Analysis of the machinability of carbon fiber composite materials in function of tool wear and cutting parameters using the artificial neural network approach

Abstract

Local delamination is the most undesirable damage associated with drilling carbonfiber reinforced composite materials (CFRPs). This defect reduces the structural integrity of thematerial, which a ects the residual strength of the assembled components. A positive correlationbetween delamination extension and thrust force during the drilling process is reported in literature.The abrasive e ect of the carbon fibers modifies the geometry of the fresh tool, which increases thethrust force and, in consequence, the induced damage in the workpiece. Using a control systembased on an artificial neural network (ANN), an analysis of the influence of the tool wear in the thrustforce during the drilling of CFRP laminate to reduce the damage is developed. The spindle speed,feed rate, and drill point angle are also included as input parameters of the study. The training andtesting of the ANN model are carried out with experimental drilling tests using uncoated carbidehelicoidal tools. The data were trained using error-back propagation-training algorithm (EBPTA).The use of the neural network rapidly provides results of the thrust force evolution in function of thetool wear and cutting parameters. The obtained results can be used by the industry as a guide tocontrol the impact of the wear of the tool in the quality of the finished workpiece.