Simplified bearing

plate computations for

post-tensioning
anchorages
Charles Rejcha, P.E.
Vice President and Chief Engineer
Freyssinet Company, Inc.
Tuckahoe, New York

Presents an analytical
method, based on tests,
for simplifying bearing
plate computations under
anchorages of post-
tensioned concrete
members.
A fully-worked numerical
example demonstrates
the design procedure.

M any post-tensioning anchorage

systems today are using an in-
dependent mild steel bearing plate to
Unfortunately, though, there is no
simple analytical method available
whereby the bearing plate stresses can
transmit the prestressing force to the be evaluated accurately.
concrete member. The purpose of this In general, the plate area is governed
plate is to distribute efficiently the high by the concrete bearing stress, which is
concentrated bearing stresses to the easy to establish if the average con-
concrete. crete stress is considered. However, the

102
 computation of the plate thickness is Fig. la shows a testing bed. The
extremely difficult because of unknown. stressing jack on the extreme left bears
factors such as load distribution, stress on the concrete test block. Further to
distribution, and arch effects. the right are two larger test bed blocks,
In practice, the bearing plate thick- one stationary and one sliding, with two
ness is usually based on tests or past rams in between. The tendon is thread-
experience. ed through the test block, through the
This paper presents a simplified de- two test bed blocks, and between the
sign procedure for calculating the bear- two rams. It is then anchored at a non-
ing stresses of plates under post-ten- stressing "super" anchor block at the far
sioning anchorages. The method is right.
based on the results of many tests, one In general, the tendon is first stressed
of which is described in this paper. and anchored using the stressing jack.
Because of the complexity of the (A jacking force of 0.75 to 0.80 of ulti-
problem, several simplifying assump- mate is usually employed.) The stress-
tions are made for both square and rec- ing jack is then removed and the ten-
tangular plates. don is subsequently restressed in steps,
A numerical example, using a square to destruction, using the two rams.
plate, is included to show how this Fig. lb shows the test block consist-
method can be used to calculate the ing of the anchorage, embedded bear-
bearing plate stresses. ing plate, and concrete block. The plate
cover, block strength, block reinforce-
ment, and tendon enclosure is designed
Bearing Plate Test to duplicate actual site conditions.
Prior to actual use, it is customary to
test the performance of a bearing plate Design criteria
anchorage assembly. The following is a The following design criteria, devel-
description of one such test: oped by the California Department-

Fig. la. Testing bed.

PCI JOURNAL/July-August 1975 103

ment of Transportation, were used for eral use, may be substantially increased
the above test: in particular cases, especially when pre-
1. The final unit compressive stress cast concrete is employed.
on the concrete directly underneath the The steel bearing stress at the con-
plate or assembly should not exceed tact area between the anchorage and
3000 psi. the plate should also be checked. It is
2. The bending stresses developed suggested that this stress be limited,
in the plates or assemblies induced by under final load, to 90 percent of the
the pull of the prestressing steel should plate yield stress.
not exceed the yield point of the ma-
terial or cause visible distortion in the
anchorage plate when 95 percent of Simplified Bearing
the specified ultimate tensile strength Plate Computations
of the tendon is applied as determined
by the engineer. This section discusses the range of ap-
To conform with Criterion 1, the plicability of this method, the design
average (uniform) concrete stress un- assumptions and design procedure for
derneath the plate is computed under both square and rectangular plates and
final load, assumed to be 60 percent of three typical stress block characteristics.
the guaranteed ultimate tensile strength Finally, the last section gives a fully
of the tendon (GUTS). worked numerical example to show how
It should be noted that the 3000-psi this method is used to compute the
concrete stress, recommended for gen- bearing plate stresses.

Fig. 1b. Test block detail.

