Behavior of Coupling Beams Under Load Reversals - Concrete

RESEARCH AND DEVELOPMENT BULLETIN RD066.01B

Behavior of Coupling Beams Under Load Reversals

by G. B. Barney, K. N. Shiu, B. G. Rabbat, A. E. Fiorato, H. G. Russell, and W. G. Corley

PORTLAND CEMENT ASSOCIATION Research and Development / Consiruclion Technology Laboratories

Behavior ofUnder Load

Coupling BeamsReversals

by G. B. Barney, K. N. Shiu, B. G. Rabbat,A. E. Fiorato, H. G. Russell, and W. G. Corley’

HIGHLIGHTS

Eight model reinforced concrete coup-ling beam specimens were subjected toreversing loads representing those thatwould occur in beams of coupled struc-tural walls during a severe earthquake.Effects of selected variables on hystere-tic response were determined. Con-trolled variables included shear span-to-effective-depth ratio of the beams,reinforcement details, and size of theconfined concrete core. Load versus de-flection, strength, energy dissipation,and ductility characteristics were thebasic parameters used to evaluate per-formance.

Beams had a shear span-to-effective-depth ratio of either 1.4 or 2.8, corres-ponding to span-to-total-depth ratiosof 2.5 or 5.0, respectively. Tests indi-cated that hysteretic performance ofbeams with conventional reinforcementis limited by deterioration that results insliding in the hinging region. Full-lengthdiagonal reinforcement significantlyimproved the performance of shortbeams. The improvement for long-spanbeams was less significant. Larger con-crete core size improved load retentioncapacity.

INTRODUCTION

The Portland Cement Association isconducting a program to develop designcriteria for reinforced concrete struc-tural walls used as lateral bracing inearthquake-resistant buildings. Theprogram consists of analytical and ex-perimental investigations of isolatedwalls, coupled walls, and frame-wallsystems. Of primary concern are thestrength, energy dissipation capacity,and ductility of walls and wall systems.

As part of the experimental program,tests were conducted to evaluate the be-havior of reinforced concrete couplingbeams under inelastic load reversals.These tests were made to develop infor-mation for use in the investigation ofcoupled walls.

Coupling beams are used to join ad-jacent structural walls. Therefore, un-derstanding hysteretic response ofcoupling beams is a prerequisite tounderstanding coupled wall response.This report describes the results of testson eight reinforced concrete couplingbeam specimens subjected to static re-versing lohds.

Previous Investigations

Investigations by other researchers pro-vided the background information forthe test series.

Brown and Jirsa (‘)**testeddoubly re-inforced cantilever beams. These testsindicated that under inelastic load re-versals, intersecting cracks formed ver-tical slip planes through the beams. For-mation of these planes led to an eventualbreakdown in shear transfer as loadingprogressed. The breakdown intensifiedas residual tensile strains developed inthe longitudinal reinforcement. De-creasing stirrup spacing improved hys-teretic response, but did not eliminatesliding shear as the limiting condition.The beams tested had a shear span-to-effective-depth ratio of either 6.0 or 3.0.Maximum nominal shear stresses

/ran ed from 2fi psi to 7@ psi (O.17

f: MPa to 0.58@ MPa).Tests by Paulay and Binney(2)on deep

coupling beams with conventional rein-forcement also resulted in sliding shearfailures. To prevent sliding shear,Paulay and Binney used full-lengthdiagonal reinforcement. This arrange-

ment altered the load transfer mecha-nism to that of a Mesnager hinge. Forsimilar load histories, diagonally rein-forced beams sustained their loadcapacity over a greater number of loadcycles and dissipated more energy thanconventionally reinforced beams. Thebeams each had a shear span-to-effec-tive-depth ratio of approximately 0.9.Maximum nominal shear stresses

/ran ed from 9fipsi to 14fipsi (0.75

f: MPa to 1.16@ MPa).Bertero and Popov(~) investigated

other arrangements of special reinforce-ment using cantilever beams with ashear span-to-effective-depth ratio of3.1. Maximum loads on the beams wereequivalent to nominal shear stresses ofapproximately 6fi psi (0.5fi MPa).

In addition to closely spaced confine-ment hoops with supplementary cross-ties, Bertero and Popov tested beamswith supplementary diagonal reinforce-ment within the hinging region. Theirresults showed that the ability of thebeams to maintain load and dissipateenergy was significantly improved byreducing tie spacing. They also deter-mined that diagonal web reinforcementin combination with vertical ties mini-mized stiffness loss and stabilized hys-teretic response under increasing inelas-tic loading cycles.

Wight and Sozen ‘4)tested a series ofcolumns under large deflection rever-sals. The specimens were tested as canti-levers with and without axial compres-sive forces. The shear span-to-effective-depth ratio for the tests was 3.5. Maxi-mum nominal shear ranged from 4fito 6&psi (0.33 @to 0.5@ MPa).These tests verified that a progressivedecrease in strength and stiffness occurswith cycling in the inelastic range. Thehysteretic response improved whentransverse reinforcement was used toconfine the concrete core and carry thetotal shear. However, providing trans-

*Respectively, Manager, Building Design Sec-tion; Structural Engineer and Senior StructuralEngineer, Structural Development Department;Manager, Construction Methods Section; Direc-tor, Structural Development Department; andDivisional Director, Engineering DevelopmentDivision, Portland Cement Association, Skokie,Illinois.

* *.SuPer~criPt numbers in parentheses designate

references at the end of this bulletin.

@portland Cement Association 1980

2 Behavior of Coupling Beams Under Load Reversals

TABLE 1. Test Prof )m Variables verse reinforcement to carry the entireshear did not eliminate the possibility ofshear failure with large load reversals.Because the concrete core must remainintact to carry the shear, effective con-finement of the core is essential.

Although the tests described in thisreport are an extension of previous in-vestigations, they were planned as partof an overall program on earthquakeresistance of structural walls. There-fore, the design of the specimens incor-porated details for a future test programon coupled-wall systems.

-Core width

(i:.)

2.63

2.63

2.63

3.50

3.50

3.50

3.50

3.50

Span l$ngth

(in.)

16.67

16,67

16.67

16.67

16,6?

16.67.

33.33

33.33

Specimen

cl

C2

C3

C4

C5

C6

C7

, C6

Primary reinforcement

x x

Objectives and Scope

The objectives of this investigation were1. To provide information for select-

ing details of coupling beams foruse in tests of structural wall sys-tems.

~ ,1 ~, ~ ~,25,, 2. To determine strengths of couplingbeams subjected to reversing loads.

I t+

3. To determine load-versusdeflec-tion characteristics of couplingbeams with various reinforcingdetails.

4. To determine ductility and energy-dissipation capacities of couplingbeams subjected to reversing loads.

Eight specimens were tested. Theyrepresented approximately one-third-scale models although no specific proto-types were considered, Specimens weresubjected to in-plane reversing loadssimulating those in beams of coupledstructural walls.

Controlled variables in the programwere the type and arrangement of pri-mary reinforcement, span-todepthratio, and size of confined concretecore. Details of specimens are listed inTable 1.This report includes a description ofthe experimental program and the ob-served response of the test specimens.Effects of controlled variables are ana-lyzed. Strength, load-versusdeflectionrelationships, energy dissipation, andductility were the basic parameters usedto evaluate performance.

L---L-J 1 in. = 25.4 mm

Lateral P.

LoadI ----

-.

I. .

~ ,“-1-~ . CouplingBeam,4 A-J.?---- “ ?

l“--1~--. ,f)---

-’l//~.-a?In

/,

COupling

Beom41’ -

Structuralwall

~i —

~ (b) Internal forces.(a) Deformed wall system.

Abutment

hp-~

Well

-——.

I

CouplingeeamSpecimen

Deflected Wall

7

w

\ Clear Span

\

\ OUTLINE OF EXPERIMENTALPROGRAM

This section gives a brief description ofthe test specimens and test procedure. Amore detailed description can be foundin Appendix A.

(c) End conditions of coupling beam specimen. (d) Deformed beam.

