A Neural-Network Model for Prediction of Damage Initiation Force in Mode 1 Delamination of Composite Materials

In this paper, an artificial neural network model is presented to investigate the crack propagation in composite materials under vertical loading mode. A Finite Element model is developed, using ABAQUS software, to simulate the delamination in composite structures under mode 1 loading. A series of mode 1 simulations are carried out for different initial crack lengths and specimen thicknesses. The crack initiation force is calculated for each simulation from the related force-displacement curve. To validate the Finite Element model, the force-displacement curve of a numerical test is compared with the corresponding experimental curve. These data are used later to develop the artificial neural network model, which is capable of predicting the crack initiation force in mode 1 test. The network has two input neurons, which are the initial crack length and the specimen thickness and the crack initiation force is evaluated in the output layer. The good performance of the neural network is shown by comparing its results with finite element simulations. The good agreement between the targets and outputs in the regression plots, indicates that the network response is satisfactory.