ENGINEERING

GEOLOGY
ELSEVIER Engineering Geology 39 (1995) 123-136

Textural characterisation of rocks

A. Ersoy, M.D. Waller *
Department of Mineral Resources Engineering, The University of Nottingham, University Park, Nottingham, NG7 2RD,
UK
Received 2 August 1994; accepted 30 January 1995

Abstract

Textural characteristics are a major factor in determining the mechanical behaviour of rocks and in the prediction
of performance of rock cutting and drilling equipment. The principal textural characteristics of rocks are grain size,
grain shape, grain orientation, relative proportion of grains and matrix material which were herein quantitatively
measured using a modern image analysis system. These features resulted in a texture coefficient represented by a
single number for each rock specimen. In this study a range of both textural and mechanical data for a range of rock
types are given with some of the textural determination methodology. Correlation between textural and physical
properties are also highlighted. The results of drilling tests using polycrystalline diamond compact (pin and hybrid)
and impregnated diamond core bits in the rocks are presented which demonstrate the influence of rock texture on
drillability. The rock texture can be used as a predictive factor for assessing the drillability and cuttability, mechanical
and wear performance of rocks.

1. I n t r o d u c t i o n Table 1
Classifications of rock characteristics for drillability
In all areas o f rock engineering, measurement Textural characteristics
o f the rock properties is essential in determining Grain size, grain shape, degree of grain orientation, packing
the behaviour o f the rock mass. In general, the density, relative proportion of grains, texture coefficient, min-
rock properties m a y be classified into four cate- eral content, matrix material and type, cement type and degree
gories according to drillability, cuttability and of cementation, porosity, grain boundary or grain contact rela-
tionships, bonding structure
machinability: that is textural, mechanical, struc-
tural and weathered characteristics, as shown in Mechanical characteristics
Strength, hardness, abrasiveness, density, pore pressure
Table 1. The first two categories are tightly inter-
related. The textural characteristics o f rocks sig- Struct~al characteristics
nificantly affect the mechanical performance o f a Joints, fracture, cleavages, foliations, faults, folds bedding,
banding; dip and strike
rock, drillability, cuttability and machine perfor-
mance. In rock engineering, the selection o f equip- Weathered characteristics
Alteration and water content
ment and o p t i m u m operation depends to a large
extent on the quality and quantity o f textural and
mechanical data available for the rock.
R o c k texture has been defined as "the degree o f crystallinity, grain size or granularity and the fabric
or geometrical relationship between the constitu-
*Corresponding author. ents o f a r o c k " (Williams et al., 1982). Here, the

0013-7952/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved

SSDI 0013-7952 (95) 00005-4
124 A. Ersoy. M.D Waller/Engineering Geology 39 (1995) 123 136

textural characteristics refer to the geometrical were taken from drill cores or hand specimens.
features of rock particles such as grain size, grain Two thin sections were made for each rock, one
shape, grain orientation, relative areas of grain parallel and one perpendicular to the core plane.
and matrix (packing density) and compositional These thin sections help in detecting any three-
features such as mineral content, cement type and dimensional fabric formed in the crystals during
degree of cementation or crystallisation and bond- crystallisation and cooling of igneous rocks or
ing structure. Although grain boundary relation- during sedimentation and lithification of sedi-
ships also play an important role in crack mentary rocks. The sections were examined under
propagation and mechanical performance (partic- a polarising microscope to determine the mineral
ularly in igneous rocks), little specific literature constituents contained within the rock, their pro-
exists in this area. The grain boundary or contact portions and relationships to each other. Grain
relationships are complex and have been examined shape, bonding structure, type and degree of
elsewhere (Bieniawski, 1967; Stanton, 1972; Sprunt cementation or crystallisation were also described.
and Brace, 1974). Numerous investigations have The percentage mineral analysis was determined
been conducted on the relationship between a using a point counter with an average of 3000
single parameter of rock texture and mechanical counts from two thin sections of each rock. The
properties of rock (Smordinov et al., 1970, Ehrlich total silica content for the rocks were determined
and Weinberg, 1970; Dube and Singh, 1972; by using a automated X-ray fluorescence spectro-
Olsson, 1974; Bell, 1978; Irfan and Dearman, 1978, photometer (Phillips ® PW 1400), as described by
Hugman and Friedman, 1979; Onodera and Asoka Harvey and Atkin (1983). The quartz content is
Kumara, 1980; Barret, 1980; West, 1981,West, the "free" silica content of the rock which was
1986; Howarth and Rowlands, 1986,Howarth and determined by thin section examination. The total
Rowlands, 1987; Hawkins and McConnell, 1991 ). silica content of the rock includes the free silica
However, little attention has been given to assess- (quartz) and the silica content of the silicate miner-
ing quantitatively the whole rock texture (as a als such as olivine, pyroxene, amphibole, feldspar,
single parameter) and its relationship with mechan- mica, clay minerals, epidote, garnet, etc.
ical performance and drillability or cuttability of A range of rock types were used during this
rocks. study, their geographic origin and stratigraphic
The sedimentary and igneous rock types used position, mineral constituents and proportions,
for drilling in this research were specially selected total silica content and summarised petrographic
for this study. The main objectives of the study description are presented in Table 2. In the table,
can be summarised as follows: the terminology used for the description of weath-
(a) To describe and determine mineralogical and ering, bonding structure, interlocking texture and
compositional features of the rocks. grain boundaries of the rocks is according to Irfan
(b) To measure the geometrical features of the and Dearman (1978). The details of the rocks
rock particles which result in texture coefficient. description can be found in Ersoy (1995).
(c) To present a wide range of mechanical and
intact rock properties and to establish links
between these and the rock texture. 3. Mechanical and intact properties of the rocks
(d) To illustrate the effect of texture coefficient on
the performance of polycrystalline diamond com- Knowledge of the mechanical and intact proper-
pact (PDC) and impregnated diamond core bits. ties of rocks and strata is essential to rock engineer-
ing. A wide range of these properties were
measured including mechanical, hardness and
2. Mineralogical examination of the rocks abrasivity index. The mechanical properties were
uniaxial compressive strength (UCS), indirect
Mineralogical studies were based on the exami- (Brazilian) tensile strength (BTS), Young's
nation of thin sections of rock specimens which Modulus and density which were all determined
 A. Ersoy, M.D. Waller~EngineeringGeology 39 (1995) 123-136 125

