New York Science Journal 2018;11(7) http://www.sciencepub.

net/newyork

Battered Piles under Lateral Loads using Strain Wedge Model and Current Practice

Mohamed Ashour1; Ahmed Alaa Eldin2; Mohamed G. Arab3

1
Associate Professor of Civil Engineering, Dept. of Mechanical and Civil Engineering, Alabama A & M University,
Normal, Alabama 35762.
2
Teaching Assistant at Mansoura High Institute for Engineering and Technology, Egypt.
3
Assistant Professor, Civil & Environmental Engineering Department, College of Engineering, University of
Sharjah, UAE; and Assistant Professor, Structural Engineering Department, Faculty of Engineering, Mansoura
University, Egypt
[email protected]; [email protected]; [email protected]

Abstract: The Strain Wedge (SW) model, LPILE and Finite Element program (MIDAS GTS-NX) are used to study
pile and soil typical parameters impact on the lateral response of single battered piles. The influence of pile battering
angle, sand relative density, and pile cross sectional shape are presented in addition to the prediction of the soil
wedge geometry infront of the pile. In SW model and LPILE analyses, the soil is modeled as a Beam on Elastic
Foundation (BEF) with a set of non-linear p-y curves (i.e., modulus of subgrade reaction, Es) which accounts for soil
and pile properties. Mohr-Coulomb soil failure criteria is employed in MIDAS soil modeling with a Tetrahedron
meshing. The used approaches have been compared with field test results. Negative battered piles sustain greater
resistance compared to the piles with positive battered angles. The larger the sand relative density the more the
battered pile ability to withstand lateral loads. The three techniques are used to predict the pile lateral deflection,
bending moment, and shear force along the pile length. Unlike the other two techniques, MIDAS predicts less
bending moments and shear forces for positive battered piles, which is also highly influenced by the interface
element controlling parameter (i.e., the virtual thickness, tv).
[Mohamed Ashour; Ahmed Alaa Eldin; Mohamed G. Arab. Battered Piles under Lateral Loads using Strain
Wedge Model and Current Practice. N Y Sci J 2018;11(7):20-28]. ISSN 1554-0200 (print); ISSN 2375-723X
(online). http://www.sciencepub.net/newyork. 4. doi:10.7537/marsnys110718.04.

Keywords: Strain Wedge, Battered pile, Lateral load, Finite Element, Modeling.

Introduction loads has limited studies in the literature. The negative

Battered piles are used instead of vertical piles to battered piles have greater resistance than that of
resist high lateral loads as applied to bridge piers and vertical and positive battered ones in sandy soils as
off-shore and retaining structures. Battered piles are concluded from full-scale lateral load tests (Alizadeh
classified as negative and positive battered piles in and Davisson, 1970; Nimityongskul et al., 2012) and
accordance with the directions of lateral load and pile model scale tests (Juvekar and Pise, 2008; Murthy,
battering as demonstrated in Fig. 1. 1964; Meyerhof and Ranjan, 1973; Meyerhof and
Yalcin, 1994; Manoppo, 2010). Sastry et al. (1995)
suggested that the lateral capacity, and the magnitude
and position of maximum bending moment of a
flexible battered pile can be estimated by solving an
equivalent rigid vertical pile subjected to an inclined
load. Where, the load inclination angle equals to the
angle between the pile axis and the horizontal. Based
on the model tests performed by Sastry et al. (1995),
the maximum bending moment for positive battered
piles is higher than that of negative battered ones. Ong
(2015) investigated the effect of the interface
elementsensitivityparameters used in Finite Element
(FE) modelson the response of laterally loaded vertical
piles in sandy and clayey soils. The ratio between the
interface normal stiffness (Kn) and soil Young`s
Fig. 1. Types of battered piles (a) Negative battered
modulus (E) ata reference level should be from 1 to 10
pile (b) Positive battered pile
and 10 to 100 in sandy and clayey soils, respectively,
as concluded by Ong (2015). Hazzar et al. (2017)
The behavior of battered piles subjected to lateral
investigated the response of battered piles subjected to

