Reflectance spectrum analysis of mineral flotation froths and slurries

Abstract

The global demand of mining products has increased during recent years, and there is pressure to improve the efficiency of mines and concentration processes. This thesis focuses on froth flotation, which is one of the most common concentration methods in mineral engineering. Froth flotation is used to separate valuable minerals from mined ore that has been crushed, mixed with water and ground to a small particle size. The separation is based on differences in the surface chemical properties of the minerals. Monitoring and control of flotation processes mainly relies on the on-line analysis of the process slurry streams. Traditionally, the analysis is performed using X-ray fluorescence (XRF) analyzers that measure the elemental contents of the solids in the slurries. The thesis investigates the application of visual and near-infrared (VNIR) reflectance spectroscopy to improve the on-line analysis of mineral flotation froths and slurries. In reflectance spectroscopy the sample is illuminated and the spectrum of the reflected light is captured by a spectrograph. The main benefits of VNIR reflectance spectroscopy with respect to XRF-based analysis are the relatively low cost of the equipment required and the easy and fast measurement process. As a consequence, the sampling rate of the reflectance spectrum measurement is radically faster than in the XRF analysis. Data-based modeling is applied to the measured VNIR spectra to calculate the corresponding elemental contents. The research is conducted at a real copper and zinc flotation process. The main results of the thesis show that VNIR reflectance spectroscopy can be used to measure temporal changes in the elemental contents of mineral flotation froths and slurries in the analyzed process. Especially the slurry measurements from the final concentrates provide accurate information on the slurry contents. A multi-channel slurry VNIR analyzer prototype is developed in this thesis. When combined with an XRF analyzer, it is able to measure the slurry lines with a very fast sampling rate. This considerably improves the monitoring and control possibilities of the flotation process. The proposed VNIR analyzer is adaptively calibrated with the sparse XRF measurements to compensate for the effect of changes in other slurry properties. The high-frequency slurry analysis is shown to reveal fast grade changes and grade oscillations that the XRF analyzer is unable to detect alone. Based on the new measurement, a plant-wide study of the harmful grade oscillations is conducted in order to improve the performance of the flotation process.