Environmental Earth Sciences (2023) 82:110

https://doi.org/10.1007/s12665-023-10801-w

ORIGINAL ARTICLE

Evaluation of different methods of pillar recovery in coal mining

by numerical simulation: a case study
Nima Babanouri1 · Helma Beyromvand1 · Hesam Dehghani1

Received: 17 August 2021 / Accepted: 5 February 2023 / Published online: 16 February 2023
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023

Abstract
The room-and-pillar technique is a self-supporting mining method in which part of the ore is left unmined as pillars. To
increase the mining efficiency, the pillars are later recovered partially or completely. This research aims at evaluating and
comparing four methods of pillar recovery, namely, ‘pocket and wing’, ‘open ending’, ‘split and fender’, and ‘Christmas
tree’ which are of particular importance. The performance of these methods of pillar recovery in the Tabas coal mine was
evaluated by numerical modeling. To do this, a series of three-dimensional simulations were carried out in which the pil-
lars were excavated according to the sequences specified by each method of recovery. The behavior of chain pillars, barrier
pillars, and the roof was monitored and compared for the different cutting sequences. Using the shear strength reduction
technique, the safety factor values were calculated during the different recovery methods until the verge of roof caving.
The results revealed that the method of split and fender leads to a softer drop of the safety factor, and consequently, more
controlled caving of the roof.

Keywords Room and pillar · Pillar recovery · Coal mine · Numerical modeling · Roof caving

Introduction 2019; Zingano and Weiss 2019; Feng and Wang 2020; Luo
et al. 2020; Zhou et al. 2020; Satyanarayana et al. 2021).
The major share of the world’s production of coal pertains Shaojie et al. designed strip coal pillars considering the sub-
to underground mines which are mostly mined using three sidence estimate, so that the coal pillars remained stable
methods of longwall, short wall, and room and pillar. Hence, in practice, and buildings in the villages above the mine
the room-and-pillar method is of particular importance in were not adversely influenced (Shaojie et al. 2016). Ghasemi
coal mining. In the room-and-pillar method, 30–60% of the and Shahriar presented a new method for designing the coal
coal is left unmined as the pillars whose proper recovery has pillars using the tributary area theory and abutment angle
a significant role in increasing mining productivity. concept to improve the safety of room-and-pillar mining
Various features of coal pillars have received a great (Ghasemi and Shahriar 2012). Ghasemi et al. studied the
deal of attention from different researchers (Ghasemi and risk of roof collapse during the pillar recovery in coal mines
Shahriar 2012; Fakhimi et al. 2016; Shaojie et al. 2016; Li and developed semi-quantitative techniques combining engi-
et al. 2017; Zhang et al. 2017, 2018; Zhu and Tu 2017; Chen neering judgment and examination of related literature to
et al. 2018; Gao 2018; Luo and Yang 2018; Sun et al. 2018; estimate the roof fall risk (Ghasemi et al. 2012). Esterhuizen
Zhang and Ni 2018; Jessu and Spearing 2019; Li and Chai et al. developed a relationship between the strength of pil-
lars in underground mines and parameters of the uniaxial
* Nima Babanouri compressive strength of the intact rock, major discontinuity
babanouri@hut.ac.ir properties, and pillar geometry using numerical modeling
Helma Beyromvand and site investigations (Esterhuizen et al. 2011). Kim et al.
h.beyromvand@yahoo.com provided a risk map for subsidence induced by room-and-
Hesam Dehghani pillar mining using the artificial neural network combined
dehghani@hut.ac.ir with the geographic information system (GIS) (Kim et al.
2009). Yang et al. investigated the effect of roof structure
1
Department of Mining Engineering, Hamedan University on coal pillar load-bearing capacity and stress distribution
of Technology, Hamedan, Iran

