RESEARCH ON GROUND ANCHORS IN NON-COHESIVE SOILS

Etude de tirants scellés dans des sols, pulvérulènts

H. OSTERMAYER

by

and f. SCHEELE

Institute of ,Soil Mechanics and Foundation Engineering, Technical University, Munich/Fed. Rep. Germany

SOMMAIRE

Des essais d'ancrages en vraie grandeur surchantier ont été effectués dans des massifs desables graveleux compactés à différentes densités.Des jauges de déformation collées sur des bar­reaux d'acier donnent la distribution' des contrain­tes dans la zone fixe d'ancrage. Ces dernièresaugmentent progressivement avec la force appli­quée en tête d'ancrage jusqu'à la rupture. Afind'e.stimer par ,des essais standards in situ, l'in-

'fluence du type et de la densité du sol sur lacapacité portante du tirant d'ancrage, on a pré­senté le rapport entre la charge critique et lenombre de corps d'essais de pénétration (marteaude 50 kg).

SUMMARY

Full scale field tests on a'nchors have beenperformed in one and the same gravelly sandcompacted to various densities. During the pull­out tests the distribution of skin friction alongthe fixed anchor length was determined from straingauge measurements. The results of 5 series oftests comprising 30 anchors complement a pre­viously published design chart and give an addi­tional correlation between carrying capacity ofanchors and number of blows of penetration tests.The variation of skin friction along fixed anchorlength with increasing load or with load kept cons­tant over a longer period of time, helps to explainthe influence of soil density and fixed anchorlength on the carrying éapacity and longterm beha­viour of anchors.

INTRODUCTION

The last Conference on Diaphragm Walls and An­chorages (London 1974) and the Seminar on the sametopic (London 1976) showed that current practice inthe field of ground anchors is ahead of theory and thatthere is urgent need for a proper understanding ofthe behaviour of the anchors and the surroundingground under working conditions (C.P. Wroth, 1975).Design charts, which are based on field test results ofabout 300 anchors (H. Ostermayer,1975) may help toestimate carrying capacity of anchors in relation tofixed anchor length in certain soil conditions. How­ever, the main factor of soil density influencing the

carrying behaviour of anchors in non-cohesive soil hasnot yet been tackled systematically in field tests.

In 1975-1976 a major research program was carriedout in Munich to examine the influence of soil densityon the carrying capacity of anchors. In order to havea better insight into the carrying behaviour of anchors,research' aimed at investigating the distribution ofstresses for different anchor lengths. The effect oftime on the variation of stresses was also studied. Theso sought inforn1ation should then help toexplain thevarious important factors that influence the carryingcapacity and long-term behaviour of anchors.

TEST PROGRAM

On a test site five series of six anchors each wereinstalled and pull-out tests were· performed. A sche­matic arrangement of the test pit is shown in Fig. 1.The dimensions of the test pit were about 5 X 10 X 10meters. A rigid concrete wall was used as abutmentfor the pulling jack.

The anchors of the first series were installed inthe existing soil, which was sandy gravel of highdensity. For the following four test series the soil wasreplaced by gravelly sand. The grain size distributioncurve of the test. soil is given in Fig. 1, the coefficientof uniformity was U == 8 to 10. The density of the

92

sand was varied for each series. After every testseries the soil was removed and compacted again in,layers of about 30 cm height to a desired uniformdensity with the help of vibrators. The compactedsand carried a surcharge of about 2 fi grave!. Foreach series the soil density was checked by 8 standardpenetration tests (SPT) and 4 dynamic penetrationtests (50 kg hammer weight and 15 cm2 cone area).In 'addition unit weight and density index weredetermined for at least 6 samples.

