Gray-Box Identification of Continuous-Time Models of Flexible Structures

Abstract

A noniterative algorithm is proposed to identify the continuous-time state space model of lightly damped flexible structures. This paper is motivated by the need for accurate dynamic models to be used in design of active vibration control systems for, e.g., large space structures and aeronautical structures. A gray-box approach is adopted to preserve the physical meaning of the model parameters and to naturally impose physical constraints to the model, e.g., passivity and reciprocity. The procedure consists of two stages both in the frequency domain; first, a subspace method is applied to estimate the dynamic matrix, then, the input and output matrices are identified by solving a weighted least square problem, both in the colocated and noncolocated case. A numerical case study is carried out to illustrate the features of the algorithm and experimental results are presented by reporting and validating the identified model of a stiffened aeronautical panel in a broad frequency range. The experimental case study serves also for carrying out a comparison between the proposed identification procedure and some of the already existing algorithms.