articulo 1

Bond strength of concrete plugs embedded intubular steel piles under cyclic loading

Abolghasem Nezamian, Riadh Al-Mahaidi, and Paul Grundy

Abstract: Investigation of the load transfer of concrete plugs to tubular steel piles subjected to tension and compres-sion and cyclic loading has been conducted at Monash University over the past 3 years. The work presented in this pa-per reports on the results of the combination of pull-out, push-out, and cyclic loading tests carried out on 15 steel tubespecimens filled partially with reinforced concrete with variable lengths of embedment. The pull-out force was appliedthrough steel reinforcing bars embedded in the concrete plug, and push-out forces were applied through a thick top cir-cular plate on the top of the concrete plug. Test results included the cyclic loading, ultimate pull-out and push-outforces, slip of concrete plugs, and longitudinal and hoop strains along the piles for some specimens. The tests clearlyshowed that average bond strength significantly exceeds expectations and is higher than the results of previous investi-gations using plugs without reinforcement. The test results also indicated that cyclic loading tests reduced the bondstrength due to the accumulation of damage to the plugpile interface. The push-out and pull-out tests conducted undersymmetric cyclic loading demonstrated that slip between the concrete plug and the steel tube increased with repeatedloading, and the rate of slip growth increased with an increase in the peak load.

Key words: tubular steel pile, reinforced concrete plug, bond, cyclic loading.

Rsum : Le transfert de charge des bouchons de bton aux pieux tubulaires en acier soumis des charges de tension,de compression et cycliques a t tudi lUniversit Monash au cours des trois dernires annes. Le travail prsentdans cet article examine les rsultats de la combinaison des essais darrachement, de contrainte par expulsion et dechargement cyclique effectus sur 15 tubes dacier partiellement remplis de bton arm dont les longueurs dencastre-ment varient. La force darrachement a t applique des tiges darmature en acier encastres dans le bouchon de b-ton, et les contraintes par expulsion ont t appliques sur une plaque circulaire paisse reposant sur le bouchon debton. Les rsultats des essais comprennent le chargement cyclique, les forces limites darrachement et dexpulsion, leglissement des bouchons de bton et les contraintes longitudinales et circonfrentielles le long les pieux pour certainschantillons. Les essais montrent clairement que la rsistance moyenne du lien dpasse de manire significative les at-tentes et elle est suprieure aux rsultats des tudes antrieures utilizant des bouchons sans renforcement. Les rsultatsdessais montrent galement que les essais de chargement cyclique ont rduit la rsistance du lien en raison de laccu-mulation de dommages linterface bouchonpieu. Les essais de contrainte par expulsion et darrachements effectussous des chargements cycliques symtriques ont dmontr que le glissement entre le bouchon de bton et le tube enacier augmente avec la rptition des charges et que le taux de glissement augmente avec la charge de pointe.

Mots cls : pieux tubulaires en acier, bouchon de bton arm, lien, charge cyclique.

[Traduit par la Rdaction] Nezamian et al. 125

Introduction

The legs of platforms of many offshore and coastal struc-tures are usually founded on tubular steel piles through rein-forced concrete pile caps. Wave, wind, and earthquake loadstend to induce compressive and uplift forces in the legs thatin turn subject the piles to cyclic compression and tensionloading. This transfer of forces takes place through a con-crete plug embedded in the top of the steel pile. The resis-

tance of the embedded concrete plug to slip is due to thesteelconcrete bond stress along the plug length (see Fig. 1).

In recent years, many investigations have resulted in anal-ysis and design rules for concrete-filled steel columns basedon experimental models of steel tubes filled with plain con-crete and tested in compression. The composite action insuch columns is due to the bond strength and mechanical in-terlock. Investigations have shown that these mechanismsdepend on the surface roughness of the steel tube and the

Can. J. Civ. Eng. 33: 111125 (2006) doi:10.1139/L05-091 2006 NRC Canada

111

Received 19 May 2004. Revision accepted 12 September 2005. Published on the NRC Research Press Web site at http://cjce.nrc.caon 11 February 2006.

A. Nezamian.1 School of Civil and Chemical Engineering, RMIT University, City Campus, Melbourne, VIC 3001, Australia.R. Al-Mahaidi and P. Grundy. Department of Civil Engineering, Monash University, Wellington Road, Clayton, VIC 3800,Australia.

Written discussion of this article is welcomed and will be received by the Editor until 30 June 2006.1Corresponding author (e-mail: [email protected]).

shape variation of the cross section of the steel tube. Valuesof bond strength reported in the literature varied from 0.2 to1.0 MPa. In contrast, a literature search has revealed that noinvestigations have been carried out on steel tubes filledwith reinforced concrete.

The earliest experimental study of the bond strength ofconcrete-filled steel tubes was carried out by Virdi andDowling (1975). A number of parameters were varied tostudy their effects on the bond strength between concreteand steel. It was concluded that the resistance to the push-out test in filled tubes derives primarily from the interlock-ing of concrete in two types of imperfections in steel. Thefirst relates to the surface roughness of the steel, and the sec-ond to variation in the shape of the cross section, away fromthe ideal cylindrical surface. The interlocking of concrete inthe surface roughness of steel (i.e., micro-locking) contrib-utes a useful component of the ultimate bond strengthrelated to the initially stiff region of the load deflection char-acteristics. Virdi and Dowling proposed the bond strength of1 MPa for design. An extensive investigation of the push-outstrength of concrete-filled tubular members was undertakenby Shakir-Khalil (1991, 1993a, 1993b). The main parame-ters studied were the shape of the tube, interface length, in-terface condition, and use of mechanical connectors. It wasalso noted that, in agreement with Virdi and Dowling, speci-men length was not a significant factor in the bond strength.Further, it was shown that the circular hollow section (CHS)tube had a superior load-carrying capacity compared withthe rectangular hollow section (RHS) tube. The resistance ofthe circular section is enhanced due to the much stiffer con-finement of the concrete during slip as it rides over the aspe-rities and irregularities of profile of the steel tube. A bondstrength of 0.4 MPa was proposed for design purposes. Thetest results and analytical study of Roeder et al. (1999) in anexamination of the bond stress capacity of circular concrete-

filled tube (CFT) members indicated that the maximum av-erage bond stress capacity is somewhat smaller with longercolumn lengths and larger diameter to thickness (D/t) ratiosand diameters due to the lack of the stiffness to enforce thebenefits of irregularity in the cross section.

Test results from the aforementioned studies showed thatthe average bond stress for rectangular tubes was approxi-mately 70% smaller than that for circular tubes and indi-cated that the influences of the steel tube aspect ratio (D/t)and the ratio of concrete core length to depth (L/D) on thebond strength are not completely understood.

