See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/349517111

Subgrade reaction modulus of rock masses under the load of single and
multiple footings

Article  in  Geomechanics and Geoengineering · February 2021

DOI: 10.1080/17486025.2021.1889687

CITATION READS

1 4,028

2 authors:

Saeed Shamloo Meysam Imani

Amirkabir University of Technology Amirkabir University of Technology
8 PUBLICATIONS   22 CITATIONS    24 PUBLICATIONS   152 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Earth Pressure View project

Upper bound stability analysis of crushed rock slopes considering the Hoek-Brown non linear failure criterion View project

All content following this page was uploaded by Saeed Shamloo on 25 February 2021.

The user has requested enhancement of the downloaded file.

 Geomechanics and Geoengineering
An International Journal

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tgeo20

Subgrade reaction modulus of rock masses under

the load of single and multiple footings

Saeed Shamloo & Meysam Imani

To cite this article: Saeed Shamloo & Meysam Imani (2021): Subgrade reaction modulus of rock
masses under the load of single and multiple footings, Geomechanics and Geoengineering, DOI:
10.1080/17486025.2021.1889687

To link to this article: https://doi.org/10.1080/17486025.2021.1889687

Published online: 22 Feb 2021.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tgeo20
GEOMECHANICS AND GEOENGINEERING
https://doi.org/10.1080/17486025.2021.1889687

Subgrade reaction modulus of rock masses under the load of single and multiple
footings
Saeed Shamloo and Meysam Imani
Geotechnical Engineering Group, Amirkabir University of Technology, Garmsar, Iran

ABSTRACT ARTICLE HISTORY

Subgrade reaction modulus is an essential parameter in the structural design of footings. Received 5 July 2020
Determination of this parameter is difficult, especially in rock masses containing an irregular dis­ Accepted 9 February 2021
tribution of discontinuities. In this paper, the subgrade reaction modulus of rock masses obeying KEYWORDS
Hoek-Brown failure criterion was investigated. It was assumed that the rock masses are under the Hoek-Brown; interference
pressure of a single isolated footing and also two adjacent strip footings. The subgrade reaction coefficient; multiple footing;
modulus was determined from the pressure-settlement curve of the rock masses. This study’s results rock foundation; subgrade
were presented in two forms, which include the subgrade reduction factor and the interference reaction modulus
coefficient. The obtained results show that the subgrade reaction modulus of rock masses may be
smaller than 50% of intact rocks. Likewise, based on the distance between the multiple footings, the
subgrade reaction modulus may be in the range of 0.7 to 1.3 times the subgrade reaction modulus in
the case of a single isolated footing.

1. Introduction footing width. This method was initially proposed for

a 30 cm×30 cm concrete rigid slab rested on the soil
The analysis of the interaction between footings and
bedding surface. The obtained results show that the ks
subsurface materials is a challenging problem in geo­
value depends on the dimensions of the footing.
technical engineering. In some footing-bedding inter­
Vesic (1961) showed that ks is related to both the soil
action analysis methods, the soil/rock bedding
and structural stiffness. He proposed a precise relation
material is replaced by a simple system that contains
for ks, which is presented in Equation (4):
a collection of springs. This approach has been used
extensively in the design of foundations by structural � � � �1=
0:65 Es Es � B4 12
engineers. The stiffness coefficient of these springs, ks ¼ � � (4)
which is named as the subgrade reaction modulus B 1 ν2s EI
(SRM) in the current study, can be obtained using In which, Es (in the unit of force/area, like MPa) and νs
Equation (1): are the soil modulus of elasticity and the Poisson’s ratio,
ks ¼ q=δ (1) respectively, and EI (in the unit of force multiplied by
area, like kN.m2) is the footing flexural stiffness. For
In which q is the footing pressure and δ is the corre­ practical applications, one can use Vesic’s simplified
sponding settlement. After Biot (1937), many studies relation as follows:
were performed on the SRM, i.e., ks, of soils. Among
Es
others, Terzaghi (1955) proposed ks value for sand and ks ¼ � (5)
clay as Equations (2) and (3), respectively. B 1 ν2s
� � Vlassov and Leontiev (1966) also proposed the reaction
B þ 0:3 2
ks ¼ k0:3 (2) coefficient for the beams and plates resting on an elastic
2B
semi-infinite medium as follows:
� �
0:3 Es ð1 νs Þ μ
ks ¼ k0:3 (3) ks ¼ � (6)
B ð1 þ νs Þð1 2νs Þ 2B
In which k0.3 (in the unit of force/volume, like MN/m3) The determination of the dimensionless factor μ is ambig­
is the value of ks from a plate load test, and B (m) is the uous, which makes the use of this method very complicated

CONTACT Meysam Imani [email protected]

© 2021 Informa UK Limited, trading as Taylor & Francis Group
2 S. SHAMLOO AND M. IMANI

