10
Temperature Induced Deformations in Match Cast Segments 62 Carin L. Roberts- Wollmann Ph.D., P.E. Consulting Engineer Advanced Bridge Technology Associates Raleigh, North Carolina John E. Breen Ph.D., P.E. Nasser I. AI Rashid Chair in Civil Engineering University of Texas at Austin Austin, Texas Michael E. Kreger Ph.D., P.E. Associate Professor University of Texas at Austin Austin, Texas During the match casting operations of precast segments for segmental post- tensioned concrete box girder bridges , a problem can arise due to the heat of hydration of the newly cast segment. When the segments have a particularly high ratio of wing span to segment length, warping of the match cast segment can occur due to a thermal gradient that is caused by the high temperatures in the adjacent newly cast concrete . This warping , or bowing , has subsequently caused problems during erection operations. This paper presents the results of a field study of this phenomenon. Concrete temperatures and segment deformations measured during the casting of several segments for the San Antonio "Y" project are reported along with observations from the erection operations. A simple method is presented to determine whether, for a given bridge, measures to reduce the thermal gradient during match casting should be taken. M atch casting is an essential segment production method to control alignment in precast segmental post-tensioned concrete box girder bridges with thin epoxy or dry joints (see Fig. 1). Normally, the first segment of a span is cast between one fixed and one re- movable bulkhead [see Fig. 2(a)]. The subsequent pieces are cast between the fixed bulkhead and the previously cast PCI JOURNAL

Temperature Induced Deformations in Match Cast Segments · Temperature Induced Deformations in Match Cast Segments 62 Carin L. Roberts Wollmann Ph.D., P.E. Consulting Engineer Advanced

Embed Size (px)

Citation preview

Temperature Induced Deformations in Match Cast Segments

62

Carin L. Roberts­Wollmann Ph.D., P.E.

Consulting Engineer Advanced Bridge

Technology Associates Raleigh, North Carolina

John E. Breen Ph.D., P.E. Nasser I. AI Rashid Chair in Civil Engineering University of Texas at Austin Austin, Texas

Michael E. Kreger Ph.D., P.E.

Associate Professor University of Texas at Austin

Austin, Texas

During the match casting operations of precast segments for segmental post­tensioned concrete box girder bridges, a problem can arise due to the heat of hydration of the newly cast segment. When the segments have a particularly high ratio of wing span to segment length, warping of the match cast segment can occur due to a thermal gradient that is caused by the high temperatures in the adjacent newly cast concrete . This warping, or bowing, has subsequently caused problems during erection operations. This paper presents the results of a field study of this phenomenon. Concrete temperatures and segment deformations measured during the casting of several segments for the San Antonio "Y" project are reported along with observations from the erection operations. A simple method is presented to determine whether, for a given bridge, measures to reduce the thermal gradient during match casting should be taken.

M atch casting is an essential segment production method to control alignment in precast segmental post-tensioned concrete box girder bridges with

thin epoxy or dry joints (see Fig. 1). Normally , the first segment of a span is cast between one fixed and one re­movable bulkhead [see Fig. 2(a)] . The subsequent pieces are cast between the fixed bulkhead and the previously cast

PCI JOURNAL

Fig. 1. Match casting operations.

segment acting as the removable bulk­head [see Fig. 2(b)].

A problem can arise when the rather high heat of hydration of the concrete in the new segment [Segment 2 in Fig. 2(c)] causes a thermal gradient in the match cast segment (Segment 1). This gradient can cause a bowing of the match cast Segment 1, as the higher

Remov- Q) able

Bulk-head

Shape of Thermal Gradient

Fixed Bulk­head

temperature on the face adjacent to the hydrating concrete causes it to elon­gate. Because the still plastic concrete of the new cast Segment 2 conforms to the shape of the match cast Segment 1 face, the bowing that occurs before the new cast concrete has achieved its ini­tial set becomes a permanent curvature in the new segment. The resulting seg­ments [see Fig. 2(d)] have one straight and one curved side.

For most segmental bridge projects, this slight bowing does not cause prob­lems. However, when the width of the segment (wing tip to wing tip) is very large compared to the length of the seg­ment, problems can arise. For these wide segments, the bow shape is a par­ticular problem during the epoxy appli­cation and temporary post-tensioning operations at the erection site. The de­formation of the segments required to close the gap can cause the size of the gap at the joints to increase as each joint is closed (see Fig. 3). In a recent seg­mental bridge project that had problems with bowed segments, the contractor re­ported that four consecutive joints could be closed but upon stressing to close the fifth , the first joint would reopen.'

End Form

"""""""'

-Fixed Bulk­head

This phenomenon not only poses problems in construction but it also can result in overly thick joints and raises questions about stress distribu­tions across joints. In extreme cases, it can cause cracking in the segments.2

This study was initiated to investi­gate the thermal gradients in match cast segments and measure the subse­quent deformations. This paper pre­sents a brief summary of previous studies, a description of the current measurement program, a presentation of the results, an analysis of the col­lected data, and design or construction approaches to overcome this problem.

