Abstract:Aiming at the currently shortage of the freckle criterion in the alloy constituents, a constituent-based freckle predictor for nickel-base suppalloy was modeled and verified by the experiment data. The effect of each element on the liquid density was quantitatively analyzed by calculating the solutal expansion coefficient β and the thermal expansion coefficient β_T of each element. The impact of each element was quantified by defining a freckle formation tendency factor. Based on calculations of these variables, a constituent-based freckle predictor was modeled, which can predict the influence between different alloy constituents and freckle formation. Three different equilibrium distribution coefficients k were taken into this model and verified which was the best k value to predict freckle. The results show that the fitting of the number of freckles using the k value based on binary phase diagram is poor, and the linear goodness-of-fit (R²) maximum is only 0.4; while using the k values of experiment and Thermo-Calc the fitting is good, and the maximum R² reach 0.87 and 0.95, respectively. Finally, the P values were calculated for the SX alloys, and the results show that the hafnium has the strongest influence on freckle prevention, tantalum also has beneficial effects though not as pronounced as hafnium; aluminum and rhenium have the stronger promoting effect on freckles, and tungsten is the second.

Key words: freckle; superalloy; constituent; freckle formation tendency factor; equilibrium distribution coefficient; prediction model