(Sadrekarimi and Akbarzad 2009). Likewise, Imanzadeh The method presented by Lee and Jeong (2016) is
et al. (2013) showed that soil stiffness has the most signifi­ applicable only for the jointed rocks containing specific
cant effect on the uncertainty of the SRM, while other joints and can not be used in general cases. However, in
parameters like structural properties have little importance. many practical projects, engineers are faced with rock
As another research, by comparing the results of finite masses that contain irregular fractures with no distinct
element models with those obtained from field plate load joint sets. In such situations, the Hoek-Brown failure
tests, Avci and Gurbuz (2018) showed that the modulus of criterion (Hoek and Brown 1980) has widely been
subgrade reaction decreases with the increasing settlement applied by the engineers. To the authors’ knowledge,
of the soils. Many other kinds of research were performed no precise method is available for obtaining the SRM of
to determine soil horizontal SRM in buried circular chan­ the rock masses obeying the Hoek-Brown failure criter­
nels (Meyerhof and Baikie 1963, Klopple and Glock 1979, ion, and the present paper seems to be the pioneer in
Selvadurai 1985, Ziaie Moayed and Naeini 2006). this field. This issue is of paramount importance since
Because of the complexity of rock media, rock many massive structures, like a bridge, are usually con­
masses’ load-displacement behaviour has gained little structed on crushed rock masses, and structural engi­
attention in past research. Among others, Carter and neers need the SRM of the underlying rock mass for
Kullhawy (1988) presented the load-displacement designing the foundation of the structure.
response and the corresponding ultimate bearing In this paper, by considering different properties for
capacity of rock masses subjected to loads of the rock masses, detailed sensitivity analyses were per­
socketed shafts. Also, Kullhawy and Carter (1992) formed to obtain the SRM of the rock masses. The
investigated different aspects of settlement and bear­ results of this study were presented as a factor named
ing capacity analysis of foundations on rock masses. ‘subgrade reduction factor’ that can be used for deter­
Using plasticity and finite element models, Alhossein mining the SRM of the rock masses. Simple tables were
et al. (1992) proposed solutions for strip footings on presented for obtaining the subgrade reduction factor,
a regularly jointed rock mass with one and two joint which can easily be used in practical applications.
sets. To the authors’ knowledge, the problem of Moreover, as a new subject in rock foundation pro­
determining the SRM for rock masses was not ade­ blems, the effect of multiple footings on the SRM was
quately investigated by researchers. Very few pub­ also investigated.
lished papers can be found regarding this issue in
the literature. The few available studies show that the
SRM of the rock masses is smaller than that of intact 2. Analysis method
rocks. Since the rock masses’ elastic modulus is The SRM is a parameter that can be obtained by
smaller than that of intact rocks and the SRM has dividing the footing pressure (q) by the correspond­
a positive correlation with the elastic modulus. ing settlement (δ). This parameter is not constant,
The experimental research performed by Lee and and it depends on the width and depth of the foot­
Jeong (2016) can be considered as the most recent ing. Two common methods are available for deter­
study regarding the SRM of jointed rocks. They mining this parameter, which includes the loading
constructed some artificial joints with different orien­ experiment on the base material and the numerical
tations and spacings in jointed rock specimens and analysis. The latter was used in the present study by
finally obtained the joint reduction factor (Jf) for applying the finite element software, Phase2.
each case. They suggested that the SRM of the
jointed rocks (kj) can easily be obtained from
Equation (7): 2.1. Boundary conditions
In the numerical models, the side boundaries were kept
Es fixed in the horizontal direction, while the base of the
kj ¼ Jf � ks ¼ Jf � � (7)
B 1 ν2s models was kept fixed in both the vertical and the
horizontal directions. Model meshing was performed
In which ks is the SRM of the intact rock, which can by 6 nodded triangular elements.
be obtained using Vesic’s simplified formula, i.e., An important issue in numerical modelling is the
Equation (5). This equation is widely used in prac­ distance between the footing edges to the sides and
tical applications for soil beddings. Es and νs are the the bottom boundaries of the model. Fattah et al.
elastic modulus, and the Poisson’s ratio of the intact (2014a, 2014b, 2014c) showed that these distances
rock, respectively, and B is the footing width. greatly influence the amount of footing settlement
 GEOMECHANICS AND GEOENGINEERING 3

and the corresponding bearing capacity. Lee and

Jeong (2016) considered these distances approxi­
mately equal to 3.5 times the footing width (i.e.,
3.5B). Lees (2016) proposed a distance approximately

pressure (q)
equal to 3B is enough. In the present study, to
minimise the boundary effects on the results, the
distance between the footing edge to the side bound­
SRM
aries was considered equal to 5B, while a distance
equal to 3.5B was considered from the footing base 1
to the bottom boundary of the model. Figure 1
represents the general shape of the numerical models
considered in the current research.
In the numerical models, the footing pressure was settlement (δ)

applied incrementally to the top boundary of the Figure 2. An example of a pressure-settlement curve and the
model at a rate equal to 2*10−5m/s. Lee and Jeong corresponding SRM.
(2016) were also considered the same loading rate in
their experimental tests.
Two sets of models were constructed in this was focused here. The pressure-settlement curve is
research, including the verification models and the generally nonlinear, but the curvature of its initial
sensitivity analysis models. In the verification mod­ part, before the yield point, is much smaller than the
els, the width of the footing was considered to be curvature of its post-yield point. As shown in Figure
equal to 8 cm, which is consistent with the models 2, an approximate line can be fitted to the initial part
constructed by Lee and Jeong (2016). However, in of the pressure-settlement curve. The slope of this
the sensitivity analysis models, the footing width was line was called the SRM. Lee and Jeong (2016) were
set to one metre. used this approach for obtaining the SRM of their
experimental models. In the present research, the
same approach was used for obtaining the SRM of
2.2. Determination of the SRM from the rock masses.
pressure-settlement curve
The pressure-settlement curve of the rock masses can
3. Verification of the numerical modelling
be drawn by applying an incremental footing pres­
sure and obtaining the corresponding settlements. The results of the present study were verified by the
This curve can be used to determine the bearing experimental tests performed by Lee and Jeong (2016).
capacity and the SRM of the rock masses. Most They constructed 21 jointed rock samples with dimen­
previous numerically obtained pressure-settlement sions equal to 48 cm*48 cm*28 cm. Each sample was
curves were used for determining the ultimate bear­ subjected to a distributed incremental load with a width
ing capacity of soil/rock masses [(Mabrouki et al. equal to 8 cm, and the corresponding pressure-
2010, Javid et al. 2015, Sargazi and Hosseininia settlement curve was obtained. Then, they obtained
2017, Mansouri et al. 2019). This issue is out of the the SRM as the slope of the nearly linear part of the
scope of this paper, and only determining the SRM curve occurred at the beginning of it.

Figure 1. Boundary condition for numerical model.

