Investigation of the drilling behavior of the Ti6Al4V alloy and developing prediction model using artificial neural network

This study investigates the influence of cutting parameters on drilling performance for Ti6Al4V alloy using TiN-coated carbide drill bits in a dry drilling environment. The surface roughness, feed force, cutting moment, cutting temperature, tool performance, hole diameter, circularity, and cylindricity values of the hole were studied. The optimal drilling parameters were determined using ANOVA to achieve the desired hole sizes, and prediction models were developed using artificial neural networks (ANN). The experimental results revealed that the feed rate was the most influential parameter for feed forces (92%), whereas the cutting speed had the greatest effect on tool performance (98%). It was determined as a result of the analyses that the optimal feed rate for tool performance, cutting moment, surface roughness, cutting temperature, and feed forces was 0.05 mm/rev. The diameter of the hole was positively affected at high cutting speeds and feed rates. The circularity and cylindricity of the hole were adversely affected by these cutting conditions. Adhesive wear was seen on the drill bits in all cutting conditions, and surface roughness was negatively affected by this adhesive wear. Artificial neural networks (ANN) were used to predict operating cost outputs (feed force, tool performance) and hole quality outputs (surface roughness, deviation of hole diameter). The compatibility of the experimental results and the ANN model was examined in the control experiment. ANN was able to predict the surface roughness and hole diameter with deviation values of 3% and 7%, respectively. Similarly, the deviation values for feed force and tool performance were determined at 3% and 10%, respectively.