13
Vol.:(0123456789)
110 Page 2 of 11 Environmental Earth Sciences (2023) 82:110

(Yang et al. 2015). Al Heib et al. back-analyzed a histori- ease of implementation, and simple ventilation and loading
cal collapse of an abandoned room-and-pillar mine using are also among its important features. This method is not
both an empirical approach and a numerical approach and suitable in conditions, where the roof or the pillar is prone
attributed the collapse to the sudden rupture of a hard lime- to swell (Kauffman et al. 1981).
stone bed (Al Heib et al. 2015). Fakhimi et al. studied the The open ending is somewhat similar to the pocket-and-
strain burst of mine pillars through physical and numerical wing process Fig. 1b schematically indicates the procedure
modeling and investigated the influence of different param- which is characterized by taking lifts in a sequence of inci-
eters, such as the geometry of pillars, compressive strength sions on one side of a pillar. One of the requirements for
of pillars, and pillar and roof rock densities (Fakhimi et al. using this method is that the roof must be reliable over the
2016). Luo and Yang investigated the influence of dewater- pillars and collapse properly. Usually, during pillar recovery
ing flooded abandoned room-and-pillar mines on surface in this way, piles are installed alone or together with trans-
subsidence (Luo and Yang 2018). Youfeng and Pengpeng verse wooden beams for temporary support of the immediate
studied the influence of the room-pillar layout on the stabil- roof (Kauffman et al. 1981).
ity of an underground mine by numerical modeling (Zhang In the USA, the method of split and fender is the most
and Ni 2018). Waclawik et al. studied the behavior of coal widely applied technique for pillar recovery. The fundamen-
pillars in great depth (Waclawik et al. 2018). Zhu and Tu tal idea of the procedure is to extract a series of incisions
investigated the support failure induced by repeated mining through the pillar, usually in the direction of the long side of
under gobs created by two-seam room mining (Zhu and Tu the pillar. This technique produces two coal fenders by split-
2017). ting the pillar. To provide an appropriate number of working
Although the different aspects of the room-and-pillar places, thereby preventing delays, two pillars are usually
method have been extensively studied, no systematic study concurrently mined (Fig. 1c). To establish proper safety dur-
has been carried out on different methods of pillar recov- ing the operations, the top of the splits created in the pillar
ery. This research aims at evaluating and comparing four may be bolted before the extraction of the fenders. Apart
methods of pillar recovery, namely, ‘pocket and wing’, ‘open from difficult ventilation, this method is not suitable for large
ending’, ‘split and fender’, and ‘Christmas tree’ which are of pillars and brittle roof conditions (Kauffman et al. 1981).
particular importance. Given the coal mines getting deeper The method of Christmas tree entails mining double
and the increasing need for coal exploitation, it is necessary pillars from one entry. As shown in Fig. 1d, incisions are
to examine different methods of pillar extraction. To do this, created to the left and then to the right. This method is gen-
three-dimensional numerical models of pillar recovery in a erally used when the overburden is high and large pillars
coal mine were analyzed. The behavior of chain pillars, bar- are needed to ensure stability. In some cases, temporary
rier pillars, and the roof was monitored and compared for the hydraulic support is installed in the mining area (Kauffman
different cutting sequences. et al. 1981).

Methods of pillar recovery Tabas coal mine

Most mining engineers decide to extract pillars to decrease The case studied in this research is the main panel of the
the loss of coal in room-and-pillar mining. There are a few Tabas coal mine, which is supposed to be the first room-
basic methods of pillar recovery and a number of variations and-pillar coal mine in Iran (Fig. 2). The mine is exploiting
for each, among which the four techniques of ‘pocket and seam C1 having a dip of 11° and a 2 m thickness. The main
wing’, ‘open ending’, ‘split and fender’, and ‘Christmas panel has an 85 m overburden with an average unit weight
tree’ are most widely practiced in the industry. The follow- of 25 kN/m3. The roof consists of a mudstone layer with a
ing explanations are intended to provide a summary of these thickness of 0.1–0.2 m, siltstone/sandstone interfaces, and
methods. sandstone traces up to 3 m in certain places. Downwards
The method of pocket and wing is practiced under widely the mining stope, a weak layer of seatearth/mudstone with
varying conditions mostly in northern West Virginia. With an approximate thickness of 1 m is followed by stronger
the pocket-and-wing procedure, two workstations are created mudstones, siltstones/sandstones. In the future, the pillars
in one pillar. The pockets are located on the side of the pil- will be extracted using retreat mining (ADAM Consulting
lar, and the lifts are typically located between the pockets. Engineers 2005; Ghasemi et al. 2012).
A wing or fender of coal remains between the pocket and
the gob. Figure 1a depicts a schematic pattern of operation
in which the numbers indicate the order of incisions. This
method is used for the extraction of large pillars due to its