4 S. SHAMLOO AND M. IMANI

Table 1. The properties of the jointed rocks considered in the intact rocks (ks) and the joint reduction factor, Jf.
analyses. According to Lee and Jeong (2016), the SRM of the
Parameter Magnitiude intact rock can be obtained using Vesic’s simplified
σci (MPa) 15
Sd (cm) 4, 8
relation (Equation (5)) that was initially proposed for
Id (degree) 0, 30, 60, 90 soil beddings. The Jf can also be obtained from the
Es (MPa) 860 simple chart provided in their study. As another ver­
Poisson’s ratio (ν) 0.3
ϕj (degree) 35 ification of the present research, the joint reduction
cj (MPa) 0 factor, Jf, was obtained using the numerical method
Kn (GPa/m) 100
Ks (GPa/m) 10 and was compared with the experimental tests (Lee
γ (MN/m3) 0.027 and Jeong 2016). According to Equation (7), the Jf is
the ratio of SRM of jointed rocks to the intact rocks.
All the 21 specimens constructed by Lee and Jeong
3.1. The pressure-settlement curve and the (2016) were analysed numerically in the present
corresponding SRM obtained paper, and the obtained Jf were presented in Table 3.
This table shows that the difference between the
For all the 21 samples, the numerical models were
numerical analyses and the experimental tests is 2 to
constructed in the current study, and the pressure-
9%, which is usually appropriate in practical engineer­
settlement curves were compared with those
ing applications. The experimental tests performed by
obtained by Lee and Jeong (2016). The properties
Lee and Jeong (2016) show that SRM of the jointed
considered for the jointed rock are presented in
rocks is about 43% to 91% of SRM of the intact rocks,
Table 1. As an example, for the eight models
while the numerical results reveal a magnitude of
shown in Figure 3, the results of this comparison
about 54% to 86%.
were shown in Figure 4. As shown in Figure 4, the
overall trend of the curves obtained from the numer­
ical models is approximately similar to that obtained
3.3. The effect of the joint parameters on the
from the experimental tests. The most similarity of
verification results
the curves obtained from these two methods was
seen in the approximately linear part at the begin­ As discussed previously, Lee and Jeong (2016) did
ning of the curves. It means that the slope of this not mention the unit weight of their samples and
part of the curves, which corresponds to the SRM, is also the joints cohesion, friction of angle, and the
very close to each other in numerical and experi­ normal and shear stiffness. The effect of these para­
mental models, which shows the applicability of the meters on the numerical results was checked in this
numerical models constructed in the current section for model 6 of Figure 3.
research. Table 2 presents a comparison between Lee and Jeong (2016) ignored the filling material
the SRM obtained from the present study and the in the joints for their samples. Therefore, the cohe­
Lee and Jeong (2016) experimental tests. sion along the joints is low and can be considered
Unfortunately, Lee and Jeong (2016) did not men­ equal to zero. On the other hand, different investiga­
tion five parameters of their samples, including the tions show that the unit weight of the rock mass
unit weight of the rock mass, the cohesion and the does not have a considerable effect on the strength
friction angle of these joints and also the normal and of the rock mass (Yang and Yin 2005, Saada et al.
shear stiffness along the joints. In the current section, 2008, AlKhafaji et al. 2020). As a result, only the
an approximate logical magnitude was assumed for effects of the joint friction angle and the normal
these parameters, as presented in Table 1. However, and shear stiffness were investigated here. Table 4
the sensitivity of the results to these assumed values presents the magnitudes considered for these para­
was investigated later in section 3.3. meters. In all cases, the joint shear stiffness was
taken into account equal to 0.1 times the joint nor­
mal stiffness. Other required parameters were select
3.2. The joint reduction factor from Table 1.
Determination of the SRM by using the in situ tests is Figures 5 and 6 show the effect of the joint fric­
too difficult, especially in rock masses. So, by using tion angle and the normal and shear stiffnesses,
Equation (7), one can easily obtain the SRM for the respectively. These figures reveal that the mentioned
jointed rock masses (kj) by having the SRM of the joint parameters do not have a considerable effect on
 GEOMECHANICS AND GEOENGINEERING 5

Figure 3. Configuration of some of the samples tested by .Lee and Jeong (2016)

the SRM and the bearing capacity of the jointed 4. Effect of the rock mass properties on the SRM
rocks. As an overall conclusion from the current In many practical rock engineering projects, rock
section, the numerical results obtained in this paper masses contain a considerable amount of randomly
are in reasonable conformity with the experimental distributed discontinuities. The analysis of such rock
results.
6 S. SHAMLOO AND M. IMANI

30 30

25 25

Pressure (MPa)
20 20

Pressure (MPa)
15 15

10 10
Present study Present study
5 5
Lee & Jeong (2016) Lee & Jeong (2016)
0 0
0 2 4 6 8 0 2 4 6 8
Sattlement (mm) Sattlement (mm)

Model 1 Model 2

25 30

20 25
Pressure (MPa)

20

Pressure (MPa)
15
15
10
10
Present study Present study
5 5
Lee & Jeong (2016) Lee & Jeong (2016)
0 0
0 2 4 6 8 0 2 4 6 8
Settlement (mm) Settlement (mm)

Model 3 Model 4

30 30

25 25
Pressure (MPa)

Pressure (MPa)

20 20

15 15

10 10
Present study Present study
5 5
Lee & Jeong (2016) Lee & Jeong (2016)
0 0
0 2 4 6 8 0 2 4 6 8
Settlement (mm) Settlement (mm)

Model 5 Model 6

30 35

25 30
25
20
Pressure (MPa)
Pressure (MPa)

20
15
15
10
Present study 10 Present study
5 5 Lee & Jeong (2016)
Lee & Jeong (2016)
0 0
0 2 4 6 8 0 2 4 6 8
Settlement (mm) Settlement (mm)

Model 7 Model 8

Figure 4. Comparison between the results of the present study with .Lee and Jeong (2016)
 GEOMECHANICS AND GEOENGINEERING 7

