Search

Abstract Mine flooding accidents have occurred frequently in recent years, and the predicting of mine water inflow is one of the most crucial flood warning indicators. Further, the mine water inflow is characterized by non-linearity and instability, making it difficult to predict. Accordingly, we propose a time series prediction model based on the fusion of the Transformer algorithm, which relies on self-attention, and the LSTM algorithm, which captures long-term dependencies. In this paper, Baotailong mine water inflow in Heilongjiang Province is used as sample data, and the sample data is divided into different ratios of the training set and test set in order to obtain optimal prediction results. In this study, we demonstrate that the LSTM-Transformer model exhibits the highest training accuracy when the ratio is 7:3. To improve the efficiency of search, the combination of random search and Bayesian optimization is used to determine the network model parameters and regularization parameters. Finally, in order to verify the accuracy of the LSTM-Transformer model, the LSTM-Transformer model is compared with LSTM, CNN, Transformer and CNN–LSTM models. The results prove that LSTM-Transformer has the highest prediction accuracy, and all the indicators of its model are well improved.

Abstract Water inrush originating from hidden faults in the coal seam floor is challenging to prevent due to their concealed nature. This paper develops a coupled stress–seepage–damage model for simulating rock fracture, implemented using the finite element method. The model is validated against compression-seepage tests on rock samples, capturing realistic dynamics of shear and tensile damage as well as permeability. The model is applied to the 27305 working face of a coal mine in Shandong Province, China, revealing the evolution of water inrush caused by a hidden fault. The results indicate that as the working face progresses, both the floor damage and the internal damage within the hidden fault escalate gradually. When mining reaches 80 m, the hidden fault has been activated internally, and the depth of floor damage reaches 13 m, which still has a certain distance from the hidden fault. At 100 m, the depth of the floor damage has stabilized, while the stress concentration at the hidden fault's tip increases, and it begins to expand if conditions for tensile damage are met. By the time mining reaches 110 m, the hidden fault has expanded 9.2 m in length and connected with the floor damage zone, forming a water inrush channel that links the aquifer to the working face, presenting a significant water inrush risk. This work provides an intuitive approach to understanding the evolution of water inrush from a hidden fault, aiding in the prevention of water inrush disasters in practical engineering applications.

This work focuses on the effect of the interval between two shocks on the ejecta formation from the grooved aluminum (Al1100) surface by using smoothed particle hydrodynamics numerical simulation. Two unsupported shocks are obtained by the plate-impact between sample and two flyers at interval, with a peak pressure of approximately 30 GPa for each shock. When the shock interval varies from 2.11 to 7.67 times the groove depth, the bubble velocity reduces to a constant, and the micro jetting factor RJ from spike to bubble exhibits a non-monotonic change that decreases initially and then increases. At a shock interval of 3.6 times the groove depth, micro jetting factor RJ from spike to bubble reaches its minimum value of approximately 0.6. While, the micro jetting factor RF from spike to free surface decreases linearly at first, and stabilizes around 0.25 once the shock interval surpasses 4.18 times the groove depth. When the shock interval is less than 4.18 times the groove depth, the unloading wave generated by the breakout of the first shock wave is superimpose with the unloading part of the second shock wave to form a large tensile area.

Porous materials are widely used in the field of protection because of their excellent energy absorption characteristics. In this work, a series of polyurethane microscopic models are established and the effect of porosity on the shock waves is studied with classical molecular dynamics simulations. Firstly, shock Hugoniot relations for different porosities are obtained, which compare well with the experimental data. The pores collapse and form local stress wave, which results in the complex multi-wave structure of the shock wave. The microstructure analysis shows that the local stress increases and the local velocity decreases gradually during the process of pore collapse to complete compaction. Finally, it leads to stress relaxation and velocity homogenization. The shock stress peaks can be fitted with two exponential functions, and the amplitude of attenuation coefficient decreases with the increase of density. Besides, the pore collapse under shock or non-shock are discussed by the entropy increase rate of the system. The energy is dissipated mainly through the multiple interactions of the waves under shock. The energy is dissipated mainly by the friction between atoms under non-shock.

