We shall present a ‘smart’ framework, developed to create surrogate models of complex phenomena in an automated way, based on a minimum amount of computer simulations. To do so, the framework combines Design of Experiments, Computational Modeling, and Neural Networks. While such a framework can be universally applied, we will showcase one possible application: the prediction of the ply properties of a fibre-reinforced-polymer-composite material.
Generating Statistical Volume Elements of the material at microscale, we use a state-of-the-art computational micromechanics model to compute the ply’s strength under transverse loading. By Design of Experiment principles, we then explore all geometrical and material parameters on which transverse strength depends at ply level. Finally, we obtain a simple analytical surrogate model of this complex relationship, by employing a neural network.
Aside from presenting the resulting surrogate model, challenges in the development and automation of the framework will be discussed, and the case will be made for wider application of such a framework.”