Table 2
Description of the rocks

Geographic location and stratigraphic position of the rocks

Mineral Mineral proportion Total silica content Petrographic description of the rocks
(%) (%)

Limestone (chalk),Italy,age unknown

Calcite 94 0.00 Soft, fossiliferous, porous, composed mainly of coarse
Opaque traces crystalline sparite, but cemented by fine crystalline micrite,
Voids 6 grains loose, not interlocked, weakly bonded.
Sandstone (Ridgeway quarry) Ambergate, Derbyshire, ~riassic
Quartz 79 91.22 Partially weathered (physical weathering no mineral com-
Feldspar 9 position change), coarse grained, porous, abrasive, angu-
Clay 5 lar and subrounded grains, loosely cemented mostly by
Opaque 1 clay and by detrital quartz, grains loose not interlocked,
Voids 6 weakly bonded.
Siltstone Beryhill, Mansfield, Nottingham (coal measures rock), Permo-Triassic
Quartz 59 81.62 Slightly weathered (physical weathering without mineral-
Feldspar 7 ogical change), fine grained, grains moderately sorted and
Clay 24 rounded, well cemented by quartz, opaque phases domi-
Mica 3 nant in the weathered part moderately to strong bonded
Rock fragments 3 with tight grain boundaries.
Opaque 2
Voids 2
Cornish Grey Granite (Penryn), Cornwall, Penno-Carboniferous (Hercynian)
Quartz 34 72.38 Hard, holocrystalline, coarse grained, elongated and very
Feldspar 49 angular crystals, interlocking texture with tight grain
mica 14 boundaries, strongly bonded or minerals linked by a
Chlorite 3 strong ionic bonding between the crystal faces.
Opaque negligible
Diorite (Ebony), South Africa stratigraphic position unknown
Plagioclase 51 51.57 Very hard, microcrystalline, angular and mostly anhedral
Pyroxene 41 crystals, extremely strong bonding, interlocking texture,
mica 2 very tight crystal boundaries.
Olivine 3
Opaque 3

according to the I S R M (Brown, 1981) suggested Where tr¢ is the stress value at failure, P is the
methods. Uniaxial compressive strength is the m o s t failure load ( N ) and A is the area o f specimen
widely used index o f the strength, deformation and (m2). Fifteen core samples f r o m each rock were
fracture characteristics o f rock and, therefore, o f subject to the U C S test which was carried out
cuttability. The U C S is characterised b y loading a using an Electronic Servo Controlled Stiff press
cylindrical specimen (with a length to diameter o f testing machine.
2:1) axially until the specimen fails with the Tensile strength test were c o n d u c t e d o n at least
strength o f the material given by I S R M (Brown, ten samples f r o m each rock type used. The test
1981): m e t h o d consists o f loading a disc o f the rock until
failure occurs across the diametrical axis. The disc
P was prepared f r o m 38 m m core diameter samples
trc - (1)
A with a thickness to diameter ratio o f 1:2. The rock
126 ~1. Ersov. M.D. H'a//er:E)l~meering Geology 39 ! 1995) 123 136