A temporary anchor construction (Type A) wasused with a total length of 9 m and an inclination of

Test .series No 1 21

3 1 41

5

Prestressing- 4 X 16 mm32 mm dia.steel tendon dia.

ro Casing1

76.89

~ 89ro diameter (mm) 114""0

~

0 Grouting 0 0.5 0.5 0.5~() pressure to and ands::< (MN/m2) 5.0 1.0 2.0

Bond length3.0 1 2.0 and 4.5 1 3.0of tendon (m)

1 1

Soil typesandy

1

gravelly sandgravel

Density1.1 1.14 0.76 0.28 0.82

ro index D~ --- --- ---""0 Dynamic,.......·0 penetr. test > 80 76 20 2 30Vl - (N/I0 cm)

--- --- ---Standard

1 penetr. test > 130 120 43 Il 60(N/30 cm)

TABLE 1Anchor and soit data for 5 series

of tests with 6 anchors each

A total of 9 anchors were specially prepared forthe pull-out tests. In the laboratory temperature com­pensating strain gauges were attached to the steel barat pre-selected points. A typical arrangement of thegauges along the bond length Qf tendon and the freetendon length is shown in Fig. 2. It is pointed outthat the gauges were mounted on the profiled siclesof the bar, which were carefully treated before thegauges were cemented to the steel-bar (see Detail A).To protect against the grouting, the strain gauges andwire connections were given several waterproof coatings~

Each gauge was calibrated in t4e laboratory bystressing the steel bar.

The procedure of the installation of the anchors wasas follows:

- the steel casing was driven to the depth of about8 m;

8oreholePlastic tube

Prestressing bar, 32 mm dia.

Strain gouges

Grouted body

load cell

Detail A

about 20°. As shown in Table 1 series 2, 3 and 4had bond lengths of tendon Lv of 2.0 m and 4.5 mrespectively. The bore hole diameter of 89 mm andthe grouting pressure of about 0.5 MN/m2 was keptconstant. The anchors of series 1 and 5 had a bondlength of 3.0 m. For these anchors different bore holediameters (76 mm to 114 mm) and different groutingpressures (0 to 5 MN/m2) were used.

5tressing head~ Electrical laad cel!

OISPI~~::~~~~~U~?_ ~ HOllOoW :O~ j:Cko

0 • 0 0 • ~: f- 1 0 0 0 0 0 0 0 0 0 0 : 0 0 0 ;

~ .' ,<:" 0 0 0 0 0 0 0 Gravel surcharge 0 ~ 0

'jT L:~~·en;·:~o~:::~tSi~o~~:::~~ :0.06 2.0[mm) 60· 1 U/l]

hydraulic p~mp ~~/~1i11 . ~L Autamatic electrical 50 .'

laad regulatar(closed loap) 0

Fig. 1.' - General arrangement of full-scale anchor testset-up.

l

Fig. 3. - One test series in sanddug out for careful examination.

- the anchor bars with fix.ed spacers (àlong the bond:length) were inserted into the· casing;

- the bore hale was grouted under constant pres­sure along the planned bond length while retract­ing the casing simultaneously;

- in the upper part of the bore hale the remainingcement suspension was flushed out with water;

- the pull-out tests were performed after a periodof about 14 days.

The pulling force was applied in steps by a hollowram jack until the failure load was reached. Betweeneach step, loading and unloading cycles were appliedaccording to the German Standards for FundamentalTests (DIN 4125).

For each loading step, the displacements of theanchor head were measured manually through thereadings of the dial gauge, while the measurements ofthe strain gauges were recorded automatically.

At certain loading steps, the force was kept constantwith the help of an -electric-hydraulic regulator for aperiod up to 2 months.

AlI the anchors were dug out after the pull-outtests (see Fig. 3) and the grouted bodies were carefullyexamined. The diameters and lengths of the groutedbodies were measured and the surrounding soil condi­tions were checked.

TEST RESULTS

Load Carrying Capacity

In Fig. 4 the failure loads (ultimate load carryingcapacity Tf) of the five test series are presented inrelation to the length of the fixed anchor (bond-to­ground length La). The results confirm the validityof the previously published design chart (H. Oster­mayer, 1975) and supplement the chart with additional~urves for loose gravelly sand and very dense sandygravels~ The diagram shows the smallest linearincrease of carrying capacity with increasing bondlength for loose sands. Contrary to this in case of densesands, the greatest increase is encountered for smallerlengths which then tapers off steadily with increasinglength. With lengths of more than 6 to 7 m theincrease of carrying capacity per m length will probablybe the same whether the anchors are installed in looseor in dense sands and gravels. The reason for this

behaviour is a progressive failure mechanism whichwill be investigated in the following paragraphs.

