Journal of JSCE, Vol.

4, 106-117, 2016
(Originally published in Journal of Japan Society of Civil Engineers, Ser. C, Vol. 67, No. 4, 558-568, 2011 in Japanese)

A STUDY OF A LIFT-OFF TEST METHOD

FOR GROUND ANCHORS

Yu FUJIWARA1 and Toshinori SAKAI2

1Member of JSCE, Chief, Geotechnics Division Road Research Department,
Nippon Expressway Research Institute Company Limited
(1-6-20 Tadao, Machida-shi, Tokyo 194-8508, Japan)
E-mail: y.fujiwara.ae@w-nexco.co.jp
2Member of JSCE, Graduate School of Bioresources, Mie University
(1577, Kurimamachiya, Tsu, Mie 514-8507, Japan)
E-mail: sakai@bio.mie-u.ac.jp

The ground anchoring method is a prevention work for maintaining the stability of slopes. Periodic in-
spections of ground anchor (hereinafter, anchor) are important to ensure slope stability. The residual tensile
load of anchor is confirmed by lift-off tests. However, a unified method of the lift-off test has not been es-
tablished so far. A concern is that the measured value may vary depending on the engineer carrying out the
work. It is also possible that proper maintenance is not being carried out at present. In this paper, we will
report on our studies of lift-off test methods. The lift-off tests were performed on working anchorsinstalled
in the cut slopes of an expressway. Lift-off tests varying the displacement positions, loading methods, and
other conditions were performed. At the end of our study, a lift-off test method will be proposed based on
our findings.

Key Words : ground anchor, maintenance, cut slope, lift-off test, residual displacement, hydraulic jack

1. INTRODUCTION nized jumping and lifting of the anchor head caused

The ground anchoring method was introduced in
Japan in 1957. Construction materials for ground
anchors (hereinafter, anchor) are easy to procure and
construction is also simple compared with the pre-
ventive pile method. Many anchors are constructed
every year as a way to stabilize cut slopes and pre-
vent landslides. According to a survey by the Japan
Anchor Association, in 2005, a total of 2,287 anc-
hors, with a total extension of 1,670 km, had been
constructed.
The anchors are constructed in natural ground.
Therefore, hypofunction caused by a variety of fac-
tors, such as the corrosion of steel and the effects of
earth pressure is of primary concern. Initially, ground
anchoring had been adopted as a temporary preven-
tion method. Still, protection against corrosion had
not been sufficient, particularly on anchor heads
boundary of tendon free length, and anchor fixed
length. The ingress of water and air also contribute to
the corrosion of anchors made of steel. By the
first-half of the 1980s, anchor failures caused by
corrosion had been seen. For example, we recog-
by rupture of the tendon free length. For these rea-
sons, in 1988, standards established by the Japanese
Society of Soil Mechanics and Foundation Engi-
neering (currently Japanese Geotechnical Society)
were revised1). And, the anchors protected against
corrosion started to be adopted2).
When constructing expressways, in some cases
we are forced to cut and fill weak ground owing to
the constraints of the construction site. As a result,
anchoring is often used to stabilize the steep slope of
the site.
Figure 1 shows the number of anchors con-
structed on expressways. The anchoring method was
adopted in 1969 and the number increased gradually
from around 1985. The present number of anchors
constructed is over 120,0003).
Of these anchors, there are those that have been in
place for about 40 years since the construction, and
some anchors are not able to maintain their original
function. Therefore, maintaining the anchors is an
important issue. Currently, the health levels of anc-
hors on expressways are checked periodically by tap
tone and visual observation4). However, there have

106
 start service time elapsed
（in service）

total extension of expressway-

（in construction）

120 tatal number of anchor 8

regular measurement
construction (☓1,000N)
total number of anchor

constructi on

service (☓1,000km)
100 total extension of
expressway-servi ce 6 anchor load measurement load measurement load measurement
80 construction （load cell） （lift-off test） （lift-off test）

60 4 （load cell set） （Function of load cell stops）

40
2
20
0 0
Ｓ60
1985Ｓ44
1969 Ｈ5
1993 Ｓ52Ｈ13
1977
2001
Fig.1 Construction result of ground anchor on expressway1).

