A new understanding of the mechanism controlling water inflow into mines based on the stratigraphic distribution beneath major aquifers through a case study of the Yijun Formation in the Binchang mining area, Shaanxi Province, China

doi:10.21203/rs.3.rs-4203558/v1

Download PDF

Research Article

A new understanding of the mechanism controlling water inflow into mines based on the stratigraphic distribution beneath major aquifers through a case study of the Yijun Formation in the Binchang mining area, Shaanxi Province, China

https://doi.org/10.21203/rs.3.rs-4203558/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The Binchang mining area is located in the southern Ordos Basin. The water inflow in mines and working faces is considerable due to the Jurassic aquifers and thick Cretaceous aquifer above the coal seams. The water inflow into mines caused by the thick Cretaceous aquifer is considerable, and the mechanism controlling the mine water inflow is still unknown. This case study is based on the mine water inflowand the thickness of the Yijun Formation in the Binchang mining area and reveals a significant correlation between the mine water inflow and the overlying major aquifers. A qualitative analysis was performed to study the relationship between the sedimentary facies and the distribution of mine water inflow in the Tingnan Coal Mine. A random forest (RF) model was applied to identify the key factors affecting the water inflow at working faces. The analysis results showed that in addition to the aquifer characteristics, the water inflow at working faces is negatively correlated with the sedimentary thickness of the Yijun Formation, and the thickness has the greatest influence on the distribution of water inflow. Furthermore, the Yijun Formation is not only an important aquifuge but also determines the bottom morphology of the Luohe Formation, which is an important aquifer. The area with a thin Yijun Formation is also the catchment area at the bottom of the Luohe Formation. This case study first reveals that the underlying strata have an important influence on the sedimentary environment and water abundance of the upper aquifer. The relationships and findings in this case study can provide helpful references for the prevention and control of mine water in similar mining areas.

Mining areas in western China

mine water inflow

sedimentary environment

stratigraphic distribution

random forest

The mine water hazard is one of the five natural disasters that can occur in coal mines. The inflow (outflow) of mines and working faces is the core issue for mine water hazard prevention. Some factors that influence mine water inflow have been comprehensively analyzed (Hu et al.,2005; Wu et al.,2007; Li et al.,2023), showing that mine water inflow is related to the water abundance of the aquifer (or water body) providing the water that seeps into the mine, water channels, mining methods and other factors.

With the depletion of coal resources in eastern China, China's coal production areas have gradually shifted to western China. In the western mining area, the mining of Jurassic coal seams is dominant. The Jurassic Yan'an Formation is a coal-bearing stratum, and the overburdening strata generally contain continental sandstone sedimentary aquifers. Most Jurassic coalfields face the threat of roof water hazards to different degrees, and water inflow occurs in various forms and can vary greatly. The characteristics of water disasters in the process of coal mining in the Ordos Basin were analyzed, and the differences and water inflows of different types of water hazards were summarized (Dong et al.,2020). There is a strong relationship between the groundwater flow field and water inflow in the western mining areas (Huang et al.,2022). Based on a specific analysis of the Huanglong coalfield, the water inflow is considered to be closely related to the lithology of the coal seam roof (Jin et al.,2022).

The factors influencing mine water inflow in the western mining areas have attracted the attention of many scientific researchers. In addition to the abovementioned water channels and mining methods, the water inflow in the western mining area is related to the complex sedimentary environment of the western Jurassic coalfields, which leads to complex rock assemblages, rapid lithofacies changes and various geological structures containing aquifuges. Therefore, mine water inflow has a strong correlation with the sedimentary environment.

At present, there are many studies on the influence of the sedimentary environment on the water abundance of aquifers. Through the subdivision of geomorphology, it was found that the major aquifers consisted of stacked multistoried channel bodies in the Sutlej-Yamuna plain of northwest India (Van Dijk et al.,2016). Horizons with good hydraulic properties are related to this complex interaction of sedimentary facies and diagenetic evolution but are restricted to only very local areas with no relevance for basin-wide fluid flow in the subsurface (Kunkel et al.,2018). According to the stratigraphic architecture and depositional environment of the Paleocene–lower Eocene aquifer system in the Douala onshore basin, SW Cameroon, the best potential aquifers appear to correspond to the fluvial channel-fill sands deposited during the base-level fall in the late Paleocene (early Thanetian) (Koah Na Lebogo et al.,2021).