As coal mines shift from shallow to deeper excavation, the number of mines facing the risk of rock burst disasters is gradually increasing. Rockburst, with their characteristics of vibration, suddenness, complexity, and unpredictability, make it increasingly difficult to prevent and control these disasters. Therefore, the challenges of preventing and controlling rock burst disasters are becoming more and more severe. This paper, based on the system-theoretic accident model and processes (STAMP) theory, extracts the causal factors affecting coal mine rock burst accidents. Using the interpretative structural modeling (ISM) and decision-making trial and evaluation laboratory (DEMATEL) method, the accident-causing factors are quantitatively assigned. By constructing model equations and drawing causal loop diagrams and stock-flow diagrams, the event is dynamically simulated and early warnings are issued. The results show that the control defects leading to the accident are analyzed from the perspectives of the government level, management level, grassroots level, physical layer, and the dynamic process of the accident. In the short term, safety investment in grassroots operations is the most effective control. In the long run, the most effective measure is for the management level to strengthen its supervisory work. By changing the input ratios of various variables, it can be seen that different variables in the system dynamics (SD) model have different impacts on coal mine rock burst accidents. It is necessary to continuously strengthen the implementation of the safety responsibility system, improve the work efficiency of the government and management level, and enhance the timeliness of emergency decision-making.

Abstract We have investigated the activation characteristics of mining faults and the effect of grouting reinforcement under thick loose layer and thin bedrock of the working face and evaluate their impact on the safety of mining at similar working faces. Implementing the geological conditions of the 331 working face of the Yangcun Coal Mine (China) of the Yankuang Energy Group Corporation, we have analyzed mechanically the process of fault activation at first. Subsequently, we have obtained the mechanical criterion of fault slip and the expression of relative strength of the nearby rock mass. Using numerical software we have simulated and analyzed the damage characteristics of different bedrock thicknesses on overlying rocks and faults in the fluid–solid coupling mode. In addition, we have studied the controlling effect of grouting reinforcement on fault activation, which has been verified in the field. The main results of our study show that: 1. The mechanical properties of the rock mass near the fault interface have changed and they are related to the cohesive force of the interface; 2. The water inrush mode of the working face changes under different bedrock thickness, and the thinner the bedrock, the less easily the fault is destroyed 3. The slip of the high-level fault is reduced after the grouting of the fault, the propagation of the fracture zone at the fault is suppressed, the seepage of the aquifer water is prevented, and the safe recovery is realized. The results of our study provide a scientific basis for the secure mining across the fault of the 331 working face of Yangcun Coal Mine. Based on the results of our study the working face can be mined safely from now on and in the future.

Many deep mining mines in southwestern Shandong Province of China are covered with thick loose layers. When mining near the loose layers, there is a risk of water and sand inrush, which threatens the personal safety of miners. The prediction of sudden water and sand inrush is difficult due to the comprehensive influence of many factors, and the influencing factors are fuzzy and random. To solve this problem, in this paper, a new risk assessment method of water and sand inrush based on comprehensive weight and cloud model was proposed. Seven factors are selected as indexes: the aquifer thickness, the thickness ratio of sand layer to clay layer, the thickness of bottom clay layer, the coal seam thickness, the percentage of core recovery, the geological structure, and the bedrock thickness. The assessment index system is established, and the index is divided into three grades. A comprehensive weighting method, which combines analytic hierarchy process (AHP), entropy weight method (EWM), and minimum entropy principle, is used to reasonably assign the weight of index. Based on the cloud generator equation, the membership function is obtained. The assessment result of the assessment object is obtained by combining the membership degree and the weight of index. The comprehensive weight-cloud model assessment method is applied to the risk assessment of water and sand inrush in the 6311-2 working face in the sixth mining area of Baodian Coal Mine. According to the assessment results, the following conclusions can be drawn: (1) the bedrock thickness and the coal seam thickness are the main factors of water and sand inrush under loose layer mining; (2) the assessment results obtained by the comprehensive weight-cloud model method are consistent with the actual situation. The assessment method can provide scientific reference for the safe mining under the thick loose layer in the deep mines of southwest Shandong.

Shock-induced plastic deformation and spall damage in the single-crystalline FCC Co25Ni25Fe25Al7.5Cu17.5 high-entropy alloy (HEA) under varying shock intensities were systematically investigated using large-scale molecular dynamics simulations. The study reveals the significant influence of crystalline orientation on the deformation mechanism and spall damage. Specifically, the shock wave velocities in the [110] and [111] directions are significantly higher than that in the [001] direction, resulting in a two-zone elastic-plastic shock wave structure observed in the [110] and [111] samples, while only a single-wave structure is found in the [001] sample. The plastic deformation is dominated by the FCC to BCC transformation following the Bain path and a small amount of stacking faults during the compression stage in the [001] sample, whereas it depends on the stacking faults induced by Shockley dislocation motion in the [110] and [111] samples. The stacking faults and phase transformation in the [001] sample exhibit high reversibility under release effects, while extensive dislocations are present in the [110] and [111] samples after release. Interestingly, tension-strain-induced FCC to BCC phase transformation is observed in the [001] sample during the release stage, resulting in increased spall strength compared to the [110] and [111] samples. The spall strength estimated from both bulk and free surface velocity history shows reasonable consistency. Additionally, the spall strength remains stable with increasing shock intensities. The study discusses in detail the shock wave propagation, microstructure change, and spall damage evolution. Overall, our comprehensive studies provide deep insights into the deformation and fracture mechanisms of Co25Ni25Fe25Al7.5Cu17.5 HEA under shock loading, contributing to a better understanding of dynamic deformation under extreme environments.

Both the nanoscale helium (He) bubble and grain boundaries (GBs) play important roles in the dynamic mechanical behavior of irradiated nanocrystalline materials. Using molecular dynamics simulations, we study the shock-induced deformation and spallation failure of bicrystal copper with a nanoscale He bubble. Two extreme loading directions (perpendicular or parallel to the GB plane) and various impact velocities (0.5–2.5 km/s) are considered. Our results reveal that the He bubble shows hindrance to the propagation of shock waves at lower impact velocities but will accelerate shock wave propagation at higher impact velocities due to the local compression wave generated by the collapse of the He bubble. The parallel loading direction is found to have a greater effect on He bubble deformation during shock compression. The He bubble will slightly reduce the spall strength of the material at lower impact velocities but has a limited effect on the spallation process, which is dominated by the evolution of the GB. At lower impact velocities, the mechanism of spall damage is dominated by the cleavage fracture along the GB plane for the perpendicular loading condition but dominated by the He bubble expansion and void growth for the parallel loading condition. At higher impact velocities, micro-spallation occurs for both loading conditions, and the effects of GBs and He bubbles can be ignored.

The penetration of projectiles into targets has a broad background in engineering. In this work, numerical simulations of the projectile-target penetration problem are conducted using the Finite Element Method (FEM), the Smoothed Particle Hydrodynamics (SPH) and the Finite Element–Smoothed Particle Hydrodynamics Adaptive Coupling Method (FE-SPH ACM) based on the LS-DYNA software package. First, the penetration experiments using aluminum targets and ceramic targets are simulated. The experimental and simulation results show that the FE-SPH ACM has the better accuracy in calculating the debris cloud head velocity and interface velocity, with an error of no more than 4%. Furthermore, we use the FE-SPH ACM to investigate the anti-penetration performance of aluminum/ceramic composite targets in different combinations. We find that the reasonable layout can improve the protective performance of multi-layered target, especially composite target plates with ceramic as the front layer. In addition, the ballistic limit velocities for ceramic-aluminum ratios of 3/7, 5/5 and 7/3 are approximately 1300 m/s, 1400 m/s and 1500 m/s, respectively. Obviously, increasing the proportion of ceramic materials can enhance the anti-penetration performance.

Necking in different positions of a pile body has a significant influence on the bearing performance of a single pile. The transparent soil model experiment was adopted to investigate the impacts of necking positions on the vertical bearing capacity of a single pile and the soil deformation around the pile, and subsequently, the causes of the variation of bearing capacity were analyzed. The results show that the existence of necking does not change the bearing behavior of single pile as pile friction resistance. The bearing capacity of piles decreased by 8.21% and increased by 7.30% when the necking is located in the shallow and middle pile body, respectively, while it did not change significantly when the defects were in the deep part of the pile. The necking position has a significant effect on the deformation and deformation range of soil. The soils at the top, side, and near the end of a pile were mainly influenced by shallow necking, middle necking, and deep necking, respectively. The soils at the location of necking had apparent settlement phenomenon as the piles subsided. The soils at shallow necking’s locations were relatively loose, which reduced the side friction force of piles and finally resulted in a reduction of bearing capacity of a single pile. With the increase of the load, the soil around piles gradually developed penetration phenomenon (large deformation), while the soil at the end of pile moved up against pile settlement, which made the necking soil in the middle denser. The loss of side friction resistance was less than the resistance from the necking, which gave birth to an increase in the bearing capacity of a single pile. It was finally found that the side friction resistance and the resistance caused by necking were the main factors governing the bearing capacity of necking single piles.

With the continuous improvement of machine learning methods, building the interatomic machine learning potential (MLP) based on the datasets from quantum mechanics calculations has become an effective technical approach to improving the accuracy of classical molecular dynamics simulation. The Spectral Neighbor Analysis Potential (SNAP) is one of the most commonly used machine learning potentials. It uses the bispectrum to encode the local environment of each atom in the lattice. The hyperparameter jmax controls the mapping complexity and precision between the local environment and the bispectrum descriptor. As the hyperparameter jmax increases, the description will become more accurate, but the number of parameters in the bispectrum descriptor will increase dramatically, increasing the computational complexity. In order to reduce the computational complexity without losing the computational accuracy, this paper proposes a two-level ensemble feature selection method (EFS) for a bispectrum descriptor, combining the perturbation method and the feature selector ensemble strategy. Based on the proposed method, the feature subset is selected from the original dataset of the bispectrum descriptor for building the dimension-reduced MLP. As a method application and validation, the data of Fe, Ni, Cu, Li, Mo, Si, and Ge metal elements are used to train the linear regression model based on SNAP for predicting these metals’ atomic energies and forces them to evaluate the performance of the feature subsets. The experimental results show that, compared to the features of SNAP and qSNAP, the training complexity improvement of our EFS method on the qSNAP feature is more effective than SNAP. Compared with the existing methods, when the feature subset size is 0.7 times that of the original features, the proposed EFS method based on the SSWRP ensemble strategy can achieve the best performance in terms of stability, achieving an average stability of 0.94 across all datasets. The training complexity of the linear regression model is reduced by about half, and the prediction complexity is reduced by about 30%.

We report a rare case of a 77-year-old man with a known left inguinal hernia presenting with groin pain following a blunt trauma of the left leg. Diagnosis of small bowel perforation away from the hernia was obtained only in surgery. Difficulty in preoperative diagnosis, rarity of histologic pattern, and surgical challenges make this case very interesting for surgeons and radiologists. Our conclusion upon dealing with the situation is that the diagnosis of small bowel perforation following blunt injury to a non-abdominal trauma is rare and difficult. The association between inguinal hernia and non-abdominal trauma may result in small bowel injuries that normally do not appear. Therefore, clinicians should be cautious in treating non-abdominal trauma patients with inguinal hernias.

This work reveals the interaction mechanism between He bubble and grain boundary (GB) in bicrystal copper under shear loading via molecular dynamics simulations. The influences of He/vacancy ratio RHe/V, temperature T0, and bubble diameter D0 on the interaction mechanism are clarified. Specifically, two interaction modes, i.e., the GB traverses or is pinned on He bubble, are observed by changing the initial RHe/V, T0, and D0. As RHe/V increases, the influence of He bubble on GB migration shows a decrease–increase trend. Different He bubble evolutions are demonstrated by comparing their shapes, pressure, and volume. In the cases of low RHe/V, the medium temperatures (10–300 K) are found to accelerate the GB migration, but higher temperatures (600–900 K) will lead to the change in interaction mode and deteriorate the interaction process. Furthermore, a more noticeable bubble-drag effect on GB migration is observed in the samples with larger He bubble.

Titanium carbides attract attention from both academic and industry fields because of their intriguing mechanical properties and proven potential as appealing candidates in the variety of fields such as nanomechanics, nanoelectronics, energy storage and oil/water separation devices. A recent study revealed that the presence of Ti8C5 not only improves the impact strength of composites as coatings, but also possesses significant strengthening performance as an interlayer material in composites by forming strong bonding between different matrices, which sheds light on the design of impact protection composite materials. To further investigate the impact resistance and strengthening mechanism of Ti8C5, a pilot Molecular Dynamics (MD) study utilizing comb3 potential is carried out on a Ti8C5 nanosheet by subjecting it to hypervelocity impacts. The deformation behaviour of Ti8C5 and the related impact resist mechanisms are assessed in this research. At a low impact velocity ~0.5 km/s, the main resonance frequency of Ti8C5 is 11.9 GHz and its low Q factor (111.9) indicates a decent energy damping capability, which would eliminate the received energy in an interfacial reflection process and weaken the shock waves for Ti8C5 strengthened composites. As the impact velocity increases above the threshold of 1.8 km/s, Ti8C5 demonstrates brittle behaviour, which is signified by its insignificant out-of-plane deformation prior to crack initiation. When tracking atomic Von Mises stress distribution, the elastic wave propagation velocity of Ti8C5 is calculated to be 5.34 and 5.90 km/s for X and Y directions, respectively. These figures are inferior compared with graphene and copper, which indicate slower energy delocalization rates and thus less energy dissipation via deformation is expected prior to bond break. However, because of its relatively small mass density comparing with copper, Ti8C5 presents superior specific penetration. This study provides a fundamental understanding of the deformation and penetration mechanisms of titanium carbide nanosheets under impact, which is crucial in order to facilitate emerging impact protection applications for titanium carbide-related composites.

Previous researches have presented some knowledge about the shock loading and unloading of the nanoscale He bubble. However, the He bubble will undergo a long high-pressure adiabatic relaxation process after being shocked. This work focuses on the release path of the nanoscale He bubble in single crystal Al from a stable high-pressure state by molecular dynamics method. Firstly, we consider the case that two opposite release waves meet at the center of the He bubble. Combined with the analysis of deformation mechanism and stress waveform around the He bubble, the difference of evolution law of He bubble under different pressure conditions is revealed. The evolution of the number and distribution of voids with or without the He bubble is compared. And the nucleation region gradually extends to both sides of the He bubble with the increase of initial Hugoniot pressure. Moreover, when a single unloading wave is considered to sweep through the He bubble, the microjet formation in the He bubble is found due to the pressure gradient effect. The shape and velocity distribution of internal and surface jets are discussed.

The excellent mechanical properties of graphyne (GY) have made it an appealing candidate in the field of impact protection. We assessed the deformation mechanisms of monolayer GY nanosheets of different morphologies, including α-GY, β-GY, γ-GY and 6612-GY, under supersonic-velocity impacts (from 1 to 6 km/s) based on in silico studies. Generally, cracks initiate at the geometry center and the nanosheet experiences significant out-of-plane deformation before the propagation of cracks. Tracking the atomic von Mises stress distribution, it is found that its cumulative density function has a strong correlation with the magnitude of the Young’s modulus of the GYs. For nanosheets with a higher Young’s modulus, it tends to transfer momentum at a faster rate. Thus, a better energy dissipation or delocalization is expected during impact. This study provides a fundamental understanding of the deformation and penetration mechanisms of monolayer GY nanosheets under impact, which is crucial in order to facilitate their emerging applications for impact protection.

