International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8

Contents lists available at ScienceDirect

International Journal of
Rock Mechanics and Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

Time function model of dynamic surface subsidence assessment T

of grout-injected overburden of a coal mine
⁎
Binglong Wanga,b, Jialin Xua,b, , Dayang Xuana,b
a
State Key Laboratory of Coal Resources and Safe Mining, China University of Mining and Technology, Xuzhou, Jiangsu 221116, China
b
School of Mines, China University of Mining and Technology, Xuzhou, Jiangsu 221116, China

A R T I C L E I N F O

Keywords:
Overburden grout injection
Subsidence prediction
Knothe time function model
Dynamic surface subsidence

1. Introduction a particularly simple type of single-parameter time function model

Coal-mining-induced surface subsidence is a complex spatio-
temporal phenomenon. With the progressive mining of the working
face, the surfaces of the affected areas undergo three stages of sub-
sidence, namely, initial subsidence, active subsidence, and residual
subsidence.1 Initial subsidence generally only accounts for 10–15% of
the final subsidence and causes minor ground deformation. Active
subsidence is the principal subsidence corresponding to about 75% of
the final subsidence and causes major ground deformation. Residual
subsidence may vary from 5% to 10% of the final subsidence and will
result in further ground deformation.2 The displacements and de-
formation that occur in each of these stages significantly affect and may
damage buildings and other surface infrastructure such as railways,
roads, and water facilities,1,3 resulting in substantial economic losses.4,5
Hence, to ensure the safety of surface structures, it is necessary to ob-
tain a thorough understanding of the dynamic changes induced by
surface subsidence.6,7
Time function models are important for predicting dynamic surface
subsidence processes. Numerous time function models have been pro-
posed in the literature, and they may be broadly classified into two
categories. The first category comprises single-parameter time function
models, a classic example of which is the Knothe time function model.8
The second category comprises multi-parameter time function models
such as the Sroka–Schober,9 arc tangent,10 logistic,11 and normal dis-
tribution7 models. The predictions of multi-parameter models are more
accurate because they are based on more parameters. However, the
difficulty associated with the determination of these parameters re-
duces the practicality of the models. The Knothe time function model is
because it contains only one easily determined time parameter. Thus, it
is often the first choice for use in the prediction of surface sub-
sidence.3,12 However, since the model was first put forward, a number
of researchers have noted its lack of generalisability. To address this
issue, the Knothe time function model has been augmented through
various approaches to facilitate its application to different practical
conditions. Examples of the modified versions are the segmented
Knothe model,13 exponential Knothe model,14 and Kelvin–Voigt creep
equation Knothe model,15 which have enabled extended applications.
However, most of the abovementioned time function models were
intended for conventional longwall working faces, and they also rarely
consider the effects of the specific method used to control mining-in-
duced subsidence. One of the methods used to control mining-induced
subsidence is the isolated overburden grout injection technique. The
technique involves the drilling of boreholes from the surface down to
the bedding separation cavities located at certain heights above the
ceiling of a longwall panel for the injection of fly-ash slurry. The in-
jection compresses the underlying bedding; then, a grouting pillar with
a certain width can be generated at the centre of the longwall panel.
The grouting pillar works the same way as a coal pillar in partial mining
(e.g. room and pillar mining); therefore, with the support of the
grouting pillar, surface subsidence can be better controlled (refer to
Fig. 1).16–22 The technique has been successfully applied in numerous
coal mines.17,18 However, previous studies mainly focused on the final
subsidence when the technique was used,22 with very few studies
considering the dynamic processes of subsidence, especially their
quantitative relationships with the grout injection parameters. Owing
to the limited research and knowledge in this area, subsidence analyses

⁎
Corresponding author at: State Key Laboratory of Coal Resources and Safe Mining, China University of Mining and Technology, Xuzhou, Jiangsu 221116, China.
E-mail address: cumtxjl@cumt.edu.cn (J. Xu).