The sedimentary environments of specific mines have hydrogeological significance, showing that the sedimentary environment of fluvial facies is an important reason for the high water abundance in the Luohe Formation (Mu et al.,2015; Zhao,2015a). The water abundance of a mine field can be reasonably estimated by using sedimentary environment characteristics (Wang et al.,2019). According to the sedimentary environment and sedimentary characteristics, the hydrogeological structure of a typical mine aquifer in the northern Ordos Basin was constructed (Lin,2020). The sediment control method was applied to construct a comprehensive evaluation system and establish a water richness zonation map, which improved the accuracy of the water abundance assessment and provided a reference for the analysis and evaluation of the water abundance of roof aquifers in mining areas without undergoing hydrogeological tests (Zhao,2015b). A water abundance evaluation index system was proposed based on a control pattern of groundwater sedimentation characteristics and a projection tracking model (Wu et al.,2017). The sedimentary environment is considered to be the main factor controlling the spatial distribution of aquifers and aquicludes, and the water-rich characteristics of aquifers and multisource information fusion technology have been applied to select water abundance evaluation indicators and predict the water abundance of coal seam roof sandstone aquifers (Li et al.,2017). Based on the hydrogeological characteristics of the Hongliu Coal Mine, the sedimentary environment of the sandstone aquifer in the first member of the Zhiluo Formation involved a channel bar paleoenvironment characterized by a varying water table, and this unit proved to be the key layer for preventing and controlling water hazards (Wang et al.,2023).

At present, studies on the effect of the sedimentary environment on mine water hazards have mainly discussed the effects of the sedimentary environment on the water abundance of the Yan’an and Zhiluo Formation aquifers and especially the Cretaceous Luohe Formation aquifers, which can affect mine water inflow. However, the influence of the sedimentary environment on hydrogeological conditions and mine water inflow is complex. The mine water inflow in the Binchang mining area is large, especially in the Gaojiapu coal mine in the mining area, which is currently a major resource base in China. Based on an analysis of the mine water inflow data from the Binchang mining area and the water inflow at the working faces of several mines, this study revealed that the distribution of the Yijun Formation, which is a weak water-rich aquifer or even an aquifuge, is closely related to the water inflow. This work shows that the stratigraphic distribution beneath the important aquifer has a controlling effect on the sedimentary environment and water abundance of the upper aquifer. This study has important guiding significance for research on mine water hazard factors and sedimentary environments.

2.1 Overview of the study area

The Binchang mining area is located in the Miaobin sag on the northern margin of the Weibei flexural belt in the southern Ordos Basin and on the northern flank of the Taiyu anticline. The surface is covered by a large area of loess; the strata slope ranges from southeast to northwest at a gentle angle of inclination, generally 1° to 3°. Wide and gentle folds are mainly developed in a northeast-east direction; from north to south lies the Qilipu − Xipo anticline, the Mengcun syncline, the Dongjiazhuang anticline, the Nanyuzi syncline, the Lujia − Xiaolingtai anticline, the Anhua syncline, the Qijia anticline, the Shijiadian syncline, and the Binxian anticline, among other structural features.

According to the borehole logs, the strata in the mine are as follows in terms of geological age: Lower Triassic Hujiacun Formation (T₃h); Lower Jurassic Fuxian Formation (J₁f); Middle Jurassic Yan’an Formation (J₂y), Zhiluo Formation (J₂z), Anding Formation (J₂a); Lower Cretaceous Yijun Formation (K₁y), Luohe Formation (K₁l), Huachi Formation (K₁h); Neogene deposits (N) and Quaternary deposits (Q). The Yan’an Formation is a coal-bearing stratum, and the No. 4 coal seam is the main target of mining. The Jurassic and Cretaceous sediments in this area are terrestrial sediments, mainly lake–river delta facies and alluvial fan facies (Fig. 1).

