As an important part of cable, steel wire will inevitably have corrosion and fretting wear during use. To investigate the wear morphology and surface mechanical properties of steel wire under different corrosion conditions with different fretting times, a fretting wear simulation model was constructed by abaqus based on the test results. Firstly, according to the surface roughness of the steel wire under different corrosion time obtained by the experiment, the surface node coordinates of the simulation model are changed to simulate different rough surfaces. Secondly, the wear characteristics of the steel wire under different fretting wear are obtained by setting the number of cycle jumps of the subroutine. Finally, based on the experimental data and simulation data, the machine learning database is built. Then, a machine learning model is constructed to predict the wire wear profile under different wear times. The results show that with the increase of corrosion time and wear times, the wear depth, wear area and surface stress increase. Moreover, the machine learning model can accurately predict the wear characteristics of the steel wire under different corrosion times and fretting cycles.
 

Wire, corrosion, fretting wear, simulation model, machine learning.