Assessment of Soil Nailing Performance

by Using Finite Element and Finite

Difference Methods
Ahmad Safuan A Rashid
Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310 UTM,
Skudai, Johor, Malaysia
e-mail: [email protected]

Koohyar Faizi
Researcher, Department of Geotechnics and Transportation, Universiti
Teknologi Malaysia,
81310 UTM Johor Bahru, Johor, Malaysia.
e-mail: [email protected]

Roohollah Kalatehjari
Post-Doctoral Fellow, Department of Geotechnics and Transportation,
Universiti Teknologi Malaysia,
81310 UTM Johor Bahru, Johor, Malaysia.
e-mail: [email protected]

Ramli Nazir
Associate Professor, Department of Geotechnics and Transportation,
Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia.
e-mail: [email protected]

ABSTRACT
The field performance of a geotechnical structures including its failure mechanism should be
evaluated in detail before construction in order to minimize the possibility of potential failure.
In this paper, the application of soil nailing system is documented in a detailed based on case
study in the north of Tehran, Iran. The total excavation depth for a 15 storey residual building
wall brought the cut face close to 12 m. The diameter of nails was 28 mm and their lengths
were varied between 8 to 10 m with inclination of 7 degree to the horizontal plane. A series of
numerical modelling was conducted to evaluate the deformation during the excavation
process using finite difference method (FDM - FLAC 2D) and finite element method (FEM -
PLAXIS 2D). In the modelling process, the pattern of wall deflection was simulated with
shotcrete facing. A good agreement has been observed between the results of both methods
when the pattern of deformation in the first and the last steps of excavation were simulated.
Based on the maximum vertical displacement of shotcrete facing, calculated by two methods,
the results obtained by FEM were closer to the field measurement. However, a good
agreement has been observed between the values of factor of safety which were calculated by
FLAC 2D and PLAXIS 2D, for different stages of excavation.
KEYWORDS: Soil, Nailing, Excavation, FEM, FDM, FLAC 2D, PLAXIS 2D.

- 5881 -
 Vol. 18 [2013], Bund. Y 5882

INTRODUCTION
Soil nailing has become a widely accepted method of providing temporary and permanent
earth support, underpinning, and slope stabilization on many civil projects in the world since
early of 1970s. In the early years, soil nailing was typically performed only in projects that
expertise geotechnical contractors offered as an alternative to other conventional systems. More
recently, soil nailing has been specified as the system of choice due to its overall acceptance and
effectiveness. Soil nailing consists of the passive reinforcement (i.e. no post-tensioning) of
existing ground by installing closely spaced steel bars (i.e. nails), which are subsequently encased
in grout. As construction proceeds from the top to bottom, shotcrete or concrete is also applied on
the excavation face to provide continuity (Lazarte, et al., 2003). Soil nailing is typically used to
stabilize existing slopes or excavations where top-to-bottom construction is advantageous
compared to other retaining wall systems. For certain conditions, soil nailing offers a viable
alternative from the viewpoint of technical feasibility, construction costs, and construction
duration when compared to ground anchor walls, which is another popular top-to bottom
retaining system (Zamiran and Saba, 2012). Although the theoretical engineering aspects of soil
nailing may be well understood, there is a far lesser degree of understanding, even within the
geotechnical community, due to the site conditions and construction problems (Thomas, 2013).
Among researchers who have studied in this field are Liu and Dugan (1972), Hanna and
Kurdi (1974), Anderson and Hanna (1977), Matlock, et al. (1981), Lim and Briaud (1997), Briaud
& Lim (1997), Dawkins (2001), Krabbenhoft, et al. (2008) and Tan and Paikowsky (2008). One
important issue in the design of excavation wall stabilization is to decrease the lateral
displacement of the constructed wall. This is because high lateral displacement indicates lack of
stability and is followed by collapse of soil into the wall which may damages to the surrounding
area. In that sense, the wall displacements should be carefully controlled to prevent any damage
(Bara and Qing, 2010). The purpose of this paper is to investigate the soil deformation during soil
nailing installation at site in the north of Tehran, Iran. Soil nail properties, procedures and
monitoring for case study are presented as a tool to aid in making good decisions and prediction
of failure zones. The application of a deep soil nailing system for a project of residual building is
presented in this study with the total excavation height reaches 12 m. The deformation of
excavation face was measured by surveyor at the last step of excavation and after nailing and
shotcrete. Finite element difference software (FLAC2D) and finite element software
(PLAXIS2D) have been employed to model the excavation and soil nailing system and predict
the horizontal displacements. Both of the software can determine the horizontal and vertical
displacements as well as the behaviour of retaining structure based on the plane strain theory
(Gareh, 2011; Gareh and Saidi, 2011). For the sake of accuracy, the results of numerical
modelling have been compared with the monitoring results and actual behaviour of the wall.

