Mine Economic and Management
Sarina Akbari; Reza Ghezelbash; Hamidreza Ramazi; Abbas Maghsoudi
Abstract
Natural hazards, particularly landslides, have long posed significant threats to people, buildings, and the surrounding environment. Therefore, comprehensive planning for urban and rural development necessitates the development and implementation of landslide risk zoning models. Numerous methodologies ...
Read More
Natural hazards, particularly landslides, have long posed significant threats to people, buildings, and the surrounding environment. Therefore, comprehensive planning for urban and rural development necessitates the development and implementation of landslide risk zoning models. Numerous methodologies have been proposed for generating landslide hazard maps, which can potentially aid in predicting future landslide-prone areas. This study employed an integrated approach that combines statistical and multi-criteria decision-making (MCDM) methodologies. The Frequency Ratio (FR) and Analytical Hierarchy Process (AHP) were utilized as knowledge-driven approaches, while the Support Vector Machine (SVM) using an RBF kernel, a widely recognized machine learning algorithm, was applied as a data-driven method. Ten factors influencing landslides were considered, including slope angle, aspect, altitude, geology, land use, climate, erosion, and distances from rivers, faults, and roads. The results revealed that landslides are more predictable in the southern, southwestern, and central regions of the studied area. A quantitative assessment of the different methods using prediction-rate curves indicated that the SVM method outperformed the FR and AHP-FR approaches in identifying susceptible areas. The findings of this work could be effectively employed to mitigate potential future hazards and associated damages.
Exploration
Hassanreza Ghasemi Tabar; Sajjad Talesh Hosseini; Andisheh Alimoradi; mahdi fathi; Maryam Sahafzadeh
Abstract
Estimating ore grades during the exploration phase is often time-consuming and costly due to the need for extensive drilling. Geophysical surveys, as the last indirect exploration method before drilling, offer valuable insights into subsurface mineralization. This study introduces a novel approach for ...
Read More
Estimating ore grades during the exploration phase is often time-consuming and costly due to the need for extensive drilling. Geophysical surveys, as the last indirect exploration method before drilling, offer valuable insights into subsurface mineralization. This study introduces a novel approach for simulating “identical twins” of borehole copper grade values using geophysical attributes derived from the geoelectrical method in the Kahang porphyry copper deposit, central Iran. By treating the simulated values as digital twins of actual borehole grades, we employed four machine learning algorithms—Linear Regression (LR), Gradient Boosting (GB), Random Forest (RF), and Support Vector Machine (SVM)—to model the complex relationships between geophysical inputs and copper grades. After data preprocessing with Principal Component Analysis (PCA), a refined dataset was used to train, test, and validate each model. The results demonstrate that GB yielded the highest predictive accuracy, generating grade estimates closely aligned with actual values. This identical twin modeling approach highlights the potential of machine learning to enhance early-stage mineral exploration by reducing dependence on costly drilling.
Rock Mechanics
Arun Kumar Sahoo; Debi Prasad Tripathy; Singam Jayanthu
Abstract
The mining industry needs to accept new-age autonomous technologies and intelligent systems to stay up with the modernization of technology, to benefit the shake of investors and stakeholders, and most significantly, for the nation, and to protect health and safety. An essential part of geo-technical ...
Read More
The mining industry needs to accept new-age autonomous technologies and intelligent systems to stay up with the modernization of technology, to benefit the shake of investors and stakeholders, and most significantly, for the nation, and to protect health and safety. An essential part of geo-technical engineering is doing slope stability analysis to determine the likelihood of slope failure and how to prevent it. A reliable, cost-effective, and generally applicable technique for evaluating slope stability is urgently needed. Numerous research studies have been conducted, each employing a unique strategy. An alternate method that uses machine learning (ML) techniques is to study the relationship between stability conditions and slope characteristics by analyzing the data collected from slope monitoring and testing. This paper is an attempt by the authors to comprehensively review the literature on using the ML techniques in slope stability analysis. It was found that most researchers relied on data-driven approaches with limited input variables, and it was also verified that the ML techniques could be utilized effectively to predict slope failure analysis. SVM and RF were the most popular types of ML models being used. RMSE and AUC were used extensively in assessing the performance of the ML models.
Rock Mechanics
Alireza Afradi; Arash Ebrahimabadi; Mansour Hedayatzadeh
Abstract
Tunnel Boring Machines (TBMs) are extensively used to excavate underground spaces in civil and tunneling projects. An accurate evaluation of their penetration rate is the key factor for the TBM performance prediction. In this study, artificial intelligence methods are used to predict the TBM penetration ...
Read More
Tunnel Boring Machines (TBMs) are extensively used to excavate underground spaces in civil and tunneling projects. An accurate evaluation of their penetration rate is the key factor for the TBM performance prediction. In this study, artificial intelligence methods are used to predict the TBM penetration rate in excavation operations in the Kerman tunnel and the Gavoshan water conveyance tunnels. The aim of this paper is to show the application of the Multivariate Linear Regression (MLR), Artificial Neural Network (ANN), and Support Vector Machine (SVM) for the TBM penetration rate prediction. The penetration rate parameter is considered as a dependent variable, and the Rock Quality Designation (RQD), Brazilian Tensile Strength (BTS), Uniaxial Compressive Strength (UCS), Density (D), Joint Angle (JA), Joint Spacing (JS), and Poisson's Ratio are considered as independent variables. The obtained results by the several proposed methods indicated a high accuracy between the predicted and measured penetration rates, but the support vector machine yields more precise and realistic outcomes.
