Monitoring and Analysis of Stress and Deformation Features of Boundary Part of Backﬁll in Metal Mine

: The backﬁll mining method is widely used in metal mines. A large and thick backﬁll body has formed in the No.2 zone of Jinchuan Nickel Mine, and its stability is critical for mining safety. In order to study the mechanical behavior of the boundary part of backﬁll, ground subsidence monitoring, underground ﬁled monitoring of displacement, and stress and numerical simulation were conducted to analyze stress distribution and deformation of backﬁll. According to underground monitoring, the bed separated displacement has the consistent trend with ground subsidence in the mine area. Based underground stress monitoring, both horizontal and vertical stress of the internal part of backﬁll is less than the stress boundary part of the backﬁll. The characteristic of backﬁll boundary outline is a step-proﬁle. Contact interaction between the surrounding rock and backﬁll led to complex stress distribution. According to stress monitoring of the boundary points in a numerical model, the multi-peak value of stress development is the major feature of the boundary part of backﬁll. The multi-peak stress behavior of the boundary part of backﬁll was inﬂuenced by mining depth. The boundary part of backﬁll deformation inﬂuenced the local stability of mining. This article provided a scientiﬁc basis for strength design and the support choice of a metal mine by the method of backﬁll mining.


Introduction
In deep mining engineering, the backfill mining method is widely used in many metal mines all over the world. This method has a vast number of advantages, like occupying space, restricting deformation, bearing load, and realizing the sustainable development of the mining industry [1][2][3][4][5]. Backfill is a composite material with a multiphase consisting of tailing, cement, and a wide variety of micro cracks and micro pore [6,7]. The materials of backfill are adhesive. Durability models of adhesive material using catastrophe theory and analyze the destruction mechanism [8][9][10]. Backfill material has unique properties, and these properties are crucial for backfill mechanical behavior. The backfill properties depend on their composition parameters, such as blinder ratio, grain size, and additions [11,12].
Compared with rock mass, backfill is a solid mass with lower strength. For mining safety, backfill strength and mechanical behavior have drawn the attention of many researchers. An analysis of mechanical action of backfill in deep mining was carried by Ryder [13]. Ercikdi et al. investigated the influence factor of backfill strength [14]. The backfill failure mechanism and the effect of material property were studied by some experimental tests. Finally, some studies have demonstrated that the tensile principal stress has led to backfill failure inducing spalling near the exposed surface [15]. Backfill stability is a crucial factor for mining safety. Proper design is helpful for backfill stability, which was analyzed by a sliding failure model of a confined block [16,17].
Backfill replaces ore body and occupies the mining void. Contact interaction between the surrounding rock and backfill have influenced local stability in mining engineering [18][19][20][21]. The shear behavior of joints formed between concrete or cement grout and soft, weak, or weathered rock was described by experimental analysis [22][23][24][25]. The stress state in a single filled stope was studied in some mines. The difference in stiffness and strength between rock and backfill tends to produce a load transfer along the interfaces due to the downward settlement of the backfill [26][27][28]. The load transferred along the interface was confirmed by some experiments in laboratory and mining operations in mines [29][30][31][32][33][34][35].
Previous research into backfill focused on aspects of material properties, blinder ratio, and mechanisms. The arch effect of the narrow backfill and the surrounding rock are not suitable for this area mine. First, the ratio of width to height of a single stope and single-stope backfill are large. The single-stope backfill is not narrow. Second, the overall size of the single-stope backfill is small, and its exposure time is short. The single-stope backfill goes together with an adjacent layer and stopes backfill. So, in study area, the issue of backfill mechanical behavior is analyzed from whole or local stability. This article studied backfill mechanical behavior and stability based on a field investigation, monitoring data, and numerical simulation. This study could provide suggestions for mining safety and support operation.

