High and medium-pressure roller crushers operate on the principle of inter-particle crushing by crushing material in a packed bed. Although reference in the study is made to high-pressure roller crushers, the work was done with medium–pressure roller crushers i.e. the Loesche mill and the Horomill. The difference in pressures between these equipment and high-pressure grinding rolls (HPGR) from measurements done by FCB, the supplier of the Horomill, was that the medium-pressure equipment operates at pressures of 30 MPa whilst the HPGR operates at pressures as high as 100 MPa. In this study, the differences between single particle and inter-particle crushing and the applicability of the batch press to predict the energy consumption and particle size distributions for medium-pressure roller crushers were investigated. Two phases of crushing were identified and investigated. The first phase occurs when the material is still being drawn into the gap between the rollers. The second phase, called packed bed crushing, occurs when the material is in the gap between the rollers. Crushing and milling energy requirements are discussed with specific reference to the energy models proposed by Bond and Rittinger along with the shortcomings of both these models. These models postulate that the comminution energy is an inverse function of product size. This is proven to be true in this study, but where certain constants are suggested in the aforementioned correlations, this study revealed that these constants are not fixed for all applications but varies for different types of ore. The results were determined for 80% as well as 50% mass passing size. Kick proposed that the energy requirements are a function of the reduction ratio. A model used for roller crushers that is similar to what Kick proposed was also investigated. This was also found to be valid but again, the coefficients in the model vary for the different materials. A new method for predicting the work index of an ore for inter-particle crushing was investigated, which involves using a piston press in which a bed of material is pressed to a predetermined pressure. The proposed method was evaluated using pilot test data obtained with a Horomill, as well as with a pilot Loesche mill. The results indicate that the correlation between the batch press and the pilot mills are poor. The Rosin-Rammler description for particle size distribution was applied and compared with other descriptions. The particle size descriptions of products from the batch press and pilot mills were compared and although there are differences, the batch press can be used to prepare material for initial research in a project. From this study it is clear that, especially when tests are done on an unknown ore body, a work index such as Bond’s, cannot be used for plant design and economic studies unless some pilot plant tests are done to confirm the relationship between energy consumed and product size. / Dissertation (MEng)--University of Pretoria, 2012. / gm2014 / Materials Science and Metallurgical Engineering / unrestricted