Determination of c and w of rocks from indirect

tensile strength and uniaxial compression tests

N. Sivakugan*1, B. M. Das2, J. Lovisa1 and C. R. Patra3
Cohesion (c) and friction (w) angle are the two key parameters required in numerical simulations and
designs of underground openings, excavations, and foundations in rocks. The paper discusses a
simple method, based on a theoretical framework, to determine the two parameters from uniaxial
compression test and indirect tensile strength test. Laboratory tests were carried out on 35 rock
specimens and the values of c and w were computed. The predicted values of c appear to be
more realistic than those of w, and the scatter is attributed to the anisotropy and heterogeneity
generally observed in the rock cores. The computed value of c is 1?82 times the indirect tensile
strength (st).
Keywords: Indirect tensile strength, Mohr–Coulomb, Shear strength parameters, Uniaxial compression, Correlations, Rock strength

Introduction Young’s modulus (E). For example, st/sc is expected to

lie in the range of 1/5 to 1/20, with an average value of 1/10
Mohr–Coulomb is one of the most common constitu- (Sivakugan et al., 2013). Das et al. (1995) have reported
tive models used for geo-materials such as soils and the laboratory test results of compressive strength and
rocks. In carrying out the designs and also for tensile strength of lightly cemented sand (cement con-
simulating a soil or rock related problem such as tent54–8%) which showed st/sc to be about 1/10. The
excavation or underground opening including tunnel- magnitudes of st for these tests were obtained from
ling, it is necessary to have the appropriate design Brazilian tensile strength tests along with the tensile strain
parameters. In soils, the shear strength parameters (et) at failure by using optical image analysis technique.
cohesion (c) and friction angle (w) are determined from The ratio of et to ec (compressive strain at failure) was
laboratory tests such as triaxial or direct shear tests, or about 1/20. The modulus ratio of a rock specimen, defined
indirectly from in situ tests. However, carrying out as the ratio of the Young’s modulus to the uniaxial
triaxial tests on rock specimens to determine the shear compressive strength lies in the range of 150–1000.
strength parameters is a difficult task. Carrying out a direct tensile strength test is quite
Recently, Piratheepan et al. (2012) proposed a method cumbersome due to the difficulty in holding the speci-
for deriving the shear strength parameters of a granular men at the ends. Indirect tensile strength test (also
base material used for roadwork, which comprises lightly known as Brazilian indirect tensile strength test) is a
cemented well graded sandy gravel with some fines, using simple and indirect means of estimating the tensile
the data from indirect tensile strength tests and uniaxial strength. Goodman (1989) noted that the value of st
compression tests. 1?5–5?0% fly ash or slag was used as derived from the indirect tensile strength test can be
binder in the compacted specimens that were studied. significantly higher than the actual value derived from a
They reported that the cohesion thus derived is 1?824 direct tensile strength test.
times the indirect tensile strength. Noting the usefulness of deriving the shear strength
Unlike soils, rocks demonstrate significant tensile parameters c and w of rock specimens from uniaxial
strength. In the designs of underground openings, compression test and indirect tensile strength test data, the
excavations, and foundations on rocks, the tensile strength objective of this paper is to apply this method to
is often an input parameter. There are several empirical laboratory test data from 35 rock specimens, and to
correlations reported in the literature that relate the ascertain its validity. In addition, st/sc and E/sc ratios
compressive strength (sc), tensile strength (st), and derived from laboratory test data are studied.

1
School of Engineering and Physical Sciences, James Cook University, Indirect Tensile Strength Test
Townsville, QLD 4811, Australia
2
California State University, Sacramento, USA To analyse or model geo-materials that are subjected to
3
National Institute of Technology, Rourkela, India tensile stresses in service, it is necessary to know their tensile
*Corresponding author, email: [email protected] strength, which is quite difficult to measure. This is

ß 2014 W. S. Maney & Son Ltd

Received 1 April 2013; accepted 28 May 2013 International Journal of
DOI 10.1179/1938636213Z.00000000053 Geotechnical Engineering 2014 VOL 8 NO 1 59
Sivakugan et al. Cohesion and friction angle of rocks from tests

