Exploitation
Tapan Dey; Gopinath Samanta
Abstract
Accurate grade prediction is an important step in the mining planning process. Various methods, namely the Inverse Distance Method and Kriging, are widely used. The application of machine learning is a new development in the grade estimation technique. The present study focused on the application of ...
Read More
Accurate grade prediction is an important step in the mining planning process. Various methods, namely the Inverse Distance Method and Kriging, are widely used. The application of machine learning is a new development in the grade estimation technique. The present study focused on the application of XGBoost, Random Forests (RFs), Multi Layer Perceptron (MLP), and Gradient Boosting Regression (GBR) models to predict iron ore grades in an Indian mine. An ensemble model was also applied to obtain a more stable grade prediction in the deposit. Models were trained using 4,112 sample data, which have spatial coordinates (east, north, and altitude) and iron grades. The dataset was divided into two parts: 80% (3,289 samples) of the data was used for model training, and 20% (823 samples) was used for model testing. The performance of the models was assessed through the coefficient of determination (R²), Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE). The results show that the XGBoost model performs better in the estimation process when compared with other methods, such as RFs, GBR, and MLP. The XGBoost model produced R² of 0.77, MSE of 2.87, and MAPE of 1.8%. The findings indicate that the XGBoost model is effective for predicting iron ore grades in this type of deposit. However, considering geological uncertainty, the application of an ensemble model may be beneficial for grade prediction in an iron deposit.
Exploitation
Mojtaba Rezakhah
Abstract
Optimizing short-term production in open-pit copper mines is crucial for maximizing economic returns and ensuring operational stability, yet is frequently challenged by inherent geological variability. This work presents a novel Mixed-Integer Linear Programming (MILP) framework designed to address these ...
Read More
Optimizing short-term production in open-pit copper mines is crucial for maximizing economic returns and ensuring operational stability, yet is frequently challenged by inherent geological variability. This work presents a novel Mixed-Integer Linear Programming (MILP) framework designed to address these challenges by directly integrating critical geometallurgical parameters, specifically rock hardness (SPI index) and clay content, into the short-term production planning process. The simultaneous integration of these key geometallurgical feed quality attributes within an operational MILP model distinguishes this work from previous approaches and effectively bridges geological data analytics with operational decision-making, aligning economic objectives with enhanced metallurgical performance. Utilizing real operational data from the Sarcheshmeh Copper Mine, the framework was validated over a 186-day period. It achieved optimal production conditions on 137 days (73.6% of the duration), realizing a maximum Net Present Value (NPV) of $132,000. Key outcomes included a significant 21% reduction in concentrate grade variability and a 15% decrease in flotation reagent consumption, achieved through the simultaneous control of SPI and clay content. Advanced statistical methods were employed to identify critical relationships. While the model demonstrates scalability for porphyry copper mines globally, its successful implementation depends on careful parameter customization and alignment with existing infrastructure. This research work underscores the substantial value of data-driven, integrated optimization techniques in enhancing both profitability and process stability within mineral processing circuits.
Exploitation
Fatemeh Asadi Ooriad; Javad Gholamnejad; Ali Dabagh
Abstract
Designing and planning in open-pit mining encompass a series of processes that commence with the preparation of a block model. Subsequently, upon designing the final scope, it culminates with the timing and sequencing of mining blocks, with the aim to maximize the pit's value within specific technical ...