Fig. 1. Idealized coudlna beam deformations and forces..-

PCA Research and Development Bulletin 3

Test Specimens

The configuration of the test specimenswas selected so that applied loads anddeformations represented those onbeams in coupled walls subjected tolateral forces. Beam deformations andforces were idealized as shown in Fig. 1.The test specimen is shown in Fig. 2.Each specimen consisted of two coup-ling beams framing into abutment wallsat each end, End conditions imposed bythe abutments represented those at thebeam-wall intersection of coupled walls.Axial forces in coupling beams were notconsidered in this investigation.

The beams had rectangular cross sec-tions 4 in. (102 mm) wide ‘and 6.67 in.(169 mm) deep. The effective depth ofthe main longitudinal reinforcementwas 6.1 in. (155 mm). Beam lengths were16.67 in. (423 mm) or 33.33 in. (847mm). These corresponded to shearspan-to-effectivedepth ratios of 1.4and2.8, respectively. The L-shaped abut-ment walls were 4 in. (102 mm) thick.Overall specimen dimensions are givenin Fig. 3.

Specimen Design

The short coupling beams weredesignedto carry maximum forces correspond-ing to nominal shear stresses of approx-imately 9&psi (0.75& MPa). Fordesign purposes, concrete strength wastaken as 3000 psi (20.7 MPa) and steelyield strength as 60 ksi (414 MPa), Steelstrain hardening of 50% was assumed.Primary reinforcement was selected todevelop capacities corresponding tothe desired maximum shear stress levels.To avoid anchorage failures, develop-ment lengths were 50% greater thanthose required by the 1971 ACI Build-ing Code.’5’

Transverse-hoop reinforcement wasprovided in all specimens to resist shearand to confine the concrete core. Thehoops consisted of D-3 deformed wirespaced 1.33 in. (34 mm) on centers. De-sign of the reinforcement was such thatstresses in the hoops were below yield atforces corresponding to the maximumcapacity of the beams, Shear stresseswere calculated using Equation 11-3 ofthe 1971 ACI Building Code with thecapacity reduction factor, +, equal to1.0. Shear reinforcement was propor-tioned according to Equation 11-13 ofthe 1971 ACI Building Code. Nominal

/“Abutmentwall

fRoller

Guide Coupllng Beoms

P---- /

L (

7$?7‘Abutment

wallFig. 2. Test specimen.

PinnedEnd

v“”””)m

m

k ‘:t--11’-9”

(o.53ni

,:67,, /%:’\

\

6’-0.67” (0.42m)

(1.85m)or

33.33”or

4’-1”[0.85ml u,!

(R%) ([’4 m) --IL6.67 “ !2 ‘- 4“ (0,17m) Abutment

[0.71m) Walls

Fig. 3. Specimen geometry.

concrete shear stress capacity was ne-glected. The hoops also met the require-ments for transverse reinforcement inflexural members of special ductileframes in Section A.5 of the i971 ACIBuilding Code.

Specimens C2, C5, and C7. SpecimensC2, C5, and C7 had straight longitudi-nal reinforcement, consisting of four6-mmdiameter (%-in.) bars top andbottom. Photographs of the reinforce-ment are shown in Fig. 4(a) and (b). Alarger confined concrete core was pro-vided for Specimens C5 and C7. Shearspan-to-effectivedepth ratio of Speci-mens C2 and C5 was 1.4. For SpecimenC7, the ratio was 2.8.

Specimens Cl, C3, and C4. SpecimensCl, C3, and C4 had diagonal bars in thehinging regions. This reinforcement wasprovided to reduce shear deterioration

as suggested in tests by Bertero andPopov. ‘3)Design oft hese bars was basedon the assumption that they would be-have as diagonal truss members afterthe concrete deteriorated under repeat-ed load reversals, They were designedto carry the maximum shear force with-out yielding. For Specimen C 1, two 6-mmdiameter (ti-in.) bars from bothtop and bottom were bent at 45° start-ing at the face of the wall at each end ofthe beam as shown in Fig. 4(c). Speci-mens C3 and C4 contained two addi-tional 6-mm (%in.) bars top and bot-tom bent at 45°. A photograph of thisreinforcement for Specimen C4 isshown in Fig. 4(d). Additional rein-forcement details are presented in theAppendix.Specimens C6 and C8. Specimens C6and C8 had full-length diagonal rein-forcement. This detail was used by


(a) Specimen C2 (b) Specimen C7 (c) Specimen Cl

(d) Specimen C4 (e) Specimen C6 (f) Specimen C8

Fig. 4. Reinforcement for test specimens.

Paulay and Binney(2)for coupling beamswith shear span-to-effective-depth ra-tios less than 0.9, Specimens C6 and C8had shear span-to-effective-depth ratiosof 1.4 and 2.8, respectively.

Full-1ength diagonals for short beamspecimens were designed to carry themaximum shear force corresponding toa nominal stress of 9fi psi (0.76fiMPa). The area of diagonal rein-forcement, A, was calculated as

/f,= ~2f, sm a

(1)

where V“= maximum shear forcef, = stress in diagontil rein I

forcement (90 ksi; 621MPa)

a = angle between diagonalbar and horizontal

Using this approach, a single No. 4 barwas provided in one direction and twoNo. 3 bars were used in the oppositedirection.

Transverse hoops were provided tocontain concrete in the core during re-versals and to prevent buckling of thediagonal bars. Longitudinal bars sup-porting transverse hoops were notanchored in the abutment walls.

The reinforcement details used forSpecimens C6 and C8 are shown in Fig.4(e) and (/).

Test Procedure

The test setup is shown in Fig. 5. Speci-mens were placed parallel to the labora-tory floor and supported on thrust bear-ings. Loads were applied by hydraulicrams at one end and resisted by a fixedsupport at the opposite end. The line ofaction of forces passed through the mid-length of the coupling beams,

Loading was controlled by the mag-nitude of applied force prior to yieldingand by imposed deflections after yield-ing. For each increment of applied loador deflection, three completely reversed

load cycles were applied. This is illus-trated in Fig. 6. Deflections in succes-sive increments were increased until thespecimen was destroyed.

Instrumentation was provided tomeasure applied loads, deflections,beam elongations, and reinforcing steelstrains.

GENERAL RESPONSECHARACTERISTICS

In evaluating the test results, appliedload was assumed to divide equally be-tween the two beams in each specimen.This assumption was checked by com-paring companion measurements fromeach beam and by visual observation ofthe beams during testing. Except forSpecimen C3, the performance of thetwo beams in each specimen was simi-lar. In Specimen C3, one of the beamsdeteriorated more rapidly than theother after the maximum load was

Rea0101

/ctionck

PCA

Ram Lood Cells Load Ce I Is

Fig. 5. Test setup.

Fig. 6. Load or deflection history.

reached. Test results for this specimenmust be considered accordingly.

Conventional LongitudinalReinforcement

Specimens C2, C5, and C7 had conven-tional longitudinal reinforcement. Spec-imens C2 and C5 had short-span beamsand core widths of 2.63 and 3.50 in. (67and 89 mm), respectively. Specimen C7had long-span beams and a core widthof 3.50 in. (89 mm). Load-versus-deflec-tion relationships for the specimens are

\ReactionOlock

shown in Figs. 7, 8, and 9. Plotted loadsare those fo~each beam. Deflections arethe relative displacements between endsof the beams.

Performance of the beams with con-ventional straight longitudinal rein-forcement was limited by deteriorationof the shear-resisting mechanism in thehinging region. Under reversing loads,intersecting cracks developed across theentire depth of the beams at their ends.As subsequent inelastic load reversalswere applied, concrete at the ends was

Research and Development Bulletin 5

destroyed by cracking, abrasion, andspalling. With the concrete destroyed,shear transfer by truss action was notpossible and the transverse hoops be-came ineffective. Interface shear trans-fer also was lost, Eventually, dowel ac-tion of longitudinal reinforcement pro-vided the primary shear resistance. Thisloss of shear transfer is often termedsliding-shear behavior. Because slidingdeveloped along a plane parallel to thetransverse reinforcement, even theclosely spaced hoops became inefficientin transmitting shear. However, thehoops did provide confinement of theconcrete core.

