CGJ 2018 0326

Page 1 of 44
Analysis of a Case Study Presenting Ground Anchor Load Transfer

Response in a Shale Stratum
L. Sebastian Bryson, 1 Jorge Romana Giraldo2
Downloaded from www.nrcresearchpress.com by UPPSALA UNIVERSITETSBIBLIOTEK on 08/24/19. For personal use only.
1Hardin-Drnevich-Huang Associate Professor, Department of Civil Engineering, University of

Kentucky, Lexington, KY 40506, USA; e-mail: sebastian.bryson@uky.edu
2Graduate Research Assistant, Department of Civil Engineering, University of Kentucky,
Lexington, KY 40506, USA; e-mail: jorge.romana@uky.edu
Abstract: This paper presents the analysis of the instrumentation and monitoring of ground
anchors for a landslide stabilization system located at the State Route 82 Bridge, near Brecksville,
Summit County, Ohio (SUM 82). The stabilization system consisted of two tieback walls. Site
conditions and design aspects are considered; also, the construction sequence is presented. Field
performance data were collected during and after construction, representing a total monitoring
Can. Geotech. J.
period of about 340 days. Load transfer was studied based on strain gage measurements along the
bonded length of the anchors. Load variation during preproduction and performance tests of the
anchors is described and discussed. A numerical procedure is presented to evaluate the anchor
load-deformation characteristics based on the t-z approach. Two strain softening load transfer
functions were used at the interface between the bonded length of the anchors and the rock. The
analysis showed that the prediction of the load-displacement curve for each anchor is more
influenced by the selected ultimate side frictional resistance along the anchor-rock interface than
the shear stiffness of the interface. This study concluded that it is critical that peak and residual
load characteristics are included in the load transfer models for anchors installed in rock.
Key words: tieback wall; ground anchors; load transfer; t-z curves; strain softening; shale.
1 Introduction
Ground anchors have been widely used to support retaining walls for deep excavations, stabilize
natural and man-made slopes, prevent uplift of submerged structures, and provide support for
bulkhead and wharf systems (Littlejohn 1997; Sabatini et al. 1999; Ehrlich and Silva 2015).
Although there have been significant advances in utilizing new materials to improve the
1
Page 2 of 44
performance of ground anchors, design methods have remained essentially the same (Xu and Yin
2016).
Typically, the ultimate load capacity is estimated via a limit equilibrium analysis that utilizes
presumptive ultimate bond stress along the bonded zone of the anchors. The presumptive ultimate
bond stress is typically given in various design manuals (Sabatini et al. 1999) or prescribed codes
(PTI, 2014). This approach assumes uniform bond stress along the bonded length. However, it is
recognized that this stress distribution is not uniform in reality (Benmokrane et al. 1995; Sabatini
et al. 1999; Kim 2003).
The stress distribution might be uniform in weak rock such as clay shales, but in more competent
rocks the bond stress is mobilized only in the upper portion of the anchor bonded length (Sabatini
et al. 1999). Given the recommended range of presumptive bond strength at the ground/grout
interface, selection of the appropriate value of bond strength is highly subjective to the experience
of the designer. According to Hegazy (2003), the use of the maximum recommended bond strength
may lead to inadequate designs when factors of safety of 2.0 and 3.0 are applied in soil and rock,
Can. Geotech. J.
respectively.
Several authors have used a non-linear relationship along the ground-grout interface to analyze the
behavior of ground anchors subjected to tensile loads. Kim (2003) analyzed the load transfer in
tension anchors using a purely elastic-plastic model. The analysis was divided into two parts; one
part corresponded to the anchor-soil interface and the other to the strand-grout interface. More
recently, Xu et al. (2014) proposed a hyperbolic load–displacement model to study the deformation
characteristics in a jet mixing anchor-pile support system. Another study reported by Liu et al.
(2017) used a broken line model to analyze the distribution of stress transfer along the bonded
length of ground anchors. This model assumes strain softening (i.e. the negative transition between
the ultimate skin friction and the residual stress) is given by an abrupt change in the load transfer
relationship, which might be unrealistic.
These aforementioned studies support the idea of using a more realistic load transfer function along
the bonded length of ground anchors to predict their load-deformation responses. A case history
of an instrumented anchored wall is thus presented herein to test the validity of using a nonlinear
load transfer function to better predict the ground anchor response.
2
Page 3 of 44
For this study, two softening models that assume a smooth transition between the ultimate and the
residual side resistances were considered. The first model, hereafter referred to as Model 1, was
proposed by Zhang and Zhang (2012). This model depends on three parameters, namely ultimate
skin friction, relative displacement at the ultimate skin friction and the ratio between the residual
skin friction and the ultimate skin friction. The second model, hereafter referred to as Model 2,
was proposed more recently by Ni et al. (2017). This model depends on a strain softening
parameter n and the ultimate skin friction to describe load transfer behavior.
This paper presents the analysis of the instrumentation and monitoring of a tieback wall
construction project located at the State Road 82 Bridge, near Brecksville, Summit County, Ohio
(SUM 82). This project included the construction of two tieback walls as a solution to an unstable
slope under the State Road 82 Bridge in the Cuyahoga Valley National Recreation Area (CVNRA).
Data for this project was initially reported by Liang (2000) and includes measurements from strain
gages installed in the anchor bonded length, inclinometers, load cells and piezometers recorded
during construction of the walls. Specifically, the data included anchor preproduction tests and
Can. Geotech. J.
performance test, and strain gage measurements at 0.9 m, 2.1 m and 3.4 m from the top of the
bonded length. This paper presents the analysis of two preproduction test and three performance
tests conducted during the construction of two tieback walls. From the data collected, the load
transfer mechanism was investigated numerically using load transfer models that include softening
behavior. Also, strain gage data were used to assess the long-term variation of the loads along the
bonded length of the anchors.
2 Project Description and Site Conditions
Near the SUM 82 project site, the side slopes rise sharply from the valley for a change in elevation
of 12 m to 15 m over a horizontal distance of 15 m to 18 m. The side slopes continue to rise at a
more moderate slope for an elevation change of 9 m to 12 m over a horizontal distance of 30 m to
60 m; then rise steeply with a change in elevation of 9 m over a distance of 9 m to 18 m. These
slopes are prone to ground movements that extended distances of 60 m and 120 m to the north and
south of the State Route 82 Bridge centerline, respectively. Slumps and block movement were
identified throughout the area. Also, surface erosion was observed in the lower portion of the slope.
Two soldier pile and lagging walls were constructed at the location as a means to control the
ground movement for the unstable slope and to preserve the SUM 82 bridge foundation. The
3
Page 4 of 44
soldier pile and lagging walls were also intended to prevent any soil and rock debris from flowing
unto the railroad tracks adjacent to the slope. A general site location map showing the location of
the walls is presented in Fig. 1.
A typical profile view of the project site is shown in Fig. 2. From the subsurface investigation,
three geologic strata were identified at the site. An upper soil layer consisted of brown silty clay
and clayey silt combined with loose to medium sand. These deposits varied vertically and
horizontally without a consistent pattern throughout the site. A lower soil layer consisted generally
of gray clayey silt and silty clay with a consistency that was typically stiff to very stiff. The lowest
part of the soil profile consisted of the shale stratum. The boring data showed the shale surface
dropped sharply to the west with a north-south tendency. The shale was mostly, black,
carbonaceous with numerous thin clay seams scattered throughout. The top of the shale stratum
was moderately weathered to severely weathered with the degree of weathering decreasing with
depth.
During construction of the walls, 68 rock anchors were installed through the soil layers, into the
Can. Geotech. J.
shale stratum. Three rows of anchors were installed in the upper wall and two rows of anchors
were installed in the lower wall. The anchors installed in the upper wall were inclined 45 degrees
with respect to the horizontal, while the rock anchors of the lower wall had an inclination of 15
degrees. Additionally, anchors from the lower wall were angled horizontally to avoid any
intersection with the embedded portion of the soldier piles for the upper wall. Also, anchors in the
upper wall were angled to avoid intersection with the bridge pier.
The subsurface investigation consisted of seven soil borings with standard penetration testing that
extended to the shale. The locations of the soil borings are presented in Fig 1. Based on data
presented in the boring logs, estimated composite soil characteristics were developed and are
presented in Fig. 3. The natural moisture content was fairly uniform with depths ranging between
20 and 30 percent. Blow counts showed a similar trend. Undrained shear strength was obtained
from unconfined compression tests and appeared to increase as a function of the effective
overburden stress, with a considerable increase observed near the shale layer.
4
Page 5 of 44
3 Design Approach
Prior to installation of the anchors, the clay stratum was excavated and replaced with a granular
backfill. Therefore, the Peck (1969) apparent earth pressure envelope for granular backfill was
used to determine the anchor loads.
Because the Peck (1969) envelope was intended for a horizontal ground surface and the backfill
in both walls was inclined, a modified Coulomb active earth pressure coefficient, K a , was used in
the design and is given as
cos( ) 2
Ka  2
(1)
 sin( ) sin(  w) 
1  
 cos( w) 
Equation (1) corresponds to an active earth pressure coefficient on a vertical wall with no friction
and backfill inclination of w ;  is the peak soil friction angle. Therefore, the maximum intensity
of the apparent earth pressure was Pa  0.65 K a H (  is the total unit weight of the soil and H is
Can. Geotech. J.
the retained height) as shown in Fig. 4. Based on the applied pressure distribution and the tributary
area method, the design anchor load for the upper wall was 343 kN for a separation of 2.4 m
between soldier piles. Conversely, for the lower wall, the anchor design load was 489 kN for a
given separation of 3 m.
Selection of the unbonded length of the anchors was based on the Rankine failure plane and an
offset of 1/5 the height of the excavation, as recommended by Sabatini et al. (1999). The anchor
unbonded length in both rows of anchors was equal in the lower wall. However, in the upper wall
the unbonded length was selected such that the bonded length was installed in the shale stratum.
4 Anchors and Wall Installation Procedures
The solider pile and lagging wall consisted of HP 14x73 piles placed in pre-drilled holes, with 76
mm-thick hardwood temporary lagging (Liang 2000). The upper wall soldier piles were installed
first. The lower wall soldier piles were installed after the completion of the upper wall piles
installation. The holes for both the upper and lower wall piles were pre-drilled in the shale. The
holes were 762 mm in diameter and 6.1 m deep. The length of each of the upper wall soldier piles
was approximately 15.2 m. The length of each of the lower wall soldier piles was approximately
5
Page 6 of 44
13.7 m. The piles were lifted by crane and lowered into place. The piles were then fixed to a
supporting frame to maintain alignment and then concrete was placed using tremie techniques.
A total of 68 rock anchors were installed. The holes for the anchors were cased to depths ranging
from 0.61 m to 1.5 m into the shale to prevent the boreholes from collapsing and to support the
jacking system while testing the anchors. The free length of the anchors ranged from 5.2 m to 17.1
m and the bond length was approximately 4.6 m. Among the anchors, six anchors were
instrumented each with three Geokon Model 4410 vibrating wire strandmeters along their bonded
length. The general construction sequence of the project is shown in Table 1.
As registered in the subsurface investigation, the soil conditions at the project site were variable
horizontally and vertically with respect to soil types and consistency. This variability made it
necessary to conduct preproduction tests prior to the installation of production anchors. Two
failure tests were performed on fully instrumented non-production rock anchors at the start of the
project to ensure the load carrying capacity of the rock anchors and to study the load transfer
mechanism of the anchors.
Can. Geotech. J.
One failure test was performed for the upper wall and one test was performed for the lower wall.
The upper wall test anchor (PT1) was installed between Soldier Piles 26 and 27 and the lower wall
test anchor (PT2) was installed between Soldier Piles 6 and 7. Both load tests were conducted after
installation of the solider piles in both walls. The locations of these tests are shown in Fig 1 and
are identified as PT1 and PT2, for the upper and lower wall tests, respectively.
The anchor holes for both upper and lower wall tests had a diameter of 101.6 mm. The anchor
tendons were composed of 7-wire strands. Although both test locations were relatively close, some
differences were noted in the testing conditions. The upper wall test had a 139.7 mm diameter steel
casing in the unbonded length that was penetrated into the shale stratum 1.2 m to prevent the
anchor hole from collapsing. Additionally, in the lower wall test, the unbonded and bonded length
of the anchor was installed completely in the shale stratum. The schematic of the preproduction
anchors for the upper wall test are shown in Fig. 5 (Fig. 5a). A similar anchor configuration was
used in the lower wall test but without the steel casing. Both load tests were conducted against the
ground surface using a bearing plate.
There was a concern at the site that the shale stratum would absorb water as a result of infiltration.
Consequently, the anchor holes were filled with water after drilling and before installation of the
6
Page 7 of 44
tendons, to study the effect of water in load-carrying capacity. Water remained seven days for the
lower wall test and 9 days for the upper wall test until insertion of the tendon and grouting of the
anchor. Filling the hole with water and letting it sit for seven days is similar to the common slake
test (i.e. immersing shale samples in water and observing the amount of deterioration). Anchor
load test procedures were conducted according to the Post-Tensioning Institute guidance (PTI,
2014). Specifically, after 1.33 times the design load (Table 2), the load was increased to 1.5, 1.75,
2.0, 2.25, 2.5, 2.75 and 3.0 times this load.
Characteristics of the tested anchors are presented in Table 2. In addition to the preproduction load
tests, performance tests were also conducted in several production anchors. A typical production
anchor installed in the project is presented in Fig. 5 (Fig. 5b). This figure corresponds to the
anchors installed in the upper wall. Anchors in the lower wall had a similar configuration but were
composed of four tendons.
5 Load Test Results

