The Evolution Of Tumbling Mill Simulation With The Discrete Element Method

The design of tumbling mills has progressed from the Bond work index concept of the 1950s to the population balance concept of the 1980s and then made a quantum leap forward with the discrete element method in the 1990s. The failure of a SAG mill at the Grasberg mine site set in motion the inquiry into the role of shell lifter geometry on the profile of the charge motion within the mill. Quickly, it became evident that this could only be accomplished by tracking individual motion of ore particles and grinding balls and not by study of en masse motion of the charge that had been established previously. The discrete element method (DEM) which was prevalent in the civil engineering literature at that time proved suitable to tackle this simulation problem. In 1991, the two dimensional simulation of laboratory scale mills was validated with the DEM. Later this simulation method evolved into ?MillSoft?? a simulation software package for the design of lifters in ball and SAG mills. MillSoft?simulations proved that cataracting charge motion is not conducive to the life of lifters. The old rule of thumb that the number of shell lifter rows should equal twice the diameter of mill in feet was shattered by simulation results. Instead, the number of rows equal to the diameter of the mill in feet, providing that the lead angle of lifter is around 25 degrees, was established as being more optimal. This bold design change was first implemented and tested by Mark Sherman at Alumbrera Mines in Argentina. Two 36-ft SAG mills responded extremely positively to the design change. This success was immediately repeated at Collahuasi, Pelambres and other mines in South America. The design change has so far exceeded expectation that there is not a single SAG mill around the world today that does not employ a relief angle on its shell lifter. The DEM method has advanced in the hands of the author, mining companies and universities, notably in Australia and South Africa. Today, three-dimensional simulation of charge profile and slurry transport is a standard in the design of mills. There is strong pursuit of a virtual comminution machine with DEM in Australia. In South Africa, an experimental method known as positron emission tracking is emerging as a method for tracking particles in grinding mills. The computational burden of computing with millions of particles in three-dimensional simulations has been eased to less than a day with a technique known as GPU computing. Keywords: DEM method, GPU computing, SAG mill, tumbling mill simulation, discrete element method