Professional Documents
Culture Documents
Connections in Precast Concrete Structures - Effects of Restrained Creep and Shrinkage
Connections in Precast Concrete Structures - Effects of Restrained Creep and Shrinkage
Concrete Structures
Effects of Restrained Creep
and Shrinkage
by K. T. Burton, W. G. Corley and E. Hognestad*
SYNOPSIS
This paper reports an experimental and analytical investigation of the
influence of creep and shrinkage deformations on precast-prestressed con-
crete T-beams with their ends restrained against horizontal movement. The
tests demonstrated that shrinkage and creep of the concrete introduces
horizontal forces in such members. Depending upon the degree of restraint,
these forces may be of such magnitude that their inclusion in design of
beam end details, column supports, and corbels is required. An approximate
method for estimating the magnitude of these induced forces is presented.
procedures are evaluated for compu- tion while still restrained in the same
tation of the forces acting on such way they were during the 600-day
beams and their supports. Methods period. In this manner it was feasible
that provide a conservative estimate to evaluate possible effects on flexu-
of the magnitude of the induced ral stiffness and strength of forces
forces are presented. induced by creep and shrinkage.
EXPERIMENTAL STUDY Preliminary Study
The experimental work was car- It was anticipated that horizontal
ried out to determine the effects of tensile forces of considerable magni-
full and partial restraint on pre- tude would develop during the test
stressed members. Five identical T- period. Severe cracking of the con-
beams were included in the study. crete near the supports as shown in
Two of these were mounted on sup-
ports that could be adjusted to main-
tain full restraint against shortening
of the beams. Another two were
mounted on flexible columns that
provided only partial restraint. The Reinforcing Bar
fifth beam was supported on rollers Welded To Plate
April 1967 19
Fig. 1 was also expected. Conse- and FR2, PR1 and PR2, and NR1,
quently, an anchorage system in- respectively, were each supported
tended to prevent horizontal slip be- over a clear span of 24 ft. Cross-sec-
tween the test beams and their end tional properties are shown in Fig. 3.
bearing plates was tested. All beams were prestressed by two
The anchorage system shown in 3/8 -in. diameter, 7-wire strands hav-
Fig. 2 was evaluated by means of ing a nominal breaking stress of 270
"pull-off" tests. These tests were car- ksi. Each strand was tensioned to
ried out on 9-ft. prestressed T-beams provide a force before release of
with cross-sectional properties simi- about 28,000 lb. Measured force in
lar to those used in the main investi- the strand before transfer and com-
gation of restraint forces. Various puted forces after transfer are listed
sizes of reinforcing bars were evalu- for each beam in Table II. All
ated. To ensure adequate embed- strands were straight throughout the
ment of the bar, its length was taken length of the beams.
equal to the transfer length of the With each beam specimen, nine
strand s > plus twice the embedment 6 x 12-in, cylinders were cast from
length required by the Ultimate which the 7-, 28-, and 300-day
Strength Design Section of the 1963 strengths were determined. The cyl-
ACI Building Code. inders were stored with the test
Results of the "pull-off" tests are specimens so that they were always
listed in Table I. In each specimen a subject to the same conditions of
force greater than the yield strength temperature and relative humidity.
of the bar was developed. Conse- In addition to the test beams, two
quently, each of the anchorage sys- 10-ft. long non-prestressed shrinkage
tems tested could be used to resist a beams, designated FS1 and F'S2, and
force equal to the bar yield strength. four 6 x 66-in, control cylinders were
An end anchorage system made of a cast. Cross-sectional properties of
No. 6 bar with l = 68 in. was se- the shrinkage beams are also shown
lected to resist the anticipated hori- in Fig. 3. Two of the cylindrical
zontal force resulting from creep and specimens were maintained under a
shrinkage in a fully restrained pre- constant stress to determine creep
stressed T-beam. properties of the concrete. The other
two were maintained without exter-
Test Specimens
nal load to determine shrinkage
The fully-, partially-, and non-re- strains. These cylindrical specimens
strained test beams, designated FRI had a volume-to-surface ratio almost
20 PCI Journal
Beam Symmetrical About It
C^9
3 No.3 Bars
_p
=co in
C.G. Beam ^^
r
o
i " 4 No.3
-N
C. G. Prestress
the same as that of the T-beams. the test floor the day before the
Measured shrinkage strains from the beams were mounted in position.