Deterioration of the concrete at theends of the beams was intensified byelongation of the beams, caused by re-sidual tensile strains in the longitudinalreinforcement. These strains developedwith successive load reversals into theinelastic range. In addition, partial slipof reinforcement anchored in the abut-ment walls contributed to wideningcracks at the ends of the beams. Slipdeveloped as bond between concreteand reinforcement deteriorated underinelastic load reversals.

Degradation of the shear-transfermechanism and deterioration of theconfined concrete core were associatedwith “pinching” of the load-versus-deflection loops shown in Figs. 7,8, and9. The pinching occurred because, asloads were reversed, slip along the inter-face took place with little increase inload. Eventually, concrete surfaces oneither side of the interface were broughtinto contact with each other and theload resistance increased. In addition,shear transfer by dowel action of thelongitudinal reinforcement increased.As the number of load cycles increased,continued abrasion and crushing ofconcrete in the critical region resultedin a complete breakdown of the shear-transfer mechanism at the interface.

Both short- and long-span couplingbeams exhibited pinching. However, itwas somewhat less severe for the long-span beams.

Diagonal Reinforcementin Hinging Regions

To improve the performance of short-span beams, diagonal reinforcementwas provided in the hinging region ofseveral specimens. This reinforcement


load, k(os 1

- 2r

Fig. 7. Load-versus-deflection relationship for Specimen C2.

I-20

Load, k,ps

--70

-70


Load, k,ps

0


was patterned after details tested byBertero and Popov. ‘3’However, modi-fications were made to simplify fabrica-tion.

Specimens C 1, C3, and C4 had diag-onal reinforcement in the hinging re-gions as indicated in Fig. 4 and Table 1.Specimens C 1 and C3, with the smallercore size, had single and double diag-onal bars, respectively. Specimen C4,with the larger core size, had doublediagonal bars.

The diagonal reinforcement was in-tended to eliminate the sliding shearthat limited inelastic response of thebeams with conventional horizontal re-inforcement. Diagonal reinforcementwas designed to provide an internaltruss system to resist shear and to movethe critical plastic hinge location awayfrom the face of the wall.

Use of diagonal reinforcement in thehinging regions did not result in theanticipated improvement in perform-ance. The unsatisfactory performancewas caused by factors discussed below.

The hysteretic response of SpecimensC 1, C3, and C4 is illustrated in the load-versusdeflection relationships shownin Figs. 10, 11, and 12, respectively.Comparison of these figures with thoseshown previously for short-span beamswith conventional reinforcement indi-cates that the special diagonals did notsignificantly improve the hysteretic re-sponse characteristics. The few im-provements in energy dissipation andload retention capacities that were at-tained do not seem to warrant the addedcomplexity and cost of the diagonals.

The primary reason that the diago-nals within the hinging region did notperform satisfactorily lies in the detailsfor tying the truss together. Initially, thebeams behaved as expected. Tensileyielding of the flexural reinforcementat the face of the wall and at the inter-section of the diagonals occurred almostsimultaneously. As loading cycles pro-gressed into the inelastic range, concretewithin the region of the diagonals dete-riorated by spalling and crushing.

In principle, loss of concrete shouldnot have affected the ability of the diag-onal trusses to resist load as long as thediagonal bars did not buckle. However,as loading progressed and concrete waslost, the support points for the diago-nals loosened, allowing the corners ofthe diagonal bars to become displaced.


Load, k,im

-Ir

Fig. 10. Load-versus-deflection relationship for Specimen Cl.

7

37


,,

4

4

Fig. 12. Load-versus-deflection relationshipfor Specimen C4. Fig.

(See Fig. 13.) Once the supports for thediagonal truss softened, efficient trussaction could not be developed. There-fore, the diagonals were not effective incarrying the applied shear.

It should be noted that the transversehoops located at the corners of the diag-onal were adequate for resisting the out-ward thrust of the diagonal bars. Strainmeasurements indicated that thesehoops did not yield. The problem en-countered was that the force in the diag-onal was not effectively transmitted tothe transverse hoop.

The addition of a second set of diago-nals in Specimens C3 and C4 was at-tempted in order to lower the forcelevels in the diagonal bars. However,the result was not satisfactory, It is ap-parent that in using this type of diagonalreinforcement, extreme caution must beexercised in securing the bars. The ef-fectiveness and cost-to-benefit ratio ofthis detail are questionable.

One method of improving this detailwould be to eliminate the bend in thediagonals at the intersection of the beamwith the wall. However, this wouldcause additional problems in fabrica-tion of beam and wall reinforcement.

Concrete Spalled Off,

(a) General Location

@.@din Tension (b) Detail I In Compression

13. Loss of support for diagonals in hinging region.


Load, k,ps length diagonals was similar to that ob-12T ,2*” ,4 served in the beams with conventional

lm=254mm

lk, o=4448kN

-3 -2

Em

\2 3

De election, in.


Load, k,ps

t1+


Full-Length DiagonalReinforcement

Because of the ineffectiveness of diago-nals located in the hinging regions andthe sliding shear limitation with conven-tional reinforcement, Specimens C6and C8 were tested with full-lengthdiagonal reinforcement. The beams hadshear span-to-effectivedepth ratios of1,4and 2.8, respectively. Reinforcementdetails are given in Fig. 4 and Table 1.The straight longitudinal bars support-ing the transverse hoops were not an-chored in the abutment walls and were

not considered to contribute to the load-resisting mechanism.

Load-versus-deflection relationshipsfor Specimens C6 and C8 are shown inFigs. 14 and 15. Full-length diagonalseffectively increased the load-retentionand energydissipation capacities ofthese coupling beams. The load-versus-deflection curves do not exhibit thepinching that results from deteriorationof the shear-resisting mechanism. Also,the sliding-shear mechanism leading toloss of stiffness and strength was notobserved,

Initial cracking in the beams with full-

reinforcement. As inelastic load cycleswere applied, concrete at the beam-wallinterface crushed and spalled. This dete-rioration, however, was not reflected inthe load-versusdeflection relationships.Forces on the beams were effectively re-sisted by truss action of the diagonalbars. The transverse hoops containedthe concrete core over most of the beam,thus preventing the diagonal bars frombuckling.

Performance of test beams with full-length diagonals was eventually limitedby inelastic buckling and subsequentfracture of the diagonal bars. At laterstages of the test, concrete spalling atthe beam-wall intersection ex~osed thediagonal bars within the abutment wall.In this region no reinforcement was pro-vided to prevent the diagonals frombuckling.

The mechanism that develops withuse of full-length diagonals is essentiallythat of a Mesnager hinge. It was expect-ed, based on Paulay and Binney’s tests, [z’that significant improvement in re-sponse would be observed in the short-span beams. The shear span-to-effec-tive-depth ratio of these beams was 1.4as compared to a maximum ratio of 0.9in the beams studied by Paulay and Bin-ney. ‘2)The use of full-length diagonalsin longer-span beams had not been test-ed previously. Thereforej Specimen C8was tested with a shear span-to-effec-tive-depth ratio of 2.8

Based on the behavior of SpecimensC7 and C8, the improvement obtainedusing full-length diagonals in beamswith a shear span-to-effective-depthratio of 2.8 (span-to-depth ratio of 5.0)was relatively small. In addition, gravityloads within the span take on greatersignificance for longer-span beams.These loads cannot be resisted effici-ently by diagonal reinforcement. Con-sidering these findings, full-length diag-onal bars do not appear to be justifiedfor coupling beams with shear span-to-effective-depth ratios of 2.8 or greater.

