measured transfer length of 0.6 in. prestressing strands ... journal/2015/may/measured... ·...

May–June 2015 | PCI Journal84

Structural lightweight concrete has numerous benefits when used in precast, prestressed concrete members because of its low self-weight and high durability.

Lightweight concrete can be classified as all lightweight, in which both lightweight coarse and fine aggregates are used, or sand lightweight, in which lightweight coarse aggregate is combined with conventional concrete sand.1 Structural lightweight concrete typically has an equi-librium density between 105 and 120 lb/ft3 (1680 to 1920 kg/m3).1 Self-consolidating concrete (SCC) has also become increasingly common in prestressed concrete applications because of the possibility for reduced time, labor, and noise during construction as well as for produc-ing an improved surface finish in hard-to-vibrate areas. Exact definitions vary, but SCC should flow and fill forms under its own weight without vibration, remain homoge-neous through long flow distances and vertical drops, and flow through congested areas without blockage or segrega-tion.2 Combining lightweight concrete with SCC behavior produces numerous benefits in situations where self-weight is a concern and SCC is preferred for production. Light-weight SCC for bridge girders has garnered more study in recent years due to a desire for weight reduction in long-span girders and to fit with the increasingly common production methods used in precast concrete plants.3 The differing material properties of lightweight SCC have the potential to affect the transfer of prestress in members cast

■ Lightweight self-consolidating concrete (SCC) in prestressed concrete members has the potential to increase transfer length.

■ Twenty-five rectangular concrete members prestressed with two 0.6 in. (15 mm) prestressing strands were cast using SCC made with expanded clay, expanded shale, and limestone coarse aggregates.

■ Measured transfer lengths indicated no significant difference between concrete types at the same compressive strength, but members with higher compressive strengths at release had shorter transfer lengths.

■ Average transfer lengths were shorter than the American Concrete Institute/American Association of State Highway and Transportation Officials’ code predictions for all concrete types.

Measured transfer length of 0.6 in. prestressing strands cast in lightweight self-consolidating concrete

Royce W. Floyd, W. Micah Hale, and Michael B. Howland

85PCI Journal | May–June 2015

capacity.13 Typical SCC mixtures are proportioned with either a high content of fine material, and as a result paste, or a viscosity-modifying admixture. Pullout testing by Burgueño and Haq4 on 0.5 in. (13 mm) prestressing strand indicated a lower bond strength for SCC mixtures than for conventionally consolidated mixtures and that the mixtures with higher fine-material contents had greater differences. The type of fine material in an SCC mixture was shown to have a marked effect on bonding capability by Hegger et al.,6 with better bond measured when angular powder par-ticles (which also produced a stronger matrix) were used. Longer transfer lengths or poorer bond strength have been measured for SCC mixtures containing fly ash than for con-ventional concrete with only portland cement,5,14 while SCC mixtures containing only portland cement15-17 or slag ce-ment18 exhibited similar behavior to conventional concrete. However, in some cases similar performance was observed for SCC mixtures containing fly ash and conventional concrete.13,16 The use of viscosity-modifying admixtures has in some cases been shown to lead to reduced early-age compressive strength and bond strength for SCC.5

Previous research concerning transfer length of prestress-ing strands cast in conventionally consolidated lightweight concrete and SCC indicates substantial variability in bond performance.19–21 Table 1 summarizes results for lightweight concrete specimens. Some measured transfer lengths for lightweight concrete were greater than those measured for the corresponding normalweight concrete and the ACI 318-11 and AASHTO LRFD specifications predictions,19 and some were less than those predic-tions.20,21 Specimens with higher compressive strength typically exhibited shorter transfer lengths.15,21 Transfer lengths measured for SCC mixtures in previous research also exhibited varying results (Table 2). Values for SCC were typically less than or in agreement with the code pre-dictions,6,13,15,16,18,22 but not in all cases.5 Trends in compari-sons with corresponding conventional concrete mixtures were not consistent among all studies. Some researchers measured longer transfer lengths for SCC specimens,5,6 and some measured similar or shorter transfer lengths.15,16

Transfer lengths measured by Burgueño and Haq for both SCC and conventionally consolidated specimens were less than those produced by Eq. (1), but the transfer lengths for the SCC specimens were more than 25% greater than the transfer lengths measured for the conventionally consoli-dated specimens.4 Peterman13 and Staton et al.16 measured transfer lengths that were less than the values produced by the code equations using strands that met a required prequalification. The strands used by Naito et al.18 exhibit-ed inadequate performance in the prequalification test used despite transfer lengths shorter than the predicted values.

Although a significant amount of research has been conducted on the bond of prestressing strands cast in various types of concrete, only limited work has been done concerning the bond of prestressing strands cast in

using lightweight SCC and seven-wire prestressing strand. Transfer bond, as quantified by transfer length, has been highly debated, with numerous research projects conducted concerning various reinforcement types, reinforcement configurations, and concrete types. Some previous re-search has shown longer transfer lengths for SCC than for conventional concrete,4–6 but little research has focused on the bond performance of prestressing strands in conjunc-tion with lightweight SCC. Lightweight concrete is known to have reduced shear strength compared with conventional concrete, as indicated by reduction factors in the American Concrete Institute’s (ACI’s) Building Code Requirements for Structural Concrete (ACI 318-11) and Commentary (ACI 318R-11),7 and an accurate transfer-length prediction is needed for shear-strength calculations in prestressed con-crete members.

Background

The transfer of tensile force from the strand to a compres-sive force in the concrete is accomplished over a bonded length, termed the transfer length, along which the steel stress increases from zero at the free end of the strand to the full initial prestress at the end of the transfer length.7–10 The magnitude of this transfer length is considered to be affected by initial prestress, strand diameter, strand surface condition, concrete compressive strength, and method of strand release.6–8,11 Complete agreement has not been reached concerning the contribution of each of these factors. Only strand diameter and effective prestress are included in the ACI 318-11 and American Association of State Highway and Transportation Officials’ (AASHTO’s) LRFD Bridge Design Specification12 expression for transfer length lt (Eq. [1]).

l

fdt

seb=

3 (1)

where

fse = effective prestress after all losses

db = nominal strand diameter

ACI 318-11 and AASHTO LRFD specifications also include the alternate values of 50db and 60db, respectively, for use in shear design.

Both lightweight concrete and SCC have the potential for increased transfer length. Lightweight concrete has a lower elastic modulus and greater shrinkage than normalweight concrete. SCC also has the potential for greater shrink-age and a higher paste content than other concrete, and the chemical admixtures used for SCC may also affect transfer length. An increase in fluidity for a given water-cement ratio w/c has been shown to result in reduced bond


using measured surface strain. Measured values were also compared with the predictions produced using the ACI 318-11 and AASHTO LRFD specifications equations.