20
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

lateral loads and the influence of vertical loads on the dimensional Beam on Elastic Foundation (BEF) (Eq. 1
pile lateral performance. FLAC3D (Itasca 2009), Finite and Fig. 2c) to an envisioned three-dimensional soil-
Difference computer software, was utilized by Hazzar pile interaction (Figs. 2a and b). Young’s modulus of
et al. (2017) to study battered piles in sandy soils and the soil (E) at the face of the passive soil wedge is
the results exhibited that the vertical load, pile batter related to the corresponding horizontal subgrade
angle, and the soil relative density significantly modulus (Es). It should be noted that the SW model
influence the lateral response of battered piles in sandy employs a soil stress-strain relationship ( - ∆σh)
soils. Hazzar et al. (2017) shows that the negative which is developed based on the concepts of the
batter piles lateral response is significantly dependent conventional triaxial test (Ashour et al. 1998). The
on the batter angle and sand relative density while the deflection pattern of the pile along its depth (y versus
positive battered does not seem to significantly depth x) is related to the soil strain (γ) developing in
fluctuate with batter angle and soil density. the passive wedge in front of the pile. Furthermore, the
There are limited studies from the literature that BEF line load (p) for a given deflection is related to
predict the lateral capacity and overall response of the the horizontal stress change (∆σh) acting at the face of
battered piles (i.e., pile deflection, bending moment, the mobilized passive wedge (Fig. 2b). More details
and shear force). So, this paper presents a detailed on the basics of the SW model are presented in Ashour
study on the response of battered piles subjected to et al. (1996 and 1998). Detailed SW formulations are
lateral loads in sandy soils using the SW method, presented in a different study to account for the pile
LPILE and MIDASGTS-NX as a Finite Element inclination (Ashour et al. 2018). The modified
numerical solution. formulations are employed in a FORTRAN code using
Strain Wedge Method Concepts the flowchart shown in Fig. 3.
The SW model correlates the traditional one-

Fig. 2. Strain Wedge model basic concept (Ashour et al. 2018)

21
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

Where EI is the pile flexure rigidity and Px denotes the axial load at the pile segment.

Fig. 3. The flow chart of the SW method

Due to pile battering, the applied lateral load (P) of the MIDAS model used in the analysis of the soil-
is decomposed intothe components Py1 and Px1 pile system. A pile with diameter (D) and total length
(perpendicular and axial components, respectively, (L) is embedded in the sandy soil with an inclination
Fig. 1). The geometry (i.e., the size) of the passive soil angle (β). While the total thickness of the soil stratum
wedge grows with advancing lateral load at the pile is selected as (L cos (β) + 6D) and the Tetrahedron
head. The size of the wedge changes as a function of mesh size is extended to a horizontal distance of 16D
the soil properties (i.e., friction angle, φ, effective unit from the center of the pile to release the effect of
weight, γ, and strain at 50% of stress, ε50) and pile model boundary conditions (Hazzar et al.2017). All
properties (pile diameter, D, bending stiffness, EI, and displacements are restrained at the bottom of the soil
pile-head condition). The basic SW model concepts domain while the external vertical faces are fully fixed
presented by Ashour et al. (2018)are applied in the x1 - in the X- and Y-directions.
y1 plane under lateral load Py1 where Py1 = P cos (β) LPILE and SW model are used to solve the
and Px1 = P Sin (β). β is the pile battering angle shown problem of laterally loaded piles as a BEF made up of
in Fig. 1. The pile head displacement (y1) in y1- a set of nonlinear p-y curves (i.e., non-linear springs)
direction (Fig. 1) is determined as a function of the representing the soil and pile properties contributions
pile deflection angle and associated strain in the to the resulting soil-pile interaction as demonstrated in
passive soil wedge. The pile head-deflection in the Fig. 2c. These p-y curves depend on the soil properties
(horizontal) y-direction is calculated as and the pile width (Reese et al., 1974) or account for
more pile properties such as the pile bending stiffness,
Modeling pile-cross sectional shape and pile-head fixity
Figure4portrays the general layout and meshing condition as employed in the SW model by Ashour

