Abstract
Wood-plastic composite (WPC) is being increasingly adopted in construction and furniture applications due to its durability and recyclability. This study investigates face-milling responses-resultant cutting force and cutting temperature-under systematically varied cutting parameters, and develops a radial basis function neural network for predictive modeling. Experiments were conducted on a computer numerical control machining center using a polycrystalline diamond end-milling cutter for face milling with fixed axial depth of cut. Feed speed, radial depth of cut, and spindle speed were selected as input factors. The results indicate that feed speed and radial depth of cut generally increase all force components, whereas higher spindle speed tends to reduce force magnitudes while elevating temperature. The radial basis function neural network yields acceptable accuracy for resultant cutting force (coefficient of determination R(2) ≈ 0.91) and acceptable accuracy for cutting temperature (R(2) ≈ 0.81). These findings demonstrate the feasibility of radial basis function neural network based prediction for WPC face milling and provide guidance for parameter selection.