LITERATURE REVIEW Little mention has been given in the

literature to temperature induced de­formations in match cast segments . Podolny2 describes the problem and notes that in order to eliminate the gaps, it is important to enclose both the newly cast and the match cast seg­ments in an isothermal enclosure. He states that the effect is particularly sig­nificant for segments with width-to­length (w/L) ratios exceeding 6.

Temporary Post-tensioning.----.----,

a) First segment is cast square.

b) First segment is placed in match cast position and adjusted as required for geometry control . Temperature cools.

a) First two segments are dry matched

b) First two segments are epoxied and temporary post-tensioning is applied. Segments deform to close joint.

- - - - - - -~ - - - - -A.. ~ 2A. .... ~

- - - -0 ® ® C) 0 ®\ 0

c) Heat of hydration of second segment sets up a thermal gradient in the first which causes it to bow(A).

d) Segments cool to constan­temperature. First segment returns to original shape, second segment retains one curved side.

c) Gap to be closed during second temporary post­tensioning operation is larger than first.

d) Gap size continues to increase.

Fig. 2. Bowing of match cast segments. Fig. 3. Plan view of temporary post-tensioning operations.

July-August 1995 63

6' r c

w/L = 9.33

56' (17.1m) j

" I )r ____ J]....L-"~~~~~) ~1 6' (4.9m) ..j

Elevation

t<T.s~l

D Type Ill Segments

r 24' (7.3m) .. 1

w/L: 3.0 \tb;= __ ::J-~~~~-.~~) w Elevation

Type I Segments

Section

8'

~ D Section

Fig. 4. Various types of segments.

New Cast 1 Segment I

I

t I

t 3'-0" (914 mm)

3'-0" (914 mm)

10" (254 mm)

1 ---+f--!-§§§~f~~=12" (51 mm) r 2" (51 mm)

Match Cast 1 Segment

I

Top View of Segments

9'-0" (2.74 m)

10" (254 mm) 3'-0" (914 mm)

3'-0" (914 mm)

Wing Thermocouples

Web Thermocouples

Elevation View

Fig. 5. Typical thermocouple layout in segments.

64

Figg and Muller Engineers3 simi­larly describe the problem in their Prestressed Concrete Segmental Bridge Construction Manual. They at­tribute the problem to improper heat­ing during accelerated curing, which is often used to reduce the construction cycle time. They also note that the problem is of particular significance in segments with a large width-to-length ratio. They assert that proper curing of both segments can eliminate the prob­lem completely.

Prescon Corporation' conducted a study of the bow-shaped segment phe­nomenon that was causing problems in the construction of Phase IDA and B of the San Antonio "Y" Project. The segments were very wide, 58 ft (17.7 m), and short in length, 6 ft (1.8 m) (w!L = 9.7) . The erection crews re­ported gaps in joints and difficulties in closing these gaps with the temporary post-tensioning system.

To study this problem, Prescon placed thermocouples in eight seg­ments of one span and measured the resulting segment deformations. They measured temperature induced defor­mations, .1, of up to O.I2 in. (3 mm). An analysis based on a linear thermal gradient from the match cast face to the exposed face produced calculated deformations similar to the measured values. The maximum recorded tem­perature difference between the match cast face and the open face was 33°F (18oC).

DESCRIPTION OF MEASUREMENT

PROGRAM In the present study, a total of four

pairs of segments of the San Antonio "Y" Project IIC were instrumented with thermocouples and deformation measurement systems. Fig. 4 shows the dimensions of the two types of in­strumented segments. Two pairs of segments, II C-4 and 5 and 11 C-8 and 9, were Type I segments with 24 ft (7.3 m) width and 8 ft (2.4 m) length (w/L = 3). The other two pairs, 44A-5 and 6 and 44A-14 and 15, were Type ill segments with 56 ft (17 .1 m) width and 6ft (1.8 m) length (w/L = 9.33).

Two arrays of eight thermocouples were placed in each pair of segments

PCI JOURNAL

Fig. 6. Installation of thermocouples in segment prior to casting.

(see Fig. 5). One array ran through the wing, while the other ran through the thickened top slab-web-wing juncture. Fig. 6 shows the installation of the thermocouples in the segments before casting.

The deformation measurement sys­tem (see Fig. 7) consisted of brackets at each wing tip to which a piano wire was attached. One bracket was equipped with a ratcheted spool that could pull and hold the piano wire very taut. Precision rulers and small mirrors were embedded in the match cast segment. The wire passed approx­imately 0 .5 in . (12 mm) above the rulers. Using the mirrors to ensure re­peatable readings, measurements were taken at 1-hour intervals beginning immediately after the casting was completed. The concrete temperatures in the new and match cast segments were read at hourly intervals as well.