The bond resistance of reinforced concrete plugs embed-ded in tubular steel piles under pull-out and push-out load-ings has been investigated by Nezamian et al. (2001, 2003,2002) and Al-Mahaidi et al. (1999). The pull-out bondstrength tested in specimens having a concrete plugembedment length to tube inner diameter ratio of L/Di = 1ranged from 4.3 to 6.2 MPa. It was not possible to determinethe pull-out bond strength for specimens with L/Di > 1 dueto yielding and rupture of the embedded steel bars, whichpreceded the development of full bond strength. The push-out strength of reinforced concrete plugs embedded in tubu-lar steel piles revealed capacities higher than those reportedby others. This was attributed, in part, to the presence of re-inforcement in the plug and smaller concrete plugembedment length to tube inner diameter L/Di comparedwith the other reported investigations. Bond strengths offrom 2.0 to 7.3 MPa were achieved.

Current code provisions

The provisions of British Standard BS5400, Steel, con-crete and composite bridges (BSI 1979), recommend thatshear connectors be provided where the shear stresses at thesteelconcrete interface, due to the design ultimate loads, ex-

2006 NRC Canada

112 Can. J. Civ. Eng. Vol. 33, 2006

Reinforcingbars

SECTION A - A

R10 rings

300

175

30 RL

A

A

Reinforcing bars

Conc

rete

plug

Bottom ofDolphin

Rake1:3Ra

ke1:

3

RL

Face ofmooringdolphin

with capPour concrete

mm

Fig. 1. Typical connection between a steel pile and a concrete pile cap. RL, reduced level.

ceed 0.4 MPa for concrete-filled steel sections. According toEurocode 4 (ECS 1994), the design shear strength due tobond and friction for a concrete-filled hollow section shouldbe taken as 0.4 MPa.

Both codes recommend the same value for bondshearstrength between concrete and steel regardless of concreteproperties, length of concrete embedment, shape of steelhollow section, roughness of internal surface, and loadingregime.

Experimental program for cyclic tests

A total of 15 specimens were constructed and tested forthe purpose of investigating the effect of cyclic loading onthe bond strength of concrete plugs embedded in tubularsteel piles. Only one circular steel tube size and concretestrength were used. The structural steel tube of grade 250with nominal yield strength of 250 MPa and ultimate tensilestrength of 350 MPa was used (SAA 1983). The steel tubeshad an outside average diameter Di of 237 mm and an aver-age wall thickness of 11.5 mm. The steel tubes were cut to alength of 600 mm. The inner surfaces of the steel tubes werescrubbed with a wire brush to remove any excess corrosion,dirt, or other material. The formwork was fabricated andplaced at the bottom of the specimens considering the differ-ent depths of concrete plugs. The specified reinforcing cageswere placed in the specimens and tack welded in position toensure the cage would not move during pouring of the con-crete. The reinforcing bars in all specimens consisted of sixY24 bars (24 mm diameter deformed bar with nominal yieldstrength of 400 MPa). The reinforcement was 5% of thegross area of the concrete plug based on the threaded area ofthe bars (Fig. 2).

One cubic metre of 32 MPa concrete with a slump of 80100 mm was ordered to manufacture the specimens. Generalpurpose (GP) Portland cement was used (SAA 1997), with awater to cement ratio of 0.5 and cement to aggregate ratio of1/7. This was in accordance with SAA (1993). The maxi-mum aggregate size was limited to 20 mm. The result of theslump test on the concrete batch on arrival revealed a slumpof 100 mm, and cylinder compressive strength test results in-dicated 40 MPa at an age of 38 days. The concrete was care-fully placed and then vibrated into each specimen to ensuresatisfactory compaction. The specifications of the con-structed specimens are summarized in Table 1.

The top surface was prepared with plaster to provide alevel surface and ensure even distribution of the compressiveforces. The supporting timber formwork was removed, andthe base plates were then welded to each sample. This pro-cess involved placing the samples in the test rig to ensure thereinforcement bars were correctly aligned with the testingrig. The base plate was then tack welded, removed from therig, and then fully welded with three passes.

As shown in Fig. 3, the loads were applied through a thickcircular steel plate placed on the concrete surface of the plugand bolted to the reinforcing bars. An Instron servo-controlledactuator of 1000 kN dynamic capacity and 1250 kN static ca-pacity was used to load the specimens on pull-out, push-out,and cycling loading tests. This gave a comfortable margin ofcapacity over the anticipated ultimate pull-out force of1000 kN (to prevent the failure of threaded bars). An Instron

8500 controller, which allowed load and displacement con-trol and had programmable trapezoidal control waveforms,which were utilized for the cyclic loading tests, controlledthe actuator. Displacement control was used for the pull-outand push-out tests, and load control (with displacement limitset) was used for all cyclic tests.

The monotonic tests (pull out and push out) were con-ducted at a displacement rate of 0.015 mm/s. The time takento reach the peak load was varied in the order of 510 min.

The cyclic tests were conducted with a symmetric triangu-lar cyclic loading with no holding time at the peak load.Wind and wave loading can be simulated using this loadprotocol. For every cyclic test, the loading was repeated fora predetermined number of cycles, with data continuouslyrecorded. The load range was then increased, and the newloading was repeated, usually for the same number of cycles.For the cyclic tests, the load versus time function was trian-gular. A typical function is shown in Fig. 4.

The type of tests on each specimen and the loading ratesand number of cycles per load range are summarized in Ta-bles 2 and 3.

A string linear variable displacement transducer (LVDT)was used to measure the relative movement (slip) betweenthe concrete core and the steel tube. Most of the test speci-mens were strain-gauged along the outer surface of the steeltube within the length of the concrete plug. Both longitudi-nal and hoop direction gauges were used. The purpose ofthese gauges was to determine from the axial stress in thesteel tube the distribution of shear stress along the contactarea. Additional strain gauges were used on the oppositeside of the tubes to establish whether the loading arrange-ment introduced significant eccentricity in the specimen.Strain gauge arrangements are shown in Fig. 5.

It was decided to investigate the effect of cyclic loadingson bond strength in two stages of experimental work. Thefirst stage focused on determination of the effect of the ini-tial cyclic loading on the ultimate pull-out strength. The sec-

2006 NRC Canada

Nezamian et al. 113

244 mm OD Steel Tube11 mm Thickness

6 24 mm deformed bars40 mm

Var

iabl

eCo

ncr

ete

Embe

dmen

t

Leng

th50

6 mm round barsRing Reinforcements

Concrete40 MPa Compressive Strength

6 mm plain barsRing Reinforcements

600

mm

mm

Fig. 2. Typical test specimen.

ond stage was aimed mainly at studying failure of thespecimens subjected to cyclic loadings. Determination of theeffect of shrinkage on bond strength is also explored in thisset of experiments.

Test results of stage 1

The aim of stage 1 was to evaluate the effect of initial cy-clic loading on ultimate pull-out strength. Three specimenseach of two different concrete plug lengths (1.0D and 1.5D)were tested. The first specimen of each plug length groupwas tested for static tension capacity to enable the assess-ment of cyclic load effects. This was followed by a push-outtest. The other two specimens were then initially subjectedto two sets of 10 cycles of 150 and 250 kN for series S1.0Dspecimens and 250 and 400 kN for series S1.5D specimens.