1 Indirect tensile strength test on a cylindrical sample

especially true for rocks, clays, and cemented soils. Point A (horizontal diametrical line) at distance x from O
Brazilian test is a practical alternative for deriving the
tensile strength. Here, a cylindrical specimen of diameter D sx (radial)~
is placed with its axis horizontal. A metal strip is placed at 8   2 39
> 4x2 4x2 >
> 1{ sin 2a >
the top and bottom while applying a load P uniformly 2p < D2 61{ D2
{1 6
7=
7 (1)
across the entire length t at right angle to the cross-section { {tan 4 tan a5>
p> 2 4
4x2
(Fig. 1). Failure takes place generally along the vertical :1z 8x cos 2az 16x
> 1z >
;
diameter, and it can be shown that the indirect tensile D2 D4 D2
strength st is given by 2P/pDt.
Figure 1 shows a cylindrical sample of diameter D and sy (tangential)~
length t is subjected to an indirect tensile strength test. The 8   2 39
> 4x2 4x2 >
>
< 1{ 2 sin 2a >
projected width of the loading  rim is a which is assumed to 2p D 6
{1 6
1{
D 2 7=
apply a uniform pressure p ~ at P
where P is the applied ztan 4 7
tan a5 (2)
p> 2 4
4x2 >
load. The loading rim subtends an angle of 2a at the origin :1z 8x cos 2az 16x
> 1z >
;
D 2 D 4 D 2
O. Assuming that the pressure p is radial, especially when
a is small, it can be shown that the normal stresses at
points A and B on the horizontal and vertical radial lines At point A, sx is the radial stress acting along x-
are given as follows (Hondros, 1959). Here compressive direction, which is tensile; sy is the tangential stress which
stresses are positive. acts along y-direction but is compressive (see Fig. 1).

60 International Journal of Geotechnical Engineering 2014 VOL 8 NO 1

 Sivakugan et al. Cohesion and friction angle of rocks from tests

Point B (vertical diametrical line) at distance y from O the centre is shown in Fig. 3. Here, the tensile stress and
the compressive stress at the centre are –Y and 3Y,
sy (radial)~ respectively. From the Mohr circle,
8   2 39
> 4y2 4y2 >
>
< 1{ 2 sin 2a > 2Y
2p D 6
{1 6
1z
D 2 7= sin w~ (9)
ztan 4 7
tan a5 (3) c cot QzY
p> 2 4
4y2 >
:1{ 8y cos 2az 16y
> 1{ >
; where Y is the absolute value of the indirect tensile
D 2 D 4 D 2
strength, as obtained from equation (7). While length to
sx (tangential)~ diameter ratio of 2:1 is adopted for concrete cores, rock
8   2 39 samples are tested in the form of discs with length to
> 4y2 4y2 >
> 1{ sin 2a > diameter ratios of 1:2.
2p < D2 61z D2
{1 6
7=
7 (4)
{ {tan 4 tan a 5>
p> 2 4
4y2
:1{ 8y cos 2az 16y
> 1{ >
; Uniaxial Compressive Strength Test
D2 D4 D2
Uniaxial compressive strength (UCS or sc), also known as
At point B, sx is the tangential stress acting along x-
the unconfined compressive strength (qu), is the most
direction, which is tensile; sy is the radial stress acting
common strength parameter used for quantifying rock
along y-direction, which is compressive (see Fig. 1). Along
strength. It is determined through a uniaxial compression
the horizontal and vertical diametrical lines the shear
test, also known as the unconfined compression test. A
stress is zero, making them principal planes, and the values
given by equations (1) through (4) principal stresses. uniform vertical normal stress, applied on the horizontal
The values of the stresses from equations (1) through circular cross-sections of a cylindrical sample, is increased
(4), assuming a52u are plotted in Fig. 2 in dimensionless to failure. During the loading, there is no confining
form to show the relative magnitudes of the four stresses. pressure around the sample and hence the test is quite
Figures 2a and b show the variations of sx and sy along simple and the interpretation is straightforward. The
the horizontal and vertical diametrical axes respectively. vertical stress at failure, is the uniaxial compressive
Figure 2c combines the plots in Fig. 2a and b for strength or unconfined compressive strength, denoted by
comparison. The compressive and tensile stresses are qu or sc. The Mohr circle at failure for a UCS test is shown
shown by continuous and dashed lines, respectively. in Fig. 4 where sc52X. Substituting X5sc/2 and Y5st, it
It can be seen that the vertical normal stress sy is always can be deduced that
compressive (positive) along the vertical and horizontal X
diametrical lines, irrespective of whether it is radial (along sin Q~ (10)
c cot QzX
vertical line) or tangential (along horizontal line). The
radial horizontal normal stress sx along the horizontal
diametrical line is always tensile (negative), and is the From equations (9) and (10),
maximum at the centre. The tangential horizontal normal    
X {2Y sc {4st
stress sx along the vertical diametrical line is mostly tensile Q~sin{1 ~sin{1 (11)
(negative), except for a small length near the loading X {Y sc {2st
regions where they can be compressive (positive). It can be
seen that this tangential horizontal normal stress sx along XY 0:5sc st
c~ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ~ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi (12)
the vertical diametrical line, which is in fact the indirect Y (2X {3Y ) st (sc {3st )
tensile strength, is nearly a constant (Fig. 2b).
At the centre, where x/D50 and y/D50, equations (1)
and (4) become
2p sc/st Ratio
sx (tensile)~{ ðsin 2a{aÞ (5)
p In Griffith failure criterion for rocks, when there is no
and equations (2) and (3) become confining pressure, the uniaxial compressive strength is
eight times the direct tensile strength (Griffith, 1924).
2p Fairhurst (1964) suggested that st estimated from the
sy (compressive)~ ðsin 2azaÞ (6)
p Brazilian test is often less than what is measured in a
P
When a is small, sin 2a<2a. Substituting p~ at and a~ Da , uniaxial tensile test. Similar observations were made by
the vertical and horizontal normal stresses at the centre, Yu et al. (2009). McClintock and Walsh (1962) reported
which are also the principal stresses, are given by uniaxial compressive to uniaxial tensile strength ratios in
2pa 2P the range of 5–22.
sx (tensile)~{ ~{ ~{Y (7) From equations (9) and (10), the uniaxial compressive
p pDt
6pa 6P strength sc and the indirect tensile strength st can be
sy (compressive)~ ~{ ~3Y (8) expressed in terms of the cohesion and friction angle as
p pDt

