Development of a white-box model for predicting in-process thermo-mechanical loading on PVD-coated carbide tools

PVD coatings are widely used to improve tool life and machining efficiency, yet accurate tool-life prediction remains challenging because thin layers require very fine meshes, coating wear mechanisms are complex, and the coating-substrate interface introduces difficult thermal and mechanical boundary conditions. Traditional white-box models frequently neglect coating effects due to simplified assumptions, whereas black-box models capture complexity but lack physical interpretability. Grey-box approaches offer a promising compromise but require detailed coating- and load-dependent datasets. This study develops a finite-element (FE) model to predict in-process thermo-mechanical loading on PVD-coated carbide tools. The white-box model supports the extensive experimental testing typically required to generate thermo-mechanical load tables for grey-box tool-life prediction. Monolayer TiAlCrSiN and bilayer TiAlCrSiN/TiAlCrSiON coatings-commercially established and representative of industrial thermal-barrier behaviour were tested in orthogonal cutting of C45 + N steel. Cutting forces and tool temperatures were measured using a dynamometer and an embedded two-colour pyrometer for model validation. The model reproduces forces and temperatures with deviations below 10 % and shows that tool wear strongly affects local thermo-mechanical load distributions, while global loads remain similar between the coatings. Friction characterization reveals a tribological advantage of the bilayer coating, with reduced interface friction and smoother surfaces. High-resolution stress and temperature fields are correlated with coating properties to improve wear modelling and provide structured inputs for future grey-box model development. • CEL-FEM predicts thermo-mechanical loads on PVD-coated carbide tools. • Validation shows <10 % deviation in forces and temperatures. • Tool wear shifts local heat and stress fields during cutting. • Bilayer coating reduces friction compared to monolayer. • Data framework supports grey-box tool-wear modelling.