Modeling of closed-circuit ball milling of cement clinker via a PBM with a variable Tromp curve for classification

Full-scale closed-circuit cement ball milling was modeled using a true unsteady-state simulator (TUSSIM), based on a transient cell-based population balance model (PBM) with a set of differential algebraic equations (DAEs). As a major novelty, the PBM for the mill was coupled with a dust load-dependent, variable Tromp curve for the air classifier. Results from the dynamic simulations suggest that lower air flow rate or higher rotor speed of the classifier not only led to a finer product but also increased the dust load of the classifier feed. When the dust load was too high, operational failure due to overloading of the whole circuit was detected. Finally, TUSSIM was used for process optimization with a global optimizer-DAE solver to identify the air classifier's parameters that yielded desirable cement quality while maximizing production rate. We have demonstrated that the optimization could increase production rate by 7% compared to the baseline process.