cylindrical specimens were com- This load was applied by means of a
pared with those of the T-beams to threaded steel rod passing through a
determine if members of different conduit in the center of each col-
shapes but having about the same umn.
volume-to-surface ratio would exhib- Two short columns supported the
it similar creep and shrinkage char- unrestrained test beams. One end of
acteristics. The results of this com- each beam was supported on a half-
parison were reported as a part of a
more comprehensive investigation
Steel Plate welded to Column Remtolce,,W
into this shape relationship.«) C<nte, of 0,0,010 For,,
April 1967 21
round while the other was supported half of a standard double-tee steel
on a roller. This permitted free rota- form long enough to accommodate
tion at both ends and free horizontal two 24-ft. 5-in, specimens placed end
translation at one. to end. The web and longitudinal
Each free shrinkage T-beam speci- reinforcement was tied to the strand
men was supported at 5-in, centers after it was tensioned.
along its 10-ft. length. Rubber pads The strands were tensioned indi-
were placed between the beam and vidually using a center-hole ram.
steel bearing plates. The bearing Force applied by the ram was moni-
plates were supported on rollers rest- tored by a load cell placed between
ing on another set of plates attached the ram and the temporary anchor-
to the test floor. This arrangement age used during this operation. Each
provided approximately uniform strand was overtensioned by a small
support for the shrinkage beams. amount to compensate for the draw-
The prestressing steel used in the in as the permanent anchorage took
restrained and non-restrained test up the load. The force in the strands
specimens had a yield stress at one after the load had been transferred
percent elongation of 255,000 psi, to the permanent anchor was meas-
and a strength of 274,000 psi. The ured by a load cell placed between
modulus of elasticity was about 29 the reaction block and the anchor.
million psi. All strand was free of Table II lists the measured value of
loose rust and surface oil before ten- the strand force just prior to transfer.
sioning. In each specimen the measured
Unstressed longitudinal reinforce- force was close to the intended value
ment consisted of intermediate of 28,000 lb. which corresponds to a
grade No. 3 deformed bars. Straight strand stress of 175,000 psi. Also
bars were placed longitudinally in listed in Table II is the calculated
the flange of the test beams and in strand force just after transfer.
hoth the flange and the stem of the The first two test beams, compan-
shrinkage beams. Both web rein- ion specimens subjected to full re-
forcement and lateral reinforcement straint, were cast simultaneously the
in the flange were made from No. 2 day after the strands were tensioned.
bars bent in the form of a T. Fig. 3 The beams were covered with plas-
shows the location and size of each tic and cured for three days in the
bar in the T-beam specimens. The b- forms. They were then removed
in, control cylinders contained no re- from the forms and stored at 70°F
inforcement. and 50% relative humidity. Seven
All concrete was made from a days after casting, prestress was
blend of Type I Portland cement transferred to the beams by torch-
and Elgin sand and gravel of a/4- cutting the strands. Measurements of
in. maximum size. Each batch con- beam deflection were made after
tained 4 to 5% entrained air. The transfer of prestress. On the same
concrete strengths listed in Table II
7th day the two beams were se-
are the average of three 6 x 12-in.
cured to their supports and the creep
cylinders taken as the beams were
cast. and shrinkage observations were be-
gun. This procedure was followed
Fabrication for all T-beam specimens.
The precast T-beams were pro- The four 6 x 66-in, creep and
duced in the laboratory using one shrinkage specimens were each cast
22 PCI Journal
I-
1 FR denotes full restraint; PR, partial restraint; NR, non-restrained; FS denotes free shrinkage T-beams
2 Cols. 1, 2 supported Beam PR1, Cols. 3, 4 supported PR2
3 Based on design calculations allowing for loss of prestress caused by creep and shrinkage of concrete and relaxation of steel
Beam
—Gage Line I
— 2
— 5
10"
in three 22-in, long sections and change in length between brass ref-
cured in the same manner as the T- erence points cemented to the sur-
beams. At an age of 7 days, the sec- face of the concrete near midspan of
tions were stacked and cemented to- the beams. Locations of the gage
gether in their locations to form the lines are shown in Fig. 5.
66-in, long specimen. Two of the Deflections were measured by a
columns were loaded to obtain a Wild N2 precision level sighted on
stress of about 1000 psi so that the engraved scales placed against ref-
combined creep and shrinkage de- erence points on top of the beam. A
formations could be observed. The detailed explanation of this proce-
other two were held lightly in a ver- dure is given elsewhere.( 8 ) End rota-
tical position so that shrinkage tions were measured by Ames dial
strains under zero external load gages bearing against reference
could be observed. points attached to the ends of the
beams.