Table 2. Comparison between the SRM obtained from the pre­

sent study and the experimental tests performed by Lee and
Jeong (2016).
Present study Lee and Jeong (2016)
Model (Values in MN/m3) (Values in MN/m3) Difference (%)
1 6137 7220 15
2 6740 7798 13
3 4740 4780 1
4 5390 5480 2
5 5894 6370 7
6 6672 7020 5
7 6980 8340 16
8 7868 9160 14

masses has usually been performed by considering

a homogeneous isotropic rock mass with the Hoek-
Brown failure criterion. Equation (8) defines the gen­
eral form of the Hoek-Brown failure criterion:
� �a Figure 5. Effect of joints friction angle on pressure-settlement
σ3 curve.
σ 1 ¼ σ 3 þ σ ci mb þ s (8)
σ1
In which, σ1 and σ3 are the major and minor principal 35
stresses, respectively, σci is the uniaxial compressive
strength of the intact rock, and mb, s, and a are the 30
Hoek-Brown parameters that can be obtained using
Equations (9)–(11), respectively. 25
� �
GSI 100
P ressure (MP a)

mb ¼ mi exp (9) 20
28 14D

� � 15
GSI 100
s ¼ exp (10) Kn=50GPa/m
Present study , Ks=5GPa/m
9 3D 10 Kn=100GPa/m , Ks=10GPa/m
25
Kn=150GPa/m , Ks=15GPa/m
30

1 1� GSI 20
�
5 Kn=200GPa/m , Ks=20GPa/m
40
a¼ þ e 16 e 3 (11)
2 6 Lee & Jeong (2016)
Lee&Jeong[13]

In which mi is an empirical constant for the intact rock, 0

0 2 4 6 8
GSI is the geological strength index, and D is the dis­ Settlement (mm)
turbance factor.
In the following sections, the SRM of the rock masses Figure 6. Effect of normal and shear stiffness of joints on
pressure-settlement curve.
obeying the Hoek-Brown failure criterion, i.e., ks(mass),

Table 3. Comparison between the joint reduction factor (Jf) obtained from the present study with Lee and Jeong (2016).
(Sd/B) Id = 0° Id = 30°
Lee & Jeong Present study Difference (%) Lee & Jeong Present study Difference (%)
0.5 0.63 0.68 5 0.43 0.54 9
1 0.73 0.75 2 0.58 0.65 7
1.5 0.83 0.79 4 0.63 0.69 6
2 0.85 0.80 5 0.67 0.75 4
Id = 60° Id = 90°
Lee & Jeong Present study Difference (%) Lee & Jeong Present study Difference (%)
0.5 0.58 0.64 6 0.71 0.66 5
1 0.64 0.71 7 0.78 0.75 3
1.5 0.71 0.74 3 0.86 0.84 2
2 0.73 0.75 2 0.91 0.86 5
8 S. SHAMLOO AND M. IMANI

Table 4. The properties considered for the rock joints. The elastic properties of the rock mass calculated
Parameter Magnitude using the following relations that were proposed by
ϕj (degree) 25, 30, 35, 40 Hoek and Brown (1980) and Vásárhelyi (2009),
Kn (GPa/m) 50, 100, 150, 200
Ks (GPa/m) 5, 10, 15, 20 respectively:
� �rffiffiffiffiffiffiffi
D σ ci
� 10ð 40 Þ
GSI 10
Table 5. Rock mass properties considered for sensitivity analyses. Em ðGPaÞ ¼ 1 (12)
2 100
Parameter Magnitude
σci (MPa) 5, 25, 50, 75, 100
GSI 10, 30, 50, 70, 90 νm ¼ 0:002GSI 0:003mi þ 0:457 (13)
mi 5, 10, 15
D 0, 0.5, 0.8 In which Em and νm are the rock mass deformation
modulus and the Poisson’s ratio, respectively, and the
Hoek-Brown parameters, D, σci (MPa), GSI, and mi
was investigated considering all combinations of the
were selected from Table 5.
parameters presented in Table 5. The ks(mass) is the slope
of the initial approximately linear part of the pressure-
settlement curve obtained from numerical models. The
4.1. Effect of σci
rock mass was considered to be under the pressure of
a strip footing with one-metre width. Due to a large Figure 7 shows the effect of σci on the ks(mass), considering
number of the models, only the results of some of them different Hoek-Brown parameters. It can be seen that
were presented in the following sections. increasing the σci results in increasing the ks(mass). The

Figure 7. Effect of σci on ks,mass assuming: (a) D = 0 and mi = 10 (b) D = 0 and GSI = 30, and (c) mi = 10 and GSI = 30.
 GEOMECHANICS AND GEOENGINEERING 9

Figure 8. Effect of GSI on ks,mass assuming: (a) D = 0 and mi = 10, (b) D = 0 and σci = 25 MPa and (c) mi = 10 and σci = 25 MPa.

Table 6. Comparison between the ks(mass) (MN/m3)obtained from

the present study with Lee and Jeong (2016) for the case of D = 0 rate. This conclusion means that for the rock masses
mi = 10 and GSI = 88. with GSI>60, the ks(mass) is considerably affected by GSI.
σci = 24 MPa σci = 15 MPa Additionally, it can be seen that among the Hoek-Brown
Present study 21743 12478 parameters, mi has the lowest effect on the ks(mass).
Lee and Jeong (2016) 20880 11814
Difference(%) 4 5
It is interesting to note that Figure 8(a) shows Another
verification of the current study. As can be seen, the ks
obtained by Lee and Jeong (2016) for the rock masses
with σci = 24MPa and 15 Mpa are equal to 20,880 MN/m3
ks(mass) value is less affected by mi, while D and GSI sig­ and 11,814 MN/m3, respectively, which were shown in
nificantly influence the ks(mass). Figure 8(a) as two single points. These two points were
obtained from the experimental tests performed by Lee
and Jeong (2016) on the samples, not including the joints.
4.2. Effect of GSI Naturally, such samples behave as rock masses with large
Figure 8 shows the effect of GSI on the ks(mass), consider­ GSI values, i.e., approximately GSI = 90. The ks for the
ing different Hoek-Brown parameters. It is clear that mentioned σci and GSI were also obtained from the cur­
increasing the GSI results in increasing the ks(mass). The rent study, and the results were presented in Table 6. This
rate of this increment is not considerable for the case of comparison shows a good agreement between the current
GSI<60, while for GSI>60, ks(mass) increases by a high study and the experimental tests.
10 S. SHAMLOO AND M. IMANI

Figure 9. Effect of mi on ks,mass assuming: (a) D = 0 and GSI = 30 (b) D = 0 and σci = 25 MPa (c) σci = 25 MPa and GSI = 30.

