https://jgm.birjand.ac.ir/ Journal of Geomine en daily 1 Mon, 01 Jun 2026 00:00:00 +0330 Mon, 01 Jun 2026 00:00:00 +0330 https://jgm.birjand.ac.ir/article_4160.html The depletion of high-grade iron ore reserves has increased interest in beneficiating fine-grained, low-grade iron ores, including those from the Tang Zagh deposit in southern Iran. This study evaluated multi-gravity separation (MGS), wet high-intensity magnetic separation (WHIMS), and a combined MGS–WHIMS flowsheet for upgrading a low-grade iron ore sample containing 45.56% total Fe. Individual tests were conducted on three particle size fractions with d90 values of 38, 75, and 150 µm using three-stage rougher–cleaner–recleaner circuits under fixed operating conditions. In addition, combined tests were conducted on the particle size fractions with d90​ values of 38 and 75 µm. Metallurgical performance was compared using concentrate grade, iron recovery, and the separation efficiency (S.E.) index. Both MGS and WHIMS showed a clear grade–recovery trade-off with decreasing particle size. WHIMS produced the highest concentrate grades, reaching 62.18% Fe, but showed substantially lower recovery, particularly in the finest fraction, where recovery declined to 13.60%. In contrast, MGS provided a better balance between grade and recovery, achieving 60.56% Fe at 39.13% recovery for the d90 = 38 µm fraction and outperforming WHIMS in overall separation efficiency. Relative to the three-stage WHIMS tests, the combined MGS–WHIMS circuit resulted in higher recovery and separation efficiency (S.E.), while producing concentrates with 62.10% Fe at 27.72% recovery for the d90 = 38 µm fraction and 60.79% Fe at 34.92% recovery for the d90 = 75 µm fraction. However, its overall separation efficiency remained lower than that of a standalone three-stage MGS. These results indicate that a three-stage MGS is the most effective route among the tested options for beneficiating fine-grained Tang Zagh iron ore. While, the combined circuit may be considered when a higher concentrate grade is prioritized over maximum recovery. https://jgm.birjand.ac.ir/article_4163.html Accurate ore grade estimation is critical for effective mine planning, resource evaluation, and economic feasibility in mineral deposits. Traditional geostatistical methods such as ordinary kriging (OK) and inverse distance weighting (IDW) provide robust spatial predictions but struggle to handle non-stationary and highly heterogeneous ore bodies. Machine learning algorithms (MLAs), including artificial neural networks (ANNs) and support vector regression (SVR), offer flexibility in capturing nonlinear relationships but may suffer from overfitting and unstable uncertainty quantification. This study introduces a hybrid ensemble framework that integrates geostatistical and machine learning models for enhanced ZnS grade estimation in the Gushfil Pb–Zn deposit, Irankuh district, Iran. A total of 518 composite drill-core samples taken from layer No. 3 of the deposit were analyzed, and five base learners (OK, IDW, kNN, ANN, and SVR) were implemented. Three ensemble strategies (Basic Ensemble Method (BEM), Generalized Ensemble Method (GEM), and Locally Weighted Ensemble Method (LEM)) were evaluated. The results show that individual models, particularly OK and SVR, provide competitive predictive performance (R = 0.60 and 0.58, respectively). However, the adoption of ensemble strategies leads to a further improvement in prediction accuracy, resulting in the best overall performance among the evaluated methods. The LEM approach achieved the highest correlation coefficient (R = 0.66) and the lowest mean squared error (MSE = 0.015) by adaptively weighting base learners according to local performance. These findings demonstrate that hybrid ensemble frameworks effectively combine spatial structure and nonlinear modeling, offering a robust and reproducible approach for ore grade estimation in complex and heterogeneous deposits. https://jgm.birjand.ac.ir/article_4168.html Fluid loss during drilling operations is one of the primary causes of formation damage and reduced well productivity. Therefore, minimizing drilling mud filtrate volume is essential for improving drilling efficiency and maintaining wellbore integrity. In this study, the effects of three additives, namely rice husk, natural gum (apricot gum), and xanthan gum, on the filtrate volume of water-based