Methods for modal model quality improvement - Parameter estimation process (part II)

摘要

This paper concerns quality improvement of modal models estimated by the use of "standard" modal analysis methods. The authors proposed application of the Least Squares Frequency Domain method making use of the Gauss-Newton algorithm to the quality improvement of stabilization diagrams as well as application of the Balanced Realization method to identify and remove spurious poles from stabilization diagrams. Formulated algorithms were implemented in the Matlab environment and verified for data measured on real structures. Created software supports critical stages of modal analysis, which results in quality improvement of estimated modal models and higher credibility of drawn conclusions.