The magnitude of the cyclic load was decided based on ap-proximately 0.25 and 0.40 of the static ultimate strength ofthe specimens at 2 mm slip. This was followed by mono-tonic pull-out tests. A total of six successful tests wereconducted, including two monotonic tests to determine thepull-out strength of the concrete plug and four pull-out testswith initial symmetric cyclic loadings. Table 2 lists the peakloads achieved and corresponding average bond strengths.The slip values at peak load, initial type of cyclic loadingtest, and age of the concrete on the test date are also tabu-lated. Average bond strength was calculated by dividing theultimate pull-out or push-out forces by the contact area ofsteelconcrete.

Specimen S1.0D-1 failed at an ultimate pull-out strengthof 665 kN, followed by a push-out test, which resulted in anultimate push-out capacity of 525 kN. Specimen S1.5D-1achieved a pull-out strength of 1000 kN at a slip of 1.7 mm.This was followed by a push-out test, which resulted in anultimate push-out capacity of 1000 kN at a slip of 1.5 mm.

Specimens S1.0D-2 and S1.0D-3 were then initially sub-jected to 10 symmetric cycles of 150 kN followed byanother 10 symmetric cycles of 250 in tension and com-pression. This was followed by pull-out tests, which resultedin ultimate loads of 711 and 405 kN for specimens S1.0D-2and S1.0D-3, respectively. Specimen S1.5D-2 was initiallysubjected to 10 symmetric cycles of 250 kN. This was fol-lowed by a pull-out test, which resulted in an ultimate loadof 500 kN. Specimen S1.5D-3 was initially subjected to 10symmetric cycles of 250 kN followed by another 10 sym-

2006 NRC Canada

114 Can. J. Civ. Eng. Vol. 33, 2006

SpecimenNo.

Tube length,Lp (mm)

Tube internaldiameter, Di (mm)

Tube wallthickness, t (mm)

Concrete pluglength, Lconc (mm) Lconc/Di Di/t

S1.0D-1 600 222 11 222 1.00 20.2S1.0D-2 600 222 11 222 1.00 20.2S1.0D-3 600 222 11 222 1.00 20.2S1.25D-1 600 222 11 278 1.25 20.2S1.25D-2 600 222 11 278 1.25 20.2S1.25D-3 600 222 11 278 1.25 20.2S1.5D-1 600 222 11 333 1.50 20.2S1.5D-2 600 222 11 333 1.50 20.2S1.5D-3 600 222 11 333 1.50 20.2S1.75D-1 600 222 11 389 1.75 20.2S1.75D-2 600 222 11 389 1.75 20.2S1.75D-3 600 222 11 389 1.75 20.2S2.0D-1 600 218 13 444 2.00 16.8S2.0D-2 600 218 13 444 2.00 16.8S2.0D-3 600 218 13 444 2.00 16.8

Table 1. Specifications of constructed specimens.

Instron servo-controlled actuator

Concrete plug in steel tube

Support stand

Strong Floor

Load CellLVDTLoad from the load cell

Fig. 3. Cyclic loading test arrangements. LVDT, linear variable displacement transducer.

-250-200-150-100

-500

50100150200250

0 10 20 30 40 50 60 70 80Time (s)

Fo

rce

(kN

)

Fig. 4. Typical load (force) versus time function for specimenS1.0D-2.

metric cycles of 400 kN. The specimen failed at the end ofthe cyclic loading test.

Average bond strengths of 4.3 MPa were achieved forstatic pull-out tests and 2.8 MPa for pull-out tests with a cy-clic loading effect. The test results indicated that pre-cyclicloading tests reduced the bond strength due to the prior dam-age to the plugpile interface.

Loadslip responseFigure 6 shows the loadslip response of specimen series

S1.0D and S1.5D. In pull out, specimen S1.0D-1 exhibited adecay in shearbond load as slip increased after peaking at aslip of 1.01.5 mm. This is the expected result associatedwith a plug in a properly circular straight pile. In push out,the specimen exhibited some slip at an initial load of300 kN. This is attributed to reversal of permanent slip cre-ated by the prior pull-out test. The initial slip of 2.0 mm isbelieved to be recovery of permanent pull-out slip. The spec-imen then exhibited a gradual increase in load transfer asslip increased after reaching an applied load of 450 kN at aslip of 1.0 mm. A possible explanation for this is that theinitial pull-out test prestressed the interface. Macro interlockeffects are then created when the slip becomes significant.These caused an increase in the contact stress between thesteel tube and concrete plug, which increases the frictionalresistance. Specimens S1.0D-2 and S1.0D-3 exhibited a typ-ical loadslip response in tension after initial cyclic loadingcharacterized by a gradual decrease in load transfer as slipincreased after peaking at a slip of 1.0 mm. The loadslip ofspecimen S1.0D-2 indicates that the initial cyclic loadingmay not have a significant effect on the loadslip behaviourand the pull-out strength of the specimen. On the contrary,the loadslip response of specimen S1.0D-3 shows that theinitial cyclic loading reduced the interface stiffness andshear transfer between the concrete and the steel tube. Thiswas due to the prior damage to the plugpile interface.

As can be seen in Fig. 6, the loadslip response of speci-men S1.5D-1 shows typical behaviour in the pull-out testwith a peak load of 1000 kN at a slip of 1.7 mm. The test

procedure stopped at 1000 kN, as the specimen reached thelimitation of the test instrumentation. The loadslip of thespecimen in push out shows gradual reversal slip to a loadlevel of 300 kN. This is attributed to the reversal of perma-nent slip created by the prior pull-out test. The specimenthen reached a load level of 1000 kN at a slip of 1.0 mm.Specimen S1.5D-2 exhibited a typical loadslip response intension after initial cyclic loading. It was characterized by agradual decrease in load transfer as slip increased after peak-ing at a slip of 1.0 mm.

Specimen S1.5D-3 failed at the end of the second 10 cy-cles. The loadslip response of the specimen indicates thatthe initial cyclic loading reduced the ultimate strength of thespecimen to the level of the second cyclic load of 400 kN.This was due to the significant damage to the plugpile in-terface. The post-failure response shows an almost constantshear transfer in the pull-out test after cyclic loading. Thisbehaviour continued until the slip values reached 9.2 mm.

The loadslip response of the specimens indicated that theloadslip curves of the pull-out test with a cyclic effect aresimilar to those obtained for monotonic static tests. Theshifting between these two curves in the ordinate load axis isdue to the different cyclic loading rate and concrete pluglength. The effects of cycling rate and the cyclic reductionfactor are discussed in the following sections.

Test results of stage 2

Previous test results (Nezamian et al. 2001, 2002) indi-cated that the ultimate push-out strength of the specimen ismost likely less than the ultimate pull-out strength. Stage 2of the experimental work was then planned to evaluate thefailure of the specimens, subjected to cyclic loading basedon the ultimate push-out strength of the specimen. Threespecimens each of three different concrete plug lengths(1.25D, 1.75D, and 2.0D) were then tested. The first speci-men of each plug length group was tested for static compres-sion capacity to enable the assessment of cyclic load effects,which were often followed by pull-out tests. The other two

2006 NRC Canada

Nezamian et al. 115

SpecimenNo. Type of test

Max. load(kN)

Hold time(min)

Avg. bondstrength (MPa)

Max. slip(mm)

Cycle time(min)

No. ofcycles

Concreteage (days)

S1.0D-1 Pull out 665 15 4.20 2.3 32Push out 525 19 3.31 7.5 32

S1.0D-2 Cyclic loading 150 40 0.94 0.6 4 10 26Cyclic loading 250 40 1.58 1.0 4 10 26Pull out 711 24 4.49 12.2 26


S1.5D-1 Pull outa 1000 17 4.30 1.7 29Push outa 1000 18 4.30 1.5 29

S1.5D-2 Cyclic loading 230 40 0.99 0.2 4 10 28Pull out 500 8 2.15 1.8 28Push out 400 24 1.72 6.8 28


aThe maximum load was limited by the capacity of the loading system and does not correspond to ultimate strength response.