Materials and methods

The six different SCC mixtures were made with expanded clay, expanded shale, or limestone coarse aggregate and normalweight concrete sand. One mixture using each ag-gregate had a design compressive strength at release fci

'

of 4000 psi (28 MPa) and 28-day compressive strength fc'

of 6000 psi (41 MPa). These mixtures were designated NS for normal-strength followed by C, S, or L for ex-panded clay, expanded shale, or limestone, respectively. The remaining three mixtures had a design fci

' of 6000 psi (41 MPa) and fc

' of 7000 psi (48 MPa) and were desig-nated HS for high-strength followed by C, S, or L. Table 3 gives a summary of the mixture designations. The normal-strength values were chosen based on typical values for most prestressed concrete members, and the high-strength values were chosen as representative of typical-span bridge girders. All mixtures used high paste contents and high-range water-reducing admixtures to achieve SCC behavior. Two mixtures contained Class C fly ash, while the remain-der contained only portland cement. Four specimens were cast using each concrete mixture with the exception of the normal-strength shale (NSS) mixture. An extra specimen was cast to test a trial version of this mixture in order to utilize available space in the prestressing bed on another project. Table 4 presents the mixture proportions for each concrete type, including the different mixture used for specimen NSS-1.

Specimen details

Each beam specimen had a 6.75 × 12 in. (170 × 300 mm) rectangular cross section and was 18 ft (5.5 m) long. These dimensions were based on those used for previ-

lightweight SCC. Lachemi et al.14 measured lower bond strength for lightweight SCC than for normalweight SCC having a compressive strength of 5250 to 6420 psi (36.2 to 43.6 MPa) from direct pullout tests of deformed reinforc-ing bars. Expanded-shale-aggregate mixtures had a higher bond strength than those containing blast-furnace-slag aggregate, indicating an impact of aggregate type on bond strength.14 Ward et al.17 measured transfer lengths for 0.5 in. (13 mm) prestressing strands cast in expanded clay lightweight SCC with compressive strength of concrete at release of prestress fci

' equal to 4530 psi (31.2 MPa).

These measured transfer lengths were approximately 30% less than those predicted using the ACI 318-11 and AASHTO LRFD specifications equations, indicating good bonding.

Prestress transfer performance is a difficult problem be-cause of the number of variables involved and the difficulty of obtaining consistent measurements. Previous research has led to mixed results concerning the bond performance of prestressing strands cast in SCC and lightweight SCC. Accurate comparisons among different studies are compli-cated by seemingly small differences in the materials and methods that can potentially affect the final results. Several factors have been identified as potential contributors to differences in behavior among SCC, lightweight SCC, and conventional concrete, but the variability of transfer-length measurements has made drawing general conclusions difficult. This goal of this study was to increase the data available for SCC mixtures using lightweight aggregate and a high paste content to facilitate informed decisions on the factors influencing bond in SCC members.

Experimental study

Twenty-five rectangular beam specimens were cast using six concrete mixtures to compare the transfer length of strands cast in lightweight SCC to normalweight SCC

Table 1. Previous transfer-length research using lightweight concrete

ResearchAggregate

typeConcrete

typeDiameter,

in.Specimen

type fc', psi fci

' , psi lt, in. fsedb/3, in. 50db, in. 60db, in.

Meyer and Kahn

Slate LWC 0.6AASHTO Type II

8000 6000 21.9 30.830.0 36.0

10,000 7500 15.6 34.2

Thatcher et al.

n.d. NWC

0.5AASHTO Type I

60003490 18.2 32.1

25.0 30.0Clay LWC

4900 35.8 30.7

8000 5560 34.4 30.5

Nassar Slate LWC 0.5AASHTO Type IV

6380 4775 17.2 n.d. 25.0 30.0

Note: AASHTO = American Association of Highway and Transportation Officials; db = nominal strand diameter; fc' = 28-day compressive strength of

concrete; fci'= compressive strength of concrete at release of prestress; fse = effective prestress after all losses; lt = transfer length; LWC = light-

weight concrete; NWC = normalweight concrete; n.d. = no data. 1 in. = 25.4 mm; 1 psi = 6.895 kPa.


first 4 ft (1.2 m) from each end and at 5 in. (125 mm) for the remainder of the beam for the normal-strength clay and normal-strength limestone specimens. The 3 in. (75 mm) spacing was extended to 5 ft (1.5 m) from each end for the remainder of the specimens. Figure 1 shows a diagram of the beam cross section and elevation view. Prestress losses used for transfer-length predictions were calculated using the AASHTO refined method.12 The moduli of elasticity

ous research16,17,22–27 so that comparable data would be obtained. Two 0.6 in. (15 mm) prestressing strands were located 10 in. (250 mm) from the top of each beam. Two no. 6 (20M) deformed reinforcing bars were placed 2.5 in. (65 mm) from the top to increase the available compression force at failure and to resist high tensile stresses at release. Shear reinforcement consisted of 1⁄4 in. (6.4 mm) smooth-bar closed stirrups spaced at 3 in. (75 mm) on center for the

Table 2. Previous transfer-length research using self-consolidating concrete

ResearchAggregate

typeConcrete

typeDiameter,

in.Specimen

type fc', psi fci

' , psi lt, in.fsedb

3 in. 50db, in. 60db, in.

Girgis and Tuan

Hard rock

SCC

0.6NU I-girders

8030 5980 43.0 n.d.

30.0 36.0SCC 10,890 6490 36.0 n.d.

NWC 9520 6970 20.0 n.d.

Naito et al. Hard rockSCC

0.5* Bulb tee8230 8270 15.7 26.0

25.0 30.0NWC 7360 6800 15.8 25.9

Hegger et al. Hard rock SCC 0.5 Rectangular n.d.2760 to 7400

<24 n.d. 25 30

Larsen et al. Hard rock SCC 0.6Rectangular 7500 5000 21.0

30 30.0 36.0T-beams 8000 5000 29.0

Peterman Various SCC 0.5 Rectangular5970 to 10,760

3500 to 5830

<30 30.4 25.0 30.0

Trent Hard rock

NWC

9⁄16 Rectangular

n.d. 4610 28.0 n.d.

28.1 33.8SCC

n.d. 4650 26.0 n.d.

n.d. 3940 33.0 n.d.

Staton et al. Hard rockSCC

0.6 Rectangular

12,240 7760 21.5 36.1

30.0 36.011,420 7540 19.7 35.3

HSC 12,380 9220 23.8 36.4

Ward et al. Clay SCC 0.5 Rectangular 6700 4530 20.1 28.7 25.0 30.0

Note: AASHTO = American Association of Highway and Transportation Officials; db = nominal strand diameter; fc' = 28-day compressive strength of

concrete; = fci' compressive strength of concrete at release of prestress; fse = effective prestress after all losses; lt = transfer length; HSC = high-

strength concrete; NWC = normalweight concrete; n.d. = no data; SCC = self-consolidating concrete. 1 in. = 25.4 mm; 1 psi = 6.895 kPa.* 1⁄2 in. special strand

Figure 1. Beam specimen dimensions and reinforcement details. Note: no. 6 = 19M; 1 in. = 25.4 mm.


used for prestress loss calculations were determined using the ACI relationship.7 The predicted rather than the mea-sured modulus-of-elasticity values were used to account for variation in compressive strength between beam batches and those used to measure modulus of elasticity.