Instrumentation
Horizontal forces that developed
Concrete strains were measured in the restrained specimens were ob-
by a 10-in, gage length Whittemore tained using the reaction pedestal
mechanical strain indicator. This in- shown in Fig. 6. For the fully re-
dicator was used to determine strained beams, these pedestals were
=,t7
0
"6 I /6 i'6 ^6
R R
April 1967 25
800
Theoretical "Rate Of Creep" Curve
z
Beam PR 2
0
N
o: 200
0
x
were applied at midspan so that the the increase was slight, indicating
mechanical strain gage reference that the forces were approaching
points located in that area could be asymptotic values.
used to measure concrete deforma- Partial Restraint. Fig. 8 indicates
tions during the test to destruction. that the build-up of horizontal forces
The loads were increased by in- for the companion specimens under
crements until failure occurred. Af- partial restraint followed similar,
ter each load stage was reached, though not identical, curves. The fi-
concrete strains and deflections were nal recorded magnitude of horizon-
measured. Since the beams were tal force for Beam PR1 was 505 lb.
permitted to rotate at the ends, the and for Beam PR2, 425 lb.
increasing vertical loads provided The 16% difference between the
relief of load on the cells used to re- final recorded forces for the two
cord the horizontal forces induced beams was evident during the early
by creep and shrinkage. This grad- stages of observation. Initially the
ual reduction in measured horizontal force built up more rapidly in Beam
force was recorded. PR1. After the first 100 days, the rate
TEST RESULTS of increase was about the same for
both beams.
Long-Term Observations These two beams, and their sup-
The increases in horizontal forces porting columns, were not cast si-
with time for the partially- and fully- multaneously as were the specimens
restrained test beams are plotted in under full restraint. Although an ef-
Figs. 8 and 9. During the early fort was made to ensure that they
stages of observation, there was a had similar properties, it is seen in
rapid build-up of horizontal forces. Table II that the concrete strengths
The build-up became more gradual of Beam PRI were lower than those
as the specimens approached an age attained by its companion specimen.
of about 200 days. After this time, Similarly, the initial camber in Beam
26 PCI Journal
PRI was greater than that in Beam days, the ratio of creep to elastic
PR2. strain, 0, approached 2.0, while the
In addition to the recorded test shrinkage strain approached 0.5% as
data, the increase in horizontal shown in Fig. 10.
forces with time was calculated us- Calculated forces for the partially
ing two different methods to account restrained beams used in this study
for the effects of creep and shrink- are shown in Fig. 8. Both the "Rate
age. These two methods, the "Rate of Creep" method and the "Effective
of Creep" and the "Effective Modu- Modulus" method provided a con-
lus," have been discussed by Ross("". servative estimate of the restraint
A simplified numerical procedure for force. Neither method accurately
applying the "Rate of Creep" meth- predicted the behavior of the par-
od is given in Ref. 10. tially restrained test specimens. In
The analyses used to calculate the both cases, the theoretical forces
time-dependent effects followed the were higher than those measured.
treatment presented by Mattock in Full Restraint. Fig. 9 shows that
Ref. 11. Both elastic and creep prop- the increase in horizontal force with
erties were assumed to be the same time under full restraint was nearly
for the beams as for the columns. identical for the companion speci-
The restraint forces were computed mens. Note that the force scales in
by the methods of elasticity, includ- Figs. 8 and 9 differ greatly. The de-
ing shrinkage but neglecting creep. velopment of these forces took place
The effects of creep were then intro- at a somewhat lower rate than would
duced by multiplying the elastic so- be expected by extrapolating from
lution by the functions of the creep the curves of the partially restrained
coefficient, c), discussed in Appendix specimens. This effect is seen as a
1 of Ref. 11. Numerical values for sharper break in the force-time
shrinkage and creep as a function of curve for the partially-restrained
time were obtained from the cylin- specimens than in that of the fully-
drical specimens. At an age of 600 restrained specimens.
