Development of an inverse procedure for parameters estimates of numerical models

Abstract

Discrete and continuum models, used to study fracture of concrete and rock, rely on an accurate identification of the model parameters. The determination of the parameters, that cannot be measured directly in laboratory experiments, requires the development of a procedure in the framework of Inverse Problems. The Kalman filter methodology is adopted as inverse technique to identify the parameters in the gradient-enhanced continuum damage model. The identification process exploits the experimental data generated in the cable loaded uniaxial tensile tests performed on single-edge-notched sandstone specimens in the Microlab of Delft University of Technology. Various aspects of the Kalman filter technique have been discussed.