ANALYSIS OF METHODS FOR EVALUATION OF SOIL SHEAR

STRENGTH PARAMETERS
Jonas Amšiejus1, Neringa Dirgėlienė2, Arnoldas Norkus3
123
Vilnius Gediminas technical university, Saulėtekio ave. 11, LT-10223 Vilnius, Lithuania.
E-mail: 1ajonas@vgtu.lt; 2Neringa.Dirgeliene@vgtu.lt; 3Arnoldas.Norkus@vgtu.lt

Abstract. Having performed an analysis of the methods for identifying a soil shear strength one can find four differ-
ent coordinate systems for evaluating the soil shear strength parameters. It is stated that a theoretical functional rela-
tion exist for the shear strength parameters, have been identified via different evaluation methods. Thus one must ob-
tain the same final parameters of shear strength. The investigation is assigned to identify reasons of obtained differing
magnitudes of the shear strength parameters via triaxial testing by employing the used in practice methods. The
method to identify angle of internal friction and cohesion satisfying all four coordinate systems is proposed.

Keywords: soil shear strength parameters evaluation methods, triaxial testing, angle of internal friction, cohesion.

1. Introduction tains the different magnitudes of the angle of internal

friction and the cohesion.
A determining the soil shear strength via direct shear The authors calculated shear strength parameters by
apparatus. At his case the tangent of internal friction an- common methods and performed comparative analysis,
gle tan ϕ and the cohesion c are calculated according to identified the reasons, resulting the different internal
normal and shear stresses acting at the failure plane hav- friction angle and cohesion magnitudes. The proposed
ing directly measured the acting forces (Aysen 2005; method for soil analysis via triaxial testing and that of
Craig 2004). processing of the data allows to evaluate each sample
The perpendicular to sample surfaces stresses are failure plane position for identifying the shear strength
measured when shear strength is investigated via triaxial parameters.
testing (CНиП 2.02.02-85). The stresses at failure plane
are calculated according to the above mentioned plane 2. Shear strength parameters evaluation methods
stresses having employed some assumptions. One of
them states that the full stresses at the sample surfaces are 2.1 σ1–σ3 coordinate system
the normal, id est they are principal ones. Thus, no shear
When the soil testing results are presented in σ1–σ3
stresses exist on the sample surface. Another assumption
coordinate system the angle of internal friction and the
states that the sample collapses at plane inclined by the
cohesion are obtained by following the certain proce-
angle 45 + ϕ 2 to the minor principal stress. This angle is
dures. First, the parameters tan θ and c θ of linear rela-
unknown value when processing the testing. Therefore
one faces difficulties to determine normal and shear tion σ1 = f (σ 3 ) are obtained (Bukhartsev 1988) (Fig 1).
stresses at failure plane. Several methods are applied to
solve the above mentioned problem (Parry 2004). The
essence of the methods are as follow: the parameters of
linear relationship are calculated for different coordinate
systems, then employing theoretical relationships the
magnitudes of the angle of internal friction and the cohe-
sion are obtained.
The methods are valid also for the case when func-
tional relationship exists for normal versus shear stresses
at failure plane. When processing the triaxial testing data
for more than 2 tests and when the sample strength pa- Fig 1. State of stresses via coordinates σ1–σ3
rameters have scatter versus linear relationship, one ob-
1077
 Taking into account a linear relationship of the ma-
jor σ1 and minor σ3 principal stresses (Amšiejus et al.
2009; Dirgėlienė 2007; CНиП 2.02.02-85), reading one
obtains the magnitudes of the tangent of internal friction
angle tan ϕ and the cohesion c:
2.3 t–s (Massachusetts) coordinate system
When the soil testing results are presented in t–s co-
ordinate system (see Fig 3) (Craig 2004), combining the
shear stresses

ϕ ϕ σ1 − σ 3
σ1u = σ 3 ⋅ tan 2 (45 + ) + 2 ⋅ c ⋅ tan(45 + ) , (1) t = , (9)
2 2 2

tan θ − 1 and the normal stresses

tan ϕ = , (2)
2 tan θ σ1 + σ 3
s= , (10)
2
cθ
c= . (3) the parameters tan α and cα of the linear relationship
2 tan θ
t = f ( s ) are obtained.

