BOND IN PRESTRESSED CONCRETE END SUPPORT EFFECTS

In previous studies many of these variables have beeninvestigated.However, since the principal objective in thisinvestigation was the pinching effect of the end reaction inbeams with more than one layer of strand, the vertical spacingof the strands and the embedment length were of maximum concern.Therefore, the test specimens had the following characteristics:(1)One size of 270K strand (1/2-in.) wasused in all three specimens.The ratioof surface perimeter to cross-sectionalarea decreases as the size of the strandincreases.Therefore, bond is most crit-ical in members pre-tensioned with thelar estsize strand.The 1/2-in. sizestrand is the largest size currently usedcommercially.From a previous investiga-tion by Hanson and Kaar 7,it was found thatrust or scale on the strand improves thebond characteristics of the strand.Th re fore, a rust-free strand should representthe most critical case.(2)Five strands and one strand pattern wereused in each of the test specimens.Ofthe five strands, analysis showed that bond would be critical in the four lower strands . -10-

The four lower strands were placed intwo layers, with one strand immediatelyabove the other.The center-to-centervertical spacing was 2 inches, which isthe minimum allowed by the current ACIlBuilding Code.The positioning of thestrands was such that the difference inthe strand stress between the lower twolayers of strands at ultimate load ofthe test specimen was less than 0.5%.(3)The initial strand stress was constantfor all specimens, that is 70% of theminimum ultimate strength which, for270K strand, was 189 ksi.(4)The concrete strength was governed bytwo limits:(1)the release strengthmust be more than 4500 psi, and (2) thestrength at time of test must not bemore than 6000 psi.(5 )From the results of a previous study by4Badaliance and VanHorn it was shown thatthe critical embedment length was 80 inches.Therefore, an embedment length of 48 incheswas chosen to ensure bond slip in the two-11-

lower layers of strand, prior to aflexural or shear failure . The aforementioned characteristics werespe ifiedso asto fulfill the requirements of the current specifications, and toensure a failure by bond slip in at least the lower level of strand.,-12-

»III.TEST SPECIMENS 3.1Description of Test SpecimensThe three specimens were designed as under-reinforcedbeams such that a flexural failure would be initiated by yieldingof the strand, followed by crushing of the concrete.In accordancewith the objectives, the tests were designed to indicate (1) thevalue of the component of frictional bond that results from thepinching effect of the reaction, and (2) the amount of additionalload that the beam may sustain after bond slip.The specimens consisted of rectangular prestressed beamswith a cast-in-place slab.The prestressed rectangular beamelement was 7-1/2 inches wide and 12 inches deep, the slab was6 inches deep and 20 inches wide.The beam width of 7-1/2 inchesallowed a strand pattern that ensured the minimum cover requirements as specified by the current ACI Building Code.lThe bottomlayer of strand had a cover of 1-1/2 inches, the minimum allowedby the Code, so as to maximize the component of normal pressureat the interface of the strand and the concrete resulting fromthe pinching effect of the reaction.The depth of the beam sectionand the strand pattern were designed so that the maximum allowableprestress stress of the concrete was not exceeded.270K strands were used for prestressing the beam.were located in three layers.Five 1/2-in.The strandsFour of these strands were placed-13-

near the bottom of the section, two strands in two layers witha vertical center-to-center spacing of 2 inches. The fifthstrand was placed 2 inches from the top of the prestressedelement and served to limit the prestress stress which was imposed upon the concrete.This strand pattern produced nearlyequal stresses in the two lower levels of strand.The strandswere initially tensioned to a stress of 70% of the specifiedultimate stress which is 189,000 psi for 270K strand.Figure 1includes a detailed sketch of the cross-section.The test section was reviewed to determine to expectedultimate load and the associate failure characteristics.Thisseries of calculations was carried out in accordance with themethod presented by Janney, Hognestad, and McHenry.10Theultimate moment was calculated using the factors, k , k , k ,l23and uas presented by Hognestad, Hanson, and McHenry.8Theultimate strength factors were expressed as a function of theconcrete strength as follows:k kl 3k2 3900 0.35 flc3200 flc flc0.50 80,000fl U 0.004 -C6.5 x 10-14-6