Can. Geotech. J.
5.1 Preproduction Tests
The preproduction tests were intended to go to complete pullout of the anchors, causing failure at
the grout-ground interface and facilitating the investigation of the load transfer characteristics for
the anchors in shale. However, the load was limited to 80 percent the guaranteed ultimate tensile
strength of the tendon (G.U.T.S). Therefore, the applied load for PT1 was stopped at 2.75 times
the design load and for PT2 the applied load was stopped at 3 times the design load. The load-
displacement curves for both tests are shown in Fig. 6. During the tests, load-unload cycles were
applied until 120 percent of the design load, with hold times varying with the load application. At
25 percent of the design load, the hold time was 10 minutes and at 120 percent of the design load,
the hold time was 60 minutes. A creep test was performed for both PT1 and PT2 at 133 percent of
the design load (Table 2), by using a hold time of 300 minutes. Fig. 6 shows that from the
beginning of the tests until the maximum applied load, the load-displacement curves in both tests
did not show a significant degradation in the anchor capacity. Consequently, the load-displacement
relationship was approximately linear.
Fig. 7 (Figs. 7a and 7b) presents the components of anchor movement. To ensure proper alignment
of the equipment during the tests, a load of 20 percent the design load, termed the alignment load,
7
Page 8 of 44
was applied at the beginning of the tests. The total movement measured included the elastic
elongation in the unbonded length of the anchor and the residual movement. The residual
movement was considered as the non-recoverable movement after the load application.
Analysis of the test results was mostly focused on the preproduction tests (PT1 and PT2). The
ultimate anchor load was defined using the Briaud et al. (1998) criteria;
 Failure Criterion 1 – The ultimate load is the load at which the residual movement is one-
tenth of the anchor diameter (D/10) or,
 Failure Criterion 2 – The ultimate load is the load at which the total displacement is (D/10)
plus the elastic elongation of the anchor unbonded length.
The residual movement measured until the applied load was 120 percent the design load is
presented in Fig 7 (Figs. 7a and 7b). Based on Failure Criterion 1, neither of the anchors failed
before reaching this load. However, the residual movement in the PT1 was close to this threshold.
The maximum residual movement observed in PT1 was 7.2 mm, while in PT2 this value was
Can. Geotech. J.
3.1 mm.
Fig. 7 (Fig. 7c) also shows the results of applying Failure Criterion 2 to the tests with the exception
that the reloading cycles during the test are not presented in the figure. As seen in the figure, the
criterion predicted an ultimate load that was less than the design load for PT1. Based on these test
results, it was unlikely that the ultimate capacity of the anchor was reached under the range of
applied loads because of the nearly linear load-displacement curve. For this criterion, the unbonded
length of the anchor is assumed constant. However, because of loading, debonding at the top of
the bonded length may occur (Benmokrane et al. 1995). During debonding, the free length of the
anchor increases and consequently the elastic elongation of the strands lead to a more linear load-
displacement curve.
Additionally, the application of Failure Criterion 2 to PT2 (Fig. 7d) shows the failure load was not
reached under the range of loads applied. In this case, it remains uncertain of which load could be
considered as the ultimate load. Given that the maximum applied load was about 3 times the design
load in both tests and the anchor movements were considered adequate, the bonded length was
assumed satisfactory.
8
Page 9 of 44
As stated by Barley (2005) to determine the ultimate or pullout load, larger displacements than
those calculated using the criterion proposed by Briaud et al. (1998) may be necessary. Liu et al.
(2017) presented the results of complete pullout tests in a slightly weathered limestone. The results
of two of the anchors tested by Liu et al. (2017) are shown in Fig. 8 as A1 and B1. In order to
compare the results, these tests were normalized with respect to the maximum applied load (Pmax)
and the diameter of the anchor (D). Additionally, Failure Criterion 2 proposed by Briaud et al.
(1998) is presented in the figure.
Based on the obtained results from PT2 and results from Liu et al. (2017), it is suggested to define
the ultimate load as the load corresponding to the elastic elongation of the free length plus 1/3 to
1/6 the diameter (D) of the anchor. The proposed range of evaluation of the ultimate load is shown
also in Fig. 8. This range is considered appropriate because more mobilization of shear strength
can be accepted and it is still conservative given that maximum observed load was not included.
The suggested range is given only for anchors installed in rock formations, in other ground
conditions suggestions might not apply.
Can. Geotech. J.
5.2 Performance Test
Performance tests were also conducted at the project site to verify the capacity of the anchors. As
with the preproduction tests, these tests were performed in accordance with recommendations
given by the Post-Tensioning Institute (PTI 2014). Results of the performance tests of the anchors
installed in Soldier Pile 30 in the middle row (B) and lower row (C) are presented in Fig. 9 (Fig 9a,
Fig 9b). Also, Fig. 9 (Fig. 9c) shows the result of the anchor installed in the middle row of the
Soldier Pile 31. The locations of these soldier piles were shown previously in Fig. 1.
The load-displacement response of the three performance tests was similar; maximum
displacements were observed in the middle anchors 30B and 31B. Given the same bonded length
in all the tests, these large displacements in the middle anchors could be attributed to the larger
unbonded length with respect to the row C of anchors.
The test load was held 10 minutes to investigate the creep characteristics of the anchors according
to the Post-Tensioning Institute procedure (PTI 2014). Based on the anchor movements, the creep
acceptability criterion was satisfied and the load at the anchor head was consequently reduced to
the lock-off load (load carried by the tieback after testing and/or stressing). A rough comparison
9
Page 10 of 44
of the behavior of the bonded length during the three tests can be observed in Fig. 10. The residual
movement was plotted against the anchor load in the figure. It is noted that the slope of the curves,
(i.e. anchor stiffness) was quite similar suggesting that the anchors were successfully installed in
the shale stratum and indicating a similar stress-strain response in the bonded length of the anchors.
6 Analytical Model
Based on the results of several pullout tests from different authors (Hsu and Chang 2007; Liu et
al. 2017), it is observed that during pullout of ground anchors some degree of softening is
manifested in the load-displacement response. Consequently, analytical models considering
softening were used to analyze the results of the test in the SUM82 project. Although the unbonded
length of the anchors was also grouted, the strands of the anchors along the unbonded length were
sheathed (see Fig. 5). Also, the grout along this length was not placed under pressure. Thus, it was
assumed that load transferred was negligible along the unbonded length. Load transfer was
assumed to occur only along the bonded length of the anchors. Therefore, the load at the top of the
bonded length was assumed equal to load at the top of the anchor head.
Can. Geotech. J.
Considering equilibrium of the segment of length dx shown in Fig. 11

