From elements to particles - particle fate modelling

The use of mineralogical information to improve minerals processing operations has barely moved beyond trial-and-error. The main reason is the limited information applied – usually mineral contents, only. However, materials characterization has greatly advanced using techniques to describe the size, shape, and composition of individual particles systematically and within a reasonable time.

Therefore, the advanced characterization of raw materials should be fully applied to not only better understand but also predict the particle behavior in particle separation processes, including its uncertainties, as currently researched in the interdisciplinary Particle Fate Modelling research group.

---

Current developments

Data mining platform

We developed a system that allows for consistently storing detailed raw materials characterization data, collected from different analytical methods (e.g. XRF, XRD, Automated Mineralogy, and X-ray computer tomography). By bringing all data into this unique data mining platform we want to ease the complete evaluation of material streams.

Through an in-house developed front-end R interface users will in the future be able to import, visualize and interprete case study data anywhere in the world. All in all, the system will enable the user to access and handle complete ore characterization datasets in a structured way, regardless of the analytical method applied, the number of samples used, and the complexity of their relationships.

Modelling techniques for particle behaviour

We also developed various modelling techniques (e.g., with kernel density estimates) to understand and predict the process behaviour of individual particles in any particle separation device (e.g., flotation). All particle information collected is used in order to make predictive models truly acknowledge material complexity, to help identify suitable process parameters for individual feed materials, and to minimize costs of later metallurgical studies.

Precisely, our methods helps to

• understand the processing behaviour of complex materials;
• simulate complete processing plants with particle separation circuits tailored to the specific raw material;
• and model single particle flotation kinetics.

3D characterization of individual particles

So far, particle characterization has mainly progressed in 2D. However, complete geometric information that a 3D body such as a particle posses are hardly available. That is why, we also develop techniques to characterize the size, shape, and composition of individual particles in 3D. Using X-ray computed micro-tomography to do so, does not only allow us to collect detailed 3D particle particles but to also push the boundaries of mineral identification. The collected data should additionally be suitable for a direct transfer into our particle-based separation models.

---

Outlook

Our group continously works to expand its understanding and tools for minerals processing with special focus on material complexity.

Some of our current projects include:

• particle-based comminution prediction
• researching the influence of particle shapes on the flotation process
• optimization of flotation circuits based on single particles characteristics

---

Supported research activities

Current

• MultiDimFlot
• 2D3Dscopy

Finished

• AfK
• MoCa

---

Related publications

• Computing single-particle flotation kinetics using automated mineralogy data and machine learning
  Lucas Pereira, Max Frenzel, Duong Huu Hoang, Raimon Tolosana-Delgado, Martin Rudolph, Jens Gutzmer
  Minerals Engineering, Volume 170, 2021, https://doi.org/10.1016/j.mineng.2021.107054

• Mounted Single Particle Characterization for 3D Mineralogical Analysis—MSPaCMAn
  Godinho, J.R.A.; Grilo, B.L.D.; Hellmuth, F.; Siddique, A.
  Minerals 2021, 11,  https://doi.org/10.3390/min11090947

• A self-adaptive particle-tracking method for minerals processing
  Lucas Pereira, Max Frenzel, Mahdi Khodadadzadeh, Raimon Tolosana-Delgado, Jens Gutzmer
  Journal of Cleaner Production, Volume 279, 2021, https://doi.org/10.1016/j.jclepro.2020.123711

• Multidimensional characterization of separation processes – Part 2: Comparability of separation efficiency,
  Minerals Engineering
  Markus Buchmann, Edgar Schach, Thomas Leißner, Marius Kern, Thomas Mütze, Martin Rudolph, Urs A. Peuker, Raimon Tolosana-Delgado
  Minerals Engineering, Volume 150, 2020, https://doi.org/10.1016/j.mineng.2020.106284

• R as an environment for data mining of process mineralogy data: A case study of an industrial rougher flotation bank
  Nathalie Kupka, Raimon Tolosana-Delgado, Edgar Schach, Kai Bachmann, Thomas Heinig, Martin Rudolph
  Minerals Engineering, Volume 146, 2020, https://doi.org/10.1016/j.mineng.2019.106111

• Multidimensional characterization of separation processes – Part 1: Introducing kernel methods and entropy in the context of mineral processing using SEM-based image analysis
  Edgar Schach, Markus Buchmann, Raimon Tolosana-Delgado, Thomas Leißner, Marius Kern, K. Gerald van den Boogaart, Martin Rudolph, Urs A. Peuker
  Minerals Engineering, Volume 137, 2019, https://doi.org/10.1016/j.mineng.2019.03.026