Table 2. Summary of the tests conducted in stage 1.

specimens were initially subjected to a variety of differentcyclic loadings. This was followed by monotonic pull-outtests. It was decided that the magnitude of cyclic loadingwould start from 0.6 of the ultimate strength of the specimenand would then be reduced to 0.3 of the ultimate strengthdue to early failure of specimens S1.25D-1 and S1.25D-2.This stage of the experiment took place about 2 years afterconstruction of the specimens. Therefore, determination ofthe effect of shrinkage on bond strength of concrete plugscan also be evaluated with this set of test data. A total of 12successful tests were conducted, including three monotonictests to determine the push-out strength of the concrete plug,six tests with symmetric cyclic loading of the specimens,and three tests with extra pull-out tests to evaluate the effectof cyclic loading. Table 3 lists the peak loads achieved andcorresponding average bond strengths. The slip values atpeak load, initial type of cyclic loading test, and age of con-crete on the date of the test are also tabulated.

A total of 35 tests were carried out on nine specimens.The pull-out bond strength varied between 3.3 and 1.2 MPa,

with an average of 2.3 MPa for seven pull-out tests. Thepush-out bond strength varied between 3.3 and 0.9 MPa,with an average of 2.3 MPa for three push-out tests. The cy-clic bond strength varied between 2.0 and 0.9 MPa, with anaverage of 1.3 MPa for six cyclic loading tests. The test re-sults indicated that cyclic bond strength is lower than ulti-mate static pull-out or push-out bond strengths. This is dueto the incremental damage to the plugpile interface. Theshrinkage cracks were observed at the concretesteel inter-face in most of the specimens prior to testing. It is then con-cluded that shrinkage can be very detrimental to bond stresscapacity and reduces the bond strength with time.

Loadslip responseFigures 7, 8, and 9 show the loadslip response of speci-

men series S1.25D, S1.75D, and S2.0D, respectively. Instatic push-out tests, specimen S1.25D-1 exhibited a decayshear transfer after peaking at a slip of 1.01.5 mm. This isthe expected result associated with a plug in a properly cir-cular straight pile. This fretting of the cement matrix on the

2006 NRC Canada

116 Can. J. Civ. Eng. Vol. 33, 2006

SpecimenNo. Type of test

Max.load (kN)

Hold time(min)

Avg. bondstrength (MPa)

Max. slip(mm)

Cycle time(min)

No. ofcycles

Concreteage (days)

S1.25D-1 Push out 443 51 2.29 2.75 403Pull out 460 36 2.38 24.50 403


S1.25D-3 Cyclic loading 245 7 1.26 8.02 4 1.75 409Pull out 540 12 2.79 20.90 409

S1.75D-1 Push out 395 7 1.45 7.48 410Pull out 330 12 1.21 12.00 410

S1.75D-2 Cyclic loading 100 20 0.37 0.12 4 5 716Cyclic loading 125 20 0.46 0.17 4 5 716Cyclic loading 150 20 0.55 0.23 4 5 716Cyclic loading 175 20 0.65 0.30 4 5 716Cyclic loading 200 20 0.74 0.37 4 5 716Cyclic loading 225 20 0.83 0.56 4 5 716Cyclic loading 250 20 0.93 0.84 4 5 716Cyclic loading 275 20 1.01 1.67 4 5 716Cyclic loading 300 16 1.11 14.80 4 4 716

S1.75D-3 Pull out 431 9 1.59 1.37 717Cyclic loading 150 20 0.55 0.03 4 5 717Cyclic loading 200 20 0.74 2.16 4 5 717Cyclic loading 225 20 0.83 7.72 4 5 717Cyclic loading 250 14 0.93 18.60 4 3.50 717

S2.0D-1 Push outa 1000 60 3.29 1.89 703Pull outa 1000 40 3.29 1.24 703Cyclic loading 500 44 1.64 1.75 4 11 703Cyclic loading 550 40 1.81 2.95 4 7 703Cyclic loading 600 5 1.97 2.97 4 1.25 703Cyclic loading 600 5 1.97 3.25 4 1.25 703Cyclic loading 600 30 1.97 16.00 4 7.25 703

S2.0D-2 Pull out 479 2 1.57 16.70 710S2.0D-3 Cyclic loading 200 20 0.66 0.54 4 5 713

Cyclic loading 250 20 0.82 1.43 4 5 713Cyclic loading 300 20 0.99 2.97 4 5 713Cyclic loading 350 5 1.15 30.80 4 1 713

aThe maximum load was limited by the capacity of the loading system and does not correspond to ultimate strength response.

Table 3. Summary of the tests conducted in stage 2.

steel surface has a powdering effect, removing the interlockwith asperities on the steel surface and lowering the effec-tive coefficient of friction. In the following pull-out test, thespecimen experienced a reversal slip at a load level of300 kN. The slip is believed to be a recovery of permanentpush-out slip. The specimen then showed a gradual increasein load transfer as slip increased. This is due to the initialpush-out test, which consolidated the concrete in the steeltube. Macro interlock effects were then created when theslip became significant. These raised the contact stress be-tween the steel tube and the concrete plug, which increasedthe friction resistance.

Specimen S1.25D-2 reached its ultimate strength at theend of initial symmetric cyclic loading of 310 kN. Thespecimen then showed a smooth decay shear transfer afterpeaking at a slip of 8 mm in the subsequent pull-out test.This was due to significant damage to the plugpile interfaceduring the initial cyclic loading. Specimen S1.25D-3 failedat the second cycle of the first cyclic loading range afterreaching a slip of 8 mm. In the following pull-out test, thespecimen exhibited a gradual increase in load transfer as slip

increased. The loadslip response of specimen S1.75D-1shows that the push-out load dipped after an early peak atabout 2 mm slip but then recovered, indicating some macroeffects discussed previously. In the pull-out test that fol-lowed, the plug locked into the steel tube with no reversalslip before peaking at a pull-out load of 330 kN. The sheartransfer then dipped down but recovered partially after re-versal of the permanent push-out slip. The locking of theplug was due to a mechanical interlock mechanism. Speci-mens S1.75D-2, S1.75-D3, and S2.0D-3 exhibited pinchedhysteretic behaviour and completely failed in the cyclicloading.