(load cell） （lift-off test）

unknown PC steel rod
(14%) (8%)
SFL
(5%)
FLO
(2%)
SEEE
(12%)
KTB
(12%)
EHD
(2%)
Fig.2 Breakdown of anchoring method on expressway.
been no actual incidence of slope failures caused by
degradation of anchors. Moreover, the inspection of
anchors is difficult to carry out. Therefore, preven-
tive maintenance is not performed sufficiently owing
to many factors. However, the importance of main-
taining anchors is expected to increase in the future,
as the hypofunction of anchors is thought to proceed
depending on the conditions of the construction and
ground, leading to destabilization of the ground.
In addition, the tension of anchor is not constant in
service. In some cases, they increase or decrease, by
ground deformation, tendon creep, relaxation, and
changes in the climate. The residual tensile load of
anchor is believed to function as a sensor that shows
ground condition and material condition. Therefore,
it is important to properly measure the residual ten-
sile load of anchor during maintenance. It is also
important to confirm the change in residual tensile
load of anchors. There are many types of anchoring
methods, as anchors have been improved based on
experiences on the field. The fixed-type anchor
heads are classified into wedge-fixed type and nut-
fixed type. The wedge-nut-fixed type is a kind of
wedge fix type that combines wedge and nut.
Figure 2 shows a breakdown of anchoring me-
thods constructed on expressways until 2006. As can
wedge fixed Fig.3 Management of residual tensile load of anchor
nut fixed on expressway5).
wedge-nut fixed
be seen, the VSL method of wedge-fixed type is the
most frequently built, followed by the
wedge-nut-fixed type KTB method, and the SEEE
VSL method for the nut-fixed type.
(36%) The design standards for expressways were es-
tablished in 19925). These standards had been based
on the standards of the Japanese Society of Soil
other(1%) Mechanics and Foundation Engineering, set in 1988.
SHS
(8%) Figure 3 shows the general flow for managing the
residual tensile load of anchors on expressways6). In
the construction phase, the destabilization of slopes by
cutting the ground is of primary concern, and therefore,
the construction is typically performed while checking
the changes in the residual tensile load by setting load
cells to some of the anchors. After construction, the
residual tensile load is managed by the load cells in-
stalled at the construction stage. But, in some cases
accurate data cannot be obtained because of hypo-
function caused by changes over time and loadcells
being exposed to the natural environment7). Cur-
rently, there are replaceable loadcells that can be
installed on new anchors or anchors with sufficient
extra tendon length in the anchor head. However, it is
difficult to exchange or set the loadcell on other
working anchors. Therefore, residual tensile load is
usually measured by lift-off tests. The lift-off test is
performed by setting a hydraulic jack on the anchor
head8). The methods of lift-off tests that are applied
differ depending on the anchor head used. However,
until now, sufficient verifications on how the dif-
ferent test methods affect the results have not been
performed. Therefore, it is considered important to
reveal the properties of tests and the factors of data
variability. In addition, these findings are considered
important for maintaining anchors.
In this study we carried out lift-off tests on anchors
installed in cut slopes of the expressways:
wedge-type anchors, nut-type anchors, and

107
 to hydraulic pump load P
loading point D
load P tension bar process
(Changing point
of the slope)

bearing plate
point E
hydraulic jack
(centerhole jack)
unloading point B
point C
ramuchea remaining load P e p rocess
temporary pooling head
anchor head
point A
tendon
point O point F

displacement displacement
residual displacement

Fig.4 Lift-off test by using center hole jack. Fig.6 Load-displacement curve in lift-off test.

Pe
P=0 applying load. Figure 4 shows the lift-off test for a
wedge-type anchor. First, the temporary pooling
( I. pull load is not applied )
head is set. The temporary pooling head has a func-
tion that fixes the extra tendon length of anchor head
P Pe and the tension bar. The lift-off test is then carried
P<Pe
unloadind process

out by applying load on the tension bar.

loadind process

( II. before lift-off ) In the loading process, after the residual tensile
load is confirmed, unloading is performed. Figure 5
shows the anchor in the loading and unloading
P Pe
P=Pe process of the lift-off test. In addition, Fig.6 shows a
typical load-displacement curve in a lift-off test.
( III. pull load = residual tensile load ) Before the anchor head starts to lift away from the
bearing plate, the displacement of anchor head is small.
P Pe The linear slope of load-displacement curve shows a
P>P e steep slope (Fig.5; II, Fig.6; O-A). Then, as the load
( IV. after lift-off )
on the hydraulic jack increases, the load of the hy-
draulic jack increases and equals the residual tensile
(P : Loading load of hydraulic jack, Pe : Residual tensile load) load. This is when the anchor head starts to lift away
Fig.5 Loading and unloading process of lift-off test. from the bearing plate (Fig.5; III, Fig.6; point A).
This phenomenon is generally called "lift-off." The
wedge-nut-type anchors. The tests showed the ap- load measured at lift-off in the load-displacement
propriate method to evaluate the residual tensile load curve is defined as a "lift-off value." That load is
and cycles to carry out lift-off tests for each anchor evaluated as the residual tensile load of the anchor.
type. In addition, the tests revealed the differences in With the increase of the load after lift-off, the linear
the test results depending on where displacement slope of O-A begins to gradually change (Fig.5; IV,
was measured on the anchors, test equipment, and Fig.6; A-B). When that change becomes constant,
how loads are applied on anchors. In addition, the load is transferred to the free anchor length. At
Chapter 2 shows the typical method of lift-off tests. that time, the load-displacement curve shows a gra-
Chapter 3 explains investigation and test methods, dual linear slope that depends on the elastic modulus
and Chapter 4 and Chapter 5 give test results and of the anchor (Fig.6; IV, Fig.6; B- D).
discussion. Finally, this study proposes a lift-off test After that, loading is continued until the residual
method based on its findings. tensile load of the anchor can be evaluated. After the
residual tensile load is confirmed, unloading is per-
formed. (Fig.6; D-E-F). After one cycle from loading
2. LIFT-OFF TEST to unloading is completed, residual displacement
sometimes occurs. This displacement difference of
The lift-off test is performed by setting a hydraulic anchor head is caused by lift-off tests. (Fig.6; O-F).
jack to the extra tendon length of the anchor head and

108
 Table 1 Test conditions*1
(unit : number)
temporary scattering chain block method
protect fence VSL SEEE SFL KTB FLO
test condition
Control the displacement mea-
panel
displacement hydraulic jack surement positions are 2 2 - - -
meter different
the test device is differ-
4 - - - -
tripod ent*2
pressure sensor the loading speed of pull
to generator 17 12 4 4 4
load is varied
the method of loading
single tube pipe hold of pull load is dif- 10 2 - - -
ramuchea data logger hydraulic unit
ferent
*1) There is a case that have tested multiple conditions in a
Fig.7 Mechanical arrangement in lift-off test.
single anchor.
*2) We compared a center hole jack and special jack.

point a point b

Fig.8 Lift-off test.

