Interactions of overburden failure zones due to multiple-seam mining using longwall caving

This paper presents an investigation on the interactions of overburden failure zones induced by the mining of adjacent coal seams using the longwall caving method. Overburden failure is an important factor in safety assessments in the fields of mining engineering geology and safety geology, especially when mining under water bodies. In this study, the influence of the thickness and properties of the interburden between seams on the development and interactions of caving and fractured water flow zones are investigated by using in situ measurements, scale model testing and numerical simulations. The height of the fractured water flow zones in the scale model tests and numerical simulations are basically in good agreement with measurements after mining of the upper and lower seams of Seam No. 3 in the Cuizhuang Coalmine. Therefore, the scale tests and numerical simulations in the study are verified. The results show that interaction and superposition between two close distance seams cannot be ignored when the ratio (h/M) of the interburden thickness (h) to the cutting height of the lower seam (M) is less than a defined critical value. A dividing line, Line D, has been proposed to judge whether the interactions exist. When the (M, h/M) points are located above Line D, the caving zone induced by excavation of the lower seam will not propagate to the caving zone induced by the upper seam. Otherwise, for the (M, h/M) points below Line D, the interactions and superposition of the overburden failure must be considered when predicting the heights of the caving and fractured water flow zones.


Introduction
Coal deposition is usually in the form of multiple seams with different interburden thicknesses and geologies. Much research on the interactions of strata stress and deformation due to multiple seam mining have been carried out for room-and-pillar and longwall panels since the 1980s (Mark 2007;Mark et al. 2007). Some of the earlier research mainly focused on room-and-pillar mining. It was found that subsidence and pillar load transfer are two of the typical interactions between adjacent coal beds (Chekan et al. 1986). An evaluation model for the probability of multiple seam mining success was proposed by Chanda (1989). Chekan et al. (1988) found that the ratio of the overburden to interburden thickness is a key factor influencing load transfer. Later, multiple seam interactions associated with longwall mining or interactions between room-and-pillar and longwall panels were investigated with a focus on their complex multiple-seam stress and displacement interactions (Chekan et al. 1989;Heasley and Akinkugbe 2005;Porathur et al. 2013). Results demonstrated that during multiple-seam excavations, it is important to protect the overlying workable seams and minimize surface subsidence.
The longwall top coal caving (LTCC) method is an effective mining approach that originated in Europe, and has been widely and successfully used in China. Overburden stress distribution, top coal caving and fracturing mechanisms and evaluations, and interaction between strata movement and hydraulic supports have attracted much attention and have been thoroughly studied since the LTCC method increased in popularity (Xie et al. 1999;Yasitli and Unver 2005;Humphries and Poulsen 2008;Alehossein and Poulsen 2010;Vakili and Hebblewhite 2010;Khanal et al. 2011;Swift 2013;Gao et al. 2014). In recent years, research on the influence of the LTCC method on close-distance seam mining has been carried out by investigating overburden failure and stress distributions using in situ measurements, numerical simulations and scale model testing (Yang et al. 2008;Luan et al. 2010;Zhang and Zhang 2010;Guo 2011;. Results show that interactions on overburden failure during the excavation of close-distance seams are enhanced by the superposition between the floor damage zone of the upper seam and the roof caving zone of the lower seam. This superposition results in increases in the height of the caving and fractured water flow zones of the lower seam, which are greater than those induced by single-seam mining under similar conditions. Previous studies in the literature revealed some characteristics of the interactions of overburden failure and subsidence induced by the mining of close-distance seams. The damaged interburden and/or caving rocks of the upper seam due to longwall mining constitutes part of the roof of the lower seam. It has been found that the influence of upper seam mining on the stress, deformation and failure of lower seam mining is reduced with increasing interburden thickness. However, the understanding on overburden failure due to multiple seam mining falls short of that on single seam mining, especially in terms of the LTCC method. Some issues require more profound studies and understanding, such as how multiple seams interact during caving mining, to the extent that the thickness and properties of the interburden influence interactions, and how mining sequences and directions of recovery influence interactions. To this end, the main purpose of this study is to investigate the influence of the thickness and geological engineering properties of the interburden between two close seams on the interactions of overburden caving and fracturing, based on a case study of Seam No. 3 (including the upper and lower seams) of the Permian Shanxi Formation in the Cuizhuang Coalmine, Shandong, China.

