Modeling of flow stress of AA6061 under hot compression using artificial neural network

AA6061 is widely used as structural material in aircraft and automobile industry. The hot deformation behavior of alloy is studied in the temperatures ranging from 300 degrees C to 500 degrees C with the strain rates of 0.001s(-1), 0.01s(-1),0.1s(-1),1s(-1) using a Gleeble-3800 servo-hydraulic simulator. An artificial neural network model trained with back-propagation learning algorithm has been prepared to get the values of flow stress in the interval of temperature for which it has been trained for (300 degrees C-500 degrees C). In which temperature and strain are used as input data whereas stress has been used as target data. For the evaluation of neural network model correlation coefficient (R) and relative percentage error has been calculated (eta). The predicted stress strain curve shows quite the similar behavior comparing to experimental stress strain data. The result shows that ANN model is a very effective tool to model the complex non-linear behavior of flow stress under hot deformation conditions. (C) 2017 Elsevier Ltd. All rights reserved.