Tablle 1:Minimum safete y actors recomm ended for design

individual working anc h orages

Minimum Minimum
load safety Proof load
factor factor

An chorage category y =- Tf Ti

T T
1 Tem r y anchorages where
the service life is less than six
months and failure wouldo h ave 1.4
few s eriousconsequenc es an d
wouldd nnot endanger public safety,
'
eg ' short term pile test Ioa din
Usnl anchorages as a r ion
t'ac

system.
2 Tem po raary anchorages with a
service life of up two years,
where, although the 1.6 1.25
consequences of local fail
q 'erious, there is no danger
to public safety without adequate
Routine on-site acce ptance tests for warning, eg retainin
anc orages 3 Perman
rmanent anchorageses an
and al so
by G S Littlejohn* o
tern porary anchorages where the
consequences of failure are 2.0 1.5
Introduction serious, eg temporary
'
Followin ar eviewofgroundanchora e practice it is appar t h anchorages for main cables of a
'g
e variety of te~stin procedures and criteria f suspension bridge, or as a
e acceptance of individual
vi u anchorages which ar orpo e reaction for lifting heavy
into tern structural members.

g d the associated

'n
e ederation Internationale de la
) and the British Standards Institution

o th
'' h oragesinorderto r
e eat
General consideera t ions
'
To put the subject int
in o perspective the c asses of Ass a principle, acceptance testin
, u on-site proceduresofsh rtd ype.-
't bilit t t d("') W here s ifi o e d ures and durations g p
'e or
es s are required to demon 'ein es arguments can arise over the mo s t
1 ofh'l ation ofworkin an iyan a r c ass orvariab le ground.
d fhd ign in relation to uidelines on maxim
1 dh h o sa etythatth gap o s. e tests may which'n

1 h I I

ays edirectl y corn pared, egwheresh o rtfix ed sli gh t variations in the ma d

y o a ure. been ce om one coun to
On-site suitabilit y t ests are carried o to hor t
'
ove able moveme t,
ns,
i onsto the workin the such as ge pu -in'ofth
sam e ev . T y carried out inadva nce o f
'h enco g" s are
them se ected workin hor d repeated in subseq uen t cyclesandso e aviour
'on. e period ofm heanchoragescanb e b oth th confirmed
t fl Wl Gi enincr o
d onfid r
ence and the im proved reliability of gr d
ese tests indicat
ca e the results that sh ou ld b e
ewor 'anchora ges

es s are carried
o od

ges an
the
c
anch

Th
2 providesanexample
1
d ions 'a
a 'on ve ad
c c oadmcrements and th e

ve saved timimeandmoney andT bl

of current recommeendations. In spite of th
over

1 o' t t th o term abilit of

bort pport a load sim p li'city of these test prroce d uresacce p tanc o gro
th th e d esignworkin g Ioad and the efficfe fl oad 'cyo anchora ousw encom pared with other found t'ion
tr e anchorzone. A o e s s
ort t sult s with those eac stage of loading, the dis phlace ment should be
Atteach o d d
'1 '1 o Iongertermbeha 'o
Pr

'r
'
s ground is a variable material and o age constru ctionis no'odofo i xt en dd
e toatleast15
s e edprudent tosim 1
ans 'p, it is not consider ediate displacement rea readin g at five minutes.
selec io o eanchora procedures y y o ep can be monito
select 1 ho s, th e writer's field ex ri Ins ome countries cre ep di'sp acement
ivi ngine erin g, Uni yo dford, C airman of BSI d il ft th fl d pp li
roun chora gesss, ChairmanofFIPW kin
'
Groupon i or suc displacemen -o
Ground Anchoraages. load =1 r o predict service behaviour. 37
GROUND ENGINEERINGG . MARCH 1991
 110'Yo design free length or
design free length plus 509o
tendonbond length

