International Journal of Coal Geology 66 (2006) 227 – 252

www.elsevier.com/locate/ijcoalgeo

Modelling of strata movement with a special reference to caving

mechanism in thick seam coal mining
B. Unver *, N.E. Yasitli
Hacettepe University, Department of Mining Engineering, Beytepe 06532, Ankara, Turkey
Received 7 February 2005; accepted 18 May 2005
Available online 2 February 2006

Abstract

Caving of top coal behind the face is the key factor affecting the efficiency of production at thick coal seams. During
production of top coal by caving behind the face not only is a significant amount of coal lost in the goaf but also the coal drawn
by means of caving is usually diluted considerably with surrounding rock. Therefore, it is not possible to carry out an efficient
production operation unless caving of top coal behind the face is optimized. In this paper, results of 3-D modelling of the top
coal caving mechanism at the M3 longwall panel of Omerler Underground Mine located at Tuncbilek (Turkey) are presented by
using a finite difference code called FLAC3D. According to modelling results, a 1.5–2.0-m-thick layer of coal just above the
shield supports is well fractured. However, 3.0–3.5-m-thick layer of coal above the well-fractured part is either not fractured or
fractured in the form of large blocks leading to obstruction of windows of shields during coal drawing. It is concluded that, in
order to decrease dilution and increase extraction ratio and production efficiency, the top coal should be as uniformly fractured
as much as possible. Hence, an efficient and continuous coal flowing behind the face can be maintained. A special pre-fracture
blasting strategy just sufficient enough to form cracks in the top coal is suggested by means of comparing results from
numerical modelling.
D 2005 Elsevier B.V. All rights reserved.

Keywords: Thick coal seam mining; Top coal caving; Numerical modelling; FLAC3D

1. Introduction ization differs among countries, but a thickness of 4.8

m is accepted as the lower limit (Singh, 1999). Pro-
Thick seams comprise half of the world’s coal duction methods at the former Yugoslavia, Hungary,
reserve and 70–80% of thick coal seams are produced Poland, France, India and China are generally similar,
by means of underground mining methods (Atkinson, however differences arise due to local conditions
1979; Senkal et al., 1988). Thick coal seam categor- (Schweitzer, 1977; Ahcan, 1977; Garg and Nath,
1977; Yasitli, 2002; Unver and Yasitli, 2002).
* Corresponding author. Tel.: +90 312 2977696; fax: +90 312 The production method applied at a thick coal
2992155. seam is different from conventional single slice coal
E-mail address: [email protected] (B. Unver). mining in many aspects. Although there are operating
0166-5162/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.coal.2005.05.008
228 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

longwall faces up to a height of 6 m in some parts of method is not currently used extensively in France
the world, in practice it is difficult to extract a slice of and former Yugoslavia, these countries had pio-
more than 4 m. Depending on the local characteristics neered the development of the production method
of a coal seam; such as seam inclination, continuity, for other countries.
surrounding rock and mechanical properties and sus- Although longwall mining with the top coal caving
ceptibility of coal to spontaneous combustion, a face is an attractive method due to higher production effi-
height of 2–2.5 m is preferred. ciency and lower development requirements, it has
Depending on slope and thickness variations, the significant difficulties in terms of the flow character-
application of an optimum slicing method may be istics of coal, timing and rate of coal drawing, loss of
prevented. Therefore, production is carried out by coal in the goaf and dilution. The primary purpose of
means of sublevel caving of top coal behind a face this research is to investigate the above-stated short-
formed at the bottom of the seam (Yasitli, 2002; comings and find ways to overcome difficulties
Unver and Yasitli, 2002). Production of thick seams related mainly to top coal caving.
by using top coal caving is much simpler in com- There have been numerous efforts to improve the
parison to slice mining and requires less develop- efficiency of thick coal seam mining. Jha and Kar-
ment, consequently, the efficiency of production is makar (1992) investigated the factors affecting pro-
significantly higher. The top coal caving method was duction. Strata behavior during caving and dilution
first applied in the 1940s in Russia and then subse- of top coal with surrounding rock was investigated
quently used in France, former Yugoslavia, Hungary, in situ by Singh et al. (1992) and Singh (1999) who
Romania, former Czechoslovakia and Turkey. The determined the formation and behavior of the
method has been in use in China and India since the immediate roof on physical models in the labora-
1980s (Yasitli, 2002; Tien, 1998). At present, the tory. Dian (1992) made a comparison between slice
method is used extensively in China and India. mining and top coal caving methods. Wu (1992)
There are a couple of mines using the method in tried to determine factors affecting the dilution of
former Yugoslavia and Turkey, whereas in other coal. Three researchers were mainly focused on the
countries due to depletion of reserves, geological isolation of factors and understanding the phenom-
constraints, technological and/or economical reasons enon by means of in situ measurements and obser-
the method is not used at present. Although this vations and constructing physical models. This

Sublevels
Wiremesh

Fig. 1. Sublevel caving mining of steep and thick coal seam with wiremesh method (Jeremic, 1985).
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 229

(a) A (b)
A-A'

8-15 m

2,6-2,8 m

A'
Front view Side view
Fig. 2. Soutirage method (Jeremic, 1985).

study is aimed at modelling of top coal movement other by the application of an artificial roof in the
at an operating longwall panel enabling a verifica- form of wire mesh (Fig. 1). During caving of top coal,
tion of numerical modelling results with in situ the artificial roof prevents the mixing of coal with side
conditions. rock. However, the success of this method depends on
good caving characteristics of the top coal.
The Soutirage method was used in France for
2. A brief presentation of production methods production of seams thicker than 12 m. A short retreat
incorporating top coal caving longwall face having a height of around 2.5 m is
formed and 3.5–12-m-thick top coal is produced by
There are mainly three different methods of pro- means of caving (Fig. 2).
duction with top coal caving and gravity flow The Velenje method was developed in Yugoslavia
namely; artificial roof, Soutirage and Velenje meth- resembling the Soutirage method for the production of
ods. Selection of the production method depends on a 16-m-thick coal seam having a slope of 30–608. A
local conditions. 2.5–3-m-high retreat longwall face is advanced paral-
Top coal caving using the artificial roof method has lel to strike and 7.5–10-m-thick coal is produced by
been used in Russia for production of seams having a caving behind the face (Fig. 3). The main difference
slope over 608 and a thickness over 5 m. The method between Soutirage and Velenje methods arises from
basically depends on the caving of top coal with the mechanical properties of coal seams. Since coal
gravity in the form of sublevels separated from each seams in former Yugoslavia were stronger than

Blasting
holes
7,5-10,0 m

2,5-3,0 m

Fig. 3. Velenje method (Jeremic, 1985).

