Empirical modeling and multicriteria decision-making of surface texture and tool wear in hard turning

Abstract

Dry hard turning (DHT) has emerged as an innovative alternative to conventional machining processes that rely on cutting fluids. The absence of external lubrication alters the contact mechanisms between the cutting tool and the workpiece, thereby affecting the machining behavior. This study investigates the machinability of AISI D2 tool steel heat-treated to a hardness of 55 ± 1 HRC. Machining performance is evaluated using three representative response variables: surface roughness (𝑅𝑎), cutting temperature (𝑇௖ ), and flank wear (𝑉௕ ). The cutting parameters considered are depth of cut (𝑎𝑝), feed (𝑓), and cutting speed (𝑉௖ ). The experimental results are analyzed using Pareto analysis to identify the most influential machining parameters, followed by predictive modeling based on Artificial Neural Networks (ANN) and Support Vector Machines (SVM). Finally, a multi-criteria optimization is performed using the TOPSIS method, assigning equal weights (33.33%) to each performance criterion.