Prospects of Applying the Neural-Network Modeling for Estimating the Structure and Properties of Polymer-Composite Materials with Hybrid Matrices

• Published in: Polymer science. Series D, Glues and sealing materials Vol. 15; no. 3; pp. 452 - 456
• Main Authors: Kosenko, E. A., Ostroukh, A. V., Baurova, N. I.
• Format: Journal Article
• Language: English
• Published: Moscow Pleiades Publishing 01.09.2022; Springer Nature B.V
• Subjects: Artificial neural networks; Chemistry; Chemistry and Materials Science; Classification; Composite materials; Evaluation; Mechanical properties; Neural networks; Pattern recognition; Polymer matrix composites; Polymer Sciences; Polymers; Quality control; Robotics; Training; neural-network modeling; polymer-composite materials; hybrid matrices; machine training; properties; structure
• ISSN: 1995-4212; 1995-4220
• DOI: 10.1134/S1995421222030157

Artificial neural networks are used in various fields, ranging from economics and sociological research to medicine and robotics. Neural networks are used to solve a wide variety of problems, including event prediction, associative information search, product quality control, and many others. However, perhaps the most popular task solved using neural networks is visual-pattern recognition. Pattern and symbol recognition make it possible to considerably reduce a working time and enhance the accuracy of various operating procedures. In this work, we consider the problems of enhancing the efficiency and accuracy of the evaluation of the structures and properties of polymer-composite materials (PCMs) with different types of hybrid matrices by classifying the data using neural networks. The results of training a neural-network model for classifying the PCM structures with different hybrid matrices are presented. After additional training, the proposed neural-network model can be used to evaluate the mechanical properties of PCMs with different hybrid matrices and to predict them when designing products.