230 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Thickness Lithology Formation Kutahya. The mine is a subsidiary of Turkish Coal

1m Top soil Enterprises and started production in 1985 and a fully
mechanized face was established in 1997. Average
thickness of overburden was around 240 m and the
24 m 1 Calcareous marl average 8-m-thick coal seam has a slope of 108.
A stratigraphic column of the coal seam, roof and
floor strata is presented in Fig. 4. There are three main
geological units classified as claystone, calcareous
marl and marl in the mine area (Destanoglu et al.,
2000). There is a 30–80-cm-thick clay layer called 3b
189 m 2 Marl at the roof contact of the coal seam frequently creating
stability problems due to its low strength and fractured
characteristics. There is another claystone layer named
3a Claystone as 3a just above the soft clay layer. The claystone
17 m
3b Soft claystone layer called 3a is stronger than soft clay and its natural
moisture content is lower. There is a claystone layer
denoted as 3c having an average thickness of 4 m at
the floor of the seam. The light gray 3c located at the
8m 4 Coal floor is stronger than the dark gray soft claystone 3b
and claystone 3a. The coal seam numbered 4 contains
three clay bands (15, 75 and 55 cm thick) from top to
bottom. Physical and mechanical characteristics of the
coal and surrounding rock are presented in Table 1
4m 3c Claystone (Destanoglu et al., 2000; Taskin, 1999; Yasitli, 2002;
Yasitli and Unver, 2003).
Fig. 4. A generalized stratigraphic column at Omerler Coal Mine. The region in the mine indicated in Fig. 5 was
reserved for production by the fully mechanized long-
France, drilling and blasting were applied in the wall face in the form of six panels. Panel length was
Velenje method to aid caving (Jeremic, 1985). selected according to geological conditions. M1 and
M2 panels had been worked out and M3 panel was in
production during the study (Fig. 5).
3. Omerler Underground Mine and production The 8-m-thick coal seam at M3 panel has been
method produced by means of the longwall retreat with top
coal caving production method where a 2.8-m-high
Omerler Underground Mine is located in the inner longwall face was formed at the floor of coal seam.
Aegean District of Turkey near Tuncbilek-Tavsanli, Top slice coal having a thickness of 5.2 m was caved

Table 1
Physical and mechanical properties of coal and surrounding rocks (Destanoglu et al., 2000; Taskin, 1999; Yasitli, 2002)
Formation Definition Density Porosity Uniaxial Tensile Internal Cohesion c Modulus of Poisson’s
code (MN/m3) (%) compressive strength friction (MPa) elasticity E ratio m
strength (MPa) (MPa) angle (/) (MPa)
Calcareous marl 1 0.023 13.8 29.2 3.9 47 12.5 5520 0.26
Marl 2 0.022 – 16.1 1.9 31 5.0 2530 0.25
Roof claystone 3a 0.021 21.30 14.4 2.3 32 3.18 1480 0.28
Soft claystone 3b 0.023 10.8 8.7 1.8 15–35 – 2040 –
Floor claystone 3c 0.024 21.30 26.5 3.5 40 2.90 2085 0.31
Coal 4 0.013 9.72 15.9 – 15–25 – 1733 0.25
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 231

Ex
ha
us
tF
an
Entrance
Ma

C
in G
alle

M1
ry

M2
B
M3

M4
A
M5
M6

Fig. 5. Plan view of Omerler Underground Mine.

and produced through windows located at the top of determination of mining method related parameters
the shields (Fig. 6). Production was carried out in such as geometry and dimensions of longwall panel,
three stages; face cut, maintenance and winning of face supports and production scheduling. Top coal
top coal by means of caving and drawing through usually caves in the form of blocks of various dimen-
windows of the shields. Daily production activities sions, small pieces and fine particles. Therefore, a
were scheduled as three face cuts in the first shift, next very heterogeneous particle size distribution is usually
shift was used for maintenance and during third shift present in the top coal leading to very complex caving
the top coal was produced (Fig. 7). and flowing characteristics. Flowing rules regarding
viscous materials cannot be applied to fractured solid
material. In order to facilitate flow characteristics of
4. Caving mechanism in top coal caving method the top coal, particle size distribution should be kept
as uniform as possible. The principle of broken solid
Certainly, caving of naturally fractured or blasted material flow is independent of size. However, it
top coal under gravity is the most critical factor in the becomes rather complicated if fine and coarse material

Soft Claystone

COAL SEAM
5.2 m

Direction of
advance

2.8 m

Claystone

Fig. 6. Longwall with top coal caving method at Omerler Underground Mine.
232 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

(a) Face cut (b) Advance of shields

(c) Advance of shields after 3 cuts (d) Caving and production of top
and drilling and blasting of top coal
coal

(e) Advance of face conveyor (f) Start of next cycle

Fig. 7. Production schedule in the mechanized longwall face.