The push-out loadslip curve of specimen S2.0D-1 exhib-ited a nearly bilinear response prior to peak load (set limita-tion of the test machine). The change of slope of the loadslip curve during loading was assumed to commence withthe breaking of chemical adhesion (nonslip mechanism) andactivation of the mechanical interlock mechanism (verysmall slip mechanism). In the following pull-out test, thespecimen experienced a reversal slip at a load level of 700 kN.The slip is believed to be a recovery of permanent push-out

2006 NRC Canada

Nezamian et al. 117

One longitudinal gaugeOne transverse gauge

4040

277.

560

0

AA

333

75

600

32

600

Stee

lTu

be

Spec

imen

S1.0

D

60

Stee

lTu

be

Spec

imen

S1.5

D33

Stee

lTu

be50

A

222


2x1 longitudinal gaugeOne each side

3040

40A

4040

Con

cret

ePlu

g

A 7560

30


Con

cret

ePlu

g 37.5

4040

4040

4044

600

444

Con

cret

ePlu

g

Stee

lTu

be

Con

cret

ePlu

g

4040

Stee

lTu

be

600

388.

5

Spec

imen

S1.2

5D50

A

Spec

imen

S1.7

5D

4040

50

A

50

A

2x1 longitudinal gaugeOne each sideSpecimen

S2.0D

5040

40

A40

40

50

A


Con

cret

ePlu

g

4040

68.5

4040

SECTION A-A

120


4040

4040

40

120




Fig. 5. Strain gauge arrangements. All measurements in millimetres.

slip. The specimen then showed a load transfer increase asslip increased before reaching 1000 kN. Specimen S2.0D-2unexpectedly failed at the first pull-out force of cyclic load-ing; however, the specimen exhibited a decay shear transferafter peaking at a slip of 1.01.5 mm.

The loadslip response of the specimens indicated that theloadslip curves of cyclic loading tests are similar to theloadslip curve obtained for monotonic static tests. Theshifting between these two curves in the ordinate load axis isdue to the different cyclic loading rate and concrete pluglength. The effects of cycling rate and the cyclic reductionfactor are discussed in the following sections.

Slip versus cycles results for cyclic loading

Cyclic loading reduced the bond strength and ultimate ca-pacity of the specimens. This was due to damage of the con-crete plug and pileplug interface either by progressive loss

of stiffness through the accumulation of microcracking or byprogressive plastification that appears as an irreversible re-sidual strain that increases with each additional cycle.

The slip versus cycle behaviour for specimens S1.0D-2,S1.75D-2, and S2.0D-1 is plotted in Fig. 10. It is evidentthat slip increases with an increase in the number of load cy-cles and that the rate of slip growth increased with an in-crease in the peak load (see Table 4). The nonsymmetricbehaviour in some specimens may be due to differences inthe local stiffness of the concrete plug adjacent to the testtube. A concentration of coarse aggregate or voids immedi-ately adjacent to the top or bottom of the steel tube wouldhave an effect on the concrete stiffness and the rate of slipgrowth. The different effective mechanical interlock mecha-nisms in pull out and push out may also have effects on theconcrete stiffness.

It was observed for most of the specimens that, after thefirst few cycles at any load range, the slip appeared to be ap-

2006 NRC Canada

118 Can. J. Civ. Eng. Vol. 33, 2006

(a)

-600

-400

-200

0

200

400

600

800

-8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0Slip (mm)

Fo

rce

(kN

)

(b)

-400

-200

0

200

400

600

800

-2.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0Slip (mm)

Fo

rce

(kN

)

(c)

-300-200-100

0100200300400500

-2.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0Slip (mm)

Fo

rce

(kN

)(d)

-1200

-800

-400

0

400

800

1200

-2.0 -1.0 0.0 1.0 2.0Slip (mm)

Fo

rce

(kN

)

(e)

-600

-400

-200

0

200

400

600

-8.0 -6.0 -4.0 -2.0 0.0 2.0Slip (mm)

Fo

rce

(kN

)

(f)

-500-400-300-200-100

0100200300400500

-4.0 -2.0 0.0 2.0 4.0 6.0 8.0 10.0Slip (mm)

Fo

rce

(kN

)

Fig. 6. Loadslip response in stage 1 of specimens (a) S1.0D-1, (b) S1.0D-2, (c) S1.0D-3, (d) S1.5D-1, (e) S1.5D-2, and (f) S1.5D-3.

proximately linear with the number of cycles. A line of bestfit to the rate of slip growth with number of cycles was cal-culated for every test at every load range. The rate of slipvalues and load ranges are presented in Table 4. These re-sults are plotted in Fig. 11, with the rate of slip growth plot-ted on a logarithmic scale. Although there is considerablescatter in the data, there is a clear trend that the rate of slipgrowth increased with an increase in the peak load. Thescatter in the data is probably a reflection of the variation inthe characteristics of the concrete plug and the effect ofshrinkage. A line of best fit to the data, plotted in Fig. 11,resulted in the following expression for the rate of slipgrowth per cycle for symmetric cyclic loading:[1] slip growth per cycle = 10( / )0.19 0.054uP P mm/cyclewhere P is the load, and Pu is the ultimate load.

Various forms of representing the data were trialed, in-cluding higher order functions, to fit the trend of the datagiven in Fig. 11.

Although a slightly higher correlation could be achievedusing higher order functions, the correlation was not signifi-cantly better and, in the absence of a physical model that sup-ports a particular relationship, an exponential function wasadopted as providing a simple function that could consistentlybe applied across different data series. Equation [1] does notstrictly satisfy the boundary condition for the rate of slipgrowth that when P = 0, the slip growth per cycle could bezero. This is not possible with an exponential function. Thediscrepancy arises because the equation is derived empirically,and not from the fundamental physical model of the behav-iour. When P = Pu, the slip growth per cycle, calculated fromthe equation, is finite (but large). This is consistent with theobserved behaviour, namely that the slip does not approachinfinity as the specimen approaches failure.

Cyclic reduction factor

The ultimate capacity and load response of the specimensunder pull-out, push-out, and cyclic loading are presented in

2006 NRC Canada

Nezamian et al. 119

(b)

-400-300-200-100

0100200300400500

-5 0 5 10 15 20 25Slip (mm)

Fo

rce

(kN

)

(c)

-300-200-100

0100200300400500600

-5 0 5 10 15 20Slip(mm)

Fo

rce

(kN

)

(a)

-500-400-300-200-100

0100200300400500

-5 0 5 10 15 20 25Slip (mm)

Fo

rce

(kN

)Fig. 7. Loadslip response for specimens (a) S1.25D-1,(b) S1.25D-2, and (c) S1.25D-3.

(a)

-500-400-300-200-100

0100200300400

-15 -10 -5 0 5 10Slip (mm)

Fo

rce

(kN

)

(b)

-400-300

-200-100

0

100200

300400

-15 -10 -5 0 5 10 15Slip (mm)

Fo

rce

(kN

)

(c)

-300-200-100

0100200300

400500

-15 -10 -5 0 5 10 15 20Slip (mm)

Fo

rce

(kN

)

Fig. 8. Loadslip responses for specimens (a) S1.75D-1,(b) S1.75D-2, and (c) S1.75D-3.

