The Neural Network Models Of Microstructure Prediction Of TC11 Alloy During Hot Deformation

As the development of computer technology and modern deformation theories, the research of hot plasticity deformation has carried out not only in macro-flow behaviors or local thermal parameter distribution, but also emphasized on prediction and controlling the microstructure evolution. In this paper, the quantitative relationships between thermal-mechanical parameters with microstructure of TC11 alloy have been established by artificial neural network,which can predict the distribution of microstructure during hot forming of TC11 alloy. It has important theoretical significance and practical merits for establishing reasonable hot deformation parameters, ensuring product quality and performance.In this paper, the effects of temperature, strain and strain rate on flow stress of TC11 have been investigated by isothermal constant-strain-rate compression tests on the Thermecmastor-Z thermal simulator. Based on the experimental results, a flow stress BP network model of TC11 alloy during hot deformation has been established. The genetic algorithm has been put forward to improve the stability of convergence by optimizing the weights and bias. The comparison of the experimental results with the calculated results shows that the BP network can predict the variation of the flow stress of TC11 alloy during hot deformation processes successfully.The microstructure variables of TC11 alloy, including the grain size and the volume fraction of the priorαphase, were obtained by the quantitative metallographic technology, and the effect of process parameters, including deformation temperature, strain rate and strain on microstructural variables has been investigated. Based on the experimental results, a BP network model for microstructure evolution of TC11 alloy during hot deformation has been established. Meanwhile, the prediction model is checked. The comparison of the prediction results with the experimental of the tests show that the prediction model is stable and reliable.Integrating the flow stress model and the microstructural evolution model of TC11 alloy into the FE code, a coupled simulation of deformation with heat transfer and microstructural evolution has been carried out for the disk forging. The effect of process parameters, including the deformation temperature, the deformation velocity and the deformation reduction on the microstructural changes during hot deformation has been studied. To verify the validity and effectiveness of the simulation results, the comparison of the experimental results with the calculated results at the same deformation conditions has been investigated. The comparison results show that the error is less than 10%. All these validate that the microstructural evolution model and the flow stress model of the TC11 alloy during hot deformation are reliable and stable.