13
Environmental Earth Sciences (2023) 82:110 Page 3 of 11 110

Fig. 1  Cutting sequences in different methods of pillar recovery: a pocket and wing, b open ending, c split and fender, and d Christmas tree
(Kauffman et al. 1981)

Numerical modeling of pillar recovery absence of support systems. A continuum modeling was car-
ried out without the presence of discontinuities. However,
In this study, a part of the main panel was modeled in 3DEC the reasons for choosing 3DEC were as follows.
with nine chain pillars of 15.5 × 4.5 × 1.8 m and two barrier
pillars of 80 × 20 × 1.8 m on both sides of the stope (Fig. 3). – Discontinuities were supposed to be included in the
Figure 3b shows a cross section of the model in which pil- models in the end, so that their effect would be studied.
lars 1–4 are supposed to be recovered. Only the barrier pil- Unfortunately, this was not possible due to the large
lars adjacent to the recovery area were present, and other dimensions of the models and the high burden of numeri-
barrier pillars were not modeled due to their distance from cal calculations.
the recovered pillars. In addition, since a variety of support – On the other hand, the use of 3DEC does not necessar-
systems may be accompanied or not by a pillar recovery ily require the presence of discontinuities. In the absence
operation, the focus of this research is on the incisions and of discontinuities, the code uses a finite-difference for-
their excavation sequence in pillar recovery methods in the mulation for continuum analysis. In fact, both continuum

13
110 Page 4 of 11 Environmental Earth Sciences (2023) 82:110

Fig. 2  Study area: a location and layout of Tabas coal mine, and b stratigraphic column

of pocket and wing, open ending, split and fender, and

Christmas tree. In the methods of pocket and wing and
open ending, the first pillar is fully recovered, and then the
next pillars are excavated. However, in the methods of split
and fender, and Christmas tree, the two adjacent pillars are
recovered together.
In this research, the rock blocks were considered to have
elasto-plastic behavior according to the Mohr–Coulomb
constitutive model specified by the Young modulus (E),
Poisson ration ( 𝜐 ), cohesion (c), friction angle ( 𝜙 ), uni-
axial compressive strength (σc), tensile strength (σt), and
density (ρ) (Itasca Consulting Group Inc. 2013). Over the
years of the Tabas mine life, a large number of field inves-
tigations and laboratory tests have been performed on its
different types of rocks by the mine’s rock mechanics unit;
experiments, such as index, uniaxial compression, triaxial
compression, and Brazilian tensile tests. Table 1 reports
the outcome of these tests assigned to the different blocks
in the model. Despite the slight differences, the roof and
the floor have the same lithology as indicated in Fig. 2b.
Fig. 3  Modeled part of main panel: a model regions, and b cross-sec- In fact, mudstone is a thin and scattered layer on the roof,
tional view and the roof and the floor both are considered sandstone.
After the cuts were made and the model meshed, the
analysis and discontinuum analysis can be performed boundary and initial conditions were applied. Given the
robustly with this 3DEC (Itasca Consulting Group Inc. overburden of 85 m with the average unit weight of 25 kN/
2013). m3, a gravity pressure of 2.125 MPa was applied to the
upper part of the model. The model bottom was considered
For modeling the recovery process of the pillars, pil- to be fixed. A uniform gravitational field of stress was also
lars 1–4 were cut and mined using different methods. Fig- exerted on the zones of the model as the initial stresses.
ure 4 shows the sequences of pillar recovery by methods

13
Environmental Earth Sciences (2023) 82:110 Page 5 of 11 110

Fig. 4  Cutting sequences in different methods of pillar recovery: a pocket and wing, b open ending, c split and fender, and d Christmas tree

Table 1  Geomechanical Regions E ρ c σc 𝝓 σt 𝝊

properties of different blocks in (MPa) (kg/m3) (MPa) (MPa) (˚) (MPa) (–)
numerical model (Ajamzadeh
2017) Coal seam 316 1600 0.016 6 20 0.002 0.25
Roof and floor 5281 2700 8.69 72.8 22 6.3 0.25

After the model reached the initial equilibrium, the pil- Results and discussion
lars were excavated according to the sequences specified
by each method of recovery. Figure 5 indicates the numeri- Safety factor analysis
cal model when the first pillar is fully removed and during
the simultaneous recovery of pillars 2 and 3 by the pocket To quantify the stability state of the roof during the pillar
and wing method. recovery, a safety factor analysis was performed. Dawson

