Deep Mining Research Articles

Research on work-stress recognition for deep ground miners based on depth-separable convolutional neural network

In deep mine production operations, the challenging operating environment intensifies the workload and pressure on coal miners. Long-term exposure to high-intensity operating pressure can seriously impact the physical and mental health of miners, leading to unsafe behaviors and accidents. To identify the pressure of miners' operations, this paper examines various driving scenarios, such as the deep-well tunneling machine cutting the wall and opening the alley, the shoveling machine shoveling ore, and the pickup truck driver transporting. The paper randomly collects facial images of miners during each operation using an explosion-proof CCD camera to obtain the facial expression characteristic data of miners. The Ferface2013 facial expression dataset was used to establish the dataset. The depth separable convolutional neural network MiniXception was used for training and to output the classification results of the pressure degree of deep shaft miners. A MiniXception-based miners' operating pressure recognition model was established. The training time, precision, recall, F1 score, and classification accuracy confusion matrix were selected. The study evaluated the effectiveness of the recognition model by measuring its training time, precision, recall, F1 score, and classification accuracy confusion matrix. The results indicate that the model has a correct recognition rate of 88% for the pressure state, 91% for the pleasure state, and 74% for the normal state. The overall accuracy of the model is 0.843. Therefore, the MiniXception recognition model is suitable for recognizing the pressure of miners' operations in deep mines. This can meet practical needs and is useful for preventing major accidents in mines, managing on-site safety, and managing safety in non-hazardous areas. It has important theoretical and practical significance.

Factor of safety analysis for mine pillar considering the influence of the intermediate principal stress component

Failure of mine pillars, especially in deep underground mines, significantly threatens the safety of miners and equipment. Previous studies on mine pillar stability design have used classical constitutive models that ignore the intermediate principal stress component when determining the factor of safety. In this study, we develop and implement a three-dimensional modified Hoek–Brown (HB) constitutive model that incorporates the intermediate principal stress component into the numerical simulation tool FLAC3D. Furthermore, we propose and apply a strength-reduction technique to determine a more accurate factor of safety for mine pillars. This novel approach provides a more comprehensive and realistic method for geomechanical analysis and pillar design, enhancing our understanding of pillar stability. Through numerical analysis, we illustrate the impact of the intermediate principal stress component on mine pillar plasticity. The factor of safety is calculated via the strength reduction method, revealing a substantial improvement from 1.7 with the classical HB model to 2.0 with the 3D HB model. Including the intermediate principal stress component reduces the evolution of plasticity in the mine pillar. For instance, the volume of plastic zones diminishes, and the factor of safety increases as the width-to-height ratio increases. Exemplary simulations show that ignoring the effect of the intermediate principal stress component, including underestimating safety levels, designing suboptimal pillar design, and misinterpreting in situ observations and measurements, can lead to severe consequences.

Management of Thermal Hazards in Deep Mines in China: Applications and Prospects of Mine Cooling Technology

With the continuous development of the mining industry and advancements in deep mining technology, mine environment optimization has become key to ensuring safety and improving the efficiency of mining. The high-temperature environment, particularly in deep mines, not only poses a serious threat to miners’ health but also significantly reduces operational efficiency. These issues have been determined based on the current application status and development trends of mine cooling technology, including traditional mechanical and non-mechanical cooling technologies, as well as emerging roadway insulation materials and mine cooling clothing applications. By comparing the advantages and disadvantages of each technology, the main challenges related to the use of current mine cooling technologies are pointed out, including the low energy efficiency ratio, high cost, and difficult implementation. Finally, this paper looks forward to the future development directions of mine cooling technologies, emphasizing the importance of intelligent, energy-saving, and environment-improving comprehensive system management and, in turn, promoting the progress and application of mine environment optimization technology and supporting safe and efficient deep mining.