Geology
In this article, the major study area is No.2 zone of the Jinchuan Nickel Mine. The Jinchuan Nickel Mine has the largest deposit of nickel in China. The mine is located in the city of Jinchang in northwest of China. The ore deposit is about 6.5 km long, 10s of meters to 570 m in width, and extends to more than 1000 m in depth [36][37][38]. The strike of the ore body is northwest and dips to the southwest, and the dip angle ranges from 40 • to 70 • . The profile of ore body existence is shown in Figure 1. of mechanical action of backfill in deep mining was carried by Ryder [13]. Ercikdi et al. investigated the influence factor of backfill strength [14]. The backfill failure mechanism and the effect of material property were studied by some experimental tests. Finally, some studies have demonstrated that the tensile principal stress has led to backfill failure inducing spalling near the exposed surface [15]. Backfill stability is a crucial factor for mining safety. Proper design is helpful for backfill stability, which was analyzed by a sliding failure model of a confined block [16,17]. Backfill replaces ore body and occupies the mining void. Contact interaction between the surrounding rock and backfill have influenced local stability in mining engineering [18][19][20][21]. The shear behavior of joints formed between concrete or cement grout and soft, weak, or weathered rock was described by experimental analysis [22][23][24][25]. The stress state in a single filled stope was studied in some mines. The difference in stiffness and strength between rock and backfill tends to produce a load transfer along the interfaces due to the downward settlement of the backfill [26][27][28]. The load transferred along the interface was confirmed by some experiments in laboratory and mining operations in mines [29][30][31][32][33][34][35].
Previous research into backfill focused on aspects of material properties, blinder ratio, and mechanisms. The arch effect of the narrow backfill and the surrounding rock are not suitable for this area mine. First, the ratio of width to height of a single stope and single-stope backfill are large. The single-stope backfill is not narrow. Second, the overall size of the single-stope backfill is small, and its exposure time is short. The single-stope backfill goes together with an adjacent layer and stopes backfill. So, in study area, the issue of backfill mechanical behavior is analyzed from whole or local stability. This article studied backfill mechanical behavior and stability based on a field investigation, monitoring data, and numerical simulation. This study could provide suggestions for mining safety and support operation.

Geology
In this article, the major study area is No.2 zone of the Jinchuan Nickel Mine. The Jinchuan Nickel Mine has the largest deposit of nickel in China. The mine is located in the city of Jinchang in northwest of China. The ore deposit is about 6.5 km long, 10s of meters to 570 m in width, and extends to more than 1000 m in depth [36][37][38]. The strike of the ore body is northwest and dips to the southwest, and the dip angle ranges from 40° to 70°. The profile of ore body existence is shown in Figure 1.

Mining Situation
The mechanized backfill mining technique was adopted at the Jinchuan Mine. In the mining process, once excavation is finished within an access drift, the void is filled using cement paste before the next access drift is excavated. Mining proceeds from top to bottom in a series of horizontal four-meter-high drift slices. Mining panels are approximately 100 m wide and perpendicular to the strike of the ore body [39][40][41]. To increase production, simultaneous mining is carried on two or more sub-levels. The annual output of No.2 zone of the Jinchuan Nickel Mine is over four million tonnes. The mining situation in the study area is shown in Figure 2. The current mining operation carries over 1000 m underground. The field investigation and measurement were extended in a sub-level of 1050 m.

Mining Situation
The mechanized backfill mining technique was adopted at the Jinchuan Mine. In the mining process, once excavation is finished within an access drift, the void is filled using cement paste before the next access drift is excavated. Mining proceeds from top to bottom in a series of horizontal fourmeter-high drift slices. Mining panels are approximately 100 m wide and perpendicular to the strike of the ore body [39][40][41]. To increase production, simultaneous mining is carried on two or more sublevels. The annual output of No.2 zone of the Jinchuan Nickel Mine is over four million tonnes. The mining situation in the study area is shown in Figure 2. The current mining operation carries over 1000 m underground. The field investigation and measurement were extended in a sub-level of 1050 m.