3. INVESTIGATION AND TEST

(1) Test equipment and material

Figure 7 shows an example of a typical
arrangement of the lift-off test used in this study.
In addition, Fig.8 is a photo of a lift-off test showing
how the test was implemented. We hang the hy-
draulic jack using a chain block to prevent the hy-
draulic jack from falling if the anchor breaks. The
anchor may jump during a lift-off test. Therefore, to
prevent the anchor from jumping, a temporary pro-
tective fence made of single tube pipes and control
panels was set. We mainly used the rumuchea and
the hydraulic jack of 600kN–1,000kN, depending on
the design of the anchor. To confirm the effects dif-
ferent jacks have on anchors, we used a special jack
of 700kN. The dial gauge-type displacement trans-
ducer of range 50mm from 10mm was used as the
displacement meter. The equipment was fixed by a
tripod to prevent disturbance of measured value
caused by fluctuations. In addition, measurements of
the hydraulic jack and displacement meter were si-
multaneously taken using a handheld data logger.
Fig.9 Position of displacement meter.
(2) Validation method
Table 1 shows the test conditions of each anc-
horing method and the number of lift-off tests. We
tested wedge-type anchors constructed using the
VSL method; wedge-nut types using the SFL, KTB,
and FLO methods; and the SEEE method for nut
types. Based on the test data, we studied the appro-
priate cycle to carry out lift-off tests, and how to
evaluate the residual tensile load. In addition, we
analyzed the difference in the results depending on
the position of displacement measurements, test
equipment, and the differences in the effects of dif-
ferent loading methods, etc. For the appropriate
number of times a cycle of lift-off test should be
performed, this study carried out three to five cycles
of lift-off tests on 18 VSL-type anchors and 11
SEEE-type anchors. In addition, this study used a

109

Hydraulic jack
anchorhead
wedge
Fig.10 Special jack.
center hole jack to load and unload in the lift-off test.
The displacement was measured at the tip of the tension
bar (measure point a) as shown in Fig.9.
During loading, we can confirm visually when the
anchor head starts to lift away from the bearing plate.
We can confirm the residual tensile load by measuring
the load of the hydraulic jacks at that time. However, in
confirming the moment of lifting of the anchor head in
the lift-off test, the measured value is considered to vary
depending on the engineer carrying out the test. For this
reason, the Japanese Geotechnical Society9) explains
the residual tensile load as the value measured
(0.1–1.0mm) when the anchoring device starts to lift
away from the bearing plate. Therefore, we can say
that the residual tensile load on lift-off test has not
been clearly defined.
The residual tensile load obtained by the lift-off test
exists in and around the change point of the
load-displacement curve. As shown in Fig.6, three
loads are considered as the residual tensile load. The
load at point A is the load where the change of slope
begins. The load at point B is the load where prevailing
displacement begins. The load at point C is the load of
intersection as shown in Fig.6. There are many cases
where the load at point C is evaluated as the residual
tensile load. The point of slope change in the
load-displacement curve is relatively clear in the lift-off
test for the nut-fixed type because the tendon of the
nut-fixed type is of a cable structure consisting of many
PC steel strands. On the other hand, the point of slope
change in the load-displacement curve is sometimes not
clear for wedge-fixed types, because with this type each
PC-steel strand is an independent cable structure.
110
P(kN)
600 3cycle 2cycle 1cycle
point C2,3
point D3 point D 2 point D1
500
400
pointC1
300
200
VSL(E5-5) / old type
100
residual
Td=513kN
point O displacement point F1
0
0 2 4 6 8 10 12 14 (mm)
point F 2,3
Fig.11 Residual displacement of load-displacement curve.
Therefore, the load is applied to each PC-steel strand.
We defined three points of change for the wedge-fixed
type anchor. Point A was defined as "change point 1,"
Point B as "change point 2," and Point C as "tangent
method." We analyzed three residual tensile loads of the
three points in many test cases, and verified the repro-
ducibility of the test results from the variations of the
measured values.
In the lift-off test, the displacement of the anchor
head and applied load on the hydraulic jack must be
measured with accuracy to get precise data. In order to
do that, the different effects caused by differences in the
displacement measurement position are considered. For
this purpose, as shown in Fig.9, we set a displacement
meter at the tip of the tension bar (measurement point a)
and another at the temporary pooling head (measure-
ment point b) of the center hole jack. Then, we simul-
taneously measured the displacement and confirmed the
influence on the test results caused by the elongation of
the tension bar. If the displacement is measured at the
tip of the tension bar, the influence on the test results
caused by the different tension bar lengths of the
hydraulic jack must be considered. Therefore, this
study compared the test results of the typical center
hole jack with that of the special jack developed for
maintenance (Fig.10).
In the anchor quality test, the planned maximum
load is divided and applied in 10 stages by holding
the applied load for one minute between each stage.
However, the loading method in the lift-off test for
residual tensile load is not defined. Therefore, with
the VSL-type anchor, we analyzed the influence of
loading speed on the load-displacement curve.
Loading was carried out at speeds 20kN/min,
30kN/min, and 60kN/min. In addition, to confirm the
effects of different holding methods, three cases of
lift-off tests were studied. In the first case, the load-
ing speed was set at 20kN/min and the holding time
at test maximum load was set at 10 minutes. In the
second case, the loading speed was set at 20kN/min
and the holding time set as 1 minute at each loading residual displacement (mm)
4.0
of 20kN. In the third case, the loading speed was set 1cycle 2 or more cycles
as 10kN/min and the holding time at test maximum 3.0
load was set at 10 minutes.
2.0
This study suggests a lift-off test method by syn-
thesizing the results of analysis and summarizing the 1.0
findings.
0.0

