Flowsheet Design Using Flotation Rate, Selectivity and RTD Data

"A methodology is proposed for milling flowsheet design using flotation rate, selectivity and residence time distribution data. The flotation rate data and the difference in flotation rates between the main minerals to be separated are determined from laboratory batch tests. The residence time distribution (RTD) data is calculated using the tanks-in-series model assuming each flotation cell is a perfect reactor. The rate, selectivity and RTD data can be used to calculate the optimum flotation cell number in a bank for rougher/scavenger in a commercial plant.This methodology is demonstrated in a case study of a nickel sulfide ore. It is shown that miniplant rougher/scavenger data can be closely predicted from a batch test using the proposed methodology.INTRODUCTIONIt has been the trend in the past several decades that large flotation cells are increasingly used. This is largely driven by the economics as ore grades are getting lower and the rate of ore processing must be increased in order to keep mining operations economically viable. The development and evaluation of 250 to 500 m3 WEMCO SmartCell and Outokumpu TankCell flotation machines are well in progress. There are many advantages in using large cells: fewer cells, less floor space, lower power cost (although this is now being questioned for fine particle flotation), simpler circuit, simpler control, and lower auxiliary equipment cost (Arbiter, 1999). As flotation cells are getting larger, the importance of cell hydrodynamics, namely, air/slurry dispersion, power intensity, froth transfer, residence time distribution (RTD) and short-circuiting are being fully recognized (Jonaitis, 1999; Weber, et al., 1999; Gomez and Finch, 2002). It is not the intention of this paper to discuss flotation cell scale-up. The intention is to answer what is the optimum number and size of cells in a flotation bank for a given throughput. Figure 1 illustrates this problem facing design engineers. The same mean residence time required for a given throughput can be achieved in a single large cell, or many smaller cells in a row. A few cells of large size may be more cost-effective but they may not be as effective as more cells of smaller size in achieving the optimum metallurgy."