Decision support system of unplanned dilution and ore-loss in underground stoping operations using a neuro-fuzzy system

Hyongdoo Jang, Erkan Topal, Youhei Kawamura

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Abstract Unplanned dilution and ore-loss are the most critical challenges in underground stoping operations. These problems are the main cause behind a mine closure and directly influencing the productivity of the underground stope mining and the profitability of the entire operation. Despite being aware of the significance of unplanned dilution and ore-loss, prediction of these phenomena is still unexplained as they occur through complex mechanisms and causative factors. Current management practices primarily rely on similar stope reconciliation data and the intuition of expert mining engineers. In this study, an innovative unplanned dilution and ore-loss (uneven break: UB) management system is established using a neuro-fuzzy system. The aim of the proposed decision support system is to overcome the UB phenomenon in underground stope blasting which provides quantitative prediction of unplanned dilution and ore-loss with practical recommendations simultaneously. To achieve the method proposed, an uneven break (UB) prediction system was developed by an artificial neural network (ANN) considering 1076 datasets covering 10 major UB causative factors collected from three underground stoping mines in Western Australia. In succession, the UB consultation system was established via a fuzzy expert system (FES) in reference to surveyed results of fifteen underground-mining experts. The UB prediction and consultation system were combined as one concurrent neuro-fuzzy system that is named the 'uneven break optimiser'. Because the current UB prediction systems in investigated mines were highly unsatisfactory with correlation coefficient (R) of 0.088 and limited to only unplanned dilution, the performance of the proposed UB prediction system (R of 0.719) is a remarkable achievement. The uneven break optimiser can be directly employed to improve underground stoping production, and this tool will be beneficial not only for underground stope planning and design but also for production management.

Original languageEnglish
Article number2872
Pages (from-to)1-12
Number of pages12
JournalApplied Soft Computing Journal
Volume32
DOIs
Publication statusPublished - Jul 1 2015
Externally publishedYes

Fingerprint

Stoping
Fuzzy systems
Decision support systems
Ores
Dilution
Blasting
Expert systems
Profitability
Productivity
Neural networks
Engineers
Planning

Keywords

  • Neuro-fuzzy system
  • Ore-loss
  • Underground metalliferous mining
  • Unplanned dilution

ASJC Scopus subject areas

  • Software

Cite this

Decision support system of unplanned dilution and ore-loss in underground stoping operations using a neuro-fuzzy system. / Jang, Hyongdoo; Topal, Erkan; Kawamura, Youhei.

In: Applied Soft Computing Journal, Vol. 32, 2872, 01.07.2015, p. 1-12.

Research output: Contribution to journalArticle

@article{6eb570a35ca34bc1abd0590a9c226653,
title = "Decision support system of unplanned dilution and ore-loss in underground stoping operations using a neuro-fuzzy system",
abstract = "Abstract Unplanned dilution and ore-loss are the most critical challenges in underground stoping operations. These problems are the main cause behind a mine closure and directly influencing the productivity of the underground stope mining and the profitability of the entire operation. Despite being aware of the significance of unplanned dilution and ore-loss, prediction of these phenomena is still unexplained as they occur through complex mechanisms and causative factors. Current management practices primarily rely on similar stope reconciliation data and the intuition of expert mining engineers. In this study, an innovative unplanned dilution and ore-loss (uneven break: UB) management system is established using a neuro-fuzzy system. The aim of the proposed decision support system is to overcome the UB phenomenon in underground stope blasting which provides quantitative prediction of unplanned dilution and ore-loss with practical recommendations simultaneously. To achieve the method proposed, an uneven break (UB) prediction system was developed by an artificial neural network (ANN) considering 1076 datasets covering 10 major UB causative factors collected from three underground stoping mines in Western Australia. In succession, the UB consultation system was established via a fuzzy expert system (FES) in reference to surveyed results of fifteen underground-mining experts. The UB prediction and consultation system were combined as one concurrent neuro-fuzzy system that is named the 'uneven break optimiser'. Because the current UB prediction systems in investigated mines were highly unsatisfactory with correlation coefficient (R) of 0.088 and limited to only unplanned dilution, the performance of the proposed UB prediction system (R of 0.719) is a remarkable achievement. The uneven break optimiser can be directly employed to improve underground stoping production, and this tool will be beneficial not only for underground stope planning and design but also for production management.",
keywords = "Neuro-fuzzy system, Ore-loss, Underground metalliferous mining, Unplanned dilution",
author = "Hyongdoo Jang and Erkan Topal and Youhei Kawamura",
year = "2015",
month = "7",
day = "1",
doi = "10.1016/j.asoc.2015.03.043",
language = "English",
volume = "32",
pages = "1--12",
journal = "Applied Soft Computing",
issn = "1568-4946",
publisher = "Elsevier",

