Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Application of clustering techniques for improved energy benchmarking on deep-level mines
Thesis (MEng)--Stellenbosch University, 2024.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | en_ZA en_ZA |
| Published: |
Stellenbosch : Stellenbosch University
2024
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867614080996474880 |
|---|---|
| access_status_str | Open Access |
| author | Caromba, Claudio Mauricio |
| author2 | Schutte, Cornelius Stephanus Lodewyk
|
| author_browse | Caromba, Claudio Mauricio Schutte, Cornelius Stephanus Lodewyk |
| author_facet | Schutte, Cornelius Stephanus Lodewyk
Caromba, Claudio Mauricio |
| author_sort | Caromba, Claudio Mauricio |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (MEng)--Stellenbosch University, 2024. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/130671 |
| institution | Stellenbosch University (South Africa) |
| language | en_ZA en_ZA |
| last_indexed | 2026-06-10T12:46:21.556Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2024 |
| publishDateRange | 2024 |
| publishDateSort | 2024 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/130671 Application of clustering techniques for improved energy benchmarking on deep-level mines Caromba, Claudio Mauricio Schutte, Cornelius Stephanus Lodewyk Van Laar, Jean Herman Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering. Mineral industries -- Energy consumption Energy benchmarking Clustering algorithms UCTD Thesis (MEng)--Stellenbosch University, 2024. ENGLISH ABSTRACT: The South African mining industry relies on coal-based energy to run operations, with the sector utilising 29.6 Terawatt-hours of electricity in 2018. However, constraints on energy availability and rising electricity prices pressure the industry to manage energy usage better and remain one of the main contributors to the South African economy. Energy benchmarking is a popular and effective energy management method in the mining sector. Current methods use the average energy usage of various mining shafts over different intervals to develop benchmarks. These benchmarks may lead to skewed performance evaluation if the mining shafts have vastly different modes of operation or when anomalous energy usage is present within the interval. Clustering-based benchmarking techniques have been applied successfully in other industries, to group similar energy users and identify energy savings opportunities, but remains unstudied in the mining sector. This study developed and applied a clustering-based benchmarking method to evaluate the performance of a single high-energy usage system (internal benchmarking) and different mining shafts (external benchmarking) at a deep level gold mine in South Africa. A combination of the quantitative research approach and a case study research design was employed to derive steps for each application of the method. In both cases the relevant datasets were collected and processed, before selecting the number of clusters and best suited clustering algorithm using a combination of popular clustering metrics. The clusters were used to identify benchmarks to compare with traditional methods from the industry. The K-means unsupervised learning clustering algorithm was used to group energy usage patterns and generate five typical load profiles representing common pumping energy usage on a mine shaft. These profiles were used as benchmarks to identify outlier energy usage and potential energy-saving opportunities on the mining shaft using cluster quartile and average pumping-energy intensity benchmarks. The K-means clustering algorithm was applied to four production shafts at the mining complex to identify different energy usage groups. The ordinary least squares benchmarking method was leveraged to develop expected energy usage benchmarks within each cluster to determine the scope for improvement and compare energy performance. Compared to traditional methods, the benefit of clustered energy usage benchmarks on deep-level mines is shown by successfully identifying groups that better describe the energy usage type. The five typical load profiles were 14 % better at explaining the energy usage than the day-of-week grouping, and the four cluster-based expected energy benchmarks were twice as good at energy prediction than the current method of a single equation for all the energy users. These improvements in energy benchmark development allow for a fair and accurate evaluation of energy performance and do not over- or underestimate the severity of the wastage and the opportunity for energy savings. This facilitates focused and measurable energy management that may assist mines in continuing their valuable contribution to the South African economy. AFRIKAANSE OPSOMMING: Die