CHAPTER 7

LOAD TESTING AND TRANSFER OF LOAD

TO THE ANCHORED SYSTEM

7.1 INTRODUCTION

For anchored system applications, each ground anchor is tested after installation and prior to being
put into service to loads that exceed the design load. This load testing methodology, combined with
specific acceptance criteria, is used to verify that the ground anchor can carry the design load without
excessive deformations and that the assumed load transfer mechanisms have been properly
developed behind the assumed critical failure surface. After acceptance, the ground anchor is
stressed to a specified load and the load is “locked-off.”

7.2 CONCEPTS FOR MONITORING ANCHOR BOND ZONE CAPACITY

The bond zone of an anchor develops resistance in the surrounding ground by straining in response
to tensile loads applied at the anchorage. For anchor bond lengths in tension, the strains in the
tendon are greatest at the top and decrease along the length of the anchor bond zone. The amount of
load transfer to the ground at any particular strain will depend on the stress-strain characteristics of
the ground. Figure 64 illustrates two possible skin friction versus strain diagrams for a ground
anchor. Curve A represents a soil or rock where very little strain is required to mobilize most of the
skin friction. Curve B represents a weaker soil or rock where more strain is required to mobilize a
peak skin friction and where continued straining results in a reduction of skin friction to a residual
value.

Figure 64. Skin friction versus strain diagrams for ground anchors.

137
Early concepts for anchor testing were based on a uniform propagation of load transfer down the
bond length as tensile loads were increased. Figure 65 shows how the centroid of load, referred to as
the “fictitious anchorage point” (FAP), in the grout body was assumed to migrate toward the end of
the tendon. The assumption that all load transfer was mobilized when the FAP approached the
midpoint of the bond length formed the basis for early acceptance testing. However, this concept of
uniform load transfer is not valid for soil anchors and only approximates behavior of most rock
anchors.
Movement of Fictitious Anchorage Point (FAP)

Load = 0

No Movement of Bottom of Bond Zone

P FAP

2 P FAP

3 P FAP

Bottom of Bond Zone

Begins to Move
3 P FAP

Sliding of Anchorage
(Failure)

Figure 65. Stress propagation in bond length of ground anchor.

The current approach to monitoring bond zone capacity in soils has been used since the 1970s and is
based on creep of the grouted body under a constant load. As shown in figure 66a, the rate of creep
of the bond zone is directly related to the applied load. Creep tests on numerous anchors have shown
that when the creep rate exceeds 2 mm per log cycle of time, additional loads applied to the tendon
will result in unacceptable continuing grout body movements. As shown in figure 66b, a maximum
load, Tc, defined as the critical creep tension does exist for each bond zone. This critical creep
tension corresponds to the load at which the creep rate exhibits a sharp upward break. Monitoring
small creep movements (typically less than 1 mm) under constant applied tension loads requires
appropriate testing equipment. Both the absolute value of applied load and, more importantly, the
ability to maintain a constant load for a substantial period of time must be addressed.

138
 Figure 66. Evaluation of critical creep tension.

7.3 TESTING AND STRESSING EQUIPMENT

7.3.1 General

Each ground anchor is load tested to verify its capacity. The load test is performed at the ground
surface and consists of tensioning the prestressing steel element (i.e., strand or bar) and measuring
load and movement. In this section, equipment that is commonly used for load testing is described.
A typical load test setup for a strand and bar tendon is shown in figures 67 and 68, respectively.
Typical load test equipment includes: (1) hydraulic jack and pump; (2) stressing anchorage; (3)
pressure gauges and load cells; (4) dial gauge to measure movement; and (5) jack chair.

139
 STRESSING
ANCHORAGE
DIAL GUAGE

JACK

FIXED
BASE

Figure 67. Typical equipment for load testing of strand ground anchor.