Tables 2 and 3 and Figs. 69. The ultimate capacity andloadslip response of specimens under the cyclic loadingcan be reasonably approximated from the static ultimatestrength and loadslip of the specimen by reducing the ulti-mate strength values of static testing by the cyclic reductionfactor. The cyclic reduction factor is defined as the factor bywhich the cyclic strength of the specimen can be obtainedfrom the static strength for a given displacement. The cyclicreduction factor seems to depend on the rate of load, numberof cycles, concrete characteristics and shrinkage, imperfec-tion of the steel tube, length of the plug, and perhaps thepresence of reinforcement. This rule does not apply to allspecimens, however, because of irregular peak loads. Theseirregularities might be caused by either steel tube imperfec-tions or the effect of shrinkage. Table 5 shows calculated cy-clic reduction factors for specimens with different concreteplug lengths based on the ultimate pull-out, push-out, andcyclic strength of specimens. The slip values at the peakload are also tabulated.

The cyclic reduction factors for the aforementioned 10specimens indicate that the symmetric cyclic loading reducesthe shearbond transfer between the concrete plug and thesteel tube. This is due to the accumulation of damage to theplugpile interface. The exceptions to this are specimensS1.0D-2 and S1.25D-3, possibly because of steel tube im-perfections or the effect of shrinkage. An average (mean) cy-clic reduction factor of 0.74 was achieved (standarddeviation of 0.25), however.

Bond strength and failure mechanisms

The bond strength of a reinforced concrete plug embeddedin a steel tube is a function of both chemical adhesion of thesteelconcrete interface and mechanical interlock betweenthe concrete core and the steel surface. To overcome me-chanical interlock, a small dilation of the tube occurs as it

2006 NRC Canada

120 Can. J. Civ. Eng. Vol. 33, 2006

(a)

-1000-800-600-400-200

0200400600800

1000

-1.5 -1 -0.5 0 0.5 1 1.5 2Slip (mm)

Fo

rce

(kN

)

(b)

0

100

200

300

400

500

0 3 6 9 12 15 18Slip (mm)

Fo

rce

(kN

)

(c)

-400-300-200

-1000

100200

300400

-15 -10 -5 0 5 10 15Slip (mm)

Fo

rce

(kN

)Fig. 9. Loadslip response for specimens (a) S2.0D-1,(b) S2.0D-2, and (c) S2.0D-3.

-1.5

-1.0

-0.5

0

0.5

1.0

0 5 10 15 20 25 30 35 40Cycle num ber

)m

m(pilS = 0.25

= 0.32

= 0.38

=0.44

= 0.51

=0.57

=0.63

=0.70

(b)

-2.0-1.5-1.0-0.5

00.51.01.52.02.5

0 5 10 15 20Cycle number

)m

m(pilS

P/Pu=0.50

= 0.55

(c)

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0 5 10 15 20Cycle number

P/Pu = 0.23

= 0.38

(a)

)m

m(pilS

P/Pu

P/PuP/Pu

P/Pu

P/Pu

P/PuP/Pu

P/PuP/Pu

P/Pu

Fig. 10. Slip versus number of cycles for specimens (a) S1.0D,(b) S1.75D-2, and (c) S2.0D-1.

rides over the asperities of the interface, generating radialcontact pressure, which enhances the frictional resistance. Ina push-out situation, dilation of the concrete plug at the topof the connection enhances radial pressure and thereforefrictional resistance. This is due to the Poissons effect at thetop of specimen, where the compression is high in the con-crete and low in the steel. At the base, contact pressure be-tween concrete and steel is reduced, due to the Poissonseffect, and the effective bond is therefore reduced at this lo-cation (see Fig. 12).

In the pull-out case, the reverse is expected to occur. Thatis, near the base of the concrete plug, contraction of the steeltube is much higher than that of the concrete core, causing itto grip the concrete plug. Near the top part of the plug, thetension force is transferred to the concrete through the rein-forcing bars embedded in the concrete core and the pile cap.

The tensile stresses that develop in the concrete core resultin contraction of the concrete, while contraction in the steeltube is relatively small. This should result in separation be-tween the steel tube and the concrete (Fig. 12). Consideringthe fact that deformed bars are used as reinforcement, theribs on the bars tend to impart wedge pressure on the outerconcrete layer, causing dilation of this layer. This dilationenhances the frictional stresses between the steel tube andthe concrete. The test results indicated that the ultimate aver-age bond stress increases with a decrease in the ratio of con-crete plug embedment length to tube inner diameter L/Di,which can also be explained by the aforementioned mecha-nisms. In stage 2, the shrinkage cracks at the pileplug inter-face were observed prior to testing, which is believed tocause a reduction in bond strength and scatter in the test re-sults.

Figure 13 shows measured longitudinal and hoop strainsat the outer surface of the steel tube for specimens S1.0Dand S1.5D in pull-out tests and specimens S1.75D-3 andS2.0D in push-out tests, which are in a good agreement withthe mechanisms. Longitudinal strains of specimens S1.0Dand S1.5D indicate that the loadshear transfer in the pileplug interface mainly occurs at the top because of reinforce-ment wedge action at the top of the specimens and at thebottom because of the Poissons effect. The test results ofhoop strain along the steel tube indicate that the ribs of thelongitudinal reinforcement tend to impart wedge pressure onthe outer concrete layer, causing dilation of this layer. Thisdilation enhances the radial contact pressure and causes ra-dial expansion of the steel tube. At the ultimate load level,this dilation overcomes the contraction of the steel tube dueto tensile axial force at the top of the specimen. Longitudinal

2006 NRC Canada

Nezamian et al. 121

Rate of slip growth (106 mm per cycle)Specimen No. Type of test Load (kN) P/Pu Positive slip Negative slipS1.0D-2 Symmetric cyclic 150 0.23 4.06 5.71

Symmetric cyclic 250 0.34 3.91 8.86S1.0D-3 Symmetric cyclic 150 0.23 5.11 8.64

Symmetric cyclic 250 0.34 19.07 33.04S1.5D-2 Symmetric cyclic 230 0.23 9.68 9.55S1.5D-3 Symmetric cyclic 230 0.23 4.02 4.08

Symmetric cyclic 400 0.34 75.80 131.93S1.75D-2 Symmetric cycling 100 0.25 2.19 2.19

Symmetric cycling 125 0.32 2.19 3.28Symmetric cycling 150 0.38 4.37 3.28Symmetric cycling 175 0.44 7.66 3.28Symmetric cycling 200 0.51 7.66 8.75Symmetric cycling 225 0.57 13.13 15.32Symmetric cycling 250 0.63 26.25 33.91Symmetric cycling 275 0.70 105.01 164.09

S1.75D-3 Symmetric cycling 150 0.38 0.41 1.09Symmetric cycling 200 0.51 49.51 99.55Symmetric cycling 225 0.57 903.43 1000.91

S2.0D-1 Symmetric cycling 500 0.50 25.71 33.36Symmetric cycling 550 0.55 29.69 33.91

S2.0D-3 Symmetric cycling 200 0.42 19.69 17.50Symmetric cycling 250 0.52 33.91 26.25Symmetric cycling 300 0.63 164.09 91.89

Table 4. Rate of slip growth.