2.2 q–p (Cambridge) coordinate system

When the soil testing results are presented in q–p

coordinate system (Fig 2) (Parry 2004; Aysen 2005; Ran-
jan and Rao 2005), combining the mean principal stresses
σ + σ 2 + σ3 )
( 1 σ + 2σ 3 )
( 1
p= = , (4)
3 3
and the deviatoric stresses
q = σ1 − σ 3 , (5)
The q = f ( p ) parameters tan ψ and cψ of the linear
Fig 3. State of stresses represented via coordinates t–s

relationship are calculated by (Флорин 1959): The angle of internal friction following the Mohr
c(ψ ) = q − p ⋅ tan ψ . (6) circle for a case c = 0 is described by:
σ1 − σ3
t
tan α = = 2 , (11)
s σ1 + σ3
2
that of the cohesion:
cα = t − s ⋅ tan α . (12)
The Coulomb stresses envelope at failure is tangen-
tial to the Mohr stresses circle with a slope ϕ and the
intercept c (Vervečkaitė et al. 2007). The failure line for
Fig 2. State of stresses represented via coordinates q–p t–s coordinate system crosses a point X (see Fig 4). The
line K f has a slope α and an intercept coordinate cα .
The parameters tan ψ and c ψ can be converted into
The values of α and cα by applying the t–s plot
the shear strength parameters ϕ and c . Having per-
can be converted into ϕ and c by employing the equa-
formed arrangements one obtains:
tions (13) and (14):
3 tan ψ
sin ϕ = , (7) ϕ = sin −1 (tan α) ,
6 + tan ψ
(13)

cα
c= . (14)
and the cohesion : cos ϕ
sin ϕ
c = c (ψ ) ⋅ . (8)
tan ψ

1078
 2

∑ ∑
n n
2 ⎛ ⎞
Δ= n σi − ⎜
⎜ σi ⎟ ⎟ . (20)
i =1 ⎝ i =1 ⎠

Here τ ui are the experimentally identified shear

strength magnitudes acting the normal stresses σ i ; n is
the number of performed experimental tests.

Fig 4. Mohr–Coulomb envelope and K f line

2.4 τ–σ coordinate system

The major σ1 and minor σ3 principal stresses of the

sample are measured directly during triaxial testing proc-
Fig 6. State of stresses represented via coordinates τ–σ
ess. One faces the problem when identifying the failure
angle of the plane in which the shear and the normal
stresses τ and σ act (Fig 5). Therefore the τ–σ coordi- The magnitudes of the shear strength and normal
nate system is employed rather seldom. stresses depend on the angle α of the failure plane in re-
The soil shear strength τ u at failure plane depends spect of minor principal stress. The angle α depends on
the angle of internal friction. The magnitudes of the val-
on the normal stresses σ acting on the failure plane: n n n
τ u = tan ϕ ⋅ σ + c . (15) ues ∑=
i 1
τ ui σ i , ∑=
i 1
τ ui , ∑= σ
i 1
i of the relations (18) and

The normal stress at the failure plane is described (19) employed for determining the tan φ and the c de-
by: pend on the magnitudes of the angle of internal friction.
σ = (σ1 − σ3 ) cos 2 α + σ3 . (16) 3. Evaluation of shear strength parameters via triaxial
The shear stress at the failure plane is described by: testing by employing different methods