https://doi.org/10.1016/j.ijrmms.2018.01.044
Received 27 June 2017; Received in revised form 11 December 2017; Accepted 25 January 2018
1365-1609/ © 2018 Elsevier Ltd. All rights reserved.
B. Wang et al.
Buildings
Ground surface
Grouting borehole
The injected fill
Grouting pillar Chain pillar Longwall panel
Fig. 1. Schematic cross section of the isolated overburden grout injection technique
(after22).
during practical grout injection works heavily rely on high-density
surface subsidence measurements. This makes advance predictions
difficult and hampers on-the-fly control of the grout injection para-
meters, strongly undermining the protection of surface facilities. Thus,
the development of a time function model of the dynamic processes of
the subsidence that occurs during overburden grout injection mining is
of utmost importance.
On the basis of the Knothe time function, we analysed how the
parameters of the isolated overburden grout injection technique af-
fected the time parameter, which represents the difference between
conventional longwall mining and mining using the isolated over-
burden grout injection technique. On the basis of the findings, we in-
troduced a time parameter cg into the Knothe time function model. The
new developed time function model enabled the modelling of the sur-
face subsidence during mining with overburden grout injection. Case
studies of different coal mine working faces where the injection tech-
nique was employed were used to determine the parameters of the
developed model, the reliability of which was further verified by field
measurement data.
2. Theoretical background
2.1. Knothe time function model
To describe the displacement and deformation of the surface with
time, Knothe, a Polish mining engineer, proposed the Knothe time
function model, which assumes that the rate of subsidence dW (t ) is
dt
proportional to the difference between the maximum subsidence W0
and the dynamic subsidence at some time t, W(t), which may be ex-
pressed as
dW (t )
= c (W0 − W (t ))
dt
where c is the time parameter (1/year). Given the initial boundary
condition W(t) = 0 at t = 0 (i.e. W(0) = 0), the time function model
may be derived by solving the following first-order linear differential
equation:3,8,14
W (t ) = W0 [1 − exp(−ct )]
2.2. Time parameter c and its determining factors
The time parameter c is an important parameter for characterising
the duration of the subsidence process in the Knothe time function
model. The duration of the subsidence is inversely proportional to the
time parameter c; that is, a larger value of c means that the duration of
the subsidence is shorter. Different Knothe time parameter curves may
be obtained by varying the value of c, as shown in Fig. 2, which con-
siders the case where the final subsidence W0 = −1600 mm. The
International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8
Coal seam
Fig. 2. Knothe time function curves for various values of the time parameter c.
duration of the subsidence is 9 months when c = 0.5 and increases to
23 months when c = 0.2.
Thus, accurate determination of the time parameter c is important
for the accurate prediction of dynamic subsidence. Numerous studies
have investigated the factors that determine c for conventional longwall
mining,23–25 and c has been determined to be a location-dependent
constant generally related to the rate of progression of the working face,
the lithology of the overburden, and the burial depth. For some working
faces, when these conditions are fixed, the value of c can be determined
and remains constant with time. However, because the conditions of
coal mining are highly complex and variable, c substantially differs
with the location. To address this issue, numerous equations for esti-
mating c have been proposed.3,23,25
2.3. Applicability of the Knothe time function to overburden grout injection
mining
To examine the applicability of the Knothe time function model (Eq.
(2)) when overburden grout injection is used during mining operations,
we measured the subsidence at the points of maximum subsidence for
four different working faces with grout injection in the Huaibei coal-
field and fitted them to the curves predicted by the Knothe time func-
tion model (Fig. 3). Although the correlation coefficients of the fits were
as high as 0.81–0.88, the shapes of the fitted curves substantially dif-
fered from those of the measured data. For example, the subsidence of
Longwall 8103 indicated by the fitted curves were significantly larger
than the actual measurements up to 5 months, whereas they were sig-
nificantly lower thereafter (Fig. 3(d)). This shows that the Knothe
function model does not accurately describe the dynamic processes of
the subsidence that occurs when overburden grout injection is em-
ployed in coal mining.
(1)
3. Time function model for mining with overburden grout
injection
The main cause of the failure of the Knothe time function to cor-
rectly predict the dynamic subsidence during mining with overburden
(2) grout injection is that the time parameter of the function for this mining