The aquifers in the Yan’an Formation and the Zhiluo Formation overlying the coal seams are the main aquifers directly providing water to the No. 4 coal seam, but the water inflow is not large. The conglomerate aquifer of the Yijun Formation and the extremely thick sandstone aquifer of the Luohe Formation are aquifers that indirectly provide water to the No. 4 coal seam, with the Luohe Formation being the main aquifer due to its substantial thickness and large water storage capacity.

There are 15 mine shafts distributed in the mining area (Fig. 2). In the Gaojiapu Coal Mine in the Binchang mining area, the water inflow from a single working face reaches more than 3000 m³/h. Other mines in the Binchang mining area also have relatively high rates of water inflow.

2.2 Data analysis of the relationship between mine water inflow and the distribution of the Yijun Formation

2.2.1 Data analysis of the Binchang mining area

The relevant data obtained from the Binchang mining area where the study area is located are listed in Table 1 (Dou and Xiao, 2013; Wu et al., 2018; Wu et al., 2021; Wang, 2021).

Table 1

Statistical table of the water inflow and factors influencing mine working faces in the Binchang mining area
Name of mine	Thickness of Yijun Formation (m)	Working face
Name of mine	Thickness of Yijun Formation (m)	Name	Mining height (m)	Face width (m)	Stable water inflow (m³/h)	Maximum water inflow (m³/h)
Gaojiapu Coal Mine (The No. 2, No. 3 and No. 4 panel areas)	0–20	205	9.5	180	1260	1405
		3401	12.1	146	2000	3224
		3406	10	180	1500	2437
Xiaozhuang Coal Mine	20–30	40201	12.5	174	550
		40202	14.3	174	440
		40203	15.8	200	510
		40204	15.5	195	550
Hujiahe Coal Mine	30–40	401101	12.1	175	520
		401102	13.1	180	470
		402103	15.3	180	420
		401105	15.8	180	550
		401103	15.5	190	210
		402102	15,1	180	300
		401111	15.2	200	220
Wenjiapo Coal Mine	23.3–74.9	4102	3.9	240	270	404

2.2.2 Graphical analysis of the water inflow at working faces and the distribution of the Yijun Formation in the Tingnan Coal Mine

To further analyze the relationship between mine water inflow at working faces and the distribution of the Yijun Formation in the Binchang mining area, this study selected the Tingnan Coal Mine, which has numerous working faces and abundant data.

The Tingnan Coal Mine is located in the central part of the Binchang mining area, and the stratigraphic structure is representative of the entire Binchang mining area. The development height of the water-conducting fracture zone caused by mining the No. 4 coal seam in the Tingnan Coal Mine could extend higher than the Jurassic aquifer and be as high as part of the Luohe Formation aquifer. Therefore, the phenomenon of water inflow occurs during the mining of each working face.

At present, at 25 working faces in 4 panel areas in the Tingnan Coal Mine, mining has been completed, and water inflow has occurred in each working face to various degrees. There are large differences in water inflow between different panel areas in the Tingnan Coal Mine. For example, the No. 3 panel areahas the largest overall water inflow, with a maximum of 874 m³/h, and the overall water inflow in the No. 1panel area is the lowest, with a minimum water inflow of 0 m³/h. In addition, the water inflow at different working faces in the same panel area also has certain differences. To further analyze the relationship between mine water inflow at the working faces and the thickness distribution of the Yijun Formation in the Binchang mining area, based on the actual water inflow data of 25 working faces, a water inflow distribution map was generated (Fig. 3), and based on the drillhole data, a contour map of the thickness distribution of the Yijun Formation was generated (Fig. 4).

2.3 RF model analysis of the water inflow at working faces in the Tingnan Coal Mine

Machine learning has a clear advantage in the processing of complex multidimensional data. The RF model was applied to quantitatively study the relationship between the sedimentary environment and water inflow. To better analyze the relationship between the sedimentary environment and water inflow, thickness data for the sedimentary facies and sedimentary subfacies were collected. Considering that the mining height has a significant impact on the water inflow at working faces, the mining height data were also analyzed as a major influencing factor.

