The Science of Bed Level and Rheology Measurements for Thickener Control

OVERVIEW OF THE PROCESS FLOW

Introduction to the Process

In mineral processing operations, the solid-liquid separation process carries consequences that extend well beyond tailings management alone, and its performance directly constrains the economic and environmental sustainability of a mineral processing operation. Historically, as Pullum (2007) notes, conventional tailings disposal involved pumping low-concentration slurries to large catchment dams, although this approach is generally no longer acceptable, as environmental and economic imperatives are prohibitive. The resulting paradigm shift toward thickened paste has remarkably altered pipeline behaviour. Consequently, the dewatering circuit has evolved into a more complex process, demanding the reliable measurements and monitoring of key variables for effective control and optimisation.

In this article, we will examine the limitations of the conventional approach, its assumptions, and the evidence for the need for additional data points to manage the circuit more efficiently. We will probe into the cost of not measuring bed-levels and slurry rheology, how the density-only control philosophy can be insufficient for thickening optimisation and managing the underflow pumps, limitations of laboratory rheology [Sun et al, 2026], how finer particles in clays can produce considerably different underflow rheology even when density stays the same [Bower et al, 2026], and how having a conservative approach can produce inefficient outcomes.

Understanding where these unaccounted variables sit requires a clear picture of what is happening inside the thickener and in its underflow line.

Susceptibility of the Process

The primary dewatering stage in mineral processing typically occurs in thickeners. Inadequate thickener control can cause the underflow solids concentration—and therefore the rheological load on the underflow pumps—to deviate abruptly from design limits. When the settling rate declines, the overflow carries a higher concentration of fine solids, while solids inventory accumulates in the thickener, raising the risk of rake bogging and underflow pump trips.

Moreover, an underflow that develops a high yield stress may exceed the available pump capacity. Overburdened pumps that cannot sustain the required discharge pressure allow the slurry to deposit in the pipeline, leading to line blockages from loss of consistent motion.

ISSUES WITH THE CURRENT CONTROL MEASURES AND ANALYSIS

Conventional measurements and their limitations

Most operators monitor the bed pressure, rake torque, and occasionally overflow turbidity for the thickener’s operations and control. Each captures a different aspect of thickener performance, yet none fully reveals the overall settling behaviour inside the tank. Bed pressure, under a set of assumptions, can sometimes estimate the total mass of the inventory, but it cannot denote the vertical distribution.  For some applications, the rake torque remains insensitive to moderate accumulations before spiking suddenly at high loads, leaving operators little time to respond. By the time the overflow turbidity registers higher solids, fine solids have already escaped the tank and contaminated the process water circuit, making it an inherently lagging indicator.

Beyond the thickener itself, the same information gap extends to how most operators manage slurry once it is pumped from the thickener tank. Most mineral processing sites monitor pressure, temperature, flow, and density—measurements that are necessary but insufficient for characterising the slurry's resistance to flow and its inter-particle bonding strength. Even though the site operators are often aware that a critical variable is missing from their control strategy, one that could help improve their operations and reduce the risk of process upsets, they do not have access to its solution – timely rheological data.

When rheological measurements are performed on underflow or tailings slurries, the appropriate constitutive model is the Herschel–Bulkley equation, which separates the true yield stress from non‑Newtonian viscous effects.

With missing information on these rheological variables, the operators are constrained to operate under proxies and alternatives, with certain limitations such as:

Offline (laboratory) testing of slurry rheology:

Unfortunately, offline testing rarely reflects true in-process conditions as temperature, shear rate, mixing intensity, and sample chemical structure change once the slurry is removed from the pipeline and the time taken for analysis. The equipment complexity in the laboratory and skill level required of the technician, for both the test work and data interpretation, is high, and retention of that capability can be challenging.

Any sample transported to a laboratory can undergo an unaccounted shear history during extraction and handling, so even the most careful measurement protocol makes it very challenging to fully recover the in‑process rheological signature.

In a recent study, Sun et al. (2026) showed that in some samples, the true in-situ plastic viscosity was underestimated by a factor of nearly 20 when measured after standard sampling, transportation, and homogenisation — a discrepancy large enough to fundamentally undermine any process decision based on that data.

Density/Solids Concentration Proxy:

Density is at times used as a proxy to understand the viscosity or overall rheological properties of a slurry, such as:

Density vs Viscosity: A common misconception is that higher density leads to higher viscosity. Oftentimes, operators rely on slurry density for impromptu corrective measures, such as dilution of the slurry. This approach can be counter-productive as a minimum viscosity needs to be maintained to keep material suspended in pipelines, and diluting the slurry can result in a reduction in viscosity, which can result in pipe blockages. Similarly, clay-rich or over-flocculated slurries can exhibit a high viscosity, independent of solids concentration, leading to blockages and rake-bogging in thickeners unexpectedly.

Density vs Yield Stress: Two slurries at identical density can have dramatically different yield stresses depending on particle-size distribution, mineralogy, clay content, temperature, and flocculant response.

Figure 1: The samples for Copper and Tin Tailings, with similar density but completely different rheology. Samples A (left) and B (right) have a higher yield stress than their counterparts. Density only measures mass per volume, not inter‑particle bonding strength. Similarly, the bed height may be different even at a similar density. Source: Bower et al. (2026)

Conservative sizing can be a coping mechanism:

Safety margins throughout the process, such as larger pipes, higher-rated pumps and more conservative pipeline velocities, might theoretically be an alternative approach. The concern is that oversizing to compensate for rheological uncertainty creates its own problems. Slurry transport lines can fail due to over-conservative sizing — if velocity is too low, it can cause pipe blockages due to settling and sedimentation.

As demonstrated by Pallum (2007), the stratified paste flows, which represent most real-world operations once shear is applied, the pressure drop becomes essentially independent of pipe diameter beyond a certain size. Simply installing a larger pipe does not create the expected hydraulic margin and may even promote settling by reducing shear rates that help keep particles entrained. Also, for many slurries, there exists a minimum transport velocity below which solids settle and form stationary beds, leading to blockage.

Inversely, running the operations at its maximum capacity, to push more tonnes, without the necessary upgrades to its equipment, can have a toll on the underflow pumps and slurry transport pipes due to excessive wear. Pushing more tonnes precariously close to the equipment’s capacity, coupled with changes in ore chemistry and, in turn, unaccounted slurry rheology, can cause costly process upsets such as lower equipment life and production stoppages.

Continuous, online slurry rheology can track rheological parameters over time, helping operators identify developing trends that signal an approaching process excursion and allowing corrective action to be taken before normal operating limits are exceeded.

NEW SOLUTIONS: CONTINUOUS BED LEVEL AND ONLINE RHEOLOGY DATA

What Bed Level Measurement Reveals

The behaviour of a thickener can be best understood by examining the distinct settling zones that form within a thickener.

The bed level in a thickener comprises four settling zones: (A) clarified liquor, (B) free-settling, (C) hindered settling/mud (often near the “gel point”), and (D) compression/heavy mud, where rakes assist in consolidating the settled solids and facilitating the release of trapped entrained water. Each zone plays an important role in the sedimentation and compaction process. Hence, knowing where these transition zones sit at any given moment can help with valuable insights on how a thickener can be controlled.

Figure 2: Schematics of a thickener tank

In the free-settling zone of a thickener, solids settle under gravity, concentration increases with depth and settling rates progressively slow. Flocculants enhance particle aggregation in this zone, and the height of the zone indicates dosing efficiency.

• Contracting Zone (interface approaching mud) → Faster settling → Reduce flocculant

• Expanding Zone (increasing separation) → Slower settling → Increase flocculant

Furthermore, the height of the mud bed correlates with solids residence time—a key factor in compaction and optimum underflow density. Adequate solids residence time in this zone improves dewatering efficiency.

Operationally, monitoring these bed levels helps ensure the underflow reaches target density before discharge and provides early warning of stratification loss or bed collapse.

What Rheology Measurement Reveals

Viscoelastic properties govern how efficiently material moves through every stage of a mineral processing plant. From milling, classification, flotation, thickening, pumping, and transport to final tailings deposition. When key rheological parameters, like viscosity, change unexpectedly, the consequences are tangible: poor particle size control in cyclones and milling, blocked slurry pipelines, rake over-torque events in thickeners, pump overloads, inefficient flocculant dosing, and sub-optimal tailings beach slopes and longer-term stability properties.

Viscosity describes how easily a slurry flows, and yield-stress tells you how much force is needed before a slurry starts to move. Many mineral slurries are considered non-Newtonian, meaning that their viscoelastic properties change with changes to various process variables and conditions.

In summary, with continuous on-site rheology data, the viscoelastic response from ore changes, particle size distribution, chemical addition, pipe velocity, and shear history can be tracked continuously.

Impact of Rheology on Dewatering Process

The intrinsic material behaviour of underflow slurry exemplifies the importance of measuring its viscosity and yield stress. Especially, as solids concentration increases, pipeline flow transitions from turbulence-dominated transport to regimes governed by particle interactions and non-Newtonian rheology, fundamentally altering transport mechanisms. Their flow behaviour depends more on rheology than on solids content. Particularly, with conditions such as over-flocculation and high levels of clays and fines.

With continuous viscosity and yield‑stress data, operators can control flocculant addition, underflow pumping rates, and rake torque far more precisely than before. This leads to more consistent paste quality, improved water recovery, and avoidance of rake overloads, problems that typically occur when rheological changes are not tracked.

Impact of Bed Level and Rheology on the Underflow Pumps

In thickeners, hydrostatic pressure is the gravitational force exerted by the slurry column. However, unlike pure liquids, slurries don't always transmit the entire tank’s load as hydraulic head because of how solids behave at different concentrations. The pressure at the bottom of the thickener does not equal the weight of all the slurry above. Only the hydraulically coupled portion transmits pressure at the bottom of the thickener. This is the key to understanding thickener behaviour because calculating the effective hydraulic head depends on what is inside the thickener and how it operates.