IV
III
makes a heterogeneous mixture. The effect of particle
size in the mixture on flowing characteristics is clearly II
demonstrated in Fig. 8. While a mixture of uniformly
sized particles flows at an angle of 408, a heteroge-
neous mixture containing fines and large blocks will GF I
B
BII
not tend to flow until an angle of more than 858 from BI
the horizontal plane. A
AII

4.1. Basic gravity flow modelling of granulated mate- AI

rial mixture

Caving and drawing of top coal through windows

of shields might be compared with the flow of gran-
ular material in bins and hoppers. The flow of gran-
ular material in bins and hoppers cannot be considered
exactly the same as caving and drawing of coal above
the face. However, in terms of principles of granulated
material flow under gravity and coal drawing, a good
correlation between the two processes may be estab- 90° 85° 70° 55° 40°
lished. Gravity flow of blasted coal or roof rock in top Fig. 8. Four basic types I through IV of coarse material and their
coal caving is a process much more complicated than mobility as a function of inclination of chute or orepasses (Kvapil,
flow in bins (Kvapil, 1992; Unver, 1995). Coarse 1992).
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 233

medium-sized fragments, chippings, sand and/or

clay size components. If water is present in coarse
material Type IV, fine particles could create plastic
and sticking components. In this case, the flow char-
acteristics of the material would become very unfa-
vorable (Kvapil, 1992).
White sand The principles of granular material flow in sublevel
or block caving methods can be considered as similar
to the top coal caving method. The flow process is
usually demonstrated with a simple vertical glass
model with horizontally layered white and black
sand as shown in Fig. 9. The glass bin has been
Black sand designed to observe the motion of sand layers as a
consequence of material withdrawal from an opening
at the bottom. Depending on the withdrawal of mate-
rial from the bottom opening, an elliptical zone of
motion has been observed. The deflection of the
Fig. 9. Simple model of gravity (Kvapil, 1992). originally horizontal thin layers indicates the active
zone, that is, the zone with gravity motion of the
materials can be very heterogeneous. Type I shows material. Since the motion is caused by gravity, the
coarse material with large bsphericalQ pieces of more axis of the active zone is stationary. The zone that is
or less the same size and shape (Fig. 8). Type II not affected by the gravity motion is called the passive
represents a material of almost the same size but zone. As can be clearly seen in Fig. 10b, an ellipsoidal
different shape. Type III indicates a coarse material motion zone can be observed by outlining the active
composed of large fragments, chippings and small zone. Normally, this ellipsoid consists of the horizon-
size particles. Type IV represents a coarse material tal and vertical regions of a nonsymmetrical object, in
that is characterized as a mixture of large blocks, other words, it is not exactly an ellipse but it is an

(a) (b)

Bin

Passive zone
Boundary of
ellipsoid
Active zone

Ellipsoid of
a extraction

c H
d
a
b b

Fig. 10. Successive phase of material extraction from model (Kvapil, 1992).
234 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

ellipsoidal shape and its geometry can be expressed Vertical Angle of

by eccentricity (Kvapil, 1998). wall Δ deviation
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ellipsoid of
e ¼ 1=a a2  b2 ð1Þ loosening

where a: semi-major axis of the ellipsoid, b: semi-

minor axis of the ellipsoid.
Ellipsoid of
extraction
4.2. Mathematical description of sublevel caving

The existence of the extraction ellipsoid can be

Vertical
demonstrated by using several different methods. outlet
One possibility is a three-dimensional physical
model in which marks are located in a certain pattern Fig. 12. Principle of gravity flow in an ellipsoid of extraction and
loosening.
in the granular material (Kvapil, 1992). Knowing the
original location of the marks before the extraction,
the marks extracted together with the material will with complete extraction of the extraction ellipsoid,
define the extraction ellipsoid. Fig. 11a represents the the original horizontal plane passing through the apex
model with a clearly marked ellipsoid of extraction. point N is deflected downwards forming the outflow
Apex point N is at a distance h n above the discharge funnel 1, N, 2. The points 1 and 2 intersect the
opening and it represents the height of the extraction ellipsoid of loosening at the height h n which is
ellipsoid. As can be seen from phases b, c and d (Fig. equal to the height of the extraction ellipsoid. It
11), an increase in extraction results in a successive means that the outflow funnel diameter 1, 2 is equal
decrease in the volume of the loosening ellipsoid. to the horizontal section of the ellipsoid of loosening
Evidently, when the apex point N arrives at the outlet, measured at the height of the apex point (N) of the
the entire volume of the extraction ellipsoid has been extraction ellipsoid. The volume of the outflow funnel
discharged. The gravity flow mechanism is shown is the same as volume of extraction ellipsoid. For a
diagrammatically in Fig. 12a, b, c and d, which is certain extraction height h n , the mechanism of gravity
identical to the model test of Fig. 11. From Fig. 12d, flow is defined by explicit relationships between the

Fig. 11. Ellipsoid of extraction: (a) represents the model prepared for extraction (b), (c) and (d) depict successive phases showing from the
extraction ellipsoid and simultaneous development of the ellipsoid of loosening.
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 235

discharged material, the ellipsoid of extraction, the dilution) is only obtained from the extraction ellipsoid
outflow funnel and the ellipsoid of loosening. with height h n . Any additional material discharge will
Let V C be the volume of discharged material, E E result inevitably in a rapid increase in dilution because
the extraction ellipsoid, V EE the volume of the extrac- the outflow funnel, equal to the discharged extraction
tion ellipsoid, h n the height of the extraction ellipsoid, ellipsoid of height h n is filled by waste rock.
E L the ellipsoid of loosening, V EL the volume of the
ellipsoid of loosening, h L the height of the ellipsoid of 4.3. Specific application of sublevel caving to thick
loosening, F the outflow funnel and V F the volume of coal seams
outflow funnel. By introducing certain simplification,
the characteristic relationships for gravity flow can be For a vertical sublevel front, the sublevel drift
expressed by the following formulas (Julin and Kva- forms a vertical opening, located in the plane of the
pil, 1982). sublevel front. Therefore, on a vertical section, the
geometry of sublevel caving is similar to a bin with
VEE iVC ð2Þ
discharge openings located at a level above the bot-
tom. The gravity flow zone is cut off by the vertical
VEE iVC iVF ð3Þ
bin wall but is otherwise unchanged. This means that
the vertical wall cuts off the ellipsoid of extraction and
VEL i15VEE i15VC i15VF : ð4Þ
loosening as shown in Fig. 13, the axis of gravity flow
Assuming approximately the same eccentricity of on this section deviates from vertical by a certain
the ellipsoid of loosening and the extraction ellipsoid angle D. This angle increases as the friction along
and knowing that V EL i 15V EE then the vertical wall increases. Neglecting this deviation,
one can assume that the vertical wall with a discharge
hL 2:5hn : ð5Þ
opening cuts off half of the ellipsoid of extraction and
Hence, the height h L of the ellipsoid of loosening is loosening. Naturally, when this half of the ellipsoid is
about two and a half times bigger than the height h n of inscribed in a prism, then its volume is 50% of this
the extraction ellipsoid. These relations form the basis prism. Therefore, in sublevel caving, if half of the
for understanding, explaining and justifying the dif- extraction ellipsoid is inscribed in the body of coal,
ferent phenomena involved in gravity flow of granular then a maximum of 50% of coal volume can be
material. extracted without dilution.
The maximum volume of pure ore which can be In sublevel caving applied at a thick coal seam,
discharged without any waste rock (i.e., without any the opening dimensions used for top coal drawing