τ = 0.5 sin 2α(σ1 − σ3 ) . (17) The modified angle of internal friction and the cohe-
sion have been calculated following the methods de-
scribed in sections 2.1–2.3 (see Figs 7–9) by employing
the triaxial testing experimental data. The above men-
tioned magnitudes were converted into tan φ and c by
using the equations 2–3, 7–8, 13–14. Having performed
an analysis of the results one can find small variation of
the angle of internal friction, but the cohesion varies sig-
nificantly within bounds of 5 and 10 kPa (see Table 1).
Having processed the same experimental soil testing
data by different common methods one finally obtains
different magnitudes of the tan φ and the c. A question
arises: which magnitudes are true? The validation can be
Fig 5. State of stresses performed as follows.
By employing the each common method for calcu-
Having determined the σ and τ the mean magni- lated the tan φ and the c magnitudes one can identify the
tudes of the angle of internal friction and that of the cohe- normal and shear stresses acting on the failure plane.
sion are the parameters of the shear graph, obtained by Then applying the least squares method one can identify
employing the method of the least squares (Fig 6): the magnitudes tan φ* and c* according these magnitudes
(Eqs 18–19). If they coincide with the ones, obtained by
⎛ n n n ⎞
tan ϕ = ⎜ n ∑ τ σ − ∑ τ ∑ σ ⎟ Δ, (18) the common method, one can state the method to be ap-
⎜ ui i ui i ⎟
⎝ i =1 i =1 i =1 ⎠ plicable for processing of the shear strength parameters.
But results presented in Table 1 obviously show the dif-
⎛ n n n n ⎞ ference of tan φ, c and tan φ*, c* obtained via the differ-
c=⎜ ∑ τ ∑ σ 2
− ∑ σ ∑ τ σ ⎟ Δ, (19) ent methods.
⎜ i=1 i =1
ui i i ui i ⎟
⎝ i =1 i =1 ⎠
1079
Table 1. Mean values of soil shear strength parameters tan φ and c (in kPa) calculated by common and proposed methods
Proposed method
Common methods σ1-σ3 coordinate system q-p coordinate system t-s coordinate system τ-σ coordinate
system
Modified shear strength tan ψ = 1.553 tan α = 0.618
tan θ = 4.127 c θ = 43.5 –
parameters c ψ = 11.99 c α = 4.05

Modified shear strength

parameters converted tan ϕ = 0.780
tan ϕ = 0.770 c = 10.71
into tan ϕ and c (according to tan ϕ = 0.784 c = 6.05 tan ϕ = 0.786 c = 5.16
c = 7.27
formulae
2,3,7,8,13,14,18,19)
Shear strength parameters tan ϕ∗ = 0.780 tan ϕ∗ = 0.780
calculated according to σ tan ϕ∗ = 0.780 c∗ = 7.24 tan ϕ∗ = 0.780 c∗ = 7.29
and τ acting in failure plane
∗
c = 7.28 c∗ = 7.27

The obtained magnitudes of the tan φ* and the c* 4. Analysis of methods for evaluating shear
(see Table 2) differ from the converted tan ϕ and c strength parameters
ones (see Table 2). Thus the failure plane positions are
wrongly positioned for different common methods when When one employs the data only of the two tests the
determining the soil shear strength parameters. magnitudes of the values tan ϕ and c being evaluated by
the different methods coincide. The same result is ob-
1500
tained for larger number of tests if the functional relation-
y = 4.126667x + 43.500000
1250
R 2 = 0.967559
ship between principal values exists. But when process-
1000 ing the experiments with significant variation of results
one obtains the different magnitudes of tan ϕ and c
kPa

750
1,

σ when employing the different methods.

500
The variation of the shear strength parameters being
250 identified by the different methods is influenced by the
0 failure plane location α = 45º + φ/2, treated differently
0 50 100 150 200 250 300 350 for separate methods. The different parameters are mini-
σ3, kPa mized in different coordinate systems, yielding the differ-
Fig 7. Stresses conditions at failure defined by the ent failure planes, subsequently resulting the different
coordinates σ1–σ3 magnitudes of the shear strength parameters.
At the standard σ1–σ3 coordinate system case one

∑ (σ1i − σ1u ) 2 of
1200 n
1000
minimizes the differences squares sum
y = 1.553023x + 11.991076 i =1
the major principal stress magnitudes σ1i being identified
R2 = 0.986967
800
q, kPa

600
experimentally and that of the soil shear strength σ1u
400
magnitudes being calculated by Eq (1).
200 At t–s coordinate system case one minimizes the dif-