All beam construction was conducted using a 50 ft (15 m) prestressing bed. Two beams were cast at one time in a sin-gle line with approximately 10 ft (3.0 m) clear between the ends of the two sets of forms. Precautions were taken so that the surface condition of the strand was as similar to the as-received condition as possible and the plywood forms were lined with plastic sheeting instead of using a form-release agent. Strands were tensioned to 202.5 ksi (1396 MPa), 75% of the guaranteed ultimate strength, the same day beams were cast using two hydraulic rams in parallel. The tension-ing apparatus also allowed for a gradual release of prestress. Figure 2 shows the tensioning setup and beam casting.

Concrete was placed approximately 2 hours after comple-tion of pretensioning. The lightweight aggregates were soaked in steel drums for 12 to 24 hours to ensure a consistent initial moisture condition. Two batches of concrete were required for each beam for all except the high-strength limestone specimens and one set of the high-strength shale specimens, which were cast using a single batch for each set of two beams. For all specimens requir-ing two batches, one of the batches would fill the forms to approximately half the depth, ensuring that the concrete in the area of influence of the strands would have the same properties. The time between batches was typically be-tween 20 and 30 minutes, and always less than 45 minutes. Pour lines were observed on a limited number of normal-strength clay and normal-strength limestone specimens with time between batches nearer to 45 minutes, but no evidence of structural effects was noted. Slump flow,28 J-ring,29 and density tests30 were performed on each batch of concrete or on the portion used to cast each beam for the larger batches. Desired concrete properties were

Table 3. Mixture/beam designations

Designation Aggregate fci' , psi fc

', psi Beams

NSC Expanded clay

4000 6000

1 to 4

NSS Expanded shale 1 to 5

NSL Limestone 1 to 4

HSC Expanded clay

6000 7000 1 to 4HSS Expanded shale

HSL Limestone

Note: fc' = 28-day compressive strength of concrete; fci

' = compres-sive strength of concrete at release of prestress; HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale; NSC = normal-strength clay; NSL = normal-strength limestone; NSS = normal-strength shale; SCC = self-consolidating concrete. 1 psi = 6.895 kPa.

Table 4. Concrete mixture proportions

Specimens NSC NSS-1 NSS 2-5 NSL HSC HSS HSL

Cement, lb/yd3 825 850 850 775 808* 832* 825

Fly ash, lb/yd3 0 0 0 0 142 147 0

Coarse aggregate, lb/yd3 649 764 748 1408 649 703 1392

Fine aggregate, lb/yd3 1407 1408 1437 1481 1242 1270 1403

Water, lb/yd3 329 298 298 310 333 333 330

w/cm 0.40 0.35 0.35 0.40 0.35 0.34 0.4

ADVA 575, oz/cwt 6.0 to 6.5† 6.0 5.0 to 6.0† 4.5 to 7.0† 10.0 to 14.0† 10.0 to 11.0† 6.0

Note: HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale; NSC = normal-strength clay; NSL = normal-strength lime-stone; NSS = normal-strength shale; w/c = water–cement ratio. 1 in. = 25.4 mm; 1 yd3 = 0.765 m3; 1 oz = 29.57 mL; 1 lb = 0.454 kg. *Type III cement†Dosage varied based on ambient temperature and aggregate moisture

Figure 2. Strand tensioning apparatus and beam specimen construction.


25 to 30 in. (635 to 760 mm) for slump flow, 2 to 5 seconds for the time required for slump flow to reach diameter of 20 in. (500 mm) T20, a visual stability index of 1.0 or less, a difference in slump and J-ring flow Δ of less than 4 in. (100 mm), a difference in height inside and outside the J-ring Δh of less than 1.5 in. (40 mm), and a unit weight between 115 and 120 lb/ft3 (1840 and 1920 kg/m3). No air entrainment was used. Compressive strength31 was measured for each batch at prestress release and 28 days. Modulus of elasticity was measured using the methods of ASTM C46932 at 1, 7, and 28 days for a companion batch of each mixture design.

Transfer-length determination methods

The strain at the concrete surface is a direct result of the transfer of stress from the prestressing strand and is there-fore related to the change in steel stress along the length of the member. The transfer length can be considered the length along which the surface strain is changing; a con-stant surface strain indicates full transfer of prestress to the concrete. Transfer length is therefore typically determined experimentally by measuring external strain of the concrete member.4–6,9,11,15-17,20,24,26

Instrumentation and measurements

Forms were removed at approximately 18 hours of con-crete age, and the beams were instrumented with demount-able mechanical (DEMEC) gauge points before strand tension was released at approximately 24 hours. Gauge points were placed at the level of the prestressing steel using fast-setting epoxy beginning at 1 in. (25 mm) from each end of the beam and then at 4 in. (100 mm) incre-ments for the first 60 in. (1525 mm) from each end. This allowed for overlapping 8 in. (200 mm) gauge lengths in the area of prestress transfer. Five additional gauge points spaced at 4 in. (100 mm) increments were placed with the center gauge point at the midpoint of each beam to monitor length change in the area of constant prestress.

Beams were stored outside and DEMEC measurements were taken immediately before and after prestress release and at 3, 5, 7, 14, and 28 days using the gauge shown in Fig. 3. Measured concrete surface-strain values were used along with the 95% average maximum strain method9 to determine transfer length for each end of each specimen. This method involves smoothing the data for each side of the member by averaging values over three consecutive gauge lengths and using that value at the center of the three gauge lengths. Results from the two sides are then aver-aged to produce a strain profile along the length of each end of the beam. The values within the strain plateau are averaged, and where 95% of this value crosses the curve is taken as the transfer length. Use of 95% of the average reduces the possibility of bias in determining the beginning

of the strain plateau, as the strain distribution typically does not have a clear transition point. An example of typi-cal strain profiles over time and transfer-length determi-nation using the 95% average maximum strain method at 28 days is shown for specimen NSC-2 in Fig. 4. Most previous research measured transfer length using a similar method.5,6,15,16,17,19,20,33

The standard test for strand bond, now covered under ASTM A1081, was used to examine the bonding quality of the prestressing strand.26,34,35 Each pullout value was com-pared with the preliminary minimum of 10,800 lb (48 kN) recommended during test development, and the average value for the six specimens was compared with the recom-mended 12,600 lb (56 kN).26 Recent research sponsored by PCI36 recommended a minimum average pullout value of 14,600 lb (65kN) for 0.5 in. (13 mm) strands, which could be extrapolated to 17,500lb (77.8kN) for 0.6in. (15 mm) strands.