Backfill Overview in Study Area
In the No.2 zone of the Jinchuan Nickel Mine, the elevation of the first mining layer is 1334 m, so the top of the backfill starts from this level. A horizontal mining layer is divided into six to eight panels, as shown in Figure 2. The size of each mining panel is about 100 m wide and is perpendicular to the strike of the ore body and backfill, and the void caused by mining is immediately filled. The mining operation is transferred the next layer after each panel of ore body is excavated. A layer of backfill is formed after all panels in same layers are filled. The backfill material is cement paste with tailings and some additives. Generally, the cement:sand ratio of the backfill is 1:4, and the curing time is 28 d. Slurry concentration is greater than 80.8%. The backfill replaces the ore body without pillars, so the upper layer backfill serves as the artificial roof of the next mining layer. Over 30 years of mining, a large and thick backfill body has formed, and its depth is more than 1000 m. The mechanical properties of backfill were obtained by a series of tests. The average uniaxial compression strength of the backfill is 5.09 MPa. The passion ratio of the backfill is 0.146, while the passion ratio of the surrounding rock is 0.261. The average tensile strength is 0.59 MPa.
With an accurate underground measurement and design drawing, we can build the exact boundary of the backfill body in the study area. The boundary provides the geometric basis of the following numerical and mechanical analysis. Because of the mining operation, the measurements were conducted in three sessions, and the three parts of the backfill are shown in Figure 3. The three parts of the backfill are located at 1334-1250 m, 1250-1100 m, and 1100-1050 m, respectively. The length maximum is 1600 m, and the width maximum is 250 m.

Backfill Overview in Study Area
In the No.2 zone of the Jinchuan Nickel Mine, the elevation of the first mining layer is 1334 m, so the top of the backfill starts from this level. A horizontal mining layer is divided into six to eight panels, as shown in Figure 2. The size of each mining panel is about 100 m wide and is perpendicular to the strike of the ore body and backfill, and the void caused by mining is immediately filled. The mining operation is transferred the next layer after each panel of ore body is excavated. A layer of backfill is formed after all panels in same layers are filled. The backfill material is cement paste with tailings and some additives. Generally, the cement:sand ratio of the backfill is 1:4, and the curing time is 28 d. Slurry concentration is greater than 80.8%. The backfill replaces the ore body without pillars, so the upper layer backfill serves as the artificial roof of the next mining layer. Over 30 years of mining, a large and thick backfill body has formed, and its depth is more than 1000 m. The mechanical properties of backfill were obtained by a series of tests. The average uniaxial compression strength of the backfill is 5.09 MPa. The passion ratio of the backfill is 0.146, while the passion ratio of the surrounding rock is 0.261. The average tensile strength is 0.59 MPa.
With an accurate underground measurement and design drawing, we can build the exact boundary of the backfill body in the study area. The boundary provides the geometric basis of the following numerical and mechanical analysis. Because of the mining operation, the measurements were conducted in three sessions, and the three parts of the backfill are shown in Figure 3. The three parts of the backfill are located at 1334-1250 m, 1250-1100 m, and 1100-1050 m, respectively. The length maximum is 1600 m, and the width maximum is 250 m.

Backfill Deformation and Ground Subsidence
The primary reason for ground movement is backfill deformation. Global Positioning System (GPS) monitoring points were laid out in the Jinchuan No.2 mine in 2001 in order to study overlying rock mass deformation and ground movement induced by backfill deformation. Surface fissures were also monitored. Field monitoring was carried out every six months, beginning in 2001. Monitoring result show that the ground movement is lagging behind the underground mining and filling.
The displacements of ground monitoring points can reveal the characteristics of backfill deformation to some degree. Figure 4 shows the horizontal displacement vectors of the ground monitoring points. The displacements of the monitoring points in the hanging wall are larger than the points in the foot wall. All horizontal displacement vectors pointed to a centrally located settlement. The settlement center was located on exploration line-14. This phenomenon indicated that the backfill deformation near exploration line-14 is larger than other exploration lines. Horizontal displacement vectors end in the boundary of the backfill in sub-level 1250 m and sub-level 1150 m. The stress distribution of the boundary part of the backfill is more complex.