4.3. Effect of mi B magnitudes that are common in practice were con­

sidered, and the obtained ks(mass) was shown in Figure
Figure 9 presents the effect of mi on the ks(mass). As
11. The figure shows that the footing’s width has
described in the previous subsection, mi has not
a considerable effect on the ks(mass) of the rock masses.
a significant influence on the ks(mass); only a small
By increasing B, the rock mass settlement beneath the
increase in ks(mass) can be seen by increasing mi.
footing will increase, which results in the reduction in
the ks(mass) (see Equation (1)). This reduction rate is
4.4. Effect of D paramount in small magnitudes of B, especially in
large values of σci.
Figure 10 shows the effect of D on the ks(mass). By
increasing the disturbance of the rock mass, ks(mass)
will decrease. The rate of reduction of ks(mass) is widely 6. Suggesting a new coefficient for computing
affected by σci. As shown in Figure 10(a), increasing the SRM of rock masses
D results in decreasing the ks; the larger the σci, the
more reduction in the ks. 6.1. Subgrade reduction factor, Jmass
The Hoek-Brown failure criterion is widely used in
practical rock engineering projects that can implicitly
5. Effect of the footing width on the SRM
consider the discontinuities of rock masses. Reviewing
In the sensitivity analyses performed in the previous the literature shows that no research has been per­
sections, the width of the footing was taken into account formed to obtain the SRM of the Hoek-Brown rock
equal to one metre. In order to check the effect of the masses subjected to a load of strip footings. Therefore,
footing width on the ks(mass) of the rock masses, several it is of paramount importance to have a simple method
 GEOMECHANICS AND GEOENGINEERING 11

Figure 10. Effect of D on ks,mass assuming: (a) GSI = 30 and mi = 10, (b) mi = 10 and σci = 25 MPa and (c) GSI = 30 and σci = 25 MPa.

to determine the SRM of rock masses in practical appli­ Brown parameters of the rock mass. Then, using Equation
cations without performing time-consuming and com­ (15) proposed by Hoek and Dierichs (2005), one can
plicated numerical analyses. In this section, a new obtain the modulus of elasticity of the intact rock, Ei, that
coefficient named subgrade reduction factor, Jmass, was can be used in Vesic’s simplified formula.
proposed that can be used easily to determine the SRM
of rock masses obeying the Hoek-Brown failure criter­ Em
Ei ¼ (15)
ion. The Jmass can be computed as follows: 0:02 þ 1þexp1 60þ15D
D=2
ð 11 GSI
Þ
q �
ksðmassÞ δ qB 1 ν2i
Jmass ¼ ¼ Ei ¼ (14) Using Equation (14), the Jmass factor was obtained for
ks δEi
Bð1 ν2i Þ various properties of the rock masses and were pre­
sented in Tables 7–9. As can be seen from these tables,
In which, ks(mass) is the SRM of rock masses obeying the by increasing σci, GSI, and mi, the Jmass increases, which
Hoek-Brown failure criterion that was obtained using the means increasing the SRM of the rock mass. However,
numerical analyses performed in the present study. Also, ks increasing D results in decreasing the Jmass.
is the SRM of the intact rock that can easily be computed In practical engineering projects, having the SRM of
using Vesic’s simplified formula for soils (Equation (5)). intact rocks ks from Equation (5), one can easily obtain
For using Equation (5) for intact rocks, one should have the SRM of the corresponding rock mass ks(mass) as
the modulus of elasticity and the Poisson’s ratio of the follows:
intact rock (Ei and νi, respectively) and the footing width,
B. The corresponding modulus of deformation of the rock ksðmassÞ ¼ Jmass ks (16)
mass can be obtained by Equation (12), using the Hoek-
12 S. SHAMLOO AND M. IMANI

Figure 11. Effect of B on ks,mass assuming: (a) D = 0, mi = 15 and GSI = 50 (b) σci = 25 MPa, D = 0 and mi = 15 (c) σci = 25 MPa, D = 0 and
GSI = 50.

Table 7. Values of Jmass for the case of mi = 5. Table 8. Values of Jmass for the case of mi = 10.
σci(MPa) σci(MPa)
GSI D 5 25 50 75 100 GSI D 5 25 50 75 100
10 0 0.004 0.006 0.008 0.010 0.011 10 0 0.007 0.008 0.009 0.011 0.012
0.5 0.003 0.005 0.006 0.008 0.009 0.5 0.005 0.006 0.008 0.009 0.010
0.8 0.002 0.003 0.004 0.005 0.006 0.8 0.002 0.004 0.005 0.006 0.007
30 0 0.020 0.025 0.027 0.033 0.034 30 0 0.023 0.029 0.030 0.034 0.037
0.5 0.009 0.013 0.014 0.015 0.016 0.5 0.011 0.014 0.015 0.016 0.017
0.8 0.006 0.007 0.009 0.010 0.011 0.8 0.007 0.010 0.012 0.013 0.014
50 0 0.071 0.083 0.124 0.127 0.128 50 0 0.078 0.102 0.124 0.128 0.130
0.5 0.026 0.038 0.049 0.056 0.060 0.5 0.035 0.046 0.053 0.060 0.061
0.8 0.017 0.024 0.034 0.042 0.044 0.8 0.020 0.027 0.040 0.044 0.047
70 0 0.129 0.244 0.291 0.321 0.323 70 0 0.158 0.267 0.341 0.343 0.350
0.5 0.094 0.150 0.154 0.171 0.194 0.5 0.102 0.164 0.175 0.177 0.226
0.8 0.055 0.109 0.110 0.113 0.121 0.8 0.065 0.113 0.119 0.120 0.122
90 0 0.271 0.400 0.463 0.512 0.523 90 0 0.290 0.459 0.474 0.518 0.526
0.5 0.114 0.286 0.303 0.318 0.354 0.5 0.176 0.306 0.331 0.332 0.362
0.8 0.090 0.225 0.226 0.227 0.229 0.8 0.117 0.241 0.243 0.244 0.253
 GEOMECHANICS AND GEOENGINEERING 13

Table 9. Values of Jmass for the case of mi = 15.