104
 Range of applicability of
simplified method
It should be emphasized that this pa-
per deals only with the bearing plate
computations. The concrete strength,
concrete cover on plate edges, and
mild steel reinforcement must be pro-
perly selected to develop the required
forces.
Due to the simplifying assumptions
the calculations must be considered
"empirical" and should not replace the
basic post-tensioning anchorage plate
testing of the assemblies. The proposed
analytical approach is recommended
for particular cases (e.g., tendon sizes
and plate shapes falling in between the
tested basic size tendons).
Also, the tendon manufacturer may
adjust the design assumptions to cor-
respond to the test results of his basic
size tendons.

Design assumptions
The following design assumptions ap-
ply to both square and rectangular
plates:
1. Use a stressing force of 0.95 of
GUTS.
2. If the anchorage or hole in the
plate is circular, use square shapes of
an equivalent area. This will simplify
the computations.
3. Due to the very large anchorage but disregard the holes in the anchor-
stiffness, assume a uniform bearing age.
stress, using a 45-deg arch effect, as 7. In general, a computation of the
shown in Fig. 2. maximum stresses due to combined bi-
4. Due to plate flexibility, assume axial bending and shear is not neces-
the concrete bearing stress to be vari- sary.
able as shown in Fig. 2. This assump- 8. At contact area between anchor-
tion approximates a uniform stress un- age and plate, the ratio of horizontal
der the anchorage with a parabolic shear to prestressing force should not
fade-out under the overhangs. exceed 0.5.
5. Assume that the anchorage and
the plate act as one homogeneous body.
6. In computing the section proper- Discussion of assumptions
ties, consider the opening in the plate Assumptions 1, 2, and 3 are self-ex-

PCI JOURNAL/July-August 1975 105

 BEARING STRESS LIMIT LINE ANCHORAGE HOLE ONE QUADRANT
(SHADED)

PLANS I FIG. 3

planatory. A commentary of the re-

maining assumptions is given below.
4. During a test under 0.95 of
GUTS, the plate deflects in a dish-like
shape. The maximum deflection of the
plate is at the center whereas the plate
edges sink into the concrete. (For ex-
r---t ample, in the case shown on Figs. lb
Y ^t tI
and 6, a deflection of 1/16 in. is observed
X. L
at the center of the plate and a sinking
of 1/3 2 in, at the edges.) After the load
is released the plate will go back to its
SYM.ABOUT 'i
PLAN original shape if it has been properly
dimensioned.
I BEARING The tendon manufacturer may also
111111^- -- ^
------^STR
adjust the shape of the bearing stress
blocks in accordance with his basic ten-
I don test results.
ANCHORAGE
5. A visual test can show convincing-
ly that the anchorage and plate assem-
bly perform homogeneously. This is be-
cause there exists a large prestressing
J BEARING PL. load which forces the anchorage against
the plate. See also the discussion below
under Assumption 8.
6. The anchorage is made out of a
CONCRETE
high quality steel. In the case of a
.l t t
STRESSES strand anchorage, using a male plug or
CROSS SECTION individual strand jaws, the anchorage is
WITH CORRESPONDING STRESSES subjected to large tensile hoop stresses,
and the holes are filled with strands,
I FIG. 4 jaws, or a plug. The resulting compres-

106
 P =fxbxc P =fxbx(a +c) P=fx bx(.75a t.67CI

Y= b .5a+.67C Y= b 33a *,4I C

Y=bx.67 a +C . 5a +.67C

v
t 1 /

b /

b U

ISOMETRIC VIEW OF TYP. STRESS BLOCKS FIG. 5

sion stress due to bending is actually Also, the shear stress is zero at the loca-
advantageous. tion of the maximum bending stress.
7. The maximum stresses due to bi- Therefore the maximum fb and f,,
axial bending and shear combined is stresses will not exceed the stresses
found from the following equations: computed in one direction and need
not to be checked.
fb = Ybl + fb2) 8. The contact area between the an-
{ [ (fbl f52)] 2 + fv2} chorage and the bearing plate is usual-
and ly minimal. The governing stress in this
location is the horizontal shear, which
fv = { [(fb1 fb2)1 2 + fv2} is subject to redistribution. Also, it
where f bl and fb2 are bending stresses should be recognized that Assumption
in the two directions. 2 (square anchorage and hole) and the
This condition will occur at or near the fact that the plan area will be divided
45-deg lines between quadrants. How- into four quadrants (see the "design
ever, at the critical points, the bending procedure" which follows) greatly in-
stresses in the two directions are equal. fluence the magnitude of the horizontal