Backfill Deformation and Ground Subsidence
The primary reason for ground movement is backfill deformation. Global Positioning System (GPS) monitoring points were laid out in the Jinchuan No. 2 mine in 2001 in order to study overlying rock mass deformation and ground movement induced by backfill deformation. Surface fissures were also monitored. Field monitoring was carried out every six months, beginning in 2001. Monitoring result show that the ground movement is lagging behind the underground mining and filling.
The displacements of ground monitoring points can reveal the characteristics of backfill deformation to some degree. Figure 4 shows the horizontal displacement vectors of the ground monitoring points. The displacements of the monitoring points in the hanging wall are larger than the points in the foot wall. All horizontal displacement vectors pointed to a centrally located settlement. The settlement center was located on exploration line-14. This phenomenon indicated that the backfill deformation near exploration line-14 is larger than other exploration lines. Horizontal displacement vectors end in the boundary of the backfill in sub-level 1250 m and sub-level 1150 m. The stress distribution of the boundary part of the backfill is more complex.

Underground Field Monitoring of Backfill Stress Distribution and Displacement
Points of stress and displacement monitoring were laid according to backfill distribution in underground. Monitoring instruments were set in the sublevel 1100 m. The location of the monitoring points are shown in Figure 5. .

Bed Separated Displacement Monitoring of the Backfill
One of critical controlling factors of backfill stability is bed separation displacement. In the study area, mining has proceeded from top to bottom in a series of horizontal four-meter-high drift slices. Once excavation is finished within an access drift, the void is filled using cement paste before the next access drift is excavated. To guarantee output, mining occurs at two sub-levels simultaneously. Backfill in different levels are formed at different times. The bed separation displacement occurred because of the volume shrinkage of the backfill. The bed separation displacements were monitored in sub-level 1100 m, using displacement meters in different depths of backfill.
The monitoring meters layout in sub-level 1100 m along four exploration lines, respectively. The monitoring meters were set in exploration line-12, line-14, line-16, and line-17. Bed separation displacement monitoring work was built together in groups of four meters that extend to the

Underground Field Monitoring of Backfill Stress Distribution and Displacement
Points of stress and displacement monitoring were laid according to backfill distribution in underground. Monitoring instruments were set in the sublevel 1100 m. The location of the monitoring points are shown in Figure 5.

Underground Field Monitoring of Backfill Stress Distribution and Displacement
Points of stress and displacement monitoring were laid according to backfill distribution in underground. Monitoring instruments were set in the sublevel 1100 m. The location of the monitoring points are shown in Figure 5. .

Bed Separated Displacement Monitoring of the Backfill
One of critical controlling factors of backfill stability is bed separation displacement. In the study area, mining has proceeded from top to bottom in a series of horizontal four-meter-high drift slices. Once excavation is finished within an access drift, the void is filled using cement paste before the next access drift is excavated. To guarantee output, mining occurs at two sub-levels simultaneously. Backfill in different levels are formed at different times. The bed separation displacement occurred because of the volume shrinkage of the backfill. The bed separation displacements were monitored in sub-level 1100 m, using displacement meters in different depths of backfill.
The monitoring meters layout in sub-level 1100 m along four exploration lines, respectively. The monitoring meters were set in exploration line-12, line-14, line-16, and line-17. Bed separation displacement monitoring work was built together in groups of four meters that extend to the