2.3.1 RF model

RF is a decision tree model based on the bagging framework and consists of multiple decision trees. A decision tree is a tree structure in which each nonleaf node represents a test of a feature attribute. Each leaf node stores a category. There is no association between the decision trees. When RF is applied in the classification algorithm, the category with the most votes from decision trees is the final category; when RF is applied in the regression algorithm, the arithmetic mean of the regression results obtained from decision trees is the final model output (Xiao et al.,2017; Jin et al.,2020). RF is a classic ensemble model, and therefore, the model has many hyperparameters. The main hyperparameters are max_depth, max_features, min_samples_leaf, min_samples_split, and n_estimators. For the specific definitions of the hyperparameters, please refer to the Principles of Algorithms (Youssef et al.,2016).

2.3.2 Data selection for the RF model

The factors affecting the water inflow at working faces include the water abundance of the aquifer, the water resistance performance of the aquifuge and the mining methods. Therefore, there are few data from mine pumping tests. The thickness of the Luohe Formation and the sand–stratum ratio of the Luohe Formation were used to characterize the water abundance of the Luohe Formation, the height from the coal seam to the bottom boundary of the Luohe Formation was used to characterize the aquifuge, and the mining height of the working face was used to characterize the mining method. In addition, the thickness of the Yijun Formation was considered to study the relationship between the above factors and water inflow at working faces.

Therefore, 5 factors, i.e., the distance from the coal seam to the Luohe Formation (L₁), the thickness of the Yijun Formation (L₂), the thickness of the Luohe Formation (L₃), the sand-to-ground ratio of the Luohe Formation (B) and the average mining height (C), were used to predict the stable water inflow at working faces. The mathematical model can be expressed as [L ₁, L ₂, L ₃, B, C] → [Q]. The data statistics are shown in Table 2.

Table 2

Data statistics of the mine water inflow, mining height and thickness of each strata sedimentary facies
Number	Q (m³/h)	C (m)	L₁(m)	L₂(m)	L₃(m)	B
1	320	6	149.21	29.35	345.56	97.13%
2	160	6	148.86	31.3	303	100.00%
3	240	5.3	109.55	17.45	308.3	98.28%
4	300	5.9	145.96	24.7	317.95	100.00%
5	260	7.4	146.84	18.63	311.85	100.00%
6	240	5.8	138.4	20.45	313.32	96.29%
7	320	8.9	148.15	27.9	414.6	100.00%
8	20	9.3	141.02	40.7	255.88	100.00%
9	100	5	211.25	38.1	300.25	98.77%
10	140	5	190.34	36.5	362	98.68%
11	240	7.5	176.11	28.2	305.45	98.63%
12	220	7.5	178.8	31.4	308.1	99.45%
13	60	7.8	181.32	40	230	100.00%
14	20	10	190	45	243	100.00%
15	80	9	177.94	41	267.5	100.00%
16	20	7.4	203.57	44.8	263.8	97.74%
17	40	3.1	178.55	32.95	235.7	100.00%
18	20	7.2	164.2	41.5	220	100.00%
19	100	7.8	128.88	35	260.5	100.00%
20	60	4.7	149.49	31.75	236.85	100.00%
21	320	6	137.41	24.45	306.89	100.00%
22	300	6	119.53	20.18	324.76	98.56%
23	350	6	159.9	27.27	352.52	87.08%
24	310	6	153.49	25.9	336.41	100.00%
25	180	6	168.24	24.8	359.7	87.44%

2.3.3 Hyperparameter determination of the RF model

The Bayesian optimization algorithm can be applied to determine the hyperparameters for a dataset (Li and Shami,2020). Based on the Bayesian principle, the Bayesian optimization algorithm looks for a functional relationship between the loss function and the hyperparameters and can find the optimal combination of hyperparameters in the shortest amount of time. The procedure of the Bayesian optimization algorithm is as follows:

1) Based on the existing model, find the locally optimal hyperparameters and calculate the error loss.

2) Update the hyperparameters and recalculate the loss.

3) Based on the continuous update process, find the hyperparameters that can cause the model to achieve the minimum loss until the predefined number of iterations is reached.

4) Output the optimal hyperparameters.

The value ranges of the hyperparameters and the optimal parameter values for the models are listed in Table 3. The 5-fold cross-validation method was applied to process the dataset. That is, the dataset was divided into 5 parts, among which 4 parts were applied as the training set and the other part was applied as the test set. The five cycles were performed sequentially to finally obtain the predicted value of the model, which can improve the generalization ability of the model. The influence weights of different parameters were calculated using the clf.feature.importances command.