Hydrogeological and geological engineering conditions
The Cuizhang Coalmine The Cuizhuang Coalmine is located in the Weishan county, 260 km south of the capital of the Shandong province, Jinan, in China (Fig. 1). Annual mine production was 0.6 million tons from 1998 to 2002 and its annual output has doubled since 2003. The upper and lower seams of No. 3 are the productive coal seams in the Permian Shanxi Formation. The mine was developed by using two vertical shafts and a single mining level at an elevation (EL.) of -250 m below the mean sea level (msl). The Cuizhuang Coalmine has an area of approximately 11.9 km 2 , with a length of 4.2 km in the east-west direction and a width of 2.5-3.3 km in the north-south direction. About two-thirds of the total geological coal reserves were deposited under  Fig. 1 Location of the Cuizhuang Coalmine Weishan Lake, the largest freshwater lake in northern China with an area of 1,266 km 2 , an average water depth between 1.5 and 3 m, a maximum reservoir of 4.73 billion m 3 and a mean water level of EL ?32.23 to ?34.21 m.
Lithology and hydrogeology Table 1 lists the stratigraphic and hydrogeological characteristics of the Cuizhuang Coalmine. The coal-bearing strata in the mine, the Carboniferous and Permian Systems, are typical coal-measures in the North China coal basin. The Permian Shanxi Formation bears 2-3 coal seams, including the upper and the lower seams of No. 3. The upper seam has a thickness of 0-6.57 m and the lower seam has a thickness of 0-5.17 m, where zero means sedimentary absence in some of the locations. The Shanxi Formation mainly consists of fine sandstone, mudstone, silty mudstone, siltstone.
The total water inflow in the Cuizhuang Coalmine has been stable with a maximum value of 40 m 3 /h, mainly from sandstone aquifers on the roof. The hydraulic relationships between the rock and soil aquifers are separated by several aquifuge layers. However, when coal mining takes place under the Neogene aquifers, the height of the fractured water flow zone is carefully controlled to avoid water and/or sand inrush from unconsolidated aquifers. Therefore, an accurate prediction for the height of overburden failure has become one of the critical issues for risk assessment and mining safety, especially in multiple seam mining.

Tectonic structure
The Cuizhuang Coalmine is geologically located on the southern rim of the Tengxian anticline hydrogeological setting, which tectonically belongs to the eastern margin of the Western Shandong Block of the North China Platform.
The geological structure of the mine is generally a monocline structure with a set of secondary gently undulating folds, varying in occurrence along the strike and dip directions (Fig. 2). The Cuizhuang anticline and the Chaili synclines are the main structures, accompanied by several normal faults. There are six large faults in the mine, which are boundary faults except for the Tiangang fault and SF1 fault across the mine. Among them, two are hydraulically permeable, two are impermeable, according to groundwater tests, and the remaining two have not been identified. The throw of the Tiangang fault is between 80 and 180 m, extending to the southwest beneath Weishan Lake. It is covered by the Neogene System with a bottom clay 49.89 (drilled depth) Confined karstic and thick fissured limestone aquifer, q = 0.14 L/s/m, excellent wateryielding capacity. Groundwater level declines about 3 m annually due to regional mine drainage Interactions of overburden failure zones due to multiple-seam mining using longwall caving 1021 5.31-23.18 m thick. The bottom clay layer protects the coal mine from water seepage from sand aquifers along the faults in the coalmine.
Engineering geology of overburden and interburden   -100 and -70 m msl with some topographical fluctuations in the coal mining area.

Methodologies
Overburden failure includes three zones from the roof to surface: caving zone; fractured zone; and bending zone. The permeability in the caving and fractured zones changed dramatically and formed the groundwater pathway from overburden aquifers into the panels. Therefore, the composition of the caving zone and the fractured zone is known as a ''water flow fracture zone.'' The height of the water flow fracture zone is the height from the roof of the seam to the upper boundary of the fractured zone. The caving zone consists of the lower part of the fractured water flow zone (Fig. 4).