σci(MPa)
GSI D 5 25 50 75 100
10 0 0.008 0.010 0.011 0.012 0.013
0.5 0.006 0.007 0.009 0.009 0.010
0.8 0.003 0.005 0.007 0.008 0.009
30 0 0.032 0.033 0.034 0.035 0.037
0.5 0.013 0.015 0.016 0.017 0.019
0.8 0.009 0.012 0.013 0.014 0.015
50 0 0.081 0.118 0.134 0.136 0.139
0.5 0.038 0.055 0.060 0.061 0.062
0.8 0.022 0.034 0.042 0.044 0.049
70 0 0.173 0.273 0.348 0.349 0.356
0.5 0.107 0.173 0.178 0.180 0.241
0.8 0.070 0.109 0.120 0.122 0.133
90 0 0.293 0.462 0.508 0.531 0.535
0.5 0.179 0.316 0.333 0.340 0.376
0.8 0.121 0.244 0.245 0.249 0.260
Figure 12. Comparison between the ks, mass obtained from the
present study and the Wyllie (1999) method.
6.2. Verification of the proposed method
In this section, some practical examples were solved to
show the applicability of the presented method. Wyllie Equation (16), and that obtained from the Wyllie
(1999) proposed that the settlement of rock masses can (1999) method, i.e., Equation (18). Good confor­
be calculated as follows: mity between the results of the two methods can
be seen, which implies the applicability of the pre­
Cd qBð1 ν2 Þ sented method.
δv ¼ (17)
E
In which Cd is the shape factor, q is the uniformly
distributed bearing pressure, B is the footing width, 7. The SRM of the rock masses in the case of
and E and ν are the rock mass deformation modulus two adjacent footings
and Poisson’s ratio, respectively. Therefore, using Some structures founded on the rock masses, like
Equation (1), which is the general form of the bridges, may have multiple footings that have been
SRM, one can obtain the SRM of rock masses as constructed close to each other. It is necessary to con­
follows: sider the interactive effects of the two adjacent footings
q E on the SRM of the underneath rock mass in such cases.
ks; Wyllie ð1999Þ ¼ ¼ (18) In the scope of the authors’ knowledge, there is no
δv Cd Bð1 ν2 Þ
research in this field. So this paper can be considered
In which, E and ν can be calculated using a pioneer. Figure 13 shows a schematic of two adjacent
Equations 12 and 13, respectively, and the average footings that were examined here by applying numerical
value of Cd for a strip footing is equal to 2.25 analyses. The width of each footing (B) was considered
(Wyllie 1999). Assuming B = 1 m, mi = 15 and equal to one metre, and S is the distance between them.
D = 0, Figure 12 shows a comparison between the It was assumed that the pressure of both footings was
ks, mass obtained from the present study, i.e., exerted on the rock mass simultaneously. Considering

Figure 13. Schematic of two adjacent footings.

14 S. SHAMLOO AND M. IMANI

the problem’s symmetry, the pressure-settlement curve the two footings that corresponds to the maximum
of one of the footings was used to determine the SRM. value of α was introduced as the critical spacing, Scr.
In order to consider the interference effect of the two If the spacing of the two footings becomes larger
adjacent footings, an interference coefficient, α, was than Scr, the α coefficient decreases until reaching
defined as the ratio of the SRM of the rock mass sub­ its final magnitude which is equal to one. In such
jected to a load of a footing in the presence of the a case, the footings distance is large enough to be
adjacent footing, ks(int), to the SRM of the rock mass considered as a single isolated footing without any
subjected to a load of a single isolated footing, ks(iso): influence from each other. Therefore, as a general
result, for S< Scr, the α coefficient increases by
ks ðintÞ
α¼ (19) increasing S, while for S> Scr, a reduction in α occurs
ks ðisoÞ
by increasing S. It is interesting to mention that the
It should be noted that ks(iso) is equal to the ks(mass) that bearing capacity of multiple footings on rock masses
was obtained numerically in the past section. In order to also conforms such a trend (Javid et al. 2015,
investigate the effect of various rock mass properties on Shamloo and Imani 2021) since both the SRM and
the α coefficient, several numerical models were carried the bearing capacity have a direct positive correlation
out, and the results were presented in the following with each other and both of them have been
subsections. obtained using the same tool, i.e., the pressure-
settlement curve. Figures 14 and 15 show that
increasing σci results in increasing α. Also, the Scr
7.1. Effect of σci increases slightly by increasing the σci. As an exam­
ple, by increasing the σci from 5MPa to 100MPa, the
Assuming mi = 10 and D = 0, Figures 14 and 15
Scr/B ratio changes from about 3 to 4. A maximum
show the variation of α versus S/B for different
increase approximately equal to 60% can be seen in α
values of σci considering GSI = 10 and 50, respec­
when the spacing between the two footings increases
tively. In S/B = 0, the two footings are in touch, and
from 0 to Scr. If the S magnitude exceeds the Scr and
they behave as a single footing with a width equal to
reaches about 8B~10B, the two footings’ interference
2B. In this case, the size of the stress bulb beneath
effect will be disappeared and the α value
the multiple footings increases with respect to
becomes one.
a single footing, which results in increasing the set­
tlement and decreasing the corresponding SRM.
Thus, the α coefficient becomes smaller than one.
7.2. Effect of GSI
By increasing the distance between the two footings,
the confinement of the passive zones beneath the two Figure 16 shows the variation of α versus S/B for differ­
footings increases, which results in increasing the α ent values of GSI, considering σci = 25MPa, mi = 10, and
coefficient. This increment will continue until reach­ D = 0. The overall trends of variation of the curves are
ing a maximum magnitude. The distance between similar to the previous subsection. As can be seen,

Figure 14. Effect of σci on α assuming GSI = 10, mi = 10 and D = 0.

 GEOMECHANICS AND GEOENGINEERING 15

Figure 15. Effect of σci on α assuming GSI = 50, mi = 10 and D = 0.