Figure 3. Demountable mechanical gauge used for measurement of concrete surface strain.

Figure 4. Typical strain profiles over time and transfer-length determination using the 95% average maximum strain method at 28 days for specimen NSC-2. Note: AMS = average maximum strain. 1 in. = 25.4 mm.


ues suggested by Polydorou,36 the measured pullout exceed-ed this value by approximately 10%. The mortar strength exceeded the 5000 psi (34.5 MPa) limit by an average of 70 psi (0.48 MPa) before the pullout testing and 240 psi (1.65 MPa) after the testing. This amounted to an over-strength of slightly more than 1% before and slightly less than 5% after testing. This magnitude of excess strength should not have produced pullout values 50% greater than the specified minimum but should be taken into consider-ation if the larger minimum value is considered.

Table 6 presented the average properties of the fresh and hardened concrete used to cast each set of beams. Ambient temperature and variation in aggregate moisture content required some adjustment to the dosage of high-range water-reducing admixture to meet the specifications for the fresh concrete. Some variations persisted in both the prop-erties of the fresh concrete and the compressive strength (Table 6). The first and last normal-strength limestone mix-tures had poor slump flows, J-ring values, and compressive strengths due to inadequate high-range-water-reducing-admixture dosage and high ambient temperature at the time of batching, respectively. Some surface imperfections resulted from the first batch, used in beam NSL-1, but no difference in transfer length was observed. The final batch exhibited poor consolidation and bond performance of one end of specimen NSL-4. Companion batches were used to measure modulus of elasticity, and the properties of these batches were similar to those for the beam batches.

Results and discussion

The results of the standard test for strand bond, presented in Table 5, indicate that the prestressing strand had adequate bonding quality compared with the preliminary threshold values suggested by Russell26 and research spon-sored by PCI.36 When compared with the values suggested by Russell,26 the pullout loads for the individual specimens exceeded the required 10,800 lb (48 kN) by 50% to 104%, and the average pullout value was 52% greater than the specified 12,600 lb (56 kN). When compared with the val-

Table 5. Standard test for strand-bond results for the strand used to construct beam specimens

Strand specimen

Load at slip, lb Mortar flow

Compressive strength, psi

0.01 in. 0.10 in. 22 hours 23 hours

1 13,560 18,010

112 5070 52702 14,600 19,340

3 11,960 16,170

4 14,170 18,540

100 5070 52105 15,730 21,010

6 16,930 22,020

Average 14,490 19,180 106 5070 5240

Note: 1 in. = 25.4 mm; 1 lb = 4.448 N; 1 psi = 6.895 kPa.

Table 6. Average concrete properties

Specimen NSC NSS-1 NSS 2-5 NSL HSC HSS HSL

Slump flow, in. 26.5 25.5 27.0 25.0 26.5 28.5 26.5

T20, in. 4.3 7.5 4.6 2.6 7.3 3.0 2.4

Visual stability index 0.0 1.5 1.0 0.5 0.5 1.0 1.0

J-ring Δ, in. 2.5 2.0 3.0 2.0 2.0 0.0 0.0

J-ring Δh, in. 1.75 1.75 1.75 1.5 1.5 1.0 0.75

Unit weight, lb/ft3 117.1 119.2 118.6 145.7 121.3 119.3 147.6

fci'

, psi 4420 3400 4210 4740 6210 5960 6910

fc'

, psi 5690 5000 6400 7540 7120 6840 9250

Eci, ksi 2840 n/a 3350 4780 3300 3270 4710

Ec, ksi 3570 n/a 3890 6110 3820 2940 5430

Note: The modulus of elasticity was measured from a companion batch with the same mixture proportions. Ec = modulus of elasticity of concrete at 28 days; Eci = modulus of elasticity of concrete at prestress release; fc

' = 28-day compressive strength of concrete; fci' = compressive strength of

concrete at release of prestress; HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale; n/a = not applicable; NSC = normal-strength clay; NSL = normal-strength limestone; NSS = normal-strength shale; T20 = time required for slump flow to reach diameter of 20 in.; Δ = difference between slump flow and J-ring flow; Δh = difference in height of concrete between inside and outside of J-ring. 1 in. = 25.4 mm; 1 ft3 = 0.0283 m3; 1 lb = 0.454 kg; 1 psi = 6.895 kPa; 1 ksi = 6.895 MPa.


in the strain profiles. The end of the specimens nearest the prestressing abutments had the longer transfer length for all specimens, irrespective of whether it was the live or dead end, but these differences were typically small.

Table 7 presents the 28-day transfer lengths measured for all normal-strength shale specimens and the correspond-ing predictions. The transfer lengths for the live end of specimen NSS-1 and the dead-end and average values for specimen NSS-3 exceeded the values predicted using Eq. (1) and the alternate value 50db in ACI 318-11. The dead-end transfer length for NSS-3 also exceeded alter-nate value of 60db in the AASHTO LRFD specifications. The differences could be accounted for by the variability of the data and were not significant. The average transfer length for the NSS specimens (Table 8) was less than the values calculated using each of the prediction equations. The values computed using Eq. (1) exceeded the average measured value by 16%, and the ACI 318-11 50db predic-tion was 12% greater than the average measured value. The high standard deviation of the dead-end values for the normal-strength shale specimens was a result of the large value for specimen NSS-3 and the small sample size. The other four measured transfer lengths were within a much smaller range, but NSS-3 could not be identified as a pos-sible outlier because of the small number of data points.

The measured moduli of elasticity were in agreement with those calculated using the ACI 318-11 equation.

Some difficulties were encountered in measuring transfer length. Problems with the epoxy early on during testing produced a limited number of inaccurate measurements or lost DEMEC points. Effects of these problems were tempered by the additional step of smoothing the data us-ing overlapping gauge lengths and by taking measurements on both sides of each specimen. The strain profiles for the lightweight SCC specimens were more erratic than those measured for the normalweight SCC specimens, possibly due in part to minor cracking observed at the top fiber after release for lightweight SCC members caused by the high release stresses combined with the low tensile strength of lightweight SCC. The reduced modulus of elasticity com-bined with compressive stresses at or near the ACI 318-11–defined stress limits at release may also have contributed to this erratic behavior. In almost all cases, the strain profiles were clear and allowed for simple determination of the transfer length using the 95% average maximum strain method. Measurements were taken over time with limited variation, so the results at 28 days are presented in this paper.