Comparison of artificial refrigeration cooling schemes for high temperature tunneling roadway in deep mines--the case of Yunnan Dahongshan Copper Mine

With the increasing depth of metal mine mining, the heat damage in deep mine tunneling roadway has become more and more serious, in order to effectively reduce the temperature in high temperature tunneling roadway, a mobile artificial refrigeration equipment was developed. To improve the efficiency of cold-heat exchange of air, enhance the cooling effect of the refrigeration equipment, evaluate the applicability of refrigeration equipment and determine the optimal cooling scheme, field tests under different conditions were conducted in Dahongshan Copper Mine in Yunnan Province. Additionally, numerical simulation studies were carried out by Fluent in order to analyze the temperature distribution characteristics of the roadway under different conditions. The results show that: (1) The mobile artificial refrigeration equipment can reduce the wet bulb temperature in the cooling space to below 30 °C, meeting the requirements of the "Safety regulation for metal and nonmetal mines" (GB16423-2020), and condition 4 has the best cooling effect, which can reduce the wet bulb temperature in the cooling space to below 23.64 °C; (2) The addition of cold air supply duct, wind barriers and heat exhaust ducts enhances the cooling effect in the cooling space; (3) The highest average axial temperature of the roadway is observed at the location of 6m from hot air outlets of the refrigeration equipment or heat exhaust ducts outlets, and the highest temperature is about 9 °C higher than the initial ambient temperature of the roadway; (4) The cold air in the cooling space gradually contracts towards the cold source outlet as the distance from the working face increases. The research results can provide guidance for the prevention and control of heat damage and the development of artificial refrigeration equipment for high temperature tunneling roadway in metal mines.

Application of artificial intelligence in coal mine ultra-deep roadway engineering—a review

The deep integration of computer field and coal mining field is the only way to coal mine intellectualization. A variety of artificial intelligence tools have been applied in open-pit and shallow coal mines. However, with the geometric increase of coal demand, the contradiction between supply and demand is becoming more and more serious, and the exploitation of resources from shallow layer (> 600 m) has become an inevitable trend. Well then, as a new engineering scene, the harsh conditions of “three high and one disturbance” seriously threaten the safety of personnel. The superposition of complex mining environment makes the number of input factors increase sharply, which leads to the application of artificial intelligence in deep coal mine roadway engineering. The guidance is not mature, the construction of various databases is missing, and there are still some problems in universality and applicability. To this end, this paper starts with the introduction of the operating characteristics of various artificial intelligence tools, and conducts a comprehensive study of the relevant high-level articles published in top journals. It systematically sorts out the research progress that has successfully solved the five directions of rock mechanics strength, surrounding rock stability, rock-burst, roof fall risks and micro-seismic events. While objectively evaluating the comprehensive performance of different tools, it also expounds its own views on key research and results. Literature review shows that whether as a development tool or a comparative model, the application of ANN in the field of coal mining is more than 98%, and it performs extremely well in the direction of surrounding rock stability and roof fall risk, with an accuracy rate of more than 90%. As the most mature direction of AI tools application, rock mechanical strength has experienced the development process from “SVM → ANN → DL → XGBoost → RF”. The dataset is from small samples (< 100) to big samples (> 1000), and the R2 of tree-based models can be stabilized at more than 95%. The research on rock-burst prediction mainly focuses on field factors and micro-seismic monitoring data. Whether it is a small sample or a large-scale data model, the accuracy of BN remains above 85%. The prediction and evaluation of micro-seismic events is a new direction in recent years. The image processing and application of CNN is extremely important. The signal recognition and classification accounts for more than 90%, and the research potential of source location needs to be further explored. In general, the nature of the rock itself is the first choice for almost all influencing factors. At the same time, the update iteration of monitoring methods (micro-seismic, ground sound, separation, deformation, etc.) expands the development of the database, making it possible to obtain the data due to threat to life and cost of equipment, which is very difficult to obtain before. In the process of parameter selection at the input end, the method of combining lithology conditions, geological environment and monitoring data will gradually become the first choice for research. Finally, in the follow-up work collation and on-the-spot investigation, it mainly focuses on the existing problems of deep coal mines, explores the application potential of artificial intelligence in deep coal mine roadway engineering, puts forward the possible research focus and challenging problems in the future, and gives its own opinions.

