Impact of ball size distribution, compartment configuration, and classifying liner on cement particle size in a continuous ball mill

A true unsteady-state simulator (TUSSIM), based on a cell-based Population Balance Model (PBM) with a dif-ferential algebraic equation (DAE) solver, was used for modeling a full-scale open-circuit cement ball mill for better understanding the industry best practices of employing number of mill compartments, classifying liners, and ball mixtures. Model parameters for the particle breakage and classification action with/without the clas-sifying liner were obtained from the available literature for cement clinker. Experimental residence time dis-tribution data for a full-scale cement ball mill was fitted by the cell-based PBM to determine the number of cells and axial back-mixing ratio. Dynamic simulations, conducted to determine the temporal evolution of the particle size distribution and mass hold-up, demonstrate that milling with a ball mixture outperforms milling with a single ball size. Single-compartment milling can achieve desirable product fineness if the feed is pre-milled. Having the same length, a two-compartment mill obviates the need for pre-milling and performs similarly or better than a three-compartment mill, depending on the ball sizes used. For a given set of ball sizes, a distribution with uniform mass of balls, as opposed to that with a uniform number of balls, achieves 8% increase in cement specific surface area. The use of a classifying liner achieves a negligibly finer cement product compared to uniformly mixed balls. Overall, these results agree with experimental observations, lending credence to TUSSIM, while providing rationale to best practices in the cement industry, offering various process insights and a toolbox to optimize existing open-circuit continuous ball mills.