JUsing these factors and a value of Fstress at ultimate flexural strength fcessive approximations. 1.0,the steelwas calculated by suc-suFor beams containing more than onelevel of steel, the strain compatibility equations are:aFeu (1 - d) -------;::;--de- e FesUseutcFeuThe condition of equilibrium of forces leads to:The flexural moment is given by the equation:Several trials were necessary to establish compatibility betweenthe calculated fsuand the value of esuand fsuobtained fromthe stress-strain curve of the strand.In the design calculation an estimate of the prestresslosses was required.The components of losses which were includedin the estimate were elastic shortening, shrinkage of the concrete,and creep in both the steel and concrete.-15-The loss in steel stress

due to elastic shortening was determined by:t,fLin12srecommends that concrete shrinkage loss becalculated by:t,fwhereE:ssc E:Es sis equal to the unit shrinkage strain of the oncrete.A value ofE:s 0.0003was used in this study.The loss due toconcrete creep was likewise calculated by the equation:t,fcs (Cc- 1) nfcwhere the value of C , the creep coefficient, was assumed equalcto 4.The loss due to creep in the prestressing steel wassimilarly computed by:t,fss K0 F.l -16-

where K is equal to three percent.oTable 4 is a comparisonbetween the measured and the calculated values of prestress losses.3.2MaterialsPrestressing SteelThe prestressing steel used was 1/2-in., 270K, seven-wire, uncoated, stress relieved type strand.This type of strandis commercially manufactured and tested in accordance with ASTMdesignation A416-64.The physical properties of this strand aregiven in Fig. 3 and the load-elongation curves are shown in Fig. 4.Further information concerning the static and fatigue propertiesof this strand is given in a recent report by Tide and VanHorn.15Although this type of strand is commercially available from severalmanufacturers, the strand used in this investigation was producedby John A. Roebling's Sons Division of the Colorado Fuel and IronCorporation.Shear ReinforcementThe shear reinforcement was fabricated from No. 3formed bars having a nominal yield stress fY de 50,000.ConcreteThe concrete strength was not a variable in this investigation.The concrete used in the beam element was designedto yield an ultimate strengthf -17- 6000psi, at an age of 21 days.

The mix chosen consisted of Type III (high-early strength) portland cement, sand, and crushed limestone coarse aggregate (3/4-in.maximum).The proportions of the mix, by weight, (cement-to-sand-to-coarse aggregate), were 1.00:2.64:2.98.The concrete wassupplied by a local ready-mix plant and was delivered in 1-1/2cubic yard batches.Although the concrete used in casting theslabs was of the same design as that used in the webs, tsslumpwas 2-1/2 inches compared to a slump of 3-1/2 inches for the beams.This difference in water content yielded a higher 21-day compressive strength for the slabs than for the beams.The three rectangular prestressed beams were all castfrom one batch of concrete, and all of the slabs from another.Compression tests were conducted on 6 x 12 in. cylinders, whichwere cast along with each batch of concrete, to determine thecompressive strength fT associated with the test beams at the timecof prestress release and at the time of test. Strains weremeasured on selected cylinders with a compressometer to determinethe shape of the stress-strain curve and the modulus of elasticityof the concrete at the time of test.As a measure of the tensilestrength of the concrete splitting tensile tests were conducted onstandard 6 x 12 in. cylinders.1/8-in Strips of plywood approximatelythick, l-in. wide, and 12 inches long were placed on thediametrical upper and lower bearing surfaces in the splitting test.The splitting cylinder tensile stress fT was determined by thespequation:-18-

f'sp -2PTId LcThe age and strength properties of the concrete described in the preceding paragraphs are presented in Table 2.3.3FabricationThe beams were fabricated in .a prestressing bed at theFritz Engineering Laboratory.follows:The sequence of operations was asthe strands were tensioned, strain gages were attachedto the strands, lead wires were soldered into place, the shearreinforcement waspo itionedand wired, forms were erected, thebeam concrete was placed and cured, the beam forms were removed,Whittemore targets were installed, the strands were released,the slab forms were set in place, the slab concrete was placedand cured, the slab forms were removed.The prestressing bed has been described in previousFritz Engineering Laboratory reports.16The bulkheads of theprestressing bed were spaced 40 feet apart and bolted to thelaboratory floor.The three beams were cast simultaneouslyusing the same assemblage of strands and three individual steelforms.The prestressing strands were held in the chosen patternwith two one-inch thick anchorage plates, one at each end of theprestressing bed.The projecting ends of the strands were securedby strand chucks.-19-