Data-driven dynamic inclination angle estimation of monorail crane under complex road conditions

Abstract Monorail cranes are crucial in facilitating auxiliary transportation within deep mining operations. As unmanned driving technology becomes increasingly prevalent in monorail crane operations, it encounters challenges such as low accuracy and unreliable attitude recognition, significantly jeopardizing the safety of monorail crane operations. Hence, this study proposes a dynamic inclination estimation methodology utilizing the Estimation-Focused-EKFNet algorithm. Firstly, based on the driving characteristics of the monorail crane, a dynamic inclination model of the monorail crane is established, based on which the dynamic inclination value can be calculated in real-time by the extended Kalman filter (EKF) estimator; however, given the complexity of the driving road conditions, in order to improve the dynamic inclination recognition accuracy, the CNN-LSTM-ATT algorithm combining the convolutional neural network (CNN), the long short-term memory (LSTM) neural network and the attention mechanism (ATT) is used to firstly predict the current dynamic camber is predicted by the CNN-LSTM-ATT algorithm combined with the CNN and the attention mechanism, and then the predicted dynamic inclination value is used as the observation value of the EKF estimator, which finally realizes that the EKF estimator can output the accurate dynamic inclination value in real-time. Experimental results indicate that, compared with the unscented Kalman filter, LSTM-ATT, and CNN-LSTM algorithms, the Estimation-Focused-EKFNet algorithm enhances dynamic inclination recognition in complex road conditions by at least 52.34%, significantly improving recognition reliability. Its recognition accuracy reaches 99.28%, effectively ensuring the safety of unmanned driving for monorail cranes.

Discovery of high-expressing lncRNA-derived sORFs as potential tumor-associated antigens in hepatocellular carcinoma.

This study is based on deep mining of Ribo-seq data for the identification of lncRNAs that have highly expressed sORFs in HCC. In this paper, dynamic prospects associated with sORFs acting as newly defined tumor-specific epitopes are discussed with possible improvement in strategies for tumor immunotherapy. Using ribosome profiling to identify and characterize sORFs within lncRNAs in HCC, identify potential therapeutic targets and tumor-specific epitopes applicable for immunotherapy. MetamORF performed the identification of sORFs with deep analysis of the data of ribosome profiling in lncRNAs associated with HCC. The translation efficiency in these molecules was estimated, and epitope prediction was done by pVACbind. Peptide search was done to check the presence of micropeptides translated from these identified sORFs. validated translational activity and identified potential epitopes. Higher translation efficiency was noted in the case of lncRNAs associated with HCC compared to normal tissues. Of particular note is ORF3418981, which results in the highest expression and has supporting experimental evidence at the protein level. Epitope prediction identified a putative epitope at the C-terminus of ORF3418981. This study uncovers the as-yet-unknown potential of lncRNA-derived sORFs as sources of tumor antigens, shifting the research focus from protein-coding genes to non-coding RNAs also in the HCC context. Moreover, this study highlights the contribution of a subset of lncRNAs, especially LINC00152, to the development of tumors and modulation of the immune response by its sORFs.

Development of a Deep Neural Network Model for the Relocation of Mining-Induced Seismic Event

The precise relocation of seismic events is critical for many engineering projects. Swarms of minor or micro earthquakes typically reveal stress concentration and spots of greater seismic hazards. Particularly in the context of deep underground mining, advanced techniques that can accurately relocate microseismicities are urgently in demand. Here, we developed a neural network-based modeling training method that can precisely relocate seismicities and invert for velocities at the same time, with preconfigured receiver network locations. Our model can be iteratively improved with field recorded data. We showed that, with roughly eight iterations, we can reasonably resolve for the earthquake locations for both clusters of events, namely spatially distributed with linear pattern or randomly scattered. Our initially trained model, which only focused on events that had a linear distribution pattern, was used as the base for the training of the subsequent models which could better resolve for randomly scattered event locations. Although we stopped at the eighth iteration, the process reported here can be continued, as the model will have a better performance with more iterations.

