Abstract Ultra-high molecular weight polyethylene (UHMWPE) is widely used in biomedical applications due to its excellent wear resistance, low friction, and high impact strength than other polymers. However, its challenging machinability especially under dry machining stemming from poor thermal conductivity, viscoelasticity, and susceptibility to burr formation necessitates advanced process control to ensure high surface quality and productivity. This study proposes a data-driven, multi-objective optimization framework for UHMWPE high speed slot milling, aimed at minimizing surface roughness and burr height while maintaining high material removal rates (MRR). The methodology integrates principal component analysis (PCA) for constructing a unified surface roughness index (SRI) from twelve measured topographic parameters, gaussian process regression (GPR) for surrogate modeling of process responses, non-dominated sorting genetic algorithm II (NSGA-II) for pareto-based optimization, and the technique for order preference by similarity to ideal solution (TOPSIS) for decision support. Milling experiments were conducted under three cooling strategies including dry, super critical CO 2 (scCO₂), and scCO₂+ bio MQL, with four cutting speeds and four feed per tooth levels. Results demonstrate that the PCA-based dimensionality reduction captured over 87% of the surface texture variance, enabling efficient and interpretable optimization. Pareto front analysis revealed that scCO₂+MQL dominated the low-SRI and low-burr region. Under this condition, the process achieved superior surface quality with minimum SRI of − 2.57 and a burr height as low as 0.03 mm, while sustaining high MRR. TOPSIS identified the optimal condition (f z = 0.086 mm/tooth, v c = 970 m/min), which was experimentally validated with a prediction error below 20%.
Siahsarani et al. (Mon,) studied this question.