The tension was applied to the prestressing strandsby jacking the moving bUlkhead, using two 50-ton hydraulic jacks.The strand tension was measured for each strand individually byplacing a load cell between the anchorage plate and the strandchuck at the stationary end of the prestressing bed.If required,the tension in individual strands was adjusted by means of aspecial hydraulic jacking arrangement.After pre-tensioning thestrand, the strain gages were mounted on the strand.The shear reinforcement was tied to the strand with14 gage wire.In addition, wire ties were used between successiveprojecting elements of the stirrups in the compressive flangetarea, in order to prevent movement of the stirrups during theplacement of the concrete.Steel forms made of No. 7 gage steel plates were usedto cast the test beams.Dimensional checks, made after the forms had been removed, indicated that the cross-sectional dimensions were maintained to within 1/16 in. and consequently, the nominal dimensionsof the cross-section were used in all calculations.The concrete was brought from the ready-mix truck tothe forms in steel buggies, and shoveled into the forms.Theconcrete was placed in two layers, and the tops of the beams wereleft unfinished.each beam.Ten standard concrete cylinders were cast withWaxed cardboard molds were used.-20-

The concrete in the test specimen was internally vibrated, while the cylinders were rodded.All specimens were covered with burlap and plasticsheeting for a period of two days, after which the forms wereremoved.After the surface of the test beams had become air dry,Whittemore targets were positioned on the beams.When cylindertests indicated that the specimen1s concrete compressive strengthhad reached a strength of 4500 psi, the prestress force was slowly,released.An oxy-acetylene torch was used to cut the strands.,Thebeams were then removed from the casting bed, and stored in thelaboratory.Wood forms were used for fabrication of the slabs.Theconcrete was placed in one layer and allowed to cure under thesheets of plastic and burlap for one week.At that time, theforms were removed and the beams were air-cured until the timeof test.3.4InstrumentationIn order to determine the behavior of the beams, datawas recorded in the form of strains, strand movement and deflections.-21-

3.4.1StrainsLoad-deformation data was measured in two fashions (1) strain gages mounted at various locations on the strands,and (2) strain gages mounted on the concrete surface.Load-deformation data on the strand was measured withSR-4 electrical resistance strain gages, A-12 type, mounted onthe lower four strands of all three beams.The gages were mountedupon completion of, pre-tensioning, in order to increase the effective range of strain measurements.The gages were located onthe instrumented strands at 20-in. intervals starting 20 inchesfrom the ends of the members.One gage was attached to an in-dividual wire at each gage location.Before the strain gage wasmounted the strand was cleaned using emery cloth and acetone.The gage was glued to the strand using Duco-Cement.The water-proofing consisted of a coating of Armstrong adhesive A-16,followed by a layer of liquid rubber.The location of thesestrain gages can be seen in Fig. 5.The primary objective in positioning strain gages onthe prestressing strand, was to determine the change in strandforce as load was applied to the beam.In addition, these straingages were used to measure the elastic shortening of the strandat the time of release of the pre-tensioning force.The straindata received from the gages mounted on the strands was converted to force with a calibration curve.-22-This plot of strain