discs were loaded using an Avery Denison testing denoted an order of hardness, and had no quantita-
machine. The tensile strength of the specimen is tive significance or regular mathematical relation-
calculated from the formula, ship (Borner, 1962; Milner, 1962; West, 1981;
Gerrard, 1987; and Rogers, 1991). On the other
2P
hand, the work of West (1986) in assessing Cerchar
~Dt
abrasivity index on the standard minerals defined
Where P is the load at failure (N). l) is the in the Moh's scale (2-7) showed that there was a
diameter of the test specimen (ram), and I thick- rank order of abrasiveness equally spaced in a
ness (mm). systematic manner. Accordingly, Gundlach and
Modulus of Elasticity of the rocks was deter- Park (1978), A1-Ameen and Waller (1992) have
mined from the stress/strain graphs obtained revealed that the standard minerals used in the
during the UCS test. The Young's Modulus values Moh's scale showed a systematic increase in Moh's
in this study were calculated from gradient of the scale with Vickers hardness. Hardness measure-
stress strain diagram at 50% of the mean UCS. ments are affected by mineral anisotropy, different
The density of the rocks was determined by values being obtained on different crystal/grain
taking the dimensions of core samples and calculat- ti~ces as well as quite considerable variation on
ing volume and weighing after air drying. individual faces, and by the relations between
The hardness is a function of the type, size. cleavage plane orientations and the direction of
shape and quantity of the various mineral constit u- the indentation. However, the authors of this paper
ents of the rock and the bond strength that exist believe that individual mineral hardness has a great
between the mineral grains. Generally, there are influence in the determination of abrasiveness
three types of test used to evaluate rock hardness: lCerchar index, F-abrasivity factor and DIAI) and
(a) Mineralogical (e.g. Moh's and Rosiwal's hard- drillability of sedimentary rocks particularly those
ness scale); which have been weakly bonded and loosely
(b) Indentation tests (e.g. Brinell, Rockwell, cemented by soft material. Thus, the overall rock
Knoop, Vickers hardness); and Moh's hardness was one of the parameters which
(c) Dynamic or rebound tests (e.g. Shore schlero- was included in the prediction of drill bit per-
scope test). formance using multi-variable linear regression
The first is a petrological method and the last analysis (Ersoy and Waller, 1995a). The main
two tests are mechanical. In this study, the first disadvantage of Moh's method is that it ignores
and the last methods were used. The Vickers. the bonding strength of the matrix material.
Brinell and Rockwell tests have been developed as The Schleroscope hardness works by dropping
standard metallurgical hardness tests which rely a small tungsten carbide hammer (2.3 g in weight
on the homogeneous nature of man made materials and 5.73 mm in diameter) with a small diamond
such as metals and various alloys, tbr integrit} insert from a fixed height and measuring the
and accuracy. However, natural rock is completely rebound height. The height of rebound is gradua-
different from homogeneous man made metal ted arbitrarily from 0 to 140 on a dial gauge
materials. Nevertheless, the Knoop and Vickers mounted directly above the vertical barrel contain-
tests may be adapted to petrographic microscopes ing the diamond hammer. The rock specimens
and used to determine micro hardness of individual have flat surfaces of 25 cm 2 with a thickness of
mineral grains. 5 cm. ISRM (Brown, 1981 ) recommended that the
The Moh's hardness of a mineral grain measures test should be repeated 20 times. However, in
its resistance to being scratched. The overall rock practise this number is too low, with a sample size
hardness (Moh's hardness) was determined from of 100 values being more representative. Therefore,
the summation of individual mineral hardness at least 100 readings were made for each rock
(Moh's scale) multiplied by the corresponding specimen. Most of the rock characteristics are
mineral percentage, divided by 100. Until recently involved in the results of the Shore test.
the numbers 1 to 10 on the Moh's scale merely Various factors affect the abrasiveness of rock.
 A. Ersoy, M.D. Waller~EngineeringGeology 39 (1995) 123-136 127

Particularly important factors are mineral com- Where Fis the Schimazek's wear factor (N/mm),
position, the hardness of mineral constituents, EqQtz is the equivalent quartz volume percentage,
grain characteristics such as size, shape and angu- ~b is the grain size (mm) and BTS is indirect
larity, the type of matrix material or the type and Brazilian tensile strength. It has been suggested
degree of cementation, and strength, hardness and that if the grain size is less than 0.025 mm, the
specific energy. Abrasivity tests index conducted grains have little influence on abrasivity. Tensile
include Cerchar abrasivity index, Dynamic impact strength is taken as a measure of the bond strength
abrasive index (DIAI) and Schimazek's between grains. Hardness of minerals other than
F-abrasivity factor. The Cerchar Abrasive index is quartz may be accounted for by expressing their
a simple test to conduct and is used widely to hardness relative to that of quartz. Certainly in
assess the abrasiveness of rock. The Cerchar abras- Europe the F value is used to assess abrasivity of
ivity test consists of scraping a sharp steel, coni- sedimentary rocks, especially in coal mining and
cally pointed stylus; En24 or En25 grade, tunnelling. The F value has shown a tendency to
accurately heat treated to 610 Vickers hardness, be linearly related to abrasive wear (Schimazek
ground to a cone angle of 90 °, across the rock and Knatz, 1970; Braybrooke, 1988; Verhoefet al.,
specimen for a distance of 10 mm under a normal 1990; Verhoef, 1993).
load of 7 kg. The abrasiveness of rock is deter- Mechanical and index properties of the rocks
mined by measuring the wear flat produced on the are presented in Table 3. Note that for all tests
steel stylus by using a travelling microscope fitted mentioned above, the test methodology and
with a dial micrometer with 0.001 mm resolution. description of the equipment used can be found
The abrasiveness of rock is defined by the wear more detail elsewhere (Ersoy, 1995).
flat diameter in mm multiplied by 10. A minimum
of 50 scratch tests for each rock have been made
in order to obtain reliable and representative
results. The Cerchar index is largely influenced by 4. Quantitative analysis of the rock texture
the abrasive mineral content, silica content and
the hardness of individual minerals (Table 9). 4.1. Experimental procedure
However, the Cerchar index is not applicable to
unconsolidated or weak or friable rock and soil. The rock particles were assessed using a micro-
It is useless when a rock contains large grains scopic image analysis of thin sections. A series of
(> 10 mm). thin sections were prepared from core or hand
The dynamic impact abrasive index (DIAI) specimens of each rock type with approximate
method of assessing the potential abrasiveness of dimensions of 35 by 50 mm. Image processing was
rocks was developed at the University of performed using an IBAS 2000 automatic image
Nottingham; details may be found in A1-Ameen analysis system, with video camera input of thin
and Waller (1992). The DIAI has the advantage section photographic prints. Manually, any aspect
of being applicable to a range of rocks from fully of discrimination or measurement is laborious and
competent to weak friable material. This clearly time consuming. However, with the advent of
provides the benefit of testing at the lower end of automatic measuring systems, any relevant infor-
the strength scale where other abrasive tests cannot mation can be derived rapidly often to a greater
be executed due to sample shape or condition. The degree of accuracy. The principals and procedures
DIAI is a function of hard and abrasive mineral of automatic image analysis are described in more
content of the rock. detail in Jones (1987) and Allen (1993). In prac-
F-abrasivity factor (Schimazek and Knatz, tise, direct analysis of thin section is impossible
1970) is defined as: under the image analyser. Therefore, the following
steps were developed:
F - EqQtz x ~bx BTS (a) Several photomicrographs from each rock
100 (N/mm) (3) type were taken using a polarising microscope.
128 A. Ersov, M.D. Waller~EngineeringGeology 39 (1995) 123 136

