Correlation and regression analysis in the x-ray fluorescence sorting of a low grade ore


The benefits of pre-concentration using sensor-based sorting have been widely reported with the greatest potential impacts on low grade and high tonnage operations such those mining copper porphyry deposits. Due to the non-selective nature of bulk surface mining methods, significant quantities of waste misreport to the concentrator reducing the feed grade, consuming energy during comminution and increasing water usage. In addition, significant metal misreports to the waste dump that represents a production loss and contributes to the metal content and therefore liability of the waste dump. The two main barriers to broader application of sorting are limitations in processing rates and the inability of sensors to accurately discriminate ore from waste. The present study is focussed on the latter and demonstrates the development of a more intelligent and more accurate sensing system. The X-ray fluorescence (XRF) technology was examined to explore its potential to pre-concentrate a low-grade copper ore through discarding barren particles. Simple linear correlation between the copper grades estimated using XRF and the copper grades obtained through chemical analysis did not meet the sorting requirement. Improved linear correlation and multiple linear correlation analysis were introduced in the study and better sorting results were obtained due to stronger correlations.

This paper may be downloaded from here with the authors' permission.

Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.