R2 = 0.3905

0.0001

0.001

0.01

0.1

10 0.2 0.4 0.6 0.8

Load range, P/Pu

Rate

of

sli

pg

row

th(m

m/c

ycle

)

Fig. 11. Load range versus rate of slip growth. P, load; Pu, ulti-mate load; R2, coefficient of determination.

strains of specimens S1.75D and S2.0D show very smallshear transfer between the concrete plug and the steel tube atthe bottom of the specimens and maximum shear transfer atthe top of the specimens due to the Poissons effect.

It should be noted that the steel tube was subjected to ra-dial contact forces along an arbitrary circle of the tube. Be-cause of the symmetry of such loading, every section normalto the axis will remain circular, and the radius R will un-dergo a change R y= , varying along the length of the plug.The radial displacement y can be regarded as deflection for alongitudinal element of the tube, and hence it is seen that theassumed loading will set up bending stresses in the longitu-dinal elements. This situation is analogous to the case of abeam on an elastic foundation (Hetnyni 1964). It can beseen that mechanical macro interlock mechanisms at the topand bottom of specimens caused the radial pressure on thesteel tube. The differential of radial pressure along the steel

tube applied a longitudinal bending moment on the tube.Since the bending of the tube wall is a plane strain environ-ment, it follows that M Mc y= in the circumferential ring,where My is the longitudinal bending moment, and is Pois-sons ratio for the steel tube. It should be noted that the mea-sured outer longitudinal strain on the steel tube is due to theaxial pull-out force, together with the longitudinal bendingmoment along the steel tube (Nezamian et al. 2004).

Figure 14 shows a concrete plug completely pulled outfrom the steel tube after the specimen failed in a cyclic test.The failure mechanism displayed by the specimen was at thebase of the concrete plug, where the contraction of the steeltube is much greater than that of the concrete core, causingit to grip the concrete plug. The diagonal tension crackformed in the concrete layer between the longitudinal rein-forcement and the steel tube and then extended to the end ofthe longitudinal reinforcement and from there in the hoopdirection. This crack appeared to correspond to a tensionsplitting of the concrete plug at ultimate pull-out capacity ofthe specimen. The observed damage at the top of the con-crete plug also indicated the Poissons effect at the top of thespecimen in the case of push out. These failure mechanismswere also verified by a nonlinear finite element analysis(Nezamian et al. 2004). The failure mechanisms are also inagreement with the described bond strength mechanisms.

Conclusions

An experimental study was conducted to investigate thebehaviour and bond strength of concrete plugs embedded intubular steel piles under cyclic loading. Based on this study,the following conclusions are drawn:(1) Average ultimate bond strengths of 4.25 MPa for a static

load and 2.77 MPa for a cyclic load were achieved forstage 1, and an average static bond strength of 2.37 MPaand average cyclic bond strength of 1.70 MPa wereachieved for stage 2. The higher bond strengths thanthose reported by Virdi and Dowling (1975) and Shakir-Khalil (1991, 1993a, 1993b) are due to the presence ofreinforcement and the use of a smaller ratio of concreteplug embedment length to tube inner diameter, L/Di.The observed shrinkage of the concrete plug at stage 2is believed to cause a reduction in bond strength, whichis in agreement with the findings of Virdi and Dowling(1975) and Roeder et al. (1999).

2006 NRC Canada

122 Can. J. Civ. Eng. Vol. 33, 2006

SpecimenNo. Failure regime

Ultimatestrength (kN)

Cyclic ultimatestrength (kN)

Cyclic reductionfactor

Slip at peakload (mm)

S1.0D-2 Pull out with pre-cyclic loading 665 711 1.07 12.2S1.0D-3 Pull out with pre-cyclic loading 665 410 0.62 11.7S1.25D-2 Symmetric cycling loading 460 439 0.95 24.1S1.25D-3 Pull out with pre-cyclic loading 460 540 1.17 20.9S1.5D-2 Pull out with pre-cyclic loading 1000 500 0.50 1.8S1.5D-3 Pull out with pre-cyclic loading 1000 404 0.40 9.2S1.75D-2 Symmetric cycling loading 395 300 0.76 14.8S1.75D-3 Cycling with pre-pull-out test 395 250 0.63 18.6S2.0D-1 Cycling with pre-pull-out test 1000 600 0.60 16.0S2.0D-3 Symmetric cycling loading 479 350 0.73 18.6

Table 5. Cyclic reduction factor.

Push-out Force

Stee

l Tu

be

Con

cret

ePl

ug

High compressive stress in concreteLow compressive stress in the steel tube

Low compressive stress in concreteHigh compressive stress in the steel tube

High contact pressuredue to the Poisson's effect

due to the Poisson's effectSeparation

Pull-out Force

High tensile stress in concreteLow tensile stress in the steel tube

Low tensile stress in concreteHigh tensile stress in the steel tube

due to the Poisson's effectSeparation

due to the Poisson's effectHigh contact pressure

Stee

l Tu

be

Con

cret

ePl

ug

Fig. 12. Bond strength mechanisms.

(2) The push-out and pull-out tests conducted under sym-metric cyclic loading demonstrated that slip between theconcrete plug and the steel tube increased with repeatedloading, and the rate of slip growth increased with anincrease in the peak load.

(3) Empirical relationships between the load and the rate ofslip growth for symmetric cyclic loading were obtainedfrom the experimental data as follows:

[1] slip growth per cycle= 10( / )0.19 0.054uP P mm/cycle

This equation may be used to predict the failure of thespecimen due to incremental slip between the concreteplug and the steel tube.

(4) The ultimate capacity and loadslip response of speci-mens under cyclic loading can be reasonably approxi-

2006 NRC Canada

Nezamian et al. 123

(a)

-0.0001-0.00010.00000.00010.00010.00020.00020.00030.00030.0004

0 50 100 150 200Distance from the bottom of the plug (mm)

Lo

ng

itu

din

al

Str

ain

(Mic

ros

train

)(b)

-0.0004-0.0004-0.0003-0.0003-0.0002-0.0002-0.0001-0.00010.0000

0 50 100 150 200Distance from the bottom of the plug (mm)

Ho

op

Str

ain

(Mic

ros

train

)

(c)

-0.0002-0.00010.00000.00010.00020.00030.00040.00050.00060.0007

0 50 100 150 200 250 300Distance from the bottom of the plug (mm)

Lo

ng

itu

din

al

Str

ain

(Mic

ros

train

)

(d)

-0.0003

-0.0002

-0.0002

-0.0001

-0.0001

0.0000

0 50 100 150 200 250 300Distance from the bottom of the plug (mm)

Ho

op

Str

ain

(Mic

ros

train

)

(e)

-0.0012

-0.0010

-0.0008

-0.0006

-0.0004

-0.0002

0.0000

0.0002

-50 50 150 250 350Distance from the bottom of the plug (mm)

Lo

ng

itu

din

al

Str

ain

(Mic

ros

train

)

(f)

0.00000.00010.00010.00020.00020.00030.00030.00040.00040.00050.0005

-50 50 150 250 350Distance from the bottom of the plug (mm)