Table 3
Mechanical and intact properties of the rocks

UCS (MPa) BTS (MPa) Young's Modulus Dry density Moh's Shore Cerchar DIAI(%) F-abrasivity
(GPa) (g/cm 3 ) hardness hardness abrasive (N/mm)
index

Limestone
28.17 2.86 9.74 2.23 2.85 19 0.63 1.50 0.06

Sandstone
37.45 3.2I 9.85 2.27 6.44 41 3.83 68.39 1.13

Siltstone
90.54 7.49 17.70 2.39 5.51 51 1.70 44.23 0.41

Granite
106.15 8.6(I 19.78 2.61 5.74 85 3.75 70.76 3.02

Diorite
375.20 30.26 37.38 2.99 5.95 95 3.31 72.41 2.48

(b) Colour prints ( 101 by 149 mm) were prepared. formula which was suggested by Howarth and
The prints of the very fine grained rock type Rowlands (1986,Howarth and Rowlands 1987).
(siltstone) were enlarged to A4 size.
(c) Grain boundaries were outlined using 0.4 mm
pencil on tracing paper. TC = AW L( No
N~N1
x
1)
FF0
(d) These traces were directly analysed with the
image analyser to determine geometrical parame-
ters of rock particles. + (No
N~-NlXAR1 xAF 1 )1 (4)
The data were recorded for analysis, which
consisted of percentage area of grains and matrix, where,
length, breadth, perimeter, orientation (angle) and
area of each grain in the each viewing window. As TC = Texture coefficient
a result, a single number of the texture coefficient, AW = Area weighting (grain packing density)
which models the interaction between the rock N 0 = N u m b e r of grains with aspect ratio (maxi-
structure and the drill bit, was calculated for mum Feret's diameter or length to minimum
each rock. Feret's diameter or breadth) less than 2.0
The method of quantitative analysis of geometri- N1 = Number of grains with aspect ratio greater
cal properties of rock particles or rock texture than 2.0
comprises the following components: FF0 =Arithmetic mean of form factor of all No
(a) To measure and analyse grain shape. grains
(b) To measure and analyse grain elongation (to AR1 = Arithmetic mean of aspect ratio of N1 grains
calculate grain circularity shape factor and aspect AF~ =angle factor orientation which were com-
ratio). puted for all N 1 grains
(c) To measure and quantify grain angle
(orientation). 4.2. Measurement and calculation of the
(d) To calculate total grain area to total reference geometrical parameters
area (including matrix) or weighting factors based
on the degree of grain packing. Individual analysis consisted of selecting a refer-
The results can be derived from the following ence area or observation window, containing a
 A. Ersoy, M.D. Waller~EngineeringGeology39 (1995) 123-136 129

number of grains which depended on the size of

the grains in the rock (e.g. fine grained, coarse
M~
grained). Grains touching the observation window
were not measured, because these grains were not
completed. In other words, all distinct boundaries
or completed grains in the reference area were
measured and analysed (Fig. 1).
Area and perimeter for each grain were mea- diamete~
sured directly. Length and breadth have been
defined as being maximum and minimum Feret's
diameter. The Feret's diameter represents the per-
pendicular distance between two parallel, outer
tangents to an object. The longest diameter of an
object is obtained by selecting the largest of the
Feret's diameters measured in 32 different direc- Fig. 2. I l l u s t r a t i o n o f m a x i m u m a n d m i n i m u m F e r e t ' s d i a m e t e r s
(length and breadth).
tions (e.g. at an angular resolution of 5.7°).
Minimum Feret's diameter is defined taking the
shortest from 32 Feret diameters. Maximum and or increased roughness of the grain's perimeter.
minimum Feret's diameters are illustrated in Fig. 2. The circularity shape factor (form factor) of the
The maximum and minimum diameters are not grain is defined as:
necessarily orthogonal.
There are two secondary geometrical parameters Form factor (circularity shape factor)=
in the analysis of grain shape, aspect ratio and 4H(area)
form factor (circularity shape factor). The former
(Perimeter)Z (5)
is elongation, which is best measured using the
grain's aspect ratio. This is simply defined as the The values of this parameter range from close
ratio of the grain's length to breadth. Therefore, to 0, for very elongated or rough objects, to 1 for
any increase in elongation results in increased a perfect circle. As a shape deviates from circularity
aspect ratio. The form factor is a measure of the (by elongation or increased roughness), the circu-
grain's deviation from circularity. This deviation larity shape factor decreases. The shape of particles
may occur in two ways; elongation of the shape, has long been thought to be critical in assessing
the abrasive potential of rock mass. The results
Frame boundary , _~. . . . . . L____J__. Inr.nmnh~atlor'aln
from the grain shape of the rocks show that granite
is most angular followed by diorite, limestone
aeasured
(particularly elongated particles), sandstone and
finally siltstone (Table 8). Angular, elongated and
rough particles of a rock produce a lower drilling
rate and higher wear rate than rounded particles
(Ersoy and Waller, 1995a,b,c). They observed that
a decrease in grain shape (increase in angularity)
increases wear rate and decreases rate of penetra-
tion (ROP) of PDC and impregnated diamond
core bits except to some extent in limestone,
because there were no hard and abrasive minerals
in limestone and it was weakly bonded.
To differentiate between which mode of devia-
Fig. 1. Thin section grain outline (obtained from photograph) tion was to be used in the analysis, an aspect ratio
of sandstone. discrimination level of 2.0 was introduced.
130 A. Ersoy, M.D. Waller~Engineering Geology 39 (1995) 123 136

Therefore, the arithmetic mean of form factors ANGLEDMAX

was calculated for all grains falling below 2.0.
However, the arithmetic mean of aspect ratios was
calculated for all particles above the level of 2.0.
To maintain continuity with Eq. 4, the form factor
~EDMIN
was inverted to ensure that as a grain deviates
from circular, the result from either method of
measurement, increases (Howarth and Rowlands.
1987). Fig. 3. Illustration of the parameter angle ( A N G L E D M A X and
The term area weighting (AW) is defined as: A N G L E D M I N ).