Read More
Designing and planning in open-pit mining encompass a series of processes that commence with the preparation of a block model. Subsequently, upon designing the final scope, it culminates with the timing and sequencing of mining blocks, with the aim to maximize the pit's value within specific technical and operational constraints. Mathematical programming methods have proven suitable for optimizing mine production scheduling. Previous studies have addressed various aspects, including the timing of deployment and periodic relocation of in-pit crushers. Nevertheless, significant challenges remain in integrating the in-pit crusher problem with production planning. This paper introduces a new mixed-integer linear programming model for long-term open-pit mine production planning, incorporating constrained pit deepening to enforce predominantly lateral progression throughout the planning horizon. To achieve this, the number of active benches in each time period was reduced, thereby decreasing the need for equipment movement between working benches. Furthermore, with the horizontal progression of the pit, more workspace became available for deploying in-pit crushers, reducing equipment movement costs between benches and overall transportation costs, ultimately lowering the mine's operational expenses. Finally, the proposed model was implemented at the Miduk copper mine. The results demonstrated that the proposed model successfully achieved the expected objectives, resulting in a 52.45% improvement in reducing the number of active benches and regarding execution time reduction, the model showed a 53.32% improvement.
Exploitation
Eslam Ghojoghi; Hamid Mansouri; Mohamad Ali Ebrahimi Farsangi; Esmat Rashedi
Abstract
Back break is an undesirable consequence of rock blasting, which causes explosive energy loss and reduces operation efficiency. Consequently, it is essential to forecast it to exercise control and avert the occurrence of operational cost losses. The objective of this scientific research is to utilize ...
Read More
Back break is an undesirable consequence of rock blasting, which causes explosive energy loss and reduces operation efficiency. Consequently, it is essential to forecast it to exercise control and avert the occurrence of operational cost losses. The objective of this scientific research is to utilize Deep Neural Network, Extreme Gradient Boosting, and Lasso Regression in conjunction with Gravitational Search Algorithm to make predictions and minimization regarding the occurrence of blast-induced back break at Gol-e-Gohar 4 iron ore mine, Sirjan, Kerman, Iran. The constructed models comprise a set of nine input parameters, encompassing blasting design parameters, and rock geomechanical properties and produces back break as single output. The datasets used for training and evaluation consist of 266 blasting records extracted from Gol-e-Gohar 4 iron ore mine. The results obtained showed that the Deep Neural Network model with R2 of 0.81 and MSE of 0.70 has better performance over the Extreme Gradient Boosting and Lasso regression models to predict back break. Furthermore, the application of optimization algorithm resulted in optimized parameter values, which minimize back break.
Exploitation
Rahul Shende; Omkar Navagire; Himanshu Kumar Jangir; Srinivasan V.; Anirban Mandal; Anjan Patel
Abstract
Due to its intricate challenges, Black cotton (BC) soil is dumped separately in mining areas, and this study focuses on a BC soil dump at an open-cast mine site. This soil, characterized by cohesion of 26-40 kPa and internal friction angle of 13°-17°, exhibits significant expansion and contraction ...
Read More
Due to its intricate challenges, Black cotton (BC) soil is dumped separately in mining areas, and this study focuses on a BC soil dump at an open-cast mine site. This soil, characterized by cohesion of 26-40 kPa and internal friction angle of 13°-17°, exhibits significant expansion and contraction with moisture fluctuations, swelling during wet periods, and shrinking during dry spells, posing considerable challenges in mining areas. The Ministry of Environment, Forest and Climate Change (MoEFCC) has suggested constructing a 15-20m wall to protect the village on the periphery of the BC soil dump of 36m height. This study aimed to identify sustainable and economical alternative feasible remedial solutions. Field testing, including borehole investigation, was conducted to determine the stratigraphy beneath the dump. Numerical analysis using SLOPE/W software was performed for slope optimization and to evaluate remedial measures such as stone pitching and rockfill trench. The study shows that the dump can be stabilized using the design modification and possible cost-effective measures. Based on field observations, dump material testing, and numerical analysis, alternative remedial measures were proposed and implemented. The study also includes a cost-benefit analysis of the recommended remedial measures.
Exploitation
Gebremariam Mesele; Miruts Hagos; Bheemalingeswara Konka; Tsegabrhan Gebreset; Misgan Molla; N Rao Cheepurupalli; Girmay Hailu; Negassi Debeb; Assefa Hailesilasie
Abstract
The Dallol Depression, located in the northern Danakil Depression, has a complex geological history shaped by Afar rifting, containing approximately 1.7 km of evaporite deposits. These deposits, heavily influenced by volcanic activity and extensional tectonic faulting, exhibit significant structural ...
