State-Of-The-Art Report on Anchorage to Concrete

ACI 355.1R-91

STATE-OF-THE-ART REPORT ON

(Reapproved 1997)

ANCHORAGE TO CONCRETE

Reported by ACI Committee 355

Patrick J. Creegan Harry A. ChambersChairman Secretary

Edwin A. BurdetteRobert W. CannonPeter J. CarratoPeter D. CourtoisRolf Eligehausen

Raymond R. FunkC. Raymond HaysPaul R. HollenbachGerard B. HassehvanderHarry B. Lancelot III*

Douglas D. LeeAlexander Makitka, Jr.Donald F. MeinheitRichard S. OrrMoorman L Scott

George A. SenkiwHarry WiewelJim L WilliamsRichard E. Wollmershauser

*Committee Chairman during the formative years of this report.

For the first time concrete anchoring knowledge based on worldwide test programs is presented in a state-of-the-art document. Performanceof different anchor types, including cast-in-place, grouted, expansion, torque-controlled, chemical (adhesive), and undercut anchors is presentedin both uncracked and cracked concrete. Failure modes in tension and shear, spacing and edge distance, group performance, and loaddisplacements are offered. The effect of loading conditions for structural supports, column bases, and pipe supports as well as base plateflexibility, how load is transferred to anchors, and ductility are discussed. Design criteria and existing code requirements, both domestic andforeign, are presented.

KEYWORDS: Adhesive anchors; anchorages; anchors; anchor groups; base plates; bolts; cast-in-place anchors; chemical anchors; coderequirements; combined loads; compression zone; concrete; cracked concrete; creep; deformation; design criteria; drilling; ductility;dynamic loads; edge distance; embedment; expansion anchors; failure modes; fatigue loads; fasteners; flexible base plates; grouting; loads;load transfer; load-displacement; post-installed anchors; preload; pullout; seismic loads; shear loads; slip; spacing; spalling; static loads;stiffness; studs; structural design; tensile strength; tension loads; tension zone; temperature; torque; torque-controlled anchors; ultimatestrength; undercut anchor, yield strength.

FORWARDThis state-of-the-art report on anchorage to concrete is the first of a two-volume project being undertakenby ACI Committee 355. The second volume, currently being developed, is a design manual. This firstvolume includes no design aids or procedures, per se, but with emphasis on behavior will serve as the guidefor preparation of the second volume.

Committee 355 is working with Committees 349 and 318 toward the objective of including the subject ofanchorage to concrete in ACI 318-95.

ACI Committee Reports, Guides, Standard Practices, andCommentaries are intended for guidance in designing,planning, executing, or inspecting construction, and inpreparing specifications. Reference to these documents shallnot be made in the Project Documents. If items found inthese documents are desired to be a part of the ProjectDocuments, they should be phrased in mandatory languageand incorporated into the Project Documents.

ACI 355.1R-91 became effective JuIy 1, 1991.Copyright 0 1991, American Concrete Institute.All rights reserved including rights of reproduction and use in any

form or by any means, including the making of copies by any photoprocess, or by any electronic or mechanical device, printed or written ororal, or recording for sound or visual reproduction or for use in anyknowledge or retrieval system or device, unless permission in writing isobtained from the copyright proprietors.

355.1 R-l

355.1R-2 MANUAL OF CONCRETE PRACTICE

TABLE OF CONTENTS

Chapter 1-Introduction, p 355.1R-21.1 Purpose1.2 Significance of the subject1.3 Scope

Chapter 2-Types of anchoring devices,p 355.1R-2

2.1 Introduction2.2 Scope2.3 Anchor systems2.4 Cast-in-place systems2.5 Post-installed systems

Chapter 3-Behavior of anchors, p 355.1R-93.1 Introduction3.2 Behavior of anchors in uncracked concrete3.3 Behavior of anchors in cracked concrete3.4 Behavior of cast-in-place anchor bolts in

uncracked concrete piers3.5 References

Chapter 4-Design considerations,p 355.1R-53

4.1 Introduction4.2 Functional requirements4.3 Materials4.4 Design basis4.5 Construction practices4.6 References

Chapter 5-Construction considerations,p 355.1R-60

5.1 Introduction5.2 Shop drawings/submittals5.3 Tolerances5.4 Installation of anchors5.5 Inspection5.6 Grouting5.7 Field problems

Chapter 6-Requirements in existing codesand specifications, p 355.1R-66

6.1 Introduction6.2 Existing codes and specifications6.3 Application and development of codes6.4 References

Appendix A-Conversion factors, p 355.1R-71

Appendix B-Notations, p 355.1R-71

CHAPTER 1 -INTRODUCTION1.1-Purpose

The purpose of this document is to summarizethe current state of the art in anchorage toconcrete.

1.2-Significance of the subjectTo date, anchorage to concrete has received

little attention in structural codes. Emphasis hasbeen primarily on the tensile and shear capacitiesof anchorage devices. As designs became moresophisticated and analyses more exacting, moreemphasis was placed on the transfer of loadsthrough single anchors and anchor systems. It wasrecognized that performance of anchors controlledthese load transfers, and that generally, failuremodes at ultimate anchor capacities wereimportant. There were no definitive design codesor anchorage performance criteria on whichdesigners and installers could rely. Subsequently,a myriad of approaches were developed.

1.3-ScopeThis state-of-the-art report summarizes anchor

types and provides an overview of anchor per-formance and failure modes under various loadingconditions in both uncracked and cracked con-crete. It covers design and constructionconsiderations and summarizes existing require-ments in codes and specifications. References aregiven for further review.

CHAPTER 2 -TYPES OF ANCHORINGDEVICES

2.1-IntroductionThere are many types of devices used for

anchoring structures or structural members toconcrete. The design of anchorages, involving theselection and positioning of these devices has beenbased on the Engineer’s experience and judgment,private test data, manufacturers’ data, and existing(sometimes obsolete) code requirements. It isproposed to promote a design of anchorages thatmore consistently reflects the performancepotential of each type of anchor.

2.2-ScopeThis report relates to the most widely used

types of anchor, in sizes ranging from 1/4 in. (6.35mm) to 2 l/2 in. (63.5 mm) in diameter. Includedfor consideration are only those devices which cangenerally be considered bolt and insert-type

ANCHORAGE TO CONCRETE 355.1R-3

anchors. Excluded from consideration are shearlugs, structural shapes, powder actuated fasteners,light plastic or lead inserts, hammer drivenconcrete nails, screw driven systems, and cables.These are excluded because there is a paucity oftest data regarding their performance. Theanchors included in this report are eithercommercially available or may be fabricated.

2.3-Anchor SystemsAccording to present practice, there are two

broad groups of anchoring systems: cast-in-placesystems (anchors installed before the concrete iscast) and post-installed systems (anchors installedin holes drilled after the concrete has been castand cured). Table 2.1 identifies these two groupsof anchors.

Table 2.1 -Types of anchors in concrete

Cast-in-place systems

Embedded, nonadjustable

Common boltsHooked "J" & "L" boltsThreaded rodReinforcing steel

Fig. 2.1Fig. 2.2Fig. 2.3Fig. 2.4

Threaded insertsStud-welded plates

Fig. 2.5Fig. 2.6

Bolted connections Fig. 2.7

L Steelplate

Fig. 2.7-Bolted connections

Plastic

Adjustable anchors Fig. 2.8

Post-installed systems

Bonded anchorsGrouted anchors

Headed bolts or anchor Fig. 2.9

Chemical anchorsWith threaded rodWith reinforcing steel

Fig. 2.10Fig. 2.11

Expansion anchorsTorque-controlled

Heavy-duty sleeve anchor Fig. 2.12
Sleeve anchor Fig. 2.13 Shell expansion anchor Fig. 2.14 Wedge anchor Fig. 2.15 Rock/concrete expansion
anchor Fig. 2.16

Deformation controlledDrop-in anchor Fig. 2.17
Self-drilling anchor Fig. 2.18 Stud anchor Fig. 2.19
UndercutWith predrilled under-cuthole Fig. 2.20

Self undercutting Fig. 2.20

2.4-Cast-in-place systems2 . 4 . 1 - Embedded Anchors, Non -

Adjustable - These anchors may have an endattachment, such as a coil loop, head, nut, orplate, which will enhance anchorage propertiesand develop full potential strength by means ofbond, and/or bearing, or both. Typical examplesof these anchors are:

Common bolts - structural steel boltsplaced with the head intothe concrete. (Fig. 2.1)

Hooked"J" or "L" bolts

Threaded rod

Reinforcing steel

Threaded inserts

Stud welded plates - steel plates which havesmooth bent hooked bars,deformed bars, or headedstud anchors. (Fig. 2.6)

2.4.2 Bolted connections-These anchors consistof headed bolts, as embedded or through-connectors. (Fig. 2.7).

-bent, smooth or deformedthreaded bars. Have beenknown to straighten out inpull-out tests. (Fig. 2.2)

- straight threaded rod,usually with coarsethreads. (Fig. 2.3)

- Stock or trade-name rein-forcing bar (Fig. 2.4)

- wire form or internallythreaded ferrule inserts,or coils, usually manu-factured with internal orexternal threads, with wireloop struts. Headedanchors made fromsmooth or reinforcingsteel bar also fall into thiscategory. (Fig. 2.5)

355.1R4 MANUAL OF CONCRETE PRACTICE

W a s h e r t a c k w e l d e d

Fig. 2.1- Common bolts

.b v *

Fig. 2.3 - Threaded rod

. -

D

P

N o t e : E i t h e r 'J ' or 'L ' ’ boIts c a n b e m a d e

f r o m p l a i n o r t h r e a d e d r o d

Fig. 2.2-J- and L-bolts (not recommended)

Fig. 2.4 -Reinforcing steel


a *

.‘X. .v * B .

Fig. 2.5 - Threaded inserts

We I d

Fig. 2.6 - Stud- welded plates


.

P - 4

Fig. 2.8-Adjustable anchors

i n o r* .v chemical from capsule

Fig. 2.10-Chemical anchor with threaded rod

n

2.4.3 Adjustable anchors-Adjustable anchorscan be adjusted for lateral position or depth (Fig.2.8). They are normally used for attaching largemachines or equipment bases. On thin floor slabs,the anchor bolt often goes through the concrete todevelop the required anchor capacity. When thefloor slab or foundation is very thick, the anchorcan develop full capacity and still be embedded inthe concrete. After the equipment or machinebase is installed and leveled, grout is used to fillthe void around the anchor. The anchor then actssimilar to a cast-in-place anchor.

2.5-Post-installed systemsThese anchors are installed in a hole drilled in

the cured concrete. There are two basic groups ofpost-installed systems: bonded and expansion.

2.5.1-Bonded anchors2.5.1.1 Grouted anchors-Grouted anchors are

headed or headless bolts or threaded rods. Theyare set in predrilled holes with portland cementand sand grout or other commercially availablepremixed grout. (Fig. 2.9)

2.5.1.2 Chemical anchors-Chemical anchorsare usually threaded rods (Fig. 2.10) or deformedbars (Fig. 2.11) which are bonded in place withtwo-part chemical compounds of polyesters,vinylesters, or epoxies. The chemicals areavailable in four forms: glass capsules, plasticcartridges, tubes, or bulk.

Glass capsules are inserted into the drilled hole,and then broken by the anchor rod when it is ro-tated and hammered into place, thereby mixingtwo components to cause a chemical reaction.

The plastic cartridges are used with a dispenserand a mixing nozzle which mixes the two parts,initiating a chemical reaction while installing thecompound into the drilled hole. The anchor rodis then inserted into the hole completing theinstallation. The setting time is dependent ontemperature, varying from a few minutes at 90o Fup to several hours at 30o F.

The tube or “sausage” type contains twocomponents which are mixed by kneading thetube, placing the mixture into the hole, and finally,inserting the anchor rod into the hole.

The bulk systems predominantly use epoxies,which are either premixed in a pot and usedimmediately, or pumped through a mixer andinjected into the hole. The anchor is installedimmediately afterward. Epoxies can be form-ulated to set up quickly or slowly (up to 36 hrcuring time).


BEFORE TORQUING AFTER TORQUING

Fig. 2.12 - Heavy-duty, torque-controlled sleeveanchor

..Fig. 2.13 - Sleeve anchor

S i n g l e - a c t i n g D o u b l e a c t i n g

( s h e l l e x p a n d e d ( s h e lb y s i n g l e w e d g e n u t )

e x p a n d e db y o p p o s i n g w e d g e )

Fig. 2.14 - Shell expansion anchor

2.5.2 Expansion anchors-Expansion anchorsare designed to be inserted into predrilled holesand then expanded by either tightening the nut(torque controlled expansion anchor, Sections2.5.2.1 to 2.5.2.5), hammering the anchor
(deformation controlled expansion anchor,Sections 2.5.2.6 to 2.5.2.8), or expanding into an undercut in the concrete (undercut anchors,Section 2.5.2.9). These anchors transfer the tension load from the bolt to the concrete byexpansion pressures or forces through frictionand/or keying against the side of the drilled hole.They often are supplied with a bolt, nut, andwasher. The following sections describe thevarious types of expansion anchors.
2.5.2.1 Heavy duty, torque controlled sleeveanchor-This type of anchor consists of a bolt orthreaded rod with nut and washer on one end anda cone on the embedded end, (Fig. 2.12). Aroundthe cone is a heavy expansion sleeve. Above thesleeve is a collapsible mechanism, sometimes madeof plastic. A spacer sleeve extends to the surfaceof the drilled hole. The anchor is set by tight-ening the bolt head or nut which draws the coneup through the expansion sleeve, expanding itagainst the side of the drilled hole. The anchordevelops its tensile capacity by means of a combi-nation of keying into the concrete and highfriction between the sleeve and concrete. Thespacer sleeve aids in increasing the shear capacity.Tensile capacity depends on the strength of thebolt and its depth of embedment.

2.5.2.2 Sleeve anchors- The sleeve anchorconsists of a steel stud, an expansion sleeve usuallymade of sheet metal, and a nut and washer (Fig.2.13). The bottom of the steel stud has auniformly tapered mandrel which has the samediameter at the end as the expansion sleeve. Theentire length of the bolt below the washer isenclosed in a section or sections of the steeltubing. The bottom of the expansion sleeve is slitlongitudinally to provide for expansion. When thenut is tightened, the tapered mandrel moves intoand expands the sleeve which in turn bears againstthe wall of the hole. This anchor is used formedium and light holding requirements.

2.5.2.3 Shell expansion anchors - The shellexpansion anchor, (Fig. 2.14) is available in twotypes. One type consists of a two-piece shell heldtogether by steel tabs with a tapered, internallythreaded end plug. The second type consists of atwo-piece shell section with two tapered steelcones, one at the top end and one at the bottom,which are held together by a steel spring at thecenter. The bottom cone is internally threaded toaccept a bolt or stud. By torquing the fastenerinto the anchor, the steel cones expand the shellto bear against the wall of the hole.

2.5.2.4 Wedge anchors-The wedge anchor,(Fig. 2.15) consists of a steel stud bolt with a nut


BEFORE AFTERTORQUING TORQUING

Fig. 2.15- Wedge anchor

Grout hole

Threaded rob

Nut

Air tube

Plate

Hollow bar

Grout hole

Thrust rl

Mal leablee shell

Fig. 2. I6 - Rock/concrete expansion anchor(grouted)

BEFORE AFTER

Fig. 2.17-Drop-in anchor

BEFORE

b

0.

AFTER

Fig. 2.18 -Self-drilling anchor

2.5.2.5 Rock/concrete expansion anchor-Therock/concrete expansion anchor, (Fig. 2.16) con-sists of a stud bolt that is threaded on the top endfor a hex nut. The bottom end consists of a largemechanical expansion anchor. To set the expan-sion anchor, the stud bolt is rotated in a clockwisedirection. Grouting is optional down the center ofthe bolt to fill the annular space between the rodand the drilled hole for corrosion protection.

2.5.2.6 Drop-in anchors-The drop-in anchorconsists of a steel shell and an internal steelexpander plug (Fig. 2.17). The anchor is internallythreaded at the top end while the internal end ismachined to a uniform taper, matching the shapeof the steel plug inside the anchor. The lowerportion of the shell is slit longitudinally into equalsegments to allow the anchor to expand when theinternal plug is hammered with a setting tool. Byhammering the plug into theportion of the shell expands towall of the hole.

shell, the lowerbear against the

2.5.2.7 Self-drilling anchors-The self-drillinganchor, (Fig. 2.18) consists of a steel shell and atapered steel end plug. The bottom of the shellhas teeth for cutting its own hole in the concrete.The top of the shell is internally threaded toaccept a bolt or stud. The bottom of the shell isexpanded by hammer drilling the anchor over thesteel plug. The plug expands the bottom of theshell which bears against the wall of the drilledhole.

and washer. The bottom of the steel stud has auniform tapered mandrel around which is posi-tioned an expandable steel clip or separate steelwedges with protrusions. When the nut istightened, the clip or steel wedges ride up on thetapered mandrel, wedging between the mandreland the wall of the hole.

355.1R-9

BEFORE AFTER

b 4.

. n.

D

Q *

. .

V .v

v -.

.4

..

V

’ .t

.I

v -v

A .b

Q V’ .

Fig. 2.19 - Stud anchor

BEFORE A F T E R

Fig. 2.20 - Undercut anchor

2.5.2.8 Stud anchors -The stud anchor con-sists of a steel stud, threaded at the top end, andhas a drilled hole with longitudinal slits at thebottom end, which accepts a tapered steel plug(Fig. 2.19). The top of the threaded section israised to provide a surface for hammering. Byhammering the top of the stud, the tapered plugexpands the bottom end of the bolt causing it tobear against the wall of the hole.

2.5.2.9 Undercut anchors-There are twoprimary designs of undercut anchors available(Fig. 2.20). They all operate by keying andbearing against an undercut in the concrete at thebottom of the drilled hole. They cause little or noexpansion force in the concrete, but generate hightensile-loading capacities.

The first type requires a second drillingoperation to create an undercut at the bottom ofthe first drilled hole. The anchor is installed withthe bottom of the expansion sleeve at the under-cut . When the nut is tightened, the taperedexpander plug expands the bottom of the steelexpansion sleeve into the undercut.

The second type cuts its own undercut at thebottom of the drilled hole. A sleeve is hammeredby a rotary hammer drill with a special settingtool. The bottom of the expansion sleeve is drivenover a cone at the bottom of the hole. Thebottom of the expansion sleeve has a sharp edgewhich, on expansion, cuts its own undercut intothe wall of the hole. By tightening the nut, the

bolt and tapered cone are drawn up into theexpansion sleeve, keeping the bottom of theexpansion sleeve in the undercut.

A

CHAPTER 3-BEHAVIOR OF ANCHORS3.1 - Introduction

Understanding anchor behavior is necessary inspecifying the appropriate anchorage for a givenapplication. This includes an understanding offailure modes and strengths as well as load-displacement and relaxation characteristics ofvarious anchor types. This chapter covers anchorbehavior in uncracked concrete and in crackedconcrete.

Anchors are primarily loaded throughattachments to the embedded anchor. Theloading can be in tension and shear orcombinations of tension and shear (Fig. 3.1).
They may also be subjected to bending dependingon the details of shear transfer through theattachment. The behavior of anchors in tension isof primary importance and will be discussed first.


tension loading combined tensionand shear loading

[ shear loading bending

Fig. 3.1 -Possible loadings of anchors

By far, most anchor testing to date has beenperformed in uncracked concrete. While crackingoccurs in almost all concrete, testing in uncrackedconcrete provides the basis for understandinganchor behavior.

3.2-Behavior of anchors in uncrackedconcrete

3.2.1 Load-displacement behavior and failuremodesfailure

under tension loading- The five primarymodes of anchors in tension are (Fig. 3.2):

a) steel failure b) pull-out failure c) concrete splitting failure

d) concrete cone failure e) spacing and edge cone failure

Fig. 3.2 - Typical failure modes of anchors loaded in tension

(a)(b)(c)(d)(e)

Steel failurePull-out failureConcrete splitting failureConcrete cone failureSpacing and edge cone failure

The various types of anchors have different dis-placement characteristics depending on preload,load transfer mechanism, and failure mode. Fig.3.3(a)-3.3(c) present three load-displacementgraphs. Fig. 3.3(a) gives the characteristic curvesfor headed and undercut anchors while Fig. 3.3(b)presents curves for torque-controlled, drop-in, andself-drilling expansion anchors. Fig. 3.3(c) gives
load displacement curves for adhesive anchors.The displacements shown represent the displace-ment (slip) of the embedded anchor and the de-formation of the concrete as well as the defor-mation of the anchor.
When a preload is applied to an anchor,typically by tightening the nut to a prescribedmoment torque, the displacement caused by anexternally applied load is affected. The preloaded


l o a d F [kN]

4 6 8 10Displacement s [mm]

load F [kN]

0 2 4 6 8d isp lacement s [mm]

I IIine anchor type bolt diameter anchorage depthmm mm I

Fig. 3.3(a) - Typical load-displacement relationships Fig. 3.3(b) - Typical load-displacement relationshipsof headed and undercut anchors (from Rehm, of expansion anchors under tension loading (fromEligehausen, and Mallee 1988) Eligehausen and Pusill-Wachtsmuth 1982)

d isp lacement [mm]

Fig. 3.3(c)- Typical load-displacement behavior of chemical anchors under tension and shear loading (fromEligehausen and Pusill- Wachtsmuth 1982)


anchor shows little displacement with increasingexternal loading until the preload in the anchor(and resulting clamping force on the concrete) isovercome. The preload has no effect on the ulti-mate static tensile capacity of the anchorage, butsignificantly reduces the anchor total displacement.