Total grain areas within the reference

area boundary
AW= (6)
Total area enclosed by the reference area maximum Ferers diameter (length) and the hori-
boundary (including matrix area) zontal direction. The maximum value of angle is
The area weighting is based upon the grain 180 C. Angular orientation of grains was quantified
packing density within the reference boundary. A by the development of the angle factor. This factor
typical example is given in Fig. 1. The texture was only calculated for the elongated grains where
coefficient is scaled down according to the percen- their aspect ratio was greater than 2.0. The angle
tage area of grains in the total reference area. The factor AF1 has been calculated by a class weighted
area weighting factor is only distinct in sedi- system applied to the absolute, acute angular
mentary rocks (sandstone, siltstone and limestone) differences (0 ° < fl < 90°), between each and every
as shown in Table 4. For example, the texture elongated grain (Howarth and Rowlands, 1987).
coefficient of siltstone is reduced by 70% from 1.58 Therefore, for a group of N grains the number of
to 0.47. However, this factor does not affect the unique angular difference is:
texture coefficient of crystalline rocks (granite and
diorite) since these rocks consisted entirely of N(N- 1)
(eV-- 1 ) + ( N - - 2 ) + . . . + 2 + 1 -- (7)
crystals and did not include matrix. Therefore, in 2
these rocks, total grain area is equal to the total
area enclosed by the reference boundary. Thus, four grains will have: 3 + 2 + 1 = 6 unique
The final term angle ( A N G L E D M A X , O in angular differences ( i l l The angular differences
Fig. 3), is defined as being the angle between the are grouped into nine classes, each of which are
weighted (Table 5). The angle factor has been
calculated by summing of the class weighting and
Table 4
Summary of unweighted and weighted texture coefficient for
rock types

AW area weighting Texture coefficient Table 5

(_?lasses and weightings for absolute, acute angular differences
Unweighted Weighted
Number Class range ( [ 3 ) Weighting (i)
Sandstone
0.72 1.75 1.26 I O<10 ~ 1
Siltstone 2 10' < O < 2 0 ° 2
0.30 1.58 0.47 3 20'-<0_<30 ° 3
Limestone 4 30'<0<40 ° 4
0.58 2.49 1.44 40°<0_<50 ° 5
Granite 6 50°<0_<60 ° 6
1.00 2.27 2.27 60°<0_<70 ° 7
Diorite 70°<0_<80 ' 8
1.00 2.45 2.45 9 80 ° < O_< 90' 9
 A. Ersoy, M.D. Waller~EngineeringGeology 39 (1995) 123-136 131

fractions of the total number of angular differences The second example is given from real data.
in each class. Where there are hundreds of elongated particles
in a given rock, derivation of the number of
LV x, -1. angular differences in each class or the combina-
Angle factor (AFI)=i~ILN(N-~I)/2f (8)
tions of each particle are difficult manually.
where, Therefore, it is carried out automatically and the
results are presented in Table 6. For instance,
N = total number of elongated grains calculation of the angle factor of sandstone is
Xi = number of angular differences in each class shown as follows:
i = weighting factor and class number (a) Number of elongated particles 23 (N)
In order to better understand the deviation of (b) Angular orientations from horizontal,
the angle factor, two examples are given here. The AngleDmax (ODMAX)is computerised.
first is simply illustrated in Fig. 4. The figure shows (c) Total number of unique angular differences
three elongated particles (N= 3; A, B, C), angular 253 (Table 6)
orientations from horizontal (ODMAX): A(0°), Thus, from Eq. 8, the angle factor of sandstone
B(50°), C(120 °) and number of unique angular is:
differences from Eq. 7, N(N-1)/2 = 3: [(28 x 1)+(29 x 2)+(39 x 3)+(25 x 4)+(25 x 5)+
ODMAXg - - ODMAXB = 0 ° - - 5 0 ° = 5 0 ° (27 x 6)+(29 x 7)+(21 × 8)+(30 x 9)]
ODMAXA - - ODMAXC = 0 ° __ 120° = 1 2 0 °
253

{~DMAXB - - ODMAXC = 500 - - 1 2 0 ° = 7 0 °

=4.86 dividing by 5=0.97.