technique, cg, distinctly differs from the conventional longwall mining
time parameter c. Ideally, the subsidence is caused by the propagation
of the fracture cavities in the overburden resulting from the continuous
progression of the mining space.26,27 Compared to conventional long-
wall mining, mining with the injection of grout into the overburden
eliminates some of the fracture cavities, thereby terminating their
propagation to the surface and significantly decreasing the duration of
the surface subsidence. This causes cg to be larger than c.
The determination of cg is thus key to the development of a time
function model for a longwall working face with overburden grout in-
jection. In this section, we first analyse the determining factors of cg and
2
B. Wang et al.
Fig. 3. Comparison of the measured subsidence of coalmine working faces under overburden grout injection in the Huaibei coalfield with the predictions of the Knothe time function: a)
Longwall 8281, b) Longwall 1018, c) Longwall 1044, and d) Longwall 8103.
the variational characteristics of this parameter, and then derive an
expression for cg based on the grout injection parameters. The time
parameter cg is then used to develop a new time function model of
longwall mining with overburden grout injection. Finally, the relevant
parameters of actual mining cases were investigated and used to de-
termine those of the proposed model.
3.1. Derivation of an expression for the time parameter cg
The time parameter cg of the working faces with overburden grout
injection is affected by the factors that affect the conventional longwall
mining time parameter c. However, cg is additionally affected by the
thickness of the injected fill, h, owing to the filling of the overburden
fracture cavities. Since cg is inversely correlated with the surface sub-
sidence duration, the relationship between cg and h may be determined
by analysing the relationship between the surface subsidence duration
and h. Incidentally, the surface subsidence duration is closely related to
the dimensions of the mining-induced overburden fracture cavities.
Grout injection into the overburden decreases the total number of
mining-induced fracture cavities as well as the propagation time of the
fracture cavities toward the surface, thus decreasing the duration of the
surface subsidence. The decrease in the duration of the subsidence
becomes more pronounced as the thickness of the fill increases. This
indicates that the time parameter cg increases in proportion to the
thickness of the fill h.
Because the number of mining-induced fracture cavities varies with
the depth, the effect of the cavities on the duration of surface sub-
sidence would also vary with the depth. Theoretically, the number of
mining-induced fracture cavities decreases as the vertical horizon from
the coal seam increases. The effect of the fracture cavities on the
duration of the subsidence thus decreases as the depth of the cavities
decreases. Similarly, for a fill with a given thickness, the duration of the
surface subsidence decreases as the distance of the fill from the coal
seam increases. Hence, the time parameter cg increases as the distance
3
International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8
of the fill from the coal seam increases. An impact coefficient n may be
introduced on this basis to assess the impact of the height of the stratum
of the grout injection (i.e. the distance of the injected fill from the coal
seam) on the time parameter cg.
In summary, it may be inferred that cg is affected by the factors that
affect the conventional time parameter c and is proportionally related
to the n-th power of the fill thickness h. Thus, it is supposed that
cg = achn (3)
We have previously established a model of the fill distribution in an
overburden for the purpose of determining the value of h.21 In the cross
section, the fill may be considered to have a graded distribution (Fig. 4),
and the relationship between the thickness of the fill within the over-
burden (at the centre of the working face) and the volume of the
compacted fill, Q, may be expressed as
Q = h (W − H1/ tan φ) L (4)
where W is the width of the working face (m), φ is the angle of full
subsidence, and the empirical value of 60° was used for the calculation
in the Huaibei coalfield28; H1 is the distance between the fill and the
ceiling of the coal seam (m); and L is the advancement length of the
working face in the grout injection area (m). The volume of the fill may
be expressed as
Q = Vf (t + t0) (5)
where Vf is the filling rate of the compacted ash (m3/d); t is the duration
of the subsidence at an arbitrary point on the surface (d); and t0 is the
duration of the grout injection prior to the subsidence (d). It should be
noted that Vf is actually a variable, but because it only varies within a
relatively small range, it may be reasonably approximated by its
average value.
An expression for the time parameter for mining with overburden
grout injection may be obtained by combining Eqs. (3)–(5), as follows:
B. Wang et al. International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8

