A two-stage static structural system identification by observability method

Abstract

Structural system identification (SSI) is a powerful tool for the assessment of the current condition of structures in operation. The basic assumption is that the deterioration of structures is reflected in the change of structural parameters. In SSI, the observability of these parameters depends on the location and the number of sensors. If the available sensors are less than the required ones, some parameters might not be observed. Furthermore, the incapability of identifying all parameters might occur when the sensors are sufficient but placed improperly. An investigation was carried out using SSI by observability method when then number of measurements is the same as the number of sensors. It is seen that a large proportion of the studied measurement sets cannot ensure the observability of all parameters. To improve the observability of structural parameters, a two-stage static SSI method is presented to fully exploit the information in the measurements. In the first stage, the SSI problem is treated in a linear manner and thereby the computation is greatly reduced. In the following stage, the nonlinear relations among the variables in the formulated system are recovered and thus the capability of the method to observe those structural parameters is enhanced.