Acute, absolute, unique angular differences (fl)

can be calculated by subtracting 180° from any
absolute angular difference greater than 90 °. Thus, 5. Results and discussion of the analysis
the final result is:
A summary of the geometrical properties of
ODMAXA - - ODMAXB = 5 0 ° rock particles, and texture coefficient derivation,
together with the average grain size and grain
ODMAXA - - ODMAXC = 1 2 0 ° - 1 8 0 ° = 6 0 °
shape factor for the rock types are presented in
ODMAXB - - ODMAXC = 7 0 ° Tables 7 and 8, respectively. A graph showing
Consequently, the angle factor from Eq. 8 and texture coefficient against rock type is given in
Table5 is derived as: ~ ( 1 / 3 x 5 ) + ( 1 / 3 x 6 ) + Fig. 5. From Fig. 5, it can be seen that the crystal-
(1/3 x 7 ) = 6 (angle factor). This value is divided line rocks have high texture coefficients, whereas
by 5, (AFt = 6/5 = 1.20), because this ensures that the sedimentary rocks tested have medium to low
the factor is numerically very similar to the other texture coefficients. From Tables 7 and 8, a high
factors (Eq. 4) and does not affect the texture texture coefficient indicates one or a combination
coefficient disproportionately. of the following:
(a) A large grain area relative to matrix area.
(b) A long grain perimeter in relation to grain
area which suggests jagged particles with a high
A ~ 0 o ~ ~ degree of interlocking.
50° ~ C = 120° (c) Long, thin particles or crystals with orienta-
tions such that the resultant mass is tightly
interlocked.
(d) An increase in the number of grains, shape
complexity and randomness of orientation.
Fig. 4. Illustration o f the derivation of the angle factor. (e) High grain volume.
132 A. Ersoy, M.D. Waller~Engineering Geology 39 (1995) 123 136

Table 6
Derivation of the number of angular differences in each class in the analysed rock types

Number Class range Weighting Number of angular difference in each class

Sandstone Siltstone Limestone Granite Diorite

l O ' < O N 10" I 28 991 105 37 1145

2 IO<ON2 ~ 2 29 81B 98 47 1152
3 20: < O ~ 3 0 ' 3 39 962 113 54 1379
4 30 < O N 4 0 4 25 711 73 49 1104
5 40:'< O N 50~' 5 25 693 93 49 1195
6 50° < 0 ~ 6 0 '~ 6 27 567 66 41 1112
7 60 ~< 0 ~ 7 0 ° 7 29 714 77 49 1267
8 70'~< O ~ 80' 8 21 556 57 40 1173
9 80~< O ~90 ° 9 30 558 59 40 1058

Table 7
Summary of the geometrical parameters of rock particles t TPa: total particle area, TRA: total reference area including particles and
matrix areas)

Geometrical parameters Rock types

Sandstone Siltstone Limestone Granite Diorite

Aspect ratio < 2 No 92 343 33 43 145

FF0 (~t) 0.62 071 0.64 0.51 0.55
Aspect ratio <2 Nj 23 114 39 29 180
AR~ (~1 2.46 2.44 3.75 2.90 3.08
TP A (g-') 14416947 2504564 17196159 25674860 12503632
TRA (ta2) 19968743 8166667 29333944 25674860 12503632

Table 8
Texture coefficient derivation and average grain size and grain shape lactor determined for the rocks

No l Nt Average grain Grain shape

AW No+Nj FF0 No+Nl AR~ AF~ TC size (mm) factor

Sandstone
0.72 0.80 1.61 0.20 2.46 0.97 1.26 0.410 0.597
Siltstone
0.30 0.75 1.40 0.24 2.44 0.92 0.47 0.085 0.677
Limestone
0.58 0.45 1.54 0.54 375 0.89 1.44 0.744 0.512
Granite
1.00 0.59 1.94 0.40 2.90 0.98 2.27 0.703 0.468
Diorite
1.00 0.44 1.79 0.55 3.08 0.99 2.45 0.24t 0.480

The statistical correlation matrix of textural, r e l a t i o n s h i p b e t w e e n s t r e n g t h p r o p e r t i e s a n d tex-

m e c h a n i c a l a n d i n t a c t r o c k p r o p e r t i e s are p r e - t u r e c o e f f i c i e n t ( T C ) w a s in e v i d e n c e ( R = 0 . 6 2 ) .
s e n t e d in T a b l e 9. M o s t o f t h e r o c k p r o p e r t i e s H o w e v e r , d a t a w e r e n o t s t r o n g l y r e l a t e d in all
were f o u n d to be c o r r e l a t e d with each other. A cases. F o r e x a m p l e , w h e n s a n d s t o n e a n d l i m e s t o n e
 A. Ersoy, M.D. Waller/EngineeringGeology 39 (1995) 123-136 133

2.5.7 ,~ 0
I I
20 I

I
i 1.0

0,5J ~.a I
0.0
Silt~me Immm~ Granite Diorite

Fig, 5. A graph of texture coefficient against rock types. II

are compared to the other rocks in terms of TC

and strength properties, both rocks have moderate
TC and low strength values. However, in fact,
high TC should indicate high strength. The texture
coefficients of the two rocks are higher than II

expected. This and the low strength values of the ba

two rocks can be explained by one, or several
combinations of the following factors: Firstly, both I I

rocks contain large grains relative to their matrix

area and they are weakly bonded. In addition, .~ .~ . . . ~ " q . ~ ' q .
[-
limestone is cemented by very soft calcite, thus I I I
type and degree of cementation is very important
for the strength property in a rock. Secondly, both
rocks are highly porous. The effects of porosity on II
rock properties are not addressed in this study,
but is important in terms of abrasivity. Previous ~ . . . . . . . . . . .
studies (Price, 1960; Smordinov et al., 1970; Dube 8
and Singh, 1972) pointed out that for a given
sedimentary rock type all strength properties
reduce with increase in porosity. Porosity can I I I I I I I
influence mechanical performance because poros-
ity assists the networking of stress-induced micro-
fractures (Sangha et al., 1974). This process ulti- I I I I I I I I I
mately leads to the breakdown of the rock struc-
ture, which is manifested in a series of fractured ..,.e
e8
surfaces (Howarth and Rowlands, 1986). Thirdly,
individual mineral hardness is also important for
the strength properties of rocks. This is true in the
case of limestone, because it consisted of soft
III
calcite crystals and matrix, but the first factor must
be taken into account for the other cases.
A strong correlation exists between TC and o
grain shape factor (R=0.94), because the grain
shape factor is involved in the derivation of TC.
There is good correlation between TC and abra-
134 A. Ersoy, M.D. Waller, k~)tgineermg Geology 39 (1995) 123 136

