Application of Artificial Neural Network for Prediction of the Cyclic Oxidation Behavior of Electrical Resistance Sintered Gamma-TiAl Intermetallics

TiAl based intermetallics are widely used for structural applications in aviation, chemical engineering, automotive and sports equipment. In this study, the electrical resistance sintering (ERS) technology used in the production of gamma-TiAl intermetallics is based on the principle of applying pressure simultaneously with a high-density electric current. The purpose of this study was to investigate the cyclic oxidation resistance of Ti-44Al-3Mo and Ti-44Al-3Nb alloys (at.%) and the applicability of artificial neural network (ANN) modeling for the forecast of the oxidation behavior of these alloys. In order to obtain this aim, the alloys sintered by ERS were oxidized at 900°C for 360 h and then the oxidation behaviors of them are evaluated by plotting a graph between weight change as a function of time. The data collected after the oxidation experiments were used to construct the prediction models. The modelling results show that a good agreement between experimental results and prediction results was found. The oxidized alloys were characterized using XRD and SEM-EDS. The XRD patterns revealed the oxidation products are composed of TiO2 and Al2O3 oxides. SEM-EDS analysis indicated that the oxide scales of alloys are made up of a multilayered structure.