Identification of ductile damage and fracture parameters from the small punch test using neural networks

This paper presents a method for the identification of deformation, damage and fracture properties of ductile materials. The small punch test is used to obtain the material response under loading. The resulting load displacement curve contains information about the deformation and failure behavior of the tested material. The finite element method is used to compute the load displacement curve depending on the parameters of the Gurson–Tvergaard–Needleman damage law. Via a systematic variation of the material parameters a data base is built up, which is used to train neural networks. This neural network can be used to predict the load displacement curve of the SPT for a given material parameter set. The identification of the material parameters is done by using a conjugate directions algorithm, which minimizes the error between an experimental load displacement curve and one predicted by the network function. The identified material parameters are validated by independent tests on notched tensile specimens. Furthermore, these parameters can be used to compute the crack growth in fracture specimens, which finally leads to a prediction of classical fracture toughness parameters.