13
110 Page 6 of 11 Environmental Earth Sciences (2023) 82:110

Table 2  Calculations of NPR during pillar recovery by pocket and

wing along with SF
Stage Cut volume (­ m3) Cumulative NPR (%) SF (–)
volume of cuts
­(m3)

1 0 0 0 1.89
2 70.20 70.20 16.23 1.87
3 70.20 140.40 32.47 1.85
4 70.20 210.60 48.70 1.78
5 70.20 280.80 64.93 1.73
6 70.20 351 81.17 1.62
7 70.20 421.20 97.40 1.50
8 11.13 432.33 99.97 1.49
9 32.15 464.48 107.41 1.45
10 37.44 501.92 116.06 1.34
11 25.94 527.86 122.06 1.29
12 4.94 532.80 123.20 1.29
13 4.94 537.74 124.35 1.26
14 25.94 563.68 130.35 1.23
15 37.08 600.76 138.92 1.09
Fig. 5  Numerical model during recovery of pillars 2 and 3 by pocket 16 32.29 633.05 146.39 1.02
and wing
17 22.48 655.53 151.59 Roof caving

et al. (1999) proposed the method of shear strength reduc-

tion (SSR) to calculate the safety factor in numerical anal- is done for the other methods of pillar recovery, and the val-
yses (Dawson et al. 1999). To determine the safety factor ues of SF were plotted versus NPR values (Fig. 6).
(SF) of the model at each stage of recovery, the model is As can be seen, the higher the NPR, the lower the SF.
run for a series of trial factors of safety (Ftrial) with c and The methods of open ending and pocket and wing yielded
𝜙 of the blocks adjusted according to the equations below similar results, while the methods of Christmas tree and split
(Dawson et al. 1999): and fender resulted in different SF-NPR plots. In the method
of Christmas tree, the roof caving roughly occurred after
1
ctrial = c, (1) NPR = 125% equivalent to the recovery of one and a quarter
Ftrial
pillar, while the method of split and fender results in the roof
caving after recovering one and a half pillars (NPR = 150%).
In the methods of open ending and pocket and wing, the roof
( )
1
𝜙trial = arctan tan𝜙 , (2)
Ftrial caves roughly after NPR of 107%, and 100%, respectively.

Ftrial is then varied using the bracketing approach until the

model starts to fail. In this case, the value of Ftrial is consid-
ered as the SF of the model. More details on the procedure 2.00
for implementing the strength reduction technique in 3DEC 1.90 Pocket and wing
Open ending
can be found in the 3DEC user’s guide (Itasca Consulting 1.80 Split and fender
1.70 Christmas tree
Group Inc. 2013).
1.60
The ratio of the cumulative volume of the parts already
1.50
SF

excavated to the volume of one pillar is defined as the nor- 1.40

malized volume of pillar recovery (NPR). Table 2 shows 1.30
the calculations of NPR at each stage of pillar recovery by 1.20
the method of pocket and wing. This normalized measure 1.10
was used to compare the model behavior during the dif- 1.00
0 25 50 75 100 125 150
ferent methods of pillar recovery with different schemes
NPR (%)
of excavation.
Table 2 shows the results of the SSR analysis for the
Fig. 6  Values of SF versus NPR values during different methods of
method of pocket and wing before the roof caving. The same pillar recovery

13
Environmental Earth Sciences (2023) 82:110 Page 7 of 11 110

The drop of SF in the methods of Christmas tree and has happened to them in the history of the numerical simu-
split and fender is gentler in comparison with the other two lation. As can be seen, the pattern and extent of yielded
methods. Consequently, the methods of Christmas tree and zones around the pillar recovery are different for the dif-
split and fender result in more controlled caving of the roof, ferent methods. In the methods of open ending and pocket
while the methods of open ending and pocket and wing lead and wing, the barrier pillars are yielded at the beginning
to the dramatic drop of SF, and hence, inappropriate caving of recovery (NPR = 50%), whereas the barrier pillars in the
of the roof. methods of Christmas tree and split and fender begin to yield
at the later stage of recovery (NPR = 75%).
The state of neighboring areas
Maximum shear strain rate in pillars and roof
Figures 7 and 8 show the state of the model at the top of pil-
lars and in the vicinity of the roof for the different methods Figures 9 and 10 show the color-scale plot of the maxi-
of recovery when the NPR is about 50% and 75%, respec- mum shear strain rate at the top of pillars and in the vicin-
tively. In all the models, the shear yielding occurred and ity of the roof for the different methods of recovery when
no tension yielding was observed. It is worth mentioning the NPR is about 50% and 75%, respectively. As can be
that the up-to-date state of the zones is displayed not what seen, at 50% of the recovery process the methods of open

