Lifetime prediction methodology for variable fretting fatigue loading: Plasticity effect

• Published in: International journal of fatigue Vol. 92; no. 2; pp. 531 - 547
• Main Authors: Gandiolle, Camille, Fouvry, Siegfried, Charkaluk, Eric
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01-11-2016; Elsevier
• Subjects: Cracking; Engineering Sciences; FEM; Fretting fatigue; Plasticity; Variable loading; Cracking; Variable loading; Fretting fatigue; Plasticity; FEM
• ISSN: 0142-1123; 1879-3452
• DOI: 10.1016/j.ijfatigue.2016.05.025

•Plastic fretting fatigue variable loading is studied experimentally and numerically.•Good results obtained by dissociating crack nucleation and crack propagation.•Decreasing crack nucleation condition is the most detrimental block order.•Cyclic plastic laws are the best for describing plastic fretting fatigue loading.•More easily acquired monotonic laws provide adequate conservative prediction.
Both experimental and numerical approaches were developed to follow cracking damage in steel cylinder/plane fretting fatigue contact subjected to variable loading conditions.
For the experimental approach, the potential drop technique was implemented on the fretting fatigue test device to observe on-line crack propagation. A calibration curve relating crack length and potential provided direct knowledge of cracking status.
Numerically, total life was separated between crack nucleation and crack propagation. To formalize crack nucleation prediction, Crossland multiaxial fatigue behavior was applied at a critical distance to take into account the severe fretting fatigue gradients. The crack propagation rate was formalized using Kujawski’s fatigue crack driving force parameter, K‾∗, and coupling the Paris law of the material.
Using this combined crack-nucleation/crack-propagation analysis, constant fretting fatigue conditions and variable fretting fatigue sequences were investigated. Good correlations were observed.
Finally, the influence of different plastic laws was investigated. Both monotonic and cyclic approximations provided good correlation with experimental data, the cyclic description being closer.