ISSN 3141-6632
Tool breakage remains a critical limitation in dry drilling because unexpected failure can compromise machining accuracy, damage the workpiece, increase production downtime, and raise tooling costs. However, fracture is preceded by deformation of the generated chip, suggesting that chip strain may provide a measurable precursor to impending tool failure. This study investigates chip strain at the onset of tool breakage as an early-warning response variable and develops predictive and optimization models relating its behaviour to controllable drilling parameters. Feed rate, cutting speed, and depth of cut were investigated during dry drilling of mild steel using a central composite experimental design. Response Surface Methodology (RSM) was employed to quantify the individual, interaction, and quadratic effects of the drilling parameters, while an Artificial Neural Network (ANN) was developed as an alternative data-driven predictive model. The quadratic RSM model exhibited strong predictive performance, with R² = 0.9627, Adjusted R² = 0.9291, and an adequate precision of 19.198. Analysis of variance showed that the interaction between cutting speed and depth of cut exerted the strongest influence on chip strain (F = 31.77), (p < 0.0002), followed by the quadratic effects of depth of cut (F = 51.45), (p < 0.0001), and cutting speed (F = 31.77), (p = 0.0002). Numerical optimization identified a minimum predicted chip strain of 0.0253804 at a feed rate of 0.30 mm/rev, cutting speed of 146.22 m/min, and depth of cut of 0.50 mm, with an overall desirability of 93.1%. The ANN yielded R² = 76.56%, with the experimental-to-predicted relationship expressed as (Experimental = 0.004602 + 0.8654 ANN). The superior performance of the RSM model further demonstrates the importance of interaction and quadratic effects in describing strain development under the investigated drilling conditions.
Keywords: Chip Strain; Tool Breakage; Dry Drilling; Fracture Prediction; Response Surface Methodology; Artificial Neural Network; Machining Optimization.