Estimation of kinetic parameters of a grinding-liberation model

A method of parameter estimation is outlined for a two-component grinding-liberation model. The large number of model variables is reduced by using functional forms. The parameters for the functional forms are estimated using a modified Gauss-Newton algorithm. The technique is applied to both synthetic and real experimental data. Parameters estimated from wet grinding batch tests of Illinois No. 6 coal were used to predict mill output for a larger pilot-plant mill. The predictions were found to match the actual mill output within the limits of experimental error.