Table 3

The hyperparameter values of the model
Hyperparameters	n_estimators	max_features	max_depth	min_sample_split	min_sample_leaf
Value range	(1,200)	(1, 7)	(1, 20)	(2, 20)	(1, 20)
Optimal value	50	4	5	5	1

3.1 Results of the data analysis and graphical analysis

Table 1 shows that the water inflow from each working face is relatively large where the Yijun Formation is the thinnest in the Gaojiapu coal mine, while the water inflow from the working face is relatively small where the Yijun Formation is relatively thick in the Dafosi coal mine, which shows that the Yijun Formation has a significant controlling effect on mine water inflow.

Figure 3 shows that the overall water inflow at working faces is lower in the east and higher in the west, and the water inflow is the highest in the Zhongyuan-North Tingkou Sag, where the No. 3 panel areais located. Figure 4 shows that the Yijun Formation in the Tingnan Coal Mine is thick in the east and thin in the west. Figures 3 and 4 show that there is greater water inflow at working faces in the area with a thinner Yijun Formation.

3.2 Results of the RF model

In this research, two error analysis indicators, the root mean square error (RMSE) and the coefficient of determination (R²), were used to evaluate model performance. The RMSE reflects the difference between the actual measurement and predicted data, and R² reflects the overall fitting degree of the model. The calculation formulas are shown in Equations 1 and 2.

$RMSE=\sqrt{\frac{1}{n}\sum _{i=1}^{n}{\left({y}_{i}-{\widehat{y}}_{i}\right)}^{2}}$ ,$\in [0,+\infty$) (1)

$${R}^{2}=1-\frac{\sum _{\varvec{i}=1}^{\varvec{n}}{\left({\varvec{y}}_{\varvec{i}}-{\widehat{\varvec{y}}}_{\varvec{i}}\right)}^{2}}{\sum _{\varvec{i}=1}^{\varvec{n}}{\left({\varvec{y}}_{\varvec{i}}-\stackrel{-}{{\varvec{y}}_{\varvec{i}}}\right)}^{2}},\in \left[\text{0,1}\right]$$

A smaller RMSE means that the smaller the error between the predicted data and the actual data, the better the prediction result of the model; a larger R² means that the better the curve fitting degree of the model is, the better the performance of the model.

The evaluation results of the model are R² = 0.94 and RMSE = 39.77, indicating that the prediction accuracy of the model is relatively high. The influence weights of the main factors in the model were calculated and are shown in Fig. 5. The Yijun Formation (L₂) accounts for 68%, verifying that the Yijun Formation is the most important factor affecting the average water inflow at a working face. The weight of the Luohe Formation thickness (L₃) is 24%, and the total weight of the Luohe Formation thickness and sand-to-ground ratio is 26%. The weight distribution shows that the Yijun Formation has a significant effect on the water inflow at a working face, which is greater than the sum of the other factors. Therefore, theoretical analysis of this finding is needed.

The RF model shows that the correlation of the Yijun Formation with mine water hazards reaches 68%, indicating that the Yijun Formation has a great effect on the water inflow at mine working faces. This phenomenon has not been mentioned in previous articles on correlation; therefore, this phenomenon is the focus of this research.

4.1 Effect of the aquifer (aquifuge) in the coal seam roof on the water inflow at working faces in the study area

In the study area, the No. 4 coal seam is the main target of mining. The No. 4 coal seam is located in the upper member of the Yan’an Formation. In addition to the Yan’an Formation aquifer, the Zhiluo Formation and Luohe Formation aquifers also have an impact on the water damage associated with the No. 4 coal seam. Among them, the Jurassic aquifers of the Yan’an and Zhiluo Formations have low water abundance and limited water inflow. The Cretaceous Luohe Formation has moderate to high water abundance and is the main source of the water flowing into working faces. This is why the Luohe Formation has a greater weight in the RF model.

