Modeling of constitutive relationships and microstructural variables of Ti–6.62Al–5.14Sn–1.82Zr alloy during high temperature deformation

The modeling of constitutive relationships and microstructural variables of the Ti–6.62Al–5.14Sn–1.82Zr alloy during high temperature deformation by using a fuzzy set and artificial neural network (FNN) technique with a back-propagation learning algorithm is the basis of this research. To obtain experimental results for the modeling, the isothermal compression of the titanium alloy in different deformation scenarios was conducted and quantitative metallography was thus obtained. The predicted results of flow stress and microstructural variables, including grain size and volume fraction of the α phase, are compared with the experimental data and the difference is less than 15%. The predicted results are consistent with the experimental data. Furthermore, the comparison between the predicted results of flow stress based on the FNN approach and those by using the regression method has illustrated that the FNN approach is efficient in predicting the flow stress of the alloy.