In the case of steel failure (Fig. 3.3(a), Line 3)the ductility depends on the relationship betweentensile strength and yield strength of the steel andthe anchor length. Inelastic displacements ofheaded anchors due to concrete deformationsunder the head may be expected at relatively lowloads unless preloaded. Increasing the bearingarea under the head may reduce inelastic displace-ments but will have little influence on the failureload [compare Lines 1 and 2 in Fig. 3.3(a)].Headed anchors that fail due to fracture of theconcrete will exhibit a brittle failure (Fig. 3.3(b),line 2).

The behavior of drop-in anchors is dependenton the magnitude of the expansion force createdin setting the anchor. When expanded properlyduring installation, high expansion forces areinduced and the load displacement curve mayremain almost linear up to failure [Fig. 3.3(b),Line 2).

The expansion force, at installation, of torque-controlled expansion anchors is smaller than thatof drop-in anchors and, therefore, the displace-ments are larger for equal loads. If the externalload exceeds the preloading force in the boltgenerated by the torquing during installation, thespreading cone is pulled further into the sleeve,leading to increased displacement. At failure thedeformations are much larger than for comparabledrop-in anchors [Fig. 3.3(b)].

Self-drilling anchors show larger displacementsin the total load range than torque-controlledexpansion and drop-in anchors [Fig. 3.3(b)]. Thishappens because load transfer is mainly bymechanical interlock which causes high pressureon the concrete and large concrete deformations.

The displacement behavior of undercut anchorsdepends primarily on the bearing area (undercutarea) and the installation torque. Thereforerelatively large deformations may be expected withsome undercut anchors while others exhibit elasticbehavior well above service load [Fig. 3.3(a)].

Adhesive anchors exhibit elastic behavior up tonearly maximum load [Fig. 3.3(c)]. While theload-displacement curves of adhesive anchorsexhibit relatively low coefficients of variation in

comparison to torque-controlled expansion anddrop-in anchors, the bond strengths vary con-siderably depending on the adhesive componentmix used and the installation procedure.

Under working loads all categories of anchorsshould behave elastically with little additionaldisplacement after installation. However, atultimate load a plastic behavior and in the case ofcyclic loading only a limited strength degradationis desired. Fig. 3.3(a)-3.3(c) show that the actualload-displacement behavior of the currentlyavailable expansion, undercut, adhesive, andheaded anchors differs somewhat from this plasticbehavior.

Under sustained loads displacements willincrease with time due to creep of concrete in thehighly stressed load transfer area (bearing area inthe case of headed or undercut anchors, contactarea in the case of expansion anchors, bondedarea in the case of adhesive or grouted anchors).

As an example, in Fig. 3.4 (see Seghezzi and

10* 10 102

Duration [Days]

Fig. 3.4 -Increase of displacement during sustainedloading

Vollmer, 1982) the displacements of a torque-controlled expansion anchor loaded with aconstant tensile force corresponding to approx-imately 70 percent of the static ultimate strength,are plotted as a function of load duration on adouble logarithmic scale. It can be seen that thedisplacement veloci ty ( tangent to thedisplacement-time curve) decreases with increasingtime and, therefore, the displacements approach alimiting final value. The increase in displacementsis smaller for lower sustained loads. If the load isincreased after a sustained load test, the displace-ment curve is rather steep until it reaches thestatic envelope which is followed thereafter. Fail-ure load and displacement at maximum load arenot negatively influenced by a previous sustainedload smaller than about 70 to 80 percent of thestatic failure load.


In principle, the same behavior is valid forcyclic loadings with up to 1 x lo6 load repetitionsand an upper load (where the cyclic load rangesbetween an upper and lower value, both of whichare tension) smaller than about 50 percent of thestatic failure load (provided no fatigue failure ofthe bolt occurs). For higher upper loads the dis-placements may increase significantly and a fatiguefailure of the concrete might occur (Rehm,Eligehausen, and Mallee 1988).

Sustained and cyclic loadings in the working-load range have the same influence on dis-placements and ultimate loads of headed anchorsas for expansion and undercut anchors.

3.2.2 Relaxation -If headed anchors are pre-loaded, the initial force induced in the anchor isreduced with time due to creep of the highlystressed concrete under the anchor head. Thefinal value of the tension force in the anchordepends primarily on the value of bearing stressesunder the head, the concrete deformation and theanchorage depth. In typical cases the value ofthat final force will approach 40 to 80 percent ofthe initial preload (40 percent for short anchors,80 percent for long anchors).

Torque-controlled expansion anchors areusually preloaded by tightening the nut duringinstallation. This preload is essential for theproper performance of such anchors. In a typicalinstallation, locally high concrete stresses arecreated around the embedded anchor wedges orexpansion devices as the anchor is preloaded.Creep of concrete under these high stresses resultsin a slight movement of the embedded anchor,and in turn, in a reduction in the load in the bolt.Fig. 3.5 shows a typical load-relaxation test

0 I I I I0 10 20 30 40 50 60 70

Time [Days]

Fig. 3.5 -Reduction of preload as a function of time(after Burdette, Perry, and Funk 1987)

(Burdette, Perry, and Funk 1987). Preload isplotted as a function of time. The shape of thecurve is essentially the same for all anchors(including headed anchors). There is anexponential drop-off of load immediately after theapplied tension is released, followed by acontinued gradual diminishing of the load over anindefinite period. It is estimated that the finalpreload will be about 40 to 60 percent of theinitial value. This is confirmed by other test data(Seghezzi and Vollmer 1982, and Wagner-Grey1976). After retorquing the anchors, the processof load relaxation starts again, however, the finalvalue of the preload is increased (Fig. 3.6).

1 i .I ITorque Controlled Expansion Anchor M12I I

0 I0 2,5 5,0 7,5 10 12,5

Time [h]Fig. 3.6 -Influence of retorquing on the final valueof preload (from Seghezzi and Vollmer 1982)

Retorquing even a short time after anchor in-stallation can be effective (Wagner-Grey 1976).

Chemical anchors are usually preloaded byapplying a predefined torque. Because of the highstresses in the adhesive bond, the preload force inthe anchor declines faster and the final value isless than for torque-controlled expansion andheaded anchors.

Long-term relaxation and creep has beeninvestigated in several studies. Four Ml6diameter polyester anchors tested at loads of 25,30, 38, and 40 kN (6, 7, 8.5, and 9 kips), showeddisplacements still increasing after 5 years, butranging from 0.090 to 0.140 mm (0.0036 to 0.056in.)(Elfgren, Anneling, Eriksson, and Granlund1988). Creep tests were also performed on 26Ml6 anchors for 3 years at various loads and


environmental conditions. At allowable workingloads of 15 kN (3.4 kips), anchors tested indoorsshowed small creep, 0.10 to 0.40 mm (0.004 to0.016 in.). However, anchors tested outdoorsexhibited continually increasing creep. Thosetested indoors at 30- and 45- kN (7 and 10 kips),loads exhibited continually increasing creep. A 4month test on epoxy anchors showed creep lessthan 0.009 in. (0.2 mm) (Wiewel 1989).

The U.S. Army Corps of Engineers performedcreep tests on polyester and epoxy anchors,subjecting the anchors to 60 percent of the anchorsteel yield strength for 6 months. Cement andepoxy grouted specimens exhibited low slippage,0.0013 to 0.0008 in. (0.03 to 0.02 mm), whilepolyester anchors exhibited approximately 30 timesas much movement, 0.008 to 0.024 in. (0.2 to 0.6mm) (Best, Floyd, and McDonald 1989).

3.2.3 Ultimate strength in tension3.2.3.1 Steel failure -The strength of anchor

steel controls failure when the embedment of theanchor is sufficient to preclude concrete failureand when the spreading forces are sufficiently high(expansion anchors) or the bearing area is suffi-ciently large (headed and undercut anchors) topreclude an anchor slip failure. The failure mode[Fig. 3.2(a)] is rupture of the anchor steel withductility dependent on the type of anchor steeland embedment length. The ultimate strength canbe determined from Eq. 3.1.

F u = 4 x f,,, lb (3 .1)

whereAs = tensile stress area, in.*fut = ultimate tensile strength of steel, psi

For given material properties and anchordimensions this case defines the upper limit forthe tensile-load-carrying capacity.

Fig. 3.7 shows a comparison of the failure loads

number

of specimens

10 -

STEEL FAILURE 5 -

Fig. 3.7-Ratio of actual to predicted tensile capacityaccording to Eq. (3.1) for steel failure (after Klingnerand Mendonca 1982)

of headed anchors measured in tests to the valuespredicted by Eq. 3.1. Because the theoreticalfailure load was calculated with the nominal steelstrength, the ratios of actual to predicted tensilecapacity are larger than one.

3.2.3.2 Concrete cone failure -When theembedment of an anchor or group of anchors isinsufficient to develop the tensile strength of theanchor steel, a pullout cone failure of the concrete[see Fig. 3.2(d)] is the principal failure mode.When the spacing of anchors or location of anedge [Fig. 3.2(e)] interferes with the developmentof the full cone strength of an anchor, its capacitywill be reduced.

The angle of the failure cone, measured fromthe axis of the anchor, varies along the failuresurface and shows considerable scatter. In ACI349, Appendix B, ACI Committee 349,1985) theangle of the failure cone of headed and expansionanchors is assumed as 45’. According to Cannon*,in the case of expansion anchors the angle variesfrom about 60’ for short embedments (Id ( 2 in)to 45O for 1, 2 6 in. According to Rehm,Eligehausen, and Mallee 1988, the angle variesbetween approximately 50° and 60”, (mean value5S”) and tends to decrease with increasinganchorage depth.

The following formulas have been developed todescribe behavior of headed studs, expansion, andundercut anchors.

*Cannon, Robert W., correspondence to ACI Committee 355, Nov.1986.

Cannon, Robert W., correspondence to ACI Committee 355, Sept.1988.

This correspondence is filed at ACIACI headquarters and is available

at cost of reproduction and handling at time of request.

ANCHORAGE TO CONCRETE 355.1R-1 5

ACI 349, Appendix B, limits the tensile capacityof the cone failure of an anchor, or wup of.anchors, to a uniform stress of 4 (psi) on+d$the stress cone surface of the anchors.

(3.2)

strength reduction factor0.85 for uncracked concrete

= 0.65 in zone of potentialcracking

A = the summation of the projected areas (in.2) of individual stresscones minus the areas of over-lap and of any area, or areas,cut off by intersecting edges.

Note: Other reductions are made based onmember thickness relative to embedment and thearea of fabricated anchor heads (see Fig. 3.8).

ACI 349 has no requirements for minimumcenter-to-center spacing of single anchors oranchors belonging to a group.

Fig. 3.9 shows the frequency diagram of the

-A Frequency [%]

n = 45 tests5i = 1,14v = 26 O/o

20

10

1,5 2,0 2,5F /F u,test u,pred

Fig. 3.9 -Ratio of actual to predicted tensile capacityof headed anchors according to Eq. (3.2) (fromCannon, 1984 **)

ratio of actual to Predicted tensile capacity ofheaded anchors. Theoretical capacity wascalculated according to Eq. (3.2). The tests weredescribed by Klingner and Mendonca (1982a), andwere evaluated by Cannon*. Tested wereindividual anchors with large and small edgedistances and anchor groups. In all tests aconcrete cone failure occurred.

If an anchor is installed too close to an edge,the anchor will fail before developing the concretecone strength. Therefore, for headed anchors,ACI 349 requires that the minimum edge distancem to the center of the anchor be sufficient toprevent a side cone failure. The followingequation is suggested in the ACI 349 Commentaryfor determining this minimum value.

m , in. (3.3)

where

D = anchor diameter, in.F = ultimate tensile strength of anchor, psif 'c = compressive strength of concrete, psi

If this requirement cannot be satisfied, stirrupor tie reinforcement should be provided.

Cannon+ found that for embedments less than6 in., ACI 349 becomes increasingly conservativewith decreasing embedment. He has proposed amodification to Eq. (3.3) to provide a better fit totest data. For embedments less than 6 in., thismodification would increase the angle of thefailure cone, measured from the axis of theanchor.

For 1, c 3 in.: cy = 62 - 1.1 (l#, degFor ld 2 3 in. but < 6 in.: (Y = 45 + 0.79 (6-ld)

(3 3,

deg (3 5).

With respect to the minimum edge distance hereported the results of tests which indicated adirect relationship between anchor load and sidecone failure.** He suggested Eq. (3.6) instead ofEq. (3.3) as a more correct lower bound for theedge distance for headed anchors:

m

Fut = ASTM-specified tensile strength of theanchor bolt, kips

*Cannon, private correspondence, 1988, previously cited (seefootnote p 14).

+Cannon, private correspondence, 1986, previously cited (seefootnote p 14).

**Cannon, Robert W., Letter to ACI 355, “Comparison of TestingEdge Conditions and Anchor Spacing with Predictions”, Dec. 1984.

*EFFECTIVE STRESS AREA,

\L DEDUCT AREAOF ANCHOR HEADS

*REDUCE BY THE TOTAL BEARING AREA OF THE ANCHOR STEEL.

. A) Effective stress area for anchorage pul lout

L d*EFFECTIVE STRESS B I41

AREA

A t

P L A N

Pd

t

L EFFECTIVE STRESSAREA

Pd

t

L J

(a+2Ld-2h)

E F F E C T I V ESTRESSAREA . STRESS AREA REDUCTION FOR LIMITED DEPTH (Ar)

Ar= (a+2Ld-2h)(b+2Ld-2h)

*REDUCE BY THE TOTAL BEARING AREA OF THE ANCHOR STEEL

B) Stress area reduct ion for I imited depth A

Fig. 3.8-ACI 349 method for determining effective stress areas


The average failure load for a side cone(bursting) failure is given as:

whereF, = 15m

f- kips (3.7)35cCo’

m = actual edge distance, in.

For expansion and undercut anchors,Eligehausen, Fuchs, and Mayer (1987 and 1988),derived Eq. (3.8a) from 287 test series with singleanchors with large edge distances showingconcrete cone failure.

(3.8a)

where

Fu = average ultimate load, N‘d = embedment depth (see Fig. 3.10), mm

Fig. 3.10 -Illustration of embedment depth as usedin Eq. (3.8a) and (3.86)

(3.8b)

f’, = average compressive strength of con-crete cylinders (6 by 12 in.) at time oftesting, N/mm2

Results of an additional 196 tests on headedstuds showed a similar relationship (from Rehm,Eligehausen, and Mallee 1988).

In the original equation the concrete strengthwas measured on cubes with a side length of 200mm (8 in.). Eq. (3.8a) and (3.8b) assume f 'c(cylinder) = 0.82 f 'cc (cube).

The tests with expansion, undercut and headedstuds included anchorage depths from 40 to 525

mm (1 9/16 to 20 l/2 in.) and concrete strengthsf’, = 20 to 50 N/mm2 (2900 psi to 7150 psi). Fig.3.11 shows a histogram of the ratio of measured to
predicted failure load.
The average failure loads given in Eq. (3.8) canonly be obtained if the distances between anchorsare large enough so that concrete cones do notoverlap each other. Assuming an angle of thefailure cone cy = 55o the critical distance isapproximately three times the embedment depth.The failure load of a two-point fastening resultsin:

whereG = xcr x F,, (3.9)

Ful = ultimate failure load, singleanchor, from Eq. (3.8)

& = 1 +a/a,,it I 2 (3.10)

where

a = distance between center ofanchors

acrit = critical distance between centerof anchors

= 31,, where 1d is the depth ofembedment.

Eq. (3.9) leads to the x-method for calculatingthe ultimate capacity of multiple anchor fasten-ings. For the calculation of the ultimate load ofquadruple fastenings the xa factors can be derivedseparately for both directions and combined inproduct form as follows.

where

FIA4 = %a1 x Xd x Fur (3.11)

xai = 1 + ai(acd 5 2 (3.12)

a.I = spacing in direction i


I

0.5 1.0 1.5 2.0$, test ’ %,pred

Fig. 3.11 (a) -Ratio of acutual to predicted tensilecapacity for concrete cone failure of individualexpansion and undercut anchors away from edgesaccording to Eq. (3.8a). (from Rehm, Eligehausen,and Mallee 1988, and Eligehausen, Fuchs, andMayer 1987 and 1988)

40 Frequency [%]

n = 196 individual testssi= 1 0 0v= 1 4 %

30

20

10

0.5 1.0 1.5Fu, test /Fu, pred

Fig. 3.11(b) -Ratio of actual to predicted tensilecapacity for concrete cone failure of individualheaded anchors away from edges according to Eq.(3.86). (from Rehm, Eligehausen, and Mallee 1988)

Fig. 3.12 shows the capacity of quadruple
fastenings for headed studs, expansion andundercut anchors as a function of the ratio ofanchor spacing to embedment depth as measuredin tests and calculated according to Eq. (3.11).
Eq. (3.9) and (3.11) can also be extended formultiple anchorages with any number of anchorsin any spacing by setting the value of ai as thedistance atot between the outer anchors, and thex0- value is limited to xa I n with n = number ofanchors in one direction. This is provided that thespacings between the individual anchors aresmaller than acrit = 31, and the anchor plate issufficiently stiff to assure an even distribution oftension forces to all anchors (see Rehm,Eligehausen, and Mallee 1988). The X-methodcan also be extended to take account of loadeccentricities (Riemann 1985).

Fig. 3.13 shows the ratio of actual to predicted
tensile capacity of groups of headed studs. In thetests the number of anchors was varied between 4
and 36, the spacing of the outer anchors between100 and 875 mm and the spacing of the individualanchors between 0.541, and 2.2&. The groupswere loaded by a concentric tension load whichwas equally distributed to all anchors.

Eq. (3.13) covers the influence of edge dis-tances, a,, smaller than critical:

where

Xa?n

am,crit ==

‘=

Fu =

Fu* = a& * Fy (3.13)

= 0.3 + 0.7 am/a,crit S 1 (3.14)

critical distance from free edge1.5 1d

actual embedment lengthultimate failure load, single anchorto be taken from Eq. (3.8)

355.1R-19

05.0

4.0

3.0

2.0

1.0

I IFE according to eqn. ( 3.8 )

O

8I 0

Fig. 3.12-Ratio of actual failure load of a group of anchors to the predicted value for an individual anchor asa function of the ratio of anchor spacing to embedment depth (from Rehm, Eligehausen, and Mallee 1988)

MANUAL OF CONCRETE PRACTICE

I L

,

l-_.-


Fig. 3.14 shows a comparison of test results with
the theoretical values according to Eq. (3.13). Itshould be noted, however, that minimum distancesfrom the free edge are necessary for headed studsin order to allow proper concreting and avoidlocal spalling of concrete. Minimum edgedistances for expansion and undercut anchors arenecessary to avoid splitting of concrete duringinstallation and expansion of the anchors.
If anchors are located in a corner [see Fig.3.15(b,)], the factors xarn are calculated separatelyfor each direction and then the two x-factors aremultiplied.

Fig. 3.15 - Typical failure modes of anchors Loadedin shear (from Rehm, Eligehausen, and Mallee1988)

Bode and Roik (1987), evaluated data of 106tests with headed studs to arrive at Eq. (3.15).

F” = 12r,3/2(1 + d&,) 8, N (3.15)

where

F, =1d =d, =f’, =

average failure load, Nembedment length, mmhead diameter, mmconcrete cylinder strengthat time of testing, N/mm2

Fig. 3.16 compares the measured failure loads

kNTU’k lN/mmz I

mean value

50 75 100 125 150h [mm]

Fig. 3.16 - Measured failure loads compared to Eq. 3.15 (where p, = concrete splitting strength) (from Bode andRoik 1987)

of headed studs with the values according to Eq.

Roik (1987), assume the criticalspacing of neighboring headed

acl+ = 41, (3.16)

.50

.00.

o Anchorstuds,

concrete break-out/

/ l Headed studs,

0 local concrete failureL ( blow -out) I

00 .50 1.00 1.50 1.75

Fig. 3.14-Ratio of actual failure load of an individual anchor close to the edge to the predicted value for ananchor with large edge distance (from Rehm, Eligehausen, and Mallee 1988)

355.1R-23

With respect to the influence of free edges (seeFig, 3.15) they consider the critical distancebeyond which there is no significant influence on

load as being in the case of one free

ati1 IJ 1.21,

and in the case of two or more free edges:

acit.2 5 21,

For distances from center of headed stud to thefree edge(s) which are smaller than the criticaldistance according to Eq. (3.17) and (3.18), theyfou d that the assumption of a linear decrease ofulti

”

ate failure load in proportion to the ratio ofact al distance/critical distance gives a lowerbound of their test results, in much the same

ner as shown in Fig. 3.14.raestrup, Nielson, Jense, and Bach (1976), give

the predicted failure load as:

FM = 0.21 x 2; (1 + d,ll&f$ N (3.19)

Eq. (3.19) was deduced by applying the theoryof plasticity to headed studs embedded inco rrete.

nI

The failure load is assumed to bepro ortional to the concrete compressive strength.