22
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

and Norris (2000). For battered piles, LPILE (Reese et 1.1. MIDAS Soil Modeling
al. 2004) applies a multiplier to the p-y curves of an The non-linear behavior of the sandy soil is
equivalent vertical pile to anticipate the response of simulated using the Mohr-Coulomb model which is
the battered one. In contrast, the SW approach commonly used in geotechnical engineering practice.
generates its own p-y curves along the pile based on Table 1 summarizes the soil properties used in the
the pile and soil properties. parametric study.

Table 1. Sand Properties Used in Parametric Study

Sand State Dr (%) γ(kN/m3) φ (degree) ν* E* (kN/m2) ε50
Loose 30 17 30 0.25 15000 0.0052
Medium Dense 60 19 35 0.30 30000 0.0037
Dense 80 21 40 0.35 60000 0.0028
* Values of soil Poisson`s ratio (ν) and modulus of elasticity (E) are assumed according to Bowles (1996).

Table 2. Piles Properties Used in Parametric Study

D (mm) t (mm) Interface Parameters
No. Material Section L (m) EI (kN-m2)
B f H w tf tw Sand State Qu (kN/m2) Kt (kN/m3) Kn (kN/m3)
Loose 2700 1800 19800
1 Steel H-Section 345 371 12.8 12.8 20 70578 Medium 4600 3400 38000
Dense 6300 6600 73000
2 RC Circular 345 - 20 70353 Medium 3100 3400 38000
Bf= Flange width t = Pipe thickness tf = Flange thickness
Hw = Web height tw =Web thickness

used herein to calculate ultimate skin friction (Qs).

Using a pile element thickness of 1m, the values of Qu
are determined for axial load as in Table 2. This
assumption is assumed to limit the failure to either the
soil or the pile and limit the relative displacement
between the pile and the soil. It should be noted that
MIDAS user manual has no criteria to calculate the
value of Qu for lateral load and battered pile. Also,
MIDAS user manual suggests the following empirical
formulations to define the values of Kt and Kn.

Fig. 4. Typical mesh used for MIDAS analysis

1.2. MIDAS Pile Modeling

The pile is modeled as a linear-elastic material
with the pile properties displayed in Table 2.
1.3. Soil-Pile Interface Modeling Where, L is the pile length and R is a strength
The soil-pile interface is defined in MIDAS reduction factor represents the friction between sand
using a pile element with three parameters: 1) the and pile which ranges from 0.6 to 0.7 in case of steel
ultimate shear force (Qu), 2) the shear stiffness and 0.8 to 1 for concrete. tv is a virtual thickness for
modulus (Kt), and 3) the normal stiffness modulus the interface element that has a value between 0.01 to
(Kn), each per total pile length. The ultimate shear 0.1 as recommended in MIDAS user manual and Ong
force is calculated by dividing the pile axial load by (2015). The higher the stiffness difference between
the pile length and the pile element thickness. In the soil and pile, the smaller the value of t v. νint. is the
absence of pile test data, MIDAS user manual interface Poisson`s ratio which has a value of 0.45
recommends using design standards to calculate the (MIDAS user manual).
ultimate skin friction. The β-method (Bowles 1996) is

