Documente Academic
Documente Profesional
Documente Cultură
Rafik Y. Itani
Associate Professor and
Head Structural Engineering
Department of Civil and
Environmental Engineering
Washington State University
Pullman, Washington
I t /4 It/2 31t/4 le
DISTANCE FROM END OF MEMBER
sections. Finite element methods were tion members. The study indicated that
used to analyze the stress distributions. spalling stresses, compared to bursting
The results of the analyses were then stresses, are relatively insensitive to the
compared with existing test data, and horizontal positioning of the prestress-
with other theoretical solutions. Com- ing forces and to the geometry of the
parison with the experimental values of section. It was also found that web
Zielinski and Rowe l-' indicated that bursting stresses in I sections tend to be
the computed stresses were extremely higher over a greater depth than the cor-
low. But, when compared with the test responding stresses in the equivalent
results obtained by Taylor, "•' s the com- rectangular section. Spalling stresses
puted stresses were found to he more were found to have a smaller transverse
accurate. Comparison with the theoreti- variation than bursting stresses. The ef-
cal solutions of Guyon, Iyengar, and fect of the flange in I sections was found
Yogananda" showed considerable to reduce the spalling stresses.
discrepancy. The high surface stresses The third paper by Yettram and Rob-
predicted by the finite element analyses bins' s was published in 1971. The stress
were not predicted by their theoretical distributions in the anchorage zone of
solutions. axially post-tensioned I sections with
Late in 1970, Yettram and Robbins'' end blocks were analyzed using finite
published their second paper on an- element methods. Comparisons were
chorage zone stresses in post-tensioned made between the stresses in the mem-
members. Again they used finite ele- bers using various lengths of end block
ment methods to study the stress dis- and an I section with no end block. It
tributions for multiple and eccentric an- was concluded that short end blocks,
chorages on both rectangular and I see- with lengths that are less than three-
102
Fig. 2. I section, quarter symmetry model.
Z
W
w 0.02 /
Uniform I section b
section with rectangular end block
0- — '-- End block with short transition
I End block with long transition
a 0.01 , (~^^
W
a
H 0 4 e 12 16 20 24 28 32
DISTANCE FROM END OF MEMBER, X (inches)
Fig. 5a. Comparison of transverse tensile force, axial post-tensioned members with and
without end blocks.
0 4 8 12 16 20 24 28 32
many researchers these assumptions are tions, strand slippage, and assumed lo-
clearly questionable since no allowance calized material failures, the transfer
is made for stress redistribution due to length for pretensioned loading is di-
cracking, time-dependent effects, or vided into regions.
plastic deformation. In this analysis, the transfer length is
Yettram and Robbins found that Pois- assumed to consist of two regions: one of
son's ratio has little effect, therefore, it zero bond stress and one of linearly
has been assumed to be 0.167 through- varying bond stress.
out this paper, and Young's modulus for For simplicity, the region of zero bond
concrete has been assumed as 4.03 mil- stress has been taken as one-half the
lion psi (28,000 MPa). length of the leading elements in the
For the purpose of analysis, a % in. finite element idealization of the end
(9.5 mm) diameter strand has been as- zone. It should he noted that had it been
sumed on all but the Series 10 and assumed that transfer begins at the end
Series 14 girders for which '/s in, (12.7 face of the girder, the results would then
mm) diameter strand is assumed. he higher tensile stresses over the initial
portion of the end zone.
Fig. 1 depicts the assumed force dis-
Pretensioning Transfer Rate
tribution over the transfer length. This
Highly stressed concrete acts as a force distribution presumes that the en-
plastic material rather than as an elastic tire pretensioning force, at initial trans-
material. Therefore, highly stressed fer, is transmitted over the transfer
areas tend to have large plastic defor- length portion of the development
mations. Due to these large deforma- length.
U) 0.5
W -- Uniform I section
l — — I section with rectangular end block
—'— End block with short transition
W 0.4 "'-- End block with long transition
J
Z.