Fig. 13. Schema of ellipsoid of extraction and loosening when material is extracted through vertical outlet in sublevel caving.
236 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

are limited by the size of the shield window. In order presented in flowsheet form in Fig. 14 (Itasca,
to maintain an efficient flow of the top coal without 1997).
dilution, the dimensions of shield windows should
be as small as possible; whereas they should be 5.1. Model geometry and meshing
adequately large enough for the proper flow of top
coal. Shield with windows of 80-cm  150-cm Steps of 3-D modelling of M3 longwall face with
dimensions are used in the Omerler Underground FLAC3D are given below:
Mine. In sublevel caving, the vertical discharge
opening (width of the window of shield) is con- 1. In order to obtain proper results from models
structed according to sublevel width (face width) formed to analyze caving of the top coal, only a
and this opening must be wider than the largest small part of the face was modeled. This consisted
coal block drawn from top. of the area of two shields located at the face center.
Hence, application of a tighter mesh in this region
facilitated the effective modelling of stresses and
5. General modelling procedure displacements.
2. Only the area affected by flow was constructed.
Modelling was carried out with a widely used 3. Soft claystone and a 5-m section of roof claystone
numerical software program called FLAC3D which is above the coal were added to the model.
used for stress and deformation analyses around sur- 4. In order to properly observe the flow of top coal
face and underground structures opened in both soil caving, mesh size on the model was very tight in 3
and rock. The software is based on the finite difference main directions (x, y, z). The model was composed
numerical method with the Langragian calculation of 42 000 elements and 45 756 gridpoints.
method. FLAC3D is a commercially available software 5. Shield supports at the face were modeled in the
that is capable of modelling in three dimensions. form of elastic material, as shown in Fig. 15.
Modelling of the longwall with top coal caving is
performed in five steps. Steps identified as A, B, C, D 5.2. Assessment of material properties
and E are described as follows:
The geological units in Omerler Underground
A Determination of boundaries and material Mine are calcareous marl, marl, roof claystone, soft
properties, claystone, coal and floor claystone (Fig. 4). The phy-
B Formation of the model geometry and meshing sical and mechanical data of rock materials (Table 1)
Determination of the model behavior, obtained from laboratory test were not used directly in
C Determination of the boundary and initial the model. However, these data were converted to
conditions, rock mass data in Table 2 (Yasitli, 2002; Unver and
Initial running of the program and monitoring of Yasitli, 2002; Yasitli and Unver, 2003) by using the
the model response, Hoek–Brown Failure Criterion (Hoek and Brown,
D Re-evaluation of the model and necessary 1977), RMR (Bieniawski, 1973; Bieniawski, 1989),
modifications, the GSI System (Hoek, 1995; Sonmez, 2001; Sonmez
E Obtaining of results. and Ulusay, 1999).

Model geometry and meshing refer to physical

conditions of the district to be modeled. Model 6. Numerical modelling
behavior is the response of a model under a certain
loading condition. By means of boundary and 6.1. Caving of top coal through window of a single
initial conditions, physical limits of the model and shield
original conditions are explained. Gate roadways,
the face and other structures were later created in Caving of top coal through the window of a
the form of modifications. The modelling process is single shield was modeled first and then the same
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 237

Determination of
the general layout to be
modelled

Determination of boundaries and material properties

Model Preperation
1. Meshing
2. Definition of material behavior
3. Definition of boundary and initial condition

Running until a condition of equilibrium

is reached

Monitoring
of model
Results are unsatisfactory response

Suitable model
PROCESS OF RE-EVALUATION

Modification in the model

1. Formation of excavation
2. Change of boundary conditions
3. Top coal caving through single and
double window of shield

Running for a complete solution

Monitoring
of model
response
Need for more trials

Acceptable results

Yes Is there a need

for further
parameters?

No

FINAL OUTPUT

Fig. 14. A general flowsheet of modelling process (Yasitli, 2002; Unver and Yasitli, 2002; Itasca, 1997).

modelling was performed for two adjacent shields. respectively. Each color (or tone) corresponds to a
Fig. 16a and b present the results of modelling of top certain amount of vertical displacement in the fig-
coal caving for a single shield along AAV and BBV ures. Displacement values are given in meters such
cross-sections perpendicular and parallel to the face, as 1.95e001 meaning that the region of this color
238 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Shear-n The region failed under shear loading and