Design free length

909o design free length

'e ~
150

Figure 1I
Table 2: Recommended load increments and minimum Acceptance
periods of observation for on-site acceptance tests criteria for
Temporary anchorages Permanent anchorages dhsplacement of
load increment (%T) load increment (%T) tendon at anchor
Minimum head.
1st load 2nd load 1st load 2nd load period of
cycle* cycle cycle* cycle observation
nltn
10 10 10 10 1
50 50 50 50 1
100 100 100 100 1 hx. hx, Ilx,
l
125 125 150 150 15 Elastic
100 100 100 100 1 displacement
hx
50 50 50 50 1 Displacement of tendon
10 10 10 10 1 atanchorhead
* For this load cycle, which often includes extraneous non-
recoverable movements such as wedge 'pull-in', bearing plate
settlement and initial fixed anchor displacement, there is no pause stage, and the anchorage can be passed or failed, within a period of
other than that necessary for the recording of displacement data. one to one and a half hours. Where monitoring involves a stressing
operation, eg a single lift-off check without load cell, an accuracy of
less than 5% is unlikely and longer observation periods of one day and
With regard to design considerations related to overall stability, it beyond are required. If necessary, the accuracy of lift-off checks can
is important to confirm that the post-tensioned load is properly be improved by repeating the test several times.
transferred through the free 'decoupled'ength of the tendon into the Where displacement-time data are required, a dial gauge/tripod
fixed anchor zone. system is suitable for short duration testing, given that the tripod
To establish the actual seat of load transfer within the anchorage, base should be surveyed accurately for movement (Figure 3). In
the apparent free length of the tendon should be calculated from the practice dial gauges reading to 0.01mm are commonly used during the
load-elastic displacement curve over the proof loading range using the test, and where movement of the tripod base is anticipated, its
manufacturer's value of elastic modulus and allowing for such effects position is checked before and after the test to an accuracy of 1mm.
as bedding of the anchor head and, in exceptional circumstances, For the testing procedures outlined above, acceptance criteria
temperature. It is normally adequate simply to record the ambient based on proof load-time data, apparent free tendon length, and short
temperature during the test, unless the monitoring equipment or term service behaviour, should be established for temporary and
anchored structure is known or observed to be temperature permanent anchorages. Appropriate criteria, which are well proven
sensitive. on site and judged to be cost effective, are detailed in the following
The free length analysis should be based on the results obtained sections (see also BS8081, 1989and FIP, 1991).
during the second cycle, otherwise extraneous non-recoverable
movements may mask the reproducible behaviour of the anchorage in
service (Figure 1). Figure 2: Proof
testingof5B
strand tendon at
For simplicity in practice the following equation is employed L Lynn dam,
West Virginia
= A,E, 6X, (courtesy of
Apparent free tendon length
T Nicolson
Construction Co
where A, is the cross section of the tendon, E, is the manufacturer's ofAmerica).
elastic modulus for the tendon unit, 6
X, is the elastic displacement of
the tendon (6
X, is equated to the displacement monitored at proof
load minus the displacement at datum load, ie 10% Tsay) and T is
the proof load minus datum load.

On completion of the second cycle, the anchorage should be

reloaded in one operation to 110%Tsay, and locked-off, after which
the load is re-read to establish the initial residual load. This moment
represents zero time for monitoring load or displacement-time
behaviour during service.
Where loss of load is monitored accurately using load cells with a
relative accuracy of 0.5%, readings can be attempted within the first
50 minutes (Figure 2). This development, although demanding more
sophisticated instrumentation, permits the on-site acceptance test to
38 be carried out in one operation alongside the routine post-tensioning
GROUND ENGINEERING MARCH . 1991