In addition it should be noted that compared to thelarge influence of the soil density on the carrying·capacity the influence of grouting pressure (minimumpressure 0.5 MN/ m2) as well as the influence of thediameter of the grouted body (diameter of 10 to 15 cm)seems to be negligible.

Distribution of Skin Friction

As a result of strain gauge measurements at anchorsin dense sand a typical distribution of tensile forcesin the steel tendon is represented in Fig. 5. Th~ de­crease of forces from the front part to the rear ofthe bond l~ngth corresponds with the load transmis­sion from the tendon into the grouted body. As

. .l

2000

__ 1800z:~

..... 1600t-.:.

~- 1400uca.cu 1200CJ)

c::

Ë' 1000cU

'1J 800c0

QJ 600-cE

::;::) 400

200

2 4 6 8Bon d- t0 - 9r0und length l 0 [m]

Very dense

10

Type of soil OensitySPT

N30[bl/30cm]

• Gravelly Very dense 120

• sand Oense 60

• U=d60/ dl0 Medium d. 43

• =1,6/0,16 Loose 11

x Sandy gravelVery dense >130

U=15/0,3

.11,3 =Diameter of grouted body do=ll.3cm

JITITill Sandy gravelllJII1II u= 5~10

Gravelly sand

~U=8~10

andMedium ta coarse sand(with gravel)U= 3,5 ~ 4,5

Oiameter of grouted bodiesdo :: 10 ~ 15 cm

Fig. 4. - Carrying capacity of anchors in sandy gravel and gravelly sand showing influence of soil type, density andbond~to-ground length. .

94

---------~--------=------------:,....,.,...--------------~-------------------

Fig. 5. - Distribution of tensile load along bond length of .tendon (grouted body in dense sand).

Failureskin friction

length

Short - termm~~~--r- skin friction

Long - termskin friction

1000

C'J 900E

:z:.::tt:.

800~

c: 7000

u

.;: 600c:

.::tt:.ln

500

400

300

200

100

Test series 5Anchor Lv =3,0 maverage do =155mm

1 1 1 1 1 l--l

!~~~M--iStrain gauges . la Lv =3,Om ----~

Fig. 6. - Distribution of skin friction in the soiljgroutinterface along bond length (grouted body in dense sand).

\\

'\..-10'(Minutes)\

\\\

\

Test series 5

Anchor Lv =3,0 maverage do=155mm

1000

:z900~

'-

-c 800c0

~ 700Cf)

cQ)

600-t--

500

400

300

100

200

lengthL-- ---=~~~~

Il 1 1 1 1 ~

%f(~~a~.Slroin gougesJ.t-- Lv =3,01:1--~I

Fig. 7. - Variation of skin friction along bond lengthovera period of 300 minutes (load of 785 kN was keptconstant).

For the load of 785 kN this balancing over a periodof 300 minutes is illustrated in the three-dimensionalplot of Fig. 7. The decrease and increase of skinfriction along the bond length is shown for severalpoints of time and demonstrated with the aid of shadedareas. 1t is pointed out that for any particul~r time theshaded area showing friction decrease in the front partis equal to the shaded area showing friction increasein the rear part.

shown in Fig. 5 the forces in the tendon increase notonly when the applied test load at the anchor head isincreased (here 5 loading steps), but also when theload iskept constant for a period of time (in theexample of Fig. 5 three loading steps were keptconstant for about one day).

The difference in the values of forces measured attwo adjoining points divided by the circumferencearea of the grouted body (grout-soil interface) givesthe value of «skin friction». these calculated valuesare shown in Fig. 6. Obviously there are maximumskin friction values, the location of this maximummoves from the front -part of bond length towards theanchor end when the test load is increased. Thereason is that the elastic deformations of the steeltendon cause progressive displacements in the grout/soil interface. This progressive displacement causesthe shear resistance of the dense sand to shift beyondthe peak -point into the region of lower residualshear values.

l t is worth noting that at each loading step thesame maximum friction value is reached for a shorttime (maximum «short-term skin friction») and thatthis value tapers off with time until a certain «long­term skin friction» value is not exceeded at any pointalong the bond length. As the applied test load iskept constant during one loading step, a decrease ofskin friction in the front part of bond length willresult in a corresponding increase of skin friction inthe rear part. This kind of balancing could not beattained for the last loading step 'of 850 kN, so thatfailure occurred within 10 minutes.