Pore Morphology and Oxidation Behaviour of Deep-Loaded Granular Coal at High Initial Temperatures

ABSTRACT As coal mining proceeds deeper, fragmented coal is subjected to high geostress and elevated thermal environments, causing changes in its micro-physical properties and spontaneous combustion (SC) risks. Deep mining operations entail significant risks of coal fires. In order to investigate the effects of high stress and high thermal environments on the microstructural and physicochemical damage of fragmented coal, and to elucidate the mechanisms influencing oxidation behavior, we conducted experiments including low-temperature nitrogen adsorption, synchronous thermal analysis, and in-situ diffuse reflectance spectroscopy. These experiments aimed to explore the pore morphology and oxidation characteristics of high initial temperature unloading granular coal (HITU-GC) unloading conditions. The results indicate that with increasing initial temperature, characteristic point temperatures initially rise and then decrease, while both the mass loss and heat release during the low-temperature oxidation stage decrease. After pressure-unloading composite treatment, the oxidation process accelerates with more intense reactions, resulting in an overall increase in heat release. The pore shapes of HITU-GC show no significant difference compared to the original coal, mainly comprising non-rigid aggregates of micropores, platy or layered matrix particles, including numerous slit-shaped pores. The coal surface also exhibits ink bottle-shaped pores and cone-shaped pores. Higher pre-oxidation temperatures correlate with higher specific surface areas. Higher stress levels from pressurized unloading correlate with lower specific surface areas. The specific surface area of sample O60-P16 is reduced by 0.879 m2/g compared to sample O60-P0. Pressurized unloading treatment disrupts mesoporous pores. Following thermal environment and pressure-unloading, the coal’s aromatic hydrocarbon content decreases, while the contents of hydroxyl, aliphatic hydrocarbons, and oxygen-containing functional groups (FG) all increase. Sample O60-P4 shows the highest content of active FG. While thermal environments enhance coal oxidation activity, pressure-unloading treatment increases the extent of coal surface fragmentation. The synergistic effect of both significantly increases the SC risk of coal. This study provides a theoretical basis for controlling thermal hazards associated with high-initial-temperature unloading fragmented coal in deep goaf areas.

An effective multi-modal adaptive contextual feature information fusion method for Chinese long text classification

Chinese long text classification plays a vital role in Natural Language Processing. Compared to Chinese short texts, Chinese long texts contain more complex semantic feature information. Furthermore, the distribution of these semantic features is uneven due to the varying lengths of the texts. Current research on Chinese long text classification models primarily focuses on enhancing text semantic features and representing Chinese long texts as graph-structured data. Nonetheless, these methods are still susceptible to noise information and tend to overlook the deep semantic information in long texts. To address the above challenges, this study proposes a novel and effective method called MACFM, which introduces a deep feature information mining method and an adaptive modal feature information fusion strategy to learn the semantic features of Chinese long texts thoroughly. First, we present the DCAM module to capture complex semantic features in Chinese long texts, allowing the model to learn detailed high-level representation features. Then, we explore the relationships between word vectors and text graphs, enabling the model to capture abundant semantic information and text positional information from the graph. Finally, we develop the AMFM module to effectively combine different modal feature representations and eliminate the unrelated noise information. The experimental results on five Chinese long text datasets show that our method significantly improves the accuracy of Chinese long text classification tasks. Furthermore, the generalization experiments on five English datasets and the visualized results demonstrate the effectiveness and interpretability of the MACFM model.

Effectiveness Spatiotemporal of Pressure Unloading and Gas Extraction in Remote Upper Protected Seam Mining.

Severe gas hazards are increasing in deep mining areas, and there exists an enormous traditional challenge for protected seam mining. In this work, we conducted a multifaceted investigation into the spatiotemporal effectiveness exerted by pressure unloading from a distant safeguarded stratum in concert with the application of gas extraction methodologies. The stress and displacement evolution in the protected coal seam (PCS) are analyzed by applying Flac3D, and then positives are verified by the measurement of coal seam deformation and investigation of the unloading boundary. The results show that diminution coefficient of 0.62, and record an expansive deformation rate of 5.83%. The opportune temporal window for effective gas drainage is discerned, with its time window spans from 58 to 67 days as the expanse between 350 and 400 m progresses at the working face. Continuous and effective gas extraction occurs when transverse cracks in the PCS reach an optimal state of pressure relief. The final cumulative gas extraction was 320,300 m3 with an extraction efficiency of 34.3%. A residual gas volume of 3.68 m3/t and a residual gas pressure of 0.5 MPa reflect the efficient gas extraction.