STRENGTH CHARACTERISTICS

Table 2 is a summary of yield and maxi-mum loads for the test specimens. Ob-served and calculated values are given.They represent the load on each beam


TABLE 2. SDecimen Strengths

Yield load M

Spe&:men Observed Calculated* Observed Observed

kips VT kips G Calc. kips v?

cl 8.1 6.2 9.0 6.9 0.90 9.2 7.0C2 10.2 7.6 9.8 7.3 1.04 10.3 7.7C3 10.2 7.7 9.8 7.4 1.04 11.8 9.0

C4 10.0 7.0 9.5 6.7 1.05 11.5 8.0C5 9.2 6.7 8.8 6.5 1.05 9.4 6.6

C6 7.6 6.3 7.4 6.0 1.05 13.4 10.9

C7 4.3 2.9 4.4 3.0 0.98 5.2 3.5C6 4.3 3.0 4.0 2.8 1.08 7.5 5.3

1 kip = 4.448 kN; 1 fi(psi) = 0.0830@_(MPa)

“Based on monotonic flexural response.

ximum load

Calculated’ I Observed

ki~s

10.812.812.411.711.414.4

5.68.7

Vr6.29.69.48.28.4

11.63.86.1

Calc.

0.850.800.950.960.620.930.930.86

based on the assumption that the totalload was distributed equally. The areaused to calculate the nominal shearstress was the width of the beam timesits effective depth.

Observed Strengths

Yield loads given in Table 2 are thoseobserved when the main flexural rein-forcement first reached its yield strain.For most specimens, the main flexuralreinforcement was the straight longitu-dinal bars. For Specimens C6 and C8,which had full-length diagonals, yieldload is that observed when the diagonalbars reached their yield strain.

Maximum observed loads corre-spond to the maximum value measuredin either direction of loading. The lastcolumn in Table 2 gives the ratio of ob-served maximum load to the observedyield load for each specimen.

Several observations can be made re-garding the strengths of the test speci-mens. For the short-span beams withconventional longitudinal reinforce-ment, Specimens C2 and C5, maximumload was only 1% higher than the yieldload. This is indicative of the sliding-shear deterioration that occurred in thecycles subsequent to yield.

Beams in Specimen C7, which hadlong spans with conventional reinforce-ment, reached a maximum load 21Yogreater than yield load.

Specimens C 1,C3, and C4, which hadshort-span beams with diagonal rein-forcement in the hinging regions,reached maximum loads 13’%0to 1770

greater than yield loads.Specimens C6 and C8, with full-

length diagonals, had ultimate strengths73% to 77% greater than their yieldstrengths. This indicates that the diag-onal bars developed strain hardeningefficiently. In designing coupled-wallsystems, differences between ultimatestrength and yield strength must berecognized because forces developed inthe beams influence the performance ofthe walls.

As indicated earlier, the hoop rein-forcement was designed so that it wouldnot yield at loads corresponding to thecapacit of the beams. Strain measure-

Yments’b confirmed that the hoops didnot yield for either conventionally rein-forced beams or those with special rein-forcement. It is apparent that transversehoops cannot prevent sliding shear inshort, conventionally reinforced beamswhen repeated inelastic load reversalsare applied. However, this is dependenton load history, The laboratory loadingon the beams was intentionally severe,

Calculated Strengths

Calculated values of yield load andmaximum load, shown in Table 2, werebased on measured material properties,The calculated values are estimates ofmonotonic flexural strength. They donot account for effects of load reversals.Loads for Specimens C6 and C8, withfull-length diagonals, were estimatedusing Equation (l).

Calculated flexural yield values are inbasic agreement with observed values.

Observed maximum

Observed yield

1.131.011.171.141.011.731.211.77

Observed maximum loads range from80% to 98% of calculated loads. Besidesthe normal variations expected in suchcalculations, the differences reflect theeffects of load reversals and shear dete-rioration.

Nominal shear stresses, given in Table2, were based on the effective depth ofthe section and the overall beam width.Observations of the beams during test-ing indicated that the concrete shell sur-rounding the confined core broke awayduring the inelastic load cycles. At thisstage, nominal shear stresses based onthe area of the confined core are morerealistic. For Specimens C 1,C2, and C3,the core area was 66910of the nominalarea. Thus, nominal shear stresses basedon the area of the confined core are 1.5times those in Table 2, Based on the corearea, maximum nominal stresses ob-served in Cl C2, and C3 were 10.7fipsi, 11.7& psi, and 13.7& psi(0.89@ MPa, 0.97@ MPa, and1.14~ MPa), respectively.

For all other specimens, the core areawas SS70of the nominal area. The corre-sponding maximum nominal shearstress for short-span beams ranged from7.8& to 12.5@_ psi (0.65fi to1.04fi MPa). For long-span beams,the range was from 4.Oflto 6.1@psi(0,33~ to 0.52@ MPa).

DEFORMATIONCHARACTERISTICS

Because of potential excursions into theinelastic range of response, deformation

10 Behavior of Coupling Beams Under Load ReversaIs

Load , k I PS

, 1

-2.0 -1.5 -1,0 -0.5 0.5 1.0 15~5—.

.L

2.0

-2 J, Defleclior , in.

1 in. = 25.4 mm1 kip = 4,448 kN

, /’//

_12 -LC6 \

.— -----

Fig. 16. Load-versus-deflection envelopes for specimenswith short-span beams.

Lood, kips

12

10[

8

I/f

C8_ - _--—_____-.---

,---------6 /’

//

4

2

{) 4

-2.0 -1,5 -1,0 -0.5 / 0,5 I .0 I .5 2.0

AI -/

Deflection, in

_-- —_ --- —-—----

C8-6- -

-8

I1 in. = 25.4 mm1 kip = 4,448 khl

-lo

-12

Fig. 17. Load-versus-deflection envelopes for specimenswith long-span beams.

characteristics of coupling beams are of Load-Versus-Deflection

considerable interest, For the beams Envelopes

tested, these characteristics have been Figs, 16 and 17 show envelopes of thequantified in terms of overall-load-ver- load-versusdeflection relationships forsusdeflection relationships, energy dis- the short-span and long-span beams,sipation, and ductility. respectively.

Except for Specimen C6, beams withshear span-to-effectivedepth ratios of1.4 reached their maximum load capac-ities at deflections between 0.4 and 0.5in. (10 and 13 mm). These deflectionscorresponded to overall rotations of0.02 and 0.03 radian. Overall rotationswere calculated as deflection divided bylength of beam.

Specimen C7, which had convention-al longitudinal reinforcement and ashear span-to-effective-depth ratio of2.8, reached its maximum load at a de-flection of 0.9 in. (23 mm). This corre-sponded to an overall rotation of 0.03radian.

Specimens C6 and C8, with full-length diagonals, reached maximumload capacities at deflections of 0.8 in.(20 mm) and 1,5in. (38 mm), respective-ly. Corresponding rotations were 0.05radian for both the short-span and long-span specimens. Although beams withfull-length diagonals reached higherrotations at maximum load than didother beams, their final load loss wasmore sudden. This is because their loadcapacity was limited by bar fracture,

It is evident in Figs. 16and 17that theenvelopes for Specimens C6 and C8 arenot symmetrical for opposite directionsof loading. This is because the areas ofthe diagonal reinforcement were notequal. Two No. 3 bars were used in onediagonal direction and one No. 4 barwas used in the other direction. Theareas for these bars were 0.22 and 0.20sq in. (142 and 129 mm2), respectively.

Measured load-versusdeflection en-velopes are compared with calculatedenvelopes in Figs.. 18and 19. The calcu-lated curves represent first-order ap-proximations for monotonically loadedbeams. Only flexural deformations wereconsidered. ‘b)Calculated curves are pre-sented to illustrate that the beam defor-mations are made up of a number ofcomponents, not all of which are quan-tifiable.

Measured deflections are consider-ably in excess of calculated values, par-ticularly for the short-span beams. Thefollowing factors, not included in thecalculations, could make a significantcontribution to the deflection.

1. Shearing distortions2. Slip of main reinforcement an-

chored in abutment walls3. Load reversals

In addition, there are normal limita-


Load, kips ,2

h

,/ / Calculated

,’

8

Measured

4

I I

-2,0 -1.5 -1.0 -0.5 0.5 1,0 1.5 2.0

w :

Deflection, in.

-4

-:

)/

~12-

(a) Specimen C2

Load, kips ,2

h

, Calculated

(’

e;Measured

4

,-20 -1.5 -1.0 -0.5 0.5 1.0 1.5

xl ~ ‘o

Deflection, in.

