Experimental/Computational-Based Determination of Material Parameters for Nonlinear Simulation of UHPFRC

Abstract

The paper describes development of a soft-computing-based identification software FRCID-4PB for material parameter determination of ultra-high-performance fiber-reinforced concrete composite material. Such a determination is performed with the help of experimental data from four-point bending tests used in inverse analysis based on artificial neural networks and nonlinear computational modelling of tests. A new tensile softening model for studied composite material has been proposed and tested with the help of sensitivity analysis. The procedure also utilizes stratified statistical simulation method for the preparation of a training set for the artificial neural network. Trained network is then implemented into the software allowing routine and user-friendly identification of material parameters from the test results. The main aspects of the software implementation, its testing and application is described and discussed in the paper.