A Multiscale Framework for Predictive Wear-Resistant Alloy Design

Thesis Abstract

This study addresses the persistent gap between fundamental understanding and practical predictability in wear behavior of engineering alloys by developing a multiscale framework that links atomistic mechanisms, microstructural features, and macro-scale tribological performance. The central aim is to enable predictive design of wear-resistant alloys through integrated theoretical, computational, and experimental insights. Specific objectives are to (1) characterize phase stability, dislocation–precipitate interactions, and diffusion processes at the atomic scale under representative sliding conditions; (2) quantify how grain size, precipitate morphology, and phase distribution govern micro-scale wear mechanisms using in-situ and ex-situ characterization; (3) formulate a mesoscale model that couples crystal plasticity with tribochemical reaction kinetics to predict surface deformation, film formation, and debris generation; (4) integrate a data-driven surrogate model trained on multi-scale simulations and experimental data to forecast wear rate and scuffing propensity under varying load, speed, and environment; and (5) validate the framework against industrially relevant alloys and service conditions to derive design guidelines for improved wear resistance. The methodology adopts a hierarchical, mixed-methods design anchored in three complementary domains. At the atomistic scale, density functional theory (DFT) and molecular dynamics (MD) simulations (approximately 200–300 configurations per alloy system) will assess phase stability, stacking fault energies, and dislocation–precipitate interactions under contact-like boundary conditions. The micro-scale experiments will involve microstructure–property mapping on polished specimens of selected Ni-based and Fe-based alloys (n = 6 alloys, each with variants) using transmission electron microscopy (TEM), electron backscatter diffraction (EBSD), atom probe tomography (APT), and in-situ nanoindentation coupled with high-resolution SEM to observe wear-initiation processes. The macro-scale tribological testing will employ pin-on-disc experiments (n = 180 tests) across a factorial design of load (5–40 N), speed (0.5–3 m/s), temperature (ambient to 600°C), and lubricant chemistry, complemented by surface profilometry and wear debris analysis. Data sources integrate published databases, institute archives, and in-house measurements. The analytical framework comprises (i) regression analysis and ANOVA to identify statistically significant factors of wear rate; (ii) crystal plasticity finite element (CPFEM) simulations to elucidate strain localization and surface roughness evolution; (iii) kinetic Monte Carlo and coupled phase-field models to capture tribochemical reaction pathways; and (iv) a Gaussian process regression-based surrogate model to enable rapid wear predictions across design space. The theoretical underpinning draws on the dislocation–precipitate interaction theory, conforming to the approach of multiscale modeling linked by scale-bridging operators, and Supported by tribology theory, including Archard’s wear law with rate modifiers derived from microstructural state variables. The validity and reliability of instruments will be ensured through calibration against standard reference materials, cross-validation of the surrogate model (k-fold, k=10), and repeatability checks on a subset of experiments (n = 15 repeats). Expected findings include a quantified mapping from microstructural attributes (e.g., coherent vs. incoherent precipitates, grain boundary character distribution) to wear mechanisms (adhesive, abrasive, oxidative wear) and a validated multiscale predictive framework capable of forecasting wear rate with R2 > 0.85 and predictive intervals within ±12% under tested conditions. The study anticipates revealing critical thresholds for precipitate dissolution under high-temperature sliding and identifying microstructural configurations that suppress oxide-film spallation. The contribution to knowledge lies in delivering a cohesive, validated multiscale framework that translates atomic-scale mechanisms to macro-scale wear performance, enabling data-informed alloy design for enhanced durability in demanding tribosystems. Practical implications include a design roadmap for 6–8 alloy compositions with optimized hardness–toughness–oxidation resistance balance and a computational toolset combining CPFEM, DFT/MD-informed constitutive laws, and a data-driven surrogate to accelerate materials discovery. The study concludes that integrated multiscale modeling, anchored by rigorous empirical validation, can substantially reduce experimental iteration in wear-resistant alloy development and foster transferable guidelines for industrial alloy design under complex sliding environments. Recommendations include extending the framework to include environmental factors such as humidity and souring gases, incorporating long-term aging effects, and expanding the surrogate model to real-time condition monitoring data for adaptive maintenance strategies.