Multi-variable relationships in a batch annealing process

Abstract

Empirical models are developed to predict the mechanical properties of four groups of cold-rolled annealed steels, based on upstream processing variables in the hot-rolling process, annealing conditions, the chemical composition of the steel and product details. Various linear models are considered but the simplest approach of using a multiple linear regression model is considered the most satisfactory. This analytical tool will allow better understanding of the steel making process and the ability to vary input parameters to improve the product. In particular the number of coils which fail mechanical testing may be able to be reduced, with a subsequent fall in production costs.