Fig. 4. Schematics of the injected fill: a) three-dimensional view and b) cross-section.

Vf time function includes the mining parameters (the width of the

cg = ac [ ]n (t + t0)n
(W − H1/ tanφ) L (6) working face and the advancement length of the filled area), geolo-
gical parameters (the angle of full subsidence), grout injection para-
Eq. (6) indicates that the Knothe time parameter cg for mining with
meters (the distance of the grout injection stratum from the coal seam,
overburden grout injection is not constant but varies with the geolo-
the filling rate of compacted ash, and the grout injection time before
gical setting of the mine, the grout injection parameters, and the time t.
the onset of subsidence), and other relevant coefficients (the constant
It is noted from Eq. (5) that the time parameter and model are only
a, conventional mining time parameter c, and impact coefficient n).
suitable for calculating the dynamic subsidence of the surface points
The new time function will promote further evaluation of the effec-
that begin to move after grout injection.
tiveness of its subsidence control measures and on-the-fly adjustments
of the grout injection parameters.
3.2. Construction of the time function model
3.3. Parameter determination
Substitution of the time parameter for mining with overburden
grout injection, cg (Eq. (6)), into the Knothe hypothesis (Eq. (1)) gives The developed time function model for mining with overburden
dW (t ) Vf grout injection contains the time parameter c for conventional longwall
= ac [ ]n (t + t0)n [W0 − W (t )] mining, the constant coefficient a, and the impact coefficient n. In this
dt (W − H1/ tanφ) L (7)
study, the time parameter c was first determined using empirical
The time function model of the dynamic subsidence during mining equations available in the literature, and the other two parameters, a
with overburden grout injection may be obtained by solving Eq. (7) and n, were then determined using the subsidence data obtained from
using W(0) = 0: four working faces in the Huaibei coalfield. The geological, mining, and
grout injection parameters of the four working faces are presented in
⎧ ac [Vf /(LW − LH1/ tanφ)]n n + 1 ⎫ Table 2.
W (t ) = W0 1 − exp ⎧ [t0 − (t + t0)n + 1] ⎫
⎨ ⎨ n + 1 ⎬⎬
⎩ ⎩ ⎭⎭
3.3.1. Determination of the conventional time parameter c
(8)
The time parameter for conventional longwall mining, c, was ob-
To explain superiority of the new time function model, we fitted tained using the equation of Hu et al.3:
the subsidence at the points of maximum subsidence in Longwall 8301 v
with overburden grout injection by the new time function model and c=− ln 0.02
2Hs cot ψs + 2Hj cot ψj (9)
widely used models (Table 1). Although the correlation coefficients of
the fits were as high as 0.82–0.98, the previous models were only where ψs and ψj are respectively the angles of the full subsidence of the
mathematical formulas that did not reflect the injected fill char- unconsolidated strata and bedrock (Fig. 5), empirically determined to
acteristics of a working face with overburden grout injection. The new be 45° and 60°, respectively, for the Huaibei coalfield.28 The de-
termined parameters of the working faces were substituted into Eq. (9)
to determine the time parameter c (see Table 3).
Table 1
Fitting results for the subsidence in longwall 8301 by the new time function model and
3.3.2. Determination of the impact coefficient n and constant a
widely used models.
The impact coefficient n and constant a of the four aforementioned
Function R2 working faces could be directly determined by fitting the developed
time function (Eq. (8)) to the measured data of the subsidence with time
1. Knothe W (t ) = W0 [1 − exp(−ct )] 0.82
at the points of maximum subsidence of the considered working faces.
ξ f 0.92
2. Sroka–Schober W (t ) = W0 ⎡1 + exp(−ft ) − exp(−ξt ) ⎤ Because a is a constant, the average value of a for the four working faces
⎣ f−ξ f−ξ ⎦
W0 0.98 (0.18) was adopted as the approximate value of a for the Huaibei
3. Logistic W (t ) =
1 + b exp(−ct )
Mining Area. Fig. 6 shows the fitted curves of the developed time
4. Exponential Knothe W (t ) = W0 [1 − exp(−ct )]k 0.95
function model and the measured subsidence with respect to time of the
5. The new time function 0.99
four considered working faces in the Huaibei coalfield.
ac [Vf / (LW − LH1 / tanφ)]n
{
W (t ) = W0 1 − exp { n+1
[t0n + 1 − (t + t0 )n + 1] }} Next, we developed a method for calculating the impact coefficient
n. According to the definition of n, it is positively correlated with H1.

4
B. Wang et al. International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8

Table 2
Geological, mining, and grout injection parameters of four grout-injection mining working faces in the Huaibei coalfield.