Failure mechanism analysis and support strength determination of deep coal mine roadways – A case study

In order to solve the problems related to deformations, failures, and difficulties in support controls of deep coal mine roadways, the typical failure forms and characteristics of deep coal mine roadways were investigated and analyzed in this study. The distribution laws of plastic zones of circular roadways under different lateral pressure coefficients were expounded using a theoretical analysis method, and the influence of plastic zone scopes on the failure forms of roadways was explained. In addition, considering the distribution characteristics of plastic zones, a calculation method of the required support strength based on the plastic zone scopes was proposed. The primary and secondary factors affecting the support strength were then determined. A numerical simulation method was applied in order to analyze the distribution characteristics of the displacements and plastic zones of deep mine circular roadways under the conditions of no support scheme, an existing support scheme, and a support scheme determined by the proposed support strength calculation method (referred to in this study as the optimized support scheme), respectively. The simulation results showed that the support effects of the optimized support scheme were superior to those of the existing support scheme, and thereby more conducive to maintaining roadway stability. In addition, an on-site engineering test method was used to identify support schemes for bolting and shotcrete with wire mesh + concrete-filled steel tubular supports based on the plastic zone scopes and the support strength of roadways in order to determine the support parameters. As indicated by the field monitoring results, the proposed support strength calculation method and optimized support scheme had shown the ability to effectively control the large deformations of the tested roadway, and thereby guarantee its overall stability.

Dataset of miRNA-disease relations extracted from textual data using transformer-based neural networks.

MicroRNAs (miRNAs) play important roles in post-transcriptional processes and regulate major cellular functions. The abnormal regulation of expression of miRNAs has been linked to numerous human diseases such as respiratory diseases, cancer, and neurodegenerative diseases. Latest miRNA-disease associations are predominantly found in unstructured biomedical literature. Retrieving these associations manually can be cumbersome and time-consuming due to the continuously expanding number of publications. We propose a deep learning-based text mining approach that extracts normalized miRNA-disease associations from biomedical literature. To train the deep learning models, we build a new training corpus that is extended by distant supervision utilizing multiple external databases. A quantitative evaluation shows that the workflow achieves an area under receiver operator characteristic curve of 98% on a holdout test set for the detection of miRNA-disease associations. We demonstrate the applicability of the approach by extracting new miRNA-disease associations from biomedical literature (PubMed and PubMed Central). We have shown through quantitative analysis and evaluation on three different neurodegenerative diseases that our approach can effectively extract miRNA-disease associations not yet available in public databases. Database URL: https://zenodo.org/records/10523046.

Seepage-heat accumulation characteristics and hydrothermal transport calculation model in mining-induced fractured rock mass: Implications for geothermal water control

The aim of this paper is to explore hydrothermal transport in surrounding rocks during deep mining, focusing on the mechanisms and key factors influencing groundwater seepage and heat accumulation disasters. Stress–seepage coupling tests were conducted on fractured rock mass with varying roughness under constant-pressure loading/unloading conditions. Also, a permeability model considering joint roughness coefficient and confining pressure was established. The study further examined the evolution of heat transfer in rough fractured rock mass and variations in convective heat transfer characteristics. The heat transfer process was observed to progress through stages of surge, attenuation, and stabilization. Two calculation models were developed using multivariate linear regression and random forest regression to quantify how different factors influenced water–rock convective heat transfer in mining-induced fractured rock mass. The average correlation coefficient R2 of the proposed models exceeded 0.95. The importance scores of osmotic pressure, inlet water temperature, rock temperature, and permeability were 0.417, 0.296, 0.223, and 0.065, respectively. The flow rate and water–rock temperature difference were determined to be the key factors influencing hydrothermal transport. This study enhanced the understanding of the hydrothermal migration mechanisms in surrounding rocks during mining and provided insights into geothermal water control.