PCI JOURNAL/July-August 1975 107



16" SYM.
ABOUT
8
I
6.69 "m HOLE AT
rL

' ONE QUADRA NT

(SHADED)
^..-^ 8.5"^ ANCHORGE
AT

BEIRING PLATE
PLAN ASTM A-3G

rPU=19X58.6=1113.4K
N O
N^ (ULTIMATE)

HALF PLAN
a G
CROSS SECTION 56.4 K107.7K
.39" .58"
FIG. 6 100.4 K I
BEARING FORCES
^^ I PER QUADRANT
shear stress. Nevertheless, the total
shear force per quadrant over one-quar-
234 5 ^SECTION
ter of the stressing force will give a DESIG.
reasonable value of the friction coeffi- 1n I N

cient which prevents sliding of the an- N ^

chorage on the plate face. in

N ^-

Design procedure for

square plates
Using the design assumptions listed
previously:
4.87"
1. Divide the plan loading area into
four quadrants as shown in Fig. 2. This --. CONCRETE
procedure is to prevent the same forces REACTIONS
PER QUAD.
being considered twice in two direc- 3j
tions.
30.3 K
2. Determine the average concrete
58 1.5"
stress underneath the plate and the
steel bearing stress on the plate and 57.9 K176.3K
check the stresses with the allowable
values. CROSS SECTION
WITH CORRESPONDING FORCES
3. Compute the section properties in
the critical location. Use the full an-
FIG.7
chorage and plate widths. Take into

108
consideration the hole in the plate but SECTION PROPERTIES
16" WIDE SECTION
disregard the holes in the anchorage.
SECTION Y TIN IYB
IN I IN3
IN 4 Q
4. Determine the shears and bend- 1,21L 2.54 2.21 I 76.8 I 24.2
ing moments per quadrant. Compute 2R,3L 2.81 I
1.94 J 94.4 I 29.3
3R,4 1.125 1.125 15.2
the vertical shear and bending stresses
and check these values with the allow- SYM. 8"
able stresses. ABOUT
2R3L
2L
SECTION
5. Calculate the horizontal shear 3R5 ES 1G.
4
force per quadrant at the contact area
between the anchorage and the plate.
N } U
The ratio of this horizontal force to the
vertical prestressing force per quadrant N-
}
should not exceed 0.5. N

1.12
Design procedure for
rectangular plates
The design assumptions listed for c
^
square plates apply also to rectangular SHEAR
plates. However, in the design pro- PER QUAD.
cedure (see Step 1 in previous section)
the quadrant should be laid out as

H_
shown in Fig. 3. In a rectangular ar-
=
Y
m M NaBENDING MOM.
rangement there are two different u, a PER QUAD.
ro
quadrants per plate. However, only the
one with a larger plate overhang will N'
govern. N Q ^

In the case of a common plate ac- m

o
commodating several anchorages close- N HOR.S HEAR STRESS
ly spaced, the computations may be AT CONTACT AREA
PER QUAD.
simplified as shown in Fig. 4.
In the y direction, compute the
CROSS SECTION
bearing stresses of the plate assembly WITH CORRESPONDING DIAGRAMS
under uniform load within the strip
"j " . However, in the x direction which
governs, compute the bearing stresses
of the plates under the loads shown in
the cross section. r. l^ I N IfI

II
Stress block characteristics N

The characteristics of three typical

stress blocks employed in this paper are
shown in Fig. 5. The following notation
is used: PLAN OF CONTACT AREA
P = total force per stress block, kips SUBJECTED TO HOR.SHEAR STRESS

f = stress, ksi
a, b, c = distances, in. FIG, 8
y = location of center of gravity, in.

