Abstract
This work focuses on determining the homogenized elastic properties of concrete mixtures containing randomly oriented chopped inclusions using soft computing techniques, to identify optimal inclusion ratios. It also addresses the efficient resolution of the inverse problem, to identify the elastic properties of the constituents from those of the mix. A solution manifold is constructed using computationally efficient 2D axially-symmetrical finite element models of a typical concrete cylinder, at the mesoscale, with randomly generated structures of round aggregates and flat inclusions. To overcome the prohibitive cost of solving the reverse component property identification problem, more efficient analytical and machine learning approaches are proposed and tested for their ability to learn the numerical manifold. The best model is used to determine the elastic properties of the constituents of a set of real cross-linked polyethylene (XLPE) modified concrete mixtures, from their observed homogenized behavior. A maximum XLPE inclusion ratio of about 230 kg/m3 is also determined to maintain an apparent stiffness consistent with structural applications.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call