Working face Hs (m) Hj (m) H1 (m) Vf (m3/d) L (m) W (m) t0 (month) v (m/d)

8281 290 98 60 594 313 161 1.2 3.7

1018 285 317 99 463 498 167 0.7 1.5
1044 350 268 143 350 497 150 2.5 2.0
8103 230 350 150 488 250 184 0.8 2.4

Hs: thickness of the unconsolidated formations; Hj: thickness of the bedrock; v: advancement speed of the working face.

Fig. 5. Schematic of the angle of full subsidence.

Table 3
Parameters of the developed time function model of four working faces in the Huaibei Fig. 7. Comparison of the fitted and measured values of the impact coefficient n of the
coalfield. considered working faces.

Working face Time parameter c Impact coefficient n Constant a

Hence, the fitting of n was performed using H1/Hj, and the relationship
8281 0.65 4.14 0.21
between H1/Hj and n for the working faces of the Huaibei coalfield was
1018 0.19 0.97 0.19
1044 0.23 3.41 0.18 determined to be n = 11H1/Hj – 2.7 (Fig. 7). The fitting revealed that,
8103 0.33 1.62 0.16 for a given bedrock thickness, the effect of the fill thickness on the time
parameter cg increased as the distance of the injection stratum from the
ceiling of the coal seam increased.

Fig. 6. Fitted curves of the developed function model and measured subsidence with respect to time of the four considered working faces: a) Longwall 8281, b) Longwall 1018, c)
Longwall 1044, and d) Longwall 8103.