I-1
I-2
F-1
F-2

J-1
H-1
H-2
A-1
B-1
B-3
C-1
C-2

E-1
E-2
D-1
D-2

G-1
G-2
anchor No
4. TEST RESULTS AND DISCUSSION
(a) Wedge-fixed type (VSL type)

(1) Effects of cycle number of lift-off test residual displacement (mm)

As an example of the test results, Fig.11 shows the 3.0
1cycle 2 or more cycles
test result of VSL (E5-5) where three cycles were
2.0
applied. It has been found out from the design
document that the design anchor load is 513kN. 1.0
However the fixed load at the time of construction is
unknown. In the first cycle of the lift-off tests, the 0.0

L-1

L-2

M-1

M-2

N-1

N-2
K-1

K-2

O-1

displacement increased slightly in the early loading
stage in a constant linear gradient (Fig.11, O-C1).
The lift-off has been identified at around 490kN,
which is about 20kN less than the design anchor load
(Fig.11, point C1). As loading continued, the dis-
placement began to increase in a more gradual linear
gradient than the early loading (Fig.11, C1-D1). Then
when the unloading was performed, the residual dis-
placement of about 4mm was confirmed in the end
(Fig.11, O-F1). The second and third cycles were per-
formed like the first cycle. In early loading, both second
and third cycles confirmed a slight displacement in-
crease caused by linear gradients, which were steeper
than in the first cycle (Fig.11, O-C2,3). In addition,
lift-off was observed at around 490kN, similar to the
first cycle (Fig.11, point C2,3). When further loading
continued, the displacement increased in a more
gradual constant linear gradient than in the early
loading (Fig.11, C2-D2, C3-D3). The linear gradients
after the lift-off for cycle 2 and cycle 3 were almost the
same. When unloaded, the residual displacements in
cycle 2 and cycle 3 were 1mm or less (Fig.11, O-F2,3).
From these results, we can say that the residual ten-
sile load of this anchor exists at about 490kN.
Moreover, we can see that the residual tensile load
has been maintained close to the design anchor load,
although more than 20 years have passed since the
construction.
Figure 12 compares the residual displacements of
VSL anchor and SEEE anchor. Eighteen VSL anc-
hors and 11 SEEE anchors were measured. For most
anchors tested, the residual displacement of the first
cycle is larger than that of the second and subsequent
cycles. The residual displacements of second and
subsequent cycles are 1mm or less.
The effects of adhesion resistance between sheath
and steel material are seen to be small in the first
O-2
-1.0
anchor No
(b) Nut-fixed-type (SEEE type)
Fig.12 Relation of cycle number and residual displacement.
temporary pooling head
wedge
wedge
PC steel
anchorhead
Fig.13 Fix of temporary pooling head.
lift-off test. Also, the residual displacement is con-
sidered to be very small owing to being eliminated
after the second cycles. Further, as shown in Fig.13,
the wedge-type anchors are usually fixed by the
temporary pooling head. Therefore, in the first cycle,
the effects of the bite of the wedge are considered.
On the other hand, the residual displacement of the
nut-fixed-type is smaller than that of the wedge-fixed
type. With the nut-fixed anchor, a temporary pooling
head is set by using the threaded tendon of the anchor
head. The biting of screw into the temporary pooling
head of the nut-fixed anchor is smaller than that of the
wedge-fixed anchor. This is the reason for the differ-
ence in the residual displacements.
From these results obtained through lift-off tests, we

111
 (60 second intervals) (10 second intervals)
P P
ratio of the residual tensile load
of 1-6cycle and 2 cycle (%)
120% Pe(60s) Pe(10s)

115%

110%

105%

100%
Difference in remaining load = Pe(60s) – Pe(10s)
95%
variability ; 94%∼112% Fig.15 Difference in recording interval.
90%
1 2 3 4 5 6 (cycle)
difference of remaining load (kN)
(a) Change point 1
100
ratio of the residual tensile load ● change point a
of 1-6cycle and 2 cycle (%) 80 ● change point b
● tangent method
120%
60
115%
40
110%
20
105%
0
100% A-1 C-1 C-1 D-1 E-1 E-1 F-1 F-1 G-1 H-2 I-1 I-1
anchor No
95%
variability ; 95%∼103% Fig.16 Results of differences in recording interval.
90%
1 2 3 4 5 6 (cycle)

