| 2.25Cr-1Mo-0.25V steel is a high strength steel, it has excellent anti tempering characteristics strengths, and it is an important material for hydrogenation reactor head, a wide range of application in petroleum industry. The metal flow behavior during hot deformation play a key role in the selection of process parameters and microstructure evolution.In order to have a better understanding of2.25Cr-1Mo-0.25V steel deformation mechanism under high temperature deformation process, in this paper,we will complete the corresponding high temperature compression test through Gleeble-1500D thermal simulation testing machine, the corresponding stress strain data will be obtained, then analyzing the influence of strain,strain rate, and the change of temperature of2.25Cr-1Mo-0.25V steel in hot forming process on flow stress of this steel.With the neural network toolbox of MATLAB software, the deformation temperature, strain rate and strain as the input vectors, the flow stress as output vectors, according to the division of the training data and test data, through selection of the BP network layer, hidden layer and output layer activation function, training function, learning rate, learning function, and creating flow stress constitutive relation model of2.25Cr-1Mo-0.25V steel, genetic algorithm is used to optimize the initial weights and threshold matrix matrix, so as to enhance the accuracy of the network.On this basis, established flow stress constitutive relation model of2.25Cr-1Mo-0.25V steel based on the BP neural network technology is integrated into DEFORM software, secondary development of constitutive model based on function description method will be implemented. Through simulation analysis about hot upsetting deformation, so as to verify the reliability of the new model and accuracy of the new model be relative to the original model of DEFORM software. The work is of positive significance to improve the technical level of the large forging process. |
PDF Full Text Request