7.3.2 Equipment Used in Load Testing

7.3.2.1 Hydraulic Jack and Pump

A hydraulic jack and pump are used to apply load to the tendon either at the anchor head or at a
pulling head attached to the prestressing steel. The hydraulic jack must be capable of applying a
concentric load to the tendon. The load should be transferred to all of the prestressing elements of
the tendon simultaneously. Applying the load to a single strand of a multistrand tendon should not
be allowed. The ram travel shall be at least 152 mm and preferably not be less than the theoretical
elongation of the tendon at the maximum test load. If elongations greater than 152 mm are required,
restroking can be allowed. In addition, the hydraulic jack should be capable of:

• applying and releasing load incrementally, as required by test procedures;

• applying each load increment within 60 seconds; and

• applying the maximum test load (termed the test load) within 75 percent of the pressure
rating of the jack and pump system.

When long, high capacity ground anchors are used, it may not be possible to apply each load
increment within 60 seconds. For this case, deformation measurements should begin when the load
is achieved.

140
 Figure 68. Typical equipment for load testing of bar ground anchor.

7.3.2.2 Stressing Anchorage

A stressing anchorage is used in front of the jack head to grip the prestressing steel element during
loading. For bar tendons, the stressing anchorage generally consists of a bearing plate with a
countersunk hole and a nut. Common practice is to order the production bar tendons to the actual in-
place length and use a short bar segment and coupler to extend the bar at the stressing head for
testing. For strand tendons, the stressing anchorage can be similar to the anchor head. The stressing
anchorage rests on a bearing plate; the bearing plate sits on front of the jack.

141
7.3.2.3 Pressure Gauges and Load Cells

The standard device used to monitor load is a pressure gauge attached to the jack pump, either alone
or in concert with a center hole load cell mounted in the stressing train. The readings on the jack
pressure gauge are used to determine the absolute value of applied load. For extended load hold
periods, load cells are used as the means to monitor a constant applied load while the pump is
incrementally adjusted. Over extended periods of time, any load losses in the jack will not be
reflected with sufficient accuracy using a pressure gauge. Also, temperature changes can affect the
hydraulic jack and/or pressure gauge readings. For proof tests and for lift off tests, a pressure gauge
alone is usually used for measuring load. For all tests involving extended load hold periods (i.e., all
creep tests), a load cell should be used in concert with a pressure gauge.

Calibration of pressure gauges and load cells should be performed within 45 working days of the
date when they are submitted for approval for the project. Calibration certifications and graphs for
pressure gauges and load cells must be provided by the contractor before use. A second certified
pressure gauge should be kept on-site to be used for periodic check of jack pressure gauges. The
pressure gauge shall be graduated in increments of 690 kPa or less.

7.3.2.4 Dial Gauge to Measure Movement

Total movement of the tendon is commonly measured using a dial gauge fixed to a tripod or other
support device that is independent of the structure. Dial gauges should be capable of measuring and
being read to the nearest 0.025 mm. A dial gauge should be used that has sufficient travel to be able
to measure in excess of the maximum elongation of the tendon. Care should be taken to ensure that
the dial gauge is aligned perpendicular to the end of the tendon or other plane of measurement. Dial
gauges with travels greater than 100 mm are not recommended. Where a tendon is expected to
elongate in excess of 100 mm during a load test, two or more gauges with shorter travel lengths may
be used together with the gauges being reset at interim points during the load test.

7.3.2.5 Jack Chair

For bar anchors, a jack chair is placed over the anchor head and it rests on the bearing plate. The
jack chair enables testing to be performed on bar anchors with the nut already in place and permits
access to the nut during transfer of the lock-off load. The jack chair must be capable of transferring
100 percent of the specified minimum tensile strength (SMTS) of the prestressing steel element to
the bearing plate. Jack chairs can also be used on strand tendons to permit the wedges to be placed
on the strands and set after the completion of the test.

