Identification of operational modal parameters using multi-reference weighted Hermitian transmissibility functions

Abstract

In recent years, methods have been developed to examine the dynamic behavior of structures without affecting their performance. Among these methods, modal parameter identification based solely on output data is a cost-effective approach. Since structures are influenced by multiple input sources simultaneously, the use of poly-reference methods is essential for identifying the structural poles with minimal uncertainty. The main drawback of poly-reference methods is their dependence on loading conditions, which must differ significantly to accurately identify the structural poles. Another limitation of these methods is the presence of non-periodic harmonic components in the frequency responses of structures, which cause dispersion around the structural poles and lead to the identification of spurious poles. To address these issues, a new approach called Poly Reference Weighted Hermitian Transmissibility Operational Modal Analysis (P-WHTOMA) is employed. In this study, a 4-DOF measured structure is analyzed under three earthquake records with varying noise levels using this novel approach. This method utilizes the weight effects on the structural responses to reduce the frequency domain of the responses and minimize spectral dispersion around the structural poles. Therefore, modal parameters are identified with minimal uncertainty under similar loading conditions and correlated input sources. Furthermore, by defining transmissibility functions using Hermitian polynomials, orthogonality is established between the normal vectors of the transmissibility functions, resulting in the identification of modal parameters at the structural poles with reduced uncertainty. In this new approach, increasing noise has less impact on the identification of structural poles. Another advantage of this method is that, unlike single-reference methods, higher structural modes are identified with greater resolution.