Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings.
Thesis (PhD)--Stellenbosch University, 2023.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | en_ZA en_ZA |
| Published: |
Stellenbosch : Stellenbosch University
2023
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613940400259072 |
|---|---|
| access_status_str | Open Access |
| author | Harmse, Michael David |
| author2 | Schutte, Cornelius Stephanus Lodewyk |
| author_browse | Harmse, Michael David Schutte, Cornelius Stephanus Lodewyk |
| author_facet | Schutte, Cornelius Stephanus Lodewyk Harmse, Michael David |
| author_sort | Harmse, Michael David |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (PhD)--Stellenbosch University, 2023. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/126961 |
| institution | Stellenbosch University (South Africa) |
| language | en_ZA en_ZA |
| last_indexed | 2026-06-10T12:44:07.837Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2023 |
| publishDateRange | 2023 |
| publishDateSort | 2023 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/126961 A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. Harmse, Michael David Schutte, Cornelius Stephanus Lodewyk Van Laar, Jean Herman Pelser, Wiehan Adriaan Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering. Artificial intelligence -- Engineering applications -- South Africa Mineral industries -- Technological innovations -- South Africa Energy conservation -- Cost effectiveness -- South Africa Thesis (PhD)--Stellenbosch University, 2023. ENGLISH ABSTRACT: The South African deep-level mining industry is experiencing a reduction in profitability due to increasing electricity costs. In the last ten years, the cost of electricity has increased considerably more than the price of most deep-level mine commodity outputs such as gold and platinum. The ventilation and refrigeration system, compressed air system, and pumping system contribute 63% to a typical deep-level mine’s energy cost. Improved control of these systems can improve profitability and is mainly implemented by static time-based control. Dynamic control, as opposed to present static time-based control, can be implemented to reduce energy inefficiencies and wastage. However, a lack of measurement equipment in the South African deep-level mining industry restricts the implementation of dynamic control. The lack of dynamic control results in energy wastage due to oversupply of the mining service systems to account for all operating conditions. A need was therefore identified to implement integrated dynamic control on mining service systems to reduce energy consumption while maintaining service delivery requirements. However, shortcomings in measurement equipment prevent the successful implementation of integrated dynamic control. Artificial intelligence (AI) techniques are used to overcome shortcomings in measurement equipment. However, AI techniques have seldomly been used in a control system in the deep-level mining industry. Additionally, these techniques have not been compared to determine the ideal technique for the application and control system. As a solution, the case study methodology was used to develop a method to utilise AI techniques to improve dynamic control in deep-level mine service systems. The method consisted of a generic optimisation process and an AI technique selection guideline. The ideal AI technique was identified based on the service and control system characteristics. Two case studies were conducted to validate the developed method. The first validation case study used an artificial neural network to predict appropriate compressed air surface pressure set points based on level requirements and present demand. This reduces unnecessary system pressurisation and tardy system depressurisation. An average absolute error of 2.83% was experienced, resulting in a potential 14.2% reduction in energy consumption, which would not have been possible without AI techniques. The second validation case study used a genetic algorithm to determine the economic dispatch of a compressed air system to determine the benefit of an air receiver. Converting a 1.17 km length of inactive haulage to an air receiver will result in an energy cost reduction of 8.8% at the case study mine with a payback period of seven months. Thus, the developed method using AI techniques can successfully improve the dynamic control of deep-level mining service systems. The novel contributions of the study include using AI techniques in a deep-level mining control system and developing a method for implementing AI techniques in deep-level mining service control systems for improved dynamic control. AFRIKAANS OPSOMMING: Die Suid-Afrikaanse diepvlakmynbedryf ervaar ʼn afname in winsgewendheid as gevolg van toenemende elektrisiteitskoste. Die koste van elektrisiteit het die afgelope tien