7.4 ANCHOR LOAD TESTING

7.4.1 Introduction

A unique aspect of ground anchors, as compared to other structural systems, is that every ground
anchor that is to be part of a completed structure is load tested to verify its load capacity and load-
deformation behavior before being put into service. The acceptance or rejection of ground anchors is

142
determined based on the results of: (1) performance tests; (2) proof tests; and (3) extended creep
tests. In addition, shorter duration creep tests (as opposed to extended creep tests) are performed as
part of performance and proof tests. Proof tests are the most common and are performed on the
majority of the ground anchors for a particular project. The number of performance and extended
creep tests that are performed for a project depends upon whether the anchors are for a temporary
support of excavation or permanent application and the type of ground.

Every ground anchor is tested using one of the particular tests introduced above. The results of these
tests are compared to specified acceptance criteria to evaluate whether the ground anchor can be put
into service. The acceptance criteria are based on allowable creep and elastic movements of the
anchor during load testing. A brief discussion of each test type follows.

7.4.2 Performance Tests

7.4.2.1 General

Performance tests involve incremental loading and unloading of a production anchor. The
performance test is used to verify anchor capacity, establish load-deformation behavior, identify
causes of anchor movement, and to verify that the actual unbonded length is equal to or greater than
that assumed in the anchor design. The results of a performance test may also be used to assist in the
interpretation of the simpler proof test.

Performance tests are commonly performed on the first two or three production anchors installed and
thereafter on a minimum of two percent of the remaining production anchors. Additional
performance testing may be required where creep susceptible soils are suspected to be present or
where varying ground conditions are encountered. Where ground conditions are variable,
performance test anchors should be located near geotechnical borings, if possible, to facilitate the
interpretation of test measurements.

7.4.2.2 Procedures for Performance Test

The load schedule for a performance test is shown in the first three columns of table 21. The first
step in a performance test comprises applying a nominal load to the anchor tendon. This load,
termed the alignment load, is typically no more than five percent of the design load and its purpose is
to ensure that the stressing and testing equipment are properly aligned. The displacement measuring
equipment is zeroed upon stabilization of the alignment load, AL, as shown on figure 69. During the
first load cycle, the load is raised to 25 percent of the design load and the incremental movement is
recorded (i.e., Point 1 on figure 69). The load is then reduced back to the alignment load. This
procedure is repeated, using load increments as shown on table 21, until the maximum testing load,
referred to as the test load, is achieved. The test load may vary from 120 to 150 percent of the design
load with 133 percent being commonly used for permanent applications and 120 percent being
commonly used for temporary applications. A test load of 150 percent may be used for anchors in
potentially creeping soils or when an independent reference cannot be established for the dial gauge.

143
 Table 21. Steps for the performance test.

Step Loading Applied Load Record and Plot Record and Plot Calculate Elastic
Total Movement Residual Movement
(δti) Movement (δri) (δei)
1 Apply alignment load (AL)
2 Cycle 1 0.25DL δt1
AL δr1 δt1-=δr1==δe1
3 Cycle 2 0.25DL δ2
0.50DL δt2
AL δr2 δt2-=δr2==δe2
4 Cycle 3 0.25DL δ3
0.50DL δ3
0.75DL δt3
AL δr3 δt3-=δr3==δe3
5 Cycle 4 0.25DL δ4
0.50DL δ4
0.75DL δ4
1.00DL δt4
AL δr4 δt4-=δr4==δe4
6 Cycle 5 0.25DL δ5
0.50DL δ5
0.75DL δ5
1.00DL δ5
1.2DL δt5
AL δr5 δt5-=δr5==δe5
7 Cycle 6 0.25DL δ6
0.50DL δ6
0.75DL δ6
1.00DL δ6
1.2DL δ6
1.33DL δt6, zero reading
for creep test
8 Hold load for 10 minutes while recording movement at specified times. If the total movement measured
during the load hold exceeds the specified maximum value then the load hold should be extended to a total of
60 minutes.
9 Cycle 6 cont’d. AL δr6 Cycle 6:=δtn-=δr6==δe6
10 Adjust to lock-off load if test results satisfy acceptance criteria, otherwise see section 7.4.5.4
Notes: AL = Alignment Load, DL = Design Load, δi = total movement at a load other than maximum for cycle, i =
number identifying a specific load cycle.

