Coal Preparation » Process Control
The objective was to directly determine the measuring efficiency of a pilot Ultra-Dynamics density gauge. Direct comparisons were obtained by installing the non-nuclear gauge immediately after a standard nuclear gauge within the ACIRL Pilot Plant. That provides a direct comparison of the gauge's accuracy, reproducibility and suitability for use in dense media and tailings circuits. The gauge's initial calibration, drift with time, drift with temperature, influence of media contaminants and sensitivity to magnetics content were assessed.
The "as developed" Ultradynamics magnetic susceptibility density gauge housed the measurement device in a 25mm diameter high alumina sheath. This sheath was inserted into the dense media pipework though a valve/gland arrangement.
The effectiveness of the nucleonic gauge was confirmed with further testing, and the gauges were then recalibrated against the nucleonic gauge data. Testing was completed to determine the gauge accuracy with the new calibration.
Temperature effects on the arrangement within the system were also investigated. By controlling other variables, it was found that temperature was affecting the magnitude of the non-nuclear gauge error. That was tracked to an incorrectly applied formula within the gauge software, and a correction was applied.
A number of tests were then carried out to determine the accuracy of the non-nuclear gauges. The tests ranged from 1.3 to 1.8 in density, and the mode of altering density changed from stepwise to continuous & increasing to decreasing. Tests were also completed at constant density to assess short term drift.
The tendency of the non-nuclear gauges to over or underestimate the nucleonic density was also shown to change depending on whether the density of the system was increasing or decreasing. That was demonstrated in a number of sub-tests.
The effects of contamination on the ability of the gauges to read density were also tested. Three types of contaminants were added to the system at three concentrations. The error in the non-nuclear gauges was shown to increase with increasing system contamination. That error was of the magnitude of 0.015 RD for every 10% of contamination added to the system.
Another test was arranged to investigate the ability of the non-nuclear gauge to detect small changes in magnetite concentration. 1.6 kg of magnetite was added in 1, 0.5 and 0.1kg increments. The gauge identified these changes in magnetite concentration, and this is visible in the following chart. However, the gauge had little success in identifying the density of the mixture without that mixture containing any magnetic material.
Marcy gauge densities and normalised DMC cut-points confirmed the accuracy of the nucleonic gauge. The negatives of the non-nuclear gauges were the occasional erratic jump in signal, and the tendency of the calibration to drift with time. The positives were their ability to pick up small quantities of magnetite (0.01%) and reasonable resilience to contamination. When compared to a nucleonic gauge, the typical error from the non-nuclear gauges was +/- 0.015 RD. This figure of 0.015 RD was also the figure generated as the average increase in error for every 10% of contaminant added to the medium.
The project has proved that the concept of using magnetic susceptibility to measure medium density within a typical dense medium circuit is sound.