Quantification of microstructural features and application of artificial neural network in correlation of microstructure and property of Ti-alloys

The mechanical properties of titanium alloys are quite sensitive to their microstructures, which present the nonlinear and interactive relationship with various microstructural characteristics. In the present paper, based on the theory of stereology and quantitative metallography ,a quantitative measurement and analysis model of microstructure characterization in titanium alloys has been developed. With the help of the proposed technique,a series of microstructural features can be achieved. Also, the evaluated parameter of Feret Ratio was involved to establish the characterization model of a phase in microstructure of titanium alloys, which is benefit to the study on the evolution process of microstructure of titanium alloys. On the other hand,the correlation model of microstructure and mechanical property of titanium alloys was developed using artificial neural network. TC17 titanium alloy was selected to verify the trained neural network. It was found that the correlation coefficient is more that 0. 958,indicating that the predicted values are in a quite good agreement with the experimental values. Based on theoretical analysis, the formulae for calculating tensile strength and elongation of high strength titanium alloys were obtained. New high strength titanium alloy is designed using a quantitative method. The errors for tensile strength and elongation of the three alloys were less than 2%.