dP  x 
 2 ra  x  (2)
dx
where P( x) is the axial load along the anchor; ra is the radius of the anchor and  ( x) is the shear
stress along the anchor bonded length. The relation between the axial strain and the axial load can
be expressed as:
dw( x) P( x)
 ( x)   (3)
dx AE
where w is the displacement at a distance x measured from the bottom of the anchor; A is the
cross-sectional area of the bonded length; E is the elastic modulus of the anchor, which can be
calculated based on a composite section.
Now, by taking the derivative of Equation (3) and substituting into Equation (2), the basic
governing equation for the ground-anchor interaction can be expressed as:
d 2 w( x) 2 ra
  ( w) (4)
dx 2 AE
10
Page 11 of 44
Equation (4) can be solved using the t-z technique (Coyle and Reese 1966). In order to solve
Equation (4), two softening models were considered for the side resistance  and the displacement
w.
6.1 Softening Model 1 (Zhang and Zhang 2012)
Zhang and Zhang (2012) proposed a softening model based on three parameters a, b, and c. The
relationship between the skin friction and relative displacement is given by Equation (5) as
w( x)(a  c  w( x))
 ( x)  (5)
(a  b  w( x)) 2
The three parameters a, b, and c are expressed as
 s  1  1   s wu
a  (5a)
2  s u
1  1  s 1
b  (5b)
Can. Geotech. J.
2  s u
2  s  2 1  s 1
c  (5c)
4  s u
where  u is the limiting unit skin friction;  s is the ratio between residual skin friction and ultimate
skin friction; and wu is the relative displacement at limiting friction. Additionally, the reciprocal
value of a corresponds to the initial slope,   1/ a , of  ( x) . Oda et al. (1997) proposed that 
can be estimated using Equations (6) and (67)
r  g
 (6)
r  g
where r is the shear stiffness of the rock and g is the shear stiffness of the grout. The parameter
g is calculated as
G
g  (7)
 r r 
ra  ln 1  a b 
 rb 
11
Page 12 of 44
where G is the shear modulus of the grout; ra is the radius of the grout; and rb is the radius of the
tendon assembly. For the case where the initial slope is calculated, the value of wu will be
a
wu  (8)
b  2c
Consequently, for this model in ground anchors, either the value of wu can be assumed or the
initial slope can be estimated based on Equations (6) and (7). According to Liu et al. (2017), the
shear stiffness of rock varies from 1.5 GPa/m to 3 GPa/m for soft rock and from 5 GPa/m to 10
GPa/m for stiff rock.
6.2 Softening Model 2 (Ni et al. 2017)
More recently, Ni et al. (2017) introduced a softening relationship based on a single parameter, n
(n  0) and the limiting unit skin friction to describe the variation of the skin friction with the
relative displacement as
Can. Geotech. J.
 n