23
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

Validation parameter tv which is used with a value of 0.1 in the

present study. More details about the sensitivity of t v
presented in Ong (2015).
Parametric Study
The SW model, MIDAS and LPILEprocedures
are utilized in performing a series of analyses on
laterally loaded battered piles embedded in loose,
medium dense, and dense sands. The current study
concentrates on the influence of typical soil/pile
properties on the lateral response of battered piles.
Such properties include the pile battering angle (β),
relative density of sandy soil (Dr), and pile cross
section shape. The study also presents a comparison
between the passive wedge size obtained from SW
Fig. 5. Calculated and measured response of model and MIDAS. Tables 1 and 2 present the soil and
laterally loaded battered pile at Arkansas River pile properties used in the analysis, respectively. For
each sand density, the values suggested by Bowles
Alizadeh and Davisson (1970) performed a full- (1996) for the sand modulus of elasticity (E) and other
scale lateral load test on negatively battered steel H- soil properties are employed. The response of battered
section pile (β = -18.43˚) with a flexure stiffness EI of piles is investigated for several values of batter angles
71081kN-m2 and a pile length (L) of 13.8m at β that ranged from -20 to +20 degrees.
Arkansas river site. As reported by Meyer and Reese Analysis and Results
(1979), the pile was a free-head driven pile in sandy 1.4. Pile batter angle
soil with 18.06 kN/m3 unit weight and 9.89 kN/m3 Figure 6 presents the lateral response of battered
submerged unit weight. The soil’s angle of internal piles for β = ±20˚ and ±10˚. The utilized medium
friction () was 38° and the water table was 0.4m dense sand and pile properties are presented in Tables
below the ground surface. The pile-head (i.e., the 1 and 2, respectively. A good agreement between the
location of loading (e), Fig. 1) was 0.15m above the SW model, LPILE, and MIDAS results can be
ground surface. For MIDAS, The pile interface observed in Fig. 6a with negative battered piles, which
parameters are used to be 1300 kN/m2, 4000 kN/m3, is not the case with the positive battered piles shown
and 45000 kN/m3 for Qu, Kt, and Kn, respectevely. in Fig. 6b. MIDAS results are not in good agreement
Figure 5 presents the measured pile-head response at with the predicted response of positive battered piles
the ground surface and the LPILE, MIDAS, and SW obtained from the SW model and LPILE even by
model results. A good agreement is found between the changing the value of tv from 0.01 to 0.1 (Figs. 6b and
measured data and MIDAS, LPILE, and SW results. 7) as recommended in MIDAS user manual and Ong
However, the results of MIDAS are very sensitive to (2015).
the interface element parameters specially the

Fig. 6. Response of laterally loaded battered pile a) Negative battered b) Positive battered

24
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

MIDAS can be used to predict the response of

laterally loaded positive battered piles by using tv
equals to 0.8 as shown in Fig. 8. The resistance of the
negative battered piles is greater than that of the
positive battered ones. Likewise, the higher the value
of negative battering angle the higher the stiffness of
the pile. In contrast, the higher the value of positive
battering angle the softer the stiffness of the battered
pile.
1.5. Sand relative density (Dr)
Fig. 7. Response of laterally loaded battered pile The pile-soil lateral resistance increases by
changing the state of sand from loose to dense for both
negative and positive battered piles as presented in
Fig.9. As shown in Fig. 9a, the difference between
MIDAS's predictions of lateral response of negative
piles and those from the other two techniques
decreases by increasing the sand relative densities to
exhibit very good agreement in the case of dense sand.
On the other hand, the SW model provides moderate
predictions for the positive pile in loose, medium
dense and dense sands compared to the results
obtained from MIDAS and LPILE (Fig. 9b). It can be
noticed that LPILE provides a softer pile head
response and MIDAS predicts a stiffer one compared
to the results of the SW model.
Fig. 8. Response of laterally positive battered pile
(tv = 0.8)

(b)
Fig. 9. Influence of sand relative density on the response of laterally loaded battered piles for a) Negative
battered b) Positive battered

Pile cross section Deflection pattern

The medium dense sand and piles No. 1 and 2 Figure 11 presents the profile of pile lateral
(Tables 1 and 2) are used to study the influence of pile deflections along the pile length. A very good
cross section on the lateral response of battered piles. agreement can be observed between the three
The SW and LPILE results indicate that the circular techniques used in this paper for the negative battered
cross section has less resistance compared to the pile pile (Fig. 11a). As previously noted, MIDAS results
with an H-section. On contrast and unexpectedly, are obtained based on tv = 0.1 do not match those
MIDAS provides a stiffer pile head response for the predicted by other methods for positive battered piles
pile with a circular cross section as demonstrated in (Fig. 11b).
Fig. 10.