failure process is still in progress.
Shear-p The region failed under shear loading and
failure process is ceased due to lowered
amount of shear forces.
Tension-n The region failed under tensile loading and
failure process is still in progress.
Tension-p The region failed under tensile loading and
failure process is ceased due to lowered
amount of tensile forces.
The modelling results reveal that a complete fail-
ure (shear-n) takes place up to the first 1.5-m-thick
Fig. 15. Flowing model constructed in FLAC3D. (To view this figure layer just above window of the shield. Above this
in colour, the reader is referred to the web version of this article.) level, vertical deformation is observed occasionally
for some time and stops as a result of lowered shear
loads (shear-p). Characteristics of the failure phe-
corresponds to a vertical displacement amount of nomenon obtained from modelling show that there
195 mm. The active zone and the boundary of the would be no problem during caving and drawing of
ellipsoid of motion are shown in Fig. 16a and b. the first 1.5 m from the shield top. The coal in this
Similar to the case shown in Fig. 16a, the amount of region is well fractured as a result of failure. How-
vertical displacement is the highest at shield window, ever, above this well-fractured section, the coal does
decreasing considerable away from this region. The not fail completely. It means that the coal would be
region of high vertical displacement is called as the broken into larger pieces starting from 1.5 m above
ellipsoid of motion and the region outside the ellip- the face roof. As a result, larger pieces would block
soid of motion where lower displacements are the shield window preventing a continuous flow of
observed called as the active zone and as the name the top coal. This phenomenon is well observed at in
implies, the region that is not affected is called as the situ conditions. In the mine, there was no problem at
passive zone. As coal is drawn through the window the beginning of coal drawing. Later, as a result of
of a shield, the top coal in the active zone enters in increased block size in the drawn coal, the possibility
the ellipsoid of motion until all of the top coal is of blockage increased. Whenever such a problem is
produced. faced within the mine, the block obstructing the
The top coal failure mechanism during top coal window was blasted and then drawing operation
drawing is presented in Fig. 17. Explanations of terms resumed.
given in the legend are as follows: In Fig. 17, there seems to be no deformation (none)
behind the face. The reason of this phenomenon may be
None No-failure zone. attributed to the fact that the material behind the face in

Table 2
The input parameters regarding rock mass used in numerical modelling (Yasitli, 2002; Yasitli and Unver, 2003)
Rock definition Calcareous marl Marl Roof claystone Soft claystone Coal Floor claystone
3
Density (d, MN/m ) 0.023 0.022 0.025 0.023 0.014 0.027
Internal friction angle (/, 8) 27.5 24.8 18.8 14.0 21.8 18.4
Cohesion (c, MPa) 1.3 0.65 0.41 0.167 0.517 0.715
Tensile strength (MPa) 0.096 0.074 0.031 0.006 0.017 0.035
Bulk modulusa (K, MPa) 2269 947 698 497 637 1206
Shear modulusb ( G, MPa) 1362 568 360 298 382 525
a E
K¼ 3ð12vÞ .
b E
G¼ 2ð1þvÞ .
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 239

I
A-A Cross-Section
(m)

COAL

AI

GOAF
B B I
FACE
I
B-B Cross-Section

(a) Isometric view

COAL
Active zone

Ellipsoid of 100 cm
extraction

115 cm

Boundary of
single shield
GOAF
(b) Front view

Fig. 16. Modelling of top coal caving through single window of shield. (To view this figure in colour, the reader is referred to the web version of
this article.)

the goaf is defined as failed material at the beginning of to the face are given in Fig. 18a and b, respectively. As
modelling. Since stress build up in the failed goaf can be seen in Fig. 18, during the top coal caving
cannot be high enough, no post failure is observed in through the windows of the two adjacent shields, ver-
this region as expected. tical displacement is the highest at window discharge
regions like caving through the window of a single
6.2. Caving of top coal through windows of two shield, whereas outside this area the displacements
adjacent shields decrease. Part of the top coal between the two windows
is not affected during drawing. This stationary part is
Caving of top coal through windows of two adjacent called the dead zone. In Fig. 18, the boundary of
shields was modeled and the results of modelling along shields, ellipsoid of extraction, active, passive and
AAV and BBV cross-sections perpendicular and parallel dead zones are indicated.
240 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Fig. 17. State of failure in top coal during caving through window of a single shield. (To view this figure in colour, the reader is referred to the
web version of this article.)
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 241

(a) Isometric view

(m)
COAL

FACE

GOAF

(b) Front view

COAL

Active zone

230 cm
100 cm 100 cm
Coal dilution zone

Extraction ellipsoid
120 cm

Dead zone

Boundary of shields

GOAF

Fig. 18. Modelling of top coal caving through windows of two adjacent shields. (To view this figure in colour, the reader is referred to the web
version of this article.)

Modelling of top coal caving through windows result of high vertical and shear stresses (shear-n).
of two adjacent shields has revealed that, up to 2 Above this part, failure is observed occasionally for
m of top coal above the shields was fractured as a some time and stops as a result of decreasing shear
242 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Fig. 19. State of failure in top coal during caving through window of two adjacent shields. (To view this figure in colour, the reader is referred to
the web version of this article.)
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 243

loads (shear-p) (Fig. 19). Therefore, the coal is not the coal was increased, leading to an increase in
fractured completely. In other words, coal would be dilution and decrease in productivity and production
broken into larger pieces starting from 2 m above rate. In order to decrease dilution and increase the rate
the shield. As a result, larger blocks would obstruct of extraction, the top coal has to be fractured as
the shield windows preventing a continuous flow uniformly as possible and the shield windows have
of top coal similar to the flow of the top coal to be sequentially opened.
through window of a single shield. During the top
coal caving, flowing takes place in an ellipsoidal 6.3. Suggested method for top coal caving
shaped region. Hence, two ellipsoids intersect dur-
ing the extraction of the upper part of top coal The upper part of the thick coal seam at the
through the window of the two adjacent shields, Omerler Region is stronger than the lower part. As
thus roof rock (claystone) would be mixed with the a consequence, formations of large blocks were
intersected region leading to dilution of the top observed during caving of the upper part of the
coal. coal seam in the mine. This situation was also
The width of a shield, window and the distance verified with the numerical modelling results. As
between two adjacent windows were 150 cm, 80 cm numerical modelling results have indicated, the
and 70 cm, respectively. As can be seen in Fig. 18, the upper part of the top coal was not uniformly
coal between two adjacent windows (70 cm) cannot fractured and, consequently, during the flow of
flow to the discharge due to being outside of ellipsoid the top coal, the shield windows were usually
of extraction. The coal at this region forms a dead obstructed by large blocks. In the case of obstruc-
zone. The dilution due to this region can be expressed tion of the shield windows in the mine, large
as dead zone dilution (Fig. 20). blocks were broken into small pieces either by
Top coal caving through shield windows was drilling and blasting or by moving the relevant
performed in one shift at the Omerler Underground shield up and down. The length of the blastholes
Mine. During top coal caving, adequate fastidiousness was between 3 and 6 m. Drilling and blasting
was not applied and the orders of top coal flowing seemed to temporarily solve the obstruction pro-
through the shield windows were not observed by blems. However, large blocks are fractured into
workers. For these reasons, shield windows were not irregular shapes with a wide range of particle
opened in a proper order. When the top coal caving sizes by blasting. As can be seen in Fig. 5, the
through one window was finished, the adjacent win- flowing angle for coal with a wide range of particle
dow was opened. Therefore, mixing of waste rock in sizes is higher than the angle for coal containing

Fig. 20. Coal dilution during top coal caving.