  w( x)  n 1  w( x)  
 ( x)   u (n  1)    n  for w  x   2 wu (9)
  wu   wu  
 
For values of w( x)  2 wu , the value of  at wu is used. Based on a given ratio  s between the
range of 0.65-0.90, the parameter n used in Equation (9) can be approximated as
n  0.16  9.11 (10)
Fig. 12 presents both softening models described previously. It can be seen that for the same ratio
 s , Model 1 decreases gradually after the maximum skin friction is achieved. In contrast, Model
2 shows an abrupt change in the load-transfer curve when the relative displacement is given as
w  2 wu .
7 Algorithm for the Load Transfer Approach
A load-displacement curve for an anchor can be estimated using the t-z approach (Coyle and Reese
1966; Knellwolf et al. 2011) and calculations can be easily adapted to a spreadsheet. Considering
the segment n shown in Fig. 13, the displacement at the middle of the segment is giving by:
12
Page 13 of 44
wmn  wbn   en (11)
where wbn and wmn are the displacements at the bottom and middle of the segment, respectively;
 en is the elongation of the half segment. Assuming an average load in the middle segment,  en
is expressed as:
 Pbn  Pmn  dl / 2
 en     AE (12)
 2 
Now, considering equilibrium of the segment
dl D
Pmn  Pbn   ( wmn ) (13)
2
where D is the diameter of the anchor and   wmn  is the mobilized skin friction. Substituting
Equations (12) and (13) into (11) gives:

 dl D  dl / 2
wwn  wbn   Pbn   ( wmn )  0 (14)
 2  AE
For a known value of wbn , wmn can be calculated iteratively from Equation (14) for a given
Can. Geotech. J.
tolerance. Fig. 13 presents a flowchart with the necessary steps to obtain the load-displacement
curve.
8 Analysis and Results
8.1 Analysis of Preproduction Test Data
Load-displacement curves for the preproduction and performance tests were determined using the
aforementioned procedure. Table 3 presents the input parameters used in the analysis. The ultimate
side frictional resistance was limited to 1400 kPa based on the average ultimate bond stress
presented by Sabatini et al. (1999) for shales, also given by Xanthakos (1991). However, it was
noticed that the ultimate fictional resistances were likely greater than the assumed values, given
that the values reported were average values. Nevertheless, complete pullout was not reached
during the tests and only a portion of the load-displacement curve was compared.
The load-transfer relationship at the anchor-rock interface is presented in Fig. 14a for the two
models adopted for the calculations. The value of n for Model 2 was selected based on a  s value
13
Page 14 of 44
of 0.7. This value was chosen based on results of Liu et al. (2017) that showed considerable
softening in the anchor response.
The estimated response for PT2 is shown in Fig. 14b. As can be seen in the figure, the calculated
response is similar to the measured response during the test. Both models did well in predicting
the measured data of the preproduction test. Load-displacement curves were calculated to obtain
a representation after the peak load.
As the preproduction test did not reach complete pullout, the load-displacement measurements
during the test did not display any non-linear behavior, which might develop under ultimate load
conditions including any softening. It is noted that the predicted response did not consider any
cycles of loading during the test. Therefore, any progressive degradation of the ground-anchor
interface was assumed negligible.
8.2 Analysis of Performance Test Data
For the performance test anchors, the minimum required unbonded lengths were given by Liang
Can. Geotech. J.
(2000) and are presented in Table 2. However, actual values of the unbonded length for these
ground anchors were not reported. Given that the ground anchors were installed according to the
recommendations of the Post Tensioning Institute (PTI), the unbonded length can be estimated
based on acceptance of the minimum apparent free tendon length criterion. Estimated values of
the unbonded length for the anchors 30B, 30C and 31B are 12 m, 9 m, and 12 m, respectively.
Calculated load-displacement responses for the performance test conducted in the anchors 30B,
30C and 31B are shown in Fig. 15. A friction coefficient   0.05 was assumed for the unbonded
length of the anchors. As seen from the figure, results show satisfactory agreement with the
measured response. However, because the applied loads were likely much less than the ultimate
load, most displacement observed corresponded to elongation of the free length of the anchors.
From Fig. 15 (a) and (c), it can be observed that the measured response was stiffer than the
estimated. This difference may be attributed to more friction developed along the unbonded length
of the anchors.
Because load is transferred from the top of the bonded length towards the bottom, under small load
the bottom of the bonded length might not experience any movement. However, to approximate
14
Page 15 of 44
the response used in the presented algorithm, a small value of bottom displacement has to be
assumed.
Parametric analyses were conducted to determine the relative implication of the input parameters
used in both models. The following anchor characteristics were constant for the analyses:
lbon =5 m ; lunb =10 m ; d  0.1 m ; As  4.2 104 m 2 and Es  200 GPa . Fig. 16 presents the results
of the analysis varying the ratio  and the displacement to the ultimate load zu for Model 1.
As seen in Fig. 16 (Fig. 16b), a variation of the parameter  had a noticeable impact on the load-
deformation curve. Specifically, a lower value of resulted in a lower ultimate load for the anchors.
Also, as  increased, the displacement at the maximum load increased continuously. The
implication from the analysis is that the most adequate means to obtain  is most likely from
pullout load tests performed at the site.
Because the t-z approach presented herein is only valid for monotonic loadings, under certain
values of  , such as   0.7 , the method will result in a decrease in load and displacement for a
Can. Geotech. J.
given bottom displacement. This behavior is physically inadmissible, therefore the red portion of
the curve, shown in Fig. 16.b, was disregarded. Similar behavior was discussed by Ren et al. (2010)
and Blanco (2012).
Fig. 16 (Fig. 16c) also presents the results of load displacement when the displacement at
maximum skin friction is varied. For this parameter, Kim et al. (2007) assumed a value of 2.5 mm.
Also, Liu et al. (2017) suggested displacement values of 1.5 mm and 1.8 mm. From Fig. 16d, it
can be seen that there is minor variation in the maximum load at the anchor head. Only 3.4 percent
variation in the maximum load was observed in the range of selected values. Consequently, the
maximum load is seen to depend mostly on the ultimate value of the skin friction  u and the degree
of softening represented by  .
For the softening Model 2, Fig. 17 (Fig. 17a) shows the results of varying the parameter n . As
seen, the variation of the load-displacement response is similar to the load-displacement response
of Model 1 with variation of the parameter  . The results of this model tend to show the same
aspect of load and displacement reduction. This tendency is related to the shape of the load transfer
15
Page 16 of 44
curve. As seen in Fig. 16, after the maximum skin friction is reached, Model 1 presents a more
gentle variation for the mobilized skin friction for an equal value of  .
It can be seen from Fig. 16 and Fig. 17 that the anchor response is almost linear until the maximum
load is reached; this behavior is a consequence of the elongation of the unbonded length, which is
significantly greater than the movement, and elongation of the bonded length. After the ultimate
load is reached, a sudden drop in the load to a residual value is observed where the displacement
increases and the load remains constant.
9 Strain Gage Monitoring for Short and Long-Term Performance
The load-transfer mechanism along the bonded length was evaluated for anchor 31B using strain
gage data. During testing, monitoring was conducted using vibrating wire strain gages along the
bonded length of the anchors. Gauges were installed at 0.9 m, 2.1 m and 3.4 m from the top of the
bonded length as shown in Fig. 18. Changes in deformation were observed during performance
tests and after the tests for a period of 185 days considered herein as the long-term analysis. To
Can. Geotech. J.
protect the strain gages during grout installation, the gages were mounted to the strand tendon and
covered with a PVC grease tube, with both ends of the tube wrapped with waterproof tape.
Strains measured along the bonded length of anchor 31B during the performance test are presented
in Fig. 19a. The figure also includes the cracking strain of the concrete grout that was assumed to
occur around 100 microstrains (Neville 1996). Initially, under small loads tensile strains in the
grout and the steel strands were compatible. After increasing the load, the tensile strains in the
steel strands exceeded the cracking strain of the grout, as shown in Fig. 19a. As soon as the
cracking strain was surpassed, debonding between the strand tendon and the grout was observed
with the negative strains measured along the bond length. Similar behavior has been observed by
other researchers (Benmokrane et al. 1995; Weerasinghe and Littlejohn 1997; Krothapalli 2013)
From the strains measured in the bonded length, axial load was estimated using the axial stiffness
of the strands and assuming that the measured strain was equal for all the strands of the anchor.
Fig. 19b shows the estimated loads along the bonded length. The load was first registered in the
top gage; then with increasing applied load, some tensile load was observed in the second gage
installed at 2.1 m from the top of the bonded length. Load variation observed in these gages was
similar during the test; initially, it was observed tension and later compression. However, the
16
Page 17 of 44
estimated loads in the lowermost gage were almost zero, suggesting that the applied load was
distributed, at most, in the top 3.4 m of the bonded length. Similar behavior has been previously
reported by other authors (Ostermayer 1978; Ludwig 1984). The strain distribution at the end of
the test confirmed that the bonded length was adequate for the anchors installed in the shale stratum
and slippage did not occur at the lowest gage location.
Long-term performance of ground anchors must be considered to ensure the structural stability of
the retention system (Benmokrane and Ballivy 1991). Long-term monitoring of strains in the
bonded length after locking-off is presented in Fig. 19c. From the figure, it is observed that
between Construction Days 155 and 190, there was a gradual redistribution of the load along the
bonded length with changes observed in all the strain gages. This redistribution corresponds to the
backfilling of the upper wall. Gages installed at 0.9 m and 2.1 m from the top of the bonded length
showed a decrease in the measured strains, meaning that more relative displacement between the
tendons and the grout occurred during this period. Also, strain measured in the gage located at
3.4 m from the top of the bonded length slightly surpassed the assumed cracking strain of the grout.
Can. Geotech. J.
However, as seen from the Fig. 19c the strain remains positive indicating that deboning at the
strand/grout interface had not occurred at this location.
It was observed that after 40 days of the anchors testing, about the time when construction of the
walls was approximately finished, the anchors reached a state of equilibrium. As shown in Fig.
19c the measured strains remained constant after the construction of both walls was finished.
10 Conclusions
The use of t-z approach and the two presented softening models was a viable means of predicting
the load-displacement response for the anchors installed in this project. Unfortunately, because the
anchors were not tested to failure, the predicted response at ultimate load conditions could not be
fully evaluated for the presented methods.
Although the displacement measured at the anchor head during a performance or proof test in
anchors is mostly due to the elongation of the unbonded length, the presented softening models
tend to show that they can be used to estimate the post-peak response of ground anchors.
17
Page 18 of 44
Some slippage was observed between the strain gages and the strands. However, the load-transfer
mechanism along the bonded length was identified using strain gages. It was noted that slippage
increased as the anchor head load increased.
After the construction period and during the long-term monitoring, only minor variations in the
strain gage data were observed suggesting that the anchor bonded length reached equilibrium and
that these anchors installed in the shale stratum did not show significant creep susceptibility.
Based on the data collected in the gage installed 3.4 m, it is believed that anchor load was
distributed from the top of the bonded length to this location. This observation confirmed that the
anchor bonded length was satisfactory. Based on the presented approach, the prediction of the
load-displacement curve for an anchor is more influenced by the selected ultimate side frictional
resistance along the anchor-rock interface than the shear stiffness of the interface.
11 Acknowledgments
The authors would like to thank Dr. Robert Liang, currently at the University of Dayton School of
Can. Geotech. J.
Engineering, for providing the SUM 82 Instrumentation and Monitoring Report, which formed the
basis of this study. The authors also would like to thank Mr. Christopher Merklin, of the Office of
Geotechnical Engineering at the Ohio Department of Transportation, for providing the
construction documents and project correspondence for the SUM 82 project. These documents
were of significant value to the analysis.
18
Page 19 of 44
12 References
Benmokrane, B., and Ballivy G. 1991. "Five-year monitoring of load losses on prestressed cement-
grouted rock anchors." Canadian Geotechnical Journal, 28(5), 668-677.
Benmokrane, B., Chekired, M., and Xu, H. 1995. "Monitoring behavior of grouted anchors using
vibrating-wire gauges." Journal of Geotechnical Engineering, 121(6), 466-475.
Blanco, M. L. 2012. "Theoretical and experimental study of fully grouted rockbolts and cablebolts
under axial loads." PhD Thesis, Theoretical and experimental study of fully grouted rockbolts and
cablebolts under axial loads, Paris.
Briaud, J.-L., Powers, W. F., and Weatherby, D. E. 1998. "Should grouted anchors have short
tendon bond length?" Journal of Geotechnical and Geoenvironmental Engineering, 124(2), 110-
119.
Coyle, H. M., and Reese, L. C. 1966. "Load transfer for axially loaded piles in clay." J. Soil Mech.
Found. Div., 92(2), 1-26.
Can. Geotech. J.
Ehrlich, M., and Silva, R. C. 2015. "Behavior of a 31m high excavation supported by anchoring
and nailing in residual soil of gneiss." Engineering Geology, 191(Supplement C), 48-60.
Hegazy, Y. A. 2003. "Reliability of estimated anchor pullout resistance." Third International

Conference on Grouting and Ground Treatment, New Orleans, LA, 10-12 February 2003.
Hsu, S.-C., and Chang, C.-M. 2007. "Pullout performance of vertical anchors in gravel formation."
Engineering Geology, 90(1), 17-29.
PTI 2014. Recommendations for Prestressed Rock and Soil Anchors, Post-Tensioning Institute,
Phoenix, AZ.
Kim, N.-K. 2003. "Performance of tension and compression anchors in weathered soil." Journal
of Geotechnical and Geoenvironmental Engineering, 129(12), 1138-1150.
Kim, N.-K., Park, J.-S., and Kim, S.-K. 2007. "Numerical simulation of ground anchors."
Computers and Geotechnics, 34(6), 498-507.
Knellwolf, C., Peron, H., and Laloui, L. 2011. "Geotechnical Analysis of Heat Exchanger Piles."
Journal of Geotechnical and Geoenvironmental Engineering, 137(10), 890-902.
19
Page 20 of 44
Krothapalli, G. 2013. "Load transfer across pre-stressed tieback anchors grouted in kope bedrock
formation." MS Thesis, University of Cincinnati.
Liang, R. Y. 2000. "Instrumentation and monitoring of tieback wall on SUM82 at Brecksville."

