It might be physically impossible to peer inside an industrial mixer-settler and see what is going on — and costly to tinker with an operating solvent extraction (SX) plant to enhance its capabilities — but the Parker CRC for Integrated Hydrometallurgy Solutions has developed tools that effectively allow operators to do both.
The tools allow operators to visualise what is happening inside the reactors used in SX plants and to experiment with potential improvements without interrupting production.
The industry-funded AMIRA International P706 project saw scientists establish physical and computational fluid dynamics (CFD) models showing how different designs and operating conditions affect fluid flow patterns and droplet size distribution within mixer-settlers and pulsed columns.
CSIRO Minerals emerging technologies manager Dr Martin Houchin says it is the first time two-phase CFD models for such reactors, incorporating the droplet break-up and coalescence stages of the SX process, have been developed.
Researchers at CSIRO Minerals — working through the Parker Centre — are now building on that work with a new project, dubbed SXT (solvent extraction technology), which aims to translate the fundamental understanding produced in the P706 project into practical outcomes for SX plant operations (see breakout).
Dr Houchin says that already some aspects of P706 have been put to commercial use, with at least one major copper producer gaining positive results after using modelling work outcomes to position mixer-settler components known as ‘picket fences’.
The project has also allowed researchers to optimise droplet size distribution. Together with fluid flow patterns, this plays an important role in the efficiency of the SX process — a process that sees an aqueous solution carrying dissolved metals from leached ore mixed with an organic solution carrying an extractant. The extractant bonds with the targeted metal or metals and thus separates and purifies metals in the original aqueous solution.
Mixer-settlers and pulsed columns are used to form and disperse droplets to encourage efficient mass transfer of the metal or metals from the aqueous phase or vice versa into the organic phase before further steps are used to recover the metal.
“What the research allows us to do is optimise the droplet size distribution for a particular solvent extraction operation,” Dr Houchin says. “The smaller the droplets, the faster the metal-mass transfer between the two phases.
“But if the droplets become too small they don’t settle out into the two phases as efficiently as they should — so operators can, for example, lose some expensive organic solution through entrainment in the aqueous phase.
“We are in a situation now where we effectively understand how to manipulate that droplet size distribution through design and operational parameters.”
Scientists involved in P706 built physical models of a mixer-settler and a pulsed column — operational pilot plants — from transparent Perspex™ and glass. This allowed laser light to be beamed into the models to measure fluid flow vectors and droplet size distributions.
“We then developed theoretical mathematical CFD models to simulate the measured results,” he says. “After that, you can ask these models to predict behaviour under various conditions and verify those predictions by doing measurements on the physical models.”
As well as running clear fluid through the physical models, the project scientists experimented with copper and nickel solutions to measure the process of solvent extraction under varying conditions.
Teams of CSIRO scientists led by Dr Phil Schwarz at Clayton, Victoria, and Dr Chu Yong Cheng at Waterford, Western Australia, conducted the P706 research through the Parker Centre over the course of about two years.
Dr Houchin says the beauty of modelling is visualisation. “For the first time we are able to visualise and predict the impact of mixer design and operation on droplet size distribution, or design and positioning of picket fences on phase disengagement.
“CFD models allow us to visualise what’s going on and to perform as many ‘what if’ experiments as we like with the mathematical models. We can then verify the predicted results by running modifications on a physical model before introducing them to the plant.
“This allows many more operational scenarios to be considered and greatly reduces the risk of implementing changes to the plant.”
This article first appeared in Process (October 2007) — a publication of CSIRO’s Minerals Resources sector.
Dr Martin Houchin, CSIRO Minerals