Based on the principal stress values given by equations (7) 2c cos Q

sc ~2X ~ (13)
and (8), the Mohr circle representing the state of stress at 1{sin Q

International Journal of Geotechnical Engineering 2014 VOL 8 NO 1 61

Sivakugan et al. Cohesion and friction angle of rocks from tests

2 sx and sy values for a52u: a Along horizontal diameter; b Along vertical diameter; c Along horizontal and vertical diameters

c cos Q
Laboratory Testing Program
st ~Y ~ (14) A series of laboratory tests including uniaxial compression
2{sin Q
test, indirect tensile strength test, point load test, and
From equations (13) and (14), the ratio sc/st can be Schmidt hammer test were carried out on 35 sets of
written as specimens obtained from rock core at the same location in
sc 2(2{sin Q) the ground (Adams, 2009). With some difficulty, direct
~ (15) tensile strength tests were also performed on selected
st 1{sin Q
specimens. The tests were carried out as per ASTM and
which is independent of cohesion. ISRM standards which are very similar. The test procedures

62 International Journal of Geotechnical Engineering 2014 VOL 8 NO 1

 Sivakugan et al. Cohesion and friction angle of rocks from tests

3 Mohr circle for the state of stress at the centre in an indirect tensile strength test

are also discussed in Sivakugan et al. (2011). The specimen angle, from 4 at w50. It takes a value of 6–12 for w530u–
dimensions for the different tests are shown in Fig. 5. From 54u. The rate of increase escalates with the friction angle.
the uniaxial compression test, Young’s modulus (E) and The sc/st ratios, derived from the 35 test specimens are
uniaxial compressive strength (sc) were determined for the 35 also shown in Table 1, where they lie in the range of 3?6–
specimens, with diameters of 45–47 mm. The tensile strength 12?8. E/sc ratios lie in the range of 147–281 for all 35
(st) was determined from the Brazilian indirect tensile specimens.
strength test. Table 1 summarises the data from the uniaxial Due to high degree of anisotropy and heterogeneity, the
compression test and indirect tensile strength tests only. The rock test data are often not reproducible. There can be
data from the other tests are not given in this paper. significant variability between specimens of the same rock
The friction angle and cohesion derived from equations obtained from the same location. Therefore, it is difficult
(11) and (12) are given in Table 1. Two (16 and 25) of the to make comparisons between any of the data sets in
friction angle values are negative which is unrealistic. Four Table 1 even when they are from the same type of rock. In
more specimens (19, 21, 26, and 32) gave relatively lower the case of lightly cemented granular soil specimens
values of friction angles. The friction angles of the other 29 studied by Piratheepan et al. (2012), prepared under
specimens were in a more realistic range of 20u–55u. The controlled conditions in the laboratory, the test data are
cohesion values were relatively more consistent, and were more consistent and the tests are reproducible.
in the range of 12–50 MPa. Figure 7 shows the variation of cohesion derived from
The theoretical variation of sc/st, as followed by equation (12), plotted against the indirect tensile strength
equation (15), is shown in Fig. 6, along with the test data, derived from the laboratory test. It can be seen that they
where it is evident that the ratio increases with the friction are related by