Fractal evolution characteristics of fracture meso-damage in uniaxial compression rock masses using bonded block model

With regard to deep mining in metal mines, an investigation into the failure mode of deep fractured rock masses and their corresponding acoustic emission signal characteristics is conducted via uniaxial compression tests. Subsequently, a fractal damage renormalization group mechanical model is developed to explain the behavior of those fractured rock masses. Employing the bonded block model (BBM) numerical simulation method, fracture process in synthetic rock samples is analyzed, thereby validating the efficacy of the mechanical model. The numerical simulations highlight the critical role of fractures expansion in underlying the deterioration of rock mass strength. As the peak load decreases, the fracture fractal dimension increases, leading to a significant 14.2% reduction in compressive strength accompanied by an approximate 8.7% rise in average fracture fractal dimension. A comparative analysis of tetrahedral and voronoi block synthetic rock samples reveals the tetrahedral block samples exhibit a superior ability to depict the fracture behavior of fractured rock masses. Specifically, they offer a more accurate simulation of acoustic emission characteristics and failure modes. Furthermore, variations in the fracture fractal dimension with respect to the hole defect’s position are observed, with the maximum value occurring along the vertical axis of the hole defect. This observation underscores the potential utility of visually monitoring deep rock fracture dynamics as an effective mean for quantitatively evaluating fracture damage and strength degradation in deep rock formations.

Research on Key Parameters for Relieving Pressure on Roofs of Deep Mine Cutting and Retaining Roadways

Scientific and reasonable roof-cutting parameters are key to ensuring pressure relief of the retained roadway roof. This manuscript takes the 7135 working face of Qidong Coal Mine as the engineering background and uses theoretical analysis, numerical and on-site measurement methods to study the quantitative relationship and pressure relief effect between different roof-cutting parameters of GERRC. We established a fracture criterion based on the tensile strength of the main roof of the uncut joint along the cutting line. We analyzed the quantitative relationship between different main roof thickness, cutting height, cutting angle, and the main roof tensile stress of the uncut joint. We found that within a small range of cutting angles, as the cutting angle decreases, the tensile stress on the main roof of the uncut joint increases. When the cutting angle is 0, the main roof tensile stress of the uncut joint reaches its maximum. As the cutting height increases, the limitation of the cutting angle on the cutting height becomes smaller. Numerical simulation was conducted to study the distribution patterns of maximum and minimum principal stresses along the direction of the roadway roof during the retention period under different roof-cutting heights and angles. Based on this, the optimal unloading effect of the roadway roof and the minimum concentration of mining stress were obtained at a roof-cutting height of 9 m and a roof-cutting angle of 80°. Through on-site measurement of the stress on the reinforcement anchor cable during the retention period, the deformation of the sinking roadway roof, and the pressure relief control effect of the retained roadway roof, the pressure relief effect and scientific rationality of the design of the cutting height and cutting angle were verified.

Study on dynamic mechanical properties and microscopic damage mechanisms of granite after dynamic triaxial compression and thermal treatment

As underground mining operations gradually extend deeper, the conditions for orebody occurrence become increasingly complex, and various geological disasters occur frequently. Rock masses are prone to different degrees and types of damage, making it impractical to continue using intact rock as a reference. To study the dynamic mechanical properties of damaged rock under actual conditions, this study subjected granite samples to impact and high-temperature damage. Detailed observations were made of the samples' surface morphology and microstructure before and after damage, and the patterns of damage changes were investigated. Subsequently, uniaxial compression tests at different loading rates were conducted on the damaged samples. By calculating the loading rate effect sensitivity, it was found that as the damage increased, the rate effect gradually diminished. In addition, this study also summarized the influence of damage and loading rate on the macroscopic failure characteristics of the samples. The novelty of this study lies in focusing on damaged rock, which more closely resembles the actual rock conditions encountered in most underground mining operations today. This research can provide a reference for stability assessment and safe construction in deep mining rock engineering and offers important insights into the feasibility of non-explosive extraction of damaged rock.

