Improved Measure of Ultimate Flotation Performance

This project was concerned with the development and validation of an improved laboratory flotation test designed to "fingerprint" the ultimate flotation performance that can be achieved for a particular coal feed. It is important to note that the aim was to measure a characteristic of the coal feed itself, not of the particular type of flotation device employed.

The benefit to the industry comes from being able to measure true recovery loss, and hence minimise this through process improvement. It is estimated that 18 million tonnes of coal per annum are treated in Australia by flotation. A one percent improvement in flotation yield translates to approximately ten million dollars of export revenue. It is not possible for plant operators to be aware of their individual losses unless a suitable "ultimate" flotation performance test is available.

The problem with the existing method (Australian Standard AS4156.2.2-1994, and its variants) is that it is highly operator sensitive, very time and hence cost intensive, and totally inaccurate unless undertaken by a highly skilled operator. The problem with the method is twofold; (1) slime entrainment in concentrate, and (2) time requirement to undertake the test. A new method needs to be able to clearly and simply delineate true flotation from entrainment flotation. In practice, a 1% clay slimes contamination to product translates directly to a 1% concentrate mineral matter increase. In modern high-intensity flotation practice (Jameson Cell and Microcel), entrainment is eliminated through the use of clean counter-current wash water through the froth phase.

[It is important to note, that prior to the commencement of this project, the only recognised coal flotation characterisation test was the BHP tree method. The Australian Standard method represents an abbreviated form of this test and gives a wide variety of results. Thus the BHP method was employed as the "benchmark" against which to evaluate the new method.]

This project developed and investigated a modified bench scale mechanical flotation cell apparatus. The modified apparatus sought to solve the problems with the standard method by utilising wash water and a continuous tailings recirculation technique. A workable method was developed and tested for reproducibility using three independent operators and two independent laboratories, all of whom had nothing to do with the development of the procedure.

The procedure developed does not perfectly reproduce the "ultimate flotation response" however the results show that the main areas of interest on the yield-ash curve are determined with the following key advantages of the method:

• The test procedure is fast compared to the traditional BHP tree method. One test takes a total duration of approximately two hours to complete.
• The analysis requirements are greatly reduced, with only 12-15 samples requiring analysis compared to more than 50 in the case of the BHP tree
• The test method requires only 300-500g of coal, which means it can be employed on small samples such as those obtained from borecore samples.

The project was led by Dr Bruce Atkinson and Mr Steve Blanshard of Jetflote Pty Ltd, and the project work was primarily undertaken by Ms Tracey Rose of BHP Research - Newcastle Laboratories, with support from Mr Ray Keast-Jones.