| Summary: | This study on the effects of particle filling and size on the ball load behaviour and power
in a dry mill was initiated at the University of the Witwatersrand in 2003. The aim of the
study was to make available a better understanding of the underlying causes in the
different power draws that occur in mills when ore particles are being added to the ball
load. This mimics the process of filling an industrial grinding mill after a grind out has
been performed. Typically after a grind out, the mill operator would refill the mill with ore
up to the point where maximum mill power draw is registered. At maximum power draw
it is assumed that the void spaces within the ball load are filled with ore particles and
that the charge is well mixed.
In order to conduct the study an inductive proximity probe was used to measure the
dynamics of the load behaviour. This novel technique in measuring load behaviour was
chosen due to the fact that the probe could sense the presence of steel balls
independent of the presence of particles in the mill. The probe’s response to a load
comprised of steel balls only at the fillings of 15-45% and mill speeds of 60 – 105%
indicate that the various changes in load behaviour such as cataracting, centrifuging,
ball packing and toe and shoulder responses were easily distinguished in probes
responses. Further tests were conducted in a mill with a 20% ball filling with increasing
coarse or fine silica sand particle filling from 0 – 150% at the mill speeds of 63-98% of
the critical mill speed. These tests clearly reveal radial segregation of coarse silica
sand, increased ball cataracting and centrifuging of just silica sand or a combination of
balls and silica sand. The impacts of these phenomena have been discussed with
reference to industrial mills.
The physical parameters defining the load provided by the inductive probe made it
interesting to make use of Morrell’s C model to simulate the power drawn by the mill.
Modifications to Morrell’s model were made thus leading to a modification in the toe and
shoulder model and proposals for a segregated charge model, a centrifuged charge
model and a particle pool model. Furthermore a modelling study based on the torquearm
modelling approach was conducted. Here Moys power model was used to study
the effect increasing coarse and fine particle filling has on the power drawn by a mill. A
liner model was proposed to define N* as a function of particle filling. In both modelling
cases the models were used to account for the various conditions arising within the load
as particle filling and mill speed increases.
|
|---|