The Yijun Formation consists mostly of gravel strata, but the composition is complex, the sorting properties are poor, and the grains are completely cemented with iron cementation or colloidal cementation. The Yijun Formation acts as an aquifuge with small pores and poor connectivity. The Yijun Formation is a key layer that isolates important aquifers providing water to working faces. The thicker the Yijun Formation is, the better the blockage. Therefore, the Yijun Formation has a certain influence on the water inflow at working faces. However, the weight of the thickness of the Yijun Formation, namely, 68%, cannot be simply explained as a function of the aquifuge because the weight of the thickness from the coal seam to the Luohe Formation, which also acts as an aquifuge, is only 2%. Therefore, it is necessary to analyze the factors controlling the water inflow at working faces in the Yijun Formation from more angles.

4.2 Sedimentary evolution process of the Yan’an Formation to the Luohe Formation and its effect on the water inflow at working faces

Based on previous data and the actual situation of the Binchang mining area, especially the Tingnan Coal Mine, the Yan’an Formation in the study area is considered to be a lacustrine facies, the Zhiluo and Anding Formations are fluvial facies, the Yijun Formation is an alluvial fan, and the Luohe Formation is a desert facies. The sedimentary environment and its evolution characteristics have a great impact on mine water inflow.

Based on the Tingnan Coal Mine as an example, the study area is located on the eastern margin of the Ordos Basin. The paleoterrain was generally high in the east and low in the west. The southern part of the study area is in the middle section of the Lujia − Xiaolingtai anticline, and the Dongjiazhuang anticline is in the north. Most of the study area is located on the southern flank of the Dongjiazhuang anticline and the northern flank of the Lujia − Xiaolingtai anticline, forming two wide and gentle coal accumulation sags, including the Napo–Xingcaowan–Gongposi sag in the north and the Zhongyuan–Tingkou North sag, and the two sags merged into one in the northern Tingkou township.

An analysis of the sedimentary thickness of each strata shows that the lacustrine facies (Yan’an Formation) is generally thick in the north and thin in the south; it is characterized by concentric circles of lake sediments, and the thickest point of sediments is in the Zhongyuan–Tingkou North Sag, which has a certain effect on filling the sag. The river facies (the Zhiluo Formation and the Anding Formation) are thicker in the Zhongyuan–Tingkou North Sag in the south and thinner in the Lujia–Xiaolingtai anticline in the north, which has a further filling effect (Figs. 6–8).

The alluvial fan (Yijun Formation) is relatively thick in the east and relatively thin in the west, indicating that its provenance lay to the east. Due to the characteristics of the alluvial fan (Yijun Formation), the alluvial fan did not fill and level the low-lying area. In contrast, the sediments in the Zhongyuan–Tingkou North Sag are relatively thin, and the sediments in the Lujia–Xiaolingtai anticline area are relatively thick, indicating that the alluvial fan facies increased the degree of depression (Figs. 8–9).

During deposition of the desert facies (Luohe Formation), the sag quickly filled due to the geological effects of wind. However, the bottom morphology of desert facies sediments was determined by the top boundary of the Yijun Formation. Water catchments were more likely to form in depressions in the top boundary of the Yijun Formation, and the Luohe Formation has accumulated large amounts of water. Therefore, when mining below these areas, the water inflow at working faces could be greater. Figure 9 shows that the level of the Luohe Formation above the No. 3 panel areais low, while that of the Luohe Formation above the No. 1 panel areais high. Therefore, the water inflow at working faces in the No. 3 panel areais significantly greater than that at working faces in the No. 1 panel area.

The above analysis shows that although the Yijun Formation is not an important aquifer, it determines the morphology of the bottom boundary of the Luohe Formation as the main aquifer. From this perspective, the Yijun Formation is also an important factor affecting the water inflow at working faces. Although the Luohe Formation is an important aquifer that provides water to the mine, due to its good and stable water richness, its influence weight is less than that of the Yijun Formation.

The lower part of the Luohe Formation in the study area is composed mainly of the arroyo subfacies and the gravel desert subfacies of desert facies (Lin, 2020). It can be further considered that the distribution of these subfacies in the Luohe Formation is mainly determined by the undulation of the top boundary of the underlying Yijun Formation. The high-lying part of the Yijun Formation was denuded as a sedimentary provenance. In the low-lying part of the Yijun Formation, more arroyo deposits formed relatively water-rich areas.