Measurements
In-situ measurements have been one of the most important and direct methods for the investigation of the height of overburden failure zones. In this traditional measurement method, water loss is investigated by the drilling of boreholes. Currently, underground water pumping, ultrasonic imaging and other geophysical approaches have been adopted in overburden failure measurement. Figure 4 shows the principles of ground and underground measurement methods using boreholes. A so-called two-terminal plugging tool was used in the Cuizhuang Coalmine to measure water loss in different boreholes to determine the height of the overburden failure zone. The boreholes were designed and drilled before and after mining. The borehole was separated into 0.1 m or 0.2 m sections. The ends of each section were plugged with an expansion rubber inside the borehole. High pressure water was injected into the isolated section and water loss was measured prior to and after coal mining. The height of the fractured water flow zone was determined by comparing water loss along the boreholes pre-and post-excavation (Xu and Sui 2013).

Scale model test
Scale model testing is one of the common methods used in the research of overburden failure. In this study, the dimensions of the scale model test rig are 6 m 9 3 m 9 0.4 m (length 9 height 9 thickness). The model parameter scales are as follows: (1) 1:200 for geometry; (2) 1:1.7 for gravity; and (3) 1:200 for uniaxial compressive strength (UCS). Materials used for modeling comprised a combination of sand, barite, gypsum, mica, sawdust, etc. Mica powder was well distributed between the rock layers to Yuanta ng Fault F4 Fa ul t T i a n g a n g F a u l t X u z h u a n g F a u lt G a o m ia o F a u lt S F 1 F a u lt P a n e l N o . 3 3 0 2 P a n e l N o . 3 3 0 1 S C u i z h u a n g A n t i c l i n e Y a o q ia o F a u lt C u i z h u a n g F a u l t  Table 3 lists the four types of scale modeling conducted in this paper to simulate different mining and geological conditions. The major differences among them are found in the cutting height and the interburden thickness. Scale Model (SM) No. 1 simulates Panels No. 3301S upper and 3301S lower to verify the scale modeling through a comparison of the height of the overburden failure zone with that from the in situ measurements. SM No. 2 is basically the same as SM No. 1 except that the interburden is changed to 14 m to study the influence of the interburden thickness on the height of the overburden failure zone. The other two models, SMs No. 3 and 4, are designed to simulate fracture propagation due to mining in the close-distance seams in Panel 3302 along the dip and strike cross sections (illustrated in Fig. 5) that correspond to those in Fig. 3. In SMs No. 3 and 4, the cutting height of the upper and the lower seams is 6.0 and 4.3 m, respectively; the thickness of the interburden is 10.6 m. Excavation was carried out in the mining entrance from left to right. Every day, a 4-cm seam (in model size) of coal was mined out; this equals 8 m in the prototype. The lower seam was mined by using the same process after excavation of the upper seam was completed. The displacement of the strata during the mining simulation was measured by using linear variable differential transformers (LVDTs). For every step of the excavation, the deformation and failure of the overburden were recorded and images were taken with a high speed camera. Then, the movement and deformation due to mining were analyzed by digital image processing.

Numerical simulation
Numerical modeling needs to be verified and validated against (1) the overburden failure measurement in the study area and (2) the overburden failure induced by fracture propagation in the scale experiments before it can be used to study multiple seam interactions in mining. It was found that the location of a sudden change of vertical displacement, which shows a concentration of vertical displacement   Interactions of overburden failure zones due to multiple-seam mining using longwall caving 1025 A Fast Lagrangian Analysis of Continua (FLAC 2D) code (Itasca 2009) was used in this study. Boundary conditions for the plane strain models are free displacement and loading on the top, fixed at the bottom, fixed in the horizontal direction and free in the vertical direction on both side boundaries. Parameters of the different geological engineering types were carefully determined by using laboratory tests and geological observations in accordance with the Hoek-Brown criteria (Hoek and Brown 1997). Extraction of coal was conducted by the function of ''model null'' in the code. The maximum convergence between the roof and floor was limited to less than the cutting height of the seam.

