Development of a Machine Learning-Enhanced Computational Framework for Accelerated Discovery of High-Performance Cutting Tool Coatings for Machining Nigerian Mineral Ores

The discovery of high-performance cutting tool coatings optimised for the machining of Nigerian mineral ore components, including the hardrock and laterite-bound formations exploited in the solid minerals sector across the North Central and North West zones, currently proceeds through slow and resource-intensive trial-and-error experimentation, a limitation that a machine learning-enhanced computational framework integrating density functional theory calculations and high-throughput experimental validation could fundamentally transform. This research develops a machine learning-enhanced computational framework for accelerated discovery of cutting tool coating materials optimised for the specific thermomechanical and tribochemical conditions encountered in Nigerian mineral ore machining operations. The framework integrates a first-principles density functional theory database of hardness, thermal conductivity, oxidation resistance, and adhesion energy properties across 300 candidate transition metal nitride, carbide, and boride coating compositions, with a machine learning surrogate model trained on this dataset to predict property sets for unexplored compositions at negligible computational cost. Multi-objective Bayesian optimisation with acquisition functions balancing exploration and exploitation navigates the compositional design space to identify Pareto-optimal coating candidates. Twelve top-ranked coating compositions are synthesised by magnetron sputtering and characterised for hardness, coating adhesion, thermal stability, and tribological performance. Cutting performance validation is conducted through dry turning of representative Nigerian iron ore and chromite rock specimens. The framework achieves an 8-fold reduction in experimental iterations required to identify coatings meeting performance targets compared to conventional design of experiments, constituting an original contribution to computational materials discovery methodology. Keywords: machine learning, cutting tool coating, density functional theory, Bayesian optimisation, mineral machining.