3.2.3.3 Pullout (slip) of the anchor- Slipfailure occurs [Fig. 3.2(b)] with expansion anchorswhen the expansion force is too small to developeither the strength of the anchor steel or a shearcone failure of the concrete. This is a typicalfailure mode for wedge anchors at moderate todeep embedments in lower strength concretewhere the crushing of the concrete at the wedgesallows the bolt to “pull through”. The cause mayalso be due to an oversize hole. Slip failure mayalso occur in low strength concrete due todeformation of the wall of the hole.

The testing of wedge bolt expansion anchors byHanks (1973), clearly demonstrated that theprimary failure mode for individual anchor tests(uninhibited by edge conditions) was either conefailure of the concrete or anchor slip dependingon the depth of anchor for a given size. Only 10of 464 tension tests indicated any crackingassociated with a cone failure. The line ofdemarcation between shear cone failure and slip

failure was approximately six bolt diameters.Under conditions of poor workmanship in thefield (e.g., oversize holes) slip failure may occur ata much smaller embedment depth than ld = 6D.

Slip failure may also occur with bonded andadhesive anchors of insufficient embedment todevelop the strength of the anchor steel or tocause a concrete cone failure.

Torque-controlled wedge anchors, which fail byslip, generally fail by slipping the expansion conepast the wedges. This failure mode may alsooccur with sleeve anchors. However, in some caseanchors may fail by pulling the whole anchor(including expansion sleeve) out of the hole.Torque-controlled expansion anchors may also slipto a critical depth and fail the concrete.Deformation-controlled expansion anchors (e.g.,drop-in anchors) have a fixed expansion and mayslip to a critical depth and then fail the concrete.

The slip failure load is dependent on thecoefficient of friction between the sliding surfacesand on the spreading force at failure which is afunction of the critical expansion force producingfailure and the deformability of the concrete whichvaries with hole depth and concrete properties.All of these factors may vary with anchor type,manufacturer, and installation. The spreadingforce and thus the slip load of drop-in anchorsdecreases significantly with increasing diameter ofthe drilled hole with respect to the diameter of theanchor.

Theoretically the slip failure load F, could becalculated from Eq. (3.20).

Fit = ps (3.20)

where

I, = coefficient of frictionS = spreading force

The coefficient of friction depends mainly onthe roughness and cleanliness of the drilled holeand of the surface of the expansion sleeve orwedge as well as on the spreading pressure. FromWagner-Grey (1976), the factor p for torquecontrolled expansion anchors is in the range of 0.2to 0.3 and for drop-in anchors is approximately0.35. The difficulty in using Eq. (3.20) lies inproperly estimating the spreading force, sincecomplex mechanics are involved. For this reason


the profession relies on test data. However,equations for estimating of the spreading force aregiven by Wagner-Grey (1976).

Because of the large variability of the spreadingforces and the coefficient of friction, Eq. (3.20)gives only an approximate estimate of the pulloutload (see Eligehausen, and Pusill-Wachtsmuth1982). Furthermore, in important applications itis advisable to test expansion anchors, whichtypically fail by slip at specified embedments, indesign strength job concrete to confirm slipcharacteristics.

For pullout failures of a chemical anchor, thebond between the wall of the drilled hole and themortar is critical (see Sell 1973). Assuming auniform bond stress distribution along theanchorage length, the bond strength is in the orderof 1300 psi (9 MPa) with a coefficient of variationof 10 to 15 percent for polyester and vinylesterchemical anchors. This value is for a concretecompressive strength of 3000 psi (21 MPa) and anembedment of about nine anchor diameters. Thebond strength increases approximately with thesquare root of the concrete strength.

The pullout capacity of chemical anchorsincreases with increasing embedment depth:however, after about nine anchor diameters theincrease is not proportional to embedment. Thisis due to the high bonding effect resulting in highload transfer to the concrete at the top of theanchorage. The bond stress is no longer uniform,and if the tensile load is sufficiently high, thefailure initiates with a concrete failure in theupper portion of the concrete and then the bondfails in the remainder of the embedment.

For headed anchors local failure in front of thehead will occur when the pressure on the concreteis larger than about 12f’, to 15f’, (Rehm, Elige-hausen, and Mallee, 1988). This type of failure issomewhat similar to a pullout failure.

3.2.3.4 Splitting failure of concrete -Thisfailure mode will occur only if the dimensions ofthe concrete are too small, the anchors are placedtoo close to an edge or too close to each other[Fig. 3.2(c)], or the expansion forces are too high.The failure load is usually smaller than for aconcrete cone failure.

Torque-controlled expansion and deformation-controlled anchors (e.g., drop-in and self-drillanchors are the type anchor most likely toexperience splitting failure due to the high lateralthrust required to resist sliding by friction on the

steel wedges. Deformation-controlled expansionanchors generate higher spreading forces andrequire larger edge distances than torque-controlled expansion and undercut anchors.

The capacity of expansion anchors which fail bysplitting of the concrete has been evaluated byPusill-Wachtsmuth (1982), using theoreticalconsiderations. It was assumed that splittingoccurs when the tensile stresses averaged over acritical area reach the concrete tensile strength.The size of this area was found by evaluating theresults of tests with concentrated loads and oftests with thick concrete rings subjected to aconstant inner pressure. According to this theory,the necessary side cover or spacing to preclude asplitting failure before reaching the concrete conefailure load must be about 1.751d or 3.51,, respec-tively. For drop-in anchors a side cover m I 31dwas recommended. The validity of this evaluationwas checked by relatively few test results.

With respect to the minimum edge distanceCannon* has proposed the following criteria topreclude a splitting failure occuring at a loadlower than the capacity for concrete cone failureor pullout failure:

m = D(11.4 - 0.92& in. (3.21)

where

:= minimum edge distance= anchor bolt diameter, in.

ld = embedment depth to the bottom of theanchor, in

Eq. (3.21) is valid for anchor spacings s L 2 in.

If side cover or spacings of anchors are toosmall, splitting cracks may occur during installationof anchors. This possibility is greater for drop-inanchors and for self-drilling anchors than fortorque-controlled expansion anchors because ofthe higher initial spreading forces. The minimumedge distance and the minimum spacing to avoidsplitting during installation, as recommended byRehm, Eligehausen, and Mallee (1988), are basedon many tests and are given in Table 3.1 for the
different types of anchors.
*Cannon, Private correspondence previously cited Dec. 1984(see footnote p 14).


Table 3.1 -Minimum edge distance and minimum spacing to avoid splitting failure

Mihimum edge distance m / 1d to avoidsplitting during installation

Minimum center-to-center spacing a / 1dto avoid splitting during installation

Undercut anchors

1.0

1.0

Torque-controlled expansion anchorsI

Drop-in anchorswith one cone (recent design)

2.0I

3.0

1.0

3.2.4 Load-displacement behavior and failuremodes in shear-For anchors with an appliedpreload, the initial friction forces between thebaseplate and the concrete have to be overcomeby the shear load before there is initial anchormovement (Fig. 3.17). The baseplate slides and

Onset of bearingcrushing in the concrete

lip of loading plate intobearing on anchor stud

Load transfered byfriction to embedment

. ~~~ r r -7

0 .50 10 1.5 0 20Deformatlon

Fig. 3.17- Typical load-displacement curve forwedge anchor in shear from Meinheit andHeidbrink 1985)

the anchor moves to the side of the hole in thesecond stage of behavior. The third stage of load-displacement behavior is a pressure loadingagainst the top surface of the concrete and asurface spa1l of the concrete at the edge of thehole. Depending on edge distance and anchorembedment, the failure may be by shearing of theanchor (for deep embedments) with or without aconcrete spa11 preceding the steel failure [Fig.3.15(a)] or by shearing of the concrete (concretefailure) in the case of anchors loaded near anedge [Fig. 3.15(b1), (b2), (b3)].

Shear loading generally produces largerdisplacements than tension loading [see Fig.3.3(c)]. This can be attributed to the bending ofthe anchor rod and the deformation of theconcrete in the direction of loading. This isespecially true if the anchor is not flush with theconcrete at the hole opening (e.g., when theconcrete is spalled during drilling). For cast-in-place anchors, the behavior will depend on thetype of anchorage used, the embedment and thesteel strength.

The distribution of shear from the attachmentto anchors of a group depends on the details ofthe anchors to the attachment connection and onovercoming the frictional resistance of theattachment. The frictional resistance depends onsurface conditions, the existing preload (if any) inthe anchors and the compressive forces applied

through the attachment as a result of direct loadsor applied moments. The connection detailsconcern the treatment of connecting surfaces andthe fit and manner of connecting the anchors tothe attachment.


3.2.5 Ultimate strength in shear3.2.5.1 Steel failure - Steel failure usually occurs

after relatively large displacements and is mostcommon for deep embedments, lower strengthsteels and large edge distances. The failure loaddepends on the steel area and the steel strength and is given by Eq. (3.22).

F8l = N A,f,, lb (3.22)

where the factor N takes account of the steel“shear” strength and has the range 0.6 to 0.7[Klingner and Mendonca, (1982b)], A, is the ten-sile stress area (as defined in Eq. (3.1)) and f,t isthe ultimate tensile strength.

Eligehausen and Fuchs (1988), propose thevalue N = 0.6 on the basis of an evaluation of230 tests.

3.2.5.2 Concrete failure -Concrete failures willexhibit two modes; (1) blow out cones due to edgeproximity (Fig. 3.15) and (2) concrete spa11followed by a possible anchor pullout or steelfailure away from an edge.

3.2.5.2.1 Edge failure- For all types ofanchors loaded in shear toward an adjacent, freeedge and exhibiting a concrete failure (Fig. 3.15),the failure load is influenced by the concretetensile strength, the edge distance m and thestiffness of the anchor. Another influencing factoris the embedment depth. The failure surface hasa conical shape that may radiate from the em-bedded end of the anchor for shallow embedmentsor from the upper part of the anchorage for deepembedments.

In the following paragraphs, several formulasfor calculating the failure load for an edge failureare reviewed.

ACI 349, Appendix B, Commentary gives adesign shear strength of

vu = 24$$n2, lb (3.23)

wherecb = 0.85f’, = compressive strength of concretem = distance from anchor to free edge

(see Fig. 3.15)

Fig. 3.18, taken from Klingner and Mendonca

x

numberof specimens

”

I0,: 0.65-4

I

RATIO OF ACTUAL TOF’FQICTED CAPACITY

Fig. 3.18 -Histogram of actual to predicted capacity

(1982b) gives the ratio of actual to predicted shearcapacities for this approach.

ACI 349, Appendix B further recommends aminimum side cover or edge distance m requiredto preclude edge failures: be calculated by Eq.(3.24).

m = , in. (3.24)

whereD = anchor diameter, in.F tfr”,

= anchor ultimate tensile load, lb= concrete compressive strength, psi

Eligehausen and Fuchs (1988), have suggested,based on the evaluation of some 80 test resultswith headed and expansion anchors (anchoragedepth ld > 4D), the average ultimate failure loadof the concrete of a single fastener in shear becalculated by:

F,, = 1.4@$n1.s~~, N (3.25)

whereD

f’,

m

shank diameter (mm) of headedstuds or drill-hole diameter foranchors, D < 25mmaverage concrete compressivestrength (cylinders) at time of testing,N/mm2distance from anchor to free edge,mm


hXh =-sl

1.4mwhere

h = member thickness, mm

Eq. (3.25) is valid for 1,/D = 4 to 6.

Fig. 3.19 shows a comparison between failureloads according to Eq. (3.25) and test results. Thethickness of the test specimens was h 1 1.4m.The tests were performed in concretes withdifferent strengths and anchors ranging indiameter between 12 and 22 mm. The test resultswere normalized to a concrete strength f 'c =20N/mm2 and D = 18 mm.

If an anchor group is loaded in shear toward anedge, a common failure cone may occur [see Fig.3.15(b2)]. T he corresponding failure load mayalso be calculated as described in Section 3.2.3.2for tension loading [Eq. (3.9), (3.11), and (3.12)]according to the x-method. The x-values forshear loads, however, depend on the distance fromthe free edge measured in load direction.

The critical (minimum) distance between two ormore anchors beyond which no intersection offailure cone will happen is given by Eligehausenand Fuchs (1988), as:

where

aWit = 3.5m (3.27)

m = distance to free edge.

For a I a,,i, Eligehausen and Fuchs (1988),have proposed the calculation of the averagefailure load of a group of anchors (see Fig. 3.20)subjected to shear load by:

FI(, Group = x,F, (3.28)

where& = 1 + a/a,,i,F, is from Eq. (3.25)

Fig. 3.20 (Eligehausen and Fuchs, 1988) showsthe ratio of the failure load of a group loaded inshear towards the edge to the failure load of anindividual anchor calculated according Eq. (3.25).The failure load ratio is plotted against the ratioof spacing to edge distance.

0 3.0 ' 3.5 4.0a/a, [-]

Fig. 3.20-Ratio of actual shear failure load ofanchor group to shear failure load of an individualanchor as a function of spacing between anchors

Similar expressions are proposed for calculatingthe failure load of single fastenings or anchorgroups situated in a corner or in narrow members.The influence of load eccentricity on the failureload of an anchor group can also be taken intoaccount by the x-method (Rehm, Eligehausen,and Mallee 1988). The method has been extendedto anchor groups with an arbitrary number ofanchors.

Klingner, Mendonca, and Malik (1982),recommend a critical (minimum) edge spacing of:

mkD Fut-, in.

%@

(3.29)

where#C = 0.90 and the other terms are as given

for the ACI 349 [Eq. (3.24)].

For anchors with small embedment depthsituated away from an edge and loaded in shear,the failure mode may be a tensile cone failure asthe anchor bends under load and induces a tensileloading into the concrete. Because of ductilityrequirements and reversible load conditionsassociated with seismic design, ACI 349 does notdistinguish between embedment requirements forshear and tension. This is very conservative if onlyshear is considered (see Shaikh and Yi, 1985).


8cv 8


3.2.5.2.2 Concrete spall-Anchors away froman edge will locally spall the concrete in front ofthe anchor. The primary factors influencingconcrete spall due to shear are tensile strength ofthe concrete, stiffness of the anchorage, anchordiameter, embedment depth, and deformability ofthe concrete. The corresponding shear capacity isgiven by Klingner and Mendonca (1982), andAmerican Institute of Steel Construction (1978),as:

whereAb =

f’, =

E, =

However,

F, = 0.5 A, fit, lb (3.30)

nominal gross cross-sectional area ofanchor shank, in.2specified compressive strength ofconcrete, psielastic modulus of concrete, psi

according to Eligehausen and Fuchs(1988), the above described local concrete failuredoes not negatively influence the anchor steelcapacity (normal strength steel) and will not causesubsequent pullout of the anchor, provided theembedment depth is 1, L 4D.

3.2.6 Combined tension and shear Loading- Thebehavior of anchors under combined tension andshear loading lies in between the behavior undertension or shear loading, and for a given depth ofembedment, is dependent on the angle of theloading (Fig. 3.21).

125Load FQ[ kN ]

100

75

50

25

0I I I I

5 10 15 20 25Displacement A,[mml

Fig. 3.21- Shear load-displacement behavior ofheaded studs for different tension loads (from Bodeand Hanenkamp 1985)

To calculate the failure load under combinedtension and shear loadings three approaches are inuse; a straight-line function, a trilinear functionand an elliptical function.

There are two types of straight-line functions.The first is a shear friction approach used by ACI349, Appendix B, and given by Eq. (3.31).

(3.31)

whereTL? = applied tension load

r”= 4 F,= 0.85

F, according to Eq. (3.22)cr = coefficient of friction

= 0.55 to 0.9, depending on thelocation of the anchor plate inrelation to the concrete surface

Tall = allowable anchor tensile load

A second straight-line equation is given by Eq.(3.32).

T,/T, + VJVu s 1.0 (3.32)

where

Ta* va = applied tensile and shear load,respectively

T”, Vu = ultimate tensile and shear load,respectively

These straight-line methods give a conservativeapproach to combined loading analyses.

Bode and Roik (1987), propose for headedstuds a trilinear function:

vp, s 1

TJT,, + VJV,, s 1.2 (3.33c)

whereT,, Va, TU and Vu as defined for Eq. (3.32).

According to Meinheit and Heidbrink (1985),Eq. (3.33) is valid also for expansion anchors (seeFig. 3.22).


Many investigators have concluded that shearand tension combine in an elliptical function asgiven by Eq. (3.34).

(T,/T,,Y + CV,lVJ s 1.0 (3.34)

where exponents x and y are determined fromtests and the other terms are as previously definedfor the straight-line equations.

The PCI Design Handbook (1978) uses x = y =4/3 for precast anchors, while the TeledyneEngineering Services report (1979) gives x = y =5/3 as a good fit for expansion anchors.

Fig. 3.22 shows a comparison between testresults with expansion anchors and the differentapproaches as described above.

3.3-Behavior of anchors in cracked concrete3.3.1 Introduction -When anchors are installed

in the tension zone of reinforced concrete mem-bers, it must be assumed that cracks will occur inthe concrete because of the rather low concretetensile strength. The concrete tensile strength maybe totally or partially consumed by the restraint ofinduced deformations due to shrinkage, tempera-ture, or flexure, or from the anchorage itself.Cracks run either in one direction (single cracks)or in two directions (intersecting cracks, in thecase of slabs spanning two directions).

If concrete cracks, experience has shown thatthere is a high probability that the crack willpropagate through the anchor location (seeCannon 1981 and Eligehausen, Fuchs, Lotze, andReuter 1989). Theoretical considerations alsoindicate that cracks should propagate through theanchor location. When the anchor is loaded, theanchor creates splitting (tensile) forces at theanchor embedded end. These tensile stresses inthe concrete would add to other tensile stressesfrom locally high bending moments. (i.e., flexuralstresses and restrained shrinkage stresses). Forthe case when expansion or undercut anchors areused, the drilled hole can also act as a notch orproduce a cross section in the concrete memberwith reduced concrete area.

The theoretical considerations discussed above,were confirmed by testing Ml2 (12 mm) torque-controlled expansion anchors and undercutanchors in a slab reinforced with welded wiremesh (AJbd = 0.004) (see Eligehausen, Fuchs,Lotze, and Reuter 1989). The test anchors wereinstalled with 1d = 80mm (3.2 in.) and inuncracked concrete. The anchorage holes were

drilled either 40 mm (1.6 in.) or 80 mm (3.2 in.)away from the transverse acting wires, [spacing of250 mm (10 in.), in the fabric]. Bending of theslab was in one direction only. All test anchorswere pretensioned or pretensioned and loadedwith their allowable load before the slab wassubjected to flexural loadings.

After preloading the anchors, the concrete slabwas loaded to its service load. Observationsduring this part of the testing often showed thatcracking started at the section with transversereinforcement but then deviated from that sectionto the section that contained the anchor hole.The cracks propagating through the anchor holealso were to the depth of the hole (Fig. 3.23 and
3.24). Testing showed that the displacement characteristics of these anchors remainedessentially unchanged until the slab load was about40 percent of the slab service load. Beyond thatpoint, significant increased displacement occurred(Fig. 3.25). The increased displacement characteristics of the anchor in cracked concrete arecaused by the crack propagating through the loadtransfer zone of the anchor (see Cannon 1981).
The crack width can vary over the depth of themember (bending cracks) or can be of constantwidth (parallel cracks, e.g. due to tension loading).In the worst case the anchor can lie in the inter-section of two cracks with constant width over themember depth. If anchors are situated in or besidethese cracks, their load displacement behavior andstrength may be significantly influenced.

3.3.2 Load-displacement-behavior and failuremodes in tension -Fig. 3.26 presents typical load-
displacement curves of torque-controlled expan-sion anchors which were set in uncracked concreteand in cracks, and loaded statically to failure. Thedisplacements of anchors located in cracks behavesimilarly to anchors in uncracked concrete up to acritical load. This critical load depends on thetype of crack and the crack width. For higherloads the displacements of anchors in cracks aremuch higher than the values expected inuncracked concrete and anchor capacity is sig-nificantly reduced.
The load-displacement behavior of headed orundercut anchors may be affected by cracks inconcrete but the displacements at maximum loadare less influenced by cracks than are expansionanchors (see Fischer 1984).


I.0

0ll 0 0

Fig. 3.22- Tension-shear interaction diagram for expansion anchors (from Meinheit and Heidbrink 1985)


IFI______ _; ! F

I K 884 (8,84cm2/mI-_,,,:‘, _ _ _ __ _ 1.2

Ia -I

kc

I

15 , 100 150 1 150 100d I 15/ l I 1 , I L

torque-contro l ledexpansion anchors -7

undercutanchors

l anchor loadedl anchor prestressed but not loadedo drill hole

--z--

i

z

ic-

Fig. 3.23 - Torque-controlled expansion anchors and undercut anchors in the cracked tensile zone of a concreteslab (from Eligehausen, Fuchs, Lotze, and Reuter 1989)


ttension

I-A

Section A - A

jr expansionarea

A

Fig. 3.24- Crack pattern in a drilled hole with expansion anchor (from Eligehausen, Fuchs, Lotze, and Reuter1989)


-4-lF .