SITE INFORMATION
The city of Tehran is located at the foothill of the southern slopes of the Alborz Mountain. It
sits on an alluvial plain formed over time by flood erosion of the mountains. As a result of this
process, large and small particles have settled respectively on high and low elevations, resulting
in varying geological conditions. The source rocks, slope of the mountains, and climate
conditions are important factors in determining the properties of the soils deposited at the foot of
mountains. The Alborz Mountains Range is steep and mainly consists of tuff, limestone and
dolomite (Fakher, et al., (2007); Uromeihy and Nassiri, 2006). The area experiences heavy rains
in some seasons and is seismically active. Therefore, non-uniform soil layers have been formed.
Figure 1 presents the site location.
 Vol. 18 [2013], Bund. Y 5883

Figure 1: Site location map pointed as red balloon A (Google map, 2013)

The system of nailing was designed for the construction of a residual building covering an
area of approximately 7000 m2 with 15 stories. Figure 2 shows overview of activities in the site.
The execution of soil nailing consists of making a 12 m vertical cut extending with a horizontal
length to be stabilized and shotcreted on the same day. It is important to note that the shotcrete
should be applied prior to installing the nails, if there is a concern over the stand-up time and
possibility of sloughing of the soil. Besides, shotcrete facing before drilling can reduce hole drill
disturbance, which is appeared in the entrance of hole drill during soil drilling. Due to the high
depth of excavation and collapsible soil in the studied project, shotcrete was applied before
nailing. In order to reinforce the concrete facing, reinforcing material as welded wire mesh was
placed along the face of the excavation prior to apply the shotcrete facing. Similar to most
temporary shotcrete walls, this process was accomplished by applying a 180 mm thick layer of
concrete. The properties of soil in excavation area are provided in Table 1. These properties were
collected from the results of a series of laboratory tests (Atterberg, limits, compaction test, sieve
analysis, hydrometer, direct shear test, and free vertical swell) on the soil samples from the
excavation area (KACE, 2012). These properties were used to analyse the project in both
PLAXIS 2D and FLAC 2D software.
 Vol. 18 [2013], Bund. Y 5884

Figure 2: Overview of activities in the site including (a) preparation and shotcreting, (b)
soil drilling (c) hole drill (d) nailing

Table 1: Properties of soil in excavation area (KACE, 2012)

Unit Elasticity
Poisson Cohesion Friction
Soil Type of Depth weight Modulus
ratio (C) Angle(Φ)
specification soil (m) (ɣ ) (E)
(ν) (kg/cm2) (degree)
(kg/cm3) (kg/cm )2

GW-GM
1.85-
Values* SW-SM 0-12 0.35 50-250 0.15-0.21 20-32
1.91
SC, CL

*Water level= 9m
 Vol. 18 [2013], Bund. Y 5885

FINITE DIFFERENCE METHOD (FDM) AND FINITE

ELEMENT METHOD (FEM)
For modelling the steps of soil nailing in the studied excavation, PLAXIS 2D and FLAC 2D
software were used based on FEM and FDM methods respectively. FDM discretization is based
on the differential form of the Partial Difference Equation to be solved. It utilizes a point-wise
approximation to a solution. The domain is discretized into a grid of rectangular cells or nodes.
The solution will be obtained at each nodal point. Although FDM is easy to implement and the
computing time for each step is fast, the number of steps required for convergence is high. The
other disadvantage is that the domain is not accurately represented if it is discontinuous or non-
rectangular in shape (Fausett, 2003). Figure 3 (a) illustrates a sample FDM mesh. FEM
discretization is based on a piecewise representation of the solution in terms of specified basis
functions. In FEM the discretization is not restricted to a grid of hexahedral cells or nodes, instead
a solution is approximated using interconnecting sub-regions or elements. These elements are
typically simple geometrical shapes as illustrated in Figure 3(b). This flexibility in construction
of elements allows FEM to accurately model complex geometries. The downside is that FEM is
difficult to implement, however there is no agreement on this between researchers (Kalla, 2010;
Zienkiewicz and Cheung, 1965).

