Modeling of microstructure and mechanical properties of heat treated components by using Artificial Neural Network

The main objective of the present work is to develop a methodology to predict the mechanical properties and microstructure of heat treated components, for a given composition and heat treatment process by using Artificial Neural Network׳s (ANN) and by using advanced thermal modeling tool FLUENT. A rotor shaft made of 30CrMoNiV5-11 steel was heat treated and the temperature profile has been measured by inserting thermocouples at different locations on shaft. Based on the obtained temperature profile and subsequent thermal modeling of shaft, the heat transfer coefficient profile was optimized. The optimized heat transfer coefficient profile was then used to determine the temperature distribution in the shaft at different locations. The temperature profiles obtained from thermal modeling of the shaft were applied on coupons made of 30CrMoNiV5-11 steel. The dataset for Neural Network modeling has been generated by studying the microstructural parameters and mechanical properties on these heat treated coupons using metallography and mechanical testing, respectively. Neural Network training was done with this experimentally generated dataset. The input parameters for the Neural Network were alloy composition, heat treatment parameters and hardness. The outputs obtained were yield strength, ultimate tensile strength, elongation, reduction in area and the volume fraction of pearlite, bainite and ferrite. A graphical user interface (GUI) is also developed for easy use of the model. A correlation coefficient (R) of over 90% was obtained to predict the mechanical properties and the microstructural behavior of heat treated steel. Moreover, the microstructural variation and mechanical properties were analyzed and the results were also found to be in a good agreement with the obtained theoretical results.