Aggregate densification in the thickening of flocculated suspensions in an un-networked bed
Introduction
Thickeners are a simple continuous solid–liquid separation device widely used in the minerals, water and waste-water industries. The feed slurry enters through a pipe or launder into a central feedwell, in which flocculation is achieved and much of the momentum/energy of the feed is dissipated. On discharge from the feedwell, the solids are concentrated through sedimentation, hindered sedimentation and then consolidation in a solids bed, with thickened slurry then exiting through a discharge or underflow point at the base. The latter is often controlled to a fixed slurry rheology to allow predictable disposal in a tailings emplacement (Sofra and Boger, 2002). The released liquor rises and a low solids overflow exits via a weir at the top of the thickener.
Enhancement of throughput whilst at the same time still achieving high overflow clarity and the desired underflow rheology or solids concentration is a key aim of many industrial thickening operations. Optimisation of the dosage of polymeric flocculants is one way to enhance throughput via improved sedimentation, although this is sometimes at the expense of the underflow solids concentration, since the addition of flocculants produces a stronger aggregate network. The result of high flocculant dosing is a lower output solids for a fixed underflow slurry rheology. The manipulation of shear processes is also known to be highly effective at improving the sedimentation rate of aggregates (Gladman et al., 2005, Gladman et al., 2010a, Loan and Arbuthnot, 2010). Linking these effects to measureable parameters at a laboratory scale in a quantitative rather than qualitative manner has proven difficult and constitutes the focus of this work.
Dewatering theory along with one-dimensional phenomenological thickener dewatering models has been successful in describing dewatering trends in thickeners. This has led to the development of analytical laboratory based tools to generate dewatering parameters relevant to the modelling of thickening at an industrial scale (de Kretser et al., 2001, Landman et al., 1995, Lester et al., 2005, Usher et al., 2001). However, comparison of the predictions from these models with industrial scale data has shown discrepancies between the two and suggests flaws in the dewatering model (Usher and Scales, 2005). These discrepancies have been attributed to a number of factors, but for the sake of producing a correlation between laboratory and field studies, the difference was offset through use of an empirical parameter referred to as the “performance enhancement factor” (Gladman et al., 2010a).
The performance enhancement factor (PEF) relates to a relative increase in observed solids flux through a device versus the predicted flux to achieve a given underflow solids concentration. The observation for flocculated suspensions is that field based devices consistently outperform model predictions based on dewatering parameters extracted from un-sheared laboratory batch settling tests (i.e. the PEF is >1). These flocculated suspensions are typical of the minerals industry (exemplified here) although nearly all thickening processes involve the addition of polymer flocculant to increase aggregate size and simultaneously aid overflow clarity and improve aggregate sedimentation rates (which as noted, also results in improved throughput). Values for the PEF are observed to cover the range of 1–100, although more typically lie between 5 and 20 (Usher and Scales, 2005). The use of an empirical factor to describe an un-quantified parameter is highly unsatisfactory as it clearly does not aid the predictive capacity that theoreticians and practitioners alike may wish to achieve through modelling of such devices.
Shear processes in full-scale thickeners are ubiquitous, since nearly all thickeners have a raking mechanism. Traditionally, rakes were designed to transport material to the discharge point, as it is important that a continuous supply of sediment makes its way to the underflow in order to prevent both caking (locally high solids concentrations) and channel formation. However, it is known that raking also plays a role in enhancing dewatering by applying shear (Rudman et al., 2008). Many thickeners also employ dewatering rods that penetrate to the top of the thickener to aid the process although the design basis for these devices is more empirical than predictive. The expectation when shear is applied through a mechanical force such as a rake or through buffeting of aggregates in sedimentation is that the local pressure gradients produced will result in the expulsion of water from the flocculated aggregates and subsequent aggregate densification. Furthermore, as the aggregates decrease in size, the tortuosities around the aggregates also decrease (at fixed solids), leading to a decrease in the resistance to fluid flow around and an increase in the resistance to fluid flow within or through the aggregates. A significant amount of such dewatering is believed to occur, and given that particle buffeting as a result of flow and not just mechanical shear is important, it is possible that the enhanced dewatering can be achieved before the aggregates even reach the raking zone of the thickener.
Experiments that look to characterise enhanced dewatering behaviour as a result of shear have shown that there is an optimum shear rate beyond which any further increase is detrimental to aggregate densification (Gladman et al., 2005). This is envisaged as a trade off between densification and breakage of aggregates, the latter being contrary to achieving enhanced sedimentation rates. It is also possible that aggregate buffeting could lead to further aggregation although this is not considered likely herein since studies on polymer bridging show that contact driven aggregation stops very quickly once free polymer is no longer available in the system (Owen et al., 2008).
An aggregate densification model was developed by Usher et al. (2009) to provide a model framework for the development of analysis methodologies that will close the gap between laboratory predictions and industrial observations due to shear effects. The model provides an estimate of the expected performance enhancement for a given state of aggregate densification but does not consider the role of the rate of densification (i.e. it is a steady state and not transient model). Zhang et al. (2013) have since presented a transient densification model at steady state flux. Given that full-scale thickeners are permeability limited (there is not enough residence time to achieve significant compressional dewatering), the performance enhancement could be thought of as resulting from permeability enhancement. This work presents techniques developed to infer shear-induced permeability enhancement data from laboratory and pilot thickening studies. They are then applied to case studies where they are used to more accurately predict thickener performance. The work provides a quantitative background to the dewatering parameter extraction techniques, observations and modelling that aim to fill the gap between laboratory and full-scale work in the thickening arena (Grassia et al., 2014, Loan and Arbuthnot, 2010, Usher et al., 2009, van Deventer et al., 2011, Zhang et al., 2013).
Section snippets
Background dewatering theory
The origin of gravity thickening theory can be traced back to Coe and Clevenger (1916), who took into account the operations of settling tanks, where it was observed that the settling rate was dependent on the solids concentration. Thus, it was recognised that with increasing solids concentration, Stokes’ Law for settling of a single particle in a continuum was unable to describe the settling behaviour of mineral suspensions. Furthermore, the existence of distinct sedimentation and
Methodology
Two types of experiment are investigated herein to provide data on enhancement of the rate of thickening. The first involved continuous flocculation in a pilot-scale thickener that has been described in some detail in earlier publications (Gladman et al., 2010a, Gladman et al., 2010b). The second was a novel laboratory experiment involving the fluidisation of a flocculated suspension in the annulus of a Couette device, such that mechanical shear could be imparted in a uniform way whilst holding
Fluidisation analysis
Height versus fluidisation time data, obtained from the fluidisation rig, is depicted in Fig. 3 (top). Un-sheared R(ϕ) data of the flocculated Omyacarb 2 obtained from a batch settling test is presented in Fig. 3 (bottom).
Comparisons of the hindered settling function data obtained from the fluidisation rig to that from settling tests of the same material under the same flocculation conditions (depicted in Fig. 3) enables the extent of aggregate densification to be determined via Eqs. (5), (6).
Pilot thickener analysis
Along with the laboratory fluidisation rig, the pilot thickener was operated to analyse shear densification of aggregates. The steady-state solids concentration profiles for each bed height were used to determine the change in the hindered settling function, R(ϕ), along the height of the column and these were compared to R(ϕ) values calculated from batch settling tests of un-sheared samples taken from the pilot thickener feed.
For both cases, significant permeability enhancement within the pilot
Conclusions
In this paper, novel techniques have been assessed to quantify the extent and time dependence of shear enhancement within a pilot thickener using both data from the thickener and a laboratory based fluidisation rig. Work was conducted in the absence of mechanical shear in the case of the pilot-thickener work and both with and without mechanical shear in the fluidisation rig study. The level of solids permeability enhancement inferred as a consequence of aggregate densification was measured and
Nomenclature
- Dagg
scaled aggregate diameter (dagg/dagg,0)
- dagg
aggregate diameter (m)
- dagg,0
initial aggregate diameter (m)
- dthick
thickener diameter (m)
- f
solids flux (m s−1)
- g
magnitude of gravitational acceleration (m s−2)
- h
bed height (m)
- h0
initial bed height (m)
- hi
bed height at time i (m)
- hf
final bed height (m)
- L
fluidisation rig dimension (cm)
- Py
compressive yield stress (Pa)
- q
steady-state solids flux (m s−1)
- R
hindered settling function (kg s−1 m−3)
- R1
fluidisation rig dimension (cm)
- R2
fluidisation rig dimension (cm)
- t
time (s)
- tres
Acknowledgements
This work was conducted as part of AMIRA International Limited collaborative research project P266E: Improving Thickener Technology and supported by the Australian Research Council (ARC) under the Linkage Scheme (Grant no. LP0989733). The sponsors of this project are gratefully acknowledged. The authors also acknowledge the support of the PFPC (Particulate Fluids Processing Centre), a Special Research Centre of the ARC and technical inputs to, advice and operational help in conducting the pilot
References (42)
- et al.
Phenomenological model of filtration processes: 1. Cake formation and expression
Chem. Eng. Sci.
(2001) On boundary conditions and solutions for ideal clarifier-thickener units
Chem. Eng. J.
(2000)- et al.
Consolidation and aggregate densification during gravity thickening
Chem. Eng. J.
(2000) - et al.
Effect of shear on particulate suspension dewatering
Chem. Eng. Res. Des.
(2005) - et al.
The effect of shear on gravity thickening: pilot scale modelling
Chem. Eng. Sci.
(2010) - et al.
Experimental validation of a 1-D continuous thickening model using a pilot column
Chem. Eng. Sci.
(2010) - et al.
Effects of aggregate densification upon thickening of Kynchian suspensions
Chem. Eng. Sci.
(2014) - et al.
Time dependent batch settling of flocculated suspensions
Appl. Math. Model
(1990) - et al.
Using turbulent pipe flow to study the factors affecting polymer-bridging flocculation of mineral systems
Int. J. Miner. Process.
(2008) - et al.
Raking in gravity thickeners
Int. J. Miner. Process.
(2008)