Suid-Afrikaanse mynbedryf, een van die grootste bydraers tot die SuidAfrikaanse ekonomie, maak staat op steenkoolopgewekte energie om operasies te bedryf. In 2018 het die sektor 29.6 Terawatt-uur van elektrisiteit gebruik. Beperkings op energiebeskikbaarheid en stygende elektrisiteitspryse plaas egter druk op die bedryf om energieverbruik beter te bestuur. Energiemaatstawwe is ’n gewilde en effektiewe energiebestuursmetode in die mynbousektor. Die huidige metodes gebruik die gemiddelde energieverbruik van verskillende mynskagte oor verskye tydsintervalle om maatstawwe te ontwikkel. Hierdie maatstawwe kan lei tot skewe prestasie-evaluering as die mynskagte baie verskillende werkswyses het of wanneer afwykende energieverbruik binne die interval teenwoordig is. Groeperingsgebaseerde maatstaftegnieke is al suksesvol in ander nywerhede toegepas om soortgelyke energiegebruikers te groepeer en energiebesparingsgeleenthede te identifiseer, maar bly onbestudeer in die mynbousektor. Hierdie studie het ’n groeperingsgebaseerde maatstafmetode ontwikkel en toegepas om die prestasie van ’n enkele hoë-energieverbruikstelsel (interne maatstawwe) en verskillende mynskagte (eksterne maatstawwe) by ’n diepvlak-goudmyn in Suid-Afrika te evalueer. ’n Kombinasie van die kwantitatiewe navorsingsbenadering en ’n gevallestudienavorsingsontwerp is gebruik om stappe vir elke toepassing van die metode af te lei. In beide gevalle is die relevante datastelle ingesamel en verwerk voordat die aantal groepe en die mees geskikte groeperingsalgoritme met behulp van ’n kombinasie van gewilde groeperingsmetrieke gekies is. Die groepe is gebruik om vergelykingsmerke te identifiseer om met tradisionele metodes in die industrie te vergelyk. Die K-gemiddeld-sonder-toesig-leergroeperingsalgoritme is gebruik om energiegerbruikspatrone te groepeer en vyf tipiese lasprofiele op te wek wat algemene pompenergieverbruik op ’n mynskag verteenwoordig. Hierdie profiele is as maatstawwe gebruik om uitskieters in energieverbruik en potensiële energiebesparende geleenthede op die mynskag te identifiseer deur gebruik te maak van kwartiel en gemiddelde pompenergie-intensiteitsmaatstawwe. Die K-gemiddeld groeperingsalgoritme is op vier produksieskagte by die mynkompleks toegepas om verskillende energieverbruiksgroepe te identifiseer. Die gewone maatstafmetode met die minste kwadrate is gebruik om verwagte maatstawwe vir energieverbruik binne elke groep te ontwikkel om die ruimte vir verbetering te bepaal en energieprestasie te vergelyk. In vergelyking met tradisionele metodes word die voordeel van gegroepeerdemaatstawwe vir energieverbruik op diepvlakmyne getoon deur groepe suksesvol te identifiseer wat die tipe energieverbruik beter beskryf. Die vyf tipiese lasprofiele was 14% beter om die energieverbruik te verduidelik as die dag-vanweek-groepering, en die vier groepgebaseerde verwagte energiemaatstawwe was twee keer so goed in energievoorspelling as die huidige metode van ’n enkele vergelyking vir al die energiegebruikers. Hierdie verbeterings in die ontwikkeling van energiemaatstawwe maak voorsiening vir ’n regverdige en akkurate evaluering van energieprestasie en onderskat nie die erns van die vermorsing en geleentheid vir energiebesparing nie. Dit fasiliteer gefokusde en meetbare energiebestuur wat myne kan help om hul waardevolle bydrae tot die Suid-Afrikaanse ekonomie voort te sit. Masters 2024-02-13T10:32:42Z 2024-04-27T02:05:57Z 2024-02-13T10:32:42Z 2024-04-27T02:05:57Z 2024-03 Thesis https://scholar.sun.ac.za/handle/10019.1/130671 en_ZA en_ZA Stellenbosch University xvi, 111 pages : illustrations. application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Mineral industries -- Energy consumption Energy benchmarking Clustering algorithms UCTD Caromba, Claudio Mauricio Application of clustering techniques for improved energy benchmarking on deep-level mines |
| title | Application of clustering techniques for improved energy benchmarking on deep-level mines |
| title_full | Application of clustering techniques for improved energy benchmarking on deep-level mines |
| title_fullStr | Application of clustering techniques for improved energy benchmarking on deep-level mines |
| title_full_unstemmed | Application of clustering techniques for improved energy benchmarking on deep-level mines |
| title_short | Application of clustering techniques for improved energy benchmarking on deep-level mines |
| title_sort | application of clustering techniques for improved energy benchmarking on deep level mines |
| topic | Mineral industries -- Energy consumption Energy benchmarking Clustering algorithms UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/130671 |
| work_keys_str_mv | AT carombaclaudiomauricio applicationofclusteringtechniquesforimprovedenergybenchmarkingondeeplevelmines |