The compacted bed doesn't add its full weight to the hydraulic pressure because compressed solids transition to a mechanically supported regime characterised by direct particle-to-particle contact networks. This structural framework enables the settled solids to transfer their gravitational load directly to the thickener floor through inter-particle friction and compressive forces, effectively decoupling them from hydraulic pressure transmission to the underflow. Consequently, while the upper inventory of hydraulically coupled solids contributes to the available Net Positive Suction Head (NPSH) at the underflow pump inlet through fluid-mediated pressure, the lower compacted bed—despite representing significant mass and vertical height—adds negligible incremental pressure because its weight bypasses the fluid phase entirely.

This dual-regime behaviour fundamentally challenges the simplistic application of hydrostatic principles (P = ρgh) to thickener systems, as the effective pressure-transmitting height is not the total slurry column but rather only that portion where solids remain in suspension without forming load-bearing particle networks. Understanding this distinction is essential for accurate prediction of underflow pump inlet pressure conditions and, principally, for understanding thickener control strategies in mineral processing operations.

The mud bed height provides operators with the fundamental geometric parameter required to predict a thickener’s hydraulic head. The rheological properties of the slurry will determine what pressure you end up with at the pump inlet.

High-density underflows, laden with fine particles, typically exhibit non-Newtonian behaviour characterised by shear-dependent viscosity that determines the magnitude of frictional losses as the slurry moves through piping, bends, and valves. These viscosity-dependent flow-induced pressure drops can be imperative, as they can consume a sizeable portion of the available static head.

Consequently, operating decisions based solely on density-based calculations can grossly overestimate the available NPSH. Operating under inadequate NPSH causes the liquid phase to vaporise within the pump, creating cavitation that generates destructive shock waves, rapidly eroding impeller surfaces, inducing severe vibration, and ultimately leading to catastrophic mechanical failure.

AVAILABLE TECHNOLOGY AND CLOSING NOTES

Instrumentation in Practice: Science of Bed Level and Rheology Measurements

The case for combining continuous bed level data with continuous rheology measurements in mineral processing is not simply additive but synergistic. Each measurement addresses a different layer of the same operational problem. Combined, these two measurements give operators a more complete picture of thickener performance.

Bed level data informs operators what is happening inside the thickener — how material is settling and whether flocculant is being used effectively. Rheology data informs operators what is leaving the thickener at the underflow — whether the discharged slurry has reached the target viscosity, whether the yield-stress is within operational limits of the pumps, and whether it is about to drift into a condition that risks downstream blockages or compels dilution. By tracking rheological trends, operators can make deliberate adjustments with greater confidence and keep the process within the safe operating limits of the site's infrastructure.

Figure 3: A process flow for all the available sensors and data points available to optimise dewatering in mineral processing operations.

Two sensors have been developed by PLA Process Analysers to deliver these complementary data streams directly on‑site: for bed level, the SmartDiver®; for rheology, the ViscoSight™.

To implement bed level data, the SmartDiver® provides continuous real-time measurements of various bed levels within a thickener. The SmartDiver's® sensor takes readings of changes in solids concentration as it descends into the thickener and charts the clarity, interface, mud levels & heavy mud levels. The data from the SmartDiver® can interpret complex thickener behaviour into actionable insights.

ViscoSight™ is a side-stream rheometer that provides automated, continuous viscoelastic properties of non-Newtonian slurries. Using the proprietary RheologiX™ control algorithm, it calculates the relationship between shear stress and shear rate, which in turn fits the data to a rheological model and then outputs the key non-Newtonian parameters needed for process control. ViscoSight™ measurements can provide the data needed to manage underflow pumpability and to respond confidently to changes in thickener discharge conditions.

Conclusion

The measurement gap of not factoring in the true physical state of the thickener and underflow slurry carries quantifiable operational consequences and monetary implications.

This gap can be closed with continuous and reliable data, representative of the processes – bed level and rheology measurements. Bed levels quantify how the thickener is operating in the settling–compression continuum; rheology quantifies how the underflow slurry behaves.

Neither system eliminates the complexity of the dewatering process in mineral processing. What they offer is essential data — internal bed structure and continuous, representative rheological properties — from which informed decisions can be made.

References

L. Pullum (2007): Pipelining Tailings, Pastes and Backfill. Paste 2007: Proceedings of the Tenth International Seminar on Paste and Thickened Tailings, Perth.

Sun, YS, McGowan, Byrne, Parkinson, Kaminsky, DeJong (2026): In situ measurement of fluid fine tailings rheology. Paste 2026: Proceedings of the 28th International Conference on Paste, Thickened and Filtered Tailings, Australian Centre for Geomechanics, Lisbon.

Bower, M, Matinin, Bonneau (2026): Evaluation of rheology-modifying and conventional flocculants across various tailings through a dynamic thickening test. Paste 2026: Proceedings of the 28th International Conference on Paste, Thickened and Filtered Tailings, Australian Centre for Geomechanics, Lisbon.