Bolts support have become a major active support method in coal mine roadways to control roadway roof failure and improve surrounding rock structure stability. The traditional bolt force status monitoring (BFSM) method has poor anti-interference performance, is easily affected by harsh downhole environments, and cannot support remote real-time monitoring. This paper presents a fiber Bragg grating (FBG) bolt force sensor that monitors the force of roadway bolts. This sensor uses a cantilever and a diaphragm as elastic elements and two FBGs bonded on the top and bottom surfaces of the cantilever as sensing elements. The experimental results indicate that the measuring sensitivity is improved by using the center wavelength difference between the two FBGs. The sensitivity is 38.79 pm/kN within the range from 0 to 150 kN, and the correlation coefficient reaches 99.98%. The engineering applications show that the FBG sensing technology can automatically acquire, and monitoring results are of great significance in roadway anchorage engineering safety and bolt support quality evaluation. Furthermore, such a sensor is also widely used in quasi-distributed measurement and long-term online monitoring of bolt force status in such fields as geotechnical engineering, tunnel engineering, and slope engineering.

The current single gas prediction model is not sufficient for identifying and processing all the characteristics of mine gas concentration time series data. This paper proposes an ARIMA-LSTM combined forecasting model based on the autoregressive integrated moving average (ARIMA) model and the long short-term memory (LSTM) recurrent neural network. In the ARIMA-LSTM model, the ARIMA model is used to process the historical data of gas time series and obtain the corresponding linear prediction results and residual series. The LSTM model is used in further analysis of the residual series, predicting the nonlinear factors in the residual series. The prediction results of the combined model are compared separately with those of the two single models. Finally, RMSE, MAPE and R² are used to evaluate the prediction accuracy of the three models. The results of the study show that the metrics of the combined ARIMA-LSTM model are R² = 0.9825, MAPE = 0.0124 and RMSE = 0.083. The combined model has the highest prediction accuracy and the lowest error and is more suitable for the predictive analysis of gas data. By comparing the prediction results of a single model and the combined model on gas time series data, the applicability, validity and scientificity of the combined model proposed in this paper are verified, which is of great importance to accurate prediction and early warning of underground gas danger in coal mines.

The method of fully mechanized top-coal caving mining has become the main method of mining thick-seam coal. The process parameters of fully mechanized caving will affect the recovery rate and gangue content of top coal. Through numerical simulation software, the top-coal recovery rate and gangue content, under different fully mechanized caving process parameters, were simulated, and the influence law of different fully mechanized caving process parameters on top-coal recovery rate and gangue content was obtained. A decision model for top-coal caving process parameters was established with a BP neural network, and the optimal top-coal caving parameters were obtained for the actual situation of a working face. On this basis, a in-lab similarity simulation test of the particle material was carried out. The results show that the top-coal recovery rate and gangue content were 86.56% and 3.45%, respectively, and the coal caving effect was good. A BP neural network was used to study the decisions optimizing fully mechanized caving process parameters, which effectively improved the decision-making efficiency thereabout and provided a basis for realizing intelligent, fully mechanized caving mining.

Research on the stress state of rock mass is essential for revealing the distribution characteristics and evolution law of the surrounding rock stress field in the roadway, studying the coal–rock dynamic disaster and the design of roadway support. This thesis proposes a three-dimensional stress monitoring method for surrounding rocks based on fiber Bragg grating (FBG) sensing technology and a cube-shaped three-dimensional stress fiber grating sensor is developed based on the principle of this monitoring method. According to the fiber grating strain obtained by numerical simulation, the calculated three-dimensional stress value is basically consistent with the theoretical value. The margin of error was plus or minus one percentage point. The sensing performance of the sensor was tested using a uniaxial compression experiment instead of a triaxial compression experiment. The experimental results show that in the range of 0~50 Mpa, the sensor’s sensitivity to X, Y and Z axis stress is 25.51, 25.97 and 24.86 pm/Mpa, respectively. The relative error of measured stress is less than 4%. Meanwhile, the sensor has good linearity and repeatability, and has broad application prospects in the field of underground engineering safety monitoring such as in coal mines and tunnels.

Support technology faces challenges in view of the large deformation of surrounding rock in three-soft coal roadways under high horizontal stress in Zijin Coal Mine, China. Geostress near the tested working face of the mine was measured and its distribution law was analyzed through theoretical analysis, numerical simulation analysis, and field measurement. The original supporting scheme of the three-soft coal roadway on the tested working face was analyzed to discover the deformation and failure mechanism of the surrounding rock of the original supporting roadway and the control measures. An optimized support scheme of H-G (hollow grouting) anchor cables, high strength bolts, W-shaped steel belts, metal meshes, and sprayed concretes was proposed for field applications. Based on the roadway in the tested 3201 working face at Zijin Coal Mine, the technical parameters for optimizing the combined support of the roadway were determined. The following results were be obtained through field measurement. The roadway was kept intact after excavation and the optimized support scheme was adopted in the three-soft coal roadway. No obvious deformation in appearance existed in the roof, floor, and roadway coal sides. Compared with the original support scheme, the stability of the roadways was improved visibly. The displacement of the roadway roof decreased from 100 to 30 mm, and that of the roadway coal walls decreased from more than 100 mm to less than 50 mm. This work verifies the effectiveness of a combined support scheme of H-G anchor cables, high strength bolts, W-shaped steel belts, metal meshes, and sprayed concretes to control deformations of surrounding rock in three-soft coal roadways. The new support scheme has good social and economic benefits and can be used as a reference for other roadway supports under similar conditions.

In order to solve the key technical problem of accurate perception of the straightness of a scraper conveyor in intelligent mining of coal mines, a three-dimensional (3D) curvature sensor of fiber Bragg grating (FBG) is designed and developed based on the curvature sensing principle of FBG. The grating string embedded in the FBG curvature sensor is simulated under the radius of curvature of 100 m, 200 m, 300 m, 400 m and 500 m. The results show that the strain of the built-in FBG string is linearly related to the curvature at the bend of the FBG string. The simulation results are consistent with the theoretical results, which verifies the rationality of the sensor design. Based on the discrete curvature information obtained by the 3D curvature sensor of FBG, the curve reconstruction algorithm is derived by using the method of fitting recursion, and the bending shape fitting perception and curve reconstruction of the scraper conveyor is realized. A test platform for the three-dimensional bending of the scraper conveyor is set up. The measuring effect of perceived and actual shape at the straight and bending states of the scraper conveyor is verified. The test results show that the measured curve is basically the same as the perceptual curve in the three-dimensional bending configuration of the scraper conveyor, and the error of the direction of each coordinate axis is not more than ±15 mm, which can meet the requirements of the accuracy of detecting the straightness of the intelligent work.