The computation of alloy additions is of central importance for the cost- and quality-optimal control of all the secondary metallurgy processes. Its function is to determine the grades and quantities of alloy-ing agents needed to achieve the target analysis, based on the actual analysis of the melt. The current state of the art is a selection strategy geared to a minimum-cost, linear optimisation of the charge ma-terials. This approach has the disadvantage that, on the one hand, the quality-dependent and metal-lurgically relevant constraints are inadequately taken into account and that, on the other hand, the operating personnel's decision-making criteria, which are based on experiential knowledge, frequently lead to better solutions that cannot be integrated.

Building upon suitable available information, the BFI has, within the scope of a research project, de-veloped data-based models which describe the melting loss behaviour of the relevant elements and charge and alloy materials at each individual treatment stage for different groups of qualities. This has led to the development of a new technique, based on the theory of fuzzy sets, for computing the charge and alloy materials. It takes account both of all the process steps comprehensively and of the deduced melting loss uncertainties. The technique has already been put to successful use and has brought about significant cost reductions. A software system based on this technique is currently being installed for the production planning operations of a German steel plant.