Find in Library

Stress-strain curves are of great significance in studying the changes of work hardening, dynamic recrystallization and dynamic recovery of metal during hot deformation, and predicting the stress-strain curves under different thermal deformation parameters is helpful to study the machinability and instability of metal in hot working process. The thermal deformation behavior of Nb-V-Ti microalloy steel was studied by hot compression experiments at strain rates of 0.01-3 s-1 and deformation temperatures of 1000-1200 ℃ on Gleeble-3500 thermal simulation testing machine. The BP neural network model and GA improved BP neural network model were established to predict the stress-strain curves at the strain rate of 0.5 s-1 and deformation temperature of 1050 ℃, and the strain rate of 1 s-1 and deformation temperature of 1100 ℃. The results show that the BP neural network model improved by GA is in good agreement with the stress-strain curves of the test data and the experimental curves. The correlation coefficients are 0.99202 and 0.99734 respectively, and the errors are only 2.7816% and 2.1703%. The relative errors between the predicted results and the experimental results are within the range of [-2, 2]. It is proved that the model is reliable and applicable to a wide range of strain, which provides theoretical guidance for rolling process in industrial production.