Bed Separated Displacement Monitoring of the Backfill
One of critical controlling factors of backfill stability is bed separation displacement. In the study area, mining has proceeded from top to bottom in a series of horizontal four-meter-high drift slices. Once excavation is finished within an access drift, the void is filled using cement paste before the next access drift is excavated. To guarantee output, mining occurs at two sub-levels simultaneously. Backfill in different levels are formed at different times. The bed separation displacement occurred because of the volume shrinkage of the backfill. The bed separation displacements were monitored in sub-level 1100 m, using displacement meters in different depths of backfill.
The monitoring meters layout in sub-level 1100 m along four exploration lines, respectively. The monitoring meters were set in exploration line-12, line-14, line-16, and line-17.
Bed separation displacement monitoring work was built together in groups of four meters that extend to the following four depths: 20 m, 25 m, 30 m, and 35 m. The duration of monitoring was three months, and the results of the monitoring are showed in Figure 6.  The bed separated displacement monitoring results demonstrated contraction deformation of different parts of the backfill in sub-level 1100 m. All the displacement values were less than 20 mm. Separated displacement of the backfill located exploratory line-12 occurred at the initial stage. However, this displacement tended to reduce as the mining went deeper. The separated deformation was intended to be stable in this part. The trend of bed separated displacement of the backfill located in exploratory line-14 and line-16 was consistent. At the initial stage, the upper displacement (monitored by 20 m deep meter), the middle displacement (monitored by 25 m and 30 m deep meters), and lower displacement (monitored by 35 m deep meter) have the same change. After two months, the separated deformation appeared upper displacement at the upper part of backfill, and the downward displacement appeared in the lower part of the backfill. The bed separated deformation was accelerated at these parts of backfill. Separated displacement of the backfill located at exploratory line-17 was not obvious.
One of factors influencing bed separated displacement is the effect of mining at two levels simultaneously. The monitoring results indicated that separated displacement of the backfill part in exploratory line 14 was larger than in other parts. This characteristic was consistent with the ground movement. The backfill is a passive support measurement. The exterior load acted on the surrounding rock and then transferred to backfill through the rock material.
The vertical stress of the internal backfill is much less than that of the boundary part of the backfill. The point with the minimum value of stress is located on exploratory line-14. This part of the backfill is the center part of backfill along its strike.
The stress of the internal parts of the backfill was less than that in the boundary parts of the backfill. The backfill located on exploratory line 14 is the internal part, where the stress was only 1. 16 MPa. This part of the backfill is the center of the backfill along its strike. Corresponding to the ground movement, the settlement center is also located in exploratory line-14, and several main fissures were concentrated near the settle center in exploratory line-14. It should be noted that the bed separation displacement of the backfill in line-14 increased during the monitoring duration. Combined with the corresponding ground subsidence, the internal part of the backfill located on exploratory line-14 released its accumulated stress. This situation led to the stress of backfill in exploratory line-14 is much lower than other parts of the backfill. The backfill is a passive support measurement. The exterior load acted on the surrounding rock and then transferred to backfill through the rock material.

Numerical Simulation
The vertical stress of the internal backfill is much less than that of the boundary part of the backfill. The point with the minimum value of stress is located on exploratory line-14. This part of the backfill is the center part of backfill along its strike.
The stress of the internal parts of the backfill was less than that in the boundary parts of the backfill. The backfill located on exploratory line 14 is the internal part, where the stress was only 1.16 MPa. This part of the backfill is the center of the backfill along its strike. Corresponding to the ground movement, the settlement center is also located in exploratory line-14, and several main fissures were concentrated near the settle center in exploratory line-14. It should be noted that the bed separation displacement of the backfill in line-14 increased during the monitoring duration. Combined with the corresponding ground subsidence, the internal part of the backfill located on exploratory line-14 released its accumulated stress. This situation led to the stress of backfill in exploratory line-14 is much lower than other parts of the backfill.