4 Mohr circle for the state of stress at the centre in a UCS test

International Journal of Geotechnical Engineering 2014 VOL 8 NO 1 63

Sivakugan et al. Cohesion and friction angle of rocks from tests

5 Test specimens from the rock core

Table 1 Rock strength test data

Specimens Measured parameters Estimates from equations (11) and (12) Ratios

Type* ID No. r/g cm23 st/MPa sc/MPa E/GPa w/u c/MPa sc/st E/sc

1 1 2?99 19?98 169?11 25?0 43?7 36?2 8?5 147?8

1 2 2?83 21?89 139?13 23?3 32?7 38?0 6?4 167?3
1 3 2?87 15?22 117?85 24?6 40?7 27?1 7?7 208?7
2 4 2?88 13?73 108?44 22?4 41?4 24?5 7?9 206?4
2 5 3?08 16?27 99?31 22?4 30?8 28?2 6?1 225?2
2 6 3?11 15?77 95?14 19?0 30?3 27?3 6?0 200?1
2 7 3?11 19?36 169?22 29?4 44?7 35?3 8?7 173?4
2 8 3?33 22?95 117?36 28?2 21?0 40?4 5?1 239?9
3 9 2?97 14?80 160?96 30?6 50?8 28?7 10?9 189?8
3 10 2?96 19?55 125?89 25?7 33?3 33?9 6?4 204?2
3 11 2?92 5?96 76?04 15?0 54?5 12?2 12?8 197?3
3 12 3?11 20?40 136?57 22?3 35?0 35?5 6?7 163?3
3 13 3?16 19?64 91?58 16?8 14?4 35?5 4?7 183?0
3 14 2?78 21?58 109?70 23?4 20?6 38?0 5?1 212?9
3 15 2?93 25?44 150?15 27?2 29?2 44?1 5?9 181?0
3 16 2?88 20?01 79?92 12?8 20?2 40?1 4?0 160?2
3 17 3?32 17?13 98?00 20?0 27?5 29?7 5?7 204?5
3 18 2?88 16?92 112?26 24?2 34?6 29?4 6?6 215?9
3 19 3?10 19?67 86?88 21?6 9?9 36?5 4?4 248?5
3 20 3?03 12?60 112?14 19?2 45?2 23?1 8?9 170?9
3 21 3?11 19?21 77?44 15?5 0?9 38?1 4?0 200?0
3 22 2?83 23?10 175?27 28?5 39?9 40?9 7?6 162?4
3 23 2?95 24?49 158?58 24?9 33?6 42?5 6?5 157?1
3 24 2?90 16?46 183?75 27?1 51?4 32?2 11?2 147?5
3 25 2?85 19?45 69?43 18?3 215?9 46?0 3?6 263?3
4 26 2?77 23?10 98?31 25?5 6?5 43?9 4?3 259?0
5 27 2?96 15?88 153?34 28?0 47?6 29?7 9?7 182?6
5 28 2?72 22?81 151?06 28?2 34?6 39?7 6?6 186?7
5 29 2?73 28?00 203?97 29?0 38?4 49?3 7?3 142?3
5 30 2?75 26?34 279?73 32?2 50?2 50?7 10?6 115?0
5 31 3?12 11?94 95?33 21?5 41?7 21?4 8?0 225?5
5 32 2?61 11?23 47?59 13?8 6?1 21?4 4?2 290?8
6 33 3?01 20?06 116?37 25?5 28?3 34?8 5?8 219?2
6 34 3?06 10?08 88?33 23?7 44?8 18?4 8?8 268?0
6 35 2?68 16?05 81?37 22?9 20?4 28?3 5?1 281?4
*15Medium grained felsic rock, 25Medium grained igneous rock, 35Carbonate altered medium grained mafic volcanic rock,
45Carbonate altered fine grained mafic volcanic rock, 55Fine grained sedimentary rock, 65Highly brecciated rock

