Перспективы применения нейросетевого моделирования при оценке структуры и свойств полимерных композиционных материалов с гибридными матрицами

Abstract

Artificial neural networks are applied in various fields: from economics and social studies to medicine and robotics. The neural networks allow one to solve various problems: prediction of events, associative search of information, check of product quality and many others. The recognition of images by the neural networks is probably the most popular task. The recognition of images and symbols allows one considerably to decrease labor-intensiveness and increase accuracy of various operation processes. The problems of an increase in the effectiveness and accuracy for evaluation of structures and properties of polymer composite materials (PCM) with different hybrid matrices by classification of data, using the neural networks, are discussed. The results of learning a neural network model for classification of PCM structures with different hybrid matrices are presented. After addition learning the proposed neuronetwork model can be used for evaluation of mechanical properties of PCM with different hybrid matrices and prediction of them for product design.