244 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

particles of uniform size. During the flow of top front of the face. The blasting should be carried out
coal through the window of a shield, coal having a only for forming cracks in the top coal to enhance
wide range of particle sizes not only results in both caving characteristics of the top coal located at 3–
mixing of the coal with the goaf but also dilution 3.5 m above the face. In other words, the objective
of surrounding rock into the coal. Being a time- of blasting is not to fracture the top coal completely
consuming operation, blasting of large blocks but to form cracks that would maintain the forma-
would not enhance flowing characteristics of top tion of a regular particle size distribution during
coal; to the contrary, productivity is significantly caving. In this manner, the possibility of existence
decreased. In order to maintain a high rate of of large blocks in the top coal will be decreased,
advance with an increased rate of recovery, forma- leading to formation of top coal having a particle
tion of large blocks should be prevented as much size that can easily flow through the shield win-
as possible. This can only be accomplished by dows. The quality of coal produced behind the face
means of pre-fracturing or pre-cracking of top will be improved and the extraction ratio of reserve
coal prior to the caving operation. In order to will also be increased due to less dilution. Besides,
minimize time-consuming and tedious operation of blasting carried out in the top coal in front of the
the large block blasting method, to increase produc- face would be carried out independently from
tivity and daily advance rate and to decrease dilu- operations carried out at the face. Thus, the time
tion, large blocks must be fractured into small required for blasting through the shield windows
pieces by the application of pre-fracture blasting will be eliminated and consequently productivity
carried out at the boundary of the top coal; conse- would also be increased.
quently, providing a more uniform particle size
distribution in the top coal prior to caving. The 6.4. Modelling of top coal caving through the window
proposed method of blasting must be carried out of a single shield after pre-fracture blasting
at the boundary of the top coal, not at the face. The
objective of this operation is not to fracture the top Caving of top coal through the window of a single
coal completely, but to provide a more uniform size shield after blasting has been modeled and Fig. 22
distribution in the coal as a result of pre-fracture presents modelling results of top coal caving. The
blasting (Fig. 21). It is suggested to open inclines boundary of a shield, ellipsoid of extraction and active
from the maingate and tailgate to the upper bound- zone are shown in Fig. 22b. As can be seen in Fig. 22,
ary of the top coal at a distance of 20–25 m in displacement is the highest at the window discharge.

Maximum fractured
coalsize 80 cm Blasting holes

Pre-fractured
3,5 m zone

5m Easily fractured
zone

Hangingwall Footwall

Fig. 21. Suggested blasting design.

 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 245

(a) Isometric view

(m )
COAL

GOAF FACE

(b) Front view

COAL

Active zone 120 cm

Ellipsoid of extraction
175 cm

Boundary of a shield
GOAF

Fig. 22. Modelling of top coal caving through single window of a shield after pre-fracture blasting. (To view this figure in colour, the reader is
referred to the web version of this article.)

In other words, the flow rate of the top coal is higher The modelling results reveal that during top coal
at this region but outside this region the displacements flowing through the window of a shield, coal above
are low. the shield was fractured very well (shear-n) as a result
246 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Fig. 23. State of failure in top coal during caving through window of a single shield after pre-fracture blasting. (To view this figure in colour, the
reader is referred to the web version of this article.)

of decreasing the mechanical properties of coal after top coal can be continuously drawn through the win-
pre-fracture blasting (Fig. 23). It means that, the coal dow of a shield, dilution will be reduced and problems
was fractured as uniformly as possible and formation encountered due to obstruction of windows would not
of large blocks was not allowed. Hence, all parts of be faced.
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 247

(a) Isometric view

(m)

COAL

FACE

GOAF

(b) Front view

COAL

Active zone
250 cm

Coal dilution zone

200 cm
Ellipsoid of
extraction

Dead zone

Boundary of shields GOAF

Fig. 24. Modelling of top coal caving through windows of two adjacent shields after pre-fracture blasting. (To view this figure in colour, the
reader is referred to the web version of this article.)
248 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Fig. 25. State of failure in top coal during caving through windows of two adjacent shields after pre-fracture blasting. (To view this figure in
colour, the reader is referred to the web version of this article.)
 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 249

6.5. Modelling of top coal caving through windows of through the window of a single shield after pre-
two adjacent shields after pre-fracture blasting fracture blasting.

