Robust computation of the observability of large linear systems with unknown parameters

ABSTRACT The concept of observability plays an important role in structural health monitoring as it can be utilized to theoretically predict whether the system identification of a structure of interest may be successful. The Observability Rank Condition (ORC) is an efficient tool used for analyzing the observability of dynamical systems. However, there are practical constraints, in terms of processing speed and computer memory, in implementing the ORC for large systems, or systems with complicated equations. In this work, a robust approach to compute the observability matrix of linear systems with unknown parameters is proposed. The newly developed algorithm is capable of handling a linear system of substantially larger order than the original algorithm.