Normal-strength specimens

Figure 4 is a plot of all measured surface-strain profiles for specimen NSC-2 as representative of the strain profiles for the normal-strength specimens. Variations in strain magnitude were typical for the different concrete types, but the strain-profile shape was similar for each specimen. This plot also shows the 95% average maximum strain values and the transfer lengths measured at 28 days. Table 7 pres-ents the 28-day transfer lengths measured for all normal-strength clay specimens along with the predictions calcu-lated using Eq. (1). Table 7 presents the effective prestress fse used in Eq. (1) and calculated in accordance with the AASHTO LRFD specifications detailed method. The live-end DEMEC transfer lengths for specimens NSC-1 and NSC-3 exceeded the values produced by the ACI 318-11 and AASHTO LRFD specifications equation and the alter-nate value of 50db in ACI 318-11, but the difference can be attributed to the variability of the measurements and should not be taken as significant. All others were significantly shorter than the predicted values.

Table 8 summarizes the average and standard deviation of the eight or ten transfer-length measurements for each beam series. The average transfer length for all normal-strength clay specimens (Table 8) was less than that predicted using all three equations. The value produced by Eq. (1) exceeded the average measured value by 24%, and the ACI 318-11 50db prediction was 20% greater than the average measured value. The standard deviation of greater than 5.5 in. (140 mm) for the DEMEC transfer lengths correlated with the sometimes-erratic behavior observed

Table 7. Normal-strength transfer length at 28 days

Specimen fse, ksi

Transfer length, in. Predicted, in.

Live Dead Averagefsedb

3 in.

NSC-1 153.2 33.1 16.7 24.9 30.6

NSC-2 160.2 22.2 25.6 23.9 32.0

NSC-3 152.0 31.3 19.7 25.5 30.4

NSC-4 154.9 23.2 28.3 25.8 31.0

NSS-1 151.0 30.7 22.8 26.8 30.2

NSS-2 157.3 27.0 24.0 25.5 31.5

NSS-3 158.0 27.4 38.2 32.8 31.6

NSS-4 157.7 29.1 24.0 26.6 31.5

NSS-5 156.1 22.8 22.0 22.4 31.2

NSL-1 169.9 22.8 24.4 23.6 34.0

NSL-2 171.9 18.5 18.9 18.7 34.4

NSL-3 166.4 22.8 23.2 23.0 33.3

NSL-4 168.1 20.5 40.6 n/a 33.6

Note: db = nominal strand diameter; fse = effective prestress after all losses; n/a = not applicable; NSC = normal-strength clay; NSL = normal-strength limestone; NSS = normal-strength shale. 1 in. = 25.4 mm, 1 psi = 6.895 kPa.


the length of the member. The 28-day transfer lengths measured for the high-strength clay specimens (Table 9) were significantly less than those produced by all three of the code provisions. The value computed using Eq. (1) was 74% greater than the average transfer length presented in Table 8, and the ACI 318-11 50db value was 56% greater. The standard deviation of 2.6 in. (67 mm) was much smaller than that for the normal-strength clay and normal-strength shale specimen sets, indicating more consistent results. The live-end transfer lengths at 28 days were greater than those of the dead end for all specimens except HSC-1, and the live-end transfer length was greater than the dead-end transfer length for this specimen at all other ages. These differences can be attributed to the variability in the data even though the difference for specimen HSC-4 seems large.

The measured high-strength shale transfer lengths (Ta-ble 9) were significantly less than those predicted using the code equations. The prediction calculated using Eq. (1) was nearly double the average value measured using surface strain at 28 days (Table 8) and the ACI 318-11 50db prediction was 79% greater than the average value. The standard deviation of 1.9 in. (48 mm) for this set of specimens was substantially less than those of the other specimen sets, indicating more consistent results. The live-end transfer lengths were greater for all specimens except HSS-2, but these differences are small considering the variability in the data.

The high-strength limestone transfer lengths (Table 9) were significantly less than those predicted by the code equations. The prediction calculated using Eq. (1) was 68% greater than the average value measured using surface strain (Table 8) and the ACI 318-11 50db predic-tion was 44% greater than the average measured value. The standard deviation of 2.1 in. (53 mm) was less than

Similarly to the normal-strength clay specimens, the ends of the members near one of the prestressing abutments tended to have a longer transfer length than the ends near the center of the prestressing bed, but these differences were within the variability of the measurements.

Table 7 presents the transfer lengths measured for all normal-strength limestone specimens at 28 days and the corresponding code predictions. These values were all less than those produced by the various prediction equa-tions except for the dead end of specimen NSL-4, which was influenced by poor consolidation caused by extreme temperatures and delays during concrete placement. This extreme value was not included in the average or later statistical analyses. The values predicted using Eq. (1) were greater for the normal-strength limestone specimens than for the lightweight SCC specimens due to the smaller pre-stress losses and, therefore, greater effective prestress. The average normal-strength limestone transfer length (Table 8) was significantly less than the values produced by the code equations. The average prediction calculated using Eq. (1) was 56% greater than the average measured value, and the ACI 318-11 50db prediction was 39% greater than the average measured value. Without including the dead end of specimen NSL-4, the standard deviation was substantially less than those for the normal-strength clay and normal-strength shale specimens. The dead-end transfer length was greater than that of the live end for all normal-strength limestone specimens, and the values were similar for all but specimen NSL-4. These differences were again small and can be explained by the variability of the results.

High-strength specimens

The surface-strain profiles for the high-strength specimens were similar in shape to those of the normal-strength specimens but typically had a reduced magnitude along

Table 8. Average transfer length at 28 days

Beam series

Live end Dead end Total Predicted

Average, in.

Standard deviation, in.

Average, in.


Average, in.


fsedb3 in. 50db, in. 60db, in.

NSC 27.5 5.52 22.6 5.32 25.0 5.65 31.0

30.0 36.0

NSS 27.4 2.96 26.2 6.74 26.8 4.95 31.2

NSL 21.2 2.09 22.2* 2.90* 21.6* 2.30* 33.8

HSC 20.1 3.11 18.2 1.99 19.2 2.63 33.5

HSS 17.9 2.00 15.7 1.03 16.8 1.87 33.2

HSL 20.9 1.40 20.9 2.89 20.9 2.11 35.2

Note: db = nominal strand diameter; fse = effective prestress after all losses; HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale; NSC = normal-strength clay; NSL = normal-strength limestone; NSS = normal-strength shale. 1 in. = 25.4 mm.* Does not include dead end of specimen NSL-4


sion of compressive strength in the transfer-length predic-tion. Further testing is needed to investigate this possibility along with investigating sudden release of prestress and variation in confining reinforcement. Flexural-bond and development-length behavior of these members are to be covered in an additional paper by the authors.