Caving of top coal through windows of two adja- As can be seen in Fig. 26, although the width of the
cent shields after pre-fracture blasting has been mod- ellipsoid of extraction was not changed, the height of
eled and Fig. 24 presents the results of this model. The the ellipsoid of extraction was increased by means of
modelling results reveal that the coal above shields is pre-fracture blasting. The phenomenon can be
fractured and drawn through the windows completely explained by the eccentricity of coal (Eq. (1)). If the
(shear-n). In other words, the coal above the shields eccentricity is high, the ellipsoid of extraction will
was fractured as uniform as possible and it seems that become narrow; if the eccentricity is low, the ellipsoid
there would be no problem during top coal caving of extraction will become wider.
through windows. If ellipsoids of extraction obtained from modelling
During top coal caving after pre-fracture blasting, results are compared with each other, it can be seen that
the coal between two adjacent windows could not the height/width ratio (h/w) of ellipsoids of extraction is
flow due to ellipsoidal motion, which causes decreas- low due to wide range of particle size distribution of top
ing coal extraction ratio and dilution of coal (Fig. 25). coal before pre-fracture blasting. However, h/w ratio is
higher after pre-fracture blasting because of more uni-
6.6. Evaluation of modelling results of longwall with form particle size distribution of coal. These results
top coal caving method reveal that the amount of flowing material through
shield windows will be high for a given unit time.
After four modelling studies for the longwall with Results obtained from all models show that, 1.5-m-
top coal caving method, the results given below were thick coal just above the shield supports are well frac-
found: tured. However, modelling of coal flowing through a
single shield window, the coal above 1.5 m and for coal
! The width and the height of the ellipsoid of extrac- flowing through two adjacent shields, the coal above 2
tion were 100 cm and 115 cm, for flowing of top m is not fractured completely. It means that, the coal
coal through the window of a single shield. would be broken into large pieces starting from 1.5 to 2
! The width and height of the ellipsoid of extraction m above the shield canopy. For this reason, there would
were 120 cm and 175 cm, for flowing of top coal be some problems such as the interruption of produc-
through window of a single shield after pre-fracture tion due to obstruction of shield windows at the face.
blasting. Existence of this phenomenon can be eliminated with
! The width and height of the ellipsoid of extraction the application of pre-fracture blasting.
were 230 cm and 120 cm, for flowing of top coal During modelling of top coal flowing through win-
through windows of two adjacent shields. dows of two adjacent shields, a dead zone is formed
! The width and height of the ellipsoid of extraction between two windows leading to dilution of the top
were 250 cm and 200 cm, for flowing of top coal coal with surrounding rock. In order to decrease dilu-

b: width

a: height
250 cm
120 cm

100 cm 230 cm

115 cm 200 cm
175 cm 120 cm

Fig. 26. The width and height of ellipsoid of extraction before and after pre-fracture blasting.
250 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

tion and increase face advance, large blocks have to be be prevented as much as possible and the top coal
fractured into small pieces with pre-fracture blasting should be fractured as uniformly as possible. Produc-
carried out at the boundary of top coal, consequently, tion of top coal by means of drawing it through
providing a uniform particle size distribution. Blasting windows located in the canopy of the shields should
has to be carried out at the boundary of top coal not at be performed in a predetermined sequence as shown
the face and the objective of this operation is not to in Fig. 27. Although there were 60 shields at the M3
fracture the top coal completely, but to provide a uni- longwall face, only 10 of them are shown in Fig. 27
form size distribution in the coal as a result of pre- for the sake of simplicity.
fracture blasting. After pre-fracture blasting, the top coal flow
through the shield windows can be carried out in
6.7. Suggested top coal caving order three ways as follows:

As stated earlier, the formation of large blocks a. Top coal flowing through windows starting either
having a potential to obstruct shield windows should from face head to face end or vice versa.

5m
TOP COAL

FACE

Blasting

5m
TOP COAL

FACE

a11 a12 a13 a14 a15 a16 a17 a18 a19 a20
b11 b13 b15 b17 b19 b20 b18 b16 b14 b12
2,5 m c19 c17 c15 c13 c11 c12 c14 c16 c18 c20
a1 a2 a3 a4 a5 a6 a7 a8 a9 a10
b9 b7 b5 b3 b1 b2 b4 b6 b8 b10 1st Slice
c1 c3 c5 c7 c9 c10 c8 c6 c4 c2

FACE

ROOF CLAYSTONE

a11 a12 a13 a14 a15 a16 a17 a18 a19 a20
2,5 m b11 b13 b15 b17 b19 b20 b18 b16 b14 b12 2nd Slice
c19 c17 c15 c13 c11 c12 c14 c16 c18 c20

FACE

a. Top coal flowing starting either from face head to face end or from face end to face
head.
b. Top coal flowing starting either from face center to face head or face end.
c. Top coal flowing starting either through window from face head or face end to face
center.

Fig. 27. Designed top coal caving order.

 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252 251

b. Top coal flowing through windows starting either dows can be performed effectively. As a consequence,
from face center to face head or face end. dilution can be reduced leading to an increase in
c. Top coal flowing through windows starting either productivity.
from face head or face end to face center. As a final conclusion, in spite of having a relatively
high investment cost, pre-fracture blasting technique
The top coal with a uniform particle size distribu- has to be performed above the face at the boundary of
tion as possible after pre-fracture blasting operation coal to obtain a uniform size distribution in the top
has to be drawn as fast as possible with great care. It is coal prior to production. It is determined that the top
considered that there would be any major problem coal has to be produced in the form of sequential
during application of the methods given in Fig. 27. In slices. These operations will certainly reduce dilution,
order to flow top coal rapidly, two crews have to be increase production and face advance rate.
organized. Simultaneously, first team can start flowing
top coal from face head, and the second team can start
Acknowledgements
from face end or both team can start flowing from face
center to face head and face end by drawing half of
The results of this paper are based on a science and
the top coal while in advance and the other half while
research project funded by Hacettepe University,
in return to starting point (Fig. 27). The flow of top
Scientific Research Unit. The authors appreciate the
coal through either a single or multiple shield win-
cooperation and help of Omerler Colliery staff during
dows does not effect dilution too much and the top
field observation.
coal at the dead zone could be drawn by moving
shields canopy up and down.
References

