Abstract

The rock fragmentation is a generic term used to describe the size distribution of blasted material. Several controllable parameters as well as rock properties themselves, influence fragmentation. Even though there is no method or equation that gives an exact prediction, during the past few years numerous investigators have developed models and techniques to computerize simulation. An effective method to assess fragmentation presently is to acquire digital images of rock fragments and to process these images using digital image processing techniques. In the case of post-blast fragmentation, this is the only practical method to estimate fragmentation since screening is impractical on a large scale. The aim of this paper is to develop a model to predict rock fragmentation after blasting. For that purpose, GoldSize software was used to determine the size of fragmentation by capturing images of fragmented rock in muck piles. The resulting size distribution data was then compared with the results obtained from some prediction models such as Larsson, Kuz-Ram and Kuznetsov. This study shows that the Kuz-Ram model has the most accurate results and is appropriate, but the confidence of the model decreases when changing the rock type. To increase the accuracy of the results, the model was modified by determination of a confidence index. The results of model verification show that the modified Kuz-Ram model can predict rock fragmentation with an accuracy of 80 percent.