Figure 3: Discretization methods in (a) FDM (b) FEM

FINITE ELEMENT MODELLING (FEM) PROCEDURES

Numerical modelling allows designers to study the soil behaviour in various conditions without
resorting to simplified assumptions. Two dimensional finite elements modelling of the different
steps of soil nailing is performed using PLAXIS 2D software. A parametric study was performed
to arrive at the critical parameters that define the behaviour of the system. The parametric study
comprised of all the elements that influence the behaviour of the system. Moreover, various
aspects of soil excavating like steps of excavation and vertical displacement were studied. The
geometry was drawn using geometric lines and standard fixities were then used to define the
boundary conditions. The model was created, properties of different soil materials were assigned
to material model, and finite mesh was generated. The water conditions can be specified in the
geometry configuration mode by generating pore pressure in phreatic level. A constant ground
water level has been considered in this study. Figure 4 illustrates the analysed section of project
 Vol. 18 [2013], Bund. Y 5886

and the soil nail configuration used in this section.

Figure 4: illustrates the analyzed section of project and the soil nail configuration used in
this section.

The following is the detail of modelling process including the choice of material models,
finite element mesh, and boundary condition to simulate field conditions and obtain settlements
of the excavation. There are many modelling methods developed to simulate 3-D soil nailing
problem using 2-D FEM. Each method poses advantages and limitations in approximating the
true behaviour of soil nails (Olia and Liu, 2011). The method was proposed by Al-Hussaini and
Johnson (1978) where the discrete reinforcement was smeared into continues plate across the
spacing. This is achieved by factoring the Young’s modules of the plate (E) using area ratio
factors such as the axial stiffness (EA). Thus, this method is adopted in this study.
Two types of triangular elements are used in the PLAXIS as 6-node and 15-node triangular
elements. Advantages of higher order triangular elements is that they better represent the
description of continuous strain and stress variations and also provide good description of a
continuous displacement field compare with relatively few elements. The disadvantages of higher
order elements is that the failure loads may be dependent on the mesh and makes poor description
of discontinuous stress and strain. In PLAXIS, the program automatically creates unstructured
mesh as there is no possibility of making a so-called structured mesh (Gómez, 2011). The mesh is
generated based on random seeds and its size cannot be set explicitly. The size of mesh may be
changed globally by means of global coarseness or locally by means of local coarseness. Figure
5 shows a typical mesh generated for the present study. Where shotcrete wall of 18 cm thickness
with reinforced wire mesh is modelled as beam elements with axial stiffness (EA) = 5.4 x 106
 Vol. 18 [2013], Bund. Y 5887

kN/m, bending stiffness (EI) = 1.458 x 104 kNm2/m, and Poisson’s ratio (ν) = 0.2; and soil nails
are modelled as geotextile elements with EA = 6.87 x 104 kN/m and ν = 0.

Figure 5: A typical mesh generated for the present study by PLAXIS 2

FINITE DIFFERENCE MODELLING (FDM) PROCEDURES

FDM is possibly the oldest numerical technique used for the solution of sets of differential
equations, given initial values and/or boundary values (Desai and Christian, 1977). For FDM
analysis, the computer code FLAC 2D (i.e. Fast Lagrangian Analysis of Continua) was selected
because of its flexibility and wide acceptance (FLAC 4.0, 2000). It easily includes key factors and
phenomena that affect the behaviour of ground conditions. The FDM code is an explicit two
dimensional finite difference program that performs a Lagrangian analysis. Here, explicit means
it uses a time stepping procedure to solve the problem without forming the stiffness matrix. The
Lagrangian formulation enables the grid to move and deform with the material it represents, since
the incremental displacements are added to the coordinates (Fakher, et al., 2007). FLAC 2D is a
powerful tool for assessing the effect of different material properties. It is also a robust tool
because it can handle any constitutive model with no adjustment to the solution algorithm.