Fig. 7  Model state when NPR

is about 50%: a pocket and
wing, b open ending, c split and
fender, and d Christmas tree
a) b) c) d)

(NPR=53.9%) (NPR=51.47 %) (NPR= 48.70 %) (NPR=48.02 %)

Fig. 8  Model state when NPR

is about 75%: a pocket and
wing, b open ending, c split and
fender, and d Christmas tree
a) b) c) d)

(NPR=75.73 %) (NPR=75.75 %) (NPR=64.93 %) (NPR=71.22 %)

a) b) c) d)

(NPR=53.9%) (NPR=51.47 %) (NPR= 48.70 %) (NPR=48.02 %)

Fig. 9  Color-scale plot of maximum shear strain rate in pillars when NPR is about 50%: a pocket and wing, b open ending, c split and fender,
and d Christmas tree

13
110 Page 8 of 11 Environmental Earth Sciences (2023) 82:110

a) b) c) d)

(NPR=75.73 %) (NPR=75.75 %) (NPR=64.93 %) (NPR=71.22 %)

Fig. 10  Color-scale plot of maximum shear strain rate in pillars when NPR is about 75%: a pocket and wing, b open ending, c split and fender,
and d Christmas tree

ending and pocket and wing are of high strain rate, while Roof settlement velocity
the method of Christmas tree experiences a lower strain
rate. However, at 75% of the recovery process, the disper- The settlement velocity is an important indicator that shows
sion and extent of the areas with high strain rates decrease the pattern of development of the roof caving in the differ-
somewhat in the methods of open ending and pocket and ent methods of pillar extraction. Figures 13 and 14 show
wing, because the roof has already undergone large defor- the contour plot of settlement velocity in the roof for the
mation. On the other hand, the strain rate increases in the different methods of recovery when the NPR is about 50%
methods of Christmas tree and split and fender since the and 75%, respectively.
roof caves slowly. In other words, when the roof caving A comparison between Figs. 13 and 14 shows how the
develops, the shear strain concentration in the pillars is formation of roof caving proceeds exponentially once the
relieved. Since the roof caving in the methods of open NPR exceeds 50%. These charts can be used to determine
ending and pocket and wing begins earlier compared to the the level of safety alerts at each stage of pillar recovery, sug-
methods of Christmas tree and split and fender (consider gesting that most precautions need to be taken immediately
Fig. 6), at NPR = 75% the first two methods have smaller after the NPR reaches 50%.
areas with the highest shear strain rate.
Figures 11 and 12 show the color-scale plot of the maxi- Excavation sequences
mum shear strain rate in the roof for the different methods
of recovery when the NPR is about 50% and 75%, respec- The excavation sequences in the methods of open end-
tively. These figures reinforce the above conclusions. ing and pocket and wing are less geometrically dispersed
within the recovery area compared to the two other meth-
ods. Hence, these methods of pillar recovery cause abrupt

a) b) c) d)

(NPR=53.9%) (NPR=51.47 %) (NPR= 48.70 %) (NPR=48.02 %)

Fig. 11  Color-scale plot of maximum shear strain rate in roof when NPR is about 50%: a pocket and wing, b open ending, c split and fender, and
d Christmas tree

13
Environmental Earth Sciences (2023) 82:110 Page 9 of 11 110

a) b) c) d)

(NPR=75.73 %) (NPR=75.75 %) (NPR=64.93 %) (NPR=71.22 %)

Fig. 12  Color-scale plot of maximum shear strain rate in pillars when NPR is about 75%: a pocket and wing, b open ending, c split and fender,
and d Christmas tree

a) b) c) d)

(NPR=53.9%) (NPR=51.47 %) (NPR= 48.70 %) (NPR=48.02 %)

Fig. 13  Contour plot of settlement velocity in roof when the NPR is about 50%: a pocket and wing, b open ending, c split and fender, and d
Christmas tree

a) b) c) d)

(NPR=75.73 %) (NPR=75.75 %) (NPR=64.93 %) (NPR=71.22 %)

Fig. 14  Contour plot of settlement velocity in roof when the NPR is about 75%: a pocket and wing, b open ending, c split and fender, and d
Christmas tree

and uncontrolled caving of the roof. On the contrary, the stress distribution over the recovery area and postpones
methods of Christmas tree and split and fender are of the roof caving. Hence, these methods lead to the gradual
higher dispersion of cuts which results in more uniform and controlled caving of the roof.