0.3 0.2crack width [mm]

0.1

8 adm

- 1.0

- 0.8

0.6

1r(I

0.4, 1 (H 0

0.2

,, , 4 torquee controlledexpansion anchors

10.1 0.2

displacement [mm]

Fig. 3.25-Crack width and anchor displacement as a function of the ratio of applied load to allowable load ofthe slab (from Eligehausen, Fuchs, Lotze, and Reuter 1989)

Force-v

r Torque Controlled Expansion AnchorTension Loading

Uncracked

Cracked Concreter---

DisplacementFig. 3.26-Influence of cracks on the load-displacement relationship of expansion anchors - schematically (fromRehm and Lehmann 1982)

Fig. 3.27 shows the typical load-displacement
[kN]
Torque Con It rolled Expac, sion Anchor

FTER CYCLIC LOADINGI

D CYC LES

5 15Displacement [mm]

Fig. 3.27-Influence of cyclic loading on the load-displacement relationship oftorque-controlled expan-sion anchors (after Rehm and Lehmann 1982)

relationship of torque-controlled expansionanchors set in intersecting cracks and cycled up to10’ times between different load levels beforeloading to failure. For comparison the load-displacement relationship for statically loadedanchors is also plotted. Provided the upper loadduring cycling is smaller than about 50 percent ofthe static failure load, cyclic loading results in analmost linear increase of the anchor displacementas a function of the logarithm of the number ofcycles. The load-displacement curve for higherloads than the upper load during cycling is rathersteep up to the static envelope which is followedthereafter. Anchor capacity and displacement atfailure are not influenced significantly by cyclicloading with an upper load as given above.

Opening and closing of cracks by cycling thereinforced concrete while subjecting the anchor toa constant load has more influence on the anchorbehavior than cycling the anchor with the crackskept open (Rehm and Lehmann 1982).

In principle the failure modes described inSections 3.2.1 and 3.2.3.1 are also valid foranchorages in cracked concrete. However,expansion anchors which produce a concrete conefailure in uncracked concrete may slip and pull outwhen located in a crack. This possible change ofthe failure mode is due to the reduction of thespreading force as a result of the cracks (seebelow).

3.3.3 Relaxation-Expansion and undercutanchors installed in cracks will show an initialdisplacement during widening of the crack. Theamount of this displacement is dependent on thedesign of the anchor and on the crack width.Usually this initial displacement is large enough toreduce the preload to zero. This is also valid forbonded anchors.

The relaxation behavior of headed anchorsinstalled in cracks has not yet been studied.However, one may assume that the residualpreload is not significantly smaller than for headedanchors in uncracked concrete.

3.3.4 Ultimate strength in tension-Fig. 3.28
shows the influence of cracks in the concrete onthe strength of headed and undercut anchorsplaced in or close to cracks. The ratios of thefailure loads of single anchors measured incracked concrete to the value in uncrackedconcrete are plotted as a function of the crackwidth. The anchors were tested in tensionspecimens with almost constant crack width overthe member depth. After installing the anchors inuncracked concrete or concrete with hairlinecracks, the cracks were opened by loading thespecimen and then the anchors were staticallyloaded in tension with the cracks open. Failureoccurred by pulling out a concrete cone.


mm

Fu (crack) / Fu (uncracked c o n c r e t e )I,OA I

fi- 20-55N/mm2

id = 8O mm

0,4 0,8 1,2 1,6crack width A w [mm]

Fig. 3.28 -Influence of cracks on the ultimate loadof undercut and headed anchors (from Eligehausen1984)

The failure load decreases rapidly up to a crackwidth of about 0.4 mm (l/64 in.) and is almostconstant for larger cracks. The scatter of the datais relatively large. On an average, the ultimateload of anchors installed in or beside cracks witha width > 0.4 mm (l/64 in.) is about 60 percent ofthe ultimate value in uncracked concrete. Itshould be noted that, under service load, crackswith a width no greater than 0.4 mm (l/64 in.) aretolerated in reinforced concrete structures. Theinfluence of the type of anchor (headed or under-cut) on the failure load reduction is negligible.An almost similar strength reduction was alsoobserved with anchors installed deeper in thetension zone of beams for various anchor-depth-to-beam-height ratios (Rehm, Eligehausen, andMallee 1988).

The reduction of the anchor strength is due tothe change of the stress distribution in theconcrete caused by cracks (Eligehausen 1984 andEligehausen, Fuchs, and Mayer 1987 and 1988).In the case of uncracked concrete, the stresses inthe concrete are radially symmetric to the anchorand tensile hoop stresses are caused by the loadtransfer into the concrete [Fig. 3.29(a)]. If theanchor is installed in a crack, tensile stressescannot be transferred across the crack. Therefore,the area which can be used for transmitting theload into the concrete is smaller than in uncrackedconcrete [Fig. 3.29(b)].

uncracked concrete b) cracked concrete

Fig. 3.29 - Load transfer into concrete schematically for a) uncracked concrete and b) cracked concrete (fromEligehausen, Fuchs, and Mayer 1987, 1988)


Furthermore, a part of the concrete cone maybe cut off by neighboring cracks. These combinedeffects cause a strength reduction of approximately40 percent compared to uncracked concrete.Some tensile stresses can be transmitted oversmall cracks due to aggregate interlock(Eligehausen and Sawade 1985). This explains theincreasing anchor strength for crack widths lessthan 0.4 mm (l/64 in.).

In addition to the above effect, the reduction ofthe spreading forces by the crack opening must betaken into account for expansion anchors (Fig.3.30). If the anchor lies in an intersecting crack,the widening of the crack by the width w leads toa reduction of the effective expansion displace-ment around the circumference of the anchor byw/2 [Fig. 3.30(a)]. Assuming elastic behavior ofthe concrete, this reduction of the expansiondisplacement causes a slight reduction of the

spreading force from F,, to F, [Fig. 3.30(b)]. If,on the other hand, it is assumed that the concreteis subjected to purely plastic deformations duringexpansion, then theoretically the expansion sleevewill free itself around its circumference from thehole wall and the spreading force will decline tozero [Fig. 3.30(c)]. In reality the concrete isdeformed elastically and plastically. Therefore,the actual situation lies between these twoextremes. However, due to the steep gradient ofthe unloading curve, it has to be expected thateven a relatively slight increase in crack width willlead to a substantial reduction of the spreadingforce [Fig. 3.30(d)]. For anchors situated in cracksrunning in one direction, the spreading force willalso be reduced by the opening of the crack, butthe reduction will be less pronounced than in thecase shown in Fig. 3.30.

a)

b) Concrete elastic

anchor unspreadcrack openingspread anchor

Spreading ForceFO II-7 .6 -

1 Spread. Displ.c

c) Concrete plastic

Fig. 3.30-Influence of cracks on spreading force (from Eligehausen and Pusill- Wachtsmuth 1982)


Properly designed torque-controlled anchorswill expand to an upper bound when they areloaded. This causes an increase of the spreadingforce until the holding capacity is reached. If thecrack width is smaller than about 0.4 mm, theholding capacity of heavy-duty, torque-controlledsleeve anchors is often large enough to causefailure by pulling out a concrete cone. Therefore,the reduction of the failure load is ahnost thesame as for headed anchors (compare Fig. 3.31

Ku /c rock ) / Fu ( u n c r a c k e d c o n c r e t e )

0,8 1,2 1,6crack width ^ _ w [mm]

Fig. 3.31 -Influence of cracks on the ultimate loadof torque controlled expansion anchors (fromEligehausen 1984)

with Fig. 3.28). For larger cracks the expansioncones are often pulled through the expansionsleeves, because the maximum spreadingdisplacement reaches the upper bound and theholding capacity is less than the concrete conefailure load. This results in an additional decreaseof the failure load in comparison to headed orundercut anchors.

If torque-controlled expansion anchors do notproperly expand further or when the spreadingdisplacement is too small, the influence of crackson the failure load will be much more pronouncedthan shown in Fig. 3.31.

Drop-in anchors cannot expand further afterthey have been properly installed. Due to thereduction of the spreading force caused by cracks(Fig. 3.30), these anchors often fail by pulling outwithout significant damage of the concrete whilein uncracked concrete they produce a concretecone type failure. Therefore, the reduction of thefailure load caused by cracks is much larger thanfor well-designed torque-controlled expansionanchors (compare Fig. 3.32 with Fig. 3.31).

1,0

0,8

0,6

0,4

0,2

0

bcrack) / F ubncracked concrete)

0,4 0,8 1,2crack width ^ _w [mm]

Fig. 3.32-Influence of cracks on the ultimate loadof drop-in anchors (from Eligehausen, Fuchs, andMayer 1987 and 1988)

For self-drilling anchors the ratio of failure loadin cracked concrete to failure load in uncrackedconcrete seems to be independent of the anchordiameter for constant crack width to maximumexpansion displacement ratio (Fig. 3.33). Because
the maximum expansion displacement increaseswith increasing anchor diameter, the reduction ofthe failure load for constant crack width is largerfor smaller anchors than for bigger anchors.

FU (Anchor in Crack)rrU (Anchor in uncracked Concrete).

7,M12

1 Single Cracks ,

I I I I L-7 1*0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

w/a

Fig 3.33 -Relative strength of self-drilling anchors as a function of the ratio of crack width to expansiondisplacement (from Eligehausen 1987)

F, (crack) / F, (uncracked concrete)

0,4 0,6crack width w [mm]

Fig. 3.34 -Ratio of the failure load of chemicalanchors installed in cracks to the failure load inuncracked concrete as a function of crack width(from Eligehausen, Mallee, and Rehm 1984)

In the case of grouted anchors (grouted bycement-based or chemical-based mortar) cracksmay disturb the bond between the grout-concreteinterface. Therefore, the failure load of groutedanchors in cracks is significantly smaller than thevalue measured in uncracked concrete (Fig. 3.34).The large scatter of the results is caused by therandom distribution of the crack around theanchor hole and along the anchor length. If thecrack widths are changing due to fluctuating loads,the anchor failure load is even more reduced orthe anchor may even be pulled out (Cannon1981).

Under constant conditions anchors placed inthe intersection of two cracks fail at approximately20 percent lower loads than anchors set in cracksrunning in one direction only (Eligehausen, Fuchs,and Mayer 1987 and 1988). This can be explainedby the fact that the effects described above willoccur in both directions and not in one directionas in the case of single cracks.

Anchors are often installed in groups where theindividual anchors are connected by anattachment. In this case some anchors might sit inuncracked concrete while others are located in


cracks. The average strength of groups situated incracked concrete was about 30 percent lower thanthe value applicable for anchor groups set inuncracked concrete (Eligehausen, Fuchs, andMayer 1987 and 1988). Approximately the samestrength reduction was measured for singleanchors installed in cracks. Failure of allfastenings was caused by pulling out a concretecone.

The strength of the entire anchor group isconstant for one or more of the anchors in aconcrete crack. The reduction is almost the samewhether one anchor or all are in concrete cracks(Fig. 3.35). In the test, the anchor plate wasconnected flexibly (by hinges) to the hydrauliccylinder.

125

100

75

50

Fu [kN]150 -

Number of anchors in cracks

Fig. 3.35-Strength of fastenings with four anchorsas a function of the number of anchors in cracks(from Eligehausen, Fuchs, and Mayer 1987 and1988)

Theoretical studies showed that the resultsdescribed are also valid for larger groups ofanchors and for applications when the anchorplate is rigidly attached (Eligehausen, Fuchs, andMayer 1987 and 1988).

Based on these results, it can be stated that thestrength of anchor groups placed in crackedconcrete can be taken as n-times (n = number ofindividual anchors of the group) the valueexpected for one anchor if the influence of cracks

and anchor spacing is taken into accountsimultaneously. This is valid for anchors with asteadily increasing load-displacement relationshipin both uncracked and cracked concrete.

Fig. 3.36 describes the influence of the load-
displacement relationship of expansion anchorsplaced in cracks on the failure load of anchorgroups. It is assumed that three anchors of aquadruple fastening (large spacing) are located incracks and one anchor is sitting between cracks inuncracked concrete. If the anchors show asteadily increasing load displacement relationshipin uncracked and cracked concrete (Lines a1 anda2 of Fig. 3.36), the failure load of the group isabout four times the failure load of one anchorplaced in a crack. (This theoretical result is inaccordance with Fig. 3.35.) Expansion anchorslocated in cracks may slip in the hole beforeexpanding further and take up more load (Line bof Fig. 3.36) or may be pulled out at rather lowloads (Line c of Fig. 3.36). If only one of theanchors shows a load-displacement behavioraccording to Lines b or c, the failure load of thegroup may be reduced by more than 40 percent.
Anchors which are being used in areas wherecracks may occur, such as the tension zone of aconcrete member, must be suitable for thisapplication.

3.3.4.1 Influence of tensile stresses generated bystructural action on anchor strength -In testssummarized to this point, the anchors were placedin the tension zone with constant stress of thereinforcement, and therefore, tensile stresses inthe concrete were mainly induced by the anchors.However, if the anchors are placed in the shearregion of beams and slabs and in the region ofanchorages and lap splices of deformed bars,locally high tensile stresses are already induced inthe concrete due to the loading of the structure.If anchors are placed in this region, the tensilestresses that they induce in the concrete combinewith the tensile stresses due to loading of thestructure. An example is shown in Fig. 3.37. It is
assumed that an anchor is placed in the endregion of lapped splices of large reinforcing bars.
Plotted are stresses in the concrete due tosplicing of the bar and loading of the anchor. Thetensile stresses along the failure surface of theconcrete cone overlap. Therefore, a reduction ofthe pullout load compared to anchors placed inotherwise unloaded concrete must be expectedwhich, according to tests, is up to 25 percent in


F

FUC

anchor in crack

load Fu

displacementrelationship

4 Fuc

0 a, 0,94

0 b 0,64.oC 0,50

vuc V

Fig. 3.36- Influence nf load-displacement relationships of expansion anchors on the ultimate load of an anchor - -

group (from Mayer and Eligehausen 1984)
\
a expansion anchorstresses causedby re in fo rcement \.~~

ds = 28mm-,

stresses caused 1by anchor

+ +c 1

Fig. 3.37-Anchor in the region of an overlap splice (cross section). Overlapping of stresses caused by the barsand by the anchor (from Eligehausen 1984)


the assumed case (Eligehausen 1984). Reducingthe size of the reinforcing bars, increasing theembedment depth of the anchor, or both reducesthe influence of these intersecting stresses. Insummary, the influence of these intersectingstresses on the failure load is smaller than theinfluence of cracks.

In the tests summarized in Fig. 3.28 and 3.31 to3.35, anchors were used which extended beyondthe tension reinforcement. If short anchors areused, they are anchored in the concrete cover orbetween the bars. In this circumstance, hightensile stresses are induced in the concrete coverby the bond action of the reinforcing bars. Thesestresses intersect the tensile stresses in theconcrete induced by the anchor. The strength ofthe concrete in the cover and in the region of thebars may be lower than in the core of thespecimen due to poor compaction, especially insections with closely spaced reinforcement.Furthermore, this reinforcement reduces theconcrete area available for transmitting tensileforces. Because of these conditions a significantreduction of the failure load of all types ofanchors must be expected. This was confirmed bytests with expansion and undercut anchors placedin the cover of a beam with rather heavyreinforcement (Eligehausen, Fuchs, and Mayer1987 and 1988) (Fig. 3.38). After loading the

0b t

0a

135 = 2.15 Id 180=3ld

c

dimensions in mm

Fig. 3.38 - Test specimens (from Eligehausen, Fuchs,and Mayer 1987 and 1988)

beams to service load (crack width w = 0.3 to 0.4mm) the anchors were loaded to failure. Theanchor failed when the concrete cover betweentwo adjacent cracks was pulled off (Fig. 3.39). On
an average the ratio of failure load in crackedconcrete to the value for uncracked concrete wasabout 30 percent smaller than shown in Fig. 3.28and 3.31.
3.3.4.2 Influence of load transfer into the tensionzone on the behavior of the structural element-Theoverlapping of concrete tensile stresses caused byloading the structure, and stresses induced locallyby the loaded anchor affects the strength of theanchor and may reduce the strength of themember where the anchor is placed (Rehm andEligehausen 1986). Transfer of high tensile forcesinto the concrete in the region of overlap splicesand of anchorages of reinforcing bars may becritical especially if the splice reinforcement is notenclosed by stirrups (Rehm and Eligehausen1986). Another critical application is the transferof forces into the tension zone in the shear regionof slabs without shear reinforcement.

Investigations of this case are described byEligehausen and Reuter (1986) and Lieberum,Reinhardt, and Walvaren (1987).

Slabs 300 mm thick were tested by Eligehausenand Reuter without shear reinforcement. Theshear-span ratio a/h ranged from 3 to 4.5. Afraction of the total load was transmitted byanchors into the tension zone and the rest byloading plates into the compression zone. Typesof anchors examined were expansion, undercut,and headed studs. The embedment depth (40 to130 mm) and the ratio of anchor load to total load(0 to 100 percent) were varied. In all cases theslabs failed by an inclined shear crack.

Fig. 3.40 shows the cracking pattern of one
specimen at about 95 percent of the failure loadof the member. The anchor loads must betransmitted over the tip of the inclined crack tothe supports. This causes high tensile stresses atthe crack tip. Therefore, the failure crack (shownas a broken line) will occur at a lower total shearforce than loading the slab in the compressionzone only. In the tests a reduction of the shearcarrying capacity of the slabs up to between 15and 20 percent was found when all the loads weretransmitted into the tension zone and not into thecompression zone. The strength reduction wassmaller when only a fraction of the total load wastransferred into the tension zone. A similarstrength reduction was found by Lieberum,Reinhardt, and Walvaren (1987), under


_I

.b.. b

Fig. 3.39 - Concrete failure of an anchor group (from Eligehausen, Fuchs, and Mayer 1987 and 1988)

a =3d1 L7 I

Fig. 3.40- Crack pattern. of a slab without shear reinforcement (from Eligehausen and Reuter 1986)


these test conditions. If the anchors are placedclose to the support the strength reduction will bemuch more significant.

This reduction of the shear capacity may,depending on the design of the slab, significantlychange the type of failure from a ductile bendingfailure to a brittle shear failure (Eligehausen andReuter 1986). To avoid this problem, it isrecommended that the shear forces transmitteddirectly into the tension zone should be limited toabout 40 percent of the total shear force, oralternatively, the shear stress should be limited toabout 80 percent of allowable values.

Composite structures (precast concreteelements with bonded cast-in-place concrete)without reinforcement connecting the precast andcast-in-place concrete, are especially critical.Failure of this type of structure will often becaused by a crack in the contact area between theprecast and the cast-in-place concrete. If the loadis transmitted into the precast concrete element,high tensile stresses are generated in the contactarea. Therefore, the shear stress at failure issignificantly lower than in the case of loading thespecimen in the usual way at the top (Fig. 3.41).

join f

-72 4 6

a/d

Fig. 3.41 -Shear stress failure of a composite slabwithout connecting reinforcement between precastand cast in place concrete (after Rehm andEligehausen 1986)

3.3.5 Shear loading-Little investigation of theinfluence of cracks on the behavior of anchorsloaded in shear has been conducted. The fewavailable test results can be summarized asfollows.

Anchors placed in cracked concrete and loadedin shear will fail the concrete (small edgedistances), or the bolt (large edge distances), or acombination of both. Under otherwise constantconditions, the failure load of anchors with a smalledge distance and loaded towards the edge will besmaller in cracked concrete than in uncrackedconcrete due to the disturbance of the distributionof stresses in the concrete by cracks. It can beassumed that the strength reduction is almost thesame as for tension loading (reduction by about 40percent). The strength reduction will be smallerif edge reinforcement is present. The ultimateload of anchors with large edge distances (steelfailure) is not significantly influenced by cracks.The edge distance required to insure a steelfailure of the anchor is about 30 to 40 percentlarger in cracked concrete than in uncrackedconcrete.

3.4-Behavior of cast-in-place anchor bolts inuncracked concrete piers

3.4.1 Introduction -Anchor bolts are commonlyused in highway and bridge structures to connectlight standards, sign supports, and traffic signalpoles. They are also used to connect steelcolumns in industrial structures to structuralconcrete members. The anchor bolt installationdiscussed in this section is one of the most widelyused cast-in-place anchorage systems. The anchorbolts used typically have long embedment lengthsand small edge distances. Such installation shouldbe distinguished from bolts embedded for shortlengths in mass concrete with very large edgedistances. The supporting concrete membersassociated with this installation are usually piers,drilled shafts, or other foundation elements withlimited plan dimensions; however, the concrete isusually well confined by reinforcement.

The structural behavior of cast-in-place anchorbolts with long embedment lengths installed insupporting members with limited dimensions isdistinctly different from that described in thepreceding sections. This section summarizes somesignificant results from extensive research con-ducted for this type of anchor bolt application atthe University of Texas at Austin (see Breen 1964;


Lee and Breen 1966; Lee and Breen 1970; Hasse lwander , J i r sa , and Breen 1974; Hasselwander, Jirsa, Breen, and Lo 1977; and Jirsa, Cichy, Calzadilla, Smart, Pavluvcik, and Breen 1984). The test results and design recommendations are valid for anchors in well- confined concrete.