(b) Change point 2
ratio of the residual tensile load
of 1-6cycle and 2 cycle (%)
120%
115%
110%
105%
100%
95%
variability ; 95%∼104%
90%
1 2 3 4
(c) Tangent method
Fig.14 Variations in residual tensile load.
propose that the first cycle test should be considered as
pre-loading. It is also necessary to perform the actual
test after the convergence of the residual displacement
is confirmed in the second and later cycles.
(2) Evaluation of residual tensile load
a) The effects of different measurement methods
This study verified the effects of different mea-
surement methods by performing many lift-off tests.
The lift-off tests we performed were for six cycles.
The first cycle was performed as pre-loading. Cycles
2–6 were performed as the actual tests. We likewise
compared and analyzed the variations in measure-
ment at each point in cycles 3–6 against those in
cycle 2.
Figure 14 shows the ratios of measured values in
cycles 3–6 versus those in cycle 2 for change point 1,
change point 2, and the tangent method. The varia-
tions in cycles 3–6 at change point 1 are as large as
94% to 112%; however, the variations at change
point 2 are 95% to 103%, and the variations in the
tangent method are 95% to 104%, and are smaller
than those at change point 1. Depending on the anc-
hor, at stages when the load is small, which is before
lift-off, there are some cases where the linear gradient of
load-displacement curve starts to change slightly. In
such cases, it is sometimes difficult to capture the
5 6 (cycle) change point in the linear gradient. From these facts, it
is thought that the measured values vary greatly at
change point 1. On the other hand, measurements at
change point 2 are relatively easier to capture than at
change point 1, because the load at change point 2 is the
same load as that of the stage when the change in the
linear gradient of the load-displacement curve becomes
constant. In the tangent method, the variation in the
measured value is small. The reason is that in this me-
thod it is not necessary to measure subtle change points
such as in change point 1 and change point 2, and the
residual tensile load is determined conveniently by the
intersection of a linear gradient before and after lift-off.
b) The effect of the recording interval of data
It is thought that the interval between recordings
of data is a factor in the varied residual tensile loads
measured in the load-displacement curve. If the re-
cording interval of the data is shortened, it is possible
to draw a precise load-displacement curve. However,
the amount of data increases by setting the short re-
cording interval, and the data analysis becomes com-

112
 P(kN)
stop nut hydraulic jack
800

700
Point C2
600
Point C1
500

400

300 tension bar

temporary pooling head anchorhead
200
VSL(E5-12) / oldtype Fig.18 SAAM jack.
100 Td(unknown)
Point O
0
0 1 2 3 4 5 6 (mm) P(kN)

(a) Wedge-fixed type (VSL type) 300

large displacement
250
P(kN) small displacement
300 200 point
C3,4 point C1,2
250 point C1 150
point C2
N-1,2 : centerhole jack
200 100 N-3,4 : special jack

150 VSL(E5-4) / old type

measurement on tension bar 50
Td=359kN
measurement on temporary pooling head point O
100 0
0 1 2 3 4 (mm)
SEEE(F50TA) / new type
50 Td(unknown)
p oint O Fig.19 Difference in loading jack.
0
-2 0 2 4 6 8 10 (mm)

(b) Nut-fixed-type (SEEE type)
Fig.17 Difference in displacement measurement position.
plicated. Therefore, this study compared the varia-
tions caused by changing the recording interval of
the data for change point 1, change point 2, and the
tangent method. In this study, as shown in Fig.15, we
compared the load-displacement curves of the in-
terval recording at 60 seconds and 10 seconds. In
addition, we measured the three residual tensile
loads at each interval. We also measured the differ-
ences in residual tensile load between values meas-
ured for 60-second intervals and 10-second intervals.
Figure 16 shows the differences in residual tensile
load for 12 wedge-fixed anchors. The difference in
residual tensile load was the largest for change point
1, and smallest for the tangent method. The residual
tensile load of anchor is loaded at the time when the
anchor head starts to lift away from the bearing
plate9). It is thought that residual tensile load meas-
ured at change point 1 is the actual residual tensile
load10), 11). However, test results also revealed that the
variation in the measurement at change point 1 is
largest. In addition, in the actual tests at the site, it
was difficult to measure exactly the point when the
anchor head started to lift away from the bearing
plate. If the lift-off test is to be carried out regularly
for the same anchor, reproducibility of the test results
is important. Therefore, it is better to evaluate the
value obtained from the load-displacement curve by
the tangent method as the residual tensile load be-
cause the measured value is more consistent and the
reproducibility is high for the tangent method12). Up
to now, the measured residual tensile load has been
evaluated by the tangent method empirically. This
study, however, showed that it can be evaluated us-
ing actually measured data.
(3) Evaluation of displacement measurement
method
a) The effects of the different displacement
measurement methods
Figures 17(a) and 17(b) show an example of the
results, which confirmed the effects of different
displacement measurement positions. The residual
tensile load is evaluated using the tangent method.
On the VSL anchor, no significant differences in the
observed residual tensile loads of the two setting
displacement meters placed at the tip of tension bar
(Fig.17(a), point C1) and the temporary pooling head
(Fig.17(a), point C2) were found. However, at the tip
of the tension bar (measurement point a) (Fig.17(a),
O-C1), a linear gradient before the lift-off is slightly
more gentle than that at the displacement meter set at
the temporary pooling head (measurement point b)
(Fig.17(a), O-C2). As for the SEEE anchor, there is
also not a large difference in the residual tensile load
between the two. However, at the tip of the tension
bar (measurement point a) (Fig.17(b), O-C1), a linear