vs. strand force was determined by averaging strains measuredby three gages mounted on individual wires of a strand sample.Figure 4 compares the calibration curve to the load-deformation'14curve determined by Tide and VanHorn.The load-deformation data of the concrete surface wasmeasured (1) with a S-in. Whittemore gage, and (2) with SR-4 gageslocated around the cross-section at midspan of the test specimens.a.Whittemore strain gage data was used todetermine concrete surface strains at alevel midway between the lower two layers of prestressing strand.The Whittemoretargets were fabricated from brass plugs,7/32-in. in diameter and 1/16-in. inthickness.These targets were center-drilled and cemented at the prescribedlevel on both sides of the beam, afterthe forms were removed but before releaseof the pre-tensioning force.The purposeof the Whittemore data was to determine thetransfer length, and the loss of prestressat various intervals up until the time oftest.These gage readings were lsousedas an indicator of initial cracking of theconcrete .-23-

b.Twelve SR-4 gages, A-9 type, weremounted on the cross-section atmidspan of the beam.Figure 5 showsthe placement of these gages.Thestrain information from these gageswas used to determine the location ofthe neutral axis as cracking becamesevere, and the maximum concrete strainat ultimate load.3.4.2Strand SlipStrand slip was measured by the use of Ames dial gageswith a least count of l/lO,OOO-inch.These dial gages wereattached to a steel mounting bracket that clamped to the rectangular beam portion of the specimens.The gages indicated themovement of four small steel plates that were attached to thefour lower strands by means of collars and set screws.(SeeFig. 16)3.4.3Deflection MeasurementThe midspan deflection was measured by level readingson scales graduated to the nearest 0.01 inch.located at each of the supports and at midspan . -24-These scales were

IV.4.1TESTSTest ProcedureThree specimens were tested in the 300 kip hydraulictesting machine located in the Fritz Engineering Laboratory.The loading beam and associated apparatus were arranged to provide a symmetric two-point loading for all three specimens.The shear span was 42 inches and the overhang was 6 inches, resUlting in a total embedment length of 48 inches.the testing arrangement is shown in Fig. 2. A sketch ofThe test specimensin all cases were initially loaded in increments of 10 kips,which was approximately 8 percent of the computed ultimate load.When cracking of the concrete became visible, the loading increment was reduced to 5 kips until failure occurred.The in-ternal A-12 type, SR-4 gages mounted on the individual wires ofthe strand were used in conjunction with two different types ofrecording equipment.Six gages of Beam Z-l were connected tochannels of a Brush direct-writing recorder to provide a continuous record of the variation of the load in the strands.The remaining internal gages, as well as the A-9 type, SR-4 gagesmounted on the surface of the concrete at the midspan cross-section,were connected to a Budd Datran digital strain indicator.Allelectric resistance strain gages were read following the application of each increment of load.A review of the test results from-25-

Beam 2-1 revealed that little was gained from the use of theBrush recorder. Therefore, in the remaining tests the instru-ment was not used .The Whittemore targets mounted on the surface of theconcrete were read at selected load increments until crackingbecame severe.Midspan deflection readings were taken at eachload increment until failure.The strand-slip dial gages wereread continuously to detect initial bond slip.The developmentof the crack pattern was marked on the surface of the specimenafter each increment of load had been applied.After failurethe specimens were photographed . 4.24.2.1Test Results and DiscussionModes of FailureIn this investigation failure in all three test speci-mens has been associated with flexure.The tensile reinforce-ment initially yielded, followed by the eventual crushing of theconcrete.However, prior to yielding of the reinforcement, bondslip was experienced in the four lower strands in all three tests.The failure mechanism associated with bond slip is closely related to the cracking of the concrete.Cracking of the Concrete In this study two basic types of cracking were exhibited,flexural and flexural-shear cracking . -26-

Flexural cracking occurred in the high moment regionof the test specimens when the stresses in the bottom fibersreached values which are normally associated with the tensilestrength of the concrete.Flexural cracking was characterizedby the initial development of vertical cracks to a level whichvaried between the lower and upper* strand levels.The longi-tudinal spacing of these flexural cracks was approximately6 inches.However, cracks which formed closer together thanapproximately 2 inches would usually merge, or the further development of one of the two cr

prestressed concrete members, the bonding action between the prestressing element and the concrete is a result of friction as well as adhesion and mechanical resistance. In the fabrication of a pre-tensioned prestressed concrete member the prestressing strand is first tensioned to the desired stress level. Concrete is then cast about the strand and