W 0
Q-the total applied force
qz Q/4ab
I
l -• 0.3
W
H 0.1
w -
IX
W
a 0 4 8 12 16 20 24 28 32
106
Uniform I section fY 12
0.05 — --I section with rectangular end block
End block with short transition
a 0.04 i---'r End block with long transition
H 1iys loaded area of
^^ U 0.03 \. quarter symmetry model °
I d 0.0
a 4^ 2$` 32
Z a - o-ol_^ ~J
a ^
a- o
~ ~ -0.02
DISTANCE FROM END OF MEMBER ,X(inches)
^Y 12 i
Uniform I section t^'1
0.05 -- I section with rectangular end block Z"
^- — End block with short transition
End block with long transition
0
0.04
a Pretensioning load a
W 0.03
WM 1i Q= the total applied force
OO 0.02 ^1 q= Q^4ab Z
la1 J 0.01
[L
Or CL 8 12 16 20 24 28 32
W a 0.0
z a
M o -o.ol __-' -
i- i-
-0.02
DISTANCE FROM END OF MEMBER, X (inches)
108
WDOT Series 10 and 14 blocks typically used on these girders
Thin Web Girders are unnecessary. Therefore, a model
study was deemed useful as one means
The primary interest in this investi- of evaluating whether the end blocks on
gation was a modeling of the stress dis- these girders could be eliminated.
tribution in deep, thin web precast pre- Typical details of the State of Wash-
stressed girders. It is felt that the end ington's Department of Transportation
42
2.875 L ,Ifs
1 ^
iI-1
1 E
73.5 A 1 I A
I I
I 1 I
1IN-
END 4AX
I ! BLOCK
3
++++ ++++
g +++++ +++++
+++++ +++++ 175 SECTION
A—A
24
1.5
510 I @I.75
+l+_
-^ +I+-r
I++I
I I
I
r1-5^I
A I I A
58.0 _
{
J
16
{
I I
I I
Y
v
0
__
3 + + p MIN—
+++{I ++++- z MAX
6 +++++ +++++- w
+++++ +++++-
75
62
'14-4i .75
24 SECTION
A—A
110
applied loading was introduced by re-
solving the applied forces into vertical
and horizontal forces applied at the
center of gravity of the harped strands.
The results of the analyses are shown in
Figs. 12a and 12b for the Series 10 gird-
ers and in Figs. 13a and 13b for the
Series 14 girders,
Using the base of the girders as a ref-
erence, the maximum total tensile force
was found to occur along a horizontal
section 21.57 in. (548 mm) above the
base for the Series 10, and 28.33 in. (720
m) above the base for the Series 14. Al-
though these are not necessarily the
sections along which the absolute max-
imuins occur, they suffice for the pur-
pose of comparison. It should be noted
that the maximum tensile stress line will
not occur at mid-depth if the prestress-
ing force is applied predominantly to
one flange.
A comparison of the transverse tensile
force for a Series 10 girder with and
without end block is shown in Fig. 12a.
It is clear from the figure that a two-fold
increase in total transverse tensile force
occurs with provision of the end block.
Comparison of the transverse tensile
stress shown in Fig. 12h indicates that a
30 percent reduction occurs in the max- Fig. 11. Half symmetry models of WDOT
imum tensile stress when the end block Series 14 girder.
is present.
Similar comparisons of transverse ten-
sile force and transverse tensile stress
are made for a Series 14 girder in Figs. 1. Transverse tensile stresses can be
13a and 13b. reduced significantly for concentrically
Since the tensile strength of concrete loaded girders by provision of an end
will be exceeded for typical transfer block. A transition length equal to twice
loads, little advantage is gained by pro- the length of the end block was found to
vision of the end blocks. Therefore, the provide the best reduction in transverse
end blocks can only serve on these gird- stress.
ers to reduce the congestion of trans- 2. For eccentrically loaded girders,
verse reinforcement. the maximum transverse tensile stress is
reduced by provision of an end block.
However, the total transverse tensile
CONCLUSIONS force increases. It was also found that
the maximum transverse tensile stress
The results of the analyses outlined in tends to increase with increases in
this paper indicate the following major transition length or with increases in ec-
conclusions: centricity.
0.016 0.4
N
C
ai
C
U 0.012 0.3
_w . g A
m
U)
m
`TTF •O.13300
0.008 0.2
C
Fig. 12a. Comparison of transverse tensile f= ig. 12b. Comparison of transverse tensile
force in WDOT Series 10. stress in WDOT Series 10.
C `m
a
0 4 9 12 16 0 4 8 12 16
Distance from end of member, X (inches) Distance from end of member, X (inches)
Fig. 13a. Comparison of transverse tensile Fig. 1 3b. Comparison of transverse tensile
force in WDOT Series 14. stress in WDOT Series 14.