These studies focused on many significant factors affecting anchor bolt behavior including clear cover, embedment length, bolt diameter, bearing area, type of anchorage device, concrete strength, steel yield strength, shape of piers, and bolt group configuration. In addition, a series of exploratory and supplementary studies were made to determine the influence of cyclic loading, lateral loading, transverse reinforcement, and method of loading on the bolt behavior. Diameters of anchor bolts ranged from 1 to 3 in. Steel yield strengths ranged from 33 ksi (A7) to 105 ksi (A139). Embedment lengths ranged from 10 bolt diameters to 20 bolt diameters. A typical test specimen geometry is shown in Fig. 3.42.

8'- o”_

I'' ANCHOR 8OLl-

I J/4” ANCHOR B O L T -

SECTION A-A SECTION B-B

Fig. 3.42 - Typical specimen geometry

3.4.2 General behavior under loading-A singleanchor bolt transfers tension load to the concretemember in three successive stages: (1) steel-to-concrete bond, (2) bearing against the washer ofthe anchorage device, and (3) a wedging action bythe cone of crushed and compacted concrete infront of the anchorage device. These three stagesare not entirely distinct, but the exact nature ofthe transition from one stage to the next is highlyindeterminate and can only be discussed in ageneral manner.

Fig. 3.43 shows tail stress plotted against lead
stress for three 1 3/4 in. anchor bolts with clearcovers of 3 l/2 in. and three differentembedments: 10, 15, and 20 bolt diameters.Adhesion or bond between the bolt and concreteis the predominant load carrying mechanism forearly stages of loading; little increase in tail stressis observed with increasing lead stress. The longerthe bolt, the more load the bolt can carry by thebond mechanism. Under increasing load, bondstrength decreases along the length of the bolt and


8 l I I I s I

10 20 30 40 50 60 70

Tai l St ress, ks i

Fig. 3.43 - Tail stress versus lead stress for different embedment lengths

tail stress begins to increase. The load that waspreviously carried by a bond mechanism must betransferred to a bearing mechanism. In Fig. 3.43the bond-to-bearing transition is most clearly seenfor the bolt with 200 embedment. For a givenload increment, the tail stress increases more thanthe lead stress as the load carried by bond isunloaded into bearing on the anchorage device.The bond-to-bearing transition is dependent onthe embedment of the bolt; the shorter the bolt,the shorter and less well-defined the transition.After the bond-to-bearing transition, tail stressincreases uniformly with increasing lead stress asthe load is carried by bearing or by wedgingaction.

3.4.3 Failure modes-The failures observedduring testing can be described as: (1) bolt failure,(2) concrete cover failure by spalling, and (3)concrete cover failure by wedge-splitting. Whilethese three categories represent distinct failuremodes, combinations of these modes wereobserved in several instances.

Bolt failures occurred in several bolts bynecking in the threaded portion of the bolts.Little damage to the concrete cover over the boltwas observed at bolt failure. A relatively suddenspalling of the concrete cover over the anchoragedevice at low loads characterized the failure ofbolts with small amounts of clear cover [Fig.3.44(a)]. For larger amount of clear cover, the
failures were characterized by the splitting andspalling of the concrete cover into distinct blocksby the wedging action of a cone of crushed andcompacted concrete which formed in front of theanchorage device [Fig. 3.44(b)].
The distinguishing feature of a wedge-splittingfailure was the diagonal cracks [marked B in Fig.3.44(b)] which started just in front of the washeron the bolt centerline and extended toward thefront and each side of the specimen. Thesediagonal cracks were frequently accompanied by alongitudinal crack along the bolt axis [C in Fig.3.44(b)], a transverse crack parallel to and nearthe washer of the anchorage device [A in Fig.

Tens ion Tension

Cover SpallingF a i l u r e

Wedge-SplittingFailure

Fig. 3.44 - Concrete cover failures

3.44(b)] or both. Cracking generally started nearthe anchorage device and extended toward thefront, toward the sides of the specimen, or bothunder increasing load.

3.4.4 Lead-slip relationships (effect of clear coverand embedment length)-Bolt tension versus leadslip curves associated with different clear coversand embedments are shown in Fig. 3.45 and 3.46.
Slip of the anchor bolts was measured relative tothe front face of the specimen (lead slip). Fig.3.45 illustrates the effect of clear cover. Since theeffect f concrete strength varied approximatelywith Pd lead stress in Fig. 3.45, calculated on thebasis of the anchor bolt stress area, wasnormalized with respect to /--d and plotted againstlead slip for four 1 3/4 in. bolts each with anembedment of 15 bolt diameters (15D) and ananchorage device consisting of a nut and a 4 in.diameter, l/2 in. thick washer. As seen in Fig.3.45, the slopes of the curves are essentially thesame until each bolt approaches ultimate capacity.A definite trend of increasing ultimate strengthwith increasing clear cover is indicated.
Fig. 3.46 illustrates the effect of embedmentlength on the stress-slip relationships of three 13/4 in. bolts each with a clear cover of 3 l/2 in.and an anchorage device consisting of a nut and a4 in. diameter, l/2 in. thick washer. The initialportions of the curves are essentially the same andthere is no appreciable difference between theultimate strengths of the 15D bolt and the 20Dbolt; the ultimate strength of the 1OD bolt,however, is noticeably reduced.

The failure of the 10D bolt developed initiallyas a typical wedge-splitting mode until the

cracking propagated to the sides and front face ofthe specimen. The result was the complete loss ofa rectangular block of concrete cover extendingback to the anchorage device over the full widthof the specimen, as opposed to the usual group oftriangular wedges with a common apex over theanchorage device. Such a failure indicates thatthe wedge-splitting mechanism did not fullydevelop and therefore the ultimate strength of theanchor bolt installation was reduced.

The major effect of embedment length on theultimate strength of an anchor bolt installation isrelated to the ability of the concrete cover to resistthe wedge-splitting action of the cone of crushedand compacted concrete in front of the anchoragedevice. A certain minimum embedment length isrequired to develop this resistance. As illustratedin Fig. 3.46, increasing the embedment lengthbeyond this minimum length provides nosignificant improvement but decreasing theembedment length results in a significantreduction in ultimate strength. A 15D embedmentlength can be considered a satisfactory minimumembedment length.

3.4.5 Ultimate strength-The ultimate strength ofa bolt in a group is clearly not the same as that ofan isolated bolt with similar geometry.

3.4.5.1 Single bolt strength -Hasselwander,Jirsa, Breen, and Lo (1977), concluded that clearcover and bearing area are the main variablesgoverning the strength of single anchor bolts. Thevariables were incorporated into an equation forpredicting the strength of isolated anchor bolts,subjected to simple tension and failing in a wedge-splitting mode:

Tn = 140A, @[O-7 + ln[2C’/(D, -II)]] (3.35)

whereT, =

Ab =

D =D, =

C' =

ultimate wedge-splitting capacity ofa single bolt, lb, with anembedment length not less than 12(Dw - D)net bearing area, in.*, (r/4) $$-D”), but not greater than 4Dbolt diameter, in.diameter of anchorage device(washer), in. with minimumthickness of Dd8clear cover to the bolt, in.


d:cu

.

sd

.

s ’d

fA!!LJr-::

I500

1250_,--L~l5 0

.1000

L = 20 D

A-.._.._& --z-LL=l0D

750

C’ = 3.5 In.

1 I . . I I . . .0.02 0 . 0 4 0 . 0 6

.0 . 0 8 0.10 0.12 0.14 0.16 0.18

. .0 . 2 0 0 . 2 2 0.24

Lead Slip, inches

Fig. 3.46-Effect of embedment length


The designas:

tensile strength T, was determined

T,, si 4 Tn but < A, fy , lb (3.36)

where4 = a capacity reduction factor of 0.75

As = tensile area of the anchor bolt, asdefined in Eq. (3.1), in.2

fy = yield stress of the bolt material, psi

The design equation was developed from aregression analysis on test results of bolts failing inthe wedge-splitting mode only. A minimumembedment length of 12(D, - D) was suggested toallow the wedge-splitting mechanism to occur. Arestriction which accounted for a reduced bearingefficiency observed for large washers, limited thenet bearing area to 4D2. A minimum washerthickness, D$?, was suggested to prevent flexibilityof the washer.

Fig. 3.47 shows graphically the suggested
ultimate strength equation and the test dataplotted to illustrate the accuracy of the equation.The equation provides a reasonable estimate ofstrength, yet is simple to use and reflects thecritical parameters observed in the test program.
3.4.5.2 Bolt group strength - Jirsa, et al. (1984),evaluated the bolt group interaction and strengthreduction by comparing the average test capacitywith the predicted capacity of an isolated bolt withsimilar geometry. It was observed that as boltspacing decreased, the reduction in strengthsignificantly increased. From a least squaresanalysis of the available data, the followingmodification to Eq. (3.35) was produced for thenominal tensile capacity of an anchor bolt in abolt group based on failure of the concrete.

T,, = 140Ab@ {0.7 + ln[2C’/(D,,,-D)]}(0.02S + 0.4), in. (3.37)

whereS = bolt spacing, in.

(0.02S+0.4) 5 1.0

and other factors are the same as in Eq. (3.35).Eq. (3.37) provides an estimate of the strength

of closely spaced anchor bolts with edge covertypical of highway- related structures. The designtensile capacity, Tu, can be determined accordingto Eq. (3.36).

3.5-REFERENCES

ACI Committee 349,1990, “Code Requirements for NuclearSafety Related Concrete Structures,” (ACI 349-90) AppendixB, American Concrete Institute, Detroit.

American Institute of Steel Construction, 1978,Specifications for the Design, Fabrication and Erection of Struc-tural Steel Buildings, with Commentary, New York, 235 pp.

Best, J. FIoyd and McDonald, James E., 1989,: “Evaluationof Polyester Resin, Epoxy, and Cement Grouts for EmbeddingReinforcing Steel Bars in Hardened Concrete,” TechnicalReport REMR-CS-23, US Army Engineer Waterways Experi-ment Station, Vicksburg, MS.

Bode, H. and Roik, K., 1987,: “Headed Studs Embedded inConcrete and Loaded in Tension,” in ACI SP 103 Anchorage toConcrete, G. Hasselwander ed. , Detroit.

Bode, H. and Hanenkamp, W., 1985, “Zur Tragfshigkeit vonKopfbolzen bei Zugbeanspruchung,” (For Load BearingCapacity of Headed Bolts Under Pullout Loads), Bauingenieupp. 361-367.

Braestrup, M.W., Nielson, M.P., Jense, B.C. and Bach, F.,1976, “Axissymetric Punching of Plain and ReinforcedConcrete, Copenhagen, Technical University of Denmark,Structural Research Laboratory, Report R 75.

Breen, J.E., 1964, “Development Length for Anchor Bolts,Research Report 55-1F, Center for Highway Research, theUniversity of Texas at Austin.

Burdette, E.G., Perry, T.C. and Funk, R.R., 1987, “LoadRelaxation Tests”, ACI SP-103 Anchorage to Concrete, G.Hasselwander ed.,Detroit, pp. 297-311.

Cannon, R.W., 1981,: “Expansion Anchor Performance inCracked Concrete,” ACI-Journal, November-December, pp.471-479.

Elfgren, L., Anneling, R., Eriksson, A., and Granlund, S.,1988, “Adhesive Anchors, Tests with Cyclic and Long-TimeLoads,” Swedish National Testing Institute Report 1987:39,Bor&.

Eligehausen, R., 1987, “Anchorage to Concrete by MetallicExpansion Anchors, ACI SP 103 Anchorage to Concrete, G.Hasselwander ed., Detroit, pp.181-201.

Eligehausen, R., 1984,: “Wechselbeziehungen zwischenBefestigungstechnik und Stahlbetonbauweise”, (Interactions ofFastenings and Reinforced Concrete Constructions), in“Fortschritte im Konstruktiven Ingenieurbau”, Verlag WilhelmErnst & Sohn. Berlin.


4Q

44

l0

0

(I w a I

00000. . . . *. l . .*a ma40

4

0


Eligehausen, R. and Fuchs, W., 1988, “Tragverhalten vonDfibelbefestigungen bei Querzug-, Schrsgzug- und Biegebean-spruchung,” (bad-bearing Behaviour of Anchor FasteningsUnder Shear, Combined Tension and Shear or FlexuralLoading), Betonwerk + Fertigteil-Technik, No. 2, in Germanand English.

EIigehausen, R., Fuchs, W., Lotze, D. and Reuter, M., 1989,“Befestigungen in der Betonzugzone,” (Fastening in theConcrete Tensile Zone), Beton-und Stahlbetonbau 84, No. 2and 3.

Eligehausen, R., Fuchs, W. and Mayer, B., 1987, 1988,“Tragverhalten von Di ibe lbefes t igungen bei Zugbean-spruchung,” (Loadbearing Behavior of Anchor Fastenings inTension), Betonwerk + Fertigteil-Technik, No. 12/1987 undNo. l/1988, in German and English.

Eligehausen, R., Mallee, R. and Rehm, G., 1984, “Befest-igungen mit Verbundankern,” (Fastenings Formed withChemical Anchors), Betonwerk + Fertigteil-Technik, No. 10,pp. 686-692, No. 11, pp. 781-785, No. 12, pp. 825-829.

Eligehausen, R. and Pusill-Wachtsmuth, P., 1982,“Stand derBefestigungstechnik im Stahlbetonbau,” (Fastening Technologyin Reinforced Concrete Construction), IVBH Survey S-19/82,IVBH- Periodica l/1982, February.

Eligehausen, R. and Reuter, M., 1986, “Tragverhalten vonPlatten ohne Schubbewehrung bei Einleitung von Lasten in dieBetonzugzone”, (Load Characteristics of Plates without ShearReinforcement by Introduction of Loads in the Tensile Zoneof Concrete), Report No. l/17-86/3 of the Institut fiirWerkstoffe im Bauwesen, Universitgt Stuttgart.

Eligehausen, R. and Sawade, G., 1985, “Verhalten vonBeton auf Zug,” (Behavior of Concrete in Tension), Betonwerk+ Fertigteil-Technik, No. 5 and 6, May/June.

Fischer, A., 1984, “Befestigen mit Hinterschnittankern,”(Fastenings with Undercut Anchors), in "Fortschritte im Kon-struktiven Ingenieurbau”, Verlag Wilhelm Ernst & Sohn,Berlin.

Hanks, Abbot A., 1973,: Kwik Bolt Testing Program, AbbotHanks Testing Laboratories of San Francisco, File H2189-S1,Report No. 8783.

Hasselwander, G.B., Jirsa, J.O., Breen, J.E., and L.o, K.,1977, “Strength and Behavior of Anchor Bolts Embedded NearEclges of Concrete Piers, Research Report 29-2F, Center forHighway Research, The University of Texas at Austin, May.

Hasselwander, G.B., Jirsa, J.O., and Breen, J.E., 1974, “AGuide to The Selection of High-Strength Anchor BoltMaterials”, Research Report 29-1, Center for Highway Research,The University of Texas at Austin, October.

Jirsa, J.O., Cichy, N.T., CaIzadilla, M.R., Smart, W.H.,Pavluvcik, M.P., & Breen, J.E., 1984, “Strength and Behaviorof Bolt Installations Anchored in Concrete Piers,” ResearchReport 305-IF, Center for Highway Research, The Universityof Texas at Austin, November.

Klingner, R.E. and Mendonca, J.A., 1982a, “TensileCapacity of Short Anchor Bolts and Welded Studs: ALiterature Review,” ACI-Journal, July/August, pp. 270-279.

Klingner, R.E. and Mendonca, J.A., 1982b, “Shear Capacityof Short Anchor Bolts and Welded Studs,” A literature review,ACI Journal, Sept/Oct.

Klingner, R.E., Mendonca, J.A. and Malik, J.B., 1982,“Effect of Reinforcing Details on the Shear Resistance ofAnchor Bolts Under Reversed Cyclic Loading,," ACI Journal,Jan/Feb.

Lee, D.W. and Breen, J.E., 1966, “Factors Affecting AnchorDevelopment, “Research Report 881F,” Center for HighwayResearch, The University of Texas at Austin, August.

Lee, D.W., and Breen J.E., 1970, “Model Study of AnchorBolt Development Factors, Models for Concrete Structures, SP-29, American Concrete Institute.

Lieberum, K.H., Reinhardt, H.W. and Walraven, J.C., 1987,’ Lasteinleitung fiber Diibel in der Schubzone von Beton-Plattenstreifen,” (Fastening of Anchors in the Shear Zone ofConcrete Slabs), Betonwerk + Fertigteil-Technik, No. 10, inGerman and English.

Mayer, B., and Eligehausen, R., 1984, “Ankergruppen mitDubeln in der Betonzugzone,” (Anchor Groups with Anchorsin the Concrete Tension Zone), Werkstoffe und KonstruktionInstitut ffir Werkstoffe im Bauwesen der Universitit Stuttgartand Forschungs-und Materialpriifungsanstalt, Baden-Wiirttemberg (Eigenverlag) October, pp. 167-180.

Meinheit, D. and Heidbrink, F.D., 1985, “Behavior ofDrilled-In Expansion Anchors,” Concrete International, April,pp. 62-66.

PCI Design Handbook-Precast and Prestressed Concrete, 1978,Prestressed Concrete Institute, Chicago, 380 pp.

Pusill-Wachtsmuth, P., 1982, “Tragverhalten vonMetallspreizdiibeln unter zentrischer Zugbelastung bei denVersagensarten Betonausbruch und Spalten des Betons,”(Bearing Behavior of Metallic Expansion Anchors, Loaded inTension, for the Failure Modes of Concrete Breakage andSplitting), Doctoral Thesis, University of Stuttgart.

Rehm, G. and Eligehausen, R., 1986, “Auswirkungen dermodernen Befestigungstechnik auf die konstruktive Gestaltungim Stahlbetonbau,” (Effects of Modern Fixing Technology onStructural Design in Reinforcing Concrete Construction),Betonwerk + Fertigteil-Technik, No. 6, in German andEnglish.


Rehm, G., Eligehausen, R. and Mallee, R., 1988, “Befest-igungstechnik,” (Fastening Technique), in “Betonkalender1988”, Verlag Wilhelm Ernst & Sohn, Berlin.

Rehm, G. and Lehmann, R., 1982, “Untersuchungen mitMetallspreizdtibeln in der gerissenen Zugzone von Stahlbeton-bauteilen,” (Investigations with Metallic Expansion Anchors inthe Cracked Tension Zone of Reinforced Concrete Members)“,Research Report of the Otto-Graf- Institut, Stuttgart, July,unpublished.

Riemann, H., 1985, “Das erweiterte x-Verfahren fiirBefestigungsmittel: Bemessung an Beispielen von Kopfbolzen-verankerungen,” (The Extended X-Method for the Design ofFastening Devices as Exemplified by Headed Stud Anchor-ages), Betonwerk + Fertigteil-Technik, No. 12, pp. 806-815, inGerman and English.

Seghezzi, H.D. and Vollmer, H., 1982, “Modern AnchoringSystems for Concrete, ACI SP-103, Anchorage to Concrete,Atlanta, January.

Sell, R., 1973, “Festigkeit und Verformung von mitReaktionsharzmiirtel-Patronen versetzten Ankern,” (Strengthand Displacement of Anchors Installed with Reaction ResinMortar Cartridges), Verbindungstechnik 5, Vol. E, August, inGerman.

Shaikh, A.F. and Yi, W., 1985, “In-Place Strength of WeldedHeaded Studs,” Journal of the Prestressed Concrete Institute,March/April, pp. 56-81.

Teledyne Engineering Services, 1979, Technical Report3501-1, Revision 1, August 30.

Wagner-Grey, U., 1976,: “Experimentelle und TheoretischeUntersuchungen zum Tragverhalten von Spreizdtibeln inBeton”, (Experimental and Theoretical Investigations on thePerformance of Expansion Anchors in Concrete), DoctoralThesis, Technical University of Munich.

Wiewel, Harry, 1989,: ” Results of Long-Term Tension Testson ITW Ramset/Red Head EPCON Sys tem@ AnchorsInstalled in Hardrock Concrete,” Techmar Inc, long Beach,CA. J une.

CHAPTER 4-DESIGN CONSIDERATIONS4.1- Introduction

The purpose of this section is to discuss thevarious factors which affect the ability of concreteanchorages to perform their intended purpose.These factors should be considered in the designof anchorages. The tendency to design anchorsbased only on their tensile or shear loading isdiscouraged, when actually bending, prying action,and redistribution of loads are often involved.

4.2 -Functional requirements4.2.1 Loading Conditions-Major considerations

in determining the requirements for concreteanchorages include the type of loading which theanchorage will experience, and the potential forconcrete cracking in the vicinity of the anchors.There is a high probability of coincidental crackingwhen anchors are located in the tensile zone of aconcrete member. As described in Chapter 3, thecapacity of anchors under sustained loading in thetensile stress zone of uncracked concrete is only 60to 75 percent of static load capacity of anchors inunstressed concrete. In cracked concrete, anchorcapacity is significantly influenced by anchor typeand width of the crack in the region of theanchorage. In regions of tensile stress, since thewidth of flexural cracks is maximum at theconcrete surface and decrease with distance awayfrom the surface, the designer should use deep-seated anchors (anchored in the compression zoneof the member), or anchors which are designed toperform in cracked concrete. Anchors whichperform well, at a given load level in uncrackedconcrete, may fail completely in cracked concreteunder loads of the same magnitude. Criteria forthe design and selection of concrete anchoragesshould account for these factors.