25,000
Theoretical °Rate Of Creep' Curve
m
J 20,000 Theoretical "Effective Modulus" Curve
w
V
0
u_
15,000
Beam FR
J
Beam FR 2
Z 10,000
Cracking
0
N observed
a:
0 5,000
24 —0 /C Supports —
Reset End
---//2Ij
Bearing Plate
FACE BEAM FRI
0
I I I
Scale — Feet
Reset End
/ I
28 PCI Journal
0
0.2
I)
Iii
I
U
Z ^
- 0.1 o Unrestrained Beam
C a
r
w D o Beam PRI
m
• Beam PR2
U 0
Z
a o Beam FR
a_
in Beam FR 2
0.1
Z o
FR2
w a
cD 3
ZZ
=0 3 0.2
U 0 FR1
0 100 200 300 400 500 600 700
TIME , DAYS
900
800
F Gage Line 5
z 700
0
J
600
2
Z 500
400
Co
Q 300
H
0
-
1 200
30 PCI Journal
o e p A
p o
x 0 ^p
Z 400 p
0_ G
J
J
2
• 300 p
z
a
• Free Shrinkage T- Beam V/S = 1.46
o 6" Dia. Cylinder V/S = 1.50
W 200
4 A Top of Flange of Non-restrained
z
2 T-Beom V/S = 1.46
I
100
Fig. 14—Comparison of Shrinkage Strain for Test Specimens with Nearly the Same V/S
a, 600 r_
x Unrestrained T— Beams
I
Z 500
0 8. °• o• • o c•
400
300
200 • Beam FR I
a) ° Beam FR 2
• Beam PR I
100 o Beam PR 2
0
f-
0 100 200 300 400 500 600 700
TIME, DAYS
(a)STRAINS AT TOP OF FLANGE
800
Unrestrained T— Beams --
700
U)
Z 600
0
500
f
400 o °
Z_
Of 300
°
Q 200 • Beam FR I
° Beam FR 2
100 • eam P R I
Beam PR 2
B
32 PCI journal
12,500
m
J
c 10,000
Q
0
J
J
Q V
v 7,500
F o Unrestrained Beam
W o Beam PRI
0 • Beam PR 2
5,000 Service Load 0 Beam FR I
a A Beam FR 2
a.
p 2,500
0 I 2 3 4 5 6 7 8 9 10 11 12 13 14 15
DEFLECTION, INCHES
in Table III. Each beam failed in Shear Strength. The test beams of
flexure by fracture of the prestress this investigation were all adequate-
strands. Fig. 17 shows the central ly reinforced for shear. Had no stir-
portion of one beam after failure. rup reinforcement been provided,
Test moments for all beams ex- and particularly had the beams been
ceeded the calculated ultimate mo- reinforced without prestress, the
ments. The calculated ultimate mo- large horizontal tensions developed
ments were computed from the under full restraint could have re-
measured properties of the concrete duced shear strength substantially.
formations did `- not influence their Fig. 17—Typical Flexural Failure of Restrained
flexural strength. T-Beam after Vertical Load Tests
April 1967 - 33
TABLE III—Load Tests to Destruction
A'2 n
'2
RO
H Ecic
2 3 4 i
in
ii
(d)
Fig. 18—Single and Multi-Span Frames
April 1967 35
36
o-
34 e
O ^
32 co h
30
28
c
26 //c/
V.
24
22
ti
20
RI O`
18 Ai
16
14
12
0
O•
I0
,`L
0 I 2 3 4 5 6 7 8 9 10 11 I2
NUMBER OF SPANS,n
Fig. 19—Ratio of Ri/Ro for Any Span in Multi-Span Frame
If for any reason it would be im- ends are restrained against horizon-
practical to follow these suggestions, tal movement will create horizontal
an estimate of the anticipated forces forces that, depending on the degree
due to restrained creep and shrink- of restraint, may cause severe crack-
age should be made by the simpli- ing in the support regions. For cases
fled method outlined here or by one of severe restraint, therefore, these
of the two methods described else- forces must be included in design of
where f9.10, ' 1 Adequate provision for beam end details, column heads, cor-
horizontal forces must be made in bels, and other supports.
the design of beam end details. The "Rate of Creep" and "Effec-
tive Modulus" methods of calculat-
CONCLUSIONS ing these forces will yield reason-
It is concluded that creep and able, though conservative, estimates
shrinkage deformations of precast- of the forces created by creep and
prestressed concrete members whose shrinkage in members that are re-
36 PCI Journal
strained against horizontal move- No. 1, Feb. 1965, pp. 16-61; PCA De-
ment. velopment Department Bulletin D85.