112
3. The provision of end blocks on Se- stress and are justifiable only to reduce
ries 10 and Series 14 girders can only congestion.
serve to reduce the congestion of the 3. In reference to the Series 10 and
transverse vertical reinforcement. Series 14 girders in use by the State of
Washington's Department of Transpor-
DESIGN tation, it is recommended that end
RECOMMENDATIONS blocks be provided only if deemed nec-
essary by quality control. Proper consol-
1. Based upon the results of this idation in the anchorage zone of these
study, it is recommended that for con- girders may be difficult to achieve. If
centrically loaded members end blocks end blocks are provided to reduce the
be considered. The two-thirds reduction congestion, it is recommended that the
in transverse stress found for the mem- width of the thickening be kept to a rnin-
bers studied clearly indicates the ad- i.mum and that the transition length not
vantages, although little reduction in the exceed the length of the end block.
total transverse force occurs. Full scale testing of end block re-
2. For eccentrically loaded members, moval from a Washington State girder is
it is recommended that end blocks be now being conducted by Washington
deleted as they offer no real advantage State University for the State of Wash-
in the reduction of transverse tensile ington Department of Transportation.
REFERENCES
1. Marshall, W. T., and Mattock, A. H., chorage Zone Stresses in Axially Post-
"Control of Horizontal Cracking in the tensioned Members of Uniform Rectan-
Ends of Pretensioned Prestressed Con- gular Section," Magazine of Concrete
crete Girders," PCI JOURNAL, V. 7, Research, V. 21, No. 67, June 1969, pp.
No. 5, October 1962, pp. 56-74. 103-112.
2. Arthur, P. D., and Ganguli, S., "Tests on 8. Zielinski, J., and Rowe, R. E., "Distribu-
End-Zone Stresses in Pretensioned Con- tion of Stress in Anchor Blocks," Engi-
crete I beams," Magazine of Concrete neering, V. 189, No. 4899, March 1960,
Research, V. 17, No. 51, June 1965, pp. pp. 347-348.
-348.
85-96. 9. Zielinski, J., and Rowe, R. E., "An In-
3. Marshall, W. T., "A Theory for End Zone vestigation of the Stress Distribution in
Stresses in Pretensioned Concrete the Anchorage Zone of Post-tensioned
Beams," PCI JOURNAL, V. 11, No.2, Concrete Members," Research Report
April 1966, pp. 45-51. No. 9, Cement and Concrete Association,
4. Sievers, H., "The Computations of Seat- London, September 1960, p.32.
ings and Abutments for Bridge Piers," 10. Zielinski, J., and Rowe, R. E., "The
DerBauingenieur, V. 27, No. 6, 1952, pp. Stress Distribution Associated With
202-213. Groups of Anchorages in Post-tensioned
5. Sievers, H., "Stress Conditions in the Vi- Concrete Members," Research Report
cinity of Anchorage Plates in Prestressed No, 13, Cement and Concrete Associa-
Tendons of Prestressed Concrete Struc- tion, London, October 1962, p. 39.
tural Units," Der Bauingenieur, V. 31, 11. Taylor, S, J., Discussion of the paper, "A
No. 4, 1956, pp. 134-135. Three Dimensional Stress Distribution
6. Hawkins, N. M., "Behaviour Design of Problem in the Anchorage Zone of a
End Blocks for Prestressed Concrete Post-tensioned Concrete Beam," Maga-
Beams," Civil Engineering Transactions, zine of Concrete Research, V. 19, No. 58,
Institution of Engineers, Australia, 1967, pp. 54-57.
V. CE8, No. 2, October 1966, pp. 193- 12. Taylor, S. J., "Anchorage Bearing
202. Stresses," Conference on Prestressed
7. Yettram, A. L., and Robbins, K., "An- Concrete Pressure Vessels, 1967, Lon-
APPENDIX — NOTATION
a = one-half of section depth, in. from base of girder or centroid
b = one-half of web width, in. fN = initial stress in prestressing strand,
l„ = distribution length, in. before losses, ksi
l r = length of prestress transfer, in. fc'j = compressive strength of the con-
Q = total applied prestressing force crete at time of initial prestress,
v„ = transverse tensile stress, psi ksi
q = Q14ah db = nominal diameter of the prestress-
Y = distance of maximum tensile force ing strand, in.
114