Froth flotation is the single most important means by which valuable minerals are recovered, upgraded and separated from unwanted gangue in minerals processing. While it has been employed in the mining industry for over a century, it remains to be a process with significant room for operational improvement and optimization.

Flotation is a multiphase process involving complex chemical and physical phenomena both in terms of particle and fluid (gas and liquid) dynamics. Yet industrial flotation circuits are often fitted with instrumentation that only allows stabilising control of tank levels, and gas and slurry flows - with fundamental physical and chemical parameters neither tracked nor responded to in real-time.

As part of IntelliSense.io’s Flotation Optimization, we provide Virtual Sensors for some of these unmeasured variables that directly impact the performance of each flotation cell/bank, and therefore the overall flotation circuit and ultimately, the entire plant.

Figure 1: Virtual Sensors part of the Fluorite ‘20 Release of the IntelliSense.io Flotation Optimization Application.

Gas Hold-up is defined as the volumetric fraction of the slurry in a flotation cell that is displaced by the gas phase (Cortes-Lopez, 1998). Or, the volumetric fraction of gas bubbles in the flotation slurry. It is well-known that the kinetics and carrying capacity of a flotation cell - and therefore its overall performance - is a function of the Gas Hold-up. This is shown in Figure 1 above. Gas Hold-up is a parameter used in the design of flotation cells, but has as-yet been underutilised in the dynamic optimization of operating flotation cells.

Similarly, the mean Gas Bubble Size in a flotation cell directly impacts not only the Gas Hold-up, but - along with it - the bubble surface area flux within the pulp phase. That is, the gas-liquid surface area moving up through the flotation cell. Gas Bubble Size is impacted by several parameters, including frother dosage, gas flow and bubble generation mechanism. Finer bubbles are generally preferred as they provide significantly more bubble surface area (per m3 of total gas flow) for the increased probability of particle contact. However, in some fast kinetic flotation duties, small bubbles can negatively impact froth mobility and hence mass pull and valuable mineral recovery. Therefore, the ability to measure this key flotation parameter, in real time using our Virtual Sensor, allows operations to target the ideal mean Gas Bubble Size for their specific needs by controlling the various parameters affecting this variable.

The ranges of these variables - Gas Hold-up and Gas Bubble Size - that lead to desirable performance can be determined, and thereafter they can be tracked in real-time to ensure that the flotation circuit is operated in such a way that they remain within these desired bands.

These two Virtual Sensors, delivered using a combination of computational fluid dynamics (CFD) and machine learning models, provide much-needed insight into the real-time performance of the cells/banks/columns in a flotation circuit. This not only enables better root cause analysis, but allows operators to make better decisions when performance is not what they want it to be.

Cortez-Lopez, F. (1998). ‘Design of a Gas Holdup Sensor for Flotation Diagnosis’, Master of Engineering Thesis. McGill University, August 1998. National Library of Canada.