Figure 16. Effect of GSI on α assuming σci = 25 MPa, mi = 10 and D = 0.

increasing the GSI results in decreasing α, but the cri­ the mi results in increasing α. For the rock masses with
tical spacing (Scr) is always approximately equal to 4B, small values of mi, the interference of the two footings
which was not considerably affected by the GSI value. will be disappeared at S/B > 8, while for larger magni­
The effect of interference of the footings on the α is tudes of mi, this interference effect will be diminished at
paramount for the rock masses with small values of GSI. about S/B > 10.
The interference effect of the two footings will disappear
at S/B > 8 for the rock masses with large values of GSI,
while for low GSI magnitudes, this interference effect 7.4. Effect of D
will be diminished at S/B > 10.
Figure 20 shows the variation of α versus S/B for different
values of D, considering σci = 25 MPa, GSI = 50, and mi
= 10. As can be seen, increasing the D results in decreasing
7.3. Effect of mi
α and Scr. For the rock masses with D = 0, the interference
Assuming σci = 25 MPa, and D = 0, Figures 17–19 show of the two footings will be disappeared at S/B > 9, while for
the variation of α versus S/B for different values of mi D = 0.8, this interference effect will be diminished at about
considering GSI = 10, 50, and 90, respectively. It was S/B > 6. This conclusion means that by increasing D, the
found that for S> Scr, α will reduce; the higher the mi, the two adjacent footings’ interference effect will be disap­
larger the rate of this reduction. Moreover, increasing peared in a smaller spacing between them.
16 S. SHAMLOO AND M. IMANI

Figure 17. Effect of mi on α assuming σci = 25 MPa, GSI = 10 and D = 0.

Figure 18. Effect of mi on α assuming σci = 25 MPa, GSI = 50 and D = 0.

Figure 19. Effect of mi on α assuming σci = 25 MPa, GSI = 90 and D = 0.

 GEOMECHANICS AND GEOENGINEERING 17

Figure 20. Effect of D on α assuming σci = 25 MPa, GSI = 50 and mi = 10.

8. Conclusion this coefficient would be less than one. By increasing

S from zero to Scr, the α coefficient increases. For
The following results were obtained from the numerical
S> Scr, the α coefficient decreases until reaching
analyses performed in this research:
a constant magnitude equal to one.
● The presence of two adjacent footings plays an
● The SRM values obtained from the numerical ana­
essential role in the amount of SRM. Depending
lyses performed in this research are in good accor­
on the distance between the two footings, the α
dance with the laboratory tests performed by Lee
coefficient is between 0.7 and 1.3. In multiple foot­
and Jeong (2016), with a range of differences
ings, the minimum value of the SRM occurs at S/
between 1% to 16%.
B = 0, while its maximum magnitude occurs at
● Among the Hoek-Brown parameters, GSI has the
2 ≤ S/B ≤ 4. Likewise, depending on the strength
highest, and mi has the lowest influence on the
properties of the rock mass, if the distance between
SRM of rock masses. The effects of the rock mass
the two footings becomes larger than 8 to 10 times
Hoek-Brown parameters on the SRM are more
the footing width, the interference of the footings
paramount in the case of significant Hoek-Brown
does not have any effect on the SRM of the rock
parameters than in the case of low magnitudes.
mass, and the multiple footings behave as two
Moreover, increasing σci, GSI, and mi result in
isolated ones.
increasing the SRM, while increasing D results in
● The width of the footing has an adverse correlation
decreasing the SRM.
with the SRM of rock masses. The obtained results
● A new coefficient named the subgrade reduction
show that by increasing footing width from 0.5 to 3
factor, Jmass, was introduced in this research,
metres, the SRM of the rock mass decreases in the
which was defined as the ratio of the SRM of
range of 13% to 67%, depending on the rock mass
the Hoek-Brown rock mass to the SRM of the
properties.
intact rock. Depending on the rock mass prop­
erties, this coefficient may vary from 0.2% to
54%. Notations
● To take into account the effect of two adjacent foot­
ings on the SRM of rock masses, the interference The following symbols were used in this paper.
coefficient, α, was introduced as the ratio of the SRM SRM: Subgrade reaction modulus;
in the case of two nearby footings to that of one ks: SRM of the intact rock;
isolated footing. The magnitude of this coefficient kj: SRM of the jointed rock;
depends on the rock mass properties and the dis­ ks(mass): SRM of the rock mass obeying the Hoek-
tance between the two footings. By increasing σci Brown failure criterion;
Jf: Joint reduction factor;
and mi, the α coefficient will increase, while increas­
Jmass: Subgrade reduction factor;
ing GSI results in decreasing α. Moreover, if the Scr: Critical spacing;
distance between the two footings becomes zero, S: Distance between two footings;
18 S. SHAMLOO AND M. IMANI