Figure 6 shows the 95% confidence intervals for the mean of the normal-strength clay, normal-strength shale, and normal-strength limestone surface-strain transfer lengths. The data points at each time increment have been slightly offset to show the confidence intervals. The overlap of these confidence intervals indicates no significant difference between the normal-strength clay and normal-strength shale specimen sets or between the normal-strength clay and the normal-strength limestone specimen sets, even though the normal-strength limestone transfer lengths were shorter in all cases. A significant difference was observed between the normal-strength shale and normal-strength limestone trans-fer lengths at all ages except 28 days. The high standard deviations for the normal-strength clay and normal-strength shale specimen sets compared with the normal-strength limestone, high-strength clay, high-strength shale, and high-strength limestone specimens (Table 8) were attributed to the lower compressive strength at release and a reduced stiffness compared with the normal-strength limestone and high-strength limestone mixtures. The high standard devia-tions for these small specimen sets contributed to the lack of a statistically significant difference. The average normal-strength clay transfer length was nearly 16% greater and the normal-strength shale 24% greater than the normal-strength limestone transfer length at 28 days. This difference could be explained by the substantially lower modulus of elastic-ity for the normal-strength clay and normal-strength shale mixtures. The measured normal-strength clay modulus was 41% (1940 ksi [13,400 MPa]) and normal-strength

that of all the normal-strength specimen sets. The average high-strength limestone transfer length was 8% and 24% greater than those measured for the high-strength clay and high-strength shale specimens, respectively.

Discussion

Figure 5 compares the average transfer lengths with the code equation predictions. Similarity between the average normal-strength clay and normal-strength shale values and a difference of both from the smaller normal-strength lime-stone average value are apparent. The high-strength clay and high-strength shale specimens also have similar average transfer lengths but have values less than the average value for the high-strength limestone specimens. The longer average transfer lengths measured for the normal-strength clay and normal-strength shale specimens would lead to a lower compressive stress than used for design at the ends of members, which would lead to reduced shear capacity in a material with an already reduced shear strength resulting from the low tensile strength of lightweight concrete. The shorter average transfer lengths measured for the high-strength clay and high-strength shale specimens would lead to increased tensile stresses at the ends of members, which would likely cause cracking, and increased compressive stresses that could exceed allowable limits. The variation in average transfer length with compressive strength for the lightweight SCC specimens, not seen for the normalweight concrete specimens, indicates a greater need for the inclu-

Figure 5. Surface-strain transfer lengths and code predictions for all beam series at 28 days. Note db = nominal strand diameter; HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale; NSC = normal-strength clay; NSL = normal-strength limestone; NSS = normal-strength shale. 1 in. = 25.4 mm.

Table 9. High-strength transfer length at 28 days

Specimen fse, ksi

Surface strain, in. Predicted, in.

Live Dead Averagefsedb

3 in.

HSC-1 167.4 16.5 18.9 17.7 33.5

HSC-2 166.8 19.3 15.7 17.5 33.4

HSC-3 167.9 20.7 20.5 20.6 33.6

HSC-4 167.4 24.0 17.7 20.9 33.5

HSS-1 166.4 18.3 15.9 17.1 33.3

HSS-2 164.5 15.9 16.3 16.1 32.9

HSS-3 163.3 20.5 16.3 18.4 32.7

HSS-4 170.2 16.7 14.2 15.5 34.0

HSL-1 176.0 22.8 19.3 21.1 35.2

HSL-2 176.4 20.9 25.2 23.0 35.3

HSL-3 176.3 20.5 19.3 19.9 35.3

HSL-4 176.2 19.5 19.7 19.6 35.2

Note: db = nominal strand diameter; fse = effective prestress after all losses; HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale. 1 in. = 25.4 mm; 1 ksi = 6.895 MPa.


high-strength shale mixtures had a substantially lower inherent stiffness than that of the high-strength limestone mixture, as indicated by modulus-of-elasticity values ap-proximately 30% (1400 ksi [9650 MPa]) less than that of the high-strength limestone mixture at prestress release and 30% (1600 ksi [11,000 MPa]) and 45% (2500 ksi [17,200 MPa]) less at 28 days for the high-strength clay and high-strength shale mixtures, respectively. This lower elastic modulus should allow more deformation of the concrete caused by the outward pressure due to the Hoyer effect, thus reducing the contribution of this mechanism to the transfer bond. The mild-steel confining reinforcement provided for shear mitigated this effect.

Transfer-length performance at a specific compressive strength was similar for all three aggregate types (Fig. 6 and 7). Differences in modulus of elasticity among con-cretes made with any two of the three aggregates at the same strength varied from 1% for the high-strength clay and high-strength shale mixtures at one day to as much as 45% for the high-strength limestone and high-strength shale mixtures at 28 days. Despite the large effects on concrete stiffness, the lack of statistically significant dif-ferences among transfer lengths for the expanded clay and shale and normalweight limestone aggregates indi-cates that aggregate type was not a significant variable for transfer length. All mixtures were proportioned with a high cementitious-materials content to produce SCC behavior. The inclusion of fly ash in the high-strength clay and high-strength shale mixtures did not adversely affect the bond of these mixtures compared with the normal-strength clay, normal-strength shale, or high-strength limestone mix-tures, which included only portland cement.

Conclusion

The research described herein examined the difference in transfer-bond performance, as measured by transfer

shale modulus 30% (1430 ksi [9860 MPa]) less than that of the normal-strength limestone mixture at 1 day of age. The normal-strength clay and normal-strength shale moduli of elasticity were 42% (2540 ksi [17,500 MPa]) and 36% (2220 ksi [15,300 MPa]) less than that of the normal-strength limestone mixture at 28 days.

The reduced stiffness of the normal-strength clay and normal-strength shale mixtures leads to more outward de-formation of the concrete as the diameter of the prestress-ing strands increases during prestress release, thus reducing the normal stress and friction forces in the transfer zone re-sulting from the Hoyer effect.11 The smaller normal stresses would lead to a reduction in the strand anchorage potential over the transfer length and longer transfer lengths for the normal-strength clay and normal-strength shale specimens compared with the normal-strength limestone specimens. The mild steel used as shear reinforcement in this study provided confinement that mitigated this effect.

Figure 7 shows the 95% confidence intervals for the mean for the high-strength clay, high-strength shale, and high-strength limestone specimens. The overlap of these inter-vals indicates no significant differences among the transfer-length measurements for any of the specimen sets made with concrete mixtures having a compressive strength of 6000 psi (41 MPa) at prestress release. The smaller stan-dard deviations measured for the high-strength specimen sets indicate greater consistency in the measurements than for the normal-strength clay and normal-strength shale specimen sets. These smaller standard deviations were attributed to the greater compressive strength at prestress release because the concrete stiffness was very similar be-tween the normal-strength and high-strength specimen sets made with the same aggregate type. The higher compres-sive strength may allow the concrete to better withstand the outward pressure of the prestressing strands and sustain the Hoyer effect.11 Conversely, the high-strength clay and

Figure 7. Average surface-strain transfer lengths and 95% confidence intervals for high-strength specimens. Note: HSC = high-strength clay; HSL = high-strength limestone; HSS = high-strength shale.