25
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

moments and shear forces, Figs. 12 and 13)

determined from MIDAS are less than those calculated
from the other two techniques.
1.7. Wedge size
Once the battered pile deflects, a mobilized
passive soil wedge develops in front of the pile. The
wedge size/geometry is controlled by the mobilized
fanning angle (φm), the width of the wedge face ( ),
and the wedge height (h) (i.e., the depth to the first
zero-deflection point) (Fig. 2). Figure14portrays the
wedge size at the ground surface predicted using SW
Fig. 10. Pile lateral deflection profile for Negative method and MIDAS. A reasonable agreement can be
battered pile observed for the negative battered pile as shown in
Fig. 14a. As in Fig. 14b, MIDAS predicts larger soil
1.6. Bending moment and shear force passive wedge which may explain the relatively stiff
Figures12and 13 display the bending moment response of the positive battered piles compared to the
and shear force diagrams along the pile, respectively. SW model one (Figs. 6b, 7and8b). The value of h of
For the same lateral load, the maximum bending the negative battered pile is equal to 3.17m and 4.0m
moments of positive battered piles are greater than for the SW model and MIDAS, respectively, as shown
those of the negative battered piles which agree with in Fig. 11a. For the positive battered pile, the SW
the results presented by Sastry et al. (1995). The model and MIDAS provide an h of 3.62m and 3.65m,
medium dense sand and pile No. 1 properties respectively, as demonstrated in Fig. 11b.
described in Tables 1 and 2, respectively, were utilized
in the analysis. The straining actions (i.e., bending

Fig. 11. Pile lateral deflection profile a) Negative battered b) Positive battered

Fig. 12. Bending moment diagram, a) Negative battered b) Positive battered

26
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

Fig. 13. Shear force diagram, a) Negative battered b) Positive battered

Fig. 14. Wedge size at ground surface predicted by SW and MIDAS a) Negative battered b) Positive battered
piles

Summary and Conclusion the predicted straining actions.

The SW model, LPILE, and MIDAS GTS-NX
are used to assess the response of battered piles References
subjected to lateral loads. The SW approach predicts 1. Alizadeh, M., and Davisson, M. T. (1970).
the behavior of laterally loaded batter piles as a “Lateral load tests on piles-Arkansas River
function of soil and pile properties (bending stiffness, project.” Journal of Soil Mechanics and
cross-sectional shape, and head fixity). As expected, Foundations Div. 96(5), PROC PAPER 7510,
the lateral behavior of batterd piles is highly 1583 - 1604.
influenced by the magnitude and direction of the pile 2. Ashour, M., Alaa Eldin, A., and Arab, M. G.
batter angle in addition to the soil properties (i.e., sand (2018). “Battered Piles in Sandy Soils under
relative density). The SW model and LPILE can be Lateral Loads.” Under review, J. Geotech. and
used to determine the pile lateral deflection and Geoenviron. Eng., ASCE, GTENG-7087.
straining actions (i.e., bending moment, and shear 3. Ashour, M., Norris, G., and Pilling, P. (1998).
force) along the pile length for negative and positive “Lateral loading of a pile in layered soil using the
battered piles unlike MIDAS which predicts less Strain Wedge model.” J. Geotech. and
bending moments and stiffer pile head response with Geoenviron. Eng., 124(4),
positive battered piles. MIDAS can be utilized to doi:10.1061/(ASCE)1090-
analyze the positive battered piles by modifying the 0241(1998)124:4(303), 303-315.
controlling parameter of the interface element (i.e., the 4. Ashour, M., and Norris, G. (2000). “Modeling
virtual thickness, tv) to be0.8. However, MIDAS lateral soil-pile response based on soil-pile
interface element needs extended sensitivity analysis interaction.” J. Geotech. and Geoenviron. Eng.,
to relate the interface parameters to soil and pile doi:10.1061/(ASCE)1090-
properties including the pile battering angle. Despite 0241(2000)126:5(420),420-428.
the good agreement of the SW results with field test, 5. Ashour, M., Pilling, P., Norris, G., and Perez, H.
additional field results are surely needed to validate (1996). “Development of a strain wedge model