In FLAC 2D, grid generation is limited to simple and regular shaped regions. There is a Fish
function which can be used to generate user defined grids with varying zones. The advantage of
using FISH function is that it can easily adjust grid boundary and zone density using “SET”
command. The plane strain model was used in this study. The size of the model was set to 25 m
in width and 15 m in height. Here the grid is composed of several zones consist of constant
 Vol. 18 [2013], Bund. Y 5888

height. A reasonably medium grid should be selected to ensure that the displacement contours
will be well-defined as it develops. Figure 6 shows a typical mesh generated for the present study
by FLAC 2D.

Figure 6: A typical mesh generated for the present study by FLAC 2D

The Mohr-Coulomb plasticity is applicable for most general engineering studies (Olia and
Liu, 2011). This model is used for materials that yield when subjected to shear loading. In this
model, the yield stress depends on the major and minor principal stresses and the intermediate
principal stress has no effect on it. Also, Mohr-Coulomb parameters (cohesion and friction angle)
are usually more readily available and more easily reachable than other properties for geo-
engineering materials. Moreover, Mohr-Coulomb models are the most computationally efficient
plasticity models (Gareh and Saidi, 2011).

RESULTS AND DISCUSSIONS

Since FLAC 2D models a nonlinear system as it evolves in time, the interpretation of its
results may be more difficult than a conventional finite-element program which produces “a
solution” at the end of its calculation phase. The calculation is continued using step command and
update interval of plot must be used to estimate the effects of time. The SOLVE command is used
to find the equilibrium state (Desai and Christian, 1997).
Failure zone contours at the first step of the numerical simulations using FDM and FEM are
presented in Figs. 7. This step represents the excavation with depth of 2 m without shotcreting or
nailing. In FLAC 2D, soil is modelled as a viscoelastic material (Mohr-Columb). The shotcrete
cover is modelled as elastic beam elements and nails are modelled as cable element.
 Vol. 18 [2013], Bund. Y 5889

(a) (b)

Figure 7: Failure zone in the first step of excavation modelled by (a) FLAC 2D and (b)
PLAXIS 2D
Displacement vectors and Horizontal displacement contours were obtained from numerical
simulations by using FDM and FEM in the last step of excavation with depth of 12 m. This step
of excavation was reinforced by shotcreting and nailing. The results are presented in Figs. 8 and
9.

(a) (b)

Figure 8: Displacement vectors in the last step of excavation modelled by (a) FLAC 2D
and (b) PLAXIS 2D
 Vol. 18 [2013], Bund. Y 5890

(a) (b)

Figure 9: Maximum displacement in the last step of excavation modelled by (a) FLAC
2D and (b) PLAXIS 2D

Shear strain zones for both models are shown in Figure 10. Based on these figures, a similar
pattern in observed in shear strain zones obtained by FEM and FDM. The shear strains within the
reinforced soil mass were developed along the potential slip surface of the active zone as a result
of soil nailing. This is fairly close to the formation of an active wedge in the retaining wall
design.

(a) (b)
Figure 10: Shear Strain zones in the last step of excavation modelled by (a) FLAC 2D
and (b) PLAXIS 2D
 Vol. 18 [2013], Bund. Y 5891

The deformation of excavation face was measured by surveyor as 14 mm. Figure 11 shows the
maximum vertical displacement of shotcrete facing prepared from field measured and the results
of FLAC 2D and PLAXIS 2D.