113
 P(kN)
800
700 point C2,3
600
東名-2
K-2(30)
500 東名-3
K-3(60)
point C4,5 K-4(20)
東名-4
400 K-5(60)
東名-5
300 ( ) : kN/min

200
VSL(E5-12) / old type
100 Td (unknown)
point O
0
0 1 2 3 4 5 6 7 8 Severe corrosion
(mm)
occurs in tendon.
(a) Difference in loading speed

P(kN)
600
point C1,3
500
400
p oint C2
300
200
100
SEEE(F50TA) / new type
Td (unknown)
point O
0 is caused by the length of the tension bar, estimated
0 5 10 15 20 25
(b) Difference in the hold method of loading
Fig.20 Test result of changing loading condition.
gradient is slightly more gentle than that at the
temporary pooling head (measurement point b)
(Fig.17(b), O-C2). At the stage before lift-off in the
loading process, the displacement of the tendon is
small, and such a phenomenon is considered to be
caused by the length of the tension bar.
b) The effect of different loading jacks
The special jack and SAAM jack13), shown in
Fig.18, were developed for the lift-off test. The ten-
sion bar of these jacks is short. The jack developed
for the lift-off test is more lightweight and compact
than the conventional center hole jack. Therefore, it
is possible to test many anchors in a short period. Of
these, by using the special jack, it is possible to test
by lifting the anchor head directly, even if there are
anchors with short extra length of tendon that are
difficult to test. This study compared the lift-off test
results using special jack and center hole jack with
VSL(E5-4). Figure 19 shows the test results. N-1, 2
shows the load-displacement curve for the center
hole jack. N-3, 4 shows the load-displacement curve
for the special jack. In addition, in both cases, the
displacement meter was set at the tip of tension bar.
The residual tensile load was evaluated by the tan-
gent method.
There is no significant difference between the re-
Fig.21 Corrosion anchor.
sidual tensile load measured using the center hole
jack (Fig.19, point C1,2) and the residual tensile load
for the special jack (Fig.19, point C3,4). However, the
D-1
D-2
linear gradient before lift-off for the center hole jack
D-3 (Fig.19, O-C1,2) is more gentle than that for the spe-
cial jack (Fig.19, O-C3,4). Sakai et al.13) stated that the
difference between center hole jack and SAAM jack
30 (mm)
from laboratory tests and field tests14). Moreover, in
recent studies, for anchors of the same type, it has
been confirmed that the linear gradient before lift-off
tends to become steep, depending on the increase of
the residual tensile load15), 16). Therefore, particularly
when the displacement is measured at the tip of the
tension bar by using the center hole jack, it is im-
portant to analyze the test results from the
load-displacement curve in consideration of the ef-
fect of elongation of the tension bar.
For working anchors, the state of the anchors can
be estimated by performing lift-off tests regularly
and comparing the changes in the incline of the
load-displacement curve. For this reason, when dif-
ferences in the load-displacement curve during pe-
riodic inspection are observed, if the past test con-
dition is not clear, it is difficult to determine whether
the difference is a result of different test conditions
or the changes in the residual tensile load and ma-
terial deterioration. Therefore, when lift-off test is
performed, it is important to record the test condi-
tion; for example, the type of hydraulic jack and the
position of displacement measurement, in addition to
the test results.
(4) Effects of different loading methods
Figure 20(a) shows the load-displacement curve
of two to five cycles when varying loading speed
for the case of VSL(E5-2). In each loading speed, the
load-displacement curves in lift-off tests are almost

