A Contemporary Review and Data-Driven Evaluation of Archard-Type Wear Laws

Abstract

Archard's Wear Law and its variants have remained fundamental to wear prediction for over 70 years, despite frequent criticism regarding their simplicity and the variability of the wear coefficient. Numerous recent sliding wear experiments were evaluated to determine whether these models still hold and to what degree and to identify areas for improvement. A total of 75 papers with detailed wear data were chosen for in-depth review, and 39 of them were selected for a regression analysis. Modified Archard models, incorporating variable exponents for load, sliding distance, and material hardness, were optimized for each independent study. A Gaussian mixture model was then used to cluster the optimized exponents into two groups: one centered around the original Archard model, and the other reflecting alternative optimized exponents. The review found that 81% of the papers referencing a wear model employed a variant of the Archard model. Models using material hardness as the primary factor influencing wear volume struggled to make physically sound or accurate predictions. Although the Archard-type model maintains its relevance in relating wear volume to applied load and sliding speed, its dependence on a constant wear coefficient and material hardness falls short in describing wear phenomena comprehensively. Further improvement of the model using mechanics is necessary to enhance the accuracy of wear predictions.