Figure 11: Horizontal displacement calculated by FLAC 2D and PLAXIS 2D for

different steps of excavation

A comparison between the values of factor of safety (FOS) obtained by FLAC 2D and
PLAXIS 2D in different steps of excavation is shown in Figure 12. The trends show that the
overall value of FOS decreased with increasing of excavation depth. However, the value of FOS
for the first step (excavation up to 2 m) is less than the second step (excavation up to 4 m). The
execution of soil nailing consisted of making an initial 2 m vertical cut supported by shotcrete
without any horizontal stabilizing (e.g. nail). Then, nailing system was applied in the second step
(excavation up to 4 m), which played a key role to affectedly recover the value of FOS. From this
point onwards, the FOS continuously decreased with increase of the depth of excavation. This
reduction in the FOS was due to decrease of lateral earth pressure by increasing the depth of
excavation (Wu, et al., 2000). The FOS predicted by numerical modelling in FEM and FDM
between excavations depths of 4 m to 10 m was almost the same. However, the difference FOS
predicted between two different methods for excavation depth of 12 m was notable. This
difference might be related to different methods applied to consider the effect of water table in
depth of 9 m. In designing of deep excavations, a major problem is often dominated by water
flow around the walls. Water flow, influences the global stability of the wall and the stability of
the excavation bottom where bulk heaving or boiling may occur (Fonte, 2010). Water flow
patterns obtained by FLAC 2D and PLAXIS 2D are shown in Figure 13.
 Vol. 18 [2013], Bund. Y 5892

Figure 12: FOS calculated by FLAC 2D and PLAXIS 2D for different steps of
excavation

(a) (b)

Figure 13: Water flow vectors (a) FLAC 2D and (b) PLAXIS 2D

CONCLUSIONS
In this paper, a real case of soil nailing in an excavation project was studied. Six steps of
excavation were studied through numerical modelling by FLAC 2D (FDM) and PLAXIS 2D
(FEM) software to better realize the soil nailing displacements occurred during excavation.
Numerical modelling were used to determine the maximum displacement, prediction of shear
strain zone, and calculate FOS for the 6 steps of excavation. Mohr-coulomb model were used in
modelling the excavation material by both software. According to the experimental measurement
and numerical analyses the following conclusions are drawn:
1) It was found that finite difference software (FLAC 2D) and finite element software (PLAXIS
2D) are powerful tools for investigating the behaviour of a wall stabilized by soil anchorage
and steel pile.
 Vol. 18 [2013], Bund. Y 5893

2) FOS decreased with increase in the excavation depth. The decreasing rate compounds with
the progress of the excavation operation. However, using soil nailing system helps to improve
factor of safety rather than unreinforced excavating.
3) A good agreement has been observed between the FOS predicted by numerical modelling in
FEM and FDM for excavation depths of 6 m to 12 m. However, in the first and second step of
excavation respectively with depth of 2 m and 4 meter, the FOS obtained by FDM was higher
than the obtained FOS by FEM. These results might be related to different method of initial
condition which was allocated by two different software.
4) Finite difference analysis needed appropriate constitutive model for a particular soil type and
took considerable time for a complete analysis. In contrast, finite element analysis was more
powerful and gave better results that match with interpretations of maximum displacement in
X direction using the measured data.

ACKNOWLEDGMENT
The work was financially supported by Universiti Teknologi Malaysia under Research
University Grant (Q.J130000.2622.06J95) and the Ministry of Education of Malaysia, MOE. The
authors also deeply appreciate the significant contribution of Khesht Azma Consulting
Engineering Company and its technical supervision office for providing the required data to make
this research possible.

REFERENCES
1. Al-Hussaini, M. M., Johnson, L.D. (1978). Numerical analysis of a reinforced earth wall, ASCE,
Symposium on Earth Reinforcement, 98-126.
2. Anderson, W. F., Hanna, T. H., Shah, S.A. (1977). Model tests on anchored walls retaining
overconsolidated sands, Canadian Geotechnical Journal , 14(2): 214-222, 10.1139/t77-022.
3. Barar, S. E., Qing, L. (2010). Time-History finite element dynamic analysis-soil nail walls a manual
casino-highland, casino. San Diego, California.
4. Briaud, J., Lim, Y. (1997). Soil-Nailed Wall Under Piled Bridge Abutment: Simulation and
Guidelines. J. Geotech. Geoenviron. Eng., 123(11), 1043–1050.
5. Briaud, J., Lim, Y. (1999). Tieback Walls in Sand: Numerical Simulation and Design Implications,
Journal of Geotechnical and Geoenvironmental Engineering, Vol. 125, No. 2, pp. 101-111, February,
ASCE, Reston, Virginia.
6. Dawkins, P. W. (2001). Investigation of Wall Friction, Surcharge Loads, and Moment Reduction
Curves for Anchored Sheet-Pile Walls, U.S. Army Corps of Engineers Information Technology
Laboratory ERDC/ITL TR-01-4.
7. Desai, C. S., Christian, J. T. (1977). Numerical methods in geotechnical engineering. Edited by.
McGraw-Hill Book Company, No. of Pages: 783.
8. Fakher, A., Cheshomi, A., Khamechiyan, M. (2007). The addition of geotechnical properties to a
geological classification of coarse-grained alluvium in a pediment zone, Quarterly Journal of
Engineering Geology and Hydrogeology.
9. Fausett, V. L. (2003). Numerical Methods Algorithms and Applications. Prentice Hall.
10. FLAC 4.0, Constitutive Model Manual, 2000.
11. Fonte, J. B. (2010). Numerical Modeling of Excavations below the Water Table, Master thesis.
Universidade Do Portu.
 Vol. 18 [2013], Bund. Y 5894

