Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Computational methods and strategies for geometallurgy
Luleå University of Technology, Department of Civil, Environmental and Natural Resources Engineering, Minerals and Metallurgical Engineering.ORCID iD: 0000-0003-4800-9533
2019 (English)Doctoral thesis, comprehensive summary (Other academic)Alternative title
Numeriska metoder och strategier för geometallurgi (Swedish)
Abstract [en]

At the interface of geology and mineral processing, geometallurgy is a powerful tool for enhancingresource efficiency. A spatial model that represents the ore body in terms ofmineralogyand physical properties is combined with a process model that describes the concentrationprocess. The performance of a given ore in the process is computed in terms of gradeand recovery of the mineral of interest in the concentrate, but also the presence of potentialpenalty elements and energy costs. The inclusion of ore performance indicators in a blockmodel yields a geometallurgical model that considers the variations in an ore body.Progress has been made in recent years to list and study different processing options interms of data requirements and implementation costs. While providing useful data, littleadvance was made to guide decision-making and to handle uncertainty. The objective has,therefore, been to develop, choose and validate computational methods that suggest optimaldecisions in the scope of geometallurgical strategies for an iron ore and a porphyry copperdeposit.The selected approach is based on an analysis of structure and regularity fromthe ore blockdown to the mineral grains. By selecting the appropriate mathematical tool for each scale,the dimension of the data is reduced and the different scales are then taken into account inmaking decisions. Methods introduced for dimension reduction include machine learningmodels, statistical models and spectral descriptors. The decision models rely on stochasticmulti-armed bandits which are a form of reinforcement learning. The presentation of thedifferent models proceeds by zooming in from coarse scale to fine scale then taking a stepback and analyze the implications. Data that was collected during sampling campaigns andindustrial plant surveys is used to design and verify the proposedmodels.iWith regard to the dimension reduction problem, results showed the method’s ability toclassify mineral textures and identify mineral phases with more than 90 percent accuracy onthe selected data sets of optical images and incorporate different physical properties into ageometallurgical ore type classification. Decision results showed that strategies in the case ofa feed grade control or when different ore types were identified, resulted in a twofold increaseof a reward function which is either Boolean (the product fulfills quality requirements ornot), or continuous (an economic objective). The cumulative value of the reward functionmeasured the optimality of a processing strategy. Quantitative methods were introduced toevaluate ore classification as well as geometallurgical strategies.The achieved results suggest the introduction of these computationalmethods in the practiceof geometallurgy. The increased knowledge of different ore type performances and appropriatemodels lead to optimal decisions for improved resource efficiency along the ore valuechain. This is achieved by bothmaximizing profit and decreasing environmental impact, forexample by choosing processing routes that minimize energy consumption.

Place, publisher, year, edition, pages
Luleå: Luleå tekniska universitet, 2019.
Series
Doctoral thesis / Luleå University of Technology 1 jan 1997 → …, ISSN 1402-1544
Keywords [en]
Geometallurgy, process simulation, minerals, machine learning, textures, decision-making
National Category
Mineral and Mine Engineering Metallurgy and Metallic Materials
Research subject
Mineral Processing
Identifiers
URN: urn:nbn:se:ltu:diva-73363ISBN: 978-91-7790-346-8 (print)ISBN: 978-91-7790-347-5 (electronic)OAI: oai:DiVA.org:ltu-73363DiVA, id: diva2:1300831
Public defence
2019-06-13, F531, F-huset, Luleå, 10:00 (English)
Opponent
Supervisors
Available from: 2019-04-03 Created: 2019-03-29 Last updated: 2019-05-24Bibliographically approved

Open Access in DiVA

fulltext(34685 kB)422 downloads
File information
File name FULLTEXT01.pdfFile size 34685 kBChecksum SHA-512
0f9df16533b496d75a76ebddd9c1dc5f9182080aef50805c3d66ba4302283ec4c1c35d4e8665f79e0e06968ade169bf43d035beee816e1f0c2a075671779ac8a
Type fulltextMimetype application/pdf

Authority records BETA

Koch, Pierre-Henri

Search in DiVA

By author/editor
Koch, Pierre-Henri
By organisation
Minerals and Metallurgical Engineering
Mineral and Mine EngineeringMetallurgy and Metallic Materials

Search outside of DiVA

GoogleGoogle Scholar
Total: 422 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 266 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf