The effective prediction of gas concentration and the reasonable formulation of corresponding safety measures have important significance for improving the level of coal mine safety. To improve the accuracy of gas concentration prediction and enhance the applicability of the models, this paper starts with actual coal mine production monitoring data, improves the accuracy of gas concentration prediction through multi-parameter fusion prediction, and constructs a recurrent neural network (RNN)-based multi-parameter fusion prediction of coal face gas concentration. We determined the performance evaluation index of the model’s prediction method; used the grid search method to optimize the hyperparameters of the batch size; and used the number of neurons, the learning rate, the discard ratio, the network depth, and the early stopping method to prevent overfitting. The gas concentration prediction models—based on RNN and PSO-SVR and PSO-Adam-BP neural networks—were compared and analyzed experimentally with the mean absolute percentage error (MAPE) as the performance evaluation index. The result show that using the grid search method to adjust the batch size, the number of neurons, the learning rate, the discard ratio, and the network depth can effectively find the optimal hyperparameter combination. The training error can be reduced to 0.0195. Therefore, Adam’s optimized RNN gas concentration prediction model had higher accuracy and stability than the BP neural network and SVR. During training, the mean absolute error (MAE) could be reduced to 0.0573, and the root mean squared error (RMSE) could be reduced to 0.0167; however, the MAPE could be reduced to 0.3384% during prediction. The RNN gas concentration prediction model and parameter optimization method based on Adam optimization can effectively predict gas concentration. This method shows high accuracy in the prediction of gas concentration time series and can be used as a reference model for predicting mine gas concentration.

The evaluation of the coal mine gas drainage effect is affected by many factors, such as flow rate, wind speed, drainage negative pressure, concentration, and temperature. This paper starts from actual coal mine production monitoring data and based on the lasso regression algorithm, features selection of multiple parameters of the preprocessed gas concentration time series to construct gas concentration feature selection based on the algorithm. The three-time smoothing index method is used to fill in the missing values. Aiming at the problem of different dimensions in the gas concentration time series, the MinMaxScaler method is used to normalize the data. The lasso regression algorithm is used to perform feature selection on the multivariable gas concentration time series, and the gas concentration time series selected by the lasso feature and the gas concentration time series without feature selection are input. The performance of the ANN algorithm for gas concentration prediction is compared and analyzed. The optimal α value and L1 norm are selected based on the grid search method to determine the strong explanatory gas concentration time series feature set of the working face, and an experimental comparison of the gas concentration prediction results before and after the lasso feature selection is performed. We verify the effectiveness of the algorithm.

The construction of gas extraction boreholes in soft coal seams is prone to collapse and deformation, and grouting reinforcement is one of the main methods to solve the problem of efficient sealing. However, the reasonable selection of key grouting parameters still needs further research. In response to the problem of selecting grouting sealing parameters for gas drainage drilling in soft coal seams, based on the “concentric ring” reinforcement sealing technology obtained in previous research, the key parameters and sealing technology of the “concentric ring” reinforcement were studied through theoretical calculation and numerical simulation experiments. The slurry diffusion morphology and range under different grouting pressures and grouting time slurry viscosity were obtained. Finally, in order to explore the application effect of key grouting parameters, on-site industrial tests were conducted in a soft and high gas coal seam. The research results indicate that the optimal grouting pressure for the “retaining wall rock hole ring” should not be less than 3 MPa, the reasonable grouting time should be 10–15 min, and the water material ratio of the grouting material should be greater than 1:1; The use of new reinforcement sealing and grouting technology can ensure long-term good extraction effect. Compared to the testing of drilling sealing effect using ordinary cloth bags with two plugs and one injection, The adoption of new reinforcement sealing technology can effectively prevent the deformation and collapse of the borehole before sealing, and due to two rounds of grouting and pre reinforcement of fractured coal, the sealing effect of the borehole is also relatively good. The research results have important theoretical value for guiding the drilling and sealing grouting engineering of gas extraction in soft coal seams.

