Thickener modelling - from laboratory experiments to full-scale prediction of what comes out the bottom and how fast

摘要

Predicting full-scale thickener performance, including the solids flux and concentration delivered from a thickener underflow based solely on laboratory-scale experiments, has long been the holy grail of thickener design and operation. A number of researchers have developed both thickener models and laboratory characterisation techniques to measure sedimentation and compressional properties of flocculated suspensions. Combining these to produce predictions of actual performance generally results in an under-estimation of the thickener solids flux, often by a factor of between 10 and 20. Consequently, a range of empirical methods and industry scalars has been developed to get around this discrepancy.Analysis of the reasons for the discrepancy shows that changes in aggregate structure in shear, due to inter-aggregate buffeting and shear induced by collisions with surfaces and rakes, causes the flocculated aggregate to change from a fractal to a denser non-fractal object as it progresses through the thickener. These changes are shear rate and solids concentration dependent and as such, very difficult to reproduce in the laboratory and then incorporate into thickener models.A method to quantify the time and shear rate dependent changes in aggregate structure is now available and a model has been developed that allows incorporation of this effect into modelling. The change in aggregate behaviour is incorporated through a shear rate dependent densification rate and final extent of aggregate densification. The latter parameter helps to define an upper limit in solids flux behaviour for a given solids underflow concentration. Using the new information, thickener models now predict a range of underflow solids flux outcomes between the upper (densified aggregate) and lower (undensified aggregate) limit for a particular underflow solids concentration, depending on the operational conditions. The difference in underflow solids flux between these two limits is significant and the actual outcome depends on the shear rate and time of shear, as well as total solids residence time in the thickener.The data indicate that for non-segregating flocculated suspensions, operational conditions that produce the optimum thickener underflow solids flux for a given flocculation condition can now be explored quantitatively without resorting to extensive pilot trials.