12. Ghareh, S. (2011). Numerical Analysis of Behavior of Pile System under Lateral Loading, 6th National
Congress on Civil Engineering.
13. Ghareh, S., Saidi, M. (2011). An Investigation on the Behavior of Retaining Structure of Excavation
Wall Using Obtained Result from Numerical Modeling and Monitoring Approach. (A Case Study of
International "Narges Razavi 2 Hotel", Mashhad), Journal of Structural Engineering and Geotechnics.
14. Gómez, G.M. (2011). Pilot and market replication projects. Agreement Number -
ECO/08/239011/SI2.535202. CIP Eco-innovation.
15. Google Map. (2013). https://maps.google.com
16. Hanna, T.H., Kurdi, I. I. (1974). Studies on anchored flexible retaining walls in sand, Journal of
Geotechnical Engineering, American Society of Civil Engineers, vol. 100, pp 1107-1122. (Discussion,
vol. 101, pp101 829-831; vol. 102, pp 111-112).
17. Kalla,E. (2010). Modeling Studies to Assess Long Term Settlement of Light Weight Aggregate
Embankment, Master thesis, The University of Texas at Arlington.
18. Khesht Azma Consulting Engineering. Co (KACE). (2012). Reports Book.8/413/2.
19. Krabbenhoft, S., Andersen, A., Damkilde, L. (2008). The tensile Capacity of bored piles in frictional
soils. Can. Geotech. J. 45:1715-1722.
20. Lazarte, C. A., Elias, V., Espinoza, D., Sabatini, P. (2003). Technical Report Documentation Page.
FHWA .Contracting Officer Technical Representative (COTR): Chien-Tan Chang (HTA-22).
21. Lie, T.K., and Dugan, J.P. (1972). An instrumented tied back deep excavation, in proceeding of ASCE,
Specialty conference on performance of earth and earth supported structures, Purdue University.
Laffayete IN.ASCE, New York, Vol.1,Pt.2, pp.1323-1339.
22. Matlock, H., Bogard, D., Lam, I. (1981). BMCOL76: A Computer Program for the Analysis of Beam-
columns under Static Axial and Lateral Loading. Program developed at the University of Texas at
Austin, under grant from Fugro, Inc., and documented at Ertec, Inc., Long Beach, CA.
23. Olia, A., Liu, J. (2011). Numerical investigation of soil nail wall during construction, Geotechnical
Conference Pan-Am CGS.
24. Tan, Y., Paikowsky, S. (2008). Performance of Sheet Pile Wall in Peat. J. Geotech. Geoenviron. Eng.,
134(4), 445–458.
25. Thomas, J. (2013). Nailing: Where, When And Why A Practical Guide. Presented at the 20th Central
Pennsylvania Geotechnical Conference Hershey, PA.
26. Uromeihy, A., Nassiri, S. (2006). Effect of landslides on the development of the districts of north
Tehran City, IAEG. Paper number 37.
27. Wu, X., Wang, M., Wang, M. (2000). Earth-Filled Wide-Base Hollow Piers for Excavation Support.
Soil-Cement and Other Construction Practices in Geotechnical Engineering: pp. 77-84.
28. Zamiran, S., Saba, H. (2012). Numerical Investigation of Seismic Behavior of Soil Nail Walls. 4th
International Conference on Seismic Retrofitting, Tabriz, Iran, 2-4 May.
29. Zienkiewicz, O. C., Cheung Y.K. (1965). Finite Elements in the solution of Field Problems, The
Engineer, September 24th.

© 2013, EJGE