114
 P(kN)
It was the SEEE-type anchor with 438kN design anchor
250 load. We observed the break point of the tendon of the
anchor. Figure 21 shows a significant corrosion in this
200 point C anchor. Figure 22(a) shows the load-displacement
curve at the time. Figure 22(b) shows the
150
load-displacement of anchor that has the same spe-
100 break in 293kN during cifications. The two load-displacement curves are
the pre-loading
significantly different. The residual tensile load of
50 SEEE(F100) / old type the anchor that broke at pre-loading is around 180kN
Td=438kN
point O (Fig. 22(a), point C). The residual tensile load of the
0
0 1 2 3
healthy anchor is around 150kN (Fig.22(b), point C).
(mm)
The residual tensile load of the broken anchor is closer
(a) Broken anchor in pre-loading to the design anchor load than to the residual tensile
load of the healthy anchor. However, the linear gra-
P(kN) dient of the broken anchor is more gentle than that of
250
the healthy anchor. Also, the displacement before
200
lift-off exceeds 1mm (Fig.22(a), O-C). On the other
hand, in the case of the healthy anchor, the dis-
150 point C placement before lift-off is as small as about
0.5mm19). There were no clear differences in the state
100 of the two anchors found in the inspection of the anchor
SEEE(F100) / old type
head before lift-off test. Therefore, it is important to
50
Td=438kN perform the lift-off test carefully by performing
0
point O preloading with a slower loading speed and con-
0 1 2 3 (mm) firming visually the status of the anchor to lower the
risk of breakage of anchor, etc.
(b) Healthy anchor
Fig.22 Comparison of the broken anchor and the healthy
anchor.
5. CONCLUSION
identical. In addition, there are no big differences in
the measured values of the residual tensile load using The study carried out lift-off tests for the
the tangent method. Figure 20(b) shows the results wedge-fixed type, nut-fixed type, and wedge-nut
of the test for SEEE (F50TA) conducted by changing combination-fixed-type anchors. Likewise, we stu-
the loading hold time. The load-displacement curves died the methods to evaluate the appropriate number
are about the same. Moreover, there are no large of cycles of lift-off tests and the residual tensile load.
differences in the measured values of the residual We also analyzed the differences caused by dis-
tensile load. Sakai et al.13) studied the effects of placement measurement positions, test equipment,
loading speed on the SEEE anchors of nut-fixed-type and loading methods. The following results were
in the lift-off test using the SAAM jack17). As a re- revealed:
sult, they concluded that it was not necessary to (1) In many anchors on which lift-off test was per-
consider the effects of different loading methods and formed, the residual displacement of the first
loading speed, because they did not observe a sig- cycle was greater compared with those of the
nificant difference in the load-displacement curve at second and subsequent cycles, and the residual
stage loading carried out conforming to the quality displacement after the second cycle was less than
assurance tests and in continuous loading conducted 1mm. The residual displacement of the first cycle
by changing the loading speed18). From these results, was greater compared with the second and sub-
it is thought that the differences in the method of sequent cycles. In addition, the residual dis-
loading speed and loading hold in lift-off test need placement after the second cycle was less than
not be considered. 1mm. Therefore, when lift-off tests are per-
The maximum load of lift-off test is 1.2 times that formed, preloading should be performed in the
of the design anchor load defined in the Geotechnical first cycle, and it is necessary to perform the
Society criteria. However, in this study, there was an actual loading after confirming the convergence
anchor that broke at the preloading stage, before the of the residual displacement in the second and
residual tensile load reached the design anchor load. subsequent cycles.

