Parameter Identification with Unknown Input and Incomplete Measurements

Abstract

Substructural identification approach has shown its advantages in terms of efficiency and accuracy due to fewer degrees-of-freedom and unknowns involved. The main challenge, however, lies in acquiring complete dynamic measurements at interface, particularly for beam and plate substructures involving angular accelerations. To address this issue, a substructural identification strategy is proposed with measurements of strains and translational accelerations, which are more easily and economically acquired than angular acceleration measurements in practice. Another challenge in structural and substructural identification lies in the fact that excitation forces are difficult or even impossible to measure accurately in practice. In this study, an iterative identification algorithm, incorporating Tikhonov regularization method and an improved genetic algorithm based on a search space reduction method, is proposed to identify parameters of structural and substructural systems without excitation measurements. The iterative strategy is further developed for substructural identification to address the simultaneous absence of complete interface measurements and excitation forces. The effectiveness of the proposed strategies is verified through numerical and experimental studies.