Modeling the correlation between microstructure and the properties of the Ti–6Al–4V alloy based on an artificial neural network

Abstract

Modeling the relationship between microstructure and mechanical properties of materials is fairly difficult in that the correlation between them presents highly non-linear and complicated interactions. In this work, the influence of microstructure on the mechanical properties of the Ti–6Al–4V alloy has been investigated using experimental data obtained from the Ti–6Al–4V alloy using forging, heat treatment experiments and tensile tests at room temperature. The microstructural feature parameters utilized were acquired with the help of Image Pro software. Furthermore, a relational model was established correlating microstructure and mechanical properties for the Ti–6Al–4V alloy using an artificial neural network (ANN) technique. In the proposed model, the input data consisted of quantitative microstructural feature parameters, including the volume fraction of α phase, the thickness of the α phase and the Ferret ratio. Whereas the tensile properties are the outputs of the model, such as ultimate tensile strength, yield strength, elongation and reduction in area. The structure of 3–16–4 in the ANN model was determined. The percentage errors between experimental and predicted values are all less than 5%, which indicates that the ANN model established in the present work possesses the desired prediction ability. In order to test the generalization capability of the ANN model, the combined influence of microstructural feature parameters on the mechanical properties of Ti–6Al–4V alloy has been studied using the established ANN model. The research results demonstrate that the proposed method utilizing an artificial neural network offers an accurate correlation between microstructure and mechanical properties for the Ti–6Al–4V alloy, which can be effectively and extensively used for other metals and alloys.