Affiliations: [a] Center for Electromagnetic Fields Engineering and High-Frequency Techniques, Faculty of Electrical Engineering, West Pomeranian University of Technology, Szczecin, Poland
Correspondence:
[*]
Corresponding author: Barbara Szymanik, Center for Electromagnetic Fields Engineering and High-Frequency Techniques, Faculty of Electrical Engineering, West Pomeranian University of Technology, Szczecin Sikorskiego 37, 70-313 Szczecin, Poland. E-mail: szymanik@zut.edu.pl
Abstract: In this article we present an approach to the quantitative evaluation of the 3D printed sample made of polyethylene terephthalate glycol (PETG) using the active infrared thermography (AIT) method with halogen lamps excitation. For this purpose, numerical and experimental studies were carried out. The numerical model solved with finite element method (FEM) was used first to create a database of signals and further to train neural networks. The networks were trained to detect the heterogeneity of the internal structure of the tested printed sample and to estimate the defects position. After training, the performance of the network was validated with the data obtained in the experiment carried out with the active thermography regime on a real 3D print identical to the modelled one.
Keywords: Active infrared thermography, deep neural networks, LSTM network, additive manufacturing, quantitative materials evaluation
