The monitoring of noise and vibration of products is becoming increasingly important and is inseparable from understanding their dynamic properties. Although many numerical methods for performing modal analysis are available, real dynamic properties must be validated through experiment. With the advances in measurement techniques and the increasing complexity of studied structures, there is a need to employ advanced signal processing methods. In this master's thesis, we present two iterative Sanathan-Koerner methods of rational approximation as a tool for identifying modal parameters. After reviewing the theoretical background, the validation of the proposed methods is done on an analytical model, which also includes a study of the influential parameters. In the second part, we implement the methods in an open-source package and use them for the experimental identification of the natural frequencies and damping ratios of the elementary structure using the stabilization diagram. The accuracy of the proposed methods on experimental data is comparable to the results of current methods. In the case of simulated increased noise presence in the experimental data, the proposed methods exhibit a higher degree of stability compared to the existing methods.
|