144
 Figure 69. Plotting of performance test data (after PTI, 1996).

At the test load, a constant load is held for ten minutes prior to reducing the load to the lock-off load.
During this ten minute load hold period, movements are measured and recorded at 1, 2, 3, 4, 5, 6,
and 10 minutes. The purpose of this load hold is to measure time-dependent (i.e., creep) movements
of the anchor. This portion of the performance test is referred to as a creep test. If the total
movement between 1 and 10 minutes exceeds the specified maximum creep movement (see section
7.4.5.2), the test load is maintained for an additional 50 minutes and total movement is recorded at
20, 30, 40, 50 and 60 minutes. If the results of a creep test for a specific anchor indicate that creep
movements are excessive relative to specified criteria, the anchor may be incorporated into the
structure at a reduced load, the anchor may be replaced, or, only in the case of postgroutable anchors,
the anchor may be regrouted and then retested.

7.4.2.3 Recording of Performance Test Data

The magnitude of each load is determined from the jack pressure gauge. During creep testing, a load
cell is monitored to insure that the jack load remains constant. The load-deformation data obtained
for each load increment in a performance test are plotted as shown in figure 70. Movement is
recorded at each load increment and for the alignment load. The total movement (δt) that is
measured consists of elastic movement and residual movement. Acceptance criteria for anchors
require that the elastic movement of the anchor be known. Elastic movements (δe) result from
elongation of the tendon and elastic movements of the ground anchor through the ground. Residual
movement (δr) includes elongation of the anchor grout and movement of the entire anchor through
the ground. The residual movement for a given increment of load is the movement that corresponds
to the net “irrecoverable” movement that occurs upon application of a load increment and the
subsequent relaxation of the load to the alignment load (see figure 69 for definition of δr6). The

145
elastic movement is therefore the arithmetic difference between the total movement measured at the
maximum load for a cycle and the movement measured at the alignment load (see table 21).
Although not used for anchor acceptance, residual movement is an indicator of the stress-strain
behavior of the ground-grout bond in the anchor bond zone.

Figure 70. Plotting of elastic and residual movement for a performance test (after PTI, 1996).

During the creep test portion of the performance test, the movement measured at specified times (i.e.,
1, 2, 3, 4, 5, 6, and 10 minutes) is recorded. The time at which the total movement is measured for
the test load (i.e., time at which point 6 on figure 69 is measured) represents the start time for the
creep test. The movement from one to ten minutes after this starting time is recorded and compared
to the acceptance criteria with respect to creep. If the creep acceptability criterion is not satisfied, the
test load is held on the anchor for an additional 50 minutes. The total amount of movement between
6 and 60 minutes is recorded and compared to specified criteria.

Creep acceptability criteria were established for anchors using bare prestressing strand. For epoxy-
coated filled strand tendons, the creep movements of the strand itself are significant during load
testing. The creep movements of the strand should be deducted from the total movement measured
during a load test so that the creep movements within the ground can be accurately calculated.

7.4.2.4 Analysis of Performance Test Data

One of the acceptability criterion for ground anchors is based on measured elastic movements of the
ground anchor during load testing. The elastic movements calculated from a load increment during a
performance test are evaluated using the equations shown in table 21. These elastic movements

146
should be calculated for each load cycle and plotted versus each load as shown on figure 70. The
residual movement curve should also be plotted. For a soil anchor to be considered acceptable with
respect to elastic movements, the elastic movement at the test load must exceed a specified minimum
value. For a rock anchor, the elastic movement must be bounded by a specified minimum and a
specified maximum value. The acceptability criteria with respect to elastic movement are described
in section 7.4.5.3.