Rep. FHWA/OH-2000/015, the University of Akron, Akron, OH.
Littlejohn, G. S. 1997. “Ground Anchorages and Anchored Structures.” Proceedings of the

International Conference organized by the Institution of Civil Engineers, London, UK, 20-21
March 1997, Thomas Telford.
Liu, X., Wang, J., Huang, J., and Jiang, H. 2017. "Full-scale pullout tests and analyses of ground
anchors in rocks under ultimate load conditions." Engineering Geology, 228, 1-10.
Ludwig, H. 1984. "Short-term and long-term behavior of tiebacks anchored in clay." PhD Thesis,
McGill University.
Neville, A. M. 1996. Properties of concrete, 4th ed, Wesley Longman Ltd, Harlow, England.
Ni, P., Song, L., Mei, G., and Zhao, Y. 2017. "Generalized nonlinear softening load-transfer model
Can. Geotech. J.
for axially loaded piles." International Journal of Geomechanics, 17(8), 04017019.
Ostermayer, H., and Scheele, F. 1978. "‘Research on ground anchors in non-cohesive soils." Rev.
Francaise Geotech., (3), 92-99.
Ren, F. F., Yang, Z. J., Chen, J. F., and Chen, W. W. 2010. "An analytical analysis of the full-
range behaviour of grouted rockbolts based on a tri-linear bond-slip model." Construction and
Building Materials, 24(3), 361-370.
Sabatini, P. J., Pass, D. G., and Bachus, R. C. 1999. "Geotechnical Engineering Circular No. 4:
Ground Anchors and Anchored Systems." Rep. FHWA-IF-99-015,Atlanta, GA.
Weerasinghe, R., and Littlejohn, G. 1997. "Load transfer and failure of anchorages in weak
mudstone." Proc., Proc. of Int. Conf. on ground anchorages and anchored structures, London, UK,
34-44.
Xanthakos, P. P. 1991. Ground anchors and anchored structures, Wiley, Washington, DC.
Xu, D.-S., and Yin, J.-H. 2016. "Analysis of excavation-induced stress distributions of GFRP
anchors in a soil slope using distributed fiber optic sensors." Engineering Geology, 213
(Supplement C), 55-63.
20
Page 21 of 44
Xu, H.-y., Chen, L.-z., and Deng, J.-l. 2014. "Uplift tests of jet mixing anchor pile." Soils and
Foundations, 54(2), 168-175.
Zhang, Q.-Q., and Zhang, Z.-M. 2012. "A simplified nonlinear approach for single pile settlement
analysis." Canadian Geotechnical Journal, 49(11), 1256-1266.

Can. Geotech. J.
21
Page 22 of 44
List of Figures.
Fig. 1. Site location map

Fig. 2. Soil profile at project site
Fig. 3. Composite soil test data at project site (from multiple boreholes)
Fig. 4. Apparent earth pressure acting on the wall

Fig. 5. Typical anchor configurations for the upper wall: (a) preproduction anchor; (b) production
anchor.
Fig. 6. Preproduction test results: (a) PT1; (b) PT2
Fig. 7. Application of the failure criterion: (a) Failure Criterion 1 for PT1 and components of
anchor movement until 120 percent of design load; (b) Failure Criterion 1 for PT2; (c) Failure
Criterion 2 for PT1 in which failure is defined the load at which the total displacement is (D/10)
plus the elastic elongation of the anchor unbonded length; (d) Failure Criterion 2 for PT2
Fig. 8. Suggested ultimate load range and pullout tests: (a) A1; (b) B1 and (C1)
Fig. 9. Performance tests in anchors: (a) 30B; (b) 30C; (c) 31B
Fig. 10. Residual movement versus load
Fig. 11. Analytical model of ground anchors: (a) ground anchor; (b) stress conditions on a
differential element
Fig. 12. Load transfer models considered: (1) Zhang and Zhang (2012) load transfer model; (2) Ni
et al. (2017) load transfer model
Fig. 13. Flow chart describing the t-z approach algorithm
Fig. 14. Analysis of the preproduction tests: (a) load transfer models used in the analysis; (b)
Can. Geotech. J.
measured and calculated response for the PT2

Fig. 15. Load-displacement response for performance test in anchors: (a) 30B; (b) 30C (c) 31B
Fig. 16. Model 1 (Zhang and Zhang, 2012) Effect of: (a)  on the load transfer curve; (b)  on the
load displacement response; (c) zu on the load transfer curve; (d) zu on the load displacement
response
Fig. 17. Model 2 (Ni et al., 2017) Effect of n on the load-displacement response (a) load transfer
curve; (b) load displacement response
Fig. 18. Locations of the vibrating wire strain gauges in bonded length for anchor 31B
Fig. 19. Load-transfer mechanism along the bonded length for anchor 31B: (a) bonded length
deformation of during performance test; (b) estimated load during performance test; (c) long-term
monitoring of anchor bonded length
22
Page 23 of 44
Table 1. General installation sequence.
Day Construction activity

0 Establishment of upper and lower benches.
20 Installation of earth inclinometer and piezometers.
27 Installation of solider piles for upper wall was started.

42 Installation of solider piles for lower wall was started.
58 Upper preproduction test.
62 Lower preproduction test.
80 Installation of anchors in row B in upper wall was started.
99 Installation of anchors in row B in lower wall was started.
105 Stressing of the anchors in row B of the upper wall was finished.
110 Stressing of the anchors in row B of the lower wall was finished.
121 Installation of anchors in row C in upper wall was started.
132 Installation of anchors in row A in lower wall was started.
135 Stressing of anchors in row C in upper wall was finished.
145 Stressing of anchors in row A in lower wall was finished.
154 Installation of anchors in row A in upper wall was started.
160 Concrete lagging installation in lower wall was completed.
Can. Geotech. J.
160 Backfilling of the lower wall was started.

