Fatigue Crack Growth on Modified CT Specimens Using Artificial Neural Networks

•Mixed mode FCG on FPB and modified CT specimens was simulated using the VCTD criterion.•The VCTD criterion accurately captured the FCG behaviour of different specimen geometries.•The resulting data was used to train ANN with up to 5 input variables, and one hidden layer with 5 neurons.•Trained ANN were able to predict the miss or sink hole behaviour of different modified CT specimens.•Both FCG paths and specimen lives were predict by the trained ANN.