Increase

E­z~

co

c.~

-0o:>

Oecrease Test seriesAnchor Lv = 3,0 maverage do =155 mm

rime

[minutes]

95

When evaluating skin friction 'ts in the grout/soilinterface from the bond stresses 'tb in the steel bar/grout interface (Fig. 8) it must be taken into consi­deration that the bond-to-ground length La is longerthan the bond length of tendon Lv. In the front partof the grouted body there is no steel/grout bond (dueto the plastic tube), so that the forces resulting fromskin friction 'ts in this area are transmitted furtherback and cause a concentration of bond stresses 'tb inthe front part of bond length of tendon. These veryhigh bond stresses which are obtained by strain gaugemeasurements are schematically shown in Fig. 8, wherethe existing 'tb is converted to an equivalent skinfriction 't's through the factor db/do, ('t/ == 'tb . db/do).

The actual skin friction is achieved by equalizing theequivalent value over the entire length including thefront part of the grouted body (see area d 'F) in orderto get the actual skin friction 'ts.

In the case of dense sands the limit values of skinfriction, max 'ts ' are effective along a .relatively shortlength. For 'short anchors this length will correspondalmost with the whole bond-to-ground length (Fig. 8 a).For long anchors this length of max 'ts is only- a partof the bond-to-ground length. The location of thispart shifts towards the anchor rear when the test loadis increased (distribution of skin friction near failureload is shown in Fig 8 b). Assuming that the limitvalue max 'ts is identical for different bond-to-groundlengths, the mean values (mean 'ts) for long anchorsare smaller than for short anchors. This has alreadybeen presented in the chart of Fig. 4 in terms of carryingcapacity versus bond length. '

For the last loading step before failure load wasreached Fig. 9 shows. the «long-term skin friction»values of all test anchors in sand, which were equippedwith strain gauges (diameter of grouted body being 9to 12 cm).

. For dense and very dense sand the skin frictionvalues obtai.ned experimentally fit very well with thequalitative distribution of skin friction of Fig. 8, therebyconfirming the assumption made. The limit values,max 't5 of shorter anchors (Lv == 2 m) ho\vever are likelyto exceed the corresponding values of longer anchors(Lv == 4 m). The difference may be partly traced backto the larger radial confining pressures in the frontpart of anchors.

In loose and medium dense sand the skin frictionis found to be more or less constant along the wholebond-to-ground length. This corresponds with thestress-strain-behaviour of the sand for these densities.

The decisive influence of soil density is obvious inthese tests when for example in the case of long anchorsone compares the limit values of skin friction for 100seand medium dense sand (about 150 and 300 kN/m2

respectively) with the limit values for very dense sand(about 800 kN/m2).

These high values of skin friction are mainly theresult of an interlocking or wedging effect due to thedilatation of soil (E. Wernick, 1977). The peak valuesof up to 1 300 kN/m2 do not represent the actualskin friction but the equivalent skin friction as alreadyexplained in Fig. 8.

Penetration Tests and, Carrying Capacityof Anchors

As the density of non-cohesive soils is in currentpractice very often indirectly determined by penetro­meter tests, it was decided to plot a diagram showing

96

Q) Short Anchor

,

Grouted bodyBar

~ ft db do

•

T's skin fr ict ion in the 9r0 ut / soi 1 inter face

Tb bond stress in the steel bar / groùt interface

Fig. 8. - Qualitative distribution of skin friction andbond stresses for short and long anchors in dense groundat ultimate load.

Soi 1densi ty Bond lengthLv

.....0-- Very dense 2.0 m-e- 4.5 m

-0-- Dense 3.0m1300

Medium l.Om-0-

1200 -.- dense 4.5 m-IJ.- Loose 2,0 m

1100 4.5 mE 1000~~

L~ 800c:

.; 700

..t: 600

.~~

V> 500

400 0 Medium.....---, dense

300 ·~.~!..-:=-::::'~·-r

'" max Ts~meanTs

~~~·~~~t~~ • moxTs",meonTs

1 . 1.0 2.0 3.0 4.0 Length [m]

=rm__?::ŒZ::@_~'0055i5555550J

Lv=2.0m Lv=3.0m Lv=4.5m

Fig. 9. - Distribution of long-term skin friction 'ts atultimate load in relation to bond length Lv and soildensity.Grouted bodies (do = 9.1 - 12.6 cm) in gravelly sand.

..

Fig. 10. - Relationship between carrying capacity, bondlength of anchors and dynamic penetration resistance intwo types of non-cohesive soils.

carrying capacity of anchors in relation· ta penetrationresistance. Fig. 10 shows the results of the 30 testanchors supplemented by additional results of otherin situ tests (fundamental tests of different anchorsystems). This chart may be used for a rough. esti­mation of the carrying capacity of anchors which areproperly installed. 1t must be emphasized however,that only for the 30 anchors in the test' pit bathdynamic. penetration tests (50 kg hammer) and stand­ard penetration tests (SPT) have 'been carried out.For the other in situ tests only dynamic penetrationtests were used. The chart may be adjusted andextended for sandy gravel soils depending upon theresults of additional future test data.

Nu mberof blows

Standard penetration test40 50 60 70 80 [N /30 cm J

20 30 40 [N/10cm]penetration test (50 kg hammer)

i-i--T--l-JiDf1 1 .~[ Sandy

illlllili gravel

+-----+---+------f-I-+--t-W'Y-+-+-+-t~~P<t__'k_l L0:: '3 mI_~

Gravellysand

~ 2000z:~--- 1800~

1:' 1600ua

1400Cl.au0) 1200c:

~1000L-

au

"0 800a~

~600

aE 400~:::>

200

0

CONCLUSIONS

Ta investigate the important influence of the densityof non-cohesive soils on the carrying behaviour ofanchors it was for the first time that full scale fieldtests were performed in one and the same sail com­pacted ta different densities. Under these controlledconditions the exact distribution of skin friction alongthe bond length could be calculated from strain gaugemeasurements.

On the basis of the results of the 5 series of testscomprising 30 anchors, the original design chart of1975, showing carrying capacity versus bond lengthfor different sail conditions, has now been cample­mented. Furthermore a diagram is presented fromwhich it is possible ta estimate carrying capacity ofanchors with different bond lengths from ': the numberof blows of standard penetration tests (SPT) anddynamic penetration tests. When using one of thesecharts it must be borne in mind that certain fluctua-

tians in test results are possible due ta the inhomo­geneity existing in the sail at site, even when theanchors have been properly installed.

The different shapes of distribution of skin frictionwhich were derived through measurements help tagive an explanation for the influence of bond lengthof anchors and density of soils on the carrying capacityas shawn in the design chart (Fig. 4).

The different shapes of distribution of skin frictionof skin friction along the bond length with increasingload paved the way for the inclusion of valid assump­tians in the calculation of carrying capacity in termsof sail constants.

In addition the variation of skin friction with res­pect ta time was measured for several loading steps.The results (shown for only one anchor in Fig. 7)will provide a basis for anticipating the long-termbehaviour of anchors.

ACKNOWLEDGEMENTS

The tests were performed at the Institute of SailMechanics and Foundation Engineering, Techniçal Uni­versity, MunichjGermany, guided by Prof. Dr-Ing. R.Jelinek. The authors are grateful to Bundesministe­rium für Raumordnung, Bauwesen und Stadtebau, taBundesministerium für Verkehr, ta Innenministeriumdes Landes Nordrhein-Westfalen and to Hauptverbandder Deutschen Bauindustrie for financial support:

REFERENCES

Deutsche Normen. - « Verpressanker für vorüber­gehende Zwecke im Lockergestein : Bemessung,Ausführung und Prüfung ». DIN 4125, Blatt 1,Beuth Verlag, Berlin (1972).

OSTERMAYER (H.). - «Construction, Carrying Beha­viaur and Creep Characteristics of Ground An­chars», Proc. Conf. on Diaphragm Walls and An­chorages, London 1974, Institution of CivilEngineers, London, pp. 141-151 (1975).

WERNICK (E.). - «Stresses and Strains on the Surfaceof Anchors», 9th Int. Conf. SMFE, SpecialtySession 4: Ground Anchors (1977).

WROTH (C.P.). - Discussion on Papers 18-21(Report). Proc. Conf. on Diaphragm Walls andAnchorages, . London 1974, Institution of CivilEngineers, London, pp. 165-169 (1975).

97