Damage response analysis combined with machine learning to investigate the effect of frequency on the impact-sliding fretting corrosion behavior of Inconel 690 alloy

Abstract Inconel 690 alloys have been widely applied in the manufacturing of steam generator tubes for pressurized water reactors at nuclear power station. However, complicated impact-sliding fretting corrosion behavior always accompanies its entire service period. This study, which is based on experimental research and numerical analysis methods, investigates the effect of impact frequency on the impact-sliding fretting corrosion behavior of Inconel 690 alloy tubes. Then, machine learning is applied to predict the evolution law of the degree of damage. The results show that different impact frequencies do not affect the damage failure mechanism of the impact-sliding fretted alloy tube surface. However, an increase in impact frequency will lead to a more severe degree of damage. The corresponding maximum wear depths of the 5, 10, and 15 Hz impact frequencies caused by the impact-sliding fretting wear scars were approximately 6.630, 11.105, and 14.485 µm, respectively. The corresponding wear volume increased from approximately 3.626×104 µm3 to 6.325×104 µm3 and 8.395×104 µm3. Furthermore, machine learning modeling demonstrates perfect robustness and precision in predicting the damage evolution rule of the impact-sliding fretting corrosion behavior of an alloy tube.