Investigation of Fiber Content and Porosity Effects on Tensile Strength in Long-Fiber-Reinforced Thermoplastics Using Artificial Neural Network

Abstract

The volume contents of composite materials directly affect the tensile strength of long-fiber-reinforced thermoplastics (LFTs). However, it is not easy to analyze how factors such as the fiber content and porosity affect the tensile strength of LFTs. With this motivation, we investigate the relationship between fiber content, porosity, and tensile strength in various LFTs using a neural network (NN) approach. In this study, polyamide 6 (PA6) and polyphenylene sulfide (PPS) are selected as the resin matrices, and glass fiber (GF) and carbon fiber (CF) are chosen as the reinforced fibers. Therefore, the LFTs invoked in this work were PA6/GF, PA6/CF, PPS/GF, and PPS/CF. The proposed NN, which can predict the tensile strength of the utilized LFTs, was trained using the experimentally measured fiber content, porosity, and tensile strength. Based on the learned NN, we then investigated the effect of fiber content and porosity on the tensile strength in each LFT case. As a result, the proposed NN can continuously express the tensile strengths of LFTs in the given ranges of the fiber content and porosity. It should be noted that the tendency of the tensile strength derived by the suggested NN matches well with the studied properties of LFTs. Consequently, through the proposed NN, it is possible to precisely analyze the tensile strengths of invoked LFTs while containing the trends of the LFTs. The detailed strategies for the experiments and NN approach are presented, and the performance of the proposed NN is evaluated through mathematical approaches and previously studied information on LFTs.