Economics or related issues may dictatedesigning for a selected mode of failure.Installations such as bridge railings and highwaysigns could potentially receive accidental loadingsthat are not reasonable design loads. In suchcases it may be prudent to design for the failure ofthe most easily replaced segment of the structure,whether it is the anchor bolt or a separate piece ofthe structure. Care must be exercised in designingfor selected failure modes to maintain the integrityof the primary structural system.

4.2.1.1 Column bases - Simply connectedcolumn bases are normally loaded in compressionof sufficient magnitude that column shear istransferred through friction and the anchorageserves only for erection purposes. It has beencommon practice for many years to use L- and J-bolts for erection anchors, which do not havesufficient embedment to develop the strength ofthe anchor steel. Headed anchors of the samesize and length as L- and J-bolts have significantlyhigher capacities. However, the increase incapacity is often not needed for the simple columnbase plate connection. Column bases which aredesigned as moment connections should require a


rigid base connection and anchors should beselected which can maintain a sufficient residualpreload to develop applied moments. Theseconditions are necessary to achieve fixity of thecolumn base.

4.2.1.2 Machine Foundations-Anchor boltsfor machinery foundations are generally specifiedby the machinery manufacturer and have beensized by experience. Their general purpose is tofix the rigid machine housing to concrete in orderto withstand machine vibrations. They aregenerally installed to a relatively low stress leveland may not have sufficient embedment todevelop the anchor steel capacity. Seismic loadingof machine foundation anchorages can be criticaland must be considered.

4.2.1.3 Structural Tension and ShearConnections-The anchorage of principalstructural connections requires carefulconsideration of all possible loading combinations.Failure of structural connections may becatastrophic, particularly when there is noredundancy in the system. It is recommended thatall structural connections be ductile.

Ductility is defined as the ratio of a structure’splastic displacement to its maximum elastic (yield)displacement. The ability of a structure to exhibithigh values of ductility (ten or greater) is anextremely desirable feature because this can allowfor an overload condition to exist withoutproducing a catastrophic failure. It can providefor highly redundant structures (i.e., structuresthat provide alternative stress paths) thatredistribute loads internally.

When designing the anchorage of a steelstructure to concrete, ductility of the structure,including the connection, should be considered.The desired ductile behavior may occur in any oneor all of the following components: the structuralsteel element being connected, the baseplateattached to the steel member, the steel anchors, orthe concrete. Steel is more ductile than concreteand it is better to proportion an anchorage so thatthe majority of the ductile displacement occurs inthe steel elements of the anchorage or in theattached structural member. In cases where thisis not possible, extra care should be taken inselecting anchor types, geometry, and safetyfactors.

Temperature changes and the shrinkage ofstructural elements should also be carefullyconsidered in determining connection details

because of the significant effect which tensileloads have on anchor stress and the manner inwhich shear is transferred to the concrete.Structural connections should also be investigatedfor cyclic loadings, vibration loads from wind ormachinery, and seismic loads.

4.2.1.4 Pipe Supports-In most structures, pipesupports are dead-load hangers or supportbrackets. Pipe supports are generally detailed toprovide free expansion and contraction of thepiping system under changing temperatures.Experience has shown these loosely supportedsystems function very well under seismicconditions without special design considerations.Vibration problems normally occur underoperating conditions and are corrected by addingor shifting supports to alter the responsefrequency of the system. Design loads for thesesupports are generally low and sizing of anchors,by experience, usually results in large safetyfactors.

In contrast to this, the pipe supports for nuclearapplications are often designed to prevent pipingsystem frequencies from coinciding with predictedstructural frequencies generated by an earthquakeof prescribed magnitude. As a result,specifications often limit support displacements tolow values under conservative combinations ofloading. Most anchorages cannot comply with theimposed displacement limitations without rigidbases and oversized anchors.

When a pipe has multiple supports and isloaded along its length, evaluation of the stiffnessof each support with respect to the longitudinalstiffness of the total support system betweenexpansion joints or bends should be made toinsure that a particular support is not overloadedto failure, thus setting up a progressive failuremechanism.

4.2.2 Anchorage Environment- Consideration ofthe service environment is essential for servicelongevity, particularly in areas where theanchorage may come in contact with saltwatersprays or deicing salts. Unprotected steel isparticularly vulnerable to corrosion when exposedto the atmosphere. For expansion anchors,vulnerability to corrosion exists in the region ofthe expansion mechanism where space is availablefor moisture collection. Corrosion will reduce theability of anchors to function correctly, especiallytorque-controlled expansion anchors.

Where steel is under a sustained high stress,


there is a higher potential for stress corrosionfailure. If the yield strength of the anchor steel isless than 120,000 psi, stress corrosion is less likelyto be a problem. However, precautions must betaken when chlorides are used in the anchoragezone either externally or as a part of the concretemix. Protective coating systems, or the use ofcorrosion resistant materials, should be consideredin corrosive environments. The use of thin zinccoatings will not provide permanent protectionagainst corrosion under normal outside exposureconditions. Proper detailing will insure that runoffwater cannot reach anchors in areas of snow andice removal. Alternate periods of wetting anddrying have been known to produce corrosioneven in the absence of chlorides.

Anchor bolts are often set in sleeves to providefor minor adjustment of the bolt to fit thefoundation base. If the foundation is exposed tofreezing temperatures, the sleeves should be filledwith grout or be otherwise protected against theintrusion of water. Gaps between a steel baseplate and the concrete surface should be sealed ifthe foundation is exposed to an aggressiveenvironment. In a similar fashion, plain sand-cement dry-pack pads which are exposed tofreezing and thawing should be coated with asealer to prevent water absorption.

Chemical adhesives, lead caulking, or othermaterials which have a high rate of creep atelevated temperatures should not be used in areasof high temperature or possible exposure to fire.Special investigations may also be necessary todetermine the possible effects of process chemicalson anchors in industrial plants. Intermittentexposure may be a more severe service conditionthan continuous exposure.

4.2.3 Behavior-The behavior of cast-in-placeand post-installed anchors is described in Chapter3. Well-designed, cast-in-place anchors performbetter than or equally as well as post-installedanchors, if for no other reason than that they arenormally set deeper into the concrete and atultimate load feature failure in the bolt ratherthan failure in the concrete. Construction logisticsand specifications that admit alternativemanufacture of the equipment to be anchored(and therefore alternative anchorage size andlocation) often make the post-installed anchormore practical. Nonetheless, the designer shouldconsider use of a cast-in-place anchor wheneverthe size and location of that anchor is known prior

to casting of the concrete.Anchor capacity may be limited by the strength

of concrete, by the strength of the anchorage steel,or by slip of the anchorage mechanism. The modeof failure is an important design consideration.

Concrete failure may occur before or duringslip of the anchor. In general, the properties ofsteel are well defined and steel failure ispredictable and controllable. In contrast to thecontrolled ductility of a steel failure mechanism,concrete is a brittle material with less well-definedproperties. Failure by slip may be either brittle orductile depending on the ability of the anchoragemechanism to maintain load during slip.

4.3 - MATERIALS4.3.1 Concrete-When the capacity of the

anchorage is controlled by the strength ofconcrete, it is generally the tensile properties ofthe concrete which control cone failures, andcrushing strength that controls slip failures.Tensile properties of concrete vary more thancompressive properties. Tensile properties of theconcrete also influence bond and affect thoseanchor types which depend on bond to developcapacity.

The tensile-compressive strength relationshipcan be complicated by the influence of grain size,type, and distribution of aggregate particles. Forthis reason, construction practices, which permitsegregation of the aggregate will increase thevariability of tensile strength more than thecompressive strength. Segregation of theaggregate is influenced by the slump of theconcrete, the height of the drop of the concrete,and the amount of vibration during placement.For this reason, the capacity of anchors may varydepending on their location in walls and in the topor bottom of slabs.

4.3.2 Steel-The type of steel used in anchors islargely dependent on the method of anchorage butcan also be influenced by the method of securingthe base plate or attachment to the anchors. It isdesirable to limit the yield strength of headedanchors to that of ASTM A 325 or lower strengthmaterial, because of the brittle nature of higherstrength steels. Zinc plating causes additionalbrittleness and reduced fatigue resistance forhigher strength steel bolts. Steel with yieldstrengths in excess of 120,000 psi have been foundto be highly susceptible to stress corrosion in mostanchorage environments.


4.4 -Design basisThe safety factor for any element in an

anchorage system should be consistent with theother elements in the system. Establishing anallowable stress or load factor must consideroverall behavior of the anchorage. The design ofconcrete anchorages is usually controlled by codesgoverning both structural steel and concrete.

4.4.1 Types of anchors4.4.1.1Headed Anchors-Headed anchors may

consist of welded studs or bolting material withanchor heads manufactured to establishedstandards. Headed anchors may also be made bywelding a rigid plate to the embedded end of theanchor or by threading a bar and using a standardnut. Once the load increases sufficiently toovercome n the shank, subsequent loading

anchor head. Headedefficiently if the shank of theThis will minimize bond andoad on the anchor by bearing

Anchors-When anchor load

inishes with depth. Thequired to fully develop the

of deformations). Undersustained loadiconcrete in theBonded anchor typically been manufactured

deformed reinforcing bars,s. The basic development

Building Code are based ond minimum spacing of an

rs. The basic developmentars with a hook or 90” bendabout 50 percent of the

of straight bars. The use ofr reinforcement was excluded

g Code in 1971 (ACI-ASCE

considered as twice thatof deformed ba

gths given in ACI 318insure that the crete capacity is higher than the

When evaluating the concretee failure modes “splitting of

e concrete between ribs”The failure mode

“concrete cone break out” was not consideredbecause typically this mode does not occur whendeveloping reinforcing bars. However, the failuremode “concrete cone break out” is quite typical forshallow anchors (see Chapter 3).

Excluding edge and spacing conditions, theyield strength of an individual reinforcing bar canbe developed in 3000 psi uncracked concrete inabout 15 bar diameters (straight bar) or 10 bardiameters (hooked bar). To preclude a concrete-cone-break-out failure, the development lengthmay increase by a factor of up to four to accountfor the effects of cover, number, and spacing ofbars. A further increase of the developmentlength by a factor of one and one-half to two isnecessary if the anchors are located in the crackedtensile zone of a reinforced concrete member.Most anchorage situations do not involveminimum values for spacing and cover. The codeprovisions will be very conservative if individualbars are anchored in uncracked concrete well awayfrom edges. However, the code provisions maynot be conservative, if a group of bars, with orwithout small edge distance, is anchored inuncracked concrete or in the (cracked) tensionzone of reinforced concrete members.

4.4.1.3 Expansion Anchors-Many patentedexpansion devices are used to mechanically fastenpost-installed anchors to the concrete. Mostexpansion anchors were originally developed forshort embedment depths to provide an anchorwhich failed in the concrete or by slip. Sinceductile steel failure had no opportunity to occur inthis situation, there were no restricting strengthsapplied to the steel in these anchors. Morerecently developed expansion anchors featureexpansion mechanisms that can fully develop thestrength of the anchor steel, when used as singleanchors. Ductile steels should be specified for thistype of anchor if a ductile failure mode is desired.

4.4.2 Concrete tensile failure -The determinationof concrete pullout strength (cone failure) ofindividual anchors and anchor groups is discussedin Section 3.2.2. Concrete cone failure will occurwhen the capacity of the anchor bolt exceeds theconcrete pullout strength. All shell type expansionanchors are designed to fail the concrete when thebolt is embedded to shell depth. Concrete failurecan also occur with wedge bolts having shallowembedment depths.

The concrete may also fail by splitting tensionwhen there is inadequate lateral confinement of

355.1R-57

the anchor. This occurs with all types ofexpansion anchors that have small edge distances.Deformation-controlled expansion anchors (drop-in, self-drilling, and stud) are especially sensitiveto edge distance because of the high expansionforces developed during anchor installation.Splitting may also occur at close edge distanceswhen the anchorage mechanism expands with loadapplication.

In the United States, most manufacturers ofexpansion anchors recommend limiting normalservice loads to 25 percent of the manufacturer’saverage test failure load. Investigations by theUnited States Nuclear Regulatory Commission(1979) indicated that installation problemsassociated with split-shell type expansion anchorswarranted increased safety factors over thoseapplied to torque-type anchors. For the split-shellanchor, and others which cause the concrete tofail, it was recommended that a minimum factorof safety of five against average test values beused.

Test results for expansion anchors differ fromjob to job and with anchor size, type, andmodifications in anchor design. Assuming acoefficient of variation of 25 percent, a factor ofsafety of five on average tested anchor strength is

appropriate.The capacities of anchors are affected by

embedment depth, edge distance, and spacing.Reinforcing steel in the concrete can be used toenhance the strength of cast-in-place anchors.When the edge distance is small, closely spacedspirals of small diameter wire or mesh may beused to resist the bursting forces. However, moreresearch is required in this area. Other solutionsmay be more effective. They consist of:

(1) Providing for deeper embedment topreclude the tensile-cone-failure mode.

(2) Using larger number of smaller anchors atcloser spacings to avoid spalling when the edgedistance is too small.

(3) Preloading the anchorage so that shear istransferred by friction at the interface of the baseplate and the concrete rather than through shearin the anchor.

4.4.3 Anchor Slip -Anchors which fail by slip,without causing the concrete to fail in tension,have load-displacement characteristics similar tothe post-yield behavior of steel. Typically, wedge

bolts and sleeve anchors with embedment depthsgreater than seven bolt diameters will fail by slip.They cannot be considered ductile, however,because the relatively wide variation in the slopeof the deflection curves and ultimate loadsdistribute loads nonuniformly to the anchors. Forthese types of anchors, most manufacturers ofpost-installed expansion anchors recommendlimiting normal service loads to 25 percent of theaverage published failure loads.

4.4.4 Tensile strength of steel - When theconcrete-failure-cone strength exceeds the tensilestrength of the anchor steel, design is controlledby the strength of steel. For structuralattachments, other than simple hangers, loaddistribution to the attachments is dependent onthe stiffness of the attachment and its degree offixi ty For rigid base connections, anchor stressmay be determined assuming that plane sectionsremain plane. However, if the load is transferredfrom the attachment to the anchors through aflexible plate, the determination of anchor stressis complicated by plate stiffness, prying action, andthe load-displacement characteristics (includingpreload) of the anchor steel.

AISC imposes a minimum safety factor of two,against ultimate, for service loads on high yieldmaterials. Considering the increased loss ofpreload in concrete anchorages (approximatelythree times that of steel to steel connections), aminimum safety factor of three for anchor boltswould provide residual service load allowablesapproximating 85 to 90 percent of the residualpreload for bolts initially preloaded close to yield.This would appear to be a reasonable limitconsidering all the other concrete and anchorvariables. Proof load for concrete anchoragesshould be approximately 110 percent of the serviceload.

For factored load design, AC I Committee 349(1990) limits maximum stress to 0.9 of yield for alltypes of connections, and with stresses based onthe net tensile area for bolted connections.Assuming an average load factor of 1.6, serviceload stresses would approximate 0.55 yield foranchors other than bolts. For ASTM A 36 steel,this also closely corresponds to a factor of safetyof 3 against tensile strength.

The capacity of welded stud anchors appears tobe affected by the thickness of the attachment,Tennessee Valley Authority (1979). Apparentlyprying action, due to the flexibility of the plate,


induces very high stress and cracking at theinterior edge of the heat-affected zone of the weldunder relatively low load applications. As a result,testing clearly indicates loss of capacity withincreasing plate flexibility.

4.4.5 Shear- Shear may be transferred frombase plate to concrete either by friction or bybearing.

4.4.5.1 Shear transfer by friction - If shear is tobe transferred by friction, no lateral translation(sliding) of the base plate can occur. The normalforce necessary to develop frictional resistancemay be caused by direct load, by the compressivereaction of the applied moment, by residualpreload in the anchors, or by any combination ofthe three. If the connection is to transfer shear byfriction, the loading combination which controlsshould be that which produces the minimumcompressive reaction in conjunction withmaximum shear.

If the connection is fastened to hardenedconcrete, the coefficient of friction used todetermine shear resistance should not exceed 0.6.If the surface of a base plate is in intimate contactwith concrete or grout, shear resistance will beincreased by the cohesion between the twosurfaces and the coefficient may be taken as 0.7.

All forces contributing to frictional resistanceshould be conservatively determined in designingfor either total or partial shear resistance byfriction. Note that:

(a) Direct loads normal to the shear planeshould be the minimum associated with the load-ing condition. For cyclic loading, this would bethe maximum direct pull-off loading includingassociated impact factors.

(b) The compression component of themoment reaction is dependent on the location ofthe center of gravity of the compressive reaction.Conservative assumptions should therefore beused concerning its location. Without test veri-fication of the analytical procedure, the locationshould not be assumed to be farther than pneplate thickness from the compressive edge of theattachment.

(c) Residual preload, if any, should be basedon conservative assumptions of preload loss.Shallow depth anchors having the capability offailing the concrete in tension may be expected toexperience a total loss of preload. When theinstallation procedure requires a positive means ofdetermining installation preload, residual preload

should not be assumed greater than 50 percent ofthe initial preload without prototype testing.When the installation load is determined bycalibrated torque wrench or other less positivemeans, a higher loss should be assumed. Lostpreload may be regained by retorquing, orretightening anchors. There appears to be littleadvantage in retorquing more than twice.Sufficient time should be allowed for the majorityof loss to occur before retorquing, but under nocondition should the time period be less thanabout 1 week. Effective preload should not beassumed without verification requirements in theinstallation procedure.

4.4.5.2 Shear transfer through bearing- Iffrictional resistance is not sufficient to resistlateral sliding, shear must be transferred by theplate bearing on anchors, shear lugs, or theconcrete at the end of a fully embedded plate. Inbearing connections, shear is distributed inproportion to the stiffnesses of the shear-resistingelements, with each element contributing its share.Failure of the stiffer elements will increase lateraltranslation. The stiffer elements then transfertheir load to the remaining elements.

4.4.6 Preload-Concern for fatigue failure is aprincipal consideration in establishing servicestresses. This is particularly true for expansionanchors. If the element is subject to frequentfluctuations in stress, the magnitude of thefluctuating stress range must be restricted toprevent eventual fatigue failure (see discussion ofbehavior under cyclic loads in Chapter 3). This isbest controlled by limiting the maximum level ofdesign stress. If the bolting system can beprestressed with sufficient load that the loadremaining after losses exceeds the maximum stressload, it is generally accepted that fatigue is notlikely to occur. Under these conditions serviceload stress should be set at a level that reflects theresidual prestress. If a sustaining (residual)prestress cannot be assured, the service load stress,under fluctuating loads, must be set at a lowenough level to assure that fatigue failure will notoccur.

Assuring a level of prestress in concreteanchorages is more complicated than steel-to-steelconnections. Preload loss occurs due to creep ofthe concrete in the highly-stressed regions of loadtransfer from steel to concrete. For mostembedments the major preload loss occurs withina few days of preloading. The loss, in percent,


diminishes each time the anchorage is retorquedsuch that losses can be minimized by retorquing atabout 1 week intervals. The prestress should notexceed the yield stress of the steel. Loss ofpreload is a function of the strain relaxation(creep) relative to the total anchor strain. Sincethe major portion of load relaxation occurs at thezone of load transfer into the concrete, the loss ofpreload, in percent, can be reduced by increasingthe total anchor elongation which increases thestrain length of the anchor. If the embedmentlength of the anchor is the minimum required todevelop its tensile strength, it will lose from 40 to50 percent of its applied preload unless retorqued(Burdette, Perry, and Funk 1987). The loss maybe more pronounced if the anchor is situated incracked concrete. Loss of preload may approach100 percent for anchors of lesser embedmentdepths which are capable of failing the concrete.This is especially true for anchors located incracked concrete. To achieve an effective residualpreload, care must be taken to exclude anybonding of the anchor to grout or concrete at theembedment surface. When bond occurs at thesurface, the confinement of the surface concreteor grout, by compression of the bearing plate onthe surface, is often sufficient to locally transferthe entire load for a limited time. When thisoccurs, stretch of the bolt may be limited to thethickness of the bearing plate or attachment. Foreffective preload, threads must be excluded frombonding to either concrete or grout. Grout hassignificantly higher bonding qualities thanconcrete, therefore the entire length of bolt abovethe anchor head should be coated to prevent bondin grouted systems.

Effective prestress requires intimate contact ofthe base plate with concrete or grout at all anchorlocations. When the base plate is bolted directlyto hardened concrete without grout, effectiveprestress can be accomplished by placing shims orwashers between the plate and concrete at theanchor locations. In most moment connections,shear is transferred to the concrete entirelythrough friction and bolts transmit tension only.If the combined effect of anchor preload andcompressive reaction of the applied moment arenot sufficient for shear transfer through friction,then shear must be transferred through theanchors. If this occurs, and shims or washers areused, the combined stress in the anchor would beincreased by the increased bending stress in the

bolts in transmitting shear through the addedspace of the washers. If intimate contact is notachieved, the danger of high stress accumulationscan be prevented by initially torquing to maximumvalues and then loosening the bolts to a minimumtorque value after the concrete has had sufficienttime to consolidate in the region of the anchorhead. This will eliminate nonlinear anchordisplacement under load and restrict peak stressaccumulation to design stress levels.