Read More
The Dallol Depression, located in the northern Danakil Depression, has a complex geological history shaped by Afar rifting, containing approximately 1.7 km of evaporite deposits. These deposits, heavily influenced by volcanic activity and extensional tectonic faulting, exhibit significant structural variability. This research focuses on the potash-bearing section of the salt sequence, which consists of several distinct layers including the marker bed, sylvinite member, upper carnallitite member, bischofitite member, lower carnallitite member, and kainitite member. Employing satellite imagery (Landsat Thematic Mapper), geological and structural mapping, borehole data, and seismic analysis, this study characterizes the sub-surface features of the evaporites and estimates their reserves. The RockWorks software facilitated the development of a subsurface stratigraphic map and a three-dimensional fence diagram for enhanced interpretation. Seismic data indicate that while the upper layers of the evaporite deposits are largely horizontal and undeformed, deeper layers exhibit considerable tectonic disturbance. Thickness variations were observed, with evaporite and alluvial deposits being thinner at the southeastern rim and thicker in the eastern concession center. The total potash reserve is estimated at approximately 2.96 billion tons, of which 877.76 million tons (29.60%) remain unexploited. Current borehole designs restrict the company's extraction capacity to 24.64%. This study recommends revising mining strategies, incorporating updated borehole designs and advanced geophysical methods to improve potash recovery and promote sustainable practices in the Dallol region.
Exploitation
Yehia Z. Darwish; Abdelrahem Khalefa Embaby; Samir Selim; Darwish El Kholy; Hani Sharafeldin; Hussin Farag
Abstract
The younger granites of Gabal Gattar area, Northern Eastern Desert of Egypt, host hydrothermal uranium mineralization at the northern segment of Gattar batholith and along its contacts with the oldest Hammamat sediments. The host rocks display many features of hydrothermal overprint results in changing ...
Read More
The younger granites of Gabal Gattar area, Northern Eastern Desert of Egypt, host hydrothermal uranium mineralization at the northern segment of Gattar batholith and along its contacts with the oldest Hammamat sediments. The host rocks display many features of hydrothermal overprint results in changing their basic engineering characteristics as a function of variations of the degree of alteration. Progression from less altered to altered and mineralized rocks as the result of the alteration processes was assessed by the chemical index of alteration (CIA). The CIA numerical values were calculated by the molecular proportion of Al to the cations Ca, Na, and K. The studied rocks were divided into five grades according to degree of alteration and strength properties including: fresh (AG-I), slightly altered (AG-II), moderately altered (AG-III), highly altered (AG-IV) and very highly altered (AG-V). The strength properties of the studied rock units correlated well with the alteration grades assigned to them. That is, as the grade increased from AG-I to AG-V, abrasion resistance and crushability index increased, whereas compressive strength, slake durability and impact strength decreased.
Exploitation
Marco Antonio Cotrina Teatino; Jairo Jhonatan Marquina Araujo; Eduardo Manuel Noriega Vidal; Jose Nestor Mamani Quispe; Johnny Henrry Ccatamayo Barrios; Joe Alexis Gonzalez Vasquez; Solio Marino Arango Retamozo
Abstract
The primary objective of this research was to apply machine learning techniques to predict the production of an open pit mine in Peru. Four advanced techniques were employed: Random Forest (RF), Extreme Gradient Boosting (XGBoost), K-Nearest Neighbors (KNN), and Bayesian Regression (RB). The methodology ...
