In order to get practical models, we've made a lot of experiments about the
rmal analysis on gray cast irons and got various models by means of differe
nt mathematical methods these years. It was proved that artificial neural n
etworks (ANNs) were fit for modeling of cast iron quality prediction. Two k
inds of BP networks for calculating chemical composition and mechanical pro
perties of gray cat iron were studied. Based on these effective models, we
can develop various types of gray cast irons quality testers. In this paper
, the basic method of applying ANNs in gray cast iron quality prediction an
d the results are introduced.