7.4.3 Proof Tests

7.4.3.1 General

The proof test involves a single load cycle and a load hold at the test load. The magnitude of the
applied load is measured using the jack pressure gauge. Load cells are only required for creep tests
in soils where the performance tests show a creep rate exceeding 1 mm per log cycle of time. The
proof test provides a means for evaluating the acceptability of anchors that are not performance
tested. Data from the proof test are used to assess the adequacy of the ground anchor considering the
same factors as for performance test data. Where proof test data show significant deviations from
previous performance test data, an additional performance test is recommended on the next adjacent
anchor.

7.4.3.2 Proof Test Procedures and Recording and Analysis of Proof Test Data

The proof test is performed in accordance with the procedure outlined in table 22. The total
movement from each load cycle in a proof test should be plotted as shown in figure 71. If an unload
cycle is included (Step 4 in table 22), residual movements and elastic movements should be
calculated for the test load. This calculation is the same as that previously described for performance
tests. If an unload cycle is not performed, an estimate of residual movement can be based on
performance tests on other production anchors from the same project.

Table 22. Test procedure for ground anchor proof test.

Step 1. Apply the alignment load at which total movement is assumed equal to zero.

Step 2. Successively apply and record total movements for the following load increments to
the test load: 0.25DL, 0.50DL, 0.75DL, 1.00DL, 1.20DL, 1.33DL (i.e., the test
load). Note that the test load for an anchor for a temporary support of excavation
application may be set at 1.20 DL.

Step 3. Hold test load for ten minutes and record total movement.

Step 4. (Optional) Unload to alignment load and record residual movement.

Step 5. If test results satisfy acceptance criteria, reduce load to the lock-off load (or if Step 4
was used, increase load to lock-off load), otherwise follow guidance provided in
section 7.4.5.4.

147
 Figure 71. Plotting of proof test data (after PTI, 1996).

7.4.4 Extended Creep Testing

7.4.4.1 General

An extended creep test is a long duration test (e.g., approximately 8 hours) that is used to evaluate
creep deformations of anchors. These tests are required for anchors installed in cohesive soil having
a plasticity index (PI) greater than 20 or liquid limit (LL) greater than 50. For these ground
conditions, a minimum of two ground anchors should be subjected to extended creep testing. Where
performance or proof tests require extended load holds, extended creep tests should be performed on
several production anchors.

7.4.4.2 Procedures for Extended Creep Test

The test arrangement for an extended creep test is similar to that used for performance or proof tests.
The increments of load for an extended creep test are the same as those for a performance test. At
each load cycle, the load is held for a specific period of time and the movement is recorded. During
this observation period, the load should be held constant. The load is assumed to remain reasonably
constant if the deviation from the test pressure does not exceed 0.35 MPa. The loading schedule and
observation periods for each load cycle in an extended creep test for a permanent anchor are
provided in table 23. Information on extended creep tests for temporary anchors is provided in
FHWA-RD-82-047 (1982).

148
Table 23. Load schedule and observation periods for extended creep test for permanent anchor.

Loading Maximum Total Observation Movements measured at following times (min)

Cycle Cycle Load Period (min)
1 0.25DL 10 1, 2, 3, 4, 5, 6, 10
2 0.50DL 30 1, 2, 3, 4, 5, 6, 10, 15, 20, 25, 30
3 0.75DL 30 1, 2, 3, 4, 5, 6, 10, 15, 20, 25, 30
4 1.00DL 45 1, 2, 3, 4, 5, 6, 10, 15, 20, 25, 30, 45
5 1.20DL 60 1, 2, 3, 4, 5, 6, 10, 15, 20, 25, 30, 45, 60
6 1.33DL 300 1, 2, 3, 4, 5, 6, 10, 15, 20, 25, 30, 45, 60, 300

7.4.4.3 Recording and Analysis of Extended Creep Test Data

The test data for an extended creep test should be plotted as shown in figure 72. The creep
movement at any time is the difference between the total movement and the movement measured at
one minute. Creep curves for a typical extended creep test are shown on figure 72. Each curve is for
a separate load hold. The creep rate is defined as the slope of the curve per log cycle of time.

