Drilling Parametric Analysis Of Perforated Copper Foil On Hemp/ Sisal Fiber Composite Using Sequential Ann

Innovation in developing optimized process parameters is vital to meet industrial demand in composite applications. Especially in the fiber composite drilling process, unexpected dimensional errors cause failure due to the anisotropic nature of cellulosic fibers and heterogeneity in reinforcement. A sequential artificial neural network (SANN) technique has been required for composite drilling to predict a wide range of operating conditions and analyze the dimensional imperfections. The work fabricates three different sisal/ copper foil/ hemp (SCH) hybrid composites with varying perforated copper foil pitches (10/20/30mm) designs. The drilling experiments were designed using a central composite design. The effect of experimental/ SANN input parameters [spindle speed, drill point angle (DPA) (92°, 112°, and 132°), composite type, feed rate] and output responses (thrust force, torque, and roundness) were analyzed. The SANN model was trained through the experimental data and verified with a test set of experiments. Finally, the proposed SANN model was used to predict the drilling responses at a lower limit to the higher operating range for 70 validation experiments. The interaction responses were analyzed using the response surface methodology technique between the drilling (Input/ Output) process parameters. The experimental results inferred that the SCH-3 type with a higher pitch between the perforations shows minor roundness error. A drill bit with higher DPA (132°) induces more delamination effects in the laminates. The validation experiments revealed that the predicted ANN data correlated with the experimental data with less than 2% mean absolute error.