Numerical Simulation
According to the monitoring data analysis, the stress value of the internal backfill is small for both horizontal and vertical stress, while the boundary part of the backfill distributed larger stress than the internal part. Contact between the boundary part of the backfill and the surrounding rock resulted in a complex stress situation. In the contact area between part of surrounding rock and the backfill (the boundary part of the backfill), the difference of stiffness and strength between the surrounding rock mass and the backfill led to different stress transfers and different types of deformation. According to our investigation, the backfill boundary in the study area is irregular. The geometric feature of the backfill boundary is a step-profile. This irregular boundary is beneficial for backfill stability, but this boundary resulted in an even more complex stress distribution and deformation in the contact area of the backfill and the surrounding rock. In this article, the discrete element method was carried out to simulate the backfill and the surrounding rock stress distribution and deformation in the contact area using software of Particle Flow Code-2D (PFC 2D , ITASCA consulting Group, Minneapolis, MN, USA) software. A numerical model was built up to analyze the characteristics of stress distribution and deformation influenced by the geometric features of the boundary part of the backfill. Geostress measurement showed that horizontal tectonic stress is larger than vertical stress. The horizontal tectonic stress is therefore the major factor. When the mining depth increases, the horizontal and vertical stresses both increase. In the study area, the horizontal stress coefficient-the ratio of horizontal stress and vertical stress-is 1.2~2. According to the practical situation in the study area, the ratio of the horizontal load and vertical load of numerical model is setting two. The sketch of numerical model is shown in Figure 8. According to the monitoring data analysis, the stress value of the internal backfill is small for both horizontal and vertical stress, while the boundary part of the backfill distributed larger stress than the internal part. Contact between the boundary part of the backfill and the surrounding rock resulted in a complex stress situation. In the contact area between part of surrounding rock and the backfill (the boundary part of the backfill), the difference of stiffness and strength between the surrounding rock mass and the backfill led to different stress transfers and different types of deformation. According to our investigation, the backfill boundary in the study area is irregular. The geometric feature of the backfill boundary is a step-profile. This irregular boundary is beneficial for backfill stability, but this boundary resulted in an even more complex stress distribution and deformation in the contact area of the backfill and the surrounding rock. In this article, the discrete element method was carried out to simulate the backfill and the surrounding rock stress distribution and deformation in the contact area using software of Particle Flow Code-2D (PFC 2D , ITASCA consulting Group, Minneapolis, MN, USA) software. A numerical model was built up to analyze the characteristics of stress distribution and deformation influenced by the geometric features of the boundary part of the backfill.

Introduction to the Numerical Model
For the issue of backfill stability in underground mining, Bloss [42] analyzed deep mining engineering and concluded that one of the main factors of backfill stability is the shear strength of the contact area of the backfill and the surrounding rock. The numerical model simulated the stress distribution of the surrounding rock and the backfill at the point of joint load bearing.
The boundary is a step profile, and the overall size of model is 16 m × 16 m. The single step height is 5.3 m, which is designed according to the mining layer height is 4~5.5 m in practical mining engineering. Particles in the surrounding rock mass have a diameter of 0.5 mm. The backfill part blended two sizes of particles with diameters of 0.6 mm and 0.8 mm as a fine aggregate and a coarse aggregate.
The parameters of the surrounding rock mass and the backfill in the numerical model are as follows: the elastic modulus of the surrounding rock mass is 2 × 10 3 MPa, and the elastic modulus of the backfill is 0.3 × 10 3 MPa. The normal and shear stiffnesses of the surrounding rock mass are 2 × 10 9 and 0.5 × 10 8 , respectively.
Geostress measurement showed that horizontal tectonic stress is larger than vertical stress. The horizontal tectonic stress is therefore the major factor. When the mining depth increases, the horizontal and vertical stresses both increase. In the study area, the horizontal stress coefficient-the ratio of horizontal stress and vertical stress-is 1.2~2. According to the practical situation in the study area, the ratio of the horizontal load and vertical load of numerical model is setting two. The sketch of numerical model is shown in Figure 8.

