Design Optimization of High Energy Ball Mills by Discrete Event Simulation

High energy ball mills are used to grind powder material down to a particle size below one micron. Up to now, the detailed physical mechanisms of the grinding process inside the mill have not been fully understood. Hence, the design of these mills as well as their process parameters is based on empirical results. The ball mill has been simulated by discrete event simulation (DES) using typical process parameters to monitor the balls motion and their collisions. The DES method – if applicable – strongly outperforms typical DEM methods. Furthermore parameter variations of the main process parameters have been made and the distribution of the balls impact velocities in normal direction has been analyzed, which is responsible for the grinding effect. Based on the new knowledge about the ball’s motion inside the grinding chamber and the probability distribution of the impact’s velocity, we developed a new grinding chamber. Its shape has been optimized by a computer simulation. The new design of the grinding chamber allows higher impact velocities and therefore finer powder of new mechanical alloyed powders. DISCRETE EVENT SIMULATION The grinding medium (about 4000 steel balls in a lab scale mill) is accelerated by rotor blades in a horizontal drum (Figure 1). The powder is grinded between two colliding balls. Hence, the collisions between balls as well as the grinding chamber are the important events to monitor. Due to the high number of objects, efficiency is a major goal of algorithm development. In the case of the ball mill observations with typical process parameters in a transparent grinding chamber using a high speed camera have shown that only pair wise collisions occur [1]. Thus, collisions can be modeled as instantaneous and pair wise. Consequently, in this work, the high energy ball mill is described by an event driven particle simulation using collisions as basic events. Figure 1: Transparent experimental grinding chamber of a high energy ball mill Event driven molecular dynamics processes in general are modeled as a series of discrete instantaneous collision events. These events are stored in an event list ordered by time.