4.4.7 Base plate flexibility-The flexibility of thebase plate connecting the attachment to theanchorage steel is a controlling factor indetermining the magnitude of anchor stress andthe distribution of stress to the anchors. If thedistance between exterior anchors and attachmentis more than two plate thicknesses, the plate maybe considered flexible, otherwise, the plate may beconsidered rigid. If the plate is rigid, anchor stressdue to moment is proportional to its distance fromthe neutral axis and a conventional summation offorces and moments can be used to determinestress. If the plate is flexible, anchor stress isdependent on plate stiffness as well as distance tothe neutral axis. It can also be influenced by theeffect of other stressed anchors in the group thatcause bending in the plate, and on any pryingforces caused by plate flexure, which may adddirectly to the anchor load. Anchor loads,determined by conventional analysis, may besignificantly in error if the plate is flexible.

4.4.7.1 Prying action-When load is trans-ferred from attachment to anchor through aflexible plate in full contact with the concrete orgrout, rotation of the plate at the anchor willinduce a prying force beyond the anchor wherethe plate bears on the concrete. The prying forceincreases the load in the anchor. Prying increaseswith plate flexibility which affects the magnitudeof potential downward displacement of the plateedge beyond the anchor. Prying decreases withincreased anchor displacement. Preload reducesthe displacement characteristics of the anchorunder applied loading and increases the counterrotation of the plate beyond the anchor. For thisreason anchor stress will increase with appliedload irrespective of preload. The rate of stressincrease, however, decreases with increasingpreload.

If the plate is not in contact with the concretebeyond the anchor, no prying will occur until thegap between plate and concrete is closed by the


downward displacement of the plate edge. If theanchor is not preloaded, the displacement of thestressed anchor will add to the gap requiringclosure to develop prying. If the anchor ispreloaded to close the gap, the preload force willadd to the anchor stress resisting applied loads.

4.5-Construction practicesDesign of anchor installations must take into

account local construction practice and expectedfield conditions. Details should be designed sothat the probability of concrete honeycombing atanchor locations is minimized. Placementtolerances may or may not be critical and shouldbe determined by the application. See Chapter 5for more information.

4.6-REFERENCES

Abbot A. Hanks, “Summary Report - Kwik-Bolt TestingProgram”, File No. H2189-S1, Report No. 8783, Abbot A.Hanks Testing Laboratories, San Francisco, CA

ACI Committee 349,1990, “Code Requirements for NuclearSafety Related Concrete Structures (AC1 349-90), “AppendixB”, American Concrete Institute, Detroit.

ACI-ASCE Committee 326, 1962, “Shear and DiagonalTension”, ACI Journal, Proceedings V 59, No. 2, Feb., pp. 277-333.

Burdette, E.G., Perry, T., Funk, R.R., 1987, “LoadRelaxation Tests”, ACI SP-103 Anchorage to Concrete, Detroit,pp. 297-311.

Cannon, Robert W., 1981, “Expansion Anchor Performancein Cracked Concrete”, ACI Journal Proceedings V. 78,November-December, pp. 471-479.

Eligehausen, Rolf, 1987, “Anchorage to Concrete byMetallic Expansion Anchors”, Anchorage to Concrete, AmericanConcrete Institute Special Publication SP-103, pp. 181-201.

Orangun, C.O., Jirsa, J.O., and Breen, J.E., 1977, “A Re-evaluation of Test Data on Development Length and Splices,”ACI Journal, Vol. 74, No. 3, pp. 114-122.

Raphael, Jerome M., 1984, “Tensile Strength of Concrete”,ACI Journal No. March - April, pp. 158-165.

Tennessee Valley Authority, 1979, “Welded Stud Anchors,Effect of Plate Flexibility on Stud Capacity”, CEB Report No.79-18, TVA, Knoxville, TN.

United States Nuclear Regulatory Commission, 1979, “PipeSupport Base Plate Designs Using Concrete Expansion AnchorBolts”, IE Bulletin No. 79-02, Office of Inspection andEnforcement, Washington, D.C.

C H A P T E R 5 - C O N S T R U C T I O N C O N -SIDERATIONS

5.1 IntroductionQuality control is the central issue among

construction considerations for anchorage toconcrete. In construction, the engineeringprofession tends to be quite meticulous withrespect to tolerance in the fabrication of structuralsteel, but somewhat less so in masonry, timberframing, and reinforced concrete. Meetingtolerances is expensive and, therefore, requiredtolerances are limited to what is practical andwhat “can be covered by the other trades” and stillyield an acceptable product.

Another concept in establishing tolerances is toweigh the consequences of constructing lessaccurately than specified. Experience has shownthat the secondary costs of compensating for thestructural skeleton being out of square or out ofplumb justify taking great care in the initialfabrication. This is also true for anchorage toconcrete. There are few details in a structurewhere care during installation pays moredividends, or where carelessness can prove morecostly. Sometimes corrective measures can be soexpensive that they are not taken and the endproduct falls far short of what the engineerintended.

Anchorage details are at the interface andprovide the connecting link between separatestructural systems. The axial load, moment, andshear required of the connection are typicallyquite well defined, and must be accommodatedbecause there is usually no alternative path forload transfer. The joint has a minimum ofredundancy to compensate for error in design orconstruction. Accordingly, it is important that thefield engineer understands the intent of thedesign, to assure that the anchorage beconstructed as specified. This relates to havingthe proper device, with the specified size andmaterial, and having it properly installed.

5.2 Shop drawings/submittalsThe first step in quality control is that the plans

and specifications must indicate clearly what isintended. The next step is the requirement forsubmittals and shop drawings for all anchorages.

5.2.1 Cast-in-place systems-For cast-in-placesystems, the submittal is the shop drawing and anyother certifications required by the constructionspecifications. With respect to each anchorage

assembly, the shop drawing should indicate thematerial of the anchoring device, its coating,length, diameter, length of threaded portion,diameter and thickness of washers, number of nuts(single, double, single or double plus leveling,etc.), and torquing requirements, if any.

The shop drawing should also indicate thelocation of the anchorage in the structure, thelocation of the bolts (or devices) in a group, andtheir projection and embedment with respect tothe finished concrete grade. When the completedanchorage is specified to be either grouted or drypacked, the dimensional details of the grout ordrypack should be shown.

When the anchorage consists of embeddeddowels of reinforcing steel, the shop drawings forthe anchorage are included in the shop drawingsfor the reinforcing steel. They should indicate thetype of steel, details of bending, location (bar orgroups of bars), embedment, and projection.

Often an anchor assembly includes embeddedstructural shapes, either as the anchor itself or asa lower template. Shop drawings for theseembedded shapes should indicate type of steel;coating; cross-sectional shape (standard desig-nation); dimensions and details of the member orgroup of members in the assembly (location, type,size, and length of welds); size and location ofholes; and embedment depth.

5.2.2 Post-Installed Systems-For post-installedsystems, the submittal should include the shopdrawings with information similar to that requiredfor cast-in-place systems, plus manufacturer’sliterature which adequately describes the deviceand its capabilities and provides instructions for itsproper installation.

5.3 TolerancesThe acceptable variation from the specified

positioning is the tolerance. The tolerancesshould be specified by the engineer and beappropriate for the application. Table 5.1 gives
suggested tolerances for anchor positioning andcan be used as a guide in determiningacceptability. Other sources such as the AmericanInstitute of Steel Construction (AISC) andPrecast/Prestressed Concrete Institute (PCI) areavailable. These requirements are rigorous, butmeeting them is judged to be more economicalthan the consequences of not meeting them.Mounting or anchoring certain special equipmentmay require even closer tolerances.
5.4 lnstatllatlon of anchors5.4.1 Cast-in-place systems

5.4.1.1 Anchors Embedded, Non-Adjustable-Anchorages that fall into this category(see Table 2.1) can be grouped as follows:

- Bolts installed in plastic concrete - Bolts incans or blockouts

- Bolts, with or without sleeves, positionedwithout template

- Bolts, embedments, weld plates, or insertsattached to the formwork

- Bolts or groups of bolts, with or withoutsleeves, positioned by top or bottom templates, orboth

- Embedded structural shapes

5.4.1.2 Bolts installed in plasticconcrete-Often in wood frame construction thebolts connecting the wood sill to the footing or awood plate to the top of a wall are installed assoon as the concrete placement is completed.This practice is not recommended because a goodbond may not be achieved.

5.4.1.3 Bolts in Cans or Blockouts-Thissystem can be used in cast-in-place or post-installed construction. Often, for machineryfoundations or in situations where it is notdesirable to have anchor bolts protruding from aslab or penetrating through a wall form, a can orblockout will be set at the approximate future boltlocation. These blockouts can be made of wood,metal, or plastic; can be cylindrical or prismatic;can provide a shear key perpendicular to the flooror wall; or be battered to provide a dovetail effect.

For flatwork, the cans or blockouts can bepositioned by wood battens or templates whichhave a soffit elevation equal to the grade at top-of-concrete and are secured to the edge forms; orthey can be wired to the reinforcement or theedge forms. For vertical surfaces, they can befastened to the wall form in their predeterminedpositions. In both cases, the blockout should bewire tied to the reinforcement so that it will notbe vibrated out of position during the placementof the concrete.

After concreting, wood blockouts are stripped.Metal or plastic units are typically left in place.The pocket is blown clean of debris, the anchorbolt positioned and the pocket grouted.

Table 5.1 -Suggested tolerances for installation of anchors in concrete

/3 Per Plans

L and P are Specified

InstalledLocation

CorrectLocation

Plan “A”

A. Cast-in-place

Type of anchorageSuggested tolerances

Vertical

Projection Positioning Alignment,

P, in. r, in. (r, deg.

1. Common bolt, J- or Gbolt, continuously threaded rods

2. Reinforcing steel

3. Embedded structural shapes

4. Weld plates

5. Troughs for adjustable anchors

6. Temporary embedded inserts

k /14

f 1/4

Flush with concrete

Flush with concrete

Flush with concrete

I 1/16 IPer recommendations of Committee 117

1/8

l/8

l/4

l/2

3.0

3.0

N/A

N/A

3.0

B. Post-installed

1. Drilled and grouted-all types

2. Expansion types

+ l/4 1/16 3.0

k 1/8 1/16 3.0

5.4.1.4 Anchors with or without sleeve,positioned without templates-This practice isgenerally not recommended, but if proper care istaken it can be successful. Tolerance criteriashould be met and maintained throughout theconcrete placement. The bolt insert or sleeveshould be rigidly tied with wire to the rein-forcement, top and bottom. Sometimes, thesleeve, the bolt head, bottom washer, or insert istack welded, according to approved procedures, tocross bars which in turn are wire tied or tackwelded to the existing reinforcement.

5.4.1.5 Bolts, embedments, weld plates or insertsattached to the formwork-This work generallyrelates to soffit and wall forms. The importantstep is to accurately scribe the inside of the formfor proper location of the anchoring unit. Theanchor unit should then be nailed or bolted to theform or wire tied to the reinforcing steel, or bothso that neither internal nor external vibration candisturb or move the anchorage unit out ofposition.

5.4.1.6 Bolts or groups of bolts, with or withoutsleeve positioned by templates -These installationsare generally used in flatwork, where the bolts arevertical. The use of templates is the besttechnique for guaranteeing that the anchorage iscorrectly positioned.

A top template is often wood, although in“loose base plate” construction (where thesuperstructure is subsequently welded or otherwiseconnected to a steel base plate), the base plateitself can be used as the template. The toptemplate for a single bolt or a group of boltsgenerally has a soffit elevation at or above the topof the finished concrete. Sometimes top templatesare plywood with the holes either laid outprecisely as the holes in the base plate, or actuallydrilled using the base plate holes as a guide forthe drill. Where only a top template is used, thereshould be nuts above and below the template tohold the anchor bolt in a plumb position.

The bottom template is a steel assembly ofangles, channels, or flat bars. Low carbon steelbolts can be precisely positioned and weldeddirectly to the steel template, or set in accuratelylocated holes in the template, and tack welded.When used in conjunction with a top template, itis the top template that controls both the boltprojection and lateral position of the group ofbolts. When there is no top template, the bottomtemplate must provide those controls and should

be wire tied or welded, according to approvedprocedures, to the reinforcement so that it will bemaintained in correct position while the concreteis being placed. Engineering approval should beobtained before welding to high-strength bolts orreinforcing bar, because material property changesmay compromise expected steel capacities.

Bottom templates are expensive and usuallyreserved for larger diameter bolt installations.They also affect the capacity of an anchorage andfor this reason should only be used where detailedor approved by the Engineer.

5.4.1.7 Embedded structural shapes -Thissystem is used mainly for transmission towers,although it has been used for other applications.The superstructure can be erected plumb, leveled,and set to grade in holes augered in the ground.Then concrete is cast around the structural shape.Alternatively, the anchoring elements, usuallyangles, are cast in the footings and the tower orsuperstructure subsequently bolted to them. Theanchor installation is straightforward but care isrequired through the use of templates and guidesto maintain proper location in plan, and at theproper grade, batter, and plane of batter.

5.4.1.8 Adjustable anchors -Anchors of thistype are patented devices used principally inflatwork. Most often they are used for machineryinstallation and are designed to compensate fornormal field tolerances in the positioning ofanchor bolts. They offer an added advantage inthat there are no bolts projecting above the floorprior to setting the machinery. The machinery canbe moved into place on its base and then the boltsset. One features a trough set flush with thesurface of the concrete and stud anchored to theconcrete below. Another features deeplyembedded pockets, housing a tapped bottomwasher plate and having a sleeve extension up tothe surface of the concrete. The devices arepositioned and held in position during concreteplacement in a manner similar to that describedfor sleeves. The principal concern is that theinsert be maintained level. The bolt is normallygrouted in place at the same time that theequipment base plate is grouted.

5.4.1.9 Common bolts pretensioned -Boltinstallation is as described in further detail inSection 5.5.1. The shank of the bolt should be
coated with bond breaker before placing concrete.After concreting, the annular space around sleeved


5.5 Inspection5.5.1 Cast-in-place systems -The inspector has

the responsibility to verify that the size andlocation of anchors or anchorage assemblies are inaccordance with the construction plans andspecifications, prior to the placement of concrete.Anchors must be located properly in plan, havethe proper projection, and be rigidly held in placeso as not to be disturbed during the placementand finishing of the concrete. Methods ofsecuring the anchorage in place include:

- Nailing to the forms (conditions applicable)- Nailing the top template to the forms- Wire tying individual bolts, or their bottom

template, to the forms or the reinforcement and- Tack welding to the reinforcement, if

approved. (High strength bolts should not bewelded)

Welding should be to the bottom washer or thebottom template of the bolt head, rather than theshank of the bolt.

In the case of bolts that are subsequently to betensioned, the inspector should verify that

bolts is grouted. When concrete and grout (anddry pack under baseplate) has cured the specifiednumber of days, screw on the nut and apply thepretensioning load with a torque wrench. Torqueshould initially be about 50 percent of desiredtorque, then to 90 percent, working from one boltto the one diagonally opposite and thusprogressing through the group. The final 10percent of torque should be applied to all bolts insequence. After 1 week verify that pretension hasheld, or retension to specified torque, if necessary.

5.4.2 Post-installed systems5.4.2.1 General anchor types -Anchors in this

group include:Common bolts, reinforcing bars, and

continuously threaded rods- Bonded (grout and chemical) anchors- Rock bolts- Expansion anchors

5.4.2.2 Common bolts, reinforcing bars,threaded rod-Section 5.5.2 applies for positioning
and drilling the hole; Section 5.6 for grouting.
5.4.2.3 Chemical anchors-These are similarto grouted anchors, with an adhesive, such asepoxy, polyester, or vinylester taking the place ofthe grout. Section 5.5.2 applies as far aspositioning and drilling the hole for the anchor.The adhesives are proprietary and installationshould follow manufacturer’s instructions.

Drilled hole diameters may vary from 1.0 to2.0 mm larger than the nominal steel diameterwithout affecting loading capacity for polyesterand vinylester anchoring systems. Storage shouldfollow manufacturer’s recommendations to preventheat, ultraviolet light, or both from shortening theshelf life of the unused product. Anchoringsystems using epoxies are not sensitive to thesesame storage requirements.

5.4.2.4 Rock bolts - Rock bolts occasionallyare used for anchoring to concrete. There aremany types available. Section 5.5.2 applies as faras positioning and drilling the lead hole. In thecase of the split end variety, bondbreaker isapplied to part of the shank and then the rockbolt is then inserted in the hole with the wedgelightly set in the split tail of the bolt. The nut isin place on the bolt, flush with the end. The boltis then rammed down over the wedge until thebolt is well set in the hole. It is then adjusted forvertical alignment and grouted per Section 5.6.

5.4.2.5 Expansion anchors-These systemsinclude a myriad of devices. They are self-drilling,

or set in predrilled holes. The wedging actionbetween the device and the sides of the hole isactuated by placing tension on the bolt, by turningthe bolt, by hammering the bolt onto a spreader(cone or wedge) in the bottom of the hole, or byhammering a spreader into the bottom expandingportion of the anchor. The manufacturer’sinstructions for installation of expansion anchorsmust be followed meticulously. This appliesparticularly to the diameter and depth of hole.Some systems afford the opportunity of using thebase plate or element being connected as atemplate in drilling the embedment hole. Othersrequire a larger hole to accommodate a sleevethat bears against the bottom of the connectedbase plate.

Expansion anchors can lose preload under acyclic loading or from concrete creep due to highlocal expansion forces unless they are sopretensioned that the bolt is always in tensionunder all loading conditions. Generally, todevelop the pretension load the wedge orexpansion device must first be “set” against theside of the hole. With certain types of anchorsthere may be an initial slip which should beanticipated and designed for. In the case ofexcessive slip, follow the recommendations inSection 5.7.2.


5.5.2 Post-installed systems -Post-installedsystems involve setting the anchor in blockouts ordrilled holes. The inspector should verify that theblockouts or holes are properly located. Withdrilled holes he should verify that the drill bit is ofthe proper diameter, that the hole is plumb to thesurface (bit guides should be used for criticalwork), that the finished hole has the properdiameter and depth, and that the appropriatedrilling equipment is used. This calls for rotarydrills (carbide tip or diamond studded bits) orhand hammered star drill bits. Jackhammeringshould not be permitted because of the damage itdoes to the concrete immediately around the hole.

Once the hole is drilled and blown clean, theanchor should be installed, preloaded, and tested(as required) in accordance with Section 5.5; orthe hole should be protected by plugging it with arag or other suitable stuffing until the time ofanchor installation.

Guidance for inspecting grouted anchors isgiven in Section 5.6.

5.6 -GroutingACI Committee 351, Foundations for Equip-

ment and Machinery, includes in its workdevelopment of information on grouting. Accord-ingly, reference is made to publications of thatcommittee. The statements which follow are in-tended to be a brief summary of grouting as itrelates to construction considerations for concreteanchorages.

5.6.1 Materials - Grouting materials fall into twobroad functional categories: nonprecision groutsand precision, “nonshrinking” grouts.

5.6.1.1 Nonprecision grouts - Nonprecisiongrouts include mixtures of cement and water, withor without the inclusion of sand or admixtures.The use of the “jobsite mixed” or packagedproducts not designed to perform as a precisiongrout has limitations. The most significantlimitation is the lack of a mechanism forovercoming drying shrinkage which occurs as freemoisture leaves the grout.

Dry packing with cement, sand, and onlyenough water to result in a stiff, but cohesivemixture has been used in grouting for many years

unsleeved bolts, or sleeved bolts that are to begrouted prior to the tensioning, have a bondbreaker (grease or other) on the shank that willprevent the bolt from bonding to the concrete orgrout.

and is an excellent method, but it is laborintensive, and in many installations is impractical.

Epoxy grouts also have been used successfullyfor a number of years. These materials offer high,early strength and provide excellent bond andprotection of steel in corrosive environments.There are, however, some limitations in the use ofthese materials. The concrete and steel surfacesto be in contact with epoxy must be cleaned and,for most epoxies, dry. Epoxies also have acoefficient of thermal expansion several times thatof the concrete or steel, which should be takeninto consideration. Epoxies can creep undersustained loading of the anchor, and some epoxygrouts lose strength when exposed to temperaturesover 120 F.

5.6.1.2 Precision, "Nonshrinking" grout-Theseportland cement based products are proprietaryand sophisticated in terms of their cementchemistry and composition. They comply with therequirements of the U.S. Army Corps of Engin-eers specifications for nonshrink grouts, CRD-C-621.

Precision grouts are proportioned to lessen theeffects of plastic and drying shrinkage in theplastic and hardened states. Accordingly they areexcellent materials to use in complex groutingsituations, such as the grouting of machinerybases.

5.6.2 Applications - Grouting of anchorages toconcrete falls into three application categories:

- Grouting of anchor bolt holes and sleevesprior to base plate installation

- Grouting or dry-packing of base plates andmachinery bases

- Grouting bolt holes after pretensioning of theanchor bolt

5.6.3 Construction procedures5.6.3.1 Preparation -Anchor bolt holes and

sleeves should be clean and free of oil, grease,dirt, or other debris. Bolt holes should preferablyhave a textured surface, thoroughly moistenedprior to grouting, but with no free moisture in thehole.

5.6.3.2 Mixing and placing-Grouts may bemixed in mortar mixers or in smaller vessels, as isappropriate to the work. When using proprietaryproducts, follow the manufacturer’s instructionsfor mixing. The “pot” life is a very importantconsideration.

