Probabilistic Finite Element Model Updating by Data Fusion of Acceleration and Angular Velocity

In a finite element model updating (FEMU), a modal analysis with acceleration data has been widely employed. However, the rotational responses may be difficult to be measured by accelerometers; thus, boundary conditions of structures which are associated with both translational and rotational degrees of freedom might be not easily calibrated with acceleration data. To this issue, this study introduces a data fusion method of acceleration and angular velocity to improve an updating accuracy by employing not only accelerometers but also gyroscopes. In addition, the proposed method also deals with uncertainties in structural properties and boundary conditions based on a maximum likelihood method. Here, a First Order Reliability Method, which is a cost-efficient structural reliability method, is utilized to reduce the computational cost for probabilistic inference. Numerical verification has been carried out, and it has been confirmed that the proposed method shows a great updating accuracy and cost-efficiency.