,/’

+.1

(c) Specimen Cl

Lood, kips

-2:0 -

,2 ;Calculated

D

i

8’ Measured

4

Lood, kips ,2

/ Calculated,/’I

e I

4 .

I 1

-2,0 -1,5 0.5 1,0 1,5 21)

Deflection, in.

,/’

+1

(b) Specimen C5

Load, hips

II

8. j

I

4.

b I-2.0 -1.5 -Lo -0.5 0.5 1.0 1.5 20

L I

Deflection, in,

-4

/

I

-w’

(d) Specimen C3

Loadt kips , Calculated12 :

r

(

8Measured

4

t I I

-2,0 -L5 -1.0 -0.5 I 0,5 1.0 1.5 2.0

1Deflection, in.

?-e

Y

1 in. = 25.4 mm

-12’1 kip = 4.448 kN

/’

(e) Specimen C4 (f) Specimen C6

Fia. 18. Calculated and measured Ioad-vereus-deflectionenielopes for short-span beama.

tions on the accuracy of deflection cal-culations.

The influence of shearing distortionscan be seen in Figs. 18and 19.Calculat-ed deflections for Specimen C7 are in

better agreement with measured valuesthan are those for Specimen C5. Itwould be expected that shearing dis-tortions would have less influence onSpecimen C7 with more slender beams.

For the short beams with conven-tional reinforcement, the contributionof shearing distortions was estimatedusing the procedure developed by Bach-mann.(’) The modified calculations’b’


Lood, hipsLoad, hips

8,8- CalcyloJe~

-----

I

6 .Calculated 6. .;

---/~-

;4. .1 4 .!

1’Measured

I

2- J

I I ~ I

-20 -1.0 I .0 20 -1.0 Lo 2.0.

De flecflon, in Deflection, in.

--”--6

I in =25.4mm--,-1 ..-6

. ..-

[klp=4448k N-+ ---

--8 --8

(a) Specimen C7 (b) Specimen C8

Fig. 19. Calculated and measured load-versus-deflection envelopes for long-span beams.

I In = 25.4mm

lkfp=4446k N

are a better approximation of the enve-lope, as shown in Figs. 20and 21. Again,load reversals were not taken into ac-count.

To be more than just a relative meas-ure of performance, the calculation ofload-versusdeflection relationships forbeams under load reversals requires ex-tensive refinement. Such refinement isoutside the scope of this project.

Ductiiity

The ductility of a structure is commonlyused as a measure of its inelastic per-formance. Deflection ductility ratio is

defined in this report as the ratio of thedeflection of the beams to the deflectionmeasured at yield. Cumulative deflec-tion ductility ratio is the summation ofductility ratios for each cycle of loading,(See Fig. 22.) Ductility ratios for posi-tive and negative loadings are addedseparately.

Comparisons of the performance ofspecimens based on cumulative ductilityare shown in Fig. 23, where the load isplotted as a percentage of the maximum,observed in either loading direction.

Specimens C2 and C5 with short-span beams and conventional reinforce-ment were tested with confined corearea equal to 66910and 88910respectively,

Load, klw

‘2 O><flC”~ModifiOd c.alcul.ath

8

k

Mnasured4

I I 4-1.5 -1.0 -0,5 0,5 Lo 1.5

‘dDeflection, In.

-4

-8

/,2 Ilm. 25,4mm/“

/’ /’ Iklp . 4448kN/,/. -12

Fig. 20. Influence of diagonal cracking on the calculationof defection for Specimen C2.

of the effective section area. As shownin Fig. 23(a), the larger confined corearea improved behavior. Under loadreversals, the concrete shell fell off.Thus, nominal shear stresses on thesmaller core were larger and lowercumulative ductility was attained.

The effects of different reinforcementarrangements on ductility are shown inFigs. 23(b) and 23(c) for the short-spanand long-span beams, respectively. Forthe short-span beams, the specimenwith full-length diagonals maintainedits load capacity for a significantly larg-er ductility than the other specimens.The difference was not as significant forthe long-span beams.

Load, kips

12

II

Calculated

----1

8 j4<-

Modlf led Calculation

4Measured

-1:5 -1:0 -0:5

1

0:5 I ,0 I:5

Deflactlon, In.

—---5+ I in.. 25.4mm—-—

I klp=4.448kN

Fig. 21. influence of diagonal cracking on the calculation ,of deflection for Specimen C5.


I II A

II

Deflection

Ai.—

AY

Cum, Deflection Ductility Ratio = #l pi

Percentage of Max Load

I —+

-400 -300 -200 -1oo 0 100 200 300 400

Cumulative Ductility

I (a) Effect of core size for short-span beams.

Fig. 22. Cumulative defection ductiiity ratio.

II Percentage of Max, Load

Energy-Dissipatlon Capacity

Energy dissipation provides a measureof the inelastic performance of a struc-ture under load reversals. The dissipat-ed energy is the difference between thatexpended during load application andthat recovered during unloading. Ener-gy dissipation is thus defined as the areaenclosed by the Ioad-versusdeflectionloops. (See Fig. 24.) Using the energydissipation derived from the measuredload-versusdeflection relationships ofthe test specimens provides a convenientreference for quantifying the specimens’performance. Essentially, for equivalentloads and deflections, specimens dissi-pating the most energy would be con-sidered to have the best performance.

Figs. 25 and 26 illustrate the “stabil-ity” of the load-versus-deflection loopsfor the specimens. In the loading se-quence, three complete cycles of load or

! deflection were applied at each load de-flection increment. Figs. 25 and 26 show

1the cumulative energy dissipated for thefirst, second, and third cycles of eachincrement. Energy dissipation is plottedagainst displacement ductility ratio.Displacement ductility ratio is definedas the deflection at each increment di-vided by the yield deflection.

If the three loops within each incre-ment had identical areas for each cycle,it would mean that there was no loss ofenergydissipation capacity. As can beseen in Figs. 25~ and 26(b), this wasnearly the case for Specimens C6 and

,A

,/

/“-./., +.

‘“\/ \

t ,\\,/,A, ,/, \, ,,/ ,

/’./ ‘,, %

‘,, C4

U-J L C5

-400 -300 -200 -100 0 100 200 300 400

Cumulative Ductiltty

(b) Effect of reinforcement for short-span beams.

Percentage of Max. Lood

25

t

-400 -300 -200 -100 ‘o 100 200 300 400

Cumulative Ductlllty

(c) Effect of reinforcement for long-span beams.

Fig. 23. Load versus cumulative defection ductiiity.


4Load

r

Load Cycle i

Deflection

~ = Energy Dissipated

Fig. 24. Energy dissipation.

Cumulative Energy Diaslpotion, in-klps

160T

cycle I

p:,

cycle 250

Cycle 3

00 2 4 6 e 10

Ois P1. Ductility Ratio

(a) Specimen C2

Cumulative Energy Dir.slpallon, In-klw

150T

i

1oo-

DIsPI DUCtlbitY Ral10

(d) Specimen C3

Cumulative Energy 0is81patl on, in-klps

150-

100,

olaP1. Ouctillty Ra710

(b) Specimen C5

Cumulative Energy Oi$al potion, in-klps

150-

CycleI1oo-.

50

0.0 2 4 6 e 10

OISPI, Ouctlllty Ratio

(e) Specimen C4

Cumulative Energy Olssi pation, in-kips

150

100I

cycle I

50

0. ——.4o 2 4 6 8 10

Oispl. Ouclility Ratio

(c) Specimen Cl

C“m.lative Enerqy Dissipation, In-kips

150- Cycle I

1oo-

50.

1 in. - kip =

o-0 2 4 6 8

DIs PI. Ouctlllty Ratio

(f) Specimen C6

0.113kN. m

10

Fig.25. Energy dissipation versus displacement ductillty ratio for short-span beams.