TEST RESULTS Fig. 8 shows a typical plot of tem­

peratures in the new and match cast segments. At the time of casting the new segment, the match cast segment was approximately 24 hours old. The concrete in the match cast segment had achieved its highest temperature several hours earlier, and at the time of the new casting it was cooling down.

Fig. 9 shows a slightly different pre-

July-August 1995

Dead End Bracket

Taut Piano Vv'ire

Match Cast Segment

Top of

New Cast Segment

Spool with Ratchet and Lock

Ruler and Mirror Embedded in Match Cast Segment (see Detail)

New and Match Cast Segments

Precision Ruler ---+-K-.. -+-.::~(~0-.0-1" increments)

Mirror i Taut Piano Wire

Detail

Fig. 7. Deformation measurement system of new and match cast segments.

Temperatures in Segment 44A-14 and 44A-15 Wing Temperatures

§

160 (71)

I.L. 140 t (60)

l 120 ~ (49)

100 (38)

;;; v.; .. ~ ~~:;;~

r-· ·'· II"

v'· ~. '-..., .... . .__ .. ~

,, ,, __

TUT18of t---··-· -··- Reading

-- - 12:00noon

---3:00pm ~---... ~ ......

--5:00pm

--7:00pm

---9:00pm

~-- --- Cast Complete at 11:00am

(27) 80 -6(-1.8} -4(-1.2) -2 (-.61} 0 2(0.61) 4(1.2) 6 (1.8)

Distance from Match Cast Joint, ft (m)

·-·-·--·- · ... · Match cast Segment Segment 44A-14

Fig. 8. Typical temperature readings in Segments 44A-14 and 44A-15 during match casting operations.

65

Temperatures in Segment 44A-14 and 44A-15 Wing Temperatures

Differences from Time of Casting ~ 80 44 u

0 0

:::: 60

0 40

-~-- -- -

--33 :::: 22 0 ,..... __ ---~ ~ .a 20 11 .a

A ...... __ -- -~ ~ ~ 0 0 ~

-Match Cast New Cast-E -20 -11 E Q) Q)

1- ... I 1- -40 -22 1--6 -4 -2 0 2 4 6

(-1.8X-1.2) (-.6) (.6) (1 .2) (1 .8) Distance from Match Cast Joint . ft (m)

Time of Reading

-12:00pm

3:00pm

5:00pm

7:00pm

----9:00pm

Cast Complete at 11 :00 am

Temperatures in Segment 44A-5 and 44A-6 Wing Tem!)eratures

Differences from Time of Casting Time of 80 44 u Reading

~ 0 r-------~~ 0

60 33 :::; -2:00pm :::: --,,--- -0 40 22 0

~ ::I ~

--t"'"

''" --~ 20 ::I

4:00pm

6:00pm

f1 0 Q) 0.

...!!}, 11

0

-11

Q) --- 8:00pm 0.

E -20 ~Match Cast New Cast- E ----1 O:OOprr Q)

I ---6 -4 -2 J

Q)

1- -40 ... I -2 4

-22 6

1- '-:=----:--=---~-:-' Cast Complete

(-1 .8X-1.2) (- .6) (.6) (1 .2) (1 .8)

Distance from Match Cast Joint. ft (m)

Temperatures in Segment 11 C-8 and 11 C-9 Wing Temperatures

Differences from Time of Casting M « u

~ 0

: M " :::; 0 40 22 0 ~ 20 11 ~ .a .a ~ 0 0 ~ 0. 0. E -20 -11 E Q) Q)

1- -40 -22 1--8 -4 0 4 8

(-2.4) (-1.2) (1.2) (2.4) Distance from Match Cast Joint. ft (m)

at 12·00 pm

Time of Reading

-3:00pm

- - 6:30pm

---- 8:30pm

---10:30pm

Cast Complete at 12:30 pm

Fig. 9. Typical temperature differences in Segments 44A-14 and 44A-1 5, 44A-5 and 44A-6, and 11 C-8 and 11 C-9.

sentation of the same data. The tem­peratures taken at each measurement station immediately after the comple­tion of casting are used as reference values and the difference between sub­sequent temperature" and the initial

66

readings are plotted. These plots illus­trate how the match cast concrete im­mediately adjacent [0 to 1 ft (0 to 0.3 m)] to the newly cast segment heats considerably as the temperature of the newly cast concrete rises due to

tQe heat of hydration. Concrete in the match cast segment more than 3 ft (0.9 m) from the new segment seems unaffected and simply continues to cool. The illustrated temperature gra­dient induces the bow shape.

Fig. 10 shows the horizontal de­formed shapes measured with the taut wire system. The precision rulers have graduations of 0.