Modeling of ?´ Precipitate Size of IN738LC Using Levenberg–Marquardt Backpropagation Neural Network

The γ´ precipitate size of IN738LC is predicted using a Levenberg-Marquard backpropagation neural network in matlab toolbox. A cast polycrystalline Ni based super alloy IN738LC (a gas turbine material) is considered and the γ´ precipitate size is described as a function of 5 variables (solutionizing temperature, solutionizing duration, ageing temperature, ageing duration, and cooling method (furnace cooling, water quenching, induction cooling, salt bath cooling, accelerated air cooling and oil quenching). The model converges very well and accurately predicts the precipitate size. Because first stage gas turbine blades operate at very high and varying temperatures for extended period of time, the prediction of their precipitate size is crucial as precipitate morphology is responsible for most high temperature properties. The model developed in this work can be useful for predicting creep and other mocrstuctural properties at high temperatures.