Research on Neural Network Prediction of Multidirectional Forging Microstructure Evolution of GH4169 Superalloy

Based on the hot compression experiment of GH4169 superalloy, the constitutive equation and the microstructure evolution model of GH4169 superalloy were established, and they were introduced into the Deform® simulation software to realize the simulation of the microstructure evolution of GH4169 superalloy in closed multidirectional forging, single-sided open multidirectional forging, and bilateral open multidirectional forging. The neural network prediction model of the microstructure of GH4169 superalloy in multidirectional forging was established using MATLAB, and the neural network prediction of microstructure evolution in three kinds of multidirectional forging was determined. Experimental results show that the prediction error of the two methods for the recrystallized grain size doesn't exceed 10%, and the error under high pass is small. For the recrystallization volume fraction, the prediction errors of the two methods are less than 3.5%, and the prediction error of the neural network for higher pass deformation is lower than the finite element simulation. The established neural network model has high prediction accuracy and can be used to predict the microstructure evolution of GH4169 during hot deformation.