∑
n
− t ui )2 of the largest shear
0
0 100 200 300 400 500 600 700 ferences squares sum (t i
p, kPa i =1
stress t i = σ1i − σ 3i / 2 and that of the soil shear strength
Fig 8. Stresses conditions at failure defined by the
coordinates q–p
t ui magnitudes being identified according to the Col-
600 oumb law t ui = si ⋅ tan α + c α .
y = 0.618178x + 4.052133
500 R2 = 0.991780 At q–p coordinate system case one minimizes

∑
n
− q ui )2 , id est the differences of principal meas-
400
(q i
t, kPa

300 i =1
200 ured stresses q i = σ1i − σ 3i and that of the qui being
100 identified according to the qui = pi ⋅ tan ψ + c ψ .
0
0 150 300 450 600 750 900

s, kPa
Fig 9. Stresses conditions at failure defined by the
coordinates t–s
1080
5. Proposed method for identifying shear strength
tan 2 ϕ
∑(
1 n
parameters due triaxial testing data 0.5 ⋅ ( )− σ1 − σ 3 ) ⋅ ((1 − 2
) / 2) 2 ⋅
1 + tan 2
ϕ i =1 tan ϕ+ 1
What one should minimize actually if the minimiz-
∑( ∑(
n 1 n
ing magnitudes differ for different coordinate systems? σ1 − σ 3 ) 2 ⋅ 0.5 ⋅ ( )− σ1 − σ 3 ) 2 ⋅ ((1 −
The proposed method offers to minimize the differ- i =1 1 + tan 2
ϕ i =1

∑ (τi − τui )
n
2
ences squares sum of the soil shear strengths tan 2 ϕ
∑(
n 1
i =1 2
) / 2) ⋅ σ1 ⋅ σ 3 ) ⋅ 0.5 ⋅ ( )−
individual magnitudes being identified experimentally tan ϕ+ 1 i =1 1 + tan 2 ϕ
and that of the soil shear strength magnitudes being cal- 2
tan 2 ϕ
∑(
n 1
culated according to Coloumb law. Here one replaces τi σ1 − σ 3 ) ⋅ 0.5 ⋅ ( ) ⋅ ((1 − ) / 2) ⋅
i =1 1 + tan 2 ϕ tan 2 ϕ + 1
by the magnitudes of shear stresses being identified by
the Eq (17), and replaces τui by the magnitudes identi-
∑ ∑( ∑
n n n 1
)⋅ 0.5 ⋅ ( )) /(n ⋅
fied by the Eq (15). σ in the Eq (15) is replaced by the
σ3 − σ1 ⋅ σ 3 σ3 ⋅
i =1 i =1 i =1 1 + tan 2 ϕ
magnitudes of the normal stresses obtained by the Eq
tan 2 ϕ
∑ ∑(
(16). The failure plane angle in respect of minor princi- n n
pal stress is calculated by α = 45º + φ/2. Thus, one ob- ( (σ1 − σ 3 ) 2 ⋅((1 − 2
) / 2) 2 + 2⋅ σ1 ⋅ σ 3 )⋅
i =1 tan ϕ+ 1 i =1
tains:
tan 2 ϕ
∑ ∑(
n n
2
2 2 ((1 − ) / 2) + σ3 )−( σ1 − σ 3 )) 2 ⋅

∑(
n tan
(tan ϕ ⋅ (n ⋅ ( σ1 − σ 3 ) ⋅ ((1 −
ϕ
) / 2) 2
+ 2⋅ tan 2 ϕ + 1 i =1 i =1
i =1 tan 2 ϕ + 1
tan 2
∑( ∑
ϕ
n n
tan 2 ϕ ) / 2) 2 ((1 −
∑ (σ1 ⋅ σ 3 ) ⋅ ((1 − ∑σ − (∑ (σ1 −
n n n ((1 − − 2⋅ σ1 − σ 3 )⋅ σ3 ⋅
) / 2) + 2
3) tan 2 ϕ+ 1 i =1 i =1
i =1 tan 2
ϕ +1 i =1 i =1

tan 2 ϕ
∑ σ3 ) 2 .
n
tan 2 ϕ
∑( ∑ )/2−(
n n
)) 2
((1 − ) / 2) 2
2⋅ )⋅ (22)
ϕ +1
σ3 ⋅ − σ1 − σ 3 σ3 ⋅ 2
tan 2 ϕ + 1 i =1 i =1 tan i =1

