Parametric study of drilling process in ramie-epoxy composites using grey relational approach and artificial neural networks

Drilling is one of the primary and most preferred machining operations for producing holes in composite materials. Unlike conventional materials, drilling of fiber reinforced polymer (FRP) composite is more challenging because of its heterogeneous and anisotropic nature. In the present study on drilling behavior, ramie-epoxy composites are fabricated and are then subjected to drilling tests following Taguchi L16 orthogonal array. The influence of spindle speed, feed rate and drill diameter on the thrust force, torque and delamination are studied using the Taguchi technique. The optimization of drilling parameters for multi-response characteristics is analyzed using grey relational approach. The results from grey relational analysis (GRA) indicate that the most dominant factor influencing the multi-performance characteristics of drilling ramie-epoxy composites is feed rate contributing 77.41% followed by drill diameter and spindle speed with individual contributions of 6.95% and 5.51% respectively. It is found that the optimal factor setting for attaining minimum thrust force, torque and delamination simultaneously can be obtained while drilling the composite at a spindle speed of 4000 rpm, feed rate of 40 mm/min and drill diameter of 4 mm respectively. Further, an artificial neural network (ANN) model is proposed as a tool to effectively predict the thrust force, torque and delamination. The results showed that the model is fairly good and found to be predicting the thrust force, torque and delamination with high accuracies of 89.593%, 95.144% and 96.186% and errors of +/- 2%, +/- 3% and +/- 0.3% respectively. Analysis of chip formation mechanisms and morphological study of drilled holes is done to study the type of chips formed during machining and to identify the possible causes of delamination.