An adaptive constitutive model in the isothermal compression of Ti600 alloy

The Ti600 alloy was isothermally compressed at the deformation temperatures ranging from 800 degrees C to 1000 degrees C with an interval of 40 degrees C, the strain rates of 0.001 s(-1), 0.01 s(-1), 0.1 s(-1), 1.0 s(-1), 10.0 s(-1) and a height reduction of 50% on a Gleeble-15000 thermo-mechanical simulator. Based on the experimental flow stress, a fuzzy neural network model (FNN model) was developed to acquire the constitutive model in the isothermal compression of Ti600 alloy. In the present constitutive model, the three inputs of FNN model are, respectively, the deformation temperature, the strain rate and the strain, and the output of FNN model is the flow stress. The predicted flow stress is in a good agreement with the experimental flow stress, meanwhile the predicted accuracy of flow stress in the isothermal compression of Ti600 alloy using the FNN model is higher than that using the regression model. (C) 2010 Elsevier B.V. All rights reserved.