Machine learning based finite element constitutive model calibration tool for composite materials

IP due diligence for this project is not yet complete. During this time this project is available only to academics, for scholarly use, on an ‘as-is’ basis. Please read the licence agreement carefully before downloading.

This software is a pipeline allowing for:

1. the automatic calibration of a mesoscopic mechanical constitutive model of composite material for a series of RVE simulations,

2. the machine learning prediction (trained with the dataset produced from the previous step) of the constitutive parameters of the mesoscopic model for any new set of inputs.

Reference:

D. Field, Y. Ammouche, J.-M. Peña , A. Jérusalem. Machine learning based multiscale calibration of mesoscopic constitutive models for composite materials: application to brain white matter. Computational Mechanics. Under review.

Note: the white matter dataset used in the reference is not provided; the user needs to provide the composite material model.