115
(2) We evaluated the residual tensile load obtained
from the lift-off tests. If the residual tensile load
is assumed as the load at the time of the lift-off,
when the fixing device starts to lift away from the
bearing plate, the variation in measured value
increases. Therefore, it is appropriate to evaluate
the measured value obtained from the intersec-
tion of a linear gradient before and after lift-off in
the load-displacement curve as the residual ten-
sile load.
(3) When the displacement at the tip of the tension
bar is measured using the center hole jack, the
effect of the length of the tension bar is, in some
cases, reflected in the linear gradient of the
load-displacement curve. The hydraulic jack
used in maintenance, such as SAAM jack and
special jack, can obtain a load-displacement
curve less affected by the length of the tension
bar, because the loading is performed by placing
the load directly on the anchor head. In addition,
the load-displacement curve obtained by lift-off
tests is a viable piece of data to evaluate the
status and health of an anchor. Therefore, in
lift-off tests, it is important to record testing de-
vices, such as hydraulic jack, and the test condi-
tions of the displacement measurement position,
along with test results.
(4) The effects of loading speed and loading hold
method on the test results were not confirmed.
For these reasons, when the lift-off test is carried
out, it is thought appropriate to set the loading
speed and the recording interval of the data,
needed to confirm the residual tensile load, in a
range that does not place burden on the anchor. In
addition, the breakage of anchor at pre-loading is
of primary concern; therefore, it is important to
carry out tests carefully by setting a low loading
speed and visually confirming the status of the
anchor.
6. CLOSING REMARKS
Recently, in view of the importance of maintain-
ing anchors, more advanced hydraulic jacks exclu-
sively for the purpose of lift-off tests have been de-
veloped20). The functions of the test equipment have
improved dramatically21). However, the method of
the lift-off test has not been defined. And therefore,
the residual tensile loads of anchors have not been
managed appropriately. From these facts, it is
thought that the significance of clarifying the test
method in this study is great for actually carrying out
maintenance.
116
It is thought that generally, several anchors are
installed on the slope and that the stability of the
slope is maintained as a plane. For this reason, even
if the function of one anchor deteriorates, it does not
lead to instability of the entire slope, immediately.
Therefore, it is important to evaluate the health of the
overall slope based on geological characteristics and
deformation in the slope. In the future, we plan to
measure the residual tensile load using the test me-
thod introduced in this paper and study methods for
evaluating the ability to deter deterioration of the
slope, as a plane.
REFERENCES
1) Okubo, K., Takemoto, M. and Amano, H. : On the ground
anchor in Expressway, Japan Road Association Miscellany,
Vol. 27, pp. 15, 2007.
2) Japanese Soc. of Soil Mechanics and Foundation Engi-
neering : Ground anchor design and construction stan-
dards, the commentary, pp. 13-15, 1988.
3) Okubo, K., Fujioka, K. and Takemoto, M. : The mainten-
ance of soil structure in the Expressway, Soil Mechanics
and Foundation Engineering, Vol. 56, No. 2, pp. 2-3, 2008.
4) Fujiwara, Y., Takemoto, M. and Yokota, S. : The evalution
of the ground anchor in Expressway, Journal of Geotech-
nical Engineering, Vol. 57, No. 6, pp. 42-43, 2009.
5) Fujiwara, Y., Takemoto, M., Yokota, S., Sakai, T. and
Tsunekawa, Y. : The effort to maintenance of ground anchor
in Expressway, The Foundation Engineering & Equipment,
Monthly, Vol. 38, No. 9, pp. 32-36, 2010.
6) Japan Highway Public Corporation : The design guidelines
for ground anchor Engineering, pp. 1-5, 1992.
7) Sakai, T., Tsunekawa, Y., Fujiwara, Y., Takemoto, M. and
Yokota, S. : The test of the existing loadcell on ground
anchor by using SAAM jack, The 64th Japan Society of
Civil Engineers Annual Scientific Conference Abstracts, pp.
713-714, 2009.
8) Fujiwara, Y., Takemoto, M., Yokota, S., Sakai, T. and
Tsunekawa, Y. : The set method of the loadcell on existing
ground anchor by using SAAM jack, The 45th Geotechnical
Engineering Research Association Annual Scientific Con-
ference Abstracts, pp. 1229-1230, 2010.
9) Japanese Geotechnical Society : Design and construction
standards for ground anchor, the commentary, pp.
187-188, 2000.
10) Sakai, T., Tsunekawa, Y., Ono, M. and Yamazaki, N. : The
effectiveness of the lift-off test using SAAM jack, Journal
of Geotechnical Engineering, Vol. 57, No. 10, pp. 46-47,
2009.
11) Sakai, T., Tsunekawa, Y., Fujiwara, Y., Takemoto, M.,
Yokota, S. and Fukuda, Y. : The relationship of the tension
of existing ground anchor and residual tensile load, The
64th Japan Society of Civil Engineers Annual Scientific
Conference Abstracts, pp. 711-712, 2009.
12) Fujiwara, Y., Takemoto, M. and Yokota, S. : On optimum
lift-off value in the lift-off test of ground anchor, The 64th
Japan Society of Civil Engineers Annual Scientific Confe-
rence Abstracts, pp. 709-710, 2009.
13) Sakai, T., Fukuda, Y., Nakamura, K. and Takeya, H. : The
development of the lightweight and compact new type
maintenance jack of ground anchor, Journal of Geotech-
nical Engineering, Vol. 55, No. 4, pp. 39-41, 2007.
14) Fukui, Y., Sakai, T. and Nakamura, K. : Lift-off test of
ground anchor using new type maintenance jack, The 62nd
 Annual Scientific Society of Civil Engineers Conference
Abstracts, pp. 139-140, 2007.
15) Sakai, T., Seki, Y., Tsunekawa, Y., Yokota, S., Takemoto,
M. and Fujiwara, Y. : The relationship of the gradient of
before lift-off and the residual tensile load in the lift-off test
of anchor, The 64th Annual Scientific Society of Civil En-
gineers Conference Abstracts, pp. 1881-1882, 2009.
16) Sakai, T., Tsunekawa, Y., Seki, Y., Fujiwara, Y., Takemoto,
M. and Yokota, S. : Relationship of the gradient of the
load-displacement curve in the lift-off test and residual
tensile load, The 44th Geotechnical Engineering Research
Association Annual Scientific Conference Abstracts, pp.
1103-1104, 2009.
17) Sakai, T., Tsunekawa, Y., Fukui, Y. and Nagano, M. : The
test method of remaining method of ground anchor by using
SAAM jack, Journal of Geotechnical Engineering, Vol. 56,
pp. 42-45, 2008.
18) Sakai, T., Nakamura, K., Fukui, Y., Yamazaki, N., Yana-
gigawa, K. and Tsunekawa, Y. : The effect of loading speed
in lift-off test of ground anchor, The 42th Geotechnical
117
Engineering Research Association Annual Scientific Con-
ference Abstracts, pp. 1389-1390, 2007.
19) Takemoto, M., Fujiwara, Y. and Yokota, S. : The evaluation
of ground anchor and the applicability of lift-off test–The
excavation survey of ground anchor–, The 45th Geotech-
nical Engineering Research Association Annual Scientific
Conference Abstracts, pp. 1879-1880, 2010.
20) Sakai, T., Yokota, S., Takemoto, M., Fujiwara, Y. and
Tsunekawa, Y. : The development of compact and
lightweight maintenance jack and the investigation as the
plane of ground anchor tension, Foundation Work, Vol. 38,
No. 9, pp. 79-82, 2010.
21) Sakai, T., Tsunekawa, Y., Yokota, S., Takemoto, M. and
Fujiwara, Y. : The evaluation of investigation of ground
anchor slope as the plane by using SAAM jack, The 65th
Annual Scientific Society of Civil Engineers Conference
Abstracts, pp. 123-124, 2010.
(Received December 7, 2015)