drilling mud were experimentally investigated. Static filtration tests were conducted using additive concentrations of 0.5, 1, 2, and 3 gr and filtration times ranging from 0.25 to 7.5 min. The results indicated that increasing the additive concentration consistently reduced filtrate volume for all tested additives. For all additives, adding 0.5 g caused the largest incremental decrease in filtrate volume, while further increases in concentration produced diminishing reductions. Xanthan gum exhibited the highest filtration-control efficiency, followed by apricot gum and rice husk. At the maximum additive concentration, filtrate volume was reduced by 60.0%, 84.0%, and 88.4% for rice husk, apricot gum, and xanthan gum, respectively, compared with the base mud. To estimate filtrate volume under arbitrary additive concentrations and filtration times, a two-layer feedforward artificial neural network (ANN) with 10 neurons in the hidden layer was developed. The ANN demonstrated great predictive capability, achieving overall correlation coefficients (R) of 0.99979, 0.99988, and 0.99991 for rice husk, apricot gum, and xanthan gum datasets, respectively. The corresponding minimum mean squared errors (MSEs) were 1.52 × 10⁻⁵, 1.08 × 10⁻⁵, and 8.56 × 10⁻⁶, while the average prediction errors were 0.59%, 1.27%, and 1.86%, respectively. These results confirm the effectiveness of the investigated additives for fluid-loss control and demonstrate the reliability of the proposed ANN model for rapid filtrate-volume prediction. https://jgm.birjand.ac.ir/article_4169.html Tunnel squeezing is one of the major challenges in mechanized longwall coal mining and is primarily caused by stress redistribution following coal extraction. In main gate tunnels, the support system is subjected to significant loads during panel extraction and longwall advancement. The Tabas Parvadeh No. 1 Coal Mine, which uses a mechanized longwall mining method, has experienced severe squeezing problems in its main gate tunnels at depths greater than 400 m. In certain sections, the magnitude of deformation following coal extraction has exceeded the capacity of the existing support system, necessitating the evaluation of alternative support designs. In this study, the squeezing behavior of the main roadways of the Parvadeh No. 1 Mine is investigated using the Hoek analytical method in conjunction with finite element modeling. Based on the results obtained, the tunnel support system has been redesigned to effectively withstand the squeezing pressures encountered at the current mining depth. It was determined that using TH40 steel arches installed at a spacing of 0.5 meters is sufficient to endure the pressures exerted by the roof and sidewalls. However, it is essential to also reinforce the tunnel floor with an additional support system, such as a concrete slab. https://jgm.birjand.ac.ir/article_4254.html The application of sodium cyanide for gold leaching remains the predominant method for gold extraction. However, the efficiency of the gold leaching process depends on several interacting performance variables, including cyanide concentration, pH, pulp density, particle size, leaching time, dissolved oxygen, agitation rate, and temperature. The combined influence of performance variables on gold recovery has not been fully optimized. Therefore, a quantitative assessment is required to understand their relational strength effects. A literature review identified eight key performance variables commonly monitored in gold leaching operations, both with and without pretreatment methods. The effects of performance variables on gold recovery were evaluated using grey relational analysis and the coefficient of variation. Artificial intelligence, particularly artificial neural networks integrated with statistical methods, was also explored to predict the effects of variables on strength and gold recovery efficiency. The analysis produced mean relational degrees of 0.59 and 0.54, with coefficients of variation of 9.0% and 10.8% for processes without and with pretreatment methods, respectively. The relational degree indicated moderate variability across the datasets. The findings confirm that input performance variables strongly influence gold leaching efficiency. The dosage ranges identified in this study can serve as valuable benchmarks for optimizing gold recovery in industrial leaching operations.