Modelling of microstructure evolution in hot deformation

This paper reviews various approaches to the modelling of microstructure ev olution in hot deformation, for the purpose of predicting the flow stress d uring deformation or for predicting the subsequent annealing behaviour. Two contrasting approaches are discussed, and illustrated for the example of h ot plane-strain compression testing of Al-Mg alloy. These approaches are (i ) physically based state variable models, in which the microstructure and p roperty evolution is modelled explicitly; and (ii) advanced statistical met hods, for linking processing conditions empirically to properties, or to an nealing rate and final microstructure. The state variable models illustrate some general features of microstructur e modelling and the level of experimental work that goes with it. Of partic ular importance are the accuracy of the data used to calibrate or validate a model, the implications that this makes on the volume of data needed, and the viable level of detail in the model that can realistically be verified . Various sensitivity analyses will be used to illustrate the need for a ba lanced view of model and experiment if a credible predictive capability is to emerge. The statistical methods provide no physical insight, but, nonetheless, warr ant further consideration for hot-deformation problems. They potentially pr ovide a means to optimize time-consuming experimental work, and may provide useful predictive capabilities for industry rather sooner than can be expe cted from complex physically based modelling.