Correlation between austempering parameters and hardness of austempered ductile iron based on artificial neural network

Xuhong Guo

COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(3) 72-76

School of Mechanical and Electric Engineering, Soochow University, Suzhou- Jiangsu, China

Mechanical properties of ADI mainly depend on the austempering parameters, which include austenitizing temperature and time, austempering time and temperature, apart from chemical composition, alloyed elements and casting parameters. In this paper, an investigation has been conducted on the prediction model of mechanical properties of ADI between austenitizing temperature and time, austempering temperature and time as inputs and Vickers hardness of samples after austempering as the outputs based on artificial neural network. There are two types of data of the model: training data and testing data. The former data come from the published literature and 12 experimental data used for network testing. The research results of the model shows that the predicted values approach to the measured data in most of the testing samples and the maximum margin of error between experimental and predicted data is 4.682%.