This also shows that the sedimentary environment of the overlying strata is strongly determined by the situation of the underlying strata, which are equivalent to the bedrock. This phenomenon should occur not only in the Binchang mining area but also in other regions with other strata. The strata beneath the aquifer affect the sedimentary subfacies of the aquifer and its water abundnce and distribution.

In this research, the Binchang mining area is the research object, and the Tingnan Coal Mine is the key case. By analyzing the relationship between the water inflow in different mines and the thickness of the Yijun Formation, the water inflow at each working face in the Tingnan Coal Mine and its influencing factors were investigated, and the RF model was applied to explore the main controlling factors affecting mine water inflow to determine the effect of the Yijun Formation on mine water hazard prevention. The following conclusions were reached:

1) Based on the water inflow data of each mine in the Binchang mining area, the important effect of the thickness of the Yijun Formation on the water inflow at working faces was revealed.

2) A qualitative study on the water inflow at each working face in the Tingnan Coal Mine and its influencing factors revealed a negative correlation between the thickness of the Yijun Formation and the water inflow at working faces.

3) The RF model was applied to analyze the influence of different influencing factors on the water inflow, and the thickness of the sedimentary facies of the Yijun Formation was determined to be the main factor with the largest weight affecting the uneven distribution of water inflow in the Tingnan Coal Mine.

4) The Yijun Formation is an important factor affecting the water inflow of the working face not only because the Yijun Formation is an aquiclude but also because the Yijun Formation is controlling the Luohe Formation, which is the main aquifer above the coal seam. The Yijun Formation increased the depression of sediments, and the morphology of the Yijun Formation determined the bottom shape and sedimentary environment of the Luohe Formation. Although the Luohe Formation is an important aquifer, its influence weight is less than that of the Yijun Formation. It can be further considered that underlying strata have an important influence on the sedimentary environment and water abundance of overlying aquifers.

ACKNOWLEDGMENT

This research was supported by the National Natural Science Foundation of China (52274243).

