Prediction of the mechanical properties of forged TC11 titanium alloy by ANN

In this paper, an artificial neural networks (ANNs) has been applied to acquire the relationships between the mechanical properties and the deformation technological parameters of TC11 titanium alloy (approximately corresponding to ASTM Ti-6Al-6V-2Sn), using the data from the isothermal compression test and the conventional tensile test of forged TC11 titanium alloy at room temperature. In establishing these relationships, the deformation temperature and the true strain were taken as the inputs, whilst the ultimate tensile strength, the yield strength, the elongation and the area reduction at fracture were taken as the outputs, respectively. The activation function in the output layer of the model obeyed a linear output, while the activation function in the hidden layer was in the form of a sigmoid function. Comparison of the predicted and experimental results shows that the ANN model used to predict the relationships of the mechanical properties of the forged TC11 titanium alloy has good learning precision and good generalization. The neural network method presented in this paper has been found to show much better agreement with the experimental data than the existing methods (for example, quadratic regression analysis), and to have the advantage of being able to treat noisy data, or data with strong non-linear relationships.