Stress Transferring in the Boundary Part of the Backfill
Vertical load was set on the top of model, and horizontal load was only acted on the surrounding rock. The horizontal load was transferred to backfill by surrounding rock mass deformation. This loading mode is the feature of backfill as a passive support measurement.
External load is acted on the model, and transferred internal part by surrounding rock mass and backfill material. In numerical model, monitoring points were set along the boundary of backfill to measure stress and displacement. Location of points was shown in Figure 9a. For each point on the boundary part (point 1~point 11), ratio of horizontal stress and vertical stress was shown in Figure 9b. According to geometric features and ratio of horizontal stress and vertical stress of boundary part of backfill, Figure 9, monitoring points could be divided into four types: type one included point 2, point 6 and point 10; type two included point 3 and pint 7; type three included point 4 and point 8; type four included point 5, point 9 and point 11. Vertical load was set on the top of model, and horizontal load was only acted on the surrounding rock. The horizontal load was transferred to backfill by surrounding rock mass deformation. This loading mode is the feature of backfill as a passive support measurement.
External load is acted on the model, and transferred internal part by surrounding rock mass and backfill material. In numerical model, monitoring points were set along the boundary of backfill to measure stress and displacement. Location of points was shown in Figure 9a. For each point on the boundary part (point 1~point 11), ratio of horizontal stress and vertical stress was shown in Figure  9b. According to geometric features and ratio of horizontal stress and vertical stress of boundary part of backfill, Figure 9, monitoring points could be divided into four types: type one included point 2, point 6 and point 10; type two included point 3 and pint 7; type three included point 4 and point 8; type four included point 5, point 9 and point 11.  Figure 10 showed the horizontal stress development of surrounding rock and backfill boundary points and these points located at the same level. At initial stage of stress curve, stress of surrounding rock was larger than the stress of the backfill boundary points. Rate of stress increase of surrounding rock remained stable and showed linear tendency. While stress backfill boundary increased slowly at first, then rapidly. This trend was remained until backfill boundary stress was lager than surrounding rock at one time. At this time the crack started initiation compared with crack numbers development ( Figure 10). Displacement monitoring of the model set was done on the adjacent rock point and backfill point on the boundary. After the post-peak stress stage, the displacements of the surrounding rock and the backfill both increased sharply. The increasing failure of the model was more and more obvious as  Figure 10 showed the horizontal stress development of surrounding rock and backfill boundary points and these points located at the same level. At initial stage of stress curve, stress of surrounding rock was larger than the stress of the backfill boundary points. Rate of stress increase of surrounding rock remained stable and showed linear tendency. While stress backfill boundary increased slowly at first, then rapidly. This trend was remained until backfill boundary stress was lager than surrounding rock at one time. At this time the crack started initiation compared with crack numbers development ( Figure 10). Vertical load was set on the top of model, and horizontal load was only acted on the surrounding rock. The horizontal load was transferred to backfill by surrounding rock mass deformation. This loading mode is the feature of backfill as a passive support measurement.
External load is acted on the model, and transferred internal part by surrounding rock mass and backfill material. In numerical model, monitoring points were set along the boundary of backfill to measure stress and displacement. Location of points was shown in Figure 9a. For each point on the boundary part (point 1~point 11), ratio of horizontal stress and vertical stress was shown in Figure  9b. According to geometric features and ratio of horizontal stress and vertical stress of boundary part of backfill, Figure 9, monitoring points could be divided into four types: type one included point 2, point 6 and point 10; type two included point 3 and pint 7; type three included point 4 and point 8; type four included point 5, point 9 and point 11.  Figure 10 showed the horizontal stress development of surrounding rock and backfill boundary points and these points located at the same level. At initial stage of stress curve, stress of surrounding rock was larger than the stress of the backfill boundary points. Rate of stress increase of surrounding rock remained stable and showed linear tendency. While stress backfill boundary increased slowly at first, then rapidly. This trend was remained until backfill boundary stress was lager than surrounding rock at one time. At this time the crack started initiation compared with crack numbers development ( Figure 10).  Displacement monitoring of the model set was done on the adjacent rock point and backfill point on the boundary. After the post-peak stress stage, the displacements of the surrounding rock and the backfill both increased sharply. The increasing failure of the model was more and more obvious as cracks developed. The displacement difference of the surrounding rock and the backfill at the post peak stage occurred suddenly. Uncoordinated deformation emerged on the boundary part of the backfill, which led to cracks and failure.

Stress Particularity of the Boundary Part of the Backfill
Some researchers have put forward a view that shear stress and the damage caused by shear stress are major factors for the local stability of the boundary part of the backfill. We set up another six monitoring points in the internal surrounding rock mass and the internal backfill, respectively. Combined with upper monitoring points in the boundary parts of the backfill, the results of the shear stress of each point are presented in Figures 11-13.
The stress monitoring was carried out in the internal part of the surrounding rock mass, the internal part of the backfill, and the boundary part of the backfill. The stress curves of internal part of the surrounding rock mass are shown in Figure 11a-c. In the upper part of the surrounding rock, the stress increased slowly in initial stage, and the stress increase rate was larger in the later stage. The stress had a sudden rise to its peak value and then collapsed. The rate of stress change remained almost stable. The stress curves of internal part of backfill are shown in Figure 12a-c. In the initial stage, the stress of the backfill increased stably. Micro-cracks were initiated, and their development lead to the unstable changing of the stress of the backfill, which then reached its peakvalue. Compared with the stress change of the surrounding rock, the phenomenon of the sudden rise of stress to its peak value was alleviated. The stress distribution of the backfill is consistent with results of the underground measurements. At the same depth, the stress value of the surrounding rock is larger than the stress of the backfill. After micro-crack initiation, the ratio of stress between the surrounding rock mass and the backfill was 1.3~2. This stress ratio increased as the depth increased. Therefore, the mining depth increase led to the larger difference between the surrounding rock stress and backfill stress.

