| Ball mills are commonly used in the mineral industry for the size reduction of large ore particles. Ore particles and steel balls are charged together in this cylindrical device. When the charge tumbles in the mill, steel balls nip large particles and compress them with sufficient energy to cause breakage. A mineral processing plant running with four to eight ball mills processes about 40,000 to 80,000 tons of ore per day. Considering the enormity of the whole process, it is realized that such a unit process requires an adequate design scheme. Especially, the energy expenditure must be understood to maximize the capacity of the mill.; In the past couple of decades, several scale-up or design methodologies have been developed. First, the parameters of the, grinding model are determined from bench and pilot scale tests. The model parameters for industrial mills are then scaled up from the laboratory data. However, all of these schemes correlated the feed and product of mill without any consideration of the charge motion inside, which is the root cause of size reduction. Therefore, the scale-up too Was based on empirical hypotheses.; The objective of this research work was to develop a practical grinding model incorporating fundamental information about the collision energy distribution and the breakage characteristics of particles subjected to these collisions. Since any modification in the design and operating conditions eventually affects the impact pattern inside the mill, it is anticipated that this will lead to a generalized approach of modeling size reduction in ball mills.; The impact energy distribution was calculated by a computer code known as 2DMILL. The core of the simulation code is the discrete element method, a numerical technique for computing the motion of numerous bodies. A drop-ball apparatus was fabricated to determine the breakage parameters. The model equation was evaluated using several sets of batch grinding tests. The model framework makes the drastic assumptions that all impacts in the mill nip a certain amount of particles and that the nipping applies force in the normal direction only. Despite these assumptions, the model predictions were in good agreement with the experimental data. Furthermore, the impact energy distribution was characterized by a single parameter and correlated quantitatively with the mill product size distribution. |
