Dimensional control of ring-to-ring casting with a data-driven approach during investment casting

The deformation behavior of the mush zone for superalloy during investment casting directly affects the dimensional control of casting has puzzled many engineers and scientists for years. Numerical simulations are not directly useful to predict the most suitable pattern allowances. A new data-driven approach to be effectively used for pattern allowance and casting process parameters prediction is proposed. The constitutive relationships and deformation parameter from high-temperature mechanical tests on superalloy K4169 is reported. The inputs are alloy temperature, shell temperature, and pattern allowance with the outputs of diameter and ovality of the ring-to-ring casting, respectively. It turns out that the shell temperature is the most momentous factor that governs the dimensional variability in ovality. An RBF-based approximation model is established and the optimized parameters are the alloy temperature 1500.5°C, shell temperature1052.5°C, and the pattern allowance 1.7258%. The optimized results agree well with the observed in practical casting and the ring-to-ring casting tolerance has been optimized as required within CT6 grade. The proposed method is believed to benefit to provide theoretical guidance for casting practice. The data-driven approach used in this research can be easily applied to different materials and different kinds of casting that are subject to dimensional control upon solidification.