This study employed data-driven machine learning models to analyze the effects of filler material composition and other process parameters on mechanical properties during the crack repair of nickel-based superalloys such as K417G using wide-gap brazing technology. First, a linear regression model was used to analyze the influence of independent variables (filler material composition and other process parameters) on the dependent variables (tensile strength and elongation). The regression results indicated that temperature and filler composition significantly affected tensile strength and elongation. Subsequently, a TabNet machine learning model was applied to simulate the relationship between parameters such as composition and mechanical properties. The experimental results showed that when four parameters, namely, the filler composition, temperature, holding time, and pressure, were used as input features, the deviation between the actual and predicted values of elongation was minimal, with a value of only 1.5650.
Cheng et al. (Fri,) studied this question.