Cumulative Energy Dissipation,, in-kips

250

I

Iln - klp= 0,113 kN, m

200

I

150

100

1

Cumulative Energy Dissipation, in-kips

250

I

I In- kip = 0.113 kN+m

3

Cycle I

200 Cycle 2

150- ,

Cycle I

Cycle 2

1oo- -

50. - 50- -

o~ t0 2 4 6 8 10 (2

0+0 2 4 6 8 10 12

Displacement Ductility Ratio Displacement Ductility Ratio

(a) Specimen C7 (b) Specimen C8

Fig. 26. Energy dissipation versus displacement ductility ratio for long-span beams.

C8, which had full-length diagonals.The wider differentials between succes-sive cycles within each load increment Normol!zed Cumulotlve Energy Disslpotion

for the other beams attested to thegreater deterioration of these beams.The curves in Figs. 25 and 26 illustratemore clearly what was indicated in theload-versus-deflection relationships 400

[~

Py, kips(k N) Af,ln. [mm)

c1 8. I (36.0)

presented earlier.0.16(4.1)

C2 10,2(45.4) 0.23(5.81While Figs. 25 and 26 show the per- C3 10.2(45.4) 0.15(3.8)

formance of each specimen, they do not C4 10.0(44$5) 0.14(3.6)

give the relationships between the vari-C5 9,2(40.9) 0.19[4.8)C6 7,8[34.7)

ous specimens. These relationships can0,11(2.8)

be shown by normalizing the cumula- 300tive energy dissipated by the product ofthe yield load and yield deflection, as ifollows:

I

‘&~,e/

Emn (2)yyt5 /

i

200.where E~n= normalized cumulative i C4

energy dissipated /’/“

Py = yield load/’ cl/’

AY = yield deflection/’

/ /’,’e, = energy dissipated in ith /“

/ /load cycle /“

i = number of load cycles 100 / ,/”/

Figs. 27 and 28 show the normalized /cumulative energy dissipated versusnumber of cycles for the short-span andlong-span beams, respectively. All load /cycles are included in the cumulative

/

energy determination. Since the speci- 0 A ,,, . . .

mens had similar load histories, these o 10 20 30 40

curves provide a comparative measureCycle Number

of the energy dissipation capacity of the Fig. 27. Normalized cumulative energy dissipation versustest specimens. cycle number for short-span baams.

16

400

3CCI

200

100

0.

Behavior of Coupling Reams Under Load Reversals

1iI

/’//’

C8 /

/’I

/’

I

0 “ 10 20 30 40 50 60

Cycle Number

Fig. 28. Normalized cumulative energy dissipation versusductility for long-span beams.

Energy -dissipation-versus-ductilityrelationships confirm the results previ-ously indicated. For short-span beams,Specimen C6 with full-length diagonalsprovided the greatest energy-dissipationcapacity. Special diagonals within thehinging regions improved energy-dissi-pation capacity somewhat, but the im-provement was not great enough to jus-tify their complexity and cost.

For the long-span beams, the full-length diagonals did not provide signif-icant additional energy-dissipation ca-pacity.

SUMMARY AND CONCLUSIONS

Eight model reinforced concrete coup-ling-beam specimens were subjected to

reversing loads representing those thatwould occur in beams of coupled struc-tural walls during a severe earthquake.The beams were constructed at approxi-mately one-third scale. Effects of se-lected variables on hysteretic responsewere determined. Controlled variablesincluded shear span-to-effectivedepthratio of the beams, reinforcement de-tails, and size of the confined concretecore. The variables are summarized inTable 1.

The beams had a shear span-to-effec-tivedepth ratio of either 1.4 or 2.8, cor-responding to a span-to-depth ratio of2.5 or 5,0. Maximum nominal shearstress on short-s an beams ranged from

/7fipsi (0.58 f{ MPa) for beams withconventional reinforcement to 11~psi (0.91 @MPa) for beams with full-Iength diagonal reinforcement. Equiva-

lent shear stresses for Ion -span beams/were 4@ psi and 5 fl psi (0.33fl

MPa and 0.42@ MPa).Load-versus-deflection relationship,

strength, energy dissipation, and ductil-ity were the basic parameters used toevaluate performance of the test spec-imens. The following conclusions arebased on the test results.

Conventional LongitudinalReinforcement

Inelastic response of coupling beamswith conventional reinforcement waslimited by sliding-shear deterioration atthe beam-wall intersection. This was thecase even though transverse hoops wereprovided to carry the entire shear with-out yielding. Since sliding-shear cracksdeveloped between transverse reinforce-ment, the hoops eventually became inef-fective. Development of sliding shear isdependent on load history. Deteriora-tion is a function of the cycle numberand the intensity of applied loads. Thus,any generalization of the results mustconsider load history. Both the numberof cycles and the load intensity used inthe laboratory tests can be consideredas representative of extremely severeearthquake conditions on the most criti-cally stressed beam in a coupled-wallsystem.

Specimen C5, which had short-spanbeams, sustained an overall rotation of0.025 radian. At this rotation incrementmore than 80~o of the maximum loadwas maintained for three complete cy-cles. Yield rotation for Specimen C5was 0.01 radian.

Specimen C7, which had long-spanbeams, sustained an overall rotation of0.04 radian. Yield rotation for this spec-imen was 0.005 radian,

Tests indicated that improved inelas-tic performance was obtained by in-creasing the size of the concrete core.The confined core of coupling beamsshould be made as large as possiblewithin the limits of cover requirements.

Diagonal ReinforcementIn Hinging Regions

Diagonal reinforcement within hingingregions at the ends of the beams im-proved performance, but not enough tojustify the added complexity and cost.

To use this type of reinforcement, ex-treme care must be exercised in theselection and construction of details,particularly at bend locations in the re-inforcement. Based on the laboratorytests, it does not appear that this detailwould be an economical solution.

Full-Length DiagonalReinforcement

Beams with full-lemzth diagonal rein-forcement had the b~st stren~th, ductil-ity, and energydissipation characteris-tics of any of those tested.

Specimen C6, which had short-spanbeams, sustained an overall rotation of0.05 radian. Yield rotation for this spec-imen was 0.01 radian.

Specimen C8, which had long-spanbeams, sustained an overall rotation of0.06 radian. Yield rotation was 0.01radian. Improvement in hysteretic re-sponse using full-length diagonals forlong-span beams was not as significantas for short-span beams. In addition,gravity loads within the span took ongreater significance for longer-spanbeams. These loads cannot be resistedefficiently by diagonal reinforcement.Thus, straight diagonal bars do not ap-pear to be justified for beams with ashear span-to-effective-depth ratio of2.8 or more. Tests using this type of re-inforcement have not been carried outon coupling beams with a shear span-to-effective-depth ratio between 1.4 and2.8.

If full-length diagonals are used, thediagonal bars must be properly an-chored in the adjoining wall. The diago-nals must be restrained over their fulllength to prevent buckling.

Since this type of detail effectivelydeveloped strain hardening of the rein-forcement, the actual capacity of thebeams should be considered in design-ing a structural wall system. A designbased on yield level would not properlyallow for the forces that can be impartedto the walls by the beams.

Final Remarks

The results of these tests have clearlyindicated the relative influence of spe-cial reinforcement details on inelastichysteretic response of coupling beams,However, the results do not justify theuse of one system over the other for all

PCA

situations. The response characteristicsand energy-dissipation capacity attain-able in the beams must be matched withthose required for the structure and thedesign conditions under consideration.

For example, for very short beamsunder severe earthquake loads, full-length diagonals may provide the bestsolution. In other situations, however,conventionally reinforced beams mightbe adequate. Also, consideration mustbe given to the fact that not all beamsover the height of a coupled-wall systemare subjected to the same load histories.

ACKNOWLEDGMENTS

Investigation reported here was carriedout at the Structural Development De-partment of the Portland Cement Asso-ciation. Fabrication and testing of thespecimens were performed by the tech-nical staff of the department under thedirection of laboratory foreman B. W.Fullhart.

The work was part of a combinedanalytical and experimental investiga-tion sponsored by the National ScienceFoundation under Grant No, ENV74-14766 and by the Portland CementAssociation. Mark Fintel was overallproject director. The opinions, findings,and conclusions expressed in this publi-cation are those of the authors and donot necessarily reflect the views of theNational Science Foundation.

