Abstract
Fretting contacts are a major source of energy dissipation, surface degradation, and fatigue damage in mechanical assemblies subjected to small amplitude oscillatory motion. Accurate prediction of fretting energy dissipation and identification of the corresponding slip regime remain challenging due to the strong correlation between contact mechanics, frictional response, and evolving slip regimes. In this work, an integrated finite element modelling and machine learning framework is proposed to predict fretting energy dissipation and slip regime classification in a ball‑on‑flat configuration, while finite element stress analysis is used to identify fatigue critical hotspots. The numerical model is first validated against Hertzian contact theory under normal loading. Fretting dissipated energy is then extracted from the area of tangential force–displacement loops, while the loop shape is used to classify the slip regime as partial slip, mixed slip, or gross slip. The generated simulation dataset is used to train machine learning models for predicting both dissipated energy per cycle and fretting slip regime. Contact mechanics-based input features, including maximum possible tangential force, frictional work, and normalized displacement amplitude, are incorporated to improve predictive consistency with the underlying contact mechanics. The trained models demonstrate the capability to rapidly estimate fretting response under new loading and frictional conditions, thereby reducing the need for repeated computationally expensive finite element simulations. The proposed framework provides a data-driven approach for interpreting fretting behavior and identifying the governing slip regime in point-contact systems. This methodology can be further extended to material-specific systems, coated surfaces, and experimentally validated fretting studies.

This publication has 18 references indexed in Scilit: