Neural networks approach for IR-heating and deformation of ABS in thermoforming

This study focuses on the interaction between an IR-heating source and material to be thermoformed, with the aim of providing an accurate description of the polymer behaviour under the conjugated effect of stress and temperature. The possibility to model material behaviour and develop a reliable and simple system to define thermoforming strategy is of great interest to improve industrial production, reducing manufacturing costs. In this investigation, both tensile tests and temperature measurements were performed on ABS subjected to IR radiation. Different values of distance polymer-lamp, sample thickness, and test rate were considered. The experimental trends were modelled by artificial neural network. A good generalisation capability and high flexibility were found for the proposed neural network solution, in accordance with the experimental results.