Implementation of ANN Modeling Techniques and Genetic Algorithm in the Diameter Prediction of MWCNTs/Epoxy Nanofibers for CFRP Structures

Electrospun multiwalled carbon nanotubes (MWCNTs)/epoxy nanofibers are placed between the layers of standard carbon fiber-reinforced polymer (CFRP) prepreg composites to improve their physical and mechanical properties. As epoxy resin is a thermosetting material, it must be electrospun carefully maintaining a specific viscosity of the epoxy, and optimizing all electrospinning parameters is both costly and time-intensive. Thus, prior to implementing the different experimental techniques, a modeling methodology is an effective tool for regulating the electrospinning process’s contributing factors. In this case, it is observed that having a smaller diameter of MWCNT/epoxy is very critical because MWCNTs stay aligned inside epoxy nanofibers with a smaller diameter than nanofibers with a bigger diameter. Those aligned MWCNTs can lead up to a 29% increase in the flexural strength of a CFRP structure. different Employing artificial neural networks (ANN) models, the present study investigates the effect of key parameters on the fiber diameter and uniformity of electrospun MWCNT/epoxy nanofibers. The goal of this work is to implement and differentiate the multilayer perceptron (MLP) feedforward backpropagation ANN, radial basis function neural network (RBFNN), and very commonly used support vector machine (SVM) methods in order to construct computational models for predicting diameter of MWCNT/epoxy nanofiber with high accuracy.