The permeability of crushed coal bodies plays a bottom neck role in seepage processes, which significantly limits the coal resource utilisation. To study the permeability of crushed coal bodies under pressure, the particle size distribution of crushed coal body grains is quantitatively considered by fractal theory. In addition, the parameters of the percolation characteristics of crushed coal body grains are calculated. Moreover, the permeability of the crushed coal body during recrushing is determined by the fractal dimension and porosity. A lateral limit compression test with the crushed coal bodies was carried out to illustrate the effect of the porosity on the permeability, In addition, a compressive crushed coal body size fractal–permeability model was proposed by combination of the fractal dimension and the non-Darcy equivalent permeability. The results show (1) the migration and loss of fine particles lead to a rapid increase in the porosity of the crushed coal body. (2) Increases in the effective stress cause the porosity and permeability to decrease. When the porosity decreases to approximately 0.375, its effect is undermined. (3) The migration and loss of fine particles change the pore structure and enhance the permeability properties of the skeleton, causing sudden seepage changes. (4) At low porosity, the permeability k is slightly larger than the non-Darcy equivalent permeability ke. Thus, the experimental data show an acceptable agreement with the present model. A particle size fractal–percolation model for crushed coal bodies under pressure provides a solution for effectively determining the grain permeability of the crushed coal bodies. The research results can contribute to the formation of more fractal-seepage theoretical models in fractured lithosphere, karst column pillars and coal goaf, and provide theoretical guidance for mine water disaster prevention.

In order to investigate the compaction and re-crushing characteristics of crushed coal gangue with different gradation, the compaction fractal test was carried out for crushed coal gangue with different Talbol power index n. The compaction deformation parameters such as displacement, stress-strain were analyzed according to the test results. And according to the test results, the compaction deformation relationship of the lateral limit crushed coal rock body considering the gradation is obtained. The test results show that: the crushing of coal rock particles exists in almost the whole compaction process of the crushed coal rock body, and the crushing of coal rock particles has a non-negligible influence on the deformation of the whole crushed rock body; the structural stability of large-size coal rock particles is not as good as that of small-size coal rock particles, and the large-size coal rock is more likely to be crushed under the same stress conditions; the distribution coefficient r of the crushed coal rock body decreases with the increase of axial stress; and before the Before the axial stress reaches 8 MPa, the distribution coefficient r of crushed rock samples tends to increase with n in general, and after reaching 8 MPa and later, the distribution coefficient r of crushed rock samples tends to decrease with n in general; the difference value of particle crushing increases with the increase of axial stress, and the weight value of particle crushing decreases with the increase of axial stress, and the changes of both are non-linear; according to the stress recovery in the compaction process of crushed coal rock body The compaction deformation model of the crushed coal rock body is constructed according to the crushing characteristics of coal rock particles in the stage of compaction, which effectively combines its fine action mechanism with macroscopic physical phenomena in a simple form and has certain practical engineering significance.

Particle loss is an important cause of water inrush catastrophes in collapsed columns. In order to study the relationship between the lost particles of different graded rock samples and the pore structure of the subsidence column filling, experiments were designed and the changes of the seepage parameters of graded rock samples during the particle migration process under different permeable water pressures P and axial loads F were determined. The results show that: (1) There will be obvious collapse, silting and particle loss behaviors in the sample during different loading processes, and the rock samples with gradation values of n = 0.3 and n = 0.5 are dominant; (2) The relationship between porosity φ and bearing pressure The exponential function can be used to fit the loads F well, and the porosity decreases with the increase of the bearing load. The water surging characteristics before and after 1.2 MPa are mainly in the turbulent water gushing stage, accompanied by instantaneous slurry. Possibility of splashing and indenter sliding; (3) After infiltration, the condition of the remaining skeleton rock samples in the cylinder generally shows a trend of first decreasing rapidly, then increasing slowly, and then decreasing; (4) The gradation value n of the sample and the bottom There is a good correlation between the damaged area and the mean value S of the maximum area of the top water inrush channel. The maximum area increase of the damaged area and the maximum area increase of the water inrush channel show an opposite trend. The permeable pores of the graded samples can be divided into There are three situations of digging and collapse, water inrush gap and scouring hole, and the pore seepage process can be divided into 4 stages of inoculation of water seepage, rapid adjustment, rapid scour and steady flow.