α: Interference coefficient; Imanzadeh, S., Denis, A., and Marache, A., 2013. Effect of
ks(int): SRM of the rock mass subjected to the load of uncertainty in soil and structure parameters for buried
a footing in the presence of the adjacent foot­ pipes. Geotech and Geophys, Site Characterization, 4,
ing (obtained from the pressure-settlement 1847–1853.
curve); Javid, A.H., Fahimifar, A., and Imani, M., 2015. Numerical
ks(iso): SRM of the rock mass subjected to the load investigation on the bearing capacity of two interfering
of a single isolated footing (obtained from strip footings resting on a rock mass. Computers and
the pressure-settlement curve); Geotechnics, 69, 514–528. doi:10.1016/j.
S d: Spacing of the joints; compgeo.2015.06.005
Id: Inclination of the joints; Klopple, K. and Glock, D., 1979. Theoretische und experi­
ϕj: Friction angle of the joints; mentelle untersuchungen zu den traglastproblemen beige­
cj: Cohesion of the joints; wiecher, in die erde eingebetteter rohre. Veroffentlichung
K n: Normal stiffness of the joints; des Instituts Statik und Stahlbau der Technischen
Ks: Shear stiffness of the joints; Hochschule Darmstadt, H–10.
γ: Unit weight of the rock mass Kullhawy, F.H. and Carter, J.P., 1992. Settlement and bearing
capacity of foundations on rock masses. In: F.G. Bell, ed.
Engineering in rock masses. Oxford, UK: Butterworth-
Heinemann, 231–245.
Disclosure statement Lee, J. and Jeong, S., 2016. Experimental study of estimating
No potential conflict of interest was reported by the author(s). the subgrade reaction modulus on jointed rock
foundations. Rock Mechanics and Rock Engineering, 49,
2055–2064. doi:10.1007/s00603-015-0905-9
Lees, A., 2016. Geotechnical finite element analysis. 1st ed.
References London: ICE Publishing.
Mabrouki, A., Benmeddour, D., and Mellas, M., 2010.
Alhossein, H., Carter, J.P., and Booker, J.R., 1992. Finite ele­
Numerical study of the bearing capacity for two interfering
ment analysis of rigid footings on jointed rock. In:
strip footing on sands. Computers and Geotechnics, 37,
Proceedings of the 3rd International Conference on
431–439. doi:10.1016/j.compgeo.2009.12.007
Computational Plasticity, Vol. 1. Barceona, Spain, 935–945.
Mansouri, M., Imani, M., and Fahimifar, A., 2019. Ultimate
AlKhafaji, H., Imani, M., and Fahimifar, A., 2020. Ultimate
bearing capacity of rock masses under square and rectan­
bearing capacity of rock mass foundations subjected to
gular footings. Computers and Geotechnics, 111, 1–9.
seepage forces using modified Hoek-Brown criterion.
doi:10.1016/j.compgeo.2019.03.002
Rock Mechanics and Rock Engineering, 53, 251–268.
Meyerhof, G.G. and Baikie, L.D., 1963. Strength of steel sheets
doi:10.1007/s00603-019-01905-6
bearing against compacted sand backfill. Highway Research
Avci, B. and Gurbuz, A., 2018. Modulus of subgrade reaction
Record, 30, 1–19.
that varies with magnitude of displacement of cohesionless
Saada, Z., Maghous, S., and Garnier, D., 2008. Bearing capa­
soil. Arabian Journal of Geosciences, 351 (11), 1–8.
city of shallow foundations on rocks obeying a modified
doi:10.1007/s12517-018-3713-1.
Hoek–Brown failure criterion. Computers and Geotechnics,
Biot, M.A., 1937. Bending of infinite beams on an elastic
38, 144–154. doi:10.1016/j.compgeo.2007.06.003
foundation. Applied Mechanics American Society of
Sadrekarimi, J. and Akbarzad, M., 2009. Comparative study of
Mechanical Engineers, 59, A1–A7.
methods of determination of coefficient of subgrade
Carter, J.P. and Kullhawy, F.H., 1988. Analysis and design of
reaction. Electronic Journal of Geotechnical Engineering,
drilled shaft foundations socketed into rock. In: Rep. EL-5918.
14, 1–14.
Palo Alto, California: Electric Power Research Institute.
Sargazi, O. and Hosseininia, E., 2017. bearing capacity of ring
Fattah, M.Y., Shlash, K.T., and Mohammad, H.A., 2014a.
footing on cohesionless soil under eccentric load.
Bearing capacity of rectangular footing on sandy soil
Computers and Geotechnics, 92, 169–178. doi:10.1016/j.
bounded by a wall. Arabian Journal for Science and
compgeo.2017.08.003
Engineering, 39, 7621–7633. doi:10.1007/s13369-014-1353-7
Selvadurai, A.P.S., 1985. Soil–pipeline interaction during
Fattah, M.Y., Shlash, K.T., and Mohammad, H.A., 2014b.
ground movement. In: F.L. Bennett and J.L. Machemehl,
Experimental study on the behavior of strip footing on sandy
eds.. Arctic, Civil engineering in the Arctic offshore. ASCE
soil bounded by a wall. Arabian Journal of Geosciences, 8 (7),
speciality conference. San Francisco, 763–773.
4779–4790. doi:10.1007/s12517-014-1564-y.
Shamloo, S. and Imani, M., 2021. Upper bound solution for
Fattah, M.Y., Shlash, K.T., and Mohammad, H.A., 2014c.
the bearing capacity of two adjacent footings on rock
Experimental study on the behavior of bounded square footing
masses. Computers and Geotechnics, 129, 1–14.
on sandy soil. Engineering and Technology Journal, University
doi:10.1016/j.compgeo.2020.103855
of Technology – Iraq, 32 (Part (A),5), 1083–1105.
Hoek, E. and Brown, E.T., 1980. Empirical strength criterion Terzaghi, K.V., 1955. Evaluation of coefficient of subgrade
for rock masses. ASCE Journal of Geotechnical Engineering, reaction. Geotechnique, 5 (4), 297–326. doi:10.1680/
106, 1013–1035. geot.1955.5.4.297.
Hoek, E. and Dierichs, M.S., 2005. Empirical estimation of Vásárhelyi, B., 2009. A possible method for estimating the
rock mass modulus. International Journal of Rock Poisson’s rate values of the rock masses. Acta Geodaetica
Mechanics and Mining Sciences, 43, 203–215. doi:10.1016/ Geophys Hungarica, 44 (3), 313–322. doi:10.1556/
j.ijrmms.2005.06.005 AGeod.44.2009.3.4.
 GEOMECHANICS AND GEOENGINEERING 19

Vesic, A.B., 1961. Beams on elastic subgrade and Winkler’s Yang, X. and Yin, J., 2005. Upper bound solution for ultimate
hypothesis. In: Proceedings of the Fifth International bearing capacity with a modified Hoek–Brown failure cri­
Conference on Soil Mechanics and Foundation Engineering. terion. International Journal of Rock Mechanics and Mining
Paris, 845–850. Sciences, 42, 550–560. doi:10.1016/j.ijrmms.2005.03.002
Vlassov, V.Z. and Leontiev, N.N., 1966. Beams, plates, and Ziaie Moayed, R. and Naeini, S.A., 2006. Evaluation of mod­
shells on elastic foundations. Translated from Russian. ulus of subgrade reaction in gravely soils based on standard
Jerusalem: Israel Program for Scientific Translations. penetration test (SPT). Chapter 115. In: Proceedings of the
Wyllie, D.C., 1999. Foundations on rock. 2nd ed. London, UK: sixth international conference on physical modelling in geo­
E & FN Spon. technics, Hong Kong.

View publication stats