Read More
The primary objective of this research was to apply machine learning techniques to predict the production of an open pit mine in Peru. Four advanced techniques were employed: Random Forest (RF), Extreme Gradient Boosting (XGBoost), K-Nearest Neighbors (KNN), and Bayesian Regression (RB). The methodology included the collection of 90 datasets over a three-month period, encompassing variables such as operational delays, operating hours, equipment utilization, the number of dump trucks used, and daily production. The data were allocated 70% for training and 30% for testing. The models were evaluated using metrics such as Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), Variance Accounted For (VAF), and the Coefficient of Determination (R2). The results indicated that the Bayesian Regression model was the most effective in predicting production in the open pit mine. The RMSE, MAPE, VAF, and R2 for the models were 3686.60, 3581.82, 4576.61, and 3352.87; 12.65, 11.09, 15.31, and 11.90; 36.82, 40.72, 1.85, and 47.32; 0.37, 0.41, 0.41, and 0.47 for RF, XGBoost, KNN, and RB, respectively. This research highlights the efficacy of machine learning techniques in predicting mine production and recommends adjusting each model's parameters to further enhance outcomes, significantly contributing to strategic and operational management in the mining industry.
Exploitation
Meisam Saleki; Reza Khaloo Kakaie; Mohammad Ataei; Ali Nouri Qarahasanlou
Abstract
One of the most critical designs in open-pit mining is the ultimate pit limit (UPL). The UPL is frequently computed initially through profit-maximizing algorithms like the Lerchs-Grossman (LG). Then, in order to optimize net present value (NPV), production planning is executed for the blocks that ...
Read More
One of the most critical designs in open-pit mining is the ultimate pit limit (UPL). The UPL is frequently computed initially through profit-maximizing algorithms like the Lerchs-Grossman (LG). Then, in order to optimize net present value (NPV), production planning is executed for the blocks that fall within the designated pit limit. This paper presents a mathematical model of the UPL with NPV maximization, enabling simultaneous determination of the UPL and long-term production planning. Model behavior is nonlinear. Thus, in order to achieve model linearization, the model has been partitioned into two linear sub-problems. The procedure facilitates the model solution and the strategy by decreasing the number of decision variables. Naturally, the model is NP-Hard. As a result, in order to address the issue, the Dynamic Pit Tracker (DPT) heuristic algorithm was devised, accepting economic block models as input. A comparison is made between the economic values and positional weights of blocks throughout the steps in order to identify the most appropriate block. The outcomes of the mathematical model, LG, and Latorre-Golosinski (LAGO) algorithms were assessed in relation to the DPT on a two-dimensional block model. Comparative analysis revealed that the UPLs generated by these algorithms are consistent in this instance. Utilizing the new algorithm to determine UPL for a 3D block model revealed that the final pit profit matched LG UPL by 97.95%.
Exploitation
Elham Lotfi; Javad Gholamnejad; Mehdi Najafi; Mohammad Sadegh Zamani
Abstract
In the context of open pit mining operations, long-term production scheduling faces significant challenges due to inherent uncertainties, particularly in commodity prices. Traditional mathematical models often adopt a single-point estimation strategy for commodity price, leading to suboptimal mine plans ...
Read More
In the context of open pit mining operations, long-term production scheduling faces significant challenges due to inherent uncertainties, particularly in commodity prices. Traditional mathematical models often adopt a single-point estimation strategy for commodity price, leading to suboptimal mine plans and missed production targets. The simultaneous effect of commodity price uncertainty on the cut-off grade and long-term production scheduling is less considered. This paper introduces a novel model for optimizing open pit mine long-term production scheduling under commodity price uncertainty considering a dynamic cut-off grade strategy, based on a two-stage Stochastic Production Programming (SPP) framework. The presented model seeks to identify optimal mining block sequences, maximizing total discounted cash flow while penalizing deviations from production targets. To illustrate the model's efficiency, it was implemented in a copper mine. First, the Geometric Brownian Motion (GBM) model is used to quantify the future commodity price. Then, both deterministic and SPP models were solved using GAMS software. The results showed that the practical NPV obtained from the SPP model is approximately 3% higher than the DPP model, while all constraints are satisfied.
