Data‐Driven Modelling of Quality and Performance Indices in Sintermaking

In developing data‐driven models of complex real‐world systems, a common problem is how to select relevant inputs from a large set of measurements. If the observations of the outputs to be predicted by the model are scarce, which may be the case if the outputs are indices determined in toilsome laboratory tests, strict constraints have to be imposed on the number of model parameters. In neural network modelling, this limitation in practice also restricts the number of hidden nodes as well as the number of input variables, since the dimension of the weight vector strongly depends on these. This paper presents a systematic method for data‐driven modelling with feedforward layered neural networks, including a method for the selection of input variables. The method is illustrated on a problem from ironmaking industry, where sinter quality indices are predicted on the basis of raw material properties. Furthermore, an inversion technique of the resulting network models is proposed, where an optimization problem is solved to maximize the performance of the sintering operation by manipulating the inputs.