Skip to main content

Increasing yield on Tungsten and Tantalum ore production by means of advanced and flexible control on crushing, milling and separation process

Periodic Reporting for period 2 - OptimOre (Increasing yield on Tungsten and Tantalum ore production by means of advanced and flexible control on crushing, milling and separation process)

Reporting period: 2016-06-01 to 2018-02-28

The main objective of the OptimOre project is to optimize the crushing, milling and separation ore processing technologies for Tungsten and Tantalum mineral processing, by means of improved fast and flexible fine tuning production process control based on new software models, advanced sensing and deeper process physical study increasing yield in 7-12% on the current best production processes, increasing energy saving on a 5% compared to the best available techniques.
The project was divided into work packages that covered the whole mineral processing domains but also the management and scientific coordination technology watch and intelligent control and design software development, system integration and validation and dissemination and exploitation objectives.
OptimOre aimed to set up the starting point of the project with the most advanced SoA technologies available. In terms of mineral processing, a full minerals characterization and the development of new models for crushing machinery and to address high precision milling models which can predicate product with a suitable mineral liberation degree. New improvements on gravity separation equipment modelling, commonly used for tungsten and tantalum processing, to develop new gravimetric separation process models and to develop and evaluate a predictive approach for froth flotation equipment and reagent regimes commonly used for the beneficiation of tungsten and tantalum ores, were also important objectives. All the process work packages had to be integrated and for this, the design of an expert control system for the tungsten and tantalum crushing, milling and separation processes, reducing energy consumption and increasing the yield was developed. The integration and validation was carried out in three main stages: (1) In a virtual simulated environment using proprietary software and expertise; (2) In a pilot test plant facility; (3) In several companies in the mining sector in Europe. Finally, the last objective was to ensure that the dissemination and exploitation strategies and project-wide training, communication and dissemination activities reached the target audience.
All significant research activity and document has been studied in order to identify and analyse any relevant technological development and innovation. Based on the collected data and the usage of mineralogical quantification, new comminution models were developed. The performance of these new models in terms of product predication and energy consumption performance have been analysed and applied in real mining sites. Using control algorithms during crushing stage, the performance increases up to 22.5%. For milling process, the energy consumption has been optimised until reached a maximum of 15% reduction. The new mineralogical model for gravity separation determines size/density classes from QEMSCAN false colour images and determines mineral recovery from size/density class recovery. Based on tungsten ore characteristics, an artificial colour vision system has been also developed using a scheelite sample.
A description of the magnetic separation process using different devices has been applied for simulation approaches. The presented modelling resulted in flexible functions and they can be readily tuned and altered to suit a variety of different magnetic separation processes or materials. The improvement on froth flotation separation is achieved through a reduction on recovery predication errors, from 3% until reach only 0.45%.
In terms of an integrated system involving an expert control system, it has been demonstrated at simulation level and in a laboratory pilot plant involving three of the five technologies researched in the project – milling, gravity separation and flotation. Crushing and magnetic separation were demonstrated in independent facilities. The benefits of the project outcomes were also demonstrated at a tungsten mine in Austria. As a result, although this plant is already highly efficient, further improvements in yield and energy savings were achieved. Similarly, it was possible to make modifications at other tungsten mine in UK to improve the degree of recovery from the gravity separation section. The tantalum mine site is recently in operation. Modification in the proposed flowsheet, to incorporate new OptimOre suggestions, is expected to yield improvements in efficiency.
An improved understanding of how tungsten and tantalum minerals are hosted in the rock matrix and how the ore minerals can be liberated has been achieved with the research carried out so far and is likely to modify established thinking on the design of the comminution stages of their processing. At the same time new approaches to sensing the mineral processes have been explored, which have the potential for enabling new approaches on process control. The control processes themselves have also being readdressed to develop intelligent decision support systems rather than ones based on mathematical functions. These incorporate best practice from the mining and other industries.
After the review of a number of processing plants visited, it can be noticed that there exists a noticeable gap between the SoA on Knowledge and SoA on Practice. One of the most evident lack of knowledge, and as a consequence, one of the potential field of progress beyond the SoA, is the usage of Process Mineralogy as a source of information to be applied in modelling and control of unitary processes and also in the evaluation the global efficiency of the plant.
As a contribution in the frame of this project, the liberation analysis conducted in the area of coarse comminution has been applied. The work shows that compressive crushing that occurs in a jaw, gyratory or cone crusher causes liberation. For milling stage, the mineralogical quantification was also integrated into the comminution model, as a way of predication of the product size taking into account the mineral composition as the particles size decreases.
In terms of gravity methods for mineral recovery, it was found to be able the predication recovery on a spiral separator at the Drakelands mine processing plant with an average error of 3.1%, well below the required 10% target of the project. New magnetic and froth flotation models were also based on mineralogical quantification, contributed to the best practice and optimal conditions for mineral recovery.