13
110 Page 10 of 11 Environmental Earth Sciences (2023) 82:110

Conclusions Esterhuizen GS, Dolinar DR, Ellenberger JL (2011) Pillar strength in

underground stone mines in the United States. Int J Rock Mech
Min Sci 48:42–50
In this research, four commonly used methods for a full Fakhimi A, Hosseini O, Theodore R (2016) Physical and numerical
recovery of mining pillars in the Tabas room-and-pillar coal study of strain burst of mine pillars. Comput Geotech 74:36–44
mine were numerically simulated. The performance of the Feng G, Wang P (2020) Simulation of recovery of upper remnant coal
pillar while mining the ultra-close lower panel using longwall top
methods was evaluated and compared from the geomechani- coal caving. Int J Min Sci Technol 30:55–61. https://​doi.​org/​10.​
cal perspectives. The findings of this research are summa- 1016/j.​ijmst.​2019.​12.​017
rized as follows. Gao W (2018) Influence of interaction between coal and rock on the
Using the SSR technique, the SF values were calculated stability of strip coal pillar. Geomechanics and Engineering
16:151–157
during the different recovery methods until the verge of roof Ghasemi E, Shahriar K (2012) A new coal pillars design method in
caving. The variation of SF versus the normalized volume order to enhance safety of the retreat mining in room and pillar
of pillar recovery revealed that among the studied methods mines. Saf Sci 50:579–585
the method of split and fender resulted in a gentler drop of Ghasemi E, Ataei M, Shahriar K et al (2012) Assessment of roof fall
risk during retreat mining in room and pillar coal mines. Int J
SF, and consequently, more controlled caving of the roof. Rock Mech Min Sci 54:80–89
Investigating the state of the zones in the vicinity of the Itasca Consulting Group Inc. (2013) 3DEC user’s guide, Minneapolis.
recovery area demonstrated that the two methods of Christ- Minneapolis
mas tree and split and fender led to yield later than the other Jessu KV, Spearing AJS (2019) Performance of inclined pillars with
a major discontinuity. Int J Min Sci Technol 29:437–443. https://​
two methods. Hence, they are safer methods to extract the doi.​org/​10.​1016/j.​ijmst.​2018.​09.​006
pillars. Kauffman PW, Hawkins SA, Thompson RR (1981) Room and pillar
The extent of the areas with the highest shear strain rates retreat mining. a manual for the coal industry.
in the methods of Christmas tree and split and fender was Kim K-D, Lee S, Oh H-J (2009) Prediction of ground subsidence in
Samcheok City, Korea using artificial neural networks and GIS.
smaller at the beginning of the pillar recovery, suggesting Environ Geol 58:61–70
them as more appropriate techniques. Li X, Chai Y (2019) Determination of pillar width to improve min-
The plots of the roof settlement velocity obtained by the ing safety in a deep burst-prone coal mine. Saf Sci 113:244–256.
numerical modeling showed the pattern of development of https://​doi.​org/​10.​1016/j.​ssci.​2018.​12.​003
Li H, Guo G, Zha J et al (2017) Stability evaluation method for hyper-
the roof caving which is of great importance in predicting bolic coal pillars under the coupling effects of high temperature
safety measures during the pillar recovery. and ground stress. Environ Earth Sci 76:704. https://​doi.​org/​10.​
1007/​s12665-​017-​7048-0
Luo Y, Yang J (2018) Effects of dewatering flooded abandoned room-
Funding No funding was received for conducting this study. and-pillar mines on surface subsidence. Min Eng 70:45–50
Luo R, Li G, Chen L et al (2020) Ground subsidence induced by pillar
Availability of data and materials Some or all data, models, or codes deterioration in abandoned mine districts. J Central South Univ
that support the findings of this study are available from the corre- 27:2160–2172. https://​doi.​org/​10.​1007/​s11771-​020-​4438-3
sponding author upon reasonable request. Satyanarayana I, Budi G, Murmu S (2021) Stability analysis of a
deep highwall slope using numerical modelling and statistical
Declarations approach—a case study. Arab J Geosci 14:179. https://​doi.​org/​
10.​1007/​s12517-​021-​06476-x
Conflict of interest The authors have no conflicts of interest to declare. Shaojie C, Hailong W, Huaiyuan W et al (2016) Strip coal pillar design
based on estimated surface subsidence in Eastern China. Rock
Mech Rock Eng 49:3829–3838
Sun W, Zhang Q, Luan Y, Zhang X-P (2018) A study of surface subsid-
References ence and coal pillar safety for strip mining in a deep mine. Environ
Earth Sci 77:627. https://​doi.​org/​10.​1007/​s12665-​018-​7810-y
ADAM Consulting Engineers (2005) Central mine design report. Waclawik P, Kukutsch R, Nemcik J (2018) Assessment of coal pillar
Ajamzadeh M (2017) Selecting the suitable coal extraction machine stability at great depth. In: 18th Coal Operators’ Conference. pp
for Parvadeh Tabas mechanized mine using Particle Flow Code. 184–194
Hamedan University of Technology Yang JX, Liu CY, Yu B, Wu FF (2015) The effect of a multi-gob, pier-
Al Heib M, Duval C, Theoleyre F et al (2015) Analysis of the histori- type roof structure on coal pillar load-bearing capacity and stress
cal collapse of an abandoned underground chalk mine in 1961 distribution. Bull Eng Geol Env 74:1267–1273. https://​doi.​org/​
in Clamart (Paris, France). Bull Eng Geol Env 74:1001–1018. 10.​1007/​s10064-​014-​0685-6
https://​doi.​org/​10.​1007/​s10064-​014-​0677-6 Zhang Y, Ni P (2018) Design optimization of room and pillar mines:
Chen S, Qu X, Yin D et al (2018) Investigation lateral deformation and a case study of the Xianglushan tungsten mine. Q J Eng Geol
failure characteristics of strip coal pillar in deep mining. Geomech Hydrogeol 51:352–364
Eng 14:421–428 Zhang J, Huang P, Zhang Q et al (2017) Stability and control of room
Dawson EM, Roth WH, Drescher A (1999) Slope stability analysis by mining coal pillars—taking room mining coal pillars of solid
strength reduction. Geotechnique 49:835–840 backfill recovery as an example. J Central South Univ 24:1121–
1132. https://​doi.​org/​10.​1007/​s11771-​017-​3515-8
Zhang P, Tulu B, Sears M, Trackemas J (2018) Geotechnical consid-
erations for concurrent pillar recovery in close-distance multiple