jaar aansienlik meer gestyg as die prys van die meeste diepvlakmynkommoditeitsuitsette soos goud en platinum. Die ventilasie- en verkoelingstelsel, saamgeperste lugstelsel en pompstelsel dra 63% by tot ʼn tipiese diepvlakmyn se energiekoste. Verbeterde beheer van hierdie stelsels kan winsgewendheid verhoog en word hoofsaaklik geïmplementeer deur statiese tydgebaseerde beheer. Dinamiese beheer, in teenstelling met huidige statiese tydgebaseerde beheer, kan geïmplementeer word om energie-ondoeltreffendheid en vermorsing te verminder. ʼn Gebrek aan meettoerusting in die Suid-Afrikaanse diepvlakmynbedryf beperk egter die implementering van dinamiese beheer. Die gebrek aan dinamiese beheer lei tot energievermorsing as gevolg van ooraanbod van die myndiensstelsels om rekening te hou met alle bedryfstoestande. 'n Behoefte is dus geïdentifiseer om geïntegreerde dinamiese beheer op myndiensstelsels te implementeer om energieverbruik te verminder terwyl diensleweringsvereistes gehandhaaf word. Tekortkominge in meettoerusting verhoed egter die suksesvolle implementering van geïntegreerde dinamiese beheer. Kunsmatige intelligensie tegnieke is gebruik om die tekortkominge in meettoerusting te oorkom. Kunsmatige intelligensie tegnieke is egter selde in beheerstelsels in die diepvlakmynbedryf gebruik. Verder is hierdie tegnieke nie vergelyk om die ideale tegniek vir die toedienings- en beheerstelsel te bepaal nie. As ʼn oplossing is ʼn gevallestudie metodologie gebruik om ʼn metode te ontwikkel om kunsmatige intelligensie tegnieke te gebruik om dinamiese beheer binne diepvlakmyndiensstelsels te verbeter. Die metode bestaan uit ʼn generiese optimaliseringsproses en ʼn kunsmatige intelligensie tegniek seleksie riglyn. Die ideale kunsmatige intelligensie tegniek word geïdentifiseer op grond van die diens- en beheerstelsel kenmerke. Twee gevallestudies is uitgevoer om die ontwikkelde metode te valideer. Die eerste valideringsgevallestudie het ʼn kunsmatige neurale netwerk gebruik om toepaslike perslug oppervlakdruk setpunte te voorspel gebaseer op die vlak vereistes en huidige aanvraag. Dit verminder onnodige en vertraagde stelseldrukverlaging. ʼn Gemiddelde absolute fout van 2.83% is gevind, wat gelei het tot ʼn potensiële vermindering van 14.2% in energieverbruik. Dit sou nie moontlik gewees het sonder kunsmatige intelligensie tegnieke nie. Die tweede valideringsgevallestudie het ʼn genetiese algoritme gebruik om die ekonomiese versending van ʼn saamgeperste lugstelsel te bepaal om die voordeel van ʼn lug ontvanger te bepaal. Die omskakeling van ʼn 1.17 km lengte van onaktiewe tonnel na ʼn lugontvanger sal lei tot ʼn energiekostevermindering van 8.8% by die gevallestudiemyn met ʼn terugbetalingstydperk van sewe maande. Die ontwikkelde metode kan dus suksesvol gebruik word om dinamiese beheer van diepvlakmyndiensstelsels te verbeter deur kunsmatige intelligensie tegnieke te gebruik. Die nuwe bydraes van die studie sluit beide die gebruik van kunsmatige intelligensie tegnieke in ʼn diepvlakmynbeheerstelsel in sowel as die ontwikkeling van ʼn metode vir die implementering van kunsmatige intelligensie tegnieke in diepvlakmyndiensbeheerstelsels vir verbeterde dinamiese beheer. Doctoral 2023-01-25T11:41:23Z 2023-05-18T06:57:43Z 2023-01-25T11:41:23Z 2023-05-18T06:57:43Z 2023-03 Thesis http://hdl.handle.net/10019.1/126961 en_ZA en_ZA Stellenbosch University xvi, 170 pages : illustrations. application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Artificial intelligence -- Engineering applications -- South Africa Mineral industries -- Technological innovations -- South Africa Energy conservation -- Cost effectiveness -- South Africa Harmse, Michael David A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. |
| title | A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. |
| title_full | A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. |
| title_fullStr | A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. |
| title_full_unstemmed | A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. |
| title_short | A method to apply artificial intelligence techniques to deep-level mine service control systems to realise energy cost savings. |
| title_sort | method to apply artificial intelligence techniques to deep level mine service control systems to realise energy cost savings |
| topic | Artificial intelligence -- Engineering applications -- South Africa Mineral industries -- Technological innovations -- South Africa Energy conservation -- Cost effectiveness -- South Africa |
| url | http://hdl.handle.net/10019.1/126961 |
| work_keys_str_mv | AT harmsemichaeldavid amethodtoapplyartificialintelligencetechniquestodeeplevelmineservicecontrolsystemstorealiseenergycostsavings AT harmsemichaeldavid methodtoapplyartificialintelligencetechniquestodeeplevelmineservicecontrolsystemstorealiseenergycostsavings |