Measurement and verification for modeling
Two panels of the Cuizhuang Coalmine were selected for overburden measurements: Panels No. 3301S upper and 3301S lower (locations shown in Fig. 2 and Panel No. 3301S denoted). The upper and the lower seams of No. 3 were separated by a 7 m thick interburden composed of grey, thin-layered silty mudstone and siltstone. Significant amounts of fossilized plant fragments were found along the bedding plane that grew on the horizontal bedding on the interburden, meaning that caving easily took place and filled the mine gob after excavation of the lower seam. The dip angle of the strata above the panels was 8°on average and varied between 0°and 11°.
The above-mentioned water loss method was used to measure overburden failure induced by upper seam mining with LTCC and carried out on Panel No. 3301S upper with 3 underground boreholes. The measurements were completed on July 2007. Results showed the maximum height of the caving zone is 19.6 m; the maximum height of the fractured water flow zone is 59.59 m above the roof of the upper seam. The height ratio of the caving and fractured water flow zones against the cutting height of 5.0 m is 3.9 and 11.9, respectively. Measurements for the lower seam were conducted on Panel No.  water flow zone against the equivalent cutting height of 7.2 m is 9.2.
where M is the equivalent thickness of the upper and lower seams; M upper and M lower are the cutting height of the upper and lower seams, respectively; h interburden is the thickness of the interburden; y lower is the ratio of the height of the caving zone to the thickness of the lower seam; this study assumes that it has the same value of 3.9 as the upper seam does. Therefore, predicting the height of the fractured water flow zone H f for mining the upper and the lower seams of No. 3 in the Cuizhuang Coalmine can be calculated by using the formula H f = 9.2 M that takes interactions into account with a rule of thumb. Here, please note that the heights of overburden failure are calculated from the roof of the upper seam.
For verification purposes, Table 4 shows a comparison of the physical and numerical simulations based on in situ measurements, numerical simulations and scale modeling. The results show the interactions lead to the superposition of overburden failure due to the mining of two close distance seams. The height of the fractured water flow zone is increased from 52.1 to 68.0 m because of superposition failure due to excavation of the upper and lower seams with an interburden thickness of 7 m, see SM No. 1. The results in Table 4 verify the feasibility of using numerical simulation and scale modeling to study overburden failure, caving and fracturing interactions in multiple-seam mining. The results from SM No. 2 show that the height of the fractured water flow zones is increased from 52.1 to 64.8 m above the roof of the upper seam, thus indicating the influence of the interburden thickness when it is changed to 14 m. These results indicate there is still interaction of rock caving and fracturing in close-seam mining when the interburden thickness reaches 14 m, but the interaction increases when the interburden thickness is 7 m. 3. The development of caving and fractured water flow zones can be interpreted through a detailed analysis of the mining process in a stepwise consecutive manner. The first caving occurred when panel length was advanced to 24 m (the prototype size, same below in this paper). After that, the roof periodically cracked and the height of the caving and fractured water flow zones gradually increased until rocks filled the mined out area behind the working face. Rocks almost fully filled the mined  Interactions of overburden failure zones due to multiple-seam mining using longwall caving 1027 out area when the recovery length reached 84 m. After mining, the height of the caving and fractured water flow zones reached 17.6 and 48.0 m, respectively. The height of the fractured water flow zone was expected to increase if recovery of coal continued. The outline of the fractured water flow zone has the shape of an asymmetric ladder. Figure 6e-h show development of the overburden failure zone during mining of the lower seam of No. 3 from SM No. 3. The interburden between the upper and lower seams did not cave until the distance of coal recovery reached 44 m; the fractured water flow zone also did not develop further until this length. The caving zone of the upper and lower seams merged into one another when the mining length was advanced to 56 m, and the height of the fractured water flow zone continuously increased until its ladder-shaped outline reached the bottom clay layer. Figure 7 shows caving height variations and fractured water flow zones with coal recovery in the coal seam. Caving zone heights reaches the maximum value first, and then the fractured water flow zone reaches its highest position with a longer recovery distance, depending on the cutting height and overburden features. The maximum   Figure 12 shows the interactions of overburden failure influenced by the interburden thickness and the cutting height of the seams. Figure 12a   Interactions of overburden failure zones due to multiple-seam mining using longwall caving 1031 obvious calculation errors. The interactions of the fractured water flow zones induced by mining of the two seams showed the same performance; that is to say, when h/M is greater than 6.8, the interactions could be neglected and the mining of the lower seam barely increases the height of the fractured water flow zone induced by mining of the upper seam. Therefore, the height of the caving zone can be calculated in terms of single-seam excavation when the h/ M ratio is greater than 6.8 in this type of geological and mining condition, while the height of the fractured water flow zone remains almost the same as that induced by upper seam mining. Figure 12b shows the results of NMs No. 7-12 for a cutting height of 4.3 m for the lower seam and the same conditions mentioned above. In this case, the turning point of h/M is 4.5 for the interactions of the caving zone and 3.9 for the fractured water flow zone. Moreover, Fig. 12c shows results for NMs No. 13-18; the h/M of the interactions in mining the upper and lower seams at a cutting height of 5.3 m is 3.7 and 3.2 for the caving and water flow fracture zones, respectively. Figure 13a summarizes some of the results of this paper and those from the literature (Zhang 2010;Guo 2011;Luan et al. 2010) to illustrate the interactions of overburden failure due to recovery of coal from multiple seams, with varying interburden thicknesses and cutting heights of the lower seam. If a point that represents (M, h/ M) falls under Line D, interaction or superposition should be taken into account to calculate the height of the overburden failure (Fig. 13b). Otherwise, the interaction can be neglected when the point drops above Line D; therefore, the height of the caving zone can be calculated by treating it single-seam mining, while the height of the fractured water flow zone will not be surpassed due to mining of the upper seam (Fig. 13c). Line D may be defined as the dividing line that determines whether interaction causes overburden failure superposition in close-seam mining.
The results of this paper could provide a helpful reference in the recovery of coal in close-distance seams by using the LTCC method. However, the interactions presented in this paper still need further investigation. An orthogonal array can be adopted to investigate the influence of factors (overburden and interburden geotechnical properties, cutting heights of multiple seams) on the interactions of multiple-seam mining in further research. Multiple-seam mining with more than two seams of different geological engineering structure, discontinuities, geo-stress environments and dip angles, especially for mining of steeply dipping seams, will also be further considered. Moreover, coupling the seepage, failure and stress fields should be taken into account for an intrinsic understanding of the mechanism of multiple mining interactions. Results need to be validated by more in situ measurements.