Figure 72. Plotting of extended creep test data (after PTI, 1996).

Extended creep test data are used in evaluating the acceptability of an anchor with respect to the
creep acceptance criteria. Creep rates should be evaluated for each of the curves shown in figure 72.
These creep rates are compared to the maximum specified rate.

149
7.4.5 Acceptance Criteria

7.4.5.1 General

An anchor may be put into service at the lock-off load following load testing if certain specified
acceptability criteria are satisfied. These criteria, which are described herein, prescribe acceptable
limits of creep (i.e., movement during load holds) and elastic movement measured during anchor
load tests. The creep and elastic movement criteria have been integrated into an acceptance decision
tree that is described in this section. This decision tree describes procedures that are to be used in the
event that a specific criterion is not satisfied.

7.4.5.2 Creep

Creep testing, either as part of a performance or proof test or as an extended creep test, is performed
on each production anchor to evaluate creep movement of the anchor grout body through the ground.
For an anchor to be accepted, total movements measured during load holds must be below a
specified limit.

For performance and proof tests, the measured total movement for the required load hold at the test
load should not exceed 1 mm between 1 and 10 minutes. If the movements are less than the 1 mm
for this period, the anchor is considered acceptable with respect to creep. As previously discussed,
for load tests in which the measured total movement exceeds the criteria described above, the load is
held for an additional 50-minute period of time. If the measured total movement over this additional
time period does not exceed 2 mm between 6 and 60 minutes, then the anchor is considered
acceptable with respect to creep.

For extended creep testing, the total movement for any load hold should not exceed 2 mm per
logarithmic cycle of time (PTI, 1996) over the final log cycle of time of each load increment.
Alternatively, the anchor load may be reduced to 50 percent of the load where acceptable creep
movements were measured over the final log cycle of time.

7.4.5.3 Apparent Free Length

The apparent free length of a tendon forms the basis for evaluating the acceptability of a ground
anchor with respect to elastic movement. The apparent free length is defined as the length of the
tendon that is, based on measured elastic movements at the test load, not bonded to the surrounding
ground or grout. The apparent free length, La, may be calculated using the following equation:

A t E sδ e 1
La = x 9 (Equation 49)
P 10

where: At is the cross sectional area of the prestressing steel, Es is the Young’s modulus of the
prestressing steel, δe is the elastic movement at the test load, and P is equal to the test load minus the
alignment load. Standard SI units are: La (m); At (mm2); Es (kPa); δe (mm); and P (kN). For proof

150
tests where the residual movement is not measured or estimated, the apparent free length may be
calculated using the total movement in place of the elastic movement.

For long multistrand tendons, it is likely that the elastic modulus of the multistrand tendon will be
less than the manufacturers elastic modulus for a single strand. Because of this, PTI (1996)
recommends that a reduction in the manufacturers reported elastic modulus of 3 to 5 percent be
allowed for satisfying apparent free length criteria.

Minimum Apparent Free Length Criterion

If the apparent free length is greater than the specified minimum apparent free length, it is assumed
that the unbonded length has been adequately developed. The minimum apparent free length is
defined as the jack length plus 80 percent of the design unbonded length. An apparent free length
less than the specified minimum apparent free length may indicate that load is being transferred
along the unbonded length and thus within the potential slip surface assumed for overall stability of
the anchored system. Alternately, an apparent free length less than the specified minimum apparent
free length may be caused by friction due to improper alignment of the stressing equipment or tendon
within the anchorage. Where test results do not satisfy this criterion, the anchor may be subjected to
two cycles of loading from the alignment load to the test load in an attempt to reduce friction along
the unbonded length. The apparent free length is then recalculated based on the elastic movement at
the test load for the reloaded anchor. A value greater than the jack length plus 80 percent of the
design unbonded length may be used to define the specified minimum apparent free length for cases
in which the redistribution of friction along the unbonded length could cause unacceptable structural
movement or where there is the potential for prestressing loads to be transferred in the unbonded
length by tendon friction.

