A Unified Micro–Macro Stochastic Framework for Tool Wear Evolution and Lifetime Prediction

Tool wear results from the interaction of numerous micro-scale contact regions subjected to mechanical, thermal and tribological loading. This paper proposes a physics-informed micro–macro stochastic framework linking contact-scale degradation processes with engineering-scale prediction of wear evolution and tool life. At the micro-contact level, tool degradation is represented by a population of interacting contacts undergoing wear, load redistribution and local instability. Model reduction leads to a state-dependent stochastic differential equation whose drift and diffusion terms describe macroscopic flank-wear evolution, while a jump component represents intermittent degradation events. Abrasion/ploughing, adhesion/BUE, thermal–oxidative and diffusion–chemical wear are represented as dominant degradation mechanisms, whereas coating protection and edge instability are treated as latent degradation states. Tool life is formulated as a first-passage problem with respect to a prescribed wear limit. Model parameters are identified from experimentally reconstructed wear increments, with different tribological systems represented using the same stochastic formulation through coefficient calibration. Validation is performed at three complementary levels: predictive assessment of wear evolution, statistical evaluation of degradation dynamics and independent physical validation. The framework reproduces nonlinear wear evolution and increasing degradation variability, while the inferred mechanism hierarchy and latent degradation states are qualitatively supported by independent experimental evidence. The proposed methodology bridges tribological processes at the micro-contact scale and stochastic reliability assessment at the engineering scale, integrating wear evolution, mechanism interpretation and tool-life prediction within a unified physics-informed framework.