Hybrid approach for modeling and optimization of hole taper during laser trepan drilling of Ti-6Al-4V alloy sheet

摘要

The study represents a hybrid approach of artificial neural network (ANN) and genetic algorithm (GA) for modeling and optimization of hole taper during laser trepan drilling (LTD) of 1.4 mm thick titanium alloy (Grade-5) sheet. A feed forward ANN model to predict the hole taper more precisely then a second order regression model, has been developed by utilizing the experimental data obtained during a well designed L_(27) orthogonal array (OA) based matrix experimentation. Further this ANN model has been formulated as an objective function to be minimized using GA tool. The results of GA optimization suggest a considerable reduction in hole taper value.