Maximum Apparent Free Length Criterion

The acceptance criterion based on maximum apparent free length was used in the past when load
transfer along the bond length was assumed to propagate at a uniform rate as the applied load was
increased (see figure 65). For that assumption, the maximum value of apparent free length was
restricted to elastic movements of 100 percent of the free length plus 50 percent of the bond length
plus the jack length. However the concept of uniform distribution of bond is not valid for soil
anchors and only approximates the behavior of most rock anchors. The primary use of this criterion
is as an alternate acceptance criterion for proof tests in sound rock where creep tests are waived.
Anchors that do not pass this preliminary criterion are subsequently creep tested to determine
acceptability before a decision is made to reject the anchor.

151
7.4.5.4 Ground Anchor Acceptance Decision Tree

PTI (1996) developed a ground anchor acceptance decision tree that is shown in figure 73. The
decision tree does not include the maximum apparent free length criterion as this criterion is not
routinely used. The purpose of the decision tree is to provide recommendations as to the field
procedures that should be followed in the event that an anchor does not satisfy specified acceptance
criteria. Anchors that do not satisfy the requirements for lock off at the design lock-off load may be
locked off at a reduced load or replaced.

Whether an anchor satisfies the minimum apparent free length criterion is the first decision to be
made using the decision tree. The ground anchor acceptance decision tree indicates that for an
anchor to be put into service at the design lock-off load, the elastic movement (i.e., minimum
apparent free length) criterion must be satisfied. The following sections provide information of the
recommended procedures to be used for an anchor that has passed the minimum apparent free length
criterion and for an anchor that has failed the minimum apparent free length criterion.

Anchors That Pass Apparent Free Length Criterion

For anchors which pass the minimum apparent free length criterion, but which do not pass the
requirements of the creep test, the anchor may, if possible, be post-grouted. Those anchors that can
be post-grouted will be retested and subject to an enhanced creep test and a more stringent
acceptance criterion as compared to creep and extended creep tests. For this enhanced creep test,
movements are monitored during a load hold at the test load for 60 minutes. The anchor may be
locked off at the design test load if the total movement does not exceed 1 mm between 1 and 60
minutes. If the anchor does not satisfy this criteria, it can be either rejected and replaced or locked
off at 50 percent of the load that the anchor holds without detectable movement. If the anchor cannot
be post-grouted, it may either be rejected and replaced or locked-off at 50 percent of the load that the
anchor holds without detectable movement.

Anchors That Fail Minimum Apparent Free Length Criterion

Anchors which fail the minimum apparent free length criterion may be either locked-off at a load no
greater than 50 percent of the maximum load attained during testing or rejected and replaced.
Replacement anchors must satisfy all project specifications. Changes in ground anchor locations
require approval from the design engineer. Where anchors are installed using prefabricated
connections to steel beams or sheet-piles, the failed anchor must be removed from the connection or
a new connection must be fabricated. Connections may not be offset from the center of a soldier
beam for a permanent anchor. Off-center connections will induce adverse bending and torsional
stress on the soldier beam and bending stresses in the tendon.

152
 ANCHOR TEST TO TEST LOAD

MINIMUM APPARENT FREE

LENGTH CRITERIA

PASS FAIL

CREEP CRITERIA

PASS FAIL

POSTGROUT NO POSTGROUT

LOCK-OFF
LOCK-OFF AT 50%
REJECT
AT DESIGN OF THE
AND OR
LOCK-OFF MAXIMUM
REPLACE
LOAD ACCEPTABLE
TEST LOAD

ENHANCED
ACCEPTANCE
REJECT
CRITERIA
AND
REPLACE

PASS FAIL
OR

OR

OR

LOCK-OFF
LOCK-OFF AT 50%
REJECT
AT DESIGN OF THE
AND
LOCK-OFF MAXIMUM
REPLACE
LOAD ACCEPTABLE
TEST LOAD

Figure 73. Ground anchor acceptance decision tree (after PTI, 1996).