176 Stressing of anchors in row A in upper wall was finished.
181 Backfilling of the upper wall was started.
185 Backfilling of the lower wall was completed.
Backfilling of the upper wall, post-grouting and construction were
188 completed.
226 Finish wiring and set up for long-term monitoring.
1
Page 24 of 44
Table 2. Anchor characteristics.
Design Total Bonded

Load Diameter Tendon length length Unbonded Inclination
Test (kN) (mm) assembly3 (m) (m) length (m) (Deg)
343
PT11 101.6 7 @ 15.24 mm 18.62 4.57 14.05 34

PT2 489 101.6 7 @ 15.24 mm 12.89 4.57 8.32 15
30B 343 101.6 3 @ 15.24 mm 10.06 4.57 5.492 45
30C 343 101.6 3 @ 15.24 mm 9.14 4.57 4.572 45
31B 343 101.6 3 @ 15.24 mm 10.06 4.57 5.492 45
1Steel casing. 2Minimum value. 3number of tendons@diameter
Can. Geotech. J.
2
Can. Geotech. J.
Downloaded from www.nrcresearchpress.com by UPPSALA UNIVERSITETSBIBLIOTEK on 08/24/19. For personal use only. Page 25 of 44
n
u
s
Es
Eg
wu
3
0.7
4.12
23GPa
2.5 mm
200 GPa
1400 kPa
Table 3. Input parameters for the analysis of the production test.
Page 26 of 44
222
N 219
219
0 10 m
Bridge SR. 82
216
Bridge pier
216
Boring location
Test Location 213
Soldier Pile
213
Upper
Wall 30 31 210
27
210 207
207
PT1
204 204
PT2 201
201
7 Lower 198
198 Wall
195
195
Railroad
Can. Geotech. J.
Fig. 1. Site location map

Page 27 of 44
EL. (m)
216
Stiff-Med Clayey
Silt (Brown)
213 212.8
45°
210 Upper
Stiff, Clayey Silt Wall
and Silty Clay
207 (Gray)
Existing Grade
Bridge Pier
204 203.3
15° Lower
201 Wall
198 197.5
Railroad
195
Shale
192 Rock
channel
189 188.9
186
Can. Geotech. J.
Fig. 2. Soil profile at project site

Page 28 of 44
Natural Moisture Content (%) Blow Counts (bpf) Shear Strength (kPa)
Elev. (m) 0 10 20 30 40 50 60 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 80
218 218 218 218
216 216 216 216

Stiff-Med Clayey
214 Silt 214 214 214
212 212 212 212
210 210 210 210

Stiff, Clayey Silt
208 and Silty Clay 208 208 208
206 206 206 206
204 204 Moisture content (w) 204 204 Unconfined

(a) Plastic Limit (b) (c) Compression Test
Shale Liquid Limit
202
202 202 202
Fig. 3. Composite soil test data at project site (from multiple boreholes)
Can. Geotech. J.
Page 29 of 44
Failure plane HP 14x73

Pa
ω
Hexc/5
φ' Lunb(min)
Lbonded Hexc
α
Ground Pa=0.65 KaγH
anchor
Embedded
length
Can. Geotech. J.
Fig. 4. Apparent earth pressure acting on the wall

(a)
Detail A Spacer
AA BB Centralizer
AA BB
Unbonded Length Bonded Length
3 12" Sch 40 Steel Trumpet 1 " Centralizer

Anchor Head Cap Pipe 30" Long - Galvanized Polystrand 2
Corrugated 15.24 mm Dia. Bare
Galvanized N tumpet
Can. Geotech. J.
Sheath Strand
Seals
Multistrand
Tendon
Corrosion
Inhibitor Filled Strand
Anchor Head
1
2" PSI Grout Tube Spacer
(b) Detail A Section A-A Section B-B
Fig. 5. Typical anchor configurations for the upper wall: (a) preproduction anchor; (b) production
anchor.
Page 31 of 44
Load (kN) Load (kN)

0 200 400 600 800 1000 0 200 400 600 800 1000 1200 1400 1600
0 0
10
20
20
Displacement (mm)
Displacement (mm)
40
30
60 40
50
80
60
100
70
(a) (b)
120 80
Fig. 6. Preproduction test results: (a) PT1; (b) PT2

Can. Geotech. J.
Page 32 of 44
Load (kN) Load (kN)

0 100 200 300 400 500 0 100 200 300 400 500 600 700
0 0
10 5
10
20
Displacement (mm)
Displacement (mm)
15
30
20
40
25
50 Total movement Total movement
Elastic movement 30 Elastic movement
1.2DL=411 kN 1.2DL=587 kN
60 Residual movement 35 Residual movement
Failure Criterion 1 (a) Failure Criterion 1 (b)
70 40
Load (kN) Load (kN)
0 200 400 600 800 1000 0 500 1000 1500
0 0
DL=343 kN (c) DL=489 kN (d)
10
20
20
Displacement (mm)
Displacement (mm)
40
30
Can. Geotech. J.
60 40
50
80
60
100 PT1 PT2
Failure Criterion 2 70 Failure Criterion 2
120 80
Fig. 7. Application of the failure criterion: (a) Failure Criterion 1 for PT1 and components of
anchor movement until 120 percent of design load; (b) Failure Criterion 1 for PT2; (c) Failure
Criterion 2 for PT1 in which failure is defined the load at which the total displacement is (D/10)
plus the elastic elongation of the anchor unbonded length; (d) Failure Criterion 2 for PT2
Page 33 of 44
P/Pmax
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.5
Suggested
failure range
1.0
d/D
1.5
A1
2.0 Liu et al. (2017)
Briaud et al. (1998)
(a) Suggested range
2.5
(a)
P/Pmax
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.5
Suggested
1.0 failure range
Can. Geotech. J.
d/D
1.5
2.0
B1
Liu et al. (2017)
2.5 Briaud et al. (1998)
Suggested range (b)
3.0
(b)
P/Pmax
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.5
Suggested
1.0 failure range
d/D
1.5
C1
2.0 Liu et al. (2017)
(c) Briaud et al. (1998)
Suggested range
2.5
(c)
Fig. 8. Suggested ultimate load range and pullout tests: (a) A1; (b) B1 and (C1)
Page 34 of 44
Load (kN)
0 100 200 300 400 500
0
10
30B
Displacement (mm)
20
30
40
50
60
70
Load (kN)
0 100 200 300 400 500
0
10
30C
Displacement (mm)
20
Can. Geotech. J.
30
40
50
60
70
Load (kN)
0 100 200 300 400 500
0
10
31B
Displacement (mm)
20
30
40
50
60
70
Fig. 9. Performance tests in anchors: (a) 30B; (b) 30C; (c) 31B
Can. Geotech. J.
Residual Movement (mm)
0
1
2
3
4
5
6
7
8
9
10
0
100
Fig. 10. Residual movement versus load

31B
30C
30B
200
Load (kN)
400
500300
Page 36 of 44
Unbonded length Bonded length
Grout
Steel Tendon
dx
(x) dx
x
(a)
(x)
P(x)+dP(x) P(x)
dx
(b)
Fig. 11. Analytical model of ground anchors: (a) ground anchor; (b) stress conditions on a
differential element
Can. Geotech. J.
Page 37 of 44

u (1)
(1)
r
(2)
(2)
wu 2wu w
Fig. 12. Load transfer models considered: (1) Zhang and Zhang (2012) load transfer model; (2) Ni
et al. (2017) load transfer model
Can. Geotech. J.
Page 38 of 44
Start 1 2
1. Divide the bonded 7. Set displacement and the load at the anchor
length into n segment of load at the bottom of head will be
length dl segment n-1 equal to the Pt1
Ptop  (20)
displacement and load at (1- )