5. LaFraugh, R. W. and Magura, D. D.,
Under the conditions imposed in "Connections in Precast Concrete
these tests, creep and shrinkage de- Structures—Column Base Plates," Jour-
formations did not significantly in- nal of the Prestressed Concrete Insti-
fluence the flexural strength of the tute, Vol. 11, No. 6 Dec. 1966, pp. 18-
members. It is expected, however, 39; PCA Development Department
Bulletin D110.
that such deformations could seri- 6. Kaar, P. H., LaFraugh, R. W. and
ously reduce shear strength, particu- Mass, M. A., "Influence of Concrete
larly for beams without stirrup rein- Strength on Strand Transfer Length,"
forcement. Journal of the Prestressed Concrete In-
stitute, Vol. 8, No. 5, Oct. 1963, pp.
ACKNOWLEDGMENTS 47-67; PCA Development Department
Bulletin D71.
The investigation reported herein 7. Hansen, T. C. and Mattock, A. H.,
was carried out in the Structural "The Influence of Size and Shape of
Laboratory of the Portland Cement Member on the Shrinkage and Creep
Association. Credit is due to B. J. of Concrete," Journal of the American
Concrete Institute, Vol. 63, No. 2, Feb.
Doepp, B. W. Fullhart, R. G. Hoff- 1966, pp. 267-290; PCA Development
man, 0. A. Kurvits and W. W. Man- Department Bulletin D103.
eck for assistance in the manufac- 8. Hognestad, E., Hanson, N. W., Kriz,
ture and testing of the specimens. L. B. and Kurvits, O. A., "Facilities
and Test Methods of PCA Structural
Prof. Alan H. Mattock, formerly with Laboratory," Journal of the PCA Re-
the PCA Laboratories and now with search and Development Laboratories,
the University of Washington, aided Vol. 1, No. 1, Jan. 1959, pp. 12-20 and
importantly in the planning of this 40-44; Vol. 1, No. 2, May 1959, pp.
investigation. 30-37; Vol. 1, No. 3, Sept. 1959, pp.
35-41; PCA Development Department
Bulletin D33.
REFERENCES
9. Ross, A. D., "Creep of Concrete Under
1. Rostasy, F., "Connections in Precast Variable Stress," Journal of the Amer-
Concrete Structures—Continuity in ican Concrete Institute, Vol. 54, No. 9,
Double-T Floor Construction," Journal March 1958, pp. 739-758.
of the Prestressed Concrete Institute, 10. Corley, W. G., Sozen, M. A. and Siess,
Vol. 7, No. 4, Aug. 1962, pp. 18-48; C. P., "Time Dependent Deflections of
PCA Development Department Bulle- Prestressed Concrete Beams," Highway
tin D55. Research Board Bulletin 307, 1962,
2. Kriz, L. B. and Baths, C. H., "Connec- pp. 1-25.
tions in Precast Concrete Structures— 11. Mattock, A. H., "Precast-Prestressed
Bearing Strength of Column Heads," Concrete Bridges, 5. Creep and Shrink-
Journal of the Prestressed Concrete In- age Studies," Journal of the PCA Re-
stitute, Vol. 8, No. 6, Dec. 1963, pp. search and Development Laboratories,
45-75; PCA Development Department Vol. 3, No. 2, May 1961, pp. 32-66;
Bulletin D73. PCA Development Department Bulle-
3. Gaston, J. R. and Kriz, L. B., "Connec- tin D46.
tions in Precast Concrete Structures— 12. Gordon, P., "Connections for Precast
Scarf Joints," Journal of the Prestressed Members Can be Trouble Free," Civil
Concrete Institute, Vol. 9, No. 3, June Engineering, Vol. 36, No. 7, July 1966,
1964, pp. 37-59; PCA Development pp. 62-64.
Department Bulletin D79. 13. Shideler, J. J., "Manufacture and Use
4. Kriz, L. B. and Raths, C. H., "Connec- of Lightweight Aggregates for Struc-
tions in Precast Concrete Structures— tural Concrete," Jan. 1961, 14 pp.,
Strength of Corbels," Journal of the PCA Development Department Bul-
Prestressed Concrete Institute, Vol. 10, letin D40.
Discussion of this paper is invited. Please forward your Discussion to PCI Headquarters
before July 1 to permit publication in the October 1967 issue of the PCI JOURNAL.
April 1967 37