The design of baffles to reduce axial mixing in flotation columns

"Large-diameter flotation columns are subject to severe axial mixing which has reduced column recovery and has generally limited it to cleaner applications. It has been assumed that columns with diameter greater than 1 m should be baffled to reduce axial mixing, but little work has been reported quantifying the effectiveness of baffling. Tests were done on a 0.2 m by 2 m column fitted with baffles of different geometry. It was generally found that the inclusion of baffles in the column worsened rather than improved mixing conditions in the column. This is explained in terms of a model for the effect of poor gas distribution on the behaviour of the column; this showed that under typical conditions a 16% excess of gas flux in the one compartment of a two-compartment column could result in one of the compartments being ""dead"" (i.e. with no net flow of slurry through it), resulting in a substantial loss of performance. The quality of mixing in the laboratory column was only reliably improved when the baffle was raised so that its top was positioned above the froth-slurry interface level. It should be possible to increase column recovery substantially with this configuration. The improved residence time distribution also means that columns can be used more effectively in roughing and scavenging applications. IntroductionFlotation columns have generally been used in cleaner applications, starting with their dramatic success at Mines Gaspe(1). Their lack of applications as roughers and scavengers is probably due to their poor residence time distribution (RID) when compared to the conventional cells they would replace. Espinosa-Gomez et al,(2) recently used them in rougher applications but installed three columns in series, presumably to cope with this disadvantage.Dobby and Finch in their seminal papersl(3,4) on the scale-up of flotation columns were the first to quantify the impact of axial mixing on the performance of industrial flotation columns. Laplante et al.(5) collected further data from the literature for column diameter D, ranging between 5 cm and 183 cm, gas superficial velocity Jg ranging between 0.7 cm/s and 2.8 cm/s and per cent solids in the feed S ranging between 3% and 52%. These data were analyzed with reference to the axial mixing model using the boundary conditions for closed vessels (the latter are more appropriate for flotation columns but the model is more difficult to use because analytical solutions for the RID are not available). The data were correlated by:"