Predicting the Bending Rigidity and Formability of Plasma-Treated Spunbond Nonwoven Fabrics Using Artificial Intelligence

Hajar Sharifi, Morteza Vadood, Aminoddin Haji

2024 / Volume 7 / Pages 1021-1038
Received 22 April 2024; Accepted 27 May 2024; Published 10 June 2024

Nonwoven fabrics are used in many industries, and surface treatment by plasma can significantly change their physical and mechanical properties by changing the surface chemistry and morphology. In this paper, oxygen/argon plasma has been applied to the spunbond nonwoven fabrics to predict the obtained properties. Therefore, by using the central composite design and considering 4 independent factors including the fabric weight, fabric direction, plasma treatment duration and oxygen ratio, 51 various samples were prepared and their bending rigidity and formability were measured. SEM images showed that the surface roughness increases due to the plasma treatment. Statistical analysis revealed that all the mentioned independent factors have a significant effect on the measured parameters directly or reciprocally. Also, the use of a neural network model with two hidden layers optimized with a method according to the genetic algorithm can predict the bending rigidity and formability based on the independent factors with errors of less than 7% and 9%, respectively. The errors resulting from the surface response method for the same parameters with the same order are about 19% and 17%. The introduced method can be used as a tool to help researchers in the field of mechanical properties.

plasma treatment, neural network, NSGAII, bending rigidity, formability, spun-bond nonwoven, AI