Sirvan Moradi; Seyed Davoud Mohammadi; Abbas Aghajani Bazzazi; Ali Aali Anvari; Ava Osmanpour
Abstract
Feasibility studies of mining and industrial investment projects are usually associated with uncertain parameters; hence, these investigations rely on prediction. In these particular conditions, simulation and modelling techniques remain the most significant approaches to reduce the decision risk. Since ...
Read More
Feasibility studies of mining and industrial investment projects are usually associated with uncertain parameters; hence, these investigations rely on prediction. In these particular conditions, simulation and modelling techniques remain the most significant approaches to reduce the decision risk. Since several uncertain parameters are incorporated in the modelling process, distribution functions are employed to explain the parameters. However, due to the usual constrain of limited data, these functions cannot significantly explain the variation of those uncertain parameters. Support vector machine, one of the efficient techniques of artificial intelligence, provides the appropriate results in the classification and regression tasks. The principal aims of this research work are to integrate the simulation and artificial intelligence methods to manage the risk prediction of an economic system under uncertain conditions. The financial process of the Halichal mine in the Mazandaran province, Iran, is considered a case study to prove the performance of the support vector machine technique. The results show that integrating the simulation and support vector machine techniques can provide more realistic results, especially when including uncertain parameters. The correlation between the net present value obtained from the simulation and the net present value is about 0.96, which shows the capability of artificial intelligence methods and the simulation process. The root mean square error of the support vector machine prediction is about 0.322, which indicates a low error rate in the net present value estimation. The values of these errors prove that this method has a high accuracy and performance for predicting a net present value in the Halichal granite mine.
Exploration
N. Mahvash Mohammadi; A. Hezarkhani
Abstract
Identification and mapping of the significant alterations are the main objectives of the exploration geochemical surveys. The field study is time-consuming and costly to produce the classified maps. Therefore, the processing of remotely sensed data, which provide timely and multi-band (multi-layer) data, ...
Read More
Identification and mapping of the significant alterations are the main objectives of the exploration geochemical surveys. The field study is time-consuming and costly to produce the classified maps. Therefore, the processing of remotely sensed data, which provide timely and multi-band (multi-layer) data, can be substituted for the field study. In this study, the ASTER imagery is used for alteration classification by applying two new methods of machine learning, including Random Forest and Support Vector Machine. The 14 band ASTER and 19 derivative data layers extracted from ASTER including band ratio and PC imagery, are used as training datasets for improving the results. Comparison of analytical results achieved from the two mentioned methods confirmed that the SVM model has sufficient accuracy and more powerful performance than RF model for alteration classification in the study area.
Rock Mechanics
S. Moshrefi; K. Shahriar; A. Ramezanzadeh; K. Goshtasbi
Abstract
A rock failure criterion is very important for prediction of the ultimate strength in rock mechanics and geotechnics; it is determined for rock mechanics studies in mining, civil, and oil wellborn drilling operations. Also shales are among the most difficult to treat formations. Therefore, in this research ...
Read More
A rock failure criterion is very important for prediction of the ultimate strength in rock mechanics and geotechnics; it is determined for rock mechanics studies in mining, civil, and oil wellborn drilling operations. Also shales are among the most difficult to treat formations. Therefore, in this research work, using the artificial neural network (ANN), a model was built to predict the ultimate strength of shale, and comparison was made with support vector machine (SVM), multiple linear regression models, and the widely used conventional polyaxial failure criteria in the stability analysis of rock structures, Drucker-Prager, and Mogi-Coulomb. For building the model, the corresponding results of triaxial and polyaxial tests have been performed on shales by various researchers. They were collected from reliable published articles. The results obtained showed that a feed forward back propagation multi-layer perceptron (MLP) was used and trained using the Levenberg–Marquardt algorithm, and the 2-4-1 architecture with root-mean-square-error (RMSE) of 24.41 exhibits a better performance in predicting the ultimate strength of shale in comparison with the investigated models. Also for further validation, triaxial tests were performed on the deep shale specimens. They were prepared from the Ramshire oilfield in SW Iran. The results obtained were compared with ANN, SVM, multiple linear regression models, and the conventional failure criterion prediction. They showed that the ANN model predicted ultimate strength with a minimum error and RMSE being equal to 43.81. Then the model was used for prediction of the threshold broken pressure shale layer in the Gachsaran oilfield in Iran. For this, a vertical and horizontal stress was calculated based on a depth of shale layer. The threshold broken pressure was calculated for the beginning and ending of a shale layer to be 154.21 and 167.98 Mpa, respectively.
R. Gholami; A. Moradzadeh
Abstract
Reservoir permeability is a critical parameter for characterization of the hydrocarbon reservoirs. In fact, determination of permeability is a crucial task in reserve estimation, production and development. Traditional methods for permeability prediction are well log and core data analysis which are ...
Read More
Reservoir permeability is a critical parameter for characterization of the hydrocarbon reservoirs. In fact, determination of permeability is a crucial task in reserve estimation, production and development. Traditional methods for permeability prediction are well log and core data analysis which are very expensive and time-consuming. Well log data is an alternative approach for prediction of permeability because they are usually available for all of the wells. Hence, attempts have been made to utilize well log data to predict permeability. However, because of complicate and non-linear relationship of well log and core permeability data, usual statistical and artificial methods are not completely able to provide meaningful results. In this regard, recent works on artificial intelligence have led to the introduction of a robust method generally called support vector machine (SVM). The term “SVM” is divided into two subcategories: support vector classifier (SVC) and support vector regression (SVR). The aim of this paper is to use SVR for predicting the permeability of three gas wells in South Pars filed, Iran. The results show that the overall correlation coefficient (R) between predicted and measured permeability of SVR is 0.97 compared to 0.71 of a developed general regression neural network. In addition, the strength and efficiency of SVR was proved by less time-consuming and better root mean square error in training and testing dataset.