Proper placement of grout is important.Whether dry packed or poured at a fluid


5.7.2 Post-installed systems-A common fieldproblem in post-installed systems is interferencewith the in-place reinforcement. The location ofthat reinforcement can be determined magneticallyor radiographically. Sometimes, it is simplydiscovered when the drill bit, drilling the hole, hitssteel. When an anchorage interferes with any in-place reinforcement, the Engineer-of-Recordshould decide on the remedy. Wherever possible,the anchorage itself should be shifted to a newlocation where there is no interference. Momentreinforcement should never be welded or cut.W i t h d u e c o n s i d e r a t i o n , t e m p e r a t u r ereinforcement can be cut.

A second problem is excessive slip inpretensioning the bolt. This can be indicative ofan oversized hole or a faulty anchoring device.When excessive slip occurs, the assembly shouldbe reinstalled in the hole and the pretensioningapplied such that the slip does not exceed theallowable limit (i.e., resulting embedment isadequate). Sometimes the entire anchor will haveto be replaced, or possibly the hole drilled to alarger size and the next larger sized anchorinstalled.

consistency, the material should be placed orpoured in a manner which will preclude theentrapment of air which produces voids in thehardened grout.

5.6.3.3 Curing-Curing is important inachieving satisfactory results in any groutinstallation. Normally this is accomplished byplacing water-saturated rags over all exposed groutsurfaces as soon as possible after grout placement.These rags should be maintained wet and in placefor at least 24 hr after which the exposed surfaceof the grout is coated with a curing compound ifsecondary grouting will not follow. Wheresecondary grouting is to follow, continue the watercuring for 7 days, or until placement of the secondgrout. Proprietary grouts should be curedaccording to the manufacturer’s recommendations.

5.7 -Field problems5.7.1 Cast-in-Place Systems-The common

problem encountered in the preconcreting stage isinterference with existing reinforcement. In thiscase a decision has to be made whether to movethe anchorage or move the reinforcement. Inweighing the consequences of each, the FieldEngineer, perhaps after consulting the Engineer-of-Record, establishes which has priority.

Another common problem is to discover, afterthe concrete has hardened, that the anchorage hasshifted during the placement of the concrete, andthat the base plate will not fit in place, or thatthere is insufficient thread projecting to fullyengage the nut. These problems can and shouldbe avoided by proper inspection, or by use ofsleeved or adjustable anchors. The specificationsshould cover these possibilities, and state that it isthe contractor’s responsibility to take necessaryprecautions and corrective measures. Actionstaken when field errors are discovered should havethe approval of the Engineer-of-Record.

Bending of protruding bolts is discouragedbecause the bending stress which results from theeccentricity of the service load, when added to thedesign axial and shear stresses, can often exceedthe yield strength of the bolt. In welding tocompensate for insufficient thread being engagedby the nut, care should be taken that the weldacting alone will develop the strength of the bolt,because the capacity of the welds and the engagedthreads are not additive. When any embeddedanchor is not installed within allowable tolerances,the structural adequacy of the installation should

be verified by the Engineer-of-Record and, ifnecessary, the design should be modified.

The single most helpful practice for avoidingthe problem of cast-in-place anchor bolts notfitting the base plates is to make holes in columnand machinery base plates oversize, and thengrout the annular space after the base plate is inplace, or use specially designed washers. Thefollowing schedule of oversize holes isrecommended.

- Bolts less than 1 in. diameter - 5/16 in.oversize

- Bolts 1 to 2 in. diameter - l/2 in. oversize- Bolts over 2 in. diameter - 1 in. oversize

CHAPTER 6-REQUIREMENTS IN EXISTINGCODES AND SPECIFICATIONS6.1 -Introduction

Sources of information relating to codes andspecifications on anchorage to concrete arepresented in this section. Sources are referencedin alphabetical order. American and internationaldocuments are included in this state-of-the-artreview.


6.2 -Existing codes and specifications6.2.1 American Association of State Highway

Transportation Officials (AASHTO)6.2.1.1 Standard Specification for Highway

Bridges -For composite bridge decks, AASHTOuses the ultimate capacity of stud shear connectorsand a reduction factor t$ of 0.85 for design.Design checks are required for horizontal shearunder working loads. Working loads arecompared to allowable loads which include areduction for fatigue.

AASHTO Section 1.7.56 bases the number,required embedment, and size of anchor bolt onthe span of the bridge, and requires that theanchor bolt be swedged or threaded to insure asatisfactory grip on material such as the grout.AASHTO requires that anchor bolts subject totension be designed to engage a mass of concretewhich will provide a resistance equal to one andone-half times the calculated uplift.

6.2.2 American Concrete Institute (ACI)6.2.2.1 ACI 318, Building Code Requirements

for Reinforced Concrete - ACI 318-63 containedallowable bond values for plain (smooth) bars.Many engineers have used these values fordetermining embedment requirements for cast-in-place anchor bolts. The current edition of ACI318 does not give allowable bond values for plainor deformed bars. Section 12.6.1 states “Anymechanical device capable of developing thestrength of reinforcement without damage toconcrete may be used as anchorage.” Section15.8.3.3 of ACI 318 states “Anchor bolts andmechanical connectors shall be designed to reachtheir design strength prior to anchorage failure orfailure of surrounding concrete.”

6.2.2.2 ACI 349, Code Requirements forN u c l e a r S a f e t y R e l a t e d C o n c r e t eStructures-Appendix B of ACI 349 givescomprehensive procedures for designinganchorages and steel embedments that are used totransmit loads from attachments to reinforcedconcrete structures governed by ACI 349. Thebasic philosophy of anchorage requirements inACI 349 is consistent with the ultimate strengthdesign philosophy of reinforced concrete. Thefailure mechanism is controlled by requiringyielding of the steel anchor prior to brittle failureof the concrete.

This design method considers not onlytraditional design parameters, i.e., steel strength,concrete strength, and anchor size, but also other

variables such as anchor type or form, spacing,edge distance, nature of the anchor load, thicknessof the concrete member, and concrete stress in theanchor zone. Concrete strength is critical toassure that the reinforced concrete structureexhibits ductile failure, which is also an ACI 318requirement. Note, however, that many of thepost-installed systems feature the brittle concrete-cone failure.

The commentary of ACI 349, Appendix B,provides an excellent source of information ontypes of anchorage devices, design requirements,modes of failure, and testing.

6.2.3 American Institute of Steel Construction

(AISC)6.2.3.1 Manual of Steel Construction -The

AISC “Specification for the Design, Fabrication,and Erection of Structural Steel for Buildings” setsallowable bolt stresses in Sections 1.5.2 and 1.6.3.These values apply to certain cast-in-place andgrouted anchor bolts and are valid for allowableanchor steel stresses, but no values are givenwhich relate to the transfer of these stresses to thesurrounding concrete.

The AISC specification gives allowable values inshear for stud shear connectors used for compositedesign in Table 1.11-4. The listed values cannotbe used for anchor bolts of the same size. Thevalues used in Table 1.11-4 are based on equationsderived from a testing program and the ultimatestrength of the composite member, using a factorof safety of 2.0.

The AISC code commentary contains thefollowing warning:

“The values of q in Table 1.11-4 must not beconfused with shear connection values suitable foruse when the required number is measured by theparameter VQ/I, where V is the total shear at anygiven cross-section. Such a misuse could result inproviding less than half the number required byFormulas 1.11-3, 1.11-4, or 1.11-5.”

The AISC specification also gives settingtolerances for bolts used to anchor structuralmembers; however, these tolerances are unsuitablefor anchoring machinery.

6.2.4 American Society for Testing and Materials(ASTM)

6.2.4.1 Annual Book of Standards - Volume04.07 contains test standard ASTM E 488,“Standard Test Methods for Strength of Anchorsin Concrete and Masonry Elements.” This teststandard describes procedures for determining the


static, dynamic, and fatigue tensile and shearstrengths of cast-in-place, chemical, grouted, andexpansion anchors.

Volume 15.08, Fasteners, contains variousASTM specifications for the steel used for bolts,including A 193, A 307, A 325, A 449, and A 490.

6.2.5 Construction Industry Research andInformation Association (CIRA) (Great Britain).

6.2.5.1 Section and Use of Fixings in Concreteand Masonry (Guide 4) - CIRA Guide 4, is acomprehensive guide on the selection and use ofanchors installed in concrete. Three maincategories of anchor types are covered. Theseinclude cast-in-place, expansion, and bondedanchors. The guide also covers behavior offastener assemblies under load, designconsiderations, limitations, durability, testing, andpractical considerations.

6.2.6 Institut fir Bautechnik (IfBT)(WestGermany)

6.2.6.1 Tests to Evaluate the Strength ofMetallic Expansion Bolts for Anchorage in Concretewith an SC of 20 MPa (2500 psi) orGreater-Approvals are based on results of testscarried out by licensed universities.

In the tests the proper functioning of theanchors under extreme conditions are checked,and tests to evaluate allowable loads for designare performed.

For evaluating allowable conditions of use (e.g.,allowable loads, required edge distance, andspacing), a sufficient number of tests have to beperformed to calculate a statistically reliableconfidence level for the failure loads [5 percentfractile (or 95 percentile) of failure loads]. Asafety factor of 3 is applied to the determined 5percent fractile of the failure loads to account forthe variations of the concrete tensile strength andof jobsite installation quality. For reasons ofsimplicity, one value for the allowable load isgiven per anchor size which is valid for all loadingdirections (tension, shear, combined tension, andshear). Expected displacements of anchors underallowable loads are given which should be takeninto account in the design of the fastened element(when appropriate).

6.2.7 International Conference of BuildingOfficials (ICBO)

6.2.7.1 Uniform Bui lding Code (1985Edition) -The Uniform Building Code (UBC),Table 26-G sets forth allowable shear and tensionloads for cast-in-place bolts of at least ASTM A

307 quality or better.The table assumes an anchor spacing of 12

anchor diameters. The spacing may be reduceddown to 6 anchor diameters with a 50 percentreduction in allowable load values. A minimumedge distance of 6 anchor diameters is required.Edge distance may also be reduced up to 50percent, provided that the listed values arereduced in equal proportion. Tension values listedin the table may be increased 100 percent when“special inspection” is provided. UBC Section2719, on anchor bolts for steel column bases, doesnot provide design values for anchor bolts, butsimply states that “Anchor bolts shall be designedto provide resistance to all conditions of tensionand shear at the bases of columns.” The sectionon steel column anchorage does not refer to TableNo. 26-G. Application of this table to steelcolumn anchorage would greatly affect currentdesign practice because of the requirement inTable No. 26-G of a minimum spacing of 6 anchordiameters.

6.2.8 Precast/Prestressed Concrete Institute (PCI)6.2.8.1 PCI Design Handbook-The handbook

gives equations for shear and tension loadallowables for headed shear stud anchors.Combined loading, as well as required edgedistances and anchor spacing for groups ofanchors, are covered.

Based on a review of past design methods andactual testing and modeling, the PCI ConnectionDetails Committee recommends the use of aprojected cone model to define the actual bolttension at which concrete failure will occur. ThePCI cone surface equation is:

Pm = 2.8 hL@ [fi 7~ ld (li + da] (6.1)

whereiz = 1.0 for normal weight concrete1 = 0.85 for sand lightweight concrete

=;=

0.75 for all lightweight concreteembedment, in.

d,, = diameter of anchor or stud head,in.

f’, = specified 28-day compressivestrength of concrete, psi

P =IlC nominal tensile capacity of anchor

as governed by concrete failure

In anchor bolt design where the concrete doesnot fail, the anchor bolt fails via a combination oftension and shear. The PCI equation for


combined tension and shear strength is:

where4 = strength reduction factorPu = applied factored tension loadPYlC = nominal tension strength of

anchorvu = applied factored shear loadKc = nominal shear strength of anchor

as governed by steel failureIn-depth discussions of these equations may be

found in Klingner and Mendonca (1982) andShaikh and Yi (1985).

6.2.9 The Agrbnent Board (Great Britain)6.2.9.1 The Assessment of Torque-Expanded

Anchor Bolts When Used in Dense AggregateConcrete (M.O.A. T. No. 19:1981) -This documentpresents the procedures for deriving designinformation and classifies ten different types ofexpansion anchors according to the mechanism forachieving expansion. It considers the effects ofdifferent types of loading conditions and typicallyrequires a minimum of 277 tests (for six differentanchor diameters) to calculate safe working loadsas the lower of:

a. The 5 percent exclusion value (or 95thpercentile, calculated by regression analysis orother statistical techniques), then divided by threeor,

b. The mean of the loads determined at adisplacement of 0.1 mm (0.004 in.) under directtension or,

c. The mean of the loads determined at adisplacement of 1.0 mm (0.039 in.) under directshear.

6.2.10 UEAtc (Union European of Agrbment)The UEAtc Directives for the Assessment of

Anchor Bolts (December, 1986) is a Europeancode for the assessment and approval of anchorbolts. The document has been adopted by theCommon Market Countries of Germany, U.K.,France, Austria, Italy, Spain, Ireland, Netherlands,Portugal, Denmark, and Belgium.

6.2.11 Nuclear Regulatory Commission (NRC)Bulletin 79-02 and 79-14).

Anchor bolt design methods have been revisedbased on the United States NRC Office ofInspection and Enforcement Bulletins No. 79-02

and 79-14. Only Class I piping (piping used tosafely shut down a nuclear power plant) wasimpacted by Bulletins 79-02 and 79-14. The NRCrequires that during anchor bolt design, thefollowing must be considered: baseplate flexibility,(i.e., baseplate prying action that increases anchorbolt loading), performance of anchors due tocyclic loading, anchor performance in masonrywalls, the effect of pipe support loads on masonrywalls, and the maximum support load consideredfor anchor bolt design. Concrete expansionanchors must have the following minimum factorof safety between the bolt design load and the boltultimate capacity determined from static loadtests, (e.g., published data from the anchor boltmanufacturer) which simulate the installationconditions, (i.e., type of concrete and its strengthproperties): (1) a safety factor of 4:1 - for wedge-and sleeve-type anchor bolts, (2) a safety factor of5:l - for shell-type anchor bolts.

The bolt ultimate capacity should account forthe effects of shear and tension interaction,minimum edge distance, and proper bolt spacing.

A summary of the USNRC criteria is found inUSNRC “Anchor Bolt Study Data Survey andDynamic Testing” by the Hanford EngineeringDevelopment Laboratory.

6.2.12 Draft 1 Regulatory Guide MS 129-4“Anchoring Component and Structural Supports inConcrete”

This draft guide from the U.S. NuclearRegulatory Commission provides the criteria foracceptance, qualification, design, installation, andinspection for steel embedments anchored inconcrete. It also provides information on theacceptability for NRC licensing actions inaccordance with Appendix B, of ACI 349-80.

6.3 -Application and development of codesASTM E 488 is the only existing American

standard exclusively and specifically concernedwith testing to determine the performance of alltypes of concrete anchors. It is not intended todescribe design procedures for anchorageconnections, nor to identify characteristics whichaffect performance in conditions other than as-tested. ICBO has also published a limited teststandard for expansion anchors only.

ACI 349, Appendix B, specifies anchoragedesign and applies ultimate strength designphilosophy to all types of anchorages. OtherAmerican codes limit their consideration to cast-


in-place or grouted anchorages. The UniformBuilding Code (UBC) allows for alternativedevices as specified in the code, generally applyingthe same conditions as specified for cast-in-placeanchors.

American codes generally base recommendeddesign procedures on ultimate strength data.European codes recommend the criterion ofdisplacement (slip) for post-concreting anchors,supported by ultimate strength data derived byregression analysis of other statistically reliabletechniques.

Codes cannot address all the conditionsapplicable to a particular design or absolve thedesigner of the responsibility to check therelevance of code data for a given design. Newand technically reliable information will inevitablybe developed between publication dates ofamendments to existing codes. Designers areencouraged to maintain familiarity with ongoingresearch and other developments and tosupplement the provisions of governing codes withsuch information as it becomes available.

6.4 - ReferencesACI Committee 318,1989, “Building Code Requirements for

Reinforced Concrete (ACI 318-89) and Commentary - ACI 318R-89, American Concrete Institute, Detroit, MI, November.

ACI Committee 349, 1990, “Code Requirements for NuclearSafety Related Concrete Structures (ACI 349-90) andCommentary - ACI 349R-90, American Concrete Institute,Detroit, MI, March.

Fasteners. 1988 Annual Book of Standards, Volume 15.08,American Society for Testing and Materials, Philadelphia, PA,January, 1988.

Klingner, R.E. and Mendonca, J.A., (1982a) “TensileCapacity of Short anchor Bolts and Welded Studs: A LiteratureReview,” ACI Journal, Proceedings, V. 79, No. 1, July-August.

Manual of Steel Construction. Eight Edition, AmericanInstitute of Steel Construction, Inc., New York, NY, 1980.

Paterson, W.S., “Selection and Use of Fixings in Concreteand Masonry”, CIRA Guide 4, Construction Industry Researchand Information Association, London, England, October, 1977.

PCI Design Handbook, Third Edition, Prestressed ConcreteInstitute, Chicago, IL, 1980.

“Prtifungen zur Beurteilung d e r Tragfghigkeit v o nzwangsweise s p r e i z e n d e n Diibeln aus MetaIl nach de rVerankerung in Normalbeton 1 Bn 250” (Tests to Evaluate theLoad Capacity of Metal Expansion Anchors Fastened intoNormal Concrete, r Bn250), Institute for Construction (IfBT),Berlin, West Germany, January 1974.

Shaikh, A.P., Yi, W., “In-Place Strength of Welded HeadedStuds,” PCI Journal, V.30, No. 2, March-April, 1985.

“Standard Specification for Highway Bridges”, TwelfthEdition, American Association of State HighwayTransportation Officials, 1977.

“Standard Test Methods for Strength of Anchors inConcrete and Masonry Elements”, (ASTM E488-88), 1988Annual Book of ASTM Standards, Volume 04.07, AmericanSociety for Testing and Materials, Philadelphia, PA, October,1988.

“The Assessment of Torque-Expanded Anchor Bolts whenused in Dense Aggregate Concrete”, M.O.A.T. No. 19:1981,Agrkment Board, Watford, Herts., England, January, 1981.

“UEAtc directives for the Assessment of Anchor Bolts”,M.O.A.T. No. 42:1986, European Union of AgrCment;December, 1986.

Uniform Building Code, International Conference ofBuilding Officials, Whittier, CA. 1985.

USNRC “Anchor Bolt Study Data Survey and DynamicTesting”, Hanford Engineering Development Laboratory,NUREG/CR-2999, December, 1982.


APPENDIX A-CONVERSION FACTORS:

Length

Area

Volume

ft

ft2

ft3

Velocity ft/s

Acceleration ft/s2

Mass “An

Force and Weight

Pressure and Stress

lb f

lb$ft2psipsi

Work and Energy

Mass Density

Weight Density

ft-lbf

lb,/ft3

lbf/ft3

INCH-POUND TO SI

Multiply By

3.048 x 10-l

9.290 x 10-2

2.832 x 1O-2

3.048 x 10-l

3.048 x 10-l

4.536 x 10-l

4.448

4.788 x 101

6.895 x lo36.895 x 1O-3

1.356

1,602 x 10

1.571 x lo2

To obtain

m2

m3

m/s

m/s2

kg

N

Pa or N/m2Pa or N/m2N/mm2

J

kg/m3

N/m3

APPENDIX B-NOTATION

0 = distance between center of anchorsA = summation of projected areas of individual stress

cones, in.2

Ab = net bearing area of head of embedded anchorage, in.’

A, = tensile stress area, psiC’ = clear cover to bolt, in.

d/I = head diameter of headed stud or boltD = anchor diameter

DUJ = diameter of anchorage device such as embedded washer,in.

E, = elastic modulus of concrete, psif, = compressive strength of concrete measured by cylinders

psi or N/mm2fee = compressive slrength of concrete measured by cubes, psi

or N/mm2

fY = yield stress of anchor or bolt, psi

F,, = ultimate strength or capacity, lb or NF,, = ultimate tensile stress of steel, psih = member thickness

1, = embedment depth of anchor1?1 = distance from anchor centerline to free unsupported edge

n = number of anchorsN = factor which takes into account steel shear strength, usu-

ally 0.6 to 0.7

pnc = bolt tension load at which concrete failure will occurP, = applied factored tension loadS = spreading force, as from expansion sleeves of an expan-

sion anchor, also anchor spacingT, = applied tension loadT,,, = allowable anchor tensile loadT,, = ultimate wedge-splitting capacity of a singlt bolt

r, = ultimate tensile load, also design tensile loadV, = applied shear loadV, = nominal shear strength of anchor as governed by steel

failure

Yl = shear strength, ultimate shear load, or applied factoredshear load

W = crack width usually measured at the concrete surface(Y = included angle of concrete spa11 cone measured from the

axis of the anchor to the failure cone surfacep = coefficient of friction4 = strength reduction factor

x = chi factor which represents a partial influencing factorsuch as a load capacity reduction based on anchorspacing interaction (x,), edge distance influence (x,),etc.

This report was submitted to letter ballot of the committeewas approved according to Institute procedures.

State-Of-The-Art Report on Anchorage to Concrete

Documents

tennessee

betonwerk

uniform building

continuously

standard test

net bearing

building code

rockconcrete