An ANN approach to predicting the impact parameters of GFRP composites under low-velocity impact

Composite materials, whose excellent strength-to-weight ratio has made them popular in the marine and aviation sectors, are an excellent substitute for metallic materials. The current work investigates the influence of ply sequence, thickness, and impact energy on low-velocity impact (LVI) analysis. The basic objective of this study is to anticipate the peak force and absorbed energy in LVI using an artificial neural network (ANN). The ANN received the experiment data as input, and then the Python program was utilized to create, train, and test the ANN structure employing a back-propagating approach. According to research results, [45/0/- 45/90]s orientation samples improved their energy absorption behaviour and peak force by 12.8%, 16.2%, and 18.6% compared to [0/0]2s, [0/90]2s, and [45/- 45/0/90]s ply orientation samples. The ANN structure predicted the absorbed energy and peak force with satisfactory accuracy. The output from the network was R2 = 0.999 for training, R2 = 0.971 for testing, and R2 = 0.977 for validation, respectively, and overall, R2 = 0.992. Furthermore the failure analysis was demonstrated by using ultrasonic C-scan and field emission scanning electron microscopy (FESEM), and the results showed extreme changes in the behaviour of fibres and matrix after impact.