tan 2 ϕ Having solved the equation (21) in respect of the

∑ ∑(
n n
((1 − )/2−( σ3 )2 − (n ⋅ ( σ1 − σ 3 ) 2 ⋅ 0.5 ⋅ single unknown tan φ, one obtains the tan ϕ = 0.780
tan 2 ϕ + 1 i =1 i =1
magnitude. Then c = 7.27 kPa is determined via equation
tan 2 ϕ
∑(
1 n (22) (see Table 2). The latter magnitudes correspond the
((1 − )/2⋅( )+ σ
1 ⋅σ
3 ) ⋅ 0.5 ⋅ magnitudes of tan φ and c, being obtained by minimiz-
tan 2 ϕ + 1 1 + tan 2 ϕ i =1
ing the differences squares sum of the individual soil
tan 2 ϕ shear strength parameters and the magnitudes calculated
∑ (σ1 − σ 3 ) 2 ⋅ 0.5 ⋅((1 −
1 n
( )− ) / 2) according the Coloumb law. The latter magnitudes are
1 + tan 2 ϕ i =1 tan 2 ϕ + 1 different from magnitudes being obtained by the common
methods (see Table 2).
The available arguments of tan φ magnitudes in-
∑( ∑
n n
⋅ (
1
)− σ1 − σ 3 )⋅ σ3 ⋅ 0.5 ⋅ cluded the equation f(tan φ) = 0 are presented in Fig 10.
1 + tan 2 ϕ i =1 i =1 One can find from the graph that equation has a single
solution.
1
( )) = 0) ; (21) x 10
6
2
1 + tan ϕ 2

∑( ∑(
n 1 n
c= σ1 − σ 3 ) ⋅ 0.5 ⋅ ( )⋅ σ1 − σ 3 ) ⋅ ((1 − -2
i =1 1 + tan 2 ϕ i =1
-4
2
∑( ∑(
tan ϕ
n n
) / 2) 2 + 2⋅ σ1 ⋅ σ 3 )⋅ σ1 − σ 3 ) ⋅ ((1 − -6

tan 2 ϕ + 1 i =1 i =1
-8

2
∑ ∑(
tan ϕ 1 n n
2 -10
) / 2) ⋅ 0.5 ⋅ ( )+ σ3 ⋅ σ1 − σ 3 )⋅ -0.2 0 0.2 0.4 0.6 0.8 1 1.2
tan 2 ϕ + 1 1 + tan 2
ϕ i =1 i =1
Fig 10. Graph of equation f(tan φ) = 0