153
7.4.5.5 Modification of Design or Installation Procedures

Multiple failures early in construction or multiple failures of adjacent anchors should be cause to
reassess subsurface conditions and/or design and installation procedures. Modifications to design and
installation procedures commonly include: (1) changing installation methods or anchor type; (2)
increasing the anchor length or anchor bond length or changing the inclination of the anchor; or (3)
reducing the anchor design load by increasing the number of anchors. A description of any proposed
changes should be submitted to the owner in writing for review and approval prior to implementing
the changes.

7.5 ANCHOR LOCK-OFF LOAD

After load testing is complete and the anchor has been accepted, the load in the anchor will be
reduced to a specified load termed the “lock-off” load. When the lock-off load is reached, the load is
transferred from the jack used in the load test to the anchorage. The anchorage transmits this load to
the wall or supporting structure.

The lock-off load is selected by the designer and generally ranges between 75 and 100 percent of the
anchor design load, where the anchor design load is evaluated based on apparent earth pressure
envelopes. Lock-off loads of approximately 75 percent of the design load may be used for temporary
support of excavation systems where relatively large lateral wall movements are permitted. Since
apparent earth pressure diagrams result in total loads greater than actual soil loads, lock-off at 100
percent of the design load typically results in some net inward movement of the wall. Lock-off loads
greater than 100 percent of the design load may be required to stabilize a landslide. For this case,
structural elements must be sized to transmit potentially large landslide forces into the ground.
Loads consistent with the required landslide restraint force to obtain a target slope stability factor of
safety are selected for the lock-off load.

When transferring the lock-off load to the anchorage, the load will inevitably be reduced owing to
mechanical losses associated with the physical transfer of load between two mechanical systems (i.e.,
the jack and the anchorage). These losses are referred to as seating losses and are generally on the
order of 1.6 mm for bar tendons and 6.4 mm for bare strand tendons (FHWA-DP-68-1R, 1988). For
strand tendons, seating losses occur as the jack ram is retracted and the wedges are pulled in around
the tendon. The wedges must be seated at a load no less than 50 percent of the ultimate load for the
tendon. This will prevent possible strand slip through the wedges if the load in the tendon increases
above the lock-off load during the service life. For epoxy coated strand, the wedges must bite
through the epoxy coating; this results in additional seating losses. To account for seating losses,
after the tendon is loaded to the lock-off load, the jack ram is extended by an amount equivalent to
the anticipated seating loss.

In the long-term, the load will also reduce due to relaxation in the prestressing steel. Long-term load
losses may be estimated as 4 percent for strand tendons and 2 percent for bar tendons (FHWA-DP-
68-1R, 1988). Specific information on relaxation losses should be obtained from the tendon
supplier. To account for these load losses, the load that is transferred to the anchorage may be

154
increased above the desired load based on results of a lift-off test. After the losses, the transferred
load will reduce presumably to the desired long-term load.

7.6 LIFT-OFF TESTING

After the load has been transferred to the anchorage, a lift-off test is performed. The purpose of a
lift-off test is to verify the magnitude of the load in the tendon. For strand tendons, the lift-off test is
performed by gradually reapplying load to the tendon until, for restressable anchor heads, the wedge
plate lifts off the bearing plate (without unseating the wedges) or, for cases where the hydraulic jack
rests on the anchor head, the wedges are lifted out of the wedge plate. For bar tendons, the lift-off
test is performed by gradually reapplying load to the tendon until the anchor nut lifts off the bearing
plate (without turning the anchor nut). Lift-off is evidenced by a sudden decrease in the rate of load
increase as observed on the jack pressure gauge. The load measured during the lift-off test should be
within five percent of the specified lock-off load. Where this criterion is not met, the tendon load
should be adjusted accordingly and the lift-off test repeated.

155