Evaluation of the statistical reliability of micro-flotation experiments using a hallimond flotation cell

Abstract

ABSTRACT This paper presents the evaluation of the statistical reliability of micro-flotation experiments using a Hallimond flotation cell. The study was conducted by three experimental analysts using different micro-flotation conditions and samples. Statistical analysis of 20 micro-flotation experiments was performed. Using the Kolmogorov–Smirnov test, the data from all experiments were found to come from a normal distribution. The coefficient of variation ranged from 0.5% to 8.2%. The high variability was mainly justified by the poor reproducibility of the bubble size distribution; this was corroborated by the high deviation of the estimated mean bubble size in the experiments, which was 742 ± 249 µm. A frequency histogram showed that although extreme values up to 21.3% could appear, 70% of the experiments had coefficient of variation less than 4%. The Asymptotic Test and the Modified Signed-likelihood Ratio Test showed that there were not statistically significant differences between the variability of the 20 micro-flotation experiments. The overall confidence interval for the coefficient of variation from 2.2% to 4.5% was established. The coefficient of variation tended to decrease when the repetitions of the micro-flotation experiments increased. The main source of variability of the micro-flotation experiments conducted in the Hallimond cell is the repeatability of the method.