Hu WY et al (2005) Theory and method of mine water disaster control. Beijing:China Coal Industry Publishing House
Wu Q et al (2007) Mine water disaster prevention. China Coal Industry Publishing House, Beijing
Li B, Wu H, Liu P et al (2023) Construction and application of mine water inflow prediction model based on multifactor weighted regression:Wulunshan Coal Mine case. Earth Sci Inf ;16(2)
Dong S, Ji YD, Wang H et al (2020) Prevention and control technology and application of roof water disaster in Jurassic coal field of Ordos Basin. J China Coal Soc 45(7):2367–2375. 10.13225/j.cnki.jccs.DZ20.0697
Huang YL, Guo YC, Qi WY et al (2022) Evolution characteristics of mining groundwater flow field and water gushing control technology of working Face in typical mining areas in Western China. Bull Natl Nat Sci Foundation China 36(06):1016–1026
Jin DW, Li CF, Liu YF et al Characteristics of roof water hazard of coal seam in Huanglong Coalfield and key technologies for prevention and control. Coal Geol Explor. 2022:1–9
Van Dijk WM, Densmore AL, Singh A et al (2016) Linking the morphology of fluvial fan systems to aquifer stratigraphy in the Sutlej-Yamuna plain of northwest India. J Geophys Research: Earth Surf 121(2):201–222. 10.1002/2015jf003720
Kunkel C, Aehnelt M, Pudlo D, Kukowski N, Totsche KU, Gaupp R (2018) Subsurface aquifer heterogeneities of Lower Triassic clastic sediments in central Germany. Mar Pet Geol 97:209–222. 10.1016/j.marpetgeo.2018.06.022
Koah Na Lebogo SP, Bisso D, Ngo Elogan Ntem J, Mvondo Ondoa J (2021) Stratigraphic architecture and depositional environment of the Paleocene-Lower Eocene aquifers system (Douala onshore basin, SW Cameroon). J Petroleum Explor Prod Technol 11(2):549–560. 10.1007/s13202-020-01057-2
Mu PF, Zhu KP, Mu L (2015) Luohe Formation water yield property pattern in Jurassic Huanglong Coalfield. Coal Geol China 27(5):42–45
Zhao BF (2015) Influence of sedimentary and structural characteristics on aquifer water abundance. Geotech Invest Surveying 43(09):51–54
Wang YG, Wang M, Liu ZP et al (2019) Water enrichment characteristics of roof aquifer under conditions of multifacie sedimentation in Bayangaole Coal Mine. Coal Geol Explor 47(S1):56–61
Lin L (2020) Study on the law of sandstone sedimentation control in Luohe Formation in Gaojiapu Coal Mine. Master. Xi’an University of Science and Technology. 10.27397/d.cnki.gxaku.2020.000540
Zhao BF (2015) Water yield property zoning based on aquifer sedimentary and structural features. Coal Geol China 27(4):30–34
Wu Q, Wang Y, Zhao DK et al (2017) Water abundance assessment method and application of loose aquifer based on sedimentary characteristics. J China Univ Min Technol 46(03):460–466
Li D, Liu SY, Zhang GD et al (2017) Typical roof water disasters and its prevention & control technology in the north of Ordos Basin. J China Coal Soc 42(12):3249–3254
Wang Y, Wu YH, Ge Q et al (2023) A sedimentary facies-based method to control water hazards in the roof of deep Jurassic coals. Mine Water Environ 42(2):240–250. 10.1007/s10230-023-00930-w
Dou GD, Xiao F Discussion on water control technology in Xiaozhuang Coal Mine of Binchang mining area. Coal mine water disaster prevention and control technology research–Shaanxi coal society academic annual meeting papers. Beijing: China Coal Industry Publishing House;2013:138–141
Wu MD, Wang JH, Hou EK et al (2018) Roof water bursting rules and influential factors of Dafosi coal mine. J Xi'an Univ Sci Technol 38(04):636–642
Wu MD, Yuan DS, Zhang LD (2021) Analysis on water inflow difference of working face in Binchang mining area. Shaanxi Coal 40(S1):42–46
Wang G (2021) Analysis of hydrogeological characteristics and water filling factors in Hujiahe Coal Mine. Min Technol 21(02):31–33
Xia YF, Liu CZ, Li YY et al (2017) A boosted decision tree approach using bayesian hyperparameter optimization for credit scoring. Expert Syst Appl 78:225–241. https://doi.org/10.1016/j.eswa.2017.02.017
Jin ZW, Shang JX, Zhu QW et al (2020) RFRSF:Employee turnover prediction based on random forests and survival analysis. Web Inform Syst Eng 12343:503–515. https://doi.org/10.1007/978-3-030-62008-0_35
Youssef AM, Pourghasemi HR, Pourtaghi ZS et al (2016) Landslide susceptibility mapping using random forest, boosted regression tree, classification and regression tree, and general linear models and comparison of their performance at Wadi Tayyah Basin, Asir Region,Saudi Arabia. Landslides.;13(5):839–856. https://doi.org/10.1007/s10346-015-0667-1
Li Y, Shami A (2020) On hyperparameter optimization of machine learning algorithms: Theory and practice. Neurocomputing 415:295–316

Download PDF

Reviewers agreed at journal
21 Apr, 2024
Reviewers invited by journal
16 Apr, 2024
Editor assigned by journal
03 Apr, 2024
First submitted to journal
01 Apr, 2024

You are reading this latest preprint version

A new understanding of the mechanism controlling water inflow into mines based on the stratigraphic distribution beneath major aquifers through a case study of the Yijun Formation in the Binchang mining area, Shaanxi Province, China

Status:

Version 1

Abstract

Figures

1 INTRODUCTION

2 MATERIALS AND METHODS

2.1 Overview of the study area

2.2 Data analysis of the relationship between mine water inflow and the distribution of the Yijun Formation

2.2.1 Data analysis of the Binchang mining area

2.3 RF model analysis of the water inflow at working faces in the Tingnan Coal Mine

2.3.1 RF model

2.3.2 Data selection for the RF model

2.3.3 Hyperparameter determination of the RF model

3 RESULTS

3.1 Results of the data analysis and graphical analysis

3.2 Results of the RF model

4. DISCUSSION

5 CONCLUSIONS

Declarations

ACKNOWLEDGMENT

References

Status:

Version 1