Figure 6. Average surface-strain transfer lengths and 95% confidence intervals for normal-strength specimens. Note: NSC = normal-strength clay; NSL = normal-strength limestone; NSS = normal-strength shale.


The results of this research do not necessarily indicate the need for a modification to the transfer-length-prediction equations, as the conditions were in some ways the best case, but measured differences suggest the need to include the effects of compressive strength in prediction of transfer length for lightweight SCC members. A shorter predicted value may also be necessary to prevent unexpectedly high stresses near member ends. The need for further testing, in-cluding sudden release of prestress, variation in confining reinforcement, and testing of full-scale lightweight SCC members, is apparent.

Acknowledgments

The researchers acknowledge the Arkansas State Highway and Transportation Department and the Mack Blackwell Rural Transportation Center for providing financial support for the research. The researchers also thank Grace Construc-tion Products, Buildex Inc., and Big River Industries Inc. for providing the materials used in this study. The views and opinions expressed in this article do not reflect those of the research sponsors or those of the material providers.

References

1. ACI (American Concrete Institute) Committee 213. 2003. Guide for Structural Lightweight-Ag-gregate Concrete. ACI 213R-03. Farmington Hills, MI: ACI.

2. Bonen, D., and S. Shah. 2005. “Fresh and Hardened Properties of Self-Consolidating Concrete.” Progress of Structural Engineering Materials 7 (1): 14–26.

3. Wall, J. 2010. “Non-Traditional Lightweight Concrete for Bridges, A Lightweight Aggregate Manufacturer’s Review of Current Practice.” In Proceedings 2010 Concrete Bridge Conference: Achieving Safe, Smart, and Sustainable Bridges. Skokie, IL: Portland Cement Association.

4. Burgueño, R., and M. Haq. 2005. “Development Length of Prestressing Strands in Precast/Prestressed Girders Using Self Compacting Concrete.” In Pro-ceedings of the 2005 Structures Congress and the 2005 Forensic Engineering Symposium, Metropolis and Beyond, 1673–1680. Reston, VA: American Soci-ety of Civil Engineers.

5. Girgis, A., and C. Tuan. 2005. “Bond Strength and Transfer Length of Pretensioned Bridge Girders Cast with Self-Consolidating Concrete.” PCI Journal 50 (6): 72–87.

6. Hegger, J., S. Bülte, and B. Kommer. 2007. “Struc-tural Behavior of Prestressed Beams Made with Self-consolidating Concrete.” PCI Journal 52 (4): 34–42.

length, of prestressing strands cast in prestressed con-crete members containing lightweight SCC and normal-weight SCC. Measured values were also used to examine the applicability of the ACI 318-11 and AASHTO LRFD specifications code predictions for transfer length. Aver-age transfer lengths measured using concrete surface strain for lightweight SCC members having a compres-sive strength of 4000 psi (28 MPa) at prestress release were shorter than those predicted using the ACI 318-11 and AASHTO LRFD specifications equations. Average transfer lengths for lightweight SCC members having a compressive strength of 6000 psi (41 MPa) at prestress release were substantially less than those predicted using the ACI 318-11 and AASHTO LRFD specifications equa-tions. The measured high-strength shale transfer lengths were approximately half of the predicted values.

The average transfer lengths measured from the expand-ed clay lightweight SCC and normalweight SCC speci-mens with the same compressive strength at prestress release did not exhibit significant differences in spite of the differences in aggregate type. The expanded shale lightweight SCC and normalweight SCC specimens with a compressive strength of 4000 psi (28 MPa) at release did exhibit a significant difference; however, the mix-tures with a 6000 psi (41 MPa) compressive strength at release did not. This result indicates that aggregate type was not a significant factor affecting transfer length. Variability, indicated by high standard deviations, may have masked differences that were noticeable in the raw data for the specimens having a compressive strength of 4000 psi (28 MPa) at release. Average measured transfer lengths for the high-strength clay specimens were 23% shorter than those measured for the normal-strength clay specimens and high-strength shale transfer lengths were 37% shorter than the normal-strength shale trans-fer lengths, indicating an effect of concrete compressive strength at both release and 28 days on transfer length in lightweight SCC members. Average transfer lengths for the normal-strength limestone and high-strength lime-stone specimens did not exhibit significant differences.

Higher compressive strength at release produced much more consistent transfer-length results, as indicated by the standard deviation, and these results did not exceed the code-predicted values. Sudden prestress release in practice may increase the scatter in results for both strengths. Expected detrimental effects on transfer length caused by reduced concrete stiffness were not evident, but the beam specimens had the shear and compression reinforcement required for subsequent development-length testing. The shear stirrups produced confinement in the transfer zone, resulting in shorter transfer lengths. No differences in performance based on mixture com-position were observed; the mixtures including fly ash performed similarly to those containing only portland cement.


Concrete Prestressed Girders and Panels.” Report FHWA/TX/1852-1. Austin, TX: Center for Transpor-tation Research.

20. Nassar, A. 2002. “Investigation of Transfer Length, Development Length, Flexural Strength and Prestress Loss Trend in Fully Bonded High Strength Light-weight Prestressed Girders.” MS Thesis, Virginia Poly-technic Institute and State University, Blacksburg, VA.

21. Meyer, K., and L. Kahn. 2004. “Transfer and Devel-opment Length of 0.6 Inch Strand in High Strength Lightweight Concrete.” In High-Performance Struc-tural Lightweight Concrete, 9–28. SP-218. Farmington Hills, MI: ACI.

22. Larson, K., R. Peterman, and A. Esmaeily. 2007. “Bond Characteristics of Self-Consolidating Concrete for Prestressed Bridge Girders.” PCI Journal 52 (4): 44–57.

23. Logan, D. 1997. “Acceptance Criteria for Bond Qual-ity of Strand for Pretensioned Prestressed Concrete Applications.” PCI Journal 42 (2): 52–90.

24. Rose, D., and B. Russell. 1997. “Investigation of Stan-dardized Tests to Measure the Bond Performance of Prestressing Strand.” PCI Journal 42 (4): 56–80.

25. Peterman, R., J. Ramírez, and J. Olek. 2000. “Design of Semi-lightweight Bridge Girders: Development Length Considerations.” Transportation Research Record 1696: 41–47.

26. Ramírez, J., and B. Russell. 2008. Transfer, Develop-ment, and Splice Length for Strand/Reinforcement in High-Strength Concrete. NCHRP (National Coopera-tive Highway Research Program) report 603. Wash-ington, DC: Transportation Research Board.

27. Floyd, R., E. Ruiz, N. Do, B. Staton, and W. Hale. 2011. “Development Lengths of High-Strength Self-Consolidating Concrete Beams.” PCI Journal 56 (1): 36–53.

