New tools for mineralogical process monitoring - X-ray diffraction (XRD) combined with statistical data analysis, U. König and N. Norberg

Ore mineralogy defines the properties during processing and metal extraction. Fast and frequent monitoring of the mineral content and related process parameters brings value to the mining operation and material processing. Decreasing ore qualities and increasing prices for raw materials require a better control of processed ore and a more efficient use of energy. Traditionally quality control in mining industries has relied on time consuming wet chemistry or the analysis of the elemental composition. The mineralogy that defines the physical properties is often monitored infrequently, if at all. The use of highspeed detectors has turned X-ray diffraction (XRD) into an important tool for fast and accurate process control. XRD data and their interpretation make the difference in the identification of minerals, in describing their distribution in ore bodies and in their beneficiation during processing. The use of statistical techniques such as Partial Least-Squares Regression (PLSR) or Principal Component Analysis (PCA) on XRD raw data has been successfully trialled to determine raw materials and processed ores, in addition to well established methods for mineral identification and quantification. PLSR can be used to correlate process parameters directly from the XRD pattern without investigating the mineral content. Hidden information present in the XRD pattern but not determined with traditional methods can be used to improve the characterisation of the analysed materials. The practical use for the direct determination of process relevant parameters during exploration and ore beneficiation will be illustrated in several case studies from mining and ore processing. Keywords: Geometallurgy, mineral monitoring, X-ray diffraction, process efficiency