Proof load-time acceptance criteria
If the proof load has not reduced during the 15 minute observation
period by more than 5% after allowing for any movement of the
anchored structure, the anchorage may be deemed satisfactory. If a
greater loss of prestress is recorded the anchorage should be subject
to two further proof load cycles and the behaviour recorded. If the 5%
criterion is not exceeded on both occasions the anchorage may be
deemed satisfactory. If the 5% criterion is exceeded on either cycle
the proof load should be reduced to a value at which compliance with
the 5% criterion can be achieved. Thereafter, the anchorage may be
accepted at a derated proof load, if appropriate.
The 5% loss limit merely serves to illustrate that the anchorage will
not yield significantly at proof load and no attempt is made to ascertain
the proof load-time characteristic of the anchorage at this abnormally
high stress level.
As an alternative to these recommendations, the proof load can be
maintained by jacking and the anchor head monitored after 15 minutes
in which case the creep criterion is 5% 6 Xie the displacement
which would cause a 5% loss of proof load.
:, II V=
Figure 3:Displacement-time monitoring at Delli in
Switzerland(courtesy of VSL International).
In some countries limiting creep displacements are specified
irrespective of free tendon length eg 2mm (0.5 5 min.) for the US
Department of Transport (1984), 1.5mm (1 10 min.) for the Bureau
Securitas (1989)and 0. 5mm (5 15 min. ) for the Deutsche Industrie
Norm (1974 and 1976). These figures illustrate the ad hoc nature of
curr'ent creep criteria and at the present time few countries provide
correlations to permit either load or creep monitoring to be adopted.
For anchorages that do fail a proof load criterion it is noteworthy
that tendon unit stressing (mono jacking) may help to ascertain
location of failure (Figure 4). For standard bonded tendons the
pull-out of individual tendon units by mono jack indicates debonding at
the grout-tendon interface, whereas, if all tendon units hold their
individual proof loads, attention is directed towards failure of the fixed
anchor at the ground-grout interface. In this diagnostic test there is
no preferred sequence for stressing individual strands. For long high
capacity anchorages, mono jacking may also be useful in establishing a
uniform initial loading of strands, prior to cyclic loading by multi
tendon unit stressing (multi jack).
Where a multiple encapsulation system or an encapsulation with a
4Q multi-unit load transfer mechanism is employed the tendon load is
GROUND ENGINEERING . MARCH 1991
Sirrdi ~ Hrcnd
lol
Hydr wIN Orrmh
crmtwc liorr
SINGLE STRAND HOLLOW RAM JACK
IMorroiock)
Hydr wlrc Orrmd
ccmrrodII oil
~+-SO.~
MULTI.STRAND SOLID RAM JACK
Figure 4: Typical jacks for stressing steel tendons.
transferred to the adjacent grout though discrete lengths uniformly
distributed at intervals along the encapsulation or fixed anchor length.
In such circumstances individual strands or groups of strands have
different free tendon lengths and care is required to avoid strand
overstressing at the tendon proof load.
For routine multi strand stressing of these anchorages all strands
are stressed to different levels at a given tendon load or displacement,
and the maximum tendon proof load will be reached when the load in
the shortest strand attains 80% fpu (fpu = characteristic strength of
the strand). If it is necessary to increase further the proof load using a
multi jack, then a prestretch procedure must be introduced whereby
all strands, which are longer than the shortest strand length, are
tensioned to predetermined displacements such that when
multistrand stressing of the tendon takes place all strands attain the
same stress level at the required proof load. Alternatively, a
monojack may be used to load incrementally each strand to the same
value for proof loading.
Irrespective of the mode of stressing, which may be more time
consuming for distributed stress transfer fixed anchors, the proof
load-time acceptance criteria above still apply.
Load transfer acceptance criteria
The apparent free tendon length should be not less than 90% of the
free length intended in the design, nor more than the intended free
length plus 50% of tendon bond length or 110%of the intended free
tendon length (Figure 1).
The latter upper limit takes account of relatively short
encapsulated tendon bond lengths of 1m to 3m and fully decoupled
tendons with an end plate or nut, and the application of this upper limit
should be restricted to such circumstances.
The boundary limit of '90% free length'eflects a tightening of
tolerances over the years, bearing in mind that a limit of '80% free
length'as in common use in the 1960s. Where greased and sheathed
tendons are assembled under 'factory-controlled'onditions the more
rigorous criterion is attained without difficulty. The key to success is
to ensure that the plastics sheathing has a 'loose'it over the tendon.
If the observed free tendon length does fall outside either of the
limits a further two load cycles up to proof load should be carried out in
order to gauge reproducibility of the load-displacement data
particularly during the third and fourth cycles. For a 'long'pparent
free tendon length the comparison of load-displacement behaviour
during the second, third and fourth cycles checks for progressive
debonding within the tendon's bond length. For a 'short'pparent free
tendon length a potential explanation is friction within a greased and
sheathed decoupled free tendon length. Such a load transfer
mechanism is often visco-elastic in nature and given time, eg 6 to 12
hours, creep within the decoupled system will permit the friction load
 Table 3:Acceptance criteria for service behaviour at
residual load
Permissible loss of Permissible
load (% initial displacement
Period of residual (% of elastic extension
observation load) De of tendon at
initial residual load)