28. ASTM C1611/C1611M. 2009. Standard Test Method for Slump Flow of Self-Consolidating Concrete. West Conshohocken, PA: ASTM International.

29. ASTM C1621/C1621M. 2009. Standard Test Method for Passing Ability of Self-Consolidating Concrete by J-Ring. West Conshohocken, PA: ASTM International.

30. ASTM C138/C138M. 2010. Standard Test Method for Density (Unit Weight), Yield, and Air Content (Gravi-metric) of Concrete. West Conshohocken, PA: ASTM International.

7. ACI Committee 318. 2011. Building Code Require-ments for Structural Concrete (ACI 318-11) and Com-mentary (ACI 318R-11). Farmington Hills, MI: ACI.

8. Hanson, N., and P. Kaar. 1959. “Flexural Bond Tests of Pretensioned Prestressed Beams.” Journal of the American Concrete Institute 55 (7): 783–802.

9. Russell, B., and N. Burns. 1996. “Measured Transfer Lengths of 0.5 and 0.6 in. Strands in Pretensioned Concrete.” PCI Journal 41 (5): 44–65.

10. PCI Industry Handbook Committee. 2010. PCI Design Handbook: Precast and Prestressed Concrete. 7th ed. Chicago, IL: PCI.

11. Janney, J. 1954. “Nature of Bond in Pretensioned Pre-stressed Concrete.” Journal of the American Concrete Institute 50 (9): 717–736.

12. AASHTO (American Association of State Highway and Transportation Officials). 2007. LRFD Bridge Design Specifications. 4th ed. Washington, DC: AASHTO.

13. Peterman, R. 2007. “The Effects of As-Cast Depth and Concrete Fluidity on Strand Bond,” PCI Journal 52 (3): 72–101.

14. Lachemi, M., S. Bae, K. Hossain, and M. Sahmaran. 2009. “Steel-Concrete Bond Strength of Lightweight Self-Consolidating Concrete.” Materials and Struc-tures 42 (7): 1015–1023.

15. Trent, J. 2007. “Transfer Length, Development Length, Flexural Strength, and Prestress Loss Evalu-ation in Pretensioned Self-Consolidating Concrete Members.” MS dissertation, Virginia Polytechnic Institute and State University, Blacksburg, VA.

16. Staton, B., N. Do, E. Ruiz, and W. Hale. 2009. “Trans-fer Lengths of Prestressed Beams Cast with Self-Con-solidating Concrete.” PCI Journal 54 (2): 64–83.

17. Ward, D., R. Floyd, and W. Hale. 2009. “Bond of 0.5 in. Diameter Strands Cast in Lightweight SCC.” In Proceedings PCI-FHWA National Bridge Conference, San Antonio, TX. Chicago, IL: PCI. CD-ROM.

18. Naito, C., G. Brunn, G. Parent, and T. Tate. 2005. “Comparative Performance of High Early Strength and Self-Consolidating Concrete for Use in Precast Bridge Beam Construction.” ATLSS report 05-03. Bethle-hem, PA: Advanced Technology for Large Structural Systems.

19. Thatcher, D., J. Heffington, R. Kolozs, G. Sylva III, J. Breen, and N. Burns. 2002. “Structural Lightweight


Notation

db = nominal strand diameter

Ec = modulus of elasticity of concrete at 28 days

Eci = modulus of elasticity of concrete at prestress release

fc' = 28-day compressive strength of concrete

fci'

= compressive strength of concrete at release of prestress

fse = effective prestress after all losses

lt = transfer length

T20 = time required for slump flow to reach diameter of 20 in. (500 mm)

w/c = water-cement ratio

w/cm = water–cementitious materials ratio

Δ = difference between slump flow and J-ring flow

Δh = difference in height of concrete between inside and outside of J-ring

31. ASTM C39/C39M. 2011. Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens. West Conshohocken, PA: ASTM International.

32. ASTM C469/C469JM. 2010. Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression. West Conshohocken, PA: ASTM International.

33. Meyer, K., and L. Kahn. 2002. “Lightweight Concrete Reduces Weight and Increases Span Length of Preten-sioned Concrete Bridge Girders.” PCI Journal 47 (1): 68–75.

34. Sobin, N. 2005. “Evaluation of Strand Bond Assur-ance Tests for Pretensioned Applications.” MS thesis, University of Arkansas, Fayetteville, AR.

35. ASTM A1081/A1081M. 2012. Standard Test Method for Evaluating Bond of Seven-Wire Steel Prestressing Strand. West Conshohocken, PA: ASTM International.

36. Polydorou, T. 2014. “Determination of Acceptance Criteria for Prestressing Strand in Pre-tensioned Appli-cations.” PhD diss., Kansas State University, Manhat-tan, KS.


About the authors

Royce W. Floyd is an assistant professor in the School of Civil Engineering and Environmental Science at the University of Oklahoma in Norman, Okla. He received his bachelor’s degree and PhD from the University of

Arkansas in Fayetteville, Ark. His research inter-ests include concrete materials and prestressed concrete.

W. Micah Hale is a professor in the Department of Civil Engineer-ing at the University of Arkansas in Fayetteville. He received his bachelor’s and master’s degrees and PhD from the University of Oklahoma and is a licensed

professional engineer in Arkansas. His research interests include concrete materials, mixture propor-tioning, and prestressed concrete.

Michael B. Howland is a U.S. Navy Civil Engineering Corps officer and works as a construc-tion manager for Naval Facilities Engineering Command Mid-Atlantic at Naval Station Norfolk in Norfolk, Va. He received his

bachelor’s degree from the University of Arkansas.

Abstract

Lightweight self-consolidating concrete (SCC) has the potential to increase transfer length. Twenty-five rectangu-lar concrete members prestressed with two 0.6 in. (15 mm) prestressing strands were cast using SCC made with three different coarse aggregates. The concretes had design compressive strengths of either 4000 and 6000 psi (28 and 41 MPa) or 6000 and 7000 psi (41 and 48 MPa) at release and 28 days, respectively. Transfer lengths were not sig-nificantly affected by aggregate type, but higher compres-sive strength at release resulted in shorter transfer lengths. Average transfer lengths were shorter than predicted by the American Concrete Institute andAmerican Associa-tion of State Highway and Transportation Officials codes. Lightweight SCC exhibited very similar performance to normalweight SCC and the code equations.

Keywords

Lightweight; self-consolidating concrete; strand bond; transfer length.

Review policy

This paper was reviewed in accordance with the Precast/Prestressed Concrete Institute’s peer-review process.

Reader comments

Please address any reader comments to [email protected] or Precast/Prestressed Concrete Institute, c/o PCI Journal, 200 W. Adams St., Suite 2100, Chicago, IL 60606. J

measured transfer length of 0.6 in. prestressing strands ... journal/2015/may/measured... ·...

Documents