A FEM/AI Strategy for Thermal Characterization of an Epoxy Nanocomposite via a Laser-based Nondestructive Testing Experiment

OTERO, FERNANDO A.; ALTUNA, FACUNDO I.; CHIURO, CARLOS A.

2022 IEEE Biennial Congress of Argentina, ARGENCON 2022

Institute of Electrical and Electronics Engineers Inc.

Año: 2022

This article combines data- and knowledge- driven modeling paradigms for solving the inverse problem of thermal characterization, i.e., the estimation of mean values for the thermal conductivity k and the specific heat capacity Cp of a synthetic epoxy nanocomposite via a laser-based nondestructive testing experiment. For this purpose, a general scheme employing a Finite Element Method/Artificial Intelligence (FEM/AI) methodology expressed as a flowchart has been proposed. In this flowchart, blocks solved via FEM are combined with Deep Neural Networks (DNN) in order to calibrate estimations biased by a dimensional and a source model mismatches. Final results show improvements from previous works in the recovered values mainly in terms of Cp