Stress Particularity of the Boundary Part of the Backfill
Some researchers have put forward a view that shear stress and the damage caused by shear stress are major factors for the local stability of the boundary part of the backfill. We set up another six monitoring points in the internal surrounding rock mass and the internal backfill, respectively.

Stress Particularity of the Boundary Part of the Backfill
Some researchers have put forward a view that shear stress and the damage caused by shear stress are major factors for the local stability of the boundary part of the backfill. We set up another six monitoring points in the internal surrounding rock mass and the internal backfill, respectively.     The stress monitoring was carried out in the internal part of the surrounding rock mass, the internal part of the backfill, and the boundary part of the backfill. The stress curves of internal part of the surrounding rock mass are shown in Figure 11a-c. In the upper part of the surrounding rock, the stress increased slowly in initial stage, and the stress increase rate was larger in the later stage. The stress had a sudden rise to its peak value and then collapsed. The rate of stress change remained almost stable. The stress curves of internal part of backfill are shown in Figure 12a-c. In the initial stage, the stress of the backfill increased stably. Micro-cracks were initiated, and their development lead to the unstable changing of the stress of the backfill, which then reached its peakvalue. Compared with the stress change of the surrounding rock, the phenomenon of the sudden rise of stress to its peak value was alleviated. The stress distribution of the backfill is consistent with results of the underground measurements. At the same depth, the stress value of the surrounding rock is larger than the stress of the backfill. After micro-crack initiation, the ratio of stress between the surrounding rock mass and the backfill was 1.3~2. This stress ratio increased as the depth increased. Therefore, the mining depth increase led to the larger difference between the surrounding rock stress and backfill stress.
The stress distribution of the boundary part of the backfill is more complex due to the interaction between the backfill and the surrounding rock. Along the step-profile of the boundary, the stress The stress distribution of the boundary part of the backfill is more complex due to the interaction between the backfill and the surrounding rock. Along the step-profile of the boundary, the stress monitoring points were set, as shown in Figure 9 (point-2~point-11). The stress development is shown in Figure 13a-j. The stress value of the boundary part was smaller than the stress value in the internal surrounding rock and lager than the stress value in the internal backfill. The major feature of some boundary stress developments is multi-peak strength-normally, they have two peak values. These two peak stress values divided the stress curve into the following five stages: Stage 1: stress increased linearly and slowly; Stage 2: stress accelerated growth; Stage 3: post first peak stress stage, the stress decreased slightly then increased; Stage 4: stress reached the second peak value; and Stage 5: stress decreased sharply in post second peak stage.
The step-profile boundary had five inflection nodes located at monitoring point-3, point-5, point-7, point-9, and point-11, respectively. The boundary points have been divided into four types. As previously mentioned, multi-peak stress is a major feature of boundary points stress distribution. The multi-peak stress was influenced by the depth and inflection location, and as depth increased, the multi-peak stress became more obvious.

Conclusions
This article aimed to study backfill stress distribution and deformation and analyzed its stability. The conclusions of study are as follows: (1) The settlement of ground subsidence corresponded to the center exploratory line of the backfill.
This part of the backfill has the largest bed separated displacement. (2) Field stress monitoring of the underground backfill showed that both horizontal and vertical stresses of the internal parts of the backfill are smaller than those of the boundary part of the backfill. The stress distribution of the boundary part of the backfill was more complex.