ONLINE PARAMETER INDENTIFICATION OF CONCRETE CONSTITUTIVE MODEL

In order to ensure the reliability and to improve the accuracy of a numerical substructure in a substructure pseudo dynamic test, parameter identification and modal updating have been proposed and recently received more attention. This paper proposes a method based on unscented Kalman filter for online identifying the parameters of a concrete constitutive model to solve the problem that the parameters are difficult to determine when establishing the numerical substructure with a fiber model. Firstly, the constitutive equations of confined and unconfined concrete are unified by defining constitutive parameters and non-constitutive parameters. Then, on the condition that the stress of concrete could be measured, the proposed method is validated through numerical simulation. Finally, some parts of the OpenSees source codes are modified to implement this identification method on an actual situation that the restoring forces of the experimental substructures could be measured. The results show that the method possesses good stability and accuracy. Thereby, it largely improved the reliability of the numerical model.