to be transferred to the tendon bond length. In these circumstances
the reproducibility check is only relevant to the load-displacement
behaviour during the third and fourth cycles.
Where the anchorage behaves consistently in an elastic manner,
the anchorage need not be abandoned, provided the reason can be
diagnosed and accepted. In this regard, it is noteworthy that the
elastic modulus E of a long strand tendon may be less than the
manufacturer's E value for a single strand, which has been measured
over a short gauge length between rigid platens (see also Janische,
1968 and Leeming, 1974). A reduction in the manufacturer's E value
of up to 10% should be allowed in any field diagnosis of the load-elastic
displacement behaviour of single or multi-strand tendons.
In the case of distributed stress transfer fixed anchors the above
load transfer criteria apply to an analysis of the apparent free length of
each individual strand or group of strands of equal length, bearing in
mind the designed decoupled and bonded lengths. To obtain the
appropriate strand load-displacement data it is usually necessary to
use multijacking with prestretch or monojacking. Analysis of an
average apparent free tendon length is not recommended since this
parameter masks the true load transfer behaviour of individual
strands.
Short term service acceptance criteria
Using accurate load cell and logging equipment, the residual load may
be monitored at 5, 15 and 50 minutes. If the rate of load loss reduces
to 1% or less per time interval for these specific observation periods
after allowing for temperature (where necessary), structural
movements and relaxation of the tendon, the anchorage may be
deemed satisfactory in relation to this serviceability criterion. If the
rate of load loss exceeds 1%, further readings should be taken at
observation periods up to 10 days (Table 3).
If, after 10 days, the anchorage fails to hold its load as given in Table
3, the anchorage is not satisfactory and following an investigation as to
the cause of failure, the anchorage should be (i) abandoned and
replaced, (ii) reduced in capacity or (iii) subjected to a remedial
stressing programme.
Where prestress gains are recorded, monitoring should continue to
ensure stabilisation of prestress within a load increment of 10% T.
Should the gain exceed 10% T, a careful analysis is required and it
will be prudent to monitor the overall structure/ground/anchorage
system. If, for example, overloading progressively increases due to
insufficient anchorage capacity in design or failure of a slope, then
additional support is required to stabilise the overall anchorage
system. Destressing to working loads should be carried out as
prestress values approach proof loads, accepting that movement may
continue until additional support is provided.
As an alternative to load monitoring, displacement-time data at the
residual load may be obtained at the specific observation periods in
Table 3, in which case the rate of displacement should reduce to 1%
6 e or less per time interval. To ensure compatibility of the
6
acceptance criteria 1% e is the displacement equivalent to the
amount of tendon shortening caused by a prestress loss of 1% initial
residual load, ie.
initial residual load x apparent free tendon length
ke =
area of tendon x elastic modulus of tendon
If the anchorages are to be used in the work and, on completion of
the on-site acceptance test, the cumulative relaxation or creep has
6
exceeded 5% initial residual load or 5% e, respectively, the
anchorage should be restressed and locked-off at 110%T, say. This
5 1 1
'7
15 2
50 3 3
150 4 4
500 5 5
1500 (= 1 day) 6 6
5000 (= 3 days) 7 7
15000 (= 10 days) 8 8
procedure ensures that a contingency overload is locked into the
ground anchorage at the start of its service.
As a general guide, either acceptance criterion for short term
service, ie rate of prestress loss or rate of displacement may be
applied quite independently for the common range of free tendon
lengths. For short free tendon lengths ( 5m), loss of prestress
becomes the more appropriate criterion, while for long free tendon
lengths ( 30m) it is clear that creep displacement may be more
important to limit and therefore more appropriate as an acceptance
criterion.
Records
Details of all forces, displacements, seating and other losses
observed during all stressing operations and the times at which the
data were monitored should be recorded in an appropriate form for
every anchorage. The completion of the record sheet and graphical
plot of load displacement during a stressing operation allows on-going
assessment of the anchorage performance and immediate
confirmation regarding compliance with the acceptance criteria (load
transfer, and percentage load or displacement change).
Safety
During stressing safety precautions are essential and operatives and
observers should stand to one side of the tensioning equipment and
never pass behind when it is under load. Notices should also be
displayed stating 'DANGER Tensioning in Progress'r similar
wording.
Reference should be made to published guidelines eg Concrete
Society (1980) and FIP (1989).
References
I 'Ground anchorages'S8081. British Standards Institution, 2 Park Street, London.
(1989).
2 Bureau Securitas 'Recommendations for the design, calculation, construction and
monitoring of ground anchorages'A Balkema, Rotterdam. (1989).
3 RS Cheney 'Permanent ground anchors'S Department of Transport, Federal
Highway Administration Report FHWA-DP-68-1R, Washington DC. (1984).
4 'Safety precautions for prestressing operations (post-tensioning) Notes for guidance'.
The Concrete Society, Terminal House, Grosvenor Gardens, London. (1980).
5 'Soil and rock anchors; bonded anchors for temporary use, DIN 4125 Part I; Permanent
anchors, Part 2 (1976), Deutsche Industrie Norm Fachnormen-ausschus Bauwesen,
Berlin. (1974).
de la Precontrainte 'Prestressed concrete safety
6 Federation Internationale
precautions in post-tensioning'. Thomas Telford Ltd., London. (1989).
7 Federation Internationale de la Precontrainte 'Recommendations for the design and
construction of prestressed ground anchorages'. Thomas Telford Ltd, London. (1991 to
bepublished).
8 W)aniche 'Recent improvements in the manufacture and properties of prestressing
steels', in FIP Proc. of Symposium 'Steel for Prestressing', Madrid, 1-4. (1968).
9 MB Leeming 'Discussion to prestressing steels by KW Longbottom and CP Mallet. The
Structural Engineer, Vol 52 (9), 357-362. (1974).
Acknowledgement
Extracts from BS8081: 1989 are reproduced with the permission of
BSI. Complete copies can be obtained through national standard
bodies. 43
GROUND ENGINEERING MARCH 1991