Ho

op

Str

ain

(Mic

ros

train

)

(g)

-0.0030

-0.0025

-0.0020

-0.0015

-0.0010

-0.0005

0.0000

-50 50 150 250 350 450Distance from the bottom of the plug (mm)

Lo

ng

itu

din

al

Str

ain

(Mic

ros

train

)

(h)

0.0000

0.0002

0.0004

0.0006

0.0008

0.0010

0.0012

-50 50 150 250 350 450Distance from the bottom of the plug (mm)

Ho

op

Str

ain

(Mic

ros

train

)

Fig. 13. Longitudinal strain of specimens S1.0D (a) and S1.5D (c) at ultimate pull out and S1.75D-3 (e) and S2.0D (g) at ultimate pushout, and hoop strain of specimens S1.0D (b) and S1.5D (d) at ultimate pull out and S1.75D-3 (f) and S2.0D (h) at ultimate push out.

mated from the static ultimate strength and loadslip ofthe specimen by reducing the ultimate strength values ofthe static test by the cyclic reduction factor. An averagecyclic reduction factor of 0.74 was achieved.

(5) The loadslip curves of the specimens showed a nearlybilinear response. The change of slope of the loadslipcurves during loading is assumed to occur with thebreaking of the microchemical adhesion and the activa-tion of the mechanical macro interlocking mechanism.

The main mechanism that is believed to have contrib-uted to the bond strength in pull out was the dilation ofthe concrete due to the wedging action exerted by thedeformed steel bars against the concrete layer betweenthe bars and the steel tube. This dilation increased thecontact pressure, which enhanced the friction resistance.A secondary factor was the pronounced Poissons ratioeffect increasing the radial contact stress at the base ofthe concrete plug.

(6) Further tests are required to account for variations insome parameters such as steel tube diameter and aspectratio, concrete strength, and steel tube surface condition.Tests are also needed using different cyclic load re-gimes. A more detailed investigation of the effect ofconcrete shrinkage on the ultimate pull-out strength isalso required.

References

Al-Mahaidi, R., Grundy, P., and Bean, W. 1999. Pullout strength ofconcrete plugs in tubular piles. In Proceedings of the 9th Inter-national Offshore and Polar Engineering Conference, Brest,France, 30 May 4 June 1999. Edited by J.S. Chung, T. Matsuiand W. Koterayama. International Society of Offshore and PolarEngineers, Cupertino, Calif. pp. 2429.

BSI. 1979. Steel, concrete and composite bridges: code of practicefor design of composite bridges. British Standard BS5400, Brit-ish Standards Institution (BSI), London, UK.

ECS. 1994. Eurocode 4: adopted European prestandard EVN 1994-1-1:1992. European Committee for Standardization (ECS),Brussels, Belgium.

Hetnyi, M. 1964. Beams on elastic foundation. University ofMichigan Press, Ann Arbor, Mich.

Morishita, Y., Tommil, L., and Yoshimura, K. 1979. Experimentalstudies on bond strength in concrete filled steel tubular columnssubject to axial loads. Transactions of the Japan Concrete Insti-tute, 1: 359366.

Morishita, Y., Tommil, L., and Yoshimura, K. 1980. A method ofimproving bond strength between steel tube and concrete corecast in circular steel tubular columns. Transactions of the JapanConcrete Institute, 2: 319326.

Nezamian, A., Al-Mahaidi, R., Grundy, P., and Bean, W. 2001.Bond strength mechanisms in the pull out of concrete plugs intubular steel piles. In Proceedings of the 2nd International Con-ference on Mechanics of Structures, Materials and Systems,MSMS 2001, Wollongong, Australia. pp. 6773.

Nezamian, A., Al-Mahaidi, R., Grundy, P., and OLoughlin, B.2002. Push out strength of concrete plugs in tubular steel piles.In Proceedings of the 12th International Offshore and Polar En-gineering Conference, Kitakyushu, Japan, 2631 May 2002.CD-ROM. International Society of Offshore and Polar Engi-neers, Cupertino, Calif. pp. 6065.

Nezamian, A., Al-Mahaidi, R., and Grundy, P. 2003. The effect ofcyclic loading on the bond strength of concrete plugs embeddedin tubular steel piles. In Advances in Structures: Proceedings ofthe International Conference on Advances in Structures(ASSCCA03), Sydney, Australia, 2225 June 2003. Edited byG.J. Hancock, M.A. Bradford, T.J. Wilkinson, B. Uy and K.J.R.Rasmussen. A.A. Balkema, Exton, Pa. pp. 11251131.

Nezamian, A., Al-Mahaidi, R., and Grundy, P. 2004. Finite elementanalysis of concrete plugs embedded in tubular steel piles. InProceedings of the 14th International Offshore and Polar Engi-neering Conference, Toulon, France, 2328 May 2004. CD-

2006 NRC Canada

124 Can. J. Civ. Eng. Vol. 33, 2006

(a)

(b)

(c)

Fig. 14. Pulled out concrete plug from the steel tube: (a) com-plete plug, (b) top of plug, and (c) bottom of plug.

ROM. International Society of Offshore and Polar Engineers,Cupertino, Calif.

Roeder, C.W., Cameron, B., and Brown, C.B. 1999. Composite ac-tion in concrete filled tubes. ASCE Journal of the Structural Di-vision, 125(5): 477484.

SAA. 1983. Carbon steel tubes for mechanical and structural use.Australian Standard SAA-AS-1450, Standards Association ofAustralia (SAA), Homebush, NSW, Australia.

SAA. 1993. Specification and methods of for packaged concretemix. Australian Standard SAA-AS-3648, Standards Associationof Australia (SAA), Homebush, NSW, Australia.

SAA. 1997. Portland and blended cement. Australian StandardSAA-AS-3972, Standards Association of Australia (SAA),Homebush, NSW, Australia.

Sakino, K., and Tomii, M. 1981. Hysteretic behaviour of concretefilled square steel tube beamcolumns failed in flexure. Transac-tions of the Japan Concrete Institute, 3: 439446.

Shakir-Khalil, H. 1991. Bond strength in concrete-filled steel hol-low sections. In Composite Steel Structures: Recent Researchand Developments: Proceedings of the International Conferenceon Steel and Aluminium Structures, Singapore. Edited by S.L.Lee and N.E. Shanmugam. Elsevier Applied Science, London,UK. pp. 157168.

Shakir-Khalil, H. 1993a. Push-out tests on concrete-filled steel hol-low sections. The Structural Engineer, 71(13): 230233.

Shakir-Khalil, H. 1993b. Push-out tests on concrete-filled steel hol-low sections. The Structural Engineer, 71(13): 234243.

Taplin, G. 1999. The behaviour of composite beams under repeatedloading. Ph.D. thesis, Department of Civil Engineering, MonashUniversity, Melbourne, Australia.

Virdi, K.S., and Dowling, P.J. 1975. Bond strength in concretefilled circular steel tubes. CESLIC Report CC11, Civil Engi-neering Structures Laboratory, Imperial College, London, UK.

2006 NRC Canada

Nezamian et al. 125