Current artificial intelligence models for evaluating goaf stability in underground metal mines need a large amount of sample data for training, and their accuracy declines with small sample size. With the aim of solving this problem, this paper proposes an improved TrAdaBoost algorithm based on transfer learning theory. The scope of the TrAdaBoost algorithm is extended from the two-level classification to multi-level classification problems, which makes it suitable for evaluating goaf stability. The isolated forest method is used to filter the bad points of the auxiliary training set, thereby eliminating the interference of abnormal data. The dynamic factor concept is introduced to solve the problem that the weight of source domain data decreases too quickly and irreversibly, and this enhances the generalization performance of the algorithm for different goaf samples. To test the accuracy of the proposed model in predicting goaf stability, an evaluation model is constructed and the performance compared with other algorithms in current use. The prediction accuracy and generalization ability of the model are evaluated by mean square error and F1 measurements, which prove that the performance of the model is excellent. The most obvious finding to emerge from this paper is that, with suitable improvements, goaf evaluation models can still maintain a high level of accuracy with small sample size.

This study is aiming at the problems of slow collision detection speed of bounding box, low efficiency of triangle intersection detection, and triangulation in regions where complex intersecting lines are segmented that exist in Boolean operation of triangular mesh model in underground space engineering. The orientation calculation method of underground space engineering is put forward, and the precise definition of bounding box is realized. A simple method to realize the fast intersection detection algorithm between triangles is presented, the process of transforming the projection of intersecting lines into scalar and then judging is eliminated, and the efficiency of triangle intersection detection is improved. On the basis of Delaunay triangulation robust algorithm, numbering method of intersection position is improved. Central axis region division method and hierarchical subdivision triangulation method are proposed, which can divide all the regions of mixed complex intersections in one traversal. A subdivision triangulation method by comparing the cosine value of internal tension angle is proposed, and the subdivision triangulation of polygon is optimized. The feasibility of the algorithm is verified by practical engineering application.

To accurately and quantitatively analyze the pollutant gas concentration in tunneling roadways, a prediction model of the pollutant gas concentration was proposed and established. Through downhole gas composition data acquisition and correlation analysis, the prediction variables of downhole gas pollution are obtained with both short-term and long-term memory neural network prediction methods and random forest regression modeling methods, making full use of historical target gas concentration data for the future in a short period of time to evaluate the model performance and prediction results. Compared with the results of the stochastic forest regression prediction and the long- and short-term memory neural network prediction, the stochastic forest regression prediction model has a good prediction effect and better generalization effect and is a reliable method with excellent performance for downhole gas concentration prediction. The analysis of the predicted results shows that the change in CO concentration is strongly correlated with CH₄ and CO₂ and strongly correlated with N₂, making it possible to obtain the potential influencing factors of the target gas. These results provide a scientific basis for the prediction of underground pollution gas concentration and the protection and treatment of the atmospheric environment in mining areas.

In this paper, the mechanical parameters of the grouting consolidation body on the rock block are studied using the Acoustic velocity test case and the revised Hoek-Brown failure criterion in underground metal mines. In light of the hidden resources in the collapsed area of the underground metal mine, the intact rock drilling core cannot be obtained to perform the mechanical test because of extreme damage or because it is completely broken. Therefore, pure cement grouting is used to reinforce the extremely broken rock mass within the collapsed area of the hidden resources. The RQD value and water permeability are tested to indicate that cement pastes can fill the cracks in broken rock blocks and can improve the integrity of the broken rock mass. To detect the grouting consolidation effect of the fractured rock mass and evaluate the mechanical parameters of the consolidation body, a nondestructive test is applied to measure the Acoustic velocity of the consolidation body after grouting. Initially, the measured Acoustic velocity is adopted to calculate the damage index (or integrity index), which is approximately equivalent to the disturbance factor. Then, the modified Hoek-Brown failure criterion is used to evaluate the mechanical parameters of the grouting consolidation body. Finally, to verify this method, the mechanical parameters of the rock mass are also obtained by a laboratory test. The outcomes show that the current research can serve as a practical method when evaluating rock mass quality by the grouting treatment.

At present, many mines adopt the filling method. It is particularly important to solve the problem of the long-distance transportation of slurry during the filling process. Based on the high-density filling material of Sanning mine, the experiments were designed to add sodium abietate (SA), triterpene saponin (SJ) and sodium dodecyl sulfonate (K₁₂) with concentrations of 0.0%, 0.2%, 0.4% and 0.6%, respectively, which were used as air entraining agents (AEA). The filling body with the curing age of 7 and 28 days was prepared for various tests, including nuclear magnetic resonance (NMR), and alternating current (AC) impedance tests. The effects of the air entraining agent and curing time on the physical properties, pore structure and AC impedance properties of the filling were obtained. The results show that: (1) within the frequency range of 10⁻¹–10⁵ Hz, the variation trend of AC impedance of the filling body cannot be changed by adding the air entraining agent, and the filling body with the same ratio had a similar topological structure. (2) The filling body with different AEA and curing times can be represented by the same equivalent circuit model, while the maximum chi-square coefficient was 0.46%. (3) Under the condition of a high frequency of 10⁵ Hz, the porosity and uniaxial compressive strength of the filling body with 7 day curing age were linearly correlated with the AC impedance. However, the porosity and uniaxial compressive strengths of the filling body with 28 days curing time were affected by the type of AEA at a high frequency of 10⁵ Hz.

The risks associated with underground mining at Fankou Lead/Zinc Mine in South China are growing due to the large-scale mining activities there. To recognize mining-induced earthquakes and assess the risk per mining level, a microseismic monitoring system, which is used to record microseismic events, is installed at multiple mining levels in Fankou Lead/Zinc Mine. The purpose of this study is to identify mining-induced earthquakes and to evaluate the risk per mining level by analyzing the spatiotemporal characteristics of microseismic activities in the Fankou Lead/Zinc Mine. In this study, the Gutenberg-Richter (G-R) relationship is applied to compute the b-value, which is used to obtain the maximum magnitude (M (max)) of microseismic event that probably occurs at each mining level. Then, the evaluation of the recurrence period for M (max) and the probability of the microseismic event with the magnitude M (max) is carried out and the M (max) at each mining level is determined based on the recording period of microseismic events. The results show that factors such as the maximum rock vibration velocity, source parameters, displacement, microseismic waveform and energy ratio (ES/EP) can be used to distinguish whether a recorded microseismic event is mining-induced earthquake. Additionally, we propose a method to assess the possibility of mining-induced earthquake at each mining level based on M (max) and predict the recurrence time of microseismic event with the magnitude M (max). The of two years results of microseismic events monitoring demonstrate that the current study is promising for identifying mining-induced earthquakes, assessing the risk of mining-induced earthquakes, predicting the potential maximum microseismic event in a region and estimating its recurrence period and its probability in the Fankou Lead/Zinc Mine.