It is shown that metal recovery in mineral processing plants is often linearly correlated with feed or concentrate grade, particularly in flotation. This correlation can be used to analyse the data from plant trials in which two operating conditions are being compared, such as different reagent regimes or circuit configurations.
The method involves the statistical comparison of the two linear recovery-grade regression lines corresponding to the two operating conditions. Details of the method and examples of its use are given.
Although not as efficient as a formal experimental design, the method can be used where such designs are impractical, or in the analysis of historical data.