Modelling the Hot-Deformation of Austenite

Master of Philosophy thesis by Mathew Peet, University of Cambridge, 2001


It is known that the hot deformation behaviour of austenite in steels is a complicated process, dependent on chemical composition, microstructure, temperature and strain rate. While many models have been developed to represent the flow stress as a function of these variables, it is however not yet possible to predict the behaviour for a new alloy. The effects of the different variables on the flow stress are investigated. Linear regression techniques are not capable of representing the data, however neural networks are capable of modelling highly non-linear data. A neural network model has been developed using a large database of compositions and meaningful inputs, including composition. The model allows the calculation of error bars that depend upon the position of a prediction in the input space and the level of perceived noise in the data. The validity of the model was evaluated by testing against six compositions of carbon-manganese steels.