Research and Development Bulletin 17

4. Wight, J. K,, and Sozen, M. A.,“Strength Decay of RC ColumnsUnder Shear Reversals,” Journal ofthe Structural Division, AmericanSociety of Civil Engineers, Vol. 101,No. ST5, May 1975, pages 1053-1065.

5. ACI Committee 318, Building CodeRequirements for Reinforced Con-crete, ACI Standard 318-71, Ameri-can Concrete Institute, Detroit, 1971,78 pages.

6. Barnev. G. B.: Shiu. K. N.: Rabbat.

7.

8.

B. G.;”Fioratoj A. E:; Russell, H. G.:and Corley, W. G., “Earthquake Re-sistant Structural Walls—Tests ofCoupling Beams,” Portland CementAssociation, January 1978, NationalTechnical Information Service, U.S.Department of Commerce, Spring-field, Va. (NTIS Accession No. PB-293 54217WB).Bachmann, H., “Influence of Shearand Bond on Rotational Capacity ofReinforced Concrete Beams; Inter-national Association for Bridge andStructural Engineering, Vol. 30, PartII, Zurich, 1970, pages 11-28.Standard Specification for Deformedand Plain Billet-Steel Bars for Con-crete Reinforcement, A61 5-76a,American Society for Testing andMaterials, Philadelphia.

9. Hognestad, E.; Hanson, N. W,; Kriz,L. B.; and Kurvits, O., Facilities andTest Methods of PCA StructuralLaboratory, Development Depart-ment Bulletin DX33, Portland Ce-ment Association, 1959.

1. Brown, R. H., and Jirsa, J. O., “Rein-forced Concrete Beams Under LoadReversals; Journal of the AmericanConcrete Institute, Proceedings, Vol.68, No. 5, May 1971, pages 380-390.

2. Paulay, T., and Binney, J. R., “Diag-onally Reinforced Coupling Beamsof Shear Walls: Shear in ReinforcedConcrete, Publication SP-42, Ameri-can Concrete Institute, Detroit, 1974,pages 579-598,

3. Bertero, V. V., and Popov, E. P.,“Hysteretic Behavior of ReinforcedConcrete Flexural Members withSpecial Web Reinforcement: Pro-ceedings, U.S. National Conferenceon Earthquake Engineering, AnnArbor, Mich., June 1975, pages 316-326.

APPENDIX: EXPERIMENTALPROGRAM

Details of the experimental programincluding specimen geometry, reinforce-ment details, material properties, fabri-cation, and testing are given in this Ap-pendix.

Test Specimens

Eight specimens representing approxi-mately one-third scale models of coup-ling beams were tested. Each specimenconsisted of two coupling beams fram-ing into rigid abutment walls at eachend as shown in Fig. 3. The end condi-tions imposed by the abutments simu-lated those in a structural wall system.


1“: 25.4 mm

n

● . .

..s3uJr~6mmBa,,

f

~4°cL. D-3 Deformed

Wire

*IL. . . . I

SECTION A-A FOR

SPECIMEN C5

l-%6.67 “

[g

. .

‘he“CL.

.

~4° CL.

O-3 Deformed

Wire

6 mm Bor

ELEVAtlON

SECTION A-A FOR

SPECIMEN Ci? (a) Specimens C2 and C5

6.67 “

25@ L33” = 33.3” I

r -A

6 mm Bars

D -3 OeformedWire

I

I b~ELEVATION

,,

Ii

Dlzlk...

D-3 Omformed

Wlra

~4° C L.

● **6 mm Bars

SECTION A-A FOR

SPECIMEN C7

Il“: 25,4mm

(b) Specimen C7

Fig. A-1. Detaiis for beams with conventional reinforcement.

Coupling beams had rectangularcross sections 4 in. (102 mm) wide and6.67 in. (169 mm) deep. The beams wereeither 16.67 in, (423 mm) or 33.33 in.(846 mm) long, corresponding to span-todepth ratios of 2.5 and 5.0, respec-tively. The L-shaped abutments were 4in. (102 mm) thick.

Details of Reinforcement

Steel reinforcement details are present-ed in Figs. A-1 through A-3.Specimens C2, C5, and C7. Primary re-inforcement in Specimens C2, C5, andC7 consisted of four straight longitudi-nal 6-mm-diameter (YGin.) bars, topand bottom. Reinforcement details areshown in Fig. A-1. Specimens C2 andCS had a shear span-to-effectivedepthratio of 1.4, The shear span-to-effective-depth ratio for Specimen C7 was 2.8.Specimens C5 and C7 had a confinedconcrete core size about 33910larger thanthat of Specimen C2. To obtain the in-crease in core size without bending thebars, it was necessary to piace thestraight flexural bars in the- couplingbeams outside the reinforcement in theabutment walls. Therefore, these barswere not anchored in confined concrete.This anchorage detail, although satis-factory for the test, is not recommendedfor field practice.Specimen Cl. Diagonal reinforeementwas provided for Specimen C 1as shownin Fig. A-2. Two 6-mm-diameter (Y&in.)bars top and bottom were bent at 45°starting at the face of the wall at eachend of the beam.Specimens C3 and C4. Diagonals wereprovided in the hinging regions of Spec-imens C3 and C4 as shown in Fig. A-2.Four 6-mmdiameter (%-in.) bars werebent at 45° top and bottom. SpecimensC3 and C4 were similar except for thesize of the confined concrete core. Spec-imen C4 had a core area 3370 greaterthan Specimen C3, as shown in Fig. A-2.The larger core size required that thetop and bottom reinforcing bars in thecoupling beams be placed outside thesteei in the rigid abutments. This an-chorage detail is not recommended forfield practice.Specimens C6 and C8. Primary rein-forcement for Specimens C6 and C8consisted of full-length diagonals asshown in Fig. A-3. Diagonals were two

PCA Research and Development BuIletin 19

l“ , 25.4 mm b16.67”

4No. 3 bars in one direction and one No.4 bar in the other. Symmetry was main-tained by passing the No, 4 bar betweenthe No. 3 bars at midspan. Two 6-mm-diameter (~-in.) longitudinal bars wereprovided for tying hoops in place, Thesebars were not anchored in the abutmentwalls.

Materiais

-1033” - 12 @ 1.33”116”

A, ‘Lo33°

b ‘8~6””C L.Iml < 6 mm Bars

/\

/ ) ( .\

I II

~ o-3 O* fOrmad

Wiro

14A4,SECTION B-B

ELEVATION

Concrete and reinforcing steel proper-ties for the test specimens are summa-rized in Table A-1.

Concrete was designed for a compres-sive strength of 3000 psi (20.7 MPa).The mix consisted of Type I cement,sand, and aggregate with a maximumsize of YEin, (9.5 mm). Material proper-ties were determined from tests on 6x 12-in. (153x305-mm) cylinders. Concreteproperties are contained in Table A-1and a representative stress-versus-straincurve is shown in Fig. A-4.

For the reinforcement, No. 3 and No.4 bars conformed to ASTM Designa-tion A615 Grade 60.(8)Deformed 6-mm(K-in.) hot-rolled bars with propertiessimilar to those of Grade 60 bars werealso used as primary reinforcement. De-formed wire, size D-3, was used fortransverse hoops. This wire was heat-treated to obtain stress-strain character-istics similar to those of Grade 60 bars.Physical properties of the reinforcementused in the test specimens are given inTable A-1. Representative stress-versus-strain relationships are shown in Fig.

t--l,,rl~l:C L

-8

O-3 D.f ormdWlrc

6.67

L- 6 mm Bors

t- 1~” CL.

SECTION A-A (a) Specimen Cl

1,

H ““’’a~o’”

6.67 “

- [@

I)Jti

,.,

V4° C L.

0.3 Deformed

mA e

6 mm Bars

A-5.

