Generating a rock breakage process from sieve data

The sieve data obtained from laboratory rock breakage impact tests are used to infer the underlying probability distribution of fragment sizes by number. This enables one to generate an explicit population of fragments from an impact of specified energy. Such fragments are needed in simulations of mills that are based on the dynamics of individual particles. thod involves generating independent fragments from the number distribution and truncating the last fragment so that total fragment mass equals the mass of the parent rock. Fairly restrictive assumptions are made about fragment shape and density. Given these assumptions, the method is shown to be exact for a particular breakage process based on Poisson statistics. It is also shown to be asymptotically exact for a large number of fragments in a broad class of fragmentation processes. Examples of fragment populations are given.