State and Input Estimation in Structural Systems

The purpose of this study was to properly describe the state and input estimation problems in the context of structural dynamic systems (Chapter 1). The selected estimation techniques were the Minimum Norm Response Corrector (MIRC) and the Augmented Kalman Filter (AKF). These algorithms where selected because they are formulated under different conditions and assumptions, e.g. the MIRC works on a batch (offline) formulation while the AKF works in a non-batch (online) formulation being the biggest one; others are pointed out in the description of the methods (Chapter 2). To validate these two methodologies a 2D six stories steel frame was modeled with the intent to estimate unmeasured states and inputs (Chapter 3). Four different scenarios were selected, in which each one postulates specific assumptions about the degree of confidence on the mathematical model (nominal model) that is being used and the description of the excitations. Another aim was to find out how proposed methodologies address these problems and discuss the assumptions made in the proposed solution. The determination of the number of independent excitations acting on a system and the localization of them was examined in the study (Chapter 4 and 5). Finally, some brief discussions about the type of measurement, inter sampling behaviour of the force of the Continuous to Discrete (C2D) assumption and some brief statistical results are presented Chapter 6).