Predicting mine water inflow volumes using a decomposition-optimization algorithm-machine learning approach

Disasters caused by mine water inflows significantly threaten the safety of coal mining operations. Deep mining complicates the acquisition of hydrogeological parameters, the mechanics of water inrush, and the prediction of sudden changes in mine water inflow. Traditional models and singular machine learning approaches often fail to accurately forecast abrupt shifts in mine water inflows. This study introduces a novel coupled decomposition-optimization-deep learning model that integrates Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Northern Goshawk Optimization (NGO), and Long Short-Term Memory (LSTM) networks. We evaluate three types of mine water inflow forecasting methods: a singular time series prediction model, a decomposition-prediction coupled model, and a decomposition-optimization-prediction coupled model, assessing their ability to capture sudden changes in data trends and their prediction accuracy. Results show that the singular prediction model is optimal with a sliding input step of 3 and a maximum of 400 epochs. Compared to the CEEMDAN-LSTM model, the CEEMDAN-NGO-LSTM model demonstrates superior performance in predicting local extreme shifts in mine water inflow volumes. Specifically, the CEEMDAN-NGO-LSTM model achieves scores of 96.578 in MAE, 1.471% in MAPE, 122.143 in RMSE, and 0.958 in NSE, representing average performance improvements of 44.950% and 19.400% over the LSTM model and CEEMDAN-LSTM model, respectively. Additionally, this model provides the most accurate predictions of mine water inflow volumes over the next five days. Therefore, the decomposition-optimization-prediction coupled model presents a novel technical solution for the safety monitoring of smart mines, offering significant theoretical and practical value for ensuring safe mining operations.

Enhancing stability graphs for stope design in deep metal mines using machine learning

Stope structural parameters, which are human-controllable, directly impact the safety and economics performance of underground mineral extraction. Current stope design still relies heavily on empirical methods such as stability graphs, due to the complex nature of rock masses and varied stope failure mechanisms. This study aims to enhance stability graphs with machine learning techniques. Firstly, a dataset of 980 records from unsupported stopes was compiled, representing perhaps the largest dataset of its kind so far in the literature. This was achieved through extensive literature review and the collation of an additional 289 records from Chinese mines which were previously not included. An analysis of data reveals that over 90 % of the records fall within a stability coefficient of 0–100 and a hydraulic radius of 0–20 m. Secondly, a stability graphs optimization process was established using Python, eliminating the subjectivity of partitioning. Nine supervised machine learning algorithms were employed and trained to test their performance in partitioning form and predicting accuracy. It was found that the neural network algorithm demonstrated the best overall performance. At last, a neural network with the Keras framework was used to establish a new multilayer perceptron model to generate safety factor probability curves, which were then used to construct the stability graph. To facilitate practical use, mathematical functions fitting safety factor curves within the unstable zone were further formulated. Compared with other empirical stability graphs, our new approach allows designers to more efficiently and reliably select safety factors to determine the stability state according to site specific conditions and technical support systems, thereby providing enhanced guidance for stope design.

Spatial modeling for detection of water retention capacity in technosols developed on carboniferous spoil heap after hard coal mining

Post-industrial areas, such as heaps, are an Anthropocene pressure on the environment but show natural potential for new ecosystem services, i.e., water retention, biodiversity, and C-sequestration. All these ecosystem functions of such sites can be developed only in the case of appropriate reclamation treatments, which are especially difficult under the conditions of carboniferous spoil heaps after coal deep mining. The study of reclamation effects has been increasingly supported in recent years by modern remote sensing techniques. This paper presents the possibilities of using selected remote sensing data and topographic indices and their effect on volumetric water content, which is crucial for habitat development. PlanetScope satellite imagery and airborne laser scanning point clouds were used to estimate the volumetric water content (VWC) of soil. This integrated analysis allows for more accurate spatial modeling of soil moisture, a key indicator of ecological restoration in degraded land. The method of generalized additive models was used to develop a predictive model. The developed predictive model was characterized by the following values of model accuracy evaluation parameters: MAE – 4.1%; RMSE – 5.6%; and R2–0.49. As a result of the analysis, topographic and spectral variables with significant influence on VWC were identified, and a map of the retention capacity of the Sosnica heap was developed.