameters

Abstract

For modal analysis (including SISO, MIMO, OMA), the accuracy of modal parameters depends mainly on the estimators used. Different parameters will be extracted even by the same user with the same method at different time. Here a criterion is presented which can compare the accuracy of the identification results. An optimization method is put forward by which all identification results of different methods can be approximately unified. The difficult problem for the optimization of MIMO test, to calculate modal participation factors by the modal frequency and damping independently, is solved in the paper. An optimization mechanism is put forward by which the mathematical poles can be deleted automatically during the optimization process. By this way, the autonomous identification of modal parameter for SISO, MIMO, OMA test can be realized, the best identification result can be obtained which only depends on the frequency bands of interest without relation to analysis methods and user. Two examples, a car brake disk of MlMO test and a real bridge of OMA test, are used to testify the optimization and autonomous analysis method and are proven successful.