Application of Artificial Neural Networks in Mineral Resource Evaluation
This paper explores the novel technique of artificial neural networks and their application to mineral resource evaluation. The primary objective of this exploratory work is concerned with cost minimization, especially in drilling. The objective is to reduce investment costs on projects before decis...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
|
| Series: | Journal of King Saud University: Engineering Sciences |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1018363918306925 |
| id |
doaj-art-28ef236c49164ac0ac44fdfd3df70458 |
|---|---|
| recordtype |
oai_dc |
| spelling |
doaj-art-28ef236c49164ac0ac44fdfd3df704582018-09-19T10:47:46ZengElsevierJournal of King Saud University: Engineering Sciences1018-3639101127138Application of Artificial Neural Networks in Mineral Resource EvaluationSaleh M. Al-Alawi0E.E. Tawo1Department of Electrical and Electronics Engineering,Sultanate of OmanDepartment of Petroleum and Mining Engineering Sultan Qaboos University,Sultanate of OmanThis paper explores the novel technique of artificial neural networks and their application to mineral resource evaluation. The primary objective of this exploratory work is concerned with cost minimization, especially in drilling. The objective is to reduce investment costs on projects before decisions on further development of a mineralization are made. A data set consisting of 163 sample locations for a bauxite deposit was collected and used for training and testing an artificial neural network model using the back propagation technique. This trained network was then used to predict known point values at specified locations. Results between actual and predictions were then compared for validation purposes. Same results are also compared with estimates obtained from a geostatistical technique (kriging). These results indicate that the ANN-based model predictions were in close agreement with actual results within the main zone of mineralization than at the boundaries. Finally, the validated model was used to predict values at unsampled locations in order to determine the feasibility of drilling in these locations. The investigation on the bauxite deposit shows that artificial neural networks can be used as a complementary decision tool in mining and earth sciences, not so much at an early stage of exploration as in the later stages of exploitation.http://www.sciencedirect.com/science/article/pii/S1018363918306925 |
| institution |
Open Data Bank |
| collection |
Open Access Journals |
| building |
Directory of Open Access Journals |
| language |
English |
| format |
Article |
| author |
Saleh M. Al-Alawi E.E. Tawo |
| spellingShingle |
Saleh M. Al-Alawi E.E. Tawo Application of Artificial Neural Networks in Mineral Resource Evaluation Journal of King Saud University: Engineering Sciences |
| author_facet |
Saleh M. Al-Alawi E.E. Tawo |
| author_sort |
Saleh M. Al-Alawi |
| title |
Application of Artificial Neural Networks in Mineral Resource Evaluation |
| title_short |
Application of Artificial Neural Networks in Mineral Resource Evaluation |
| title_full |
Application of Artificial Neural Networks in Mineral Resource Evaluation |
| title_fullStr |
Application of Artificial Neural Networks in Mineral Resource Evaluation |
| title_full_unstemmed |
Application of Artificial Neural Networks in Mineral Resource Evaluation |
| title_sort |
application of artificial neural networks in mineral resource evaluation |
| publisher |
Elsevier |
| series |
Journal of King Saud University: Engineering Sciences |
| issn |
1018-3639 |
| description |
This paper explores the novel technique of artificial neural networks and their application to mineral resource evaluation. The primary objective of this exploratory work is concerned with cost minimization, especially in drilling. The objective is to reduce investment costs on projects before decisions on further development of a mineralization are made. A data set consisting of 163 sample locations for a bauxite deposit was collected and used for training and testing an artificial neural network model using the back propagation technique. This trained network was then used to predict known point values at specified locations. Results between actual and predictions were then compared for validation purposes. Same results are also compared with estimates obtained from a geostatistical technique (kriging). These results indicate that the ANN-based model predictions were in close agreement with actual results within the main zone of mineralization than at the boundaries. Finally, the validated model was used to predict values at unsampled locations in order to determine the feasibility of drilling in these locations. The investigation on the bauxite deposit shows that artificial neural networks can be used as a complementary decision tool in mining and earth sciences, not so much at an early stage of exploration as in the later stages of exploitation. |
| url |
http://www.sciencedirect.com/science/article/pii/S1018363918306925 |
| _version_ |
1612547931947139072 |