| Abstract: | Fragmentation measurements in the form of sieve passing and mass fraction data were used to test the capability of three different distributions to fit the observed data over a wide range in fragment size and mass. These distributions were based on Rosin-Rammler, lognormal and simple sigmoidal (S-shaped) functions, having 2 input parameters for the single-component versions and 5 input parameters for the two-component versions. Provided convergence was achieved in the non-linear curve-fitting technique, the two-component versions always provided superior fits to the observed data. However, these versions were very sensitive to variations in the values chosen for the input parameters. In this particular regard, the two-component sigmoidal function was the most robust. The present results also show that the two-component lognormal function provided the best fit to the fragmentation data in a general sense, and the two-component Rosin-Rammler function provided the worst fit. However, there was not a significant difference between any of the three methods. |