5
B. Wang et al.
Fig. 8. Plan of the 7226 working face and a picture of the nearby surface buildings.
4. Experimental validation
4.1. Geological setting and mining conditions
The validity of the developed time function model was verified by
using it to predict the dynamic subsidence of the 7226 working face of
the Yuandian No. 2 coal mine in Huaibei, China. The predictions were
compared with the field measurements. Numerous buildings are located
on the surface around the mine working face (Fig. 8), which has a
mining height, an advancement length, and a mining width of 5.0 m,
860 m, and 180 m, respectively. Because surface buildings are parti-
cularly located beside the termination line of the working face, coal
recovery is performed via direct longwall caving over the first 544 m
along the advancement length. To protect the surface buildings, the
isolated overburden grout injection technique is only used over the final
316 m of the advancement length.
Four grout injection boreholes were directly drilled during extrac-
tion within 316 m of the termination line of Longwall 7226. The first
was drilled 300 m from the termination line. However, it has been
observed from experience that the grout injected through the first
borehole does not strictly diffuse in the forward direction owing to the
difficulty of preventing backward diffusion.22 The backward dispersion
of the injected grout is estimated to extend over 100 m, which makes
the length of the overburden grout injection area for the working face
approximately 400 m. Compacted-ash grout was injected through a
stratum 78 m above the coal seam with an average filling rate of
818 m3/d. The subsidence of the surface was observed throughout the
grout injection process. The measurement points within the grout
6
International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8
injection area are shown in Fig. 8. Detailed descriptions of the grout
injection process and the monitoring of surface subsidence are available
in the work of Teng et al.16
4.2. Extraction of the time function model parameters
The bedrock thickness of the 7226 working face, Hj, is 213 m, and
the distance of the fill from the coal seam, H1, is 78 m. Substituting
these values into n = 11H1/Hj – 2.7 yields an impact coefficient n of
1.3. As noted earlier, the value of the constant coefficient a is 0.18. In
addition, from Eq. (9), the conventional longwall mining time para-
meter c was determined to be 0.32.
4.3. Results and analyses of the predictions
The developed time function model was used to predict the sub-
sidence of the B4–B9 monitoring stations (measurement points) of the
7226 working face (Fig. 9). The time interval t0 for each of these points
is 0.27 months, which implies that grout injection is continuously
performed for 0.27 months beginning at the onset of surface displace-
ment at the measurement points. It should be noted that the predicted
values of the final subsidence W0 for measurement points B4–B9 agree
well with the measured values.
To quantify the accuracy of the predictions, the following equations
were respectively used to obtain the root mean square errors (RMSEs)
and relative standard errors (RSEs) of the predictions relative to the
measured values:
B. Wang et al.
Fig. 9. Comparison of the dynamic subsidence measured at points B4–B9 with the pre-
dictions of the developed time function model.
Table 4
Prediction accuracies for the different measurement points.
Measurement point RMSE (mm)
B4 7.1
B5 7.2
B6 12.3
B7 19.7
B8 30.7
B9 57.0
∑ di2
m=±
N (10)
m Interferometry (Northern France). J Appl Geophys. 2009;69(1):24–34.
f=
Wm
where m is the RMSE, f is the RSE, di is the difference between the
predicted and measured values at the i-th measurement point, Wm is the
maximum subsidence at the measurement point, and N is the number of
measurements and predications. The results for the different measure-
ment points are presented in Table 4.
The maximum RSE for the six measurement points was 8.6% with
an average value of only 6.1%, which indicates a high reliability level
for the predictions. The RMSEs for measurement points B4–B8 ranged
between 7.1 and 30.7 mm, which is much smaller than 57.0 mm for
point B9. This is because points B4–B8 were located within the area of
the working face that received an adequate amount of grout injection
and were thus minimally affected compared to the mining areas
without grout injection. Therefore, the accuracy of the predictions for
these measurement points was quite high. Conversely, measurement
point B9 was close to the un-injected area and thus benefitted mini-
mally from grout injection, resulting in poorer predictions.
5. Conclusions
To ensure the safety of surface structures above the longwall panels,
it is necessary to obtain a thorough understanding of the dynamic
surface subsidence associated with mining. In this study, a new time
function model was developed to predict the dynamic subsidence that
occurs in the grout-injected overburden of a coal mine during over-
burden grout injection mining.
This function includes the mining parameters (the width of the
working face and the advancement length of the filled area), a geolo-
gical parameter (the angle of full subsidence), grout injection para-
meters (the distance of the grout injection stratum from the coal seam,
the filling rate of compacted ash, and the grout injection time before the
onset of subsidence), and other relevant coefficients (the constant a,
conventional mining time parameter c, and impact coefficient n). The
International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8
function thus adequately reflects the injected fill characteristics of a
working face with overburden grout injection. A method for de-
termining the parameters of the developed model was established
through case studies of several overburden grout injection working