Conclusions
Multiple methods, including in situ measurements, scale model testing, numerical simulations and empirical formulas are used in this paper for the investigation of interactions on overburden failure due to mining of multiple thick coal seams using longwall caving. Included is a case study on the Cuizhuang Coalmine that uses LTCC to mine thick seams in the Shandong province of China. The paper focuses on the interactions from the development of the caving zone and fractured water flow zones, which play a significant role in the safety evaluation and technical decision-making when mining under surface water and groundwater. The influence of the interburden thickness, cutting heights of the seams and geotechnical properties of the overburden on the interactions and superposition of overburden failure have been analyzed in detail and the following conclusions are drawn.
1. The results from the in situ measurements, scale modeling and numerical simulations show a good agreement and, therefore, validate the scale and numerical modeling methods by examining the influencing effects on the interactions in the mining of multiple thick coal seams using the longwall caving method. 2. Results of the scale and numerical modeling indicate that the interburden thickness and cutting height have a significant influence on overburden failure in the multiple seam interactions of caving and fractured water flow zones. The interactions decrease with increases in the interburden thickness. The interaction cannot be neglected between two close seams when the ratio of the interburden thickness to the cutting height of the lower seam is less than a defined critical value. 3. A dividing Line D has been determined for the calculation of the interactions on overburden failure and proposed according to the results of this paper and previous studies. Line D conceptually determines whether it is necessary to take the interactions in multiple seam mining into consideration in the calculation of the heights of the caving and fractured water flow zones. When (M, h/M), where h/M represents the ratio of the interburden thickness h to the cutting height of the lower seam M, are above Line D, the caving zone induced by mining the lower seam does not propagate to the caving zone induced by mining the upper seam. The height calculations can then be neglected in the interactions of overburden failure without resulting in obvious errors. Otherwise, when (M, h/M) fall below Line D, the interaction and superposition of the overburden failure must be considered in the calculation of the heights of the caving and fractured water flow zones.