Affordable Access

Publisher Website

A robust statistical method to evaluate unit operation in coal washery

International Journal of Mineral Processing
Publication Date
DOI: 10.1016/s0301-7516(02)00016-9
  • Plant Trials
  • Data Analysis
  • Regression Line
  • Comparison
  • Statistical Methods
  • Design


Abstract The effect of the change of an operating parameter on the washing performance is conventionally evaluated through a direct comparison of the average clean coal yields obtained under normal and trial conditions. Sometimes, in an apparent refinement to the analysis, the clean coal yields are normalized for their ash levels before such a comparison is carried out. This paper illustrates how such simplistic analytical methods can often be misleading and lead to erroneous conclusions. Another method of analysis involves the comparison of two data sets on the basis of their separation efficiencies. The method is not effective in plant conditions, where the fluctuation of several process variables leads to generation of noisy data. An alternative, statistically robust method has been proposed, where a simple linear relationships is established between the response variable and the independent variable during a normal and a trial conditions. These two regression lines are then tested for their equality of slopes using analysis of covariance (ANCOVA). The separation between the regression lines, as measured by the difference of their intercepts, quantifies the change in the process performance. Unlike other methods, this method is the least affected by variables that are not of intrinsic interest and hence is more reliable. This method can be used in a two-stage two-product system and also in a multi-stage multi-product system. Although the recommended method of analysis is not as efficient as one involving a formal design of experiments using analysis of variance (ANOVA), it is nevertheless of significant practical merit for analysis of plant data or where the design of experiments would be too time consuming or difficult.

There are no comments yet on this publication. Be the first to share your thoughts.