Mass transport in wet overflow ball mills

Particle size reduction by grinding in wet overflow mills is a very common operation in the mineral industry. In spite of many decades of practice and numerous studies, a few problems of critical importance cannot be addressed with the current state of knowledge. For instance, there is no model to predict the influence of process variables on the filling level along the mill; the failure of some large-diameter mills to meet their design capacity cannot be explained; and the maximum feed rate that leads to the overload condition in a ball mill cannot be predicted. These limitations are primarily due to a lack of understanding of the mechanism of slurry flow within the mill.; Previous researchers have treated the ball mill as a black box, because it is virtually impossible to carry out direct measurements inside the mill. Residence time distribution measurements have been done more frequently with chemical or nuclear tracers. In the current study the detailed motion of ball charge and slurry is provided via appropriate physical models and their numerical solution.; The heart of the computer simulation code is the marker and cell technique for tracking the free surface of the slurry within and above the ball charge. The method itself is a numerical technique for solving the Navier-Stokes equations pertinent to the rotational motion of slurry. The grinding ball trajectories are first solved by the well-established discrete element method. The positions and velocities of balls and fluid are exchanged simultaneously between the two algorithms to arrive at the overall ball and slurry velocity profiles. This in turn allows direct calculation of slurry rotational flow, axial slurry flow through the pool and the ball charge, and slurry hold-up.; Experimental confirmation of predictions is provided in two separate mills. First a laboratory-scale mill, 0.305 m diameter by 0.305 m long, is fitted with transparent end-plates to capture the motion of thick sucrose solution on video camera. The profile of the fluid is compared with the corresponding prediction. Next a pilot-scale mill, 0.416 m diameter by 0.641 m long, mounted with its accessories on three load cells, is used for continuous measurement of hold-up weight. Indeed this experimental setup is the first of its kind, since all previously published studies have used either tracer response or weighing the mill contents after the experiment to get the hold-up weight. The response of hold-up to slurry feed rate, percent solids in the mill feed, mill speed and overflow opening are reported in this work.