27
 New York Science Journal 2018;11(7) http://www.sciencepub.net/newyork

program for pile group interference and pile cap 13. Murthy, V. N. S. (1964). “Behavior of battered
contribution effects.” Report – CCEER-96-4, piles embedded in sand subjected to lateral
FHWA, and CALTRANS, University of Nevada, loads.” Proc., Symp. on Bearing Capacity of
Reno, 274pp. Piles, CBRI, Roorkee, India, 142–153.
6. Bowles, L. E. (1996). “Foundation analysis and 14. Nimityongskul, N., Barker, P., and Ashford, S.
design.” McGraw-hill. (2012). “Effects of Soil Slope on Lateral
7. Hazzar, L., Hussien, M.N. and Karray, M., Capacity of Piles in Cohesive and Cohesionless
(2017). Numerical investigation of the lateral Soils.” Research Report of a Research Project
response of battered pile foundations. Funded by Caltrans under Facilities Contract No.
International Journal of Geotechnical 59A0645, Oregon State University, Kiewit.
Engineering, 11(4), pp.376-392. 15. Ong. Y. H, (2016). Back analysis of laterally
8. Juvekar, M. S., and Pise, P. J. (2008). “Behavior loaded pile behavior using Midas/GTS to
of rigid batter piles and pile groups subjected to determine stiffness modulus of pile-soil interface.
horizontal load in sand.” Indian Geotechnical Japanese Geotechnical Society Special
Journal, 38(2), 221-242. Publication, 2(35), 1279-1284.
9. Manoppo, F. J. (2010). “Behavior of the ultimate 16. Reese, L. C., Cox, W. R., and Koop, F. D.
bearing capacity of single flexible batter pile (1974). “Analysis of laterally loaded piles in
under horizontal loads in homogeneous sand.” Offshore Technology in Civil Engineering
sand.” Dinamika TEKNIK SIPIL 10(2), 116-119. Hall of Fame Papers from the Early Years, 95-
10. Meyerhof, G. G., and Ranjan, G. (1973). “The 105.
bearing capacity of rigid piles under inclined 17. Reese, L. C., Wang, S. T., Isenhower, W. M., and
loads in sand. II: Batter piles.” Canadian Arrellaga, J. A. (2004). Computer program Lpile
Geotechnical Journal, 10(1), 71-85. plus version 5.0 technical manual. Ensoft,
11. Meyer, B. J., and Reese, L. C (1979). “Analysis Austin, TX.
of single piles under lateral loading.“ Report (No. 18. Sastry, V.V. R., Koumoto, T. and Monoppo, F.J.
FHWA-TX-79-38+ 244-1 Intrm Rpt.), TX: “Bearing capacityand bending moment of
Center for Highway Research, University of flexible batter piles in homogeneous soilunder
Texas at Austin. horizontal loads”. Bul, fac. Agr. Saga univ. 79,
12. Meyerhof, G. G., and Yalcin, A. S. (1994). 77-85, 1995.
“Behavior of flexible batter piles under inclined 19. Technology, M. I. 2014. Midas Users Manual.
loads in layered soil.” Canadian Geotechnical GTS NXV2.1, Korea.
Journal,30(2), 247-256.

7/11/2018

28