Adaptation of Coal Grain Analysis to Improve Flotation Yield Estimation

This project involved sampling of full-scale flotation circuits at each of six different coal preparation plants, each subject to three different reagent dose rate regimes (three sampling runs per site). Coal Grain Analysis (CGA) was undertaken on feed, concentrate and tailing for each run. This project was an extension of an earlier project C24045, where three different sites had been previously investigated in the same manner. This project incorporates the data from C24045 and encompasses all nine sites.

The primary objectives of this project were to:

• Employ CGA to characterise the flotation response at a particulate level;
• Compare the CGA data and CPP flotation circuit performance to laboratory sequential (tree) flotation and fine float and sink test tests on each flotation feed; and to
• Evaluate CGA as a means to model flotation in a more reliable manner, in order to facilitate improved flotation yield estimation.

The key findings of this project include:

• Characterising flotation response at a particulate level, in terms of maceral and mineral group response, provides deep insight to the flotation process.
• Vitrinite and inertinite each behave very differently in flotation. Generally, vitrinite recovery increases with increasing particle size, and vice-versa for inertinite. The flotation rate of inertinite is generally significantly lower than that for vitrinite.
• As the flotation process relies on air-particle contact and adhesion (hydrophobicity), and the hydrophobicity depends on the nature of the coal and mineral matter in each particle, CGA is able to provide a fundamental approach to understanding coal flotation.
• Various different forms of modelling are possible using CGA, including providing a direct quantitative link to pseudo-density models.
• The use of CGA in flotation modelling overcomes the severe limitation of all alternative model approaches, in that the CGA models are based on the actual particle compositions, and not on any inferences such as flotation feed ash value.
• The most promising form of CGA model is that which was investigated in associated Project C27033, where an adaptation of CGA was employed to measure the surface composition of particles, rather than the volumetric composition, and relate the surface composition to flotation performance. The particular benefits of that approach are that all CGA data are employed without resorting to any form of particle class-grouping; and the only information that needs to be pre-procured are the flotation rate constants for each of the pure components: vitrinite, inertinite, liptinite and minerals. This project generated the pure component rate constants which could then be used by Project C27033.
• For any given flotation feed that has been characterised by CGA, the modelling methodology identified in this project and C27033 provides a quantitative basis upon which to estimate flotation yield for different flotation devices and different flotation circuit configurations. Project C27033 has specifically identified the notion of a threshold rate constant, which provides a quantitative basis to estimate the different operating performance of different flotation devices.
• Data presented in the Report provides a quantitative basis for understanding the impact of changes in reagent dose rate upon flotation rate for each of the feed components.
• The percentage of minus one micrometre material in flotation feed is often very significant. CGA does not measure particles smaller than one micrometre, and so this is a limitation of CGA which needs to be accounted for.
• The flotation feed washability measured by CGA appears to define the best possible ('ultimate') flotation performance that may be attained, although the CGA washability needs to be corrected for the minus one micrometre material.
• This project has demonstrated the application of CGA as a tool to provide a more fundamental basis for modelling flotation performance. The CGA data provide a very clear picture of where valuable product coal components are being lost to tailing, thus providing a tool for flotation cell and flotation reagent suppliers to target improved performance.