ANN based prediction model for fatigue crack growth in DP steel

An artificial neural network (ANN)-based model was developed to analyse hig h-cycle fatigue crack growth rates (da/dN) as a function of stress intensit y ranges (DeltaK) for dual phase (DP) steel. The training data consisted of da/dN at DeltaK ranges between 5 and 16 MPa rootm for DP steel with marten site contents in the range 32 to 76%. The ANN back-propagation model with G aussian activation function exhibited excellent agreement with the experime ntal results. The fatigue crack growth rate predictions were made to demons trate its practical significance in a given real-life situation. Because of the wide range of data points used during training of the model, it will p rovide a useful predictor for fatigue crack growth in DP steels.