Face-centred design approach for modelling and predicting areal surface roughness in freeform surface milling of Stavax ESR

Freeform surfaces have become a major element in high precision industrial design, such as moulds and dies. These sculptured surfaces are characterised by non-uniform and aesthetic features requiring high-quality surface finishing. Therefore, minimising surface roughness becomes critical, especially when machining Stavax, a well-known material, in moulds and dies. This study aims to optimise 3-axis CNC ball nose milling parameters, namely spindle speed ($v_c$), feed rate ($f_z$), and depth of cut ($a_p$), in minimising areal surface roughness on the B-Spline tensor product of the Stavax freeform profile. A total of 17 experimental runs were designed using a face-centred composite approach (FCD) under the response surface methodology (RSM). Areal surface roughness data were captured via Alicona InfiniteFocus and then evaluated by MountainsLab in compliance with ISO 25178. ANOVA validation indicated a highly reliable predictive model, characterised by an R2 of 0.9680, an adjusted R2 of 0.9268, and a predicted R2 of 0.8762 with the difference less than 0.2. This study also developed the FCD mathematical model for factors influencing the roughness.