The influence of rock seepage must be considered in geotechnical engineering, and understanding the fluid flow in rock fractures is of great concern in the seepage effect investigation. This study is aimed at developing a model for inversion of rock fracture domains based on digital images and further study of non-Darcy flow. The visualization model of single rock fracture domain is realized by digital images, which is further used in flow numerical simulation. We further discuss the influence of fracture domain geometry on non-Darcy flow. The results show that it is feasible to study non-Darcy flow in rock fracture domains by inversion based on digital images. In addition, as the joint roughness coefficient (JRC) increases or the fracture aperture decreases, distortion of the fluid flow path increases, and the pressure gradient loss caused by the inertial force increases. Both coefficients of the Forchheimer equation decrease with increasing fracture aperture and increase with increasing JRC. Meanwhile, the critical Reynolds number tends to decrease when JRC increases or the fracture aperture decreases, indicating that the fluid tends to non-Darcy flow. This work provides a reference for the study of non-Darcy flow through rock fractures.

Abstract Background Osteosarcoma (OS) is a rare, malignant bone tumor that primarily affects adolescents and has a high degree of malignancy and high incidence of recurrence and metastasis. Our study aimed to explore the role of miR-338-3p in OS cells. Methods qRT-qPCR was performed to quantify miR-338-3p expression levels in OS tissue samples and in three common OS cell lines. MG-63 and Saos2 cells were separately transfected with miR-338-3p or NC mimics and miR-338-3p expression levels was determined by qRT-PCR. Cell proliferation was monitored using the Cell Counting Kit-8. Flow cytometer analysis was carried out to determine the distribution of cell cycle stages and apoptosis. Transwell assay was performed to measure the migratory and invasive capacities of MG-63 and Saos2 cells. The expression of Vimentin and E-cadherin was detected by western blot. Luciferase reporter assay, qRT-PCR and western blotting were performed to confirm the target of miR-338-3p. Results Analysis by qRT-PCR revealed that miR-338-3p was downregulated in the tissue samples of 20 OS patients when compared with that in their matched adjacent non-tumor tissues. Furthermore, miR-338-3p was significantly downregulated in three common OS cell lines, namely, MG-63, Saos2, and HOS, when compared with that in the human osteoblast cell line hFOB1.19. Analysis by luciferase reporter assay, qRT-PCR, and western blotting revealed that activator of 90 kDa heat shock protein ATPase homolog 1 (AHSA1) is a direct target of miR-338-3p. miR-338-3p overexpression led to significant reduction in AHSA1 protein levels in MG63 and Saos2 cells. miR-338-3p overexpression reduced cell viability and migration and invasion behavior of MG63 and Saos2 cells. In addition, miR-338-3p overexpression suppressed epithelial–mesenchymal transition (EMT), induced a significant G1-phase arrest and did not affect the apoptosis in both MG-63 and Saos2 cells. Moreover, overexpression of AHSA1 reversed the inhibitory effect of miR-338-3p overexpression on proliferation, cell cycle, apoptosis, EMT, migration, and invasion of MG63 and Saos2 cells, thereby suggesting that miR-338-3p acts as a tumor suppressor in OS cells by targeting AHSA1. Conclusions miR-338-3p/AHSA1 can serve as a potential therapeutic target for OS therapy.

Molecular dynamic (MD) simulations were used to study threshold displacement energy (TDE) surface and Si displacement cascades in 4H-SiC system. To figure out the role of different Wyckoff sites in determining the TDE values, both Si and C atoms in 2a and 2b Wyckoff sites were separately considered as the primary knocked atoms (PKA). The initial kinetic energy was then distributed along 146 different crystallographic directions at 10 K. TDE surface appeared highly anisotropic for Si and C displacements along different crystallographic directions. The TDE was determined as 41 eV for Si and 16 eV for C. The average values of TDE over two Wyckoff sites were estimated to 66 eV for Si PKA and 24 eV for C PKA. The displacement cascades produced by Si recoils of energies spanning varied from 5 keV to 50 keV at 300 K. To count the number of point defects using Voronoi cell analysis method, the crystal structure of 4H–SiC was transformed from hexagonal to orthorhombic. It was found that the surviving defects at the end of cascades were dominated by C vacancies and interstitials due to low displacement energies of C atoms and greater number of C interstitials when compared to C vacancies.

