Data from the Rautaruukki Raahe sinter plant were analyzed with feed-forward
neural networks. The resulting models were used to investigate and optimize the sinter
plant production rate and the reduction degradation index (RDI) that is an important
sinter quality indicator for small blast furnaces. Especially, the effects of controllable
parameters such as the chemical composition of sinter, physical conditions of raw materials
and factors reflecting the sintering event were studied.