sivity factor F (Table 9, R=0.83). This is because However, inter granular fi'acture modes (along
average grain size is included in the derivation of grain boundary) are predominant through low
factor F. Cerchar abrasivity index is not influenced strength rocks (Howarth and Rowlands, 1987:
by texture to the same degree as the abrasivity Rogers, 1991). Texture coefficient models grain
factor F, since mineral content, mineral hardness. shape, orientation, degree of interlocking, in a
bonding structure and type and degree of cementa- qualitative sense and it is apparent that TC posi-
tion all influence the Cerchar index as well as the tively influences drilling performance. An inter-
grain characteristics. There is also some correlation locked texture simply presents a physical barrier
between TC, density and Shore hardness, respec- to crack propagation. It can be concluded that the
tively (Table 9, R=0.73, 0.69). drilling and cutting process and fracture mecha-
An example of the relationship between drilling nisms in the rock are to some extent related to
and dependent on aspects of the rock texture.
performance (for PDC and impregnated diamond
core bits while drilling at 1150 rpm) and the texture
coefficient is given in Fig. 6, Note that a more 6. C o n c l u s i o n
comprehensive set of drilling data tbr these rocks
and bits may be found in Ersoy and Waller A wide range of compositional, mineralogical,
(1995a,b); Waller and Ersoy (1995) and Ersoy mechanical and intact rock properties were pre-
et al. (1995). The relationship of texture coefficient sented and their relationship with the rock texture
with rate of penetration (ROP) is clearly signifi- were statistically established. Many of the rock
cant. However, sandstone behaves differently from properties used in this study were found to have
the other rocks. The polycrystalline diamond com- high degree of correlation for a given rock type.
pact (pin and hybrid) attain the highest ROP in The characteristics of rock particles are important
the sandstone. This change in ROP in the sand- as the basis of an objective quantification of rock
stone is basically caused by the quantity of silicate texture. Texture coefficient is simple in concept
matrix, porosity and weak bonding of particles and empirical. The mechanical properties of the
which produce high drilling rates tbr the bits. In test rocks are significantly influenced by their
general, the fracture mechanism and crack propa- texture. Texture coefficient and intact rock prop-
gation in high strength rocks, particularly igneous, erty relationships are linear to some degree. The
are complex and varied. It has been suggested that relationship between rock texture and drill bits
these rocks tend to fracture by the propagation of perfbrmance was shown. However, TC is not
sufficient alone to assess all the rock properties,
cracks along cleavage planes (intra granular).
because hardness and abrasiveness of rock constit-
uents, bonding structure, type and degree of
cementation are not quantitatively involved in the
determination of the texture coefficient. Therefore,
• Pin * Hybrid J- Impregnated
the effects of these features on the performance of
14"
rock properties should be considered simulta-
12. ~ o~ neously, because there is no single test which
10' ~ determines the whole of the rock properties at the
same time. The rock texture can be used as a
6 ~ predictive tool for assessing the mechanical perfor-
mance, drillability, cuttability and wear perfor-
?
mance of the rocks.
0 ; : : : : : : ; : I
04 06 08 I 12 14 I~ 18 2 22 24 2o

Texture Coefficient
Acknowledgment
Fig. 6. The effect of texture coefficienton tile drilling perfor-
mance of PDC (pin and hybrid) and impregnated core bits The authors would like to acknowledge the
at 1150rpm. technical staff and research students of the
 A. Ersoy, M.D. Waller~EngineeringGeology 39 (1995) 123-136 135

