Estimation of Breakage Function Parameters for the Andrews-Mika Grinding Model Using a Stereostochastic Model for Liberation

For the last few years, the grinding model first proposed by Andrews and Mika (1975) has regained attention from the scientific community. Though sophisticated identification algorithms have been developed, the number of its parameters that have to be estimated remains huge. Most of these constants (often more than 90%) are for the breakage function. This paper describes a new method for reducing efficiently the number of the breakage function parameters of the Andrews-Mika grinding model. It makes use of second-order statistics and of a stereostochastic model for liberation (for instance, BooKing).