Multi-directional Forging Finite Element Simulation And Neural Network Prediction Of Microstructure Evolution For GH4169 Superalloy

GH4169 superalloy has excellent hot-working and cold-working properties.It is often used as high-temperature parts in aerospace and other fields.The working environment has high requirements for the microstructure and properties of materials,so it is necessary to improve these materials by special treatment and production methods.As a typical severe plastic deformation process,multidirectional forging technology can significantly refine the grain structure by applying compression deformation in different load directions,so as to obtain significantly improved comprehensive mechanical properties.On the basis of closed multi-directional forging(CMDF)and single side open multi-directional forging(SSOMDF),this paper creatively puts forward the bilateral open multi-directional forging(BOMDF).Through the development of constitutive equation and dynamic recrystallization model and finite element simulation,the reasonable selection range of process parameters of three multi-directional forging processes was determined,and the recrystallization distribution and evolution law were obtained.Finally,it was verified by experiment.Through the hot compression test,the constitutive equation and dynamic recrystallization model suitable for large amount of hot deformation were established,the hot working diagram of GH4169 alloy was established,and the quantitative description of the microstructure evolution law of multi-directional forging deformation was realized.The secondary development of DEFORM software has realized the simulation of dynamic recrystallization volume fraction and average grain size change.The equivalent strain,recrystallization volume fraction and recrystallization grain size of forgings with different passes at 800 °C,900 °C and 1000 °C under single-open and double-open multi-directional forging process conditions were simulated by finite element method.By comparing the finite element simulation results of closed multi-directional forging,single side open multi-directional forging and bilateral open multi-directional forging,the three-dimensional distribution diagram of recrystallization grain size of forgings was drawn with the help of MATLAB software,the development and evolution characteristics of recrystallized microstructure under three process conditions were revealed,and the influence law of process parameters on recrystallization degree and recrystallization grain size was obtained.The neural network prediction model of microstructure of multi-directional forging of GH4169 superalloy was established by using MATLAB software,and the high-precision neural network prediction with small error between microstructure evolution law and test results within the reasonable process parameters of closed multi-directional forging,single side open multi-directional forging and bilateral open multi-directional forging was realized.The prediction error of recrystallization grain size by the two methods is less than10%.The simulation results were verified by single side open multi-directional forging,and the grain refinement mechanism and texture evolution law of GH4169 superalloy during isothermal multi-directional forging were obtained.After 9 passes,the average grain size decreases from 45.0 μm to 15.5 μm,and the preferred orientation of grains occurs during multi-directional forging.During multi-directional forging,a large number of twin lamellae with rotation axis of 60 ° / < 111 > are produced,which is the main mechanism leading to grain refinement.