Prediction of mechanical properties and surface roughness of FGH4095 superalloy treated by laser shock peening based on XGBoost

Abstract

Laser shock peening (LSP) is a novel metal surface modifying technology, which can enhance the mechanical properties of materials and extend fatigue performance of components evidently. The outstanding mechanical properties induced by LSP occurs through severe plastic deformation, which will cause the surface topography evolution. With the rapid development of artificial intelligence (AI), which provide an effective solution to solve complex problems when with limited experimental data. In this work, FGH4095 superalloy was selected as the experimental material. LSP experiments were conducted with a Q-switched Nd: YAG laser. Laser energy of 4 J, 6 J and 8 J were used with overlap rate of 30 % and 50 %. The residual stress, microhardness and surface topography of experimental samples treated by LSP was investigated. What is more, the AI method based on XGBoost was applied to predict the mechanical properties and surface roughness of FGH4095 treated by LSP. The laser energy, overlap rate, depth were set as input parameters, while the residual stress, microhardness and surface roughness were set as output parameters. The predicted results showed a great agreement with the experimental data. It can be indicated that XGBoost is a suitable method for accurately performing prediction of mechanical properties and surface roughness of materials treated by LSP.