}

TY - JOUR

T1 - Decision support system of unplanned dilution and ore-loss in underground stoping operations using a neuro-fuzzy system

AU - Jang, Hyongdoo

AU - Topal, Erkan

AU - Kawamura, Youhei

PY - 2015/7/1

Y1 - 2015/7/1

N2 - Abstract Unplanned dilution and ore-loss are the most critical challenges in underground stoping operations. These problems are the main cause behind a mine closure and directly influencing the productivity of the underground stope mining and the profitability of the entire operation. Despite being aware of the significance of unplanned dilution and ore-loss, prediction of these phenomena is still unexplained as they occur through complex mechanisms and causative factors. Current management practices primarily rely on similar stope reconciliation data and the intuition of expert mining engineers. In this study, an innovative unplanned dilution and ore-loss (uneven break: UB) management system is established using a neuro-fuzzy system. The aim of the proposed decision support system is to overcome the UB phenomenon in underground stope blasting which provides quantitative prediction of unplanned dilution and ore-loss with practical recommendations simultaneously. To achieve the method proposed, an uneven break (UB) prediction system was developed by an artificial neural network (ANN) considering 1076 datasets covering 10 major UB causative factors collected from three underground stoping mines in Western Australia. In succession, the UB consultation system was established via a fuzzy expert system (FES) in reference to surveyed results of fifteen underground-mining experts. The UB prediction and consultation system were combined as one concurrent neuro-fuzzy system that is named the 'uneven break optimiser'. Because the current UB prediction systems in investigated mines were highly unsatisfactory with correlation coefficient (R) of 0.088 and limited to only unplanned dilution, the performance of the proposed UB prediction system (R of 0.719) is a remarkable achievement. The uneven break optimiser can be directly employed to improve underground stoping production, and this tool will be beneficial not only for underground stope planning and design but also for production management.

AB - Abstract Unplanned dilution and ore-loss are the most critical challenges in underground stoping operations. These problems are the main cause behind a mine closure and directly influencing the productivity of the underground stope mining and the profitability of the entire operation. Despite being aware of the significance of unplanned dilution and ore-loss, prediction of these phenomena is still unexplained as they occur through complex mechanisms and causative factors. Current management practices primarily rely on similar stope reconciliation data and the intuition of expert mining engineers. In this study, an innovative unplanned dilution and ore-loss (uneven break: UB) management system is established using a neuro-fuzzy system. The aim of the proposed decision support system is to overcome the UB phenomenon in underground stope blasting which provides quantitative prediction of unplanned dilution and ore-loss with practical recommendations simultaneously. To achieve the method proposed, an uneven break (UB) prediction system was developed by an artificial neural network (ANN) considering 1076 datasets covering 10 major UB causative factors collected from three underground stoping mines in Western Australia. In succession, the UB consultation system was established via a fuzzy expert system (FES) in reference to surveyed results of fifteen underground-mining experts. The UB prediction and consultation system were combined as one concurrent neuro-fuzzy system that is named the 'uneven break optimiser'. Because the current UB prediction systems in investigated mines were highly unsatisfactory with correlation coefficient (R) of 0.088 and limited to only unplanned dilution, the performance of the proposed UB prediction system (R of 0.719) is a remarkable achievement. The uneven break optimiser can be directly employed to improve underground stoping production, and this tool will be beneficial not only for underground stope planning and design but also for production management.

KW - Neuro-fuzzy system

KW - Ore-loss

KW - Underground metalliferous mining

KW - Unplanned dilution

UR - http://www.scopus.com/inward/record.url?scp=84926485900&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84926485900&partnerID=8YFLogxK

U2 - 10.1016/j.asoc.2015.03.043

DO - 10.1016/j.asoc.2015.03.043

M3 - Article

VL - 32

SP - 1

EP - 12

JO - Applied Soft Computing

JF - Applied Soft Computing

SN - 1568-4946

M1 - 2872

ER -