Although the mechanism and influence of fault water inrush have been widely studied, there are still few studies on the migration of filling particles and the evolution process of seepage characteristics within faults. In this work, the coupling effects of water flow, particle migration, and permeability evolution are considered synthetically, and the evolution model of seepage characteristics with multifield coupling is established. This model was used to investigate the evolution process of water inrush within faults and the effects of water pressure, initial effective porosity, and initial permeability on water flow rate. The results show that the evolution of seepage characteristics can be divided into three phases: (i) low velocity seepage, (ii) drastic changes with substantial particle migration, and (iii) steady-state water flow. The multifield coupling causes the effective porosity, permeability, flow velocity, and particle concentration to accelerate each other during the dramatic phase. Moreover, the increases in initial water pressure, initial porosity, and initial permeability have different degrees of promotion on the water flow rate. Finally, the simulation results are approximately the same as the data of water inrush in the mining area, which verifies the correctness of the evolution model established in this work. This work provides new approaches to the evolution process and prevention of water inrush in faults.

Damage to mine shafts located below thick loose aquifers, caused by coal seam mining, can seriously affect the operational safety of mines. In view of such a problem, the comprehensive weight method and the fuzzy matter-element analysis method were used to analyze and evaluate the stability of mine shafts below thick, loose aquifers. Based on data relating to shaft structure parameters collected from fifteen damaged shafts in coal mines in East China, a comprehensive analysis was performed. The results showed that (1) the two indexes of surface subsidence velocity and water level drop value at the bottom of a loose aquifer have a large influence on the stability of shafts below thick loose aquifers, and (2) the predicted results of thirteen sets of samples were the same as the results measured during actual production; the accuracy of the model was 86.67%. The comprehensive weight method and the fuzzy matter-element analysis model both show good reliability for evaluating the stability of mine shafts below thick loose aquifers and can provide a scientific reference for the analysis of the stability of mine shafts in areas with thick loose strata, as well as for the design of comprehensive control plans that could subsequently be implemented.

Imbalanced data exist in many classification problems. The classification of imbalanced data has remarkable challenges in machine learning. The support vector machine (SVM) and its variants are popularly used in machine learning among different classifiers thanks to their flexibility and interpretability. However, the performance of SVMs is impacted when the data are imbalanced, which is a typical data structure in the multi-category classification problem. In this paper, we employ the data-adaptive SVM with scaled kernel functions to classify instances for a multi-class population. We propose a multi-class data-dependent kernel function for the SVM by considering class imbalance and the spatial association among instances so that the classification accuracy is enhanced. Simulation studies demonstrate the superb performance of the proposed method, and a real multi-class prostate cancer image dataset is employed as an illustration. Not only does the proposed method outperform the competitor methods in terms of the commonly used accuracy measures such as the F-score and G-means, but also successfully detects more than 60% of instances from the rare class in the real data, while the competitors can only detect less than 20% of the rare class instances. The proposed method will benefit other scientific research fields, such as multiple region boundary detection.

Water inrush is a major geological hazard for safe mining and tunnel construction. For the water inrush channel containing mud, sand, and other sediments, it is difficult to predict the change of permeability and water surge, which makes disaster prevention difficult. As a typical water inrush channel, a filled fracture under seepage-induced erosion needs to be focused. In this work, a numerical model for the evolution of flow in a filled fracture under seepage-induced erosion was established, which included the seepage velocity, hydraulic erosion, and permeability of the filling medium. The effects of joint roughness coefficient (JRC) and homogeneity of the filling medium on the seepage evolution are discussed. The results showed that the fracture seepage properties experienced a non-linear change process, and the evolution can be divided into three phases: the slowly varying phase, the rapidly varying phase, and the stable phase. The increase of the JRC hindered the development in flow velocity and erosion. Compared with low homogeneous filling medium, pores in the high homogeneous filling medium were easier to expand and connect, and the seepage characteristics evolved faster. The model established in this study will help to understand the seepage evolution of filled fractures, and can be used to predict the permeability of filled fractures in engineering geology.