1081
6. Data of bearing resistance calculation
The characteristics values according to the require-
ments of EC 7 of converted shear strength parameters
obtained by the common methods of processing of the
triaxial testing data were calculated for coefficients of
variation Vφ= 0.05 and Vtanφ= 0.3 (Schneider 1999; Fellin
et al. 2005; Amšiejus and Dirgėlienė 2007). The magni-
tudes of angle of internal friction were obtained to be
similar when the cohesion varied from 2 kPa till 5 kPa
(see Table 2).
The design bearing resistance R d was calculated for
drained conditions by means of the methods provided in
standard documents EC 7. A spread foundation is loaded
centrically. The foundation width (see Table 2) was cal-
culated according to the design approach 3 by employing
the design values obtained from the characteristics values
of the soil shear strength parameters. The foundation
width varies by a negligible margin, id est within bounds
of 1.70 m and 1.75 m. Obviously foundation width var-
ies more for the larger Vϕ and Vc .
Table 2. Results of calculation of design bearing resistance
according EC7
σ1-σ3 q-p coor- t-s coordi-
Methods coordinate dinate nate sys-
system system tem
Characteristics
values of shear tan ϕk = 0.705 tan ϕk = 0.719 tan ϕk = 0.721
strength parame-
ters tan φ and c,
c k = 3.1 c k = 2.6
c k = 5.4
(in kPa)
Foundation width
magnitude B, 1.70 1.74 1.75
(in m)
The characteristics values of the tan ϕ k = 0.716 ,
c k = 3.7 kPa and subsequently the foundation width
b = 1.73 m were calculated for the shear strength pa-
rameters evaluated by proposed method for variation
coefficients Vϕ = 0.05 and Vc = 0.30 .
Conclusions
1. Currently the shear strength parameters identified
by employing the triaxial testing are evaluated by the
different methods. The magnitudes of the angle of inter-
nal friction and the cohesion are obtained differently for
the same primary testing data. The internal friction angle
magnitude varies insignificantly, when comparative dif-
ference of the cohesion reaches 51 %.
2. The shear strength parameters being processed via
the different methods do not coincide as the different
magnitudes are minimized for identifying their mean
values by the least squares method, e.g. differences
squares sum of the largest principal stresses or the princi-
1082
pal stresses from their calculated magnitudes, being cal-
culated by regression equation.
3. The mean shear strength parameters obtained by
the proposed method are calculated by minimizing the
differences squares sum of the individual shear strengths
at the failure plane and that of being calculated by the
regression equation.
References
Aysen, A. 2005. Soil Mechanics: Basic Concepts and Engineer-
ing Applications. New York: Taylor & Francis. 457 p.
ISBN 0-415-38393-5.
Amšiejus, J.; Dirgėlienė, N. 2007. Probabilistic assesment of
soil shear strength parameters using triaxial test results,
The Baltic Journal of Road and Bridge Engineering 2(3):
125–131.
Amšiejus, J.; Dirgėlienė, N.; Norkus, A.; Žilionienė, D. 2009.
Evaluation of soil shear strength parameters via triaxial
testing by height versus diameter ratio of sample, The Bal-
tic Journal of Road and Bridge Engineering 4(2): 54–60.
Bukhartsev, V. N. 1988. Statistical procccesing of results of
stabilometer tests, Soil Mechanics and Foundation Engi-
neering 25(6): 562–565. doi:10.1007/BF01721620
СНиП 2.02.02-85. Основания гидротехнических сооружений
[Foundation Beds of Hydraulic Structures]. Госстрой
СССР. Москва: Стройиздат, 1986. 48 с.
Craig, R. F. 2004. Craig’s Soil Mechanics. London: Spon Press.
447 p. ISBN 0-415-32703-2.
Dirgėlienė, N. 2007. Grunto stipruminių ir deformacinių savy-
bių tyrimas triašio slėgio aparate bei jų tikimybinis
vertinimas [Research of soil shear strength in triaxial tests
and probabilistic assessment of results]. Daktaro dis-
ertacija 2007. Vilnius: Vilniaus Gedimino technikos uni-
versitetas. 113 p.
Eurocode 7: Geotechnical design. Part 1: General rules. Brus-
sels, 2004. 168 p.
Fellin, W.; Lessmann, H.; Oberguggenberger, M.; Vieider, R
(eds). 2005. Analyzing Uncertainty in Civil Engineering.
Berlin: Springer–Verlag. 242 p. ISBN 3-540-22246-4.
Флорин, В. А. 1959. Основы механики грунтов: Общие
зависимости и напреженное состояние оснований
сооружений.[Basic soil mechanics: relations and ground
stresses condition]. Ленинград-Москва: Стройиздат.
356c.
Parry, R. H. G. 2004. Mohr Circles, Stress Paths and Geotech-
nics. London: Spon Press. 264 p. ISBN 0-415-27297.
Ranjan, G.; Rao, A. S. R. 2005. Basic and Applied Soil Me-
chanics. New Delhi: New Age International (P). 759 p.
ISBN 81-224-1223-8.
Schneider, H. R. 1999. Determination of characteristic soil
properties, in Proc. XII International Conference on Soil
Mechanics and Geotechnical Engineering, vol. 4. Ed. by
Barends et al. 1999, Hamburg. Rotterdam: Balkema,
2271–2274. ISBN 90-5809-047-7.
Vervečkaitė, N.; Amšiejus, J.; Stragys, V. 2007. Stress-strain
analysis in the soil sample during laboratory testing, Jour-
nal of Civil Engineering and Management 13(1): 63–70.