7. Conclusions Ahcan, R., 1977. Mechanization and concentration of thick coal

seams mining in SFR Yugoslavia. Proc. of Int. Symp. on Thick
Seam Mining, Dhanbad, India, Paper, no. 4.
In this study, 3-D modelling of a longwall with the Atkinson, T., 1979. Thick, steep and irregular coal seam mining.
top coal caving method at the Omerler Underground The Mining Engineer, 421 – 441.
Mine was carried out. According to modelling results, Bieniawski, Z.T., 1973. Engineering classification of jointed rock
1.5–2 m parts of the top coal from the shield top are masses. Transactions of the South African Institution of Civil
well fractured at a particle size that can flow easily Engineers 15, 335 – 344.
Bieniawski, Z.T., 1989. Engineering Rock Mass Classifications: A
through the shield windows. However, the coal above Complete Manual for Engineers and Geologists in Mining, Civil
this fractured zone either could not be fractured at all or and Petroleum Engineering. Wiley, New York, p. 251.
fractured in the form of large blocks. Results of numer- Destanoglu, N., Taskin, F.B., Tastepe, M., Ogretmen, S., 2000.
ical modelling coincided with the in situ observations Application of GLI Tunable-Omerler Underground Mechaniza-
carried out in the mine. In the M3 panel, when shield tion. Kozan Publishing, Ankara, p. 211 (in Turkish).
Dian, C., 1992. The state of the art and future of China thick seam
windows were obstructed with large blocks, these mining technology. In: Singh, T.N., Dhar, B.B. (Eds.), Proc. of
blocks were fractured by means of drilling and blasting. Int. Symp. on Thick Seam Mining. Central Mining Research
At first glance, this operation seems to solve the pro- Station, Dhanbad, India, pp. 171 – 182.
blem. However, blasting lead to the formation of a wide Garg, P.C., Nath, P.D., 1977. Choice of methods for mining thick
coal seams in India. Proc. of Int. Symp. on Thick Seam Mining,
range of particle sizes in the top coal. Hence, a portion
Dhanbah, India, Paper, no. 10.
of the top coal is lost in the goaf and part of the roof Hoek, E., 1995. Strength of rock and rock masses. ISRM News
claystone is mixed with the coal causing dilution. Journal 2 (2), 4 – 16.
Pre-fracture blasting, which was designed to frac- Hoek, E., Brown, E.T., 1977. Practical estimates of rock mass
ture top coal further than 3–3.5 m from the shield top strength. International Journal of Rock Mechanics and Mining
is not aimed at fracturing top coal completely, but Sciences 34 (8), 1165 – 1186.
Itasca,, 1997. User manual for FLAC3D Ver. 2.0. Itasca Consulting
forming cracks in this part leading to formation of a Group Inc., Minnesota.
uniform size distribution. After application of pre- Jeremic, M.L., 1985. Strata Mechanics in Coal Mining. A.A. Balk-
fracture blasting, top coal caving through shield win- ema, Rotterdam, p. 556.
252 B. Unver, N.E. Yasitli / International Journal of Coal Geology 66 (2006) 227–252

Jha, S.N., Karmakar, S., 1992. Thick seam mining—some experi- Sonmez, H., 2001. Investigation on the applicability of the Hoek–
ence and exaltation. In: Singh, T.N., Dhar, B.B. (Eds.), Proc. of Brown criteria to the failure of the fissured clays. PhD thesis,
Int. Symp. on Thick Seam Mining. Central Mining Research Hacettepe University, Ankara, p. 215 (in Turkish).
Station, Dhanbad, India, pp. 191 – 202. Sonmez, H., Ulusay, R., 1999. Modifications to the Geological
Julin, P., Kvapil, R., 1982. Caving methods. In: Hustrulid, W.A. Strength Index (GSI) and their applicability to stability of
(Ed.), Underground Mining Methods Handbook. Society of slopes. International Journal of Rock Mechanics and Mining
Mining Engineers (SME), New York, pp. 871 – 893. Sciences 36 (6), 743 – 760.
Kvapil, R., 1992. Sublevel caving. SME Mining Engineering Hand- Taskin, F.B., 1999. Optimum dimensioning of pillars between long-
book, pp. 1789 – 1814. wall panels in Tuncbilek Mine, PhD thesis, Osmangazi Univer-
Kvapil, R., 1998. The mechanics and design of sublevel caving sity, Eskisehir, p. 149 (in Turkish).
systems. In: Gertsch, R.E., Bullock, R.L. (Eds.), Techniques in Tien, J., 1998. Longwall caving in thick seams. Coal Age 103 (4),
Underground Mining. SME Inc, Littleton, USA, pp. 621 – 654. 52 – 61.
Schweitzer, R., 1977. Thick seam winning methods in French coal Unver, B., 1995. A practical approach to strata control and caving
mines. Proc. of Int. Symp. on Thick Seam Mining, Dhanbah, mechanism in thick seam mining. 14th Mining Congress of
India, Paper, no. 3. Turkey, Ankara, pp. 15 – 22 (in Turkish).
Senkal, S., Kose, H., Ermisoglu, N., 1988. A study about coal losses Unver, B., Yasitli, N.E., 2002. The Simulation of Sublevel Caving
and dilution problems in the mining method applied at GLI Method Using in Thick Coal Seam by Computer. Hacettepe
Tuncbilek Colliery. Mining Journal of Turkish Mining Society University Scientific Research Unit, pp. 148 (Project No: 00 02
27 (4), 5 – 11. 602 008 (in Turkish)).
Singh, R., 1999. Mining methods to overcome geotechnical Wu, J., 1992. The movement regularity of the roof coal around a
problems during underground working of thick coal longwall with caving and its coal recovery. Proc. of Int. Symp.
seams—case studies. Transactions-Institution of Mining and on Fully Mechanized Mining Technology for High Output and
Metallurgy. Section A. Mining Industry 108, A121 – A131 Efficiency, pp. 307 – 318.
(May–August). Yasitli, N.E., 2002. Numerical modelling of longwall with top coal
Singh, T.N., Kushwaha, A., Singh, R., Singh, R., 1992. Strata caving. M.Sc. thesis, Hacettepe University, Ankara, p. 148 (in
behavior during slicing of thick seam at East Katras Colliery. Turkish).
In: Singh, T.N., Dhar, B.B. (Eds.), Proc. of Int. Symp. on Thick Yasitli, N.E., Unver, B., 2003. 3-D estimation of stresses around a
Seam Mining. Central Mining Research Station, Dhanbad, longwall face by using finite difference method. 18th Interna-
India, pp. 237 – 250. tional Mining Congress of Turkey, Antalya, pp. 83 – 88.