faces in the Huaibei Coalfield. The proposed model was experimentally
validated by field measurements.
However, it is noted that the developed model is only suitable for
working faces with a uniform grout injection rate. Moreover, owing to
some constraints that were imposed during its development, it can only
be used to predict the dynamic subsidence at a surface point that begins
to subside after the commencement of grout injection. In addition, the
values of some of the model parameters determined in this work, such
as a and n, are only suitable for the Huaibei coalfield in China, having
been calibrated on the basis of the conditions of the mining area.
Nonetheless, the proposed dynamic subsidence model promises to sig-
nificantly facilitate on-the-fly control of the grout injection parameters
and real-time assessment of surface subsidence.
Acknowledgements
RSE (%) This study was financially supported by the Natural Science
7.7
Foundation of China (Grant No. 51604258) and the Independent
4.7 Research Projects of State Key Laboratory of Coal Resources and Safe
5.3 Mining, China University of Mining and Technology (Grant No.
4.7 SKLCRSM15X06).
5.5
8.6
References
1. Liu BC, Liao GH. Surface Movements in Coal Mines. Beijing: China Industry Publishing
House; 1965.
2. Guéguen Y, Deffontaines B, Fruneau B, et al. Monitoring residual mining subsidence
of Nord/Pas-de-Calais coal basin from differential and Persistent Scatterer
3. Hu QF, Deng XB, Feng RM, Li CY, Wang XJ, Jiang T. Model for calculating the
(11) parameter of the Knothe time function based on angle of full subsidence. Int J Rock
Mech Min. 2015;78:19–26.
4. Can E, Mekik Ç, Kuşcu Ş, Akçın H. Monitoring deformations on engineering struc-
tures in Kozlu Hard Coal Basin. Nat Hazards. 2013;65(3):2311–2330.
5. Donnelly LJ, De LA, Cruz H, Asmar I, Zapata O, Perez JD. The monitoring and pre-
diction of mining subsidence in the Amaga, Angelopolis, Venecia and Bolombolo
Regions, Antioquia, Colombia. Eng Geol. 2001;59(1):103–114.
6. Cui XM, Wang JC, Liu YS. Prediction of progressive surface subsidence above long-
wall coal mining using a time function. Int J Rock Mech Min. 2001;38(7):1057–1063.
7. Gonzalez-Nicieza C, Alvarez-Fernandez MI, Menendez-Diaz A, Alvarez-Vigil AE. The
influence of time on subsidence in the Central Asturian Coalfield. Bull Eng Geol
Environ. 2007;66(3):319–329.
8. Knothe S. Effect of time on formation of basin subsidence. Arch Min Steel Ind.
1953;1(1):1–7.
9. Kwinta A, Hejmanowski R, Sroka A. A time function analysis used for the prediction
of rock mass subsidence. In: Mining Science and Technology, YG Gui & TS Golosinski,
ed, Proceeding of the International Symposium on Mining Science and Technology.
Xuzhou; 1996:419–424.
10. Nie L, Wang HF, Xu Y, Li ZC. A new prediction model for mine subsidence de-
formation: the arc tangent function model. Nat Hazards. 2015;75(3):2185–2198.
11. Xu HZ, Li XH. Time function of surface subsidence based on logistic growth model.
Rock Soil Mech. 2005;26:151–153.
12. Peng XZ, Cui XM, Zang YQ, Wang Y, Yuan DB. Time function and prediction of
progressive surface movements and deformations. J Univ Sci Technol Beijing.
2004;26(4):341–344.
13. Chang ZQ, Wang JZ. Study on time function of surface subsidence—the improved
Knothe time function. Chin J Rock Mech Eng. 2003;22(9):1496–1499.
14. Liu YC, Zhuang YH. Model for dynamic process of ground surface subsidence due to
underground mining. Rock Soil Mech. 2009;30:3406–3410.
15. Gao YF, Jia JY, Li B, Zhang QS. The attenuation function of surface subsidence and
stability analysis due to mining. J China Coal Soc. 2009;34(7):892–896.
16. Teng H, Xu JL, Xuan DY, Wang BL. Surface subsidence characteristics of grout in-
jection into overburden: case study of Yuandian No. 2 coalmine, China. Environ Earth
Sci. 2016;75:1–11.
17. Xu JL, Xuan DY, Zhu WB, Wang XZ, Wang BL, Teng H. Study and application of coal
mining with partial backfilling. J China Coal Soc. 2015;40(6):1303–1312.
18. Xu JL, Ni JM, Xuan DY, Wang XZ. Coal mining technology without village relocation
by isolated grout injection into overburden. Coal Sci Technol. 2015;43(12):8–11.
19. Xuan DY, Wang BL, Xu JL. A shared borehole approach for coal-bed methane drai-
nage and ground stabilization with grouting. Int J Rock Mech Min Sci.
2016;86:235–244.
20. Xuan DY, Xu JL, Wang BL, Teng H. Borehole investigation of the effectiveness of
grout injection technology on coal mine subsidence control. Rock Mech Rock Eng.
7
B. Wang et al.
2015;48(6):2435–2445.
21. Xuan DY, Xu JL, Wang BL, Teng H. Investigation of fill distribution in post-injected
longwall overburden with implications for grout take estimation. Eng Geol.
2016;206:71–82.
22. Xuan DY, Xu JL. Grout injection into bed separation to control surface subsidence
during longwall mining under villages: case study of Liudian coal mine, China. Nat
Hazards. 2014;73(2):883–906.
23. Hu QF, Cui XM, Kang XL, Lei B, Ma KF, Li LC. Impact of parameter on Knothe time
function and its calculation model. J Min Saf Eng. 2014;31(1):122–126.
24. Li DH. Influence of cover rock characteristics on time influencing parameters in
process of surface movement. Chin J Rock Mech Eng. 2004;23(22):3780–3784.
International Journal of Rock Mechanics and Mining Sciences 104 (2018) 1–8
25. Wu K. A programme of mining subsidence dynamic prediction and its application.
Eng Surv Map. 1995;4(3):44–48.
26. Palchik V. Experimental investigation of apertures of mining-induced horizontal
fractures. Int J Rock Mech Min Sci. 2010;47(3):502–508.
27. Palchik V. Localization of mining-induced horizontal fractures along rock layer in-
terfaces in overburden: field measurements and prediction. Environ Geol.
2005;48(1):68–80.
28. State Bureau of Coal Industry. Regulations of Coal Pillar Design and Extraction for
Buildings, Water Bodies, Railways, Main Shafts and Roadways. Beijing: Coal Industry
Press; 2000.