PREDICTION OF ESHELBY'S INCLUSION PROBLEM SOLUTION USING ARTIFICIAL NEURAL NETWORK

Abstract eng:
In this contribution we present our new approach to obtain or better estimate mechanical fields (strain, stress and displacement) inside isotropic infinite body with ellipsoidal-like inclusions. The precise solution has been given by J. D. Eshelby (1957) to internal and external points of inclusion domains and form the basis of our work. When the Eshelby’s solution is extended to take into account perturbations due to the presence of numerous adjacent inclusions (Novák et al., 2012; Oberrecht et al., 2013) the solution given for dozens of points is very time demanding. Utilizing Artificial Neural Network (ANN) trained by exact Eshelby’s solutions to predict mechanical fields can be achieved considerable speedup at the cost of approximate solution. At this state we only focus on prediction of one component of a perturbation strain tensor for single ellipsoidal inclusion.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket