Beginners guide to thickeners

Particle properties affect the degree of separation that can be achieved with sedimentation. Size, shape, surface chemistry, and solids density all influence this process. These properties must be considered when selecting a rise rate, solids loading, and bed depth for a thickener. Rise rate parametrizes the area required to recover the design flow or solids loading, the design dry tonnage. In high rate thickener (HRT) designs the rise rate or solids loading will dictate the required tank diameter for any given throughput. In high compression (HCT) and paste (PT) applications the bed depth or the mass of solids in the thickener must also be selected to ensure the desired underflow density from the thickener is achieved. 

By adding flocculants and increasing the depth of the solids bed in thickeners, higher degrees of separation can be achieved. At low concentrations solid-liquid mixtures behave like a liquid (described primarily by a viscosity) and at high concentrations they can be better described as a paste or plastic with yield stress the best descriptor. In a small region of the transition from liquid to paste the pulp displays a rapid increase in yield stress.  At high levels of separation the increase in yield stress may reach the limit of pumpability or greatly increase the rake torque, which in-practice limits the degree of separation achievable.

Despite the simple structure of a thickener it is quite difficult to see or measure what is happening inside. Many thickener installations operate with water recovery below design limits largely through conservative control and low prioritization of thickener optimization.

The operation of the thickener must accommodate fluctuations – control for disturbances – in the process feed stream to deliver a clear overflow and thickened underflow. This control is achieved by manipulating both the flocculant addition and the rate of underflow withdrawal (typically pumping). The flocculant addition should be adjusted to suit the solids input and is mostly affected by the processing rate. Simple systems with constant feed can use flocculant pump speed control, more advanced control systems utilize a ratio controller to maintain a constant “grams per ton” dose rate. Unfortunately, fluctuations in ore type can significantly change the dosage of flocculant required and require a feedback to the “grams per ton” constant.

The discharge rate of underflow slurry is manipulated to maintain a steady solids inventory in the thickener. It should be remembered that both the thickener rakes and the underflow pumps (or valves) transport the thickened underflow. The size of the solids inventory determines the bed depth and the residence time that solids will experience in the thickener. This, in turn, determines the underflow slurry rehology and the water recovery to overflow. Generally, operators directly manipulate the underflow pump speed to maintain a constant solids inventory; site process engineers determine the inventory set-point. 

Good control systems should be able to monitor the rate limiting variable and apply control strategies to maintain optimum conditions. Often this rate limiting step is defined by pumping capacity. Simple density control strategies using PID feedback loops (figure 2) do not allow for the dead time inherent in the system and may lead to cyclic loading and emptying with the consequent underflow variability. Better strategies (figure 3) keep the thickener inventory constant through feedback from either bed level or pressure, and utilize flow control to ensure that pipeline disturbances are never seen by the thickener. By controlling flocculation and solids inventory, the thickening process is stabilized and consistent, optimum; underflow density can be achieved through manual optimization. An advantage of inventory control is that inventory responds much faster than density to changes in underflow withdrawal and simple PID feedback can be effective. More advanced systems can use measurements of the solids feed rate to further improve the control system dynamics. Inventory can be indicated by a simple pressure transmitter located on the floor of the thickener, under the feedwell.

Feedwells and flocculant dosing control have a significant influence on thickener performance. Flocculation is controlled by varying the flocculant pump speed to achieve a consistent flocculant dosage rate per ton of dry solids feed. Care should be taken to ensure that flocculant dilution is maintained with changes in flocculant dosing rate. Bed level can be used as feedback to control the dosage (g/ton) set point, either manually or through a master PID loop with an output cascaded to the dose-rate (ratio) controller. In practice, with a fixed inventory (bed pressure) the bed expands or contracts depending upon the flocculation effectiveness.

Instrumentation

The measurable parameters that can be used for thickener control include thickener feed flow rate, feed density, underflow density, overflow clarity, bed level, bed mass, rake torque, rake height, solids settling  rate and underflow rehology.

Rake torque

Rake torque is the most fundamental instrument and an accurate indicator of when things have already gone really wrong! Under normal circumstances the torque is an indication of the force necessary to drive the rakes through the bed. Higher torque is an indicator of higher underflow density (Yield stress dependent) and bed depth. Increased inventory causes the rakes to interact with mud even at their tips which results in higher torques. Increased levels of coarse material can create islands that give uneven high torque levels (with peaks at 2 or 4 times the rake rotation frequency) and rotating masses or doughnuts also cause rake torque to increase (with cycles equal to the rake rotation frequency).Measurement can be by hydraulic pressure, load cells or torque arms mostly supplied by the OEM and are essential to prevent rake drive damage. Operating thickeners for extended periods at high torque levels can negatively impact on gearbox and bearing lifetimes.

Bed level

Bed level is a frequently installed instrument and can be installed and serviced during normal thickener operation. Instruments measure the height of the interface between thickened slurry at the bottom of the thickener and the clarified liquor at the top. The interface is not always a clearly defined transition and the transition point is determined by the instrument calibration. Bed level is not always reliable as probes can foul or be obscured by entrained solids. Several systems of varying cost and sophistication are available to detect this interface and their performance is affected by the process conditions:

1. Turbidity sensor, either at a fixed height or attached to a motorized cable spool.
2. Buoyancy-based electromechanical float system either static or on a motorized cable spool.
3. Submerged ultrasound transducer to sense reflections from the solid bed.

Bed level is best utilized as a feedback on flocculant dosing rate, rising levels require an increase in flocculant dosing and an increase in underflow pump speed if the underflow density is high. 

Bed mass

Bed mass is utilized to control thickener inventory. Bed mass is generally measured by a pressure transducers mounted at the bottom of the tank (figure 4). These pressure transducers’ are best located in the cone or close to the center of the thickener. The transmitter should be located in a flanged spigot with the sensing face flush with the thickener surface. The measurement is usually very reliable and indicates the amount of solids in the water column above the sensor.

Ore type

Changing ore type can have an effect from several perspectives. Lower solids sg will reduce the settling velocity and the underflow density. The presence of clays will significantly change the settling behavior – for the worse. Actions to compensate for the change can include pH changes (when allowed), flocculation rate change and coagulation can have beneficial effects.

Surface chemistry

The separation of concentrates by flotation typically means that all particles in the system have a high surface charge. This is generally detrimental to flocculation and settling where the best results are achieved when the surface charge is a minimum. Where it is not possible to change the surface chemistry coagulants and flocculants can be effective in improving settling rates.

Inventory - Introduction

Thickeners accumulate solids to allow both the residence time and the bed pressure necessary to achieve dewatering. To establish this solids inventory it is necessary to operate with the solids feed rate in excess of the solids withdrawal rate for some period during startup, then at steady state operation, the solids feed rate should match the withdrawal rate. To achieve this balanced operation it is either necessary to measure both flows with a high level of accuracy or -- more practically -- to measure the solids inventory in the thickener tank. 

Inventory - Measurement

Inventory indication is generally achieved with bed pressure transducers. When running at optimal density the normal operating point has the mud line within the cylindrical section of the thickener. This means that the rakes are fully covered by mud so the torque cannot give a timely indication of inventory. This is in contrast to low bed applications, such as concentrate thickeners where torque can sometimes be used manually to help indicate inventory.

Flocculant measurement

On-line measurement of the settling rate of a sample drawn from the feedwell is the fastest way to determine if the flocculant requirements of the feed have changed. An instrument called a “Clarometer” (figure 4) is available which provides an on-line settling rate measurement. This target settling rate cannot be calculated explicitly but is manually optimized based on the performance history. With stable feed types, turbidity and bed level are often used as indicators of insufficient flocculant addition. Over-flocculation can be indicated by the relationship between torque, density and inventory. Both of these methods are slow to indicate flocculation errors as they do not directly measure the behavior inside the feedwell.