,,

--i

6.67”

SECTION e-B FOR

SPECIMEN C3ELEVATION

—

B.,,..,......D-3 Deformed

Wre

6.67”

6mm Bars

l/~”CL

SECTION A -A FOR

SPECIMEN C3

SECTION A-A FOR SECTION B-B FORSPECIMEN C4 SPECIMEN C4 (b) Specimens C3 and C4

Fig. A-2. Details for beams with bent diagonal reinforcement.


~r’4 “

m

O-3 Wormedwlr*

*4 Bar

nil●

#4 ~

6.67 “ #:>D-3 Deformed

Wlfo

~4° C LELEVATION l“ = 25.4mm

6.6

i-lLA?!JSECTION A-A

#4

[~!!l”

●

7“

~~” C L,

● ●

y4° Cc.

SECTION B-B

(a) Specimen C6

4“ ELEVATION

t-l

1“ = 25.4mm

DI*4

6.67” ~ O-3 OotwmedWire

sECTION A-A

6,.67

#4

[Q!l’

●

,,,

~14°C L.

● *.

~4° cc.

SECTION 6.8 (b) Specimen C8

Fig. A-3. Details for beams with full-length diagonal rein forcement.

Fabrication

Specimens were cast in a horizontalposition using the forming systemshown in Fig. A-6. Reinforcing cagesfor the abutments and coupling beamswere constructed separately and thenplaced together in the form. Before cast-ing, lifting eyes and inserts for attachingexternal instrumentation were placed inposition.

Each specimen was cast in fourbatches. Concrete for both couplingbeams in each specimen was taken fromthe same batch, After casting, the speci-mens were covered with a sheet of poly-ethylene plastic and allowed to cure for4 days. Control cylinders were curedsimilarly. Specimens were then strippedand moved to the test location. Testingusually began on the fourteenth dayafter casting.

Supporting and Loading Systems

Specimens werelaboratory floor

,~laeed parallel to theand supported on

thrust bearings as illustrated in Fig. A-7.Blocks to resist applied forces werepost-tensioned to the laboratory flooron each side of the specimen. One end ofthe specimen was fixed. Hydraulic ramswere used to apply load at the oppositeend. The line of action of the appliedforces passed through the midspan ofthe coupling beams. This minimized thepossibility of axial forces occurring inthe beams.

Magnitude of the applied forces wascontrolled by a,hydraulic pump. A four-way valve was used in the hydraulic linefor directing pressure to one of two ramsto either push or pull on the specimen.Lateral movement at the live end of thespecimens was prevented by rollerguides bearing against blocks stressedto the laboratory floor.

Instrumentation

Test specimens were instrumented tomeasure applied and resisted loads, de-flections, and steel strains. Readingsfrom each sensing device were recordedby a digital data acquisition systeminterfaced with a desk-top microcom-puter. Data were stored in cassettes.

Loads were recorded by local cells(g)located at both the fixed and live ends ofthe specimens. This arrangement pro-vided a means for determining losses


TABLE A-1. Material Properties

D-3* deformed wire 6-mm** bar(ksi)

No. 3 bar(ksi)

No, 4 bar I Concrete(ksi) (ksi)

&

—

—

—

—

—

30,200—

28,600

f,

70.0

69.369.370.871.1

71.4

62.1

71.0

f:

—————

59.2—

62.8

($i)

3180

2910

3040

3170

2730

2780

3050

3100

E, f:(psi)

2940

3050

2970

3490

3140

f,.

76.3

75.0

75.1

75.0

75.1

75.1

72.5

81.7

f,

69.2

74.9

73.6

66.0

66.3—

66.5—

f,.

98.0

99.8

98.8

89.7

88.8—

87.8—

f,

—————

70.7—

82.5

fsu

—————

104.7—

125.1

f,.li

32,400

30,000

29,400

31,300

31,100

31,000

26,700

28,300

.&

31,400

30,000

30,600

30,000

30,000—

30,800—

cl

C2

C3

C4

C5

C6

C7

C8

—

—

—

——

103.2—

102.5

—

——

——

31,000—

28,500

3710

3470

.f, = yield strength of steel X = compressive strength of concretefs. = tensile strength of steel& = modulus of elasticity of steel

“Area = 0.03 aq in. (19 mmz)‘“Area = 0.05 sq in, (32 t-nrnz)E. = modtilus of elastic~ty of concrete

1000 psi = 1 ksi = 6.895 MPa

75 -

Stress,ksi

50 -

25 -- 200

3000

Stress

psi

2000

,0

StressMPa

5

)

I

“o 5000

Fig. A-5. Representativeships for reinforcement.

.10000 15000

Strain, millionths

stress-versus-strain

20000 25000

relation-

1000

0 t 1 1 1

0 500 I000 1500 2000

SIram , millionths

Fig. A-4. Representative stress-versus-strain relationshipfor concrete.

caused by friction in the thrust-bearingsupports. Loss was generally less than2% of the applied loads.

Lateral displacements in most speci-mens were recorded at three locations.The sensitivity of the gages was 0.001 in.(0.03 mm). One gage was attached tobrackets on the inside face of each rigidabutment midway between the couplingbeams, as shown in Fig. A-8. Two addi-tional gages were installed at the end of


each coupling beam. All three gagesmeasured the relative lateral displace-ment of the ends of the coupling beams.Gages were also installed to measureaxial deformations at each couplingbeam location.

A continuous record of load versusdeflection was also recorded by an XYplotter, Load was measured by calibrat-ed pressure cells. Deflections were takenas the relative lateral movement be-tween the abutment walls.

Test Procedure

Prior to yield, loading was controlled bythe magnitude of applied forces. Afteryielding, loading was controlled by theimposed deflection on specimens.

Three complete loading cycles wereapplied for each predetermined level offorce or displacement. The combinedthree cycles are termed a load increment.Each cycle started and ended with zeroforce applied.

The first cycle of each increment wasapplied in steps termed load stages.Data were recorded at each load stage.On the second and third cycles, datawere recorded only at peak loads or dis-placements.

Testing was terminated either when asignificant loss of capacity occurred orwhen specimens experienced bar frac-ture.

After each load stage, specimens wereinspected visually for cracking and evi-dence of distress. Cracks were markedwith a felt-tip pen. Photographs weretaken at the end of each load stage toprovide a record of crack development.

El El\ Loti Cell

Ad?.

Fig. A-7. Teet setup.

Fig. A-8. External instrumentation.

*Direct-current differential transformer

Test Results

Test results are given in the body of thisreport and have also been published indetail elsewhere. ‘b)

This publication is based on the facts, tests, and authorities stated herein. Itis intended for the use of professional personnel competent to evaluate thesignificance and limitations of the reported findings and who will acceptresponsibility for the application of the material it contains. The PortlandCement Association disclaims any and all responsibility for application ofthe stated principles or for the accuracy of any of the sources other than workperformed or information developed by the Association.

r --------------------- --------------------- ------- ---------1 7iII

iIIIIIIIIIIIIIII

;IIIIII

iIII1

KEYWORDS: beams, buildings, concrete (reinforced), deformation, earthquakeresistant structures (forces), loads (forces), reinforced concrete, reinforcement(structures), shear strength, shear stress, structural design, tests.

ABSTRACT Describes tests of eight model reinforced concrete coupling beamspecimens subjected to reversing loads to determine load versus deflection; strength;energy dissipation; and ductility. The effects of shear span-to-effectivedepth ratio,reinforcement details, and size of confined concrete core on hysteretic responsewere determined. Full-length diagonal reinforcement improved performance ofshort beams. Larger concrete core increased load-retention capacity.

REFERENCE: Barney, G. B., and others, Behavior of Coupling Beams UnderLoad Reversals(RD068.OIB),Portland Cement Association, 1980.

iI

jIIIIII:IIIIIIIIIIIIIIII1IIII

i ------------------------- -------------------------- -------- J

PCA R&D Ser. 1585

PORTLAND CEMENT mII ASSOCIATION

An organization of cement manufacturers to improve and extend the uses of portland cement and concrete through scientific research, engineering field work, and market development.

5420 Old Orchard Road, Skokie, Illinois 60077

Printed in U.S.A. RD068.01 B

Documents

Behavior of Coupling Beams Under Load Reversals - Concrete