From unstable raw materials to the stable clinker: a digital twin and real-time blending

The article discusses a practical way to manage the raw material handling of a cement plant when the raw material base is heterogeneous and gradually depleting, and the local optimization of the chemical composition of the batch with no account taken of the state of the kiln and its other equipment and external conditions does not yield the expected economic effect. It is shown how key restrictions were successively removed: first, a smart charge module was developed with inverse calculation of the chemical composition of components from frequent sampling of raw meal and clinker, then a digital twin of the raw material handling process was produced, restoring the path of the material from the quarry and intermediate warehouses through a circular pile storage to the inlet of the mill. Their combination into a single circuit turns the digital twin into a virtual sensor of the chemical composition and physical properties of raw materials, which makes it possible to recalculate the composition of raw meal more than once every 1.5—2 hours according to laboratory samples, and every few minutes — up to automatic operation with the transmission of setpoints to the APCS. This technical solution was worked out on the Industrial Ocean analytical platform of the industrial Internet of things and tested in commercial operation at the factories of the CEMROS holding company. As a result, the range of fluctuations in the composition decreased by 2—3 times, the kiln fuel consumption - by 2—3.5 %, the productivity of raw material mills increased to 4 % with a decrease in specific electricity costs for grinding by 1-3 %. Specifically, the "organizational" effect should be emphasized: a single digital footprint from the quarry to the mill makes it possible to transfer disagreements between process stages from the opinion plane into the data plane.