2. Assume a small displacement the top of segment n
12. Repeat steps 1-11
at the bottom of the anchor wbn
with other displacement
Axial load is assumed zero at the 8. Repeat step 3-6 until at the bottom of the
bottom segment the top segment, anchor
segment 1, of the bonded
3. Determine the displacement length.
at the middle of the segment n End
from Eq.(14) 9. Use the load on top of
segment 1 to calculate the
elastic elongation of the
4. Calculate the load at the
unbonded length as:
Ptn
middle of the segment as:
dl D Pt1Lunb
Pmn  Pbn +
2
 (wbn ) (15) wunb 
AsEs
(18)
5. The load at the top of the 10. Calculate the total  Pmn dl
segment is given by: displacement of the anchor
as:
Ptn  Pbn +2 ( Pmn - Pbn ) (16)
wtot  wunb + wt1 (19)
Can. Geotech. J.
6. Calculate the top

11. Considering a
Pbn
displacements of the segment
as: percentage of load loss,  ,
Pmn dl along the unbonded length
wtn  wbn + (17)
AE
1 2
Fig. 13. Flow chart describing the t-z approach algorithm

Page 39 of 44
Load (kN)
0 400 800 1200 1600 2000
0
20
Can. Geotech. J.
40
Displacement (mm)
60
80 Softening
response
100
120
140
Model 1
160 Model 2
180 PT2
200
Fig. 14. Analysis of the preproduction tests: (a) load transfer models used in the analysis; (b)
measured and calculated response for the PT2
Page 40 of 44
Load (kN)
0 100 200 300 400 500
0
10
20
Displacement (mm)
30
40
50
Measured
60 Model 1
Model 2 (a)
70
Load (kN)
0 100 200 300 400 500
0
5
10
Displacement (mm)
15
20
25
30
Can. Geotech. J.
35
40 Measured
Model 1
45
Model 2 (b)
50
Load (kN)
0 100 200 300 400 500
0
10
Displacement (mm)
20
30
40
50
Measured
60 Model 1
Model 2
70
Fig. 15. Load-displacement response for performance test in anchors: (a) 30B; (b) 30C (c) 31B
Page 41 of 44
3000
1.0
2500
0.8 2000
Load (kN)
1500
τ/τuu
τ/τ
0.5
β = 0.7 1000
β = 0.7
0.3 β = 0.8 β = 0.8
(a) β = 0.9 500 β = 0.9
(a) β = 0.95 (b)
(b) β = 0.95
0.0 0
0 1 2 3 4 0 50 100 150 200 250 300 350 400
w/wu Displacement (mm)
3000
1.0
2500
0.8 2000
Load (kN)
1500
τ/τu
Can. Geotech. J.
0.5
zu = 1 mm 1000 zu = 1 mm
0.3 zu = 2 mm zu = 2 mm
zu = 2.5 mm 500 zu= 2.5 mm
(c) zu= 3 mm (d)
(b) zu= 3 mm
(a)
0.0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 0 50 100 150 200 250 300 350 400
w/w
w/wuu Displacement (mm)
Fig. 16. Model 1 (Zhang and Zhang, 2012) Effect of: (a) β on the load transfer curve; (b) β on
the load displacement response; (c) zu on the load transfer curve; (d) zu on the load displacement
response
Page 42 of 44
3000
1.0
2500
0.8 2000
Load (kN)
1500
/u
u

0.5
1000
n = 0.5 n = 0.5
0.3 n = 1 n = 1
n = 4 500 n = 4
n = 8 n = 8
0.0 0
0 1 2 3 4 0 50 100 150 200 250 300 350 400
w/wu Displacement (mm)
Fig. 17. Model 2 (Ni et al., 2017) Effect of n on the load-displacement response (a) load transfer
curve; (b) load displacement response
Can. Geotech. J.
Page 43 of 44
Steel Strand
Bonded length
4.6 m
0.9 m
2.1 m
3.4 m
Strain
Gage
Grout
Fig. 18. Locations of the vibrating wire strain gauges in bonded length for anchor 31B
Can. Geotech. J.
Page 44 of 44
4000
0.9 m
Gage location 2.1 m
3000 3.4 m
εcrack
2000
Strain (µε)
1000
-1000
(a)
-2000
0 50 100 150 200 250 300
Time (min)
150
0 min
50 min
100 100 min
150 min
50 200 min
Load (kN)
0
Can. Geotech. J.
-50
-100
(b)
-150
0 1 2 3 4
Distance from top of bonded length (m)
1000
0.9 m
750 Gage location 2.1 m
500 3.4 m
εcrack
250
0
-250
Strain (µε)
-500
-750
-1000
-1250
-1500
-1750 (c)
-2000
155 180 205 230 255 280 305 330 355
Construction Day
Fig. 19. Load-transfer mechanism along the bonded length for Anchor 31B: (a) bonded length
deformation of during performance test; (b) estimated load during performance test; (c) long-term
monitoring of anchor bonded length

CGJ 2018 0326

Uploaded by

Copyright:

Available Formats

You might also like

CGJ 2018 0326

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CGJ 2018 0326

Uploaded by

Copyright:

Available Formats

Page 1 of 44

Analysis of a Case Study Presenting Ground Anchor Load Transfer

1Hardin-Drnevich-Huang Associate Professor, Department of Civil Engineering, University of

2 Project Description and Site Conditions

used to determine the anchor loads.

the design and is given as

4 Anchors and Wall Installation Procedures

5 Load Test Results

5.1 Preproduction Tests

5.2 Performance Test

Considering equilibrium of the segment of length dx shown in Fig. 11

6.1 Softening Model 1 (Zhang and Zhang 2012)

The three parameters a, b, and c are expressed as

6.2 Softening Model 2 (Ni et al. 2017)

range of 0.65-0.90, the parameter n used in Equation (9) can be approximated as

n  0.16  9.11 (10)

7 Algorithm for the Load Transfer Approach

wmn  wbn   en (11)

Equations (12) and (13) into (11) gives:

8 Analysis and Results

8.1 Analysis of Preproduction Test Data

8.2 Analysis of Performance Test Data

9 Strain Gage Monitoring for Short and Long-Term Performance

and slippage did not occur at the lowest gage location.

Hegazy, Y. A. 2003. "Reliability of estimated anchor pullout resistance." Third International

Liang, R. Y. 2000. "Instrumentation and monitoring of tieback wall on SUM82 at Brecksville."

Rep. FHWA/OH-2000/015, the University of Akron, Akron, OH.

Littlejohn, G. S. 1997. “Ground Anchorages and Anchored Structures.” Proceedings of the

for axially loaded piles." International Journal of Geomechanics, 17(8), 04017019.

analysis." Canadian Geotechnical Journal, 49(11), 1256-1266.

Fig. 1. Site location map

Fig. 4. Apparent earth pressure acting on the wall

measured and calculated response for the PT2

Table 1. General installation sequence.

Day Construction activity

27 Installation of solider piles for upper wall was started.

160 Backfilling of the lower wall was started.

Table 2. Anchor characteristics.

Design Total Bonded

PT11 101.6 7 @ 15.24 mm 18.62 4.57 14.05 34

Fig. 1. Site location map

Fig. 2. Soil profile at project site

216 216 216 216

212 212 212 212

210 210 210 210

206 206 206 206

204 204 Moisture content (w) 204 204 Unconfined

Failure plane HP 14x73

Fig. 4. Apparent earth pressure acting on the wall

3 12" Sch 40 Steel Trumpet 1 " Centralizer

(b) Detail A Section A-A Section B-B

Load (kN) Load (kN)

Fig. 6. Preproduction test results: (a) PT1; (b) PT2

Load (kN) Load (kN)

Residual Movement (mm)

Fig. 10. Residual movement versus load

Unbonded length Bonded length

displacement and load at (1- )