Department of Mineral Resources Engineering at Harvey, P.K. and Atkin, B.P., 1983. Automated X-ray fluores-
the University of Nottingham for their help during cence analysis. Spec. Vol.: Sampling and Analysis for the
Mineral Industry. Inst. Min. Metall., pp. 17-26.
this investigation. Hawkins, A. and McConnell, B.J., 1991. Influence of geology
on geomechanical properties of sandstones. Proc. 7th Int.
Cong. Rock Mechanics. Balkema, Rotterdam, Vol. I:
257-260.
Howarth, D.F. and Rowlands, J.C., 1986. Development of
an index to quantify rock texture for qualitative assessment
References of intact rock properties. Geotech. Testing J., 9:
169-179.
Allen, M., 1993. Profile analysis of bulk particulate material. Howarth, D.F. and Rowlands, J.C., 1987. Quantitative assess-
Unpublished Ph.D. thesis, University of Nottingham, ment of rock texture and correlation with drillability and
Nottingham. strength properties. Rock Mech. Rock Eng., 20:
A1-Ameen, S.I. and Waller, M.D., 1992. Dynamic Impact Abra- 57-85.
sion Index for rocks. Int. J. Rock Mech. Min. Sci., Geomech. Hugman, R.H. and Friedman, M., 1979. Effects of texture and
Abstr., 29 (6): 555-560. composition on mechanical behaviour of experimentally
AI-Ameen, S.I. and Waller, M.D., 1992. Petrographic analysis deformed carbonate rocks. Am. Assoc. Pet. Geol. Bull.,
for assessing the abrasiveness of coal measures rock. Univ. 63(9): 1478-1489.
Nottingham, Dep. Miner. Resour. Eng. Mag., 19: 29-34. Irfan, T.Y. and Dearman, W.R., 1978. Engineering petrography
Barret, P.J., 1980. The shape of rock particles, a critical review. of a weathered granite. Q. J. Eng. Geol., 11:
Sedimentology, 21: 291-303. 233-244.
Bell, F.G., 1978. The physical and mechanical properties of Fell Jones, M.P., 1987. Applied Mineralogy, a Quantitative
sandstones, Northumberland, England. Eng. Geol., 12: Approach. Graham and Trotman, London, 259 pp.
1 29. Milner, H.B., 1962. Sedimentary Petrography, 4th ed. George
Bieniawski, Z.T., 1967. Mechanism of brittle fracture in rock. Allen and Unwin, London.
Int. J. Rock Mech. Min. Sci., 4: 395-430. Olsson, W.A., 1974. Grain size dependence of yield stress in
Borner, R., 1962. Minerals and Gemstones. Oliver and Boyd, marble. J. Geophys. Res., 79(32): 4859-4861.
London. Onodera, T.F. and Asoka Kumara, H.M., 1980. Relation
Braybrooke, J.C., 1988. The state of the art of rock cuttability between texture and mechanical properties of crystalline
and rippability prediction. Proc. 5th Australian-New rocks. Bull. Int. Assoc. Eng. Geol., (22): 173 177.
Zealand Conf. Geomechanics, Sydney, pp. 13-42. Price, N.J., 1960. The compressive strength of coal measure
Brown, E.T., 1981. Rock Characterisation Testing and Moni- rocks. Colliery Eng., 37(437): 283-292.
toring ISRM Suggested Methods. Pergamon, Oxford, 211 Rogers, S.P., 1991. Rock mass characterisatiou and indexing
PP. for excavation assessment. Unpublished Ph.D. thesis, Uni-
Dube, A.K. and Singh, B., 1972. Effect of humidity on tensile versity of Nottingham, Nottingham.
strength of sandstone. J. Mines, Metals Fuels, 20(1): 8-10. Sangha, C.M., Talbot, C.J. and Dhir, R.K., 1974. Microfractur-
Ehrlich, R. and Weinberg, B., 1970. An exact method for char- ing of a sandstone in uniaxial compression. Int. J. Rock
acterisation of grain shape. J. Sediment. Petrol., 40(1): Mech. Min. Sci., Geomech. Abstr., 11: 107--113.
205-212. Schimazek, J. and Knatz, H., 1970. The influence of rock struc-
Ersoy, A., 1995. Performance analysis of polycrystalline dia- ture on the cutting speed and pick wear of heading machines.
mond compact (PDC) core bits in rocks. Ph.D. thesis, in Gluckauf, 106:274-278 (in German).
preparation. The University of Nottingham, Notting- Smordinov, M.I., Motovilov, E.A. and Volkov, V.A., 1970.
ham. Determination of correlation relationships between strength
Ersoy, A. and Waller, M.D., 1995a. Drill bit performance pre- and some physical characteristics of rocks. Proc. 2nd Congr.
diction using multi-variable linear regression analysis. Trans. Int. Soc. Rock Mechanics, Belgrade, Yugoslavia, Vol. 2:
Inst. Min. Metall., Sect. A. 35-37.
Ersoy, A. and Waller, M.D., 1995b. Wear characteristics of Sprunt, E.S. and Brace, W.F., 1974. Direct observations of
PDC pin and hybrid core bits in rock drilling. Wear. microcavities in crystalline rocks. Int. J. Rock Mech. Min.
Ersoy, A., Waller, M.D. and Clark, I., 1995. The influence of Sci., Geomech. Abstr., 11: 139-150.
rock properties on polycrystalline diamond compact (PDC) Stanton, R.L., 1972. Ore Petrology. McGraw Hill, New
drill bits performance. Ind. Diamond Rev. York, N.Y.
Gerrard, A.J., 1987. Discussion on "An observation on Moh's Verhoef, P.N.W., 1993. Abrasivity of Hawkesbury sandstone
scale of hardness". Q. J. Eng. Geol., 20: 99. (Sydney, Australia) in relation to rock dredging. Q. J. Eng.
Gundlach, R.B. and Park, J.L., 1978. Influence of abrasive Geol., 26(1): 5-17.
hardness on the wear resistance of high chromium irons. Verhoef, P.N.W., Van Den Bold, H.J. and Vermeer, Th.W.M.,
Wear, 46: 95-107. 1990. Influence of microscopic structure on the abrasivity of
136 A. Ersoy, M.D. Waller~Engineering Geology 39 (1995) 123-136

rock as determined by the pin-on-disc test. Proc. 6th Int. ling. Proc. Int. Syrup. Weak Rock, Tokyo. Balkema, Rotter-
Congr. IAEG, Amsterdam. Balkema, Rotterdam, Vol. 1: dam, pp. 585 593.
495 504. West, G., 1986. A relation between abrasiveness and quartz
Waller, M.D. and Ersoy, A., 1995. Performance analysis of content for some coal measures sediments. Int. J. Min. Geol.
PDC pin and hybrid core bits in rock drilling. 4th Int. Symp. Eng., 4:73 78.
Mine Planning and Equipment Selection, Calgary, Alta. Oct. Williams, H., Turner, F.J. and Gilbert, C.M, 1982. Petrography:
31 Nov. 3, 1995 (accepted for presentation). An Introduction to the Study of Rocks in Thin Sections.
West, G., 1981. A review of rock abrasiveness testing lbr tunnel- W.H. Freeman and Company, San Francisco, Calif.