13
Environmental Earth Sciences (2023) 82:110 Page 11 of 11 110

seams. Int J Min Sci Technol 28:21–27. https://​doi.​org/​10.​1016/j.​ Publisher's Note Springer Nature remains neutral with regard to
ijmst.​2017.​12.​012 jurisdictional claims in published maps and institutional affiliations.
Zhou Z, Wang H, Cai X et al (2020) Bearing characteristics and fatigue
damage mechanism of multi-pillar system subjected to different Springer Nature or its licensor (e.g. a society or other partner) holds
cyclic loads. J Central South Univ 27:542–553. https://d​ oi.o​ rg/1​ 0.​ exclusive rights to this article under a publishing agreement with the
1007/​s11771-​020-​4315-0 author(s) or other rightsholder(s); author self-archiving of the accepted
Zhu D, Tu S (2017) Mechanisms of support failure induced by repeated manuscript version of this article is solely governed by the terms of
mining under gobs created by two-seam room mining and pre- such publishing agreement and applicable law.
vention measures. Eng Fail Anal 82:161–178. https://​doi.​org/​10.​
1016/j.​engfa​ilanal.​2017.​08.​029
Zingano A, Weiss A (2019) Subsidence over room and pillar retreat
mining in a low coal seam. Int J Min Sci Technol 29:51–57.
https://​doi.​org/​10.​1016/j.​ijmst.​2018.​11.​022

13