64 International Journal of Geotechnical Engineering 2014 VOL 8 NO 1

 Sivakugan et al. Cohesion and friction angle of rocks from tests

6 Variation of sc/st versus w

7 Cohesion versus indirect tensile strength relation

c~1:82st (16)
which is in agreement with the relationship suggested by The cohesion estimated from equation (12) is related to the
Piratheepan et al. (2012) for cemented sands. measured value of st through the following equation

Summary and Conclusions c~1:82st (16)

2
Uniaxial compression test and indirect tensile strength test with r 50?92 for the 35 data (see Fig. 7).
are two of the common tests carried out on rock
specimens, to determine the uniaxial compressive strength References
(sc), Young’s modulus (E), and indirect tensile strength
(st), respectively. In designs and numerical modelling of Adams, S. 2009. Correlation analysis of rock strength parameters, BE
Hons thesis, James Cook University, Townsville, Australia.
underground openings, excavations, and foundations on Das, B. M., Yen, S. C. and Dass, R. N. 1995. Brazilian tensile strength of
rocks, Mohr–Coulomb shear strength parameters, cohe- a lightly cemented sand, Can. Geotech. J., 32, (1), 166–171.
sion (c), and friction angle (w), are often required. Fairhurst, C. 1964. On the validity of the Brazilian test for brittle
Following the work of Piratheepan et al. (2012), it is materials, Int. J. Rock Mech. Min. Sci. Geomech. Abstr., 1, (4), 535–
546.
shown from first principles that the cohesion and friction
Goodman, R. E. 1989. Introduction to rock mechanics, 2nd edn, New
angle can be derived from sc and st using the following York, Wiley.
equations. Griffith, A. A. 1924. The theory of rupture. Proc. 1st International
  Congress on ‘Applied mechanics’, Delft, the Netherlands, Techni-
{1 sc {4st sche Boekhandel en Drukkerij J.Waltman Jr, 54–63.
Q~sin (11)
sc {2st Hondros, G. 1959. The evaluation of Poisson’s ratio and the modulus of
materials of a low tensile resistance by the Brazilian (indirect tensile)
test with particular reference to concrete, Aust. J. Appl. Sci., 10, (3),
0:5sc st 243–268.
c~ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi (12)
st (sc {3st ) McClintock, F. A. and Walsh, J. B. 1962. Friction on Griffith cracks in
rocks under pressure. Proceedings, 4th US National Congress of
Applied Mechanics, ASME, New York, Vol. 2, 1015–1021.
Laboratory tests were carried out on 35 sets of rock Piratheepan, J., Gnanendran, C. T. and Arulrajah, A. 2012.
Determination of c and w from an IDT and unconfined compression
specimens, and the shear strength parameters determined testing and numerical analysis, J. Mater. Civ. Eng., 24, (9), 1153–
from equations (11) and (12) are summarised in Table 1. 1163.
The predicted cohesion values appear to be more realistic Sivakugan, N., Arulrajah, A. and Bo, M. W. 2011. Laboratory testing of
than the friction angle values. The scatter is attributed to soils, rocks and aggregates, Boca Raton, FL, US, J Ross Publishing.
anisotropy and heterogeneity generally observed in the Sivakugan, N., Shukla, S. K. and Das, B. M. 2013. Rock mechanics – an
introduction, Boca Raton, FL, US, CRC Press, Taylor and Francis
rock specimens. Group.
sc/st ratios of the specimens ranged from 3?6 to 12?8, Yu, Y., Zhang, J. and Zhang, J. 2009. A modified Brazilian disk tension
which is in agreement with values reported in the literature. test, Int. J. Rock Mech. Min. Sci., 46, 421–425.

International Journal of Geotechnical Engineering 2014 VOL 8 NO 1 65