PCI JOURNAL/July-August 1975 109

 Design Example

Assume we have a round anchorage for Bearing forces per quadrant:

nineteen 0.6 -in, diameter strands with a
P 112 = 11.15(3)3 = 100.4 kips
guaranteed ultimate force of 58.6 kips
per strand (or P 1113.4 kips per an- Pp22 = 11.15(0.75)(6 + 0.75)
chorage). = 56.4 kips
Let the prestressing force be distrib-
uted on an ASTM A36 square plate Pr34 = 11.15(1.12)(7.5 + 1.12)
as shown in Fig. 6 and Fig. 1b. The = 107.7 kips
dimensions of the plate are 16 x 16 in.
square and 21/4 in. thick In the mid- Concrete reaction forces per quadrant:
dle of the plate there is a 6.69 in. di- Pi23 = 5.99(0.75)(6 + 0.75)
ameter opening. 30.3 kips
First, check under the final load of P 134 = 5.99(1.12)(7.5 + 1.12)
0.6P, the average concrete stress un- = 57.9 kips
derneath the plate (f^f), and the steel
bearing stress on the plate (fpf). P.4 , = 5.99(3.13)9.40 = 176.3 kips

0.6 x 1113.4 Check the sum of forces per quadrant:

fef 16 X 16 - 0.25 ( i
r) 6.692 1p p = JP, = 0.25P = 264.5 kips
= 3.00 ksi (ok)
The location distances y of all forces
0.6 x 1113.4 are shown in Fig. 7. These values were
fpt 0.25 (vr) (8.5 2 - 6.692) computed using the stress block formu-
= 30.95 ksi < 0.9 x 36 (ok) las.

Compute the bending and shear The section properties for the 16 -in.
stresses (see Fig. 7) under a load of wide plate is shown in Fig. 8. The val-
0.95 P., making use of the assumptions ues Q represent the statical moment of
listed previously. the 7.5 x 2.5 in. area around the center
of gravity axis.
Total force: The shears and bending moment of
P = 0.95 X 1113.4 = 1058 kips one quadrant are shown in Fig. 8.

Bearing stress: Vertical shear stress at Section 3R:

fD = 1058/(4.87 x 2) 2 = 11.15 ksi f, = (3/2) [126.5/(16 X 2.25)] = 5.3 ksi
Concrete stress, using stress block for- Bending stresses at Section 3R:
mulas :
f b = -x(434/15.2)1.125 = -_*32.1 ksi
1058/4 = f [0.75(6 + 0.75) (<36 ksi, ok)
+ 1.12 (7.5 + 1.12) + 3.13z]
Bending stresses at Section 1:
where z= 0.75X9.74+0.67X3.13
= 9.40 fbt = +( 618/76.8)2.54 = +20.4 ksi
from which f0 = 5.99 ksi. f,b = -(618/76.8)2.21 = -17.8 ksi

110
Horizontal shear stresses at contact force per quadrant:
area:
fj = ( 1/2 X 3 X 1.5x21.09)+
f h = VQ/bI [0.75 x 7.5(4.15 + 5.24)/2]
= 73.9 kips
fh21 = (100.4 x 24.2)/(1.5 x 76.8)
21.09 ksi Friction coefficient at contact area:
f n2r _ (100.4 x 29.3)/(7.5 x 94.4) Cf = 73.9/(0.25 x 1058)
= 4.15 ksi = 0.28 < 0.5 (ok)
For biaxial stresses and horizontal
fh3t = (126.5 x 29.3)/(7.5 x 94.4)
= 5.24 ksi shear magnitude see the commentary
on Assumptions 7 and 8 in the "Design
The sum of fh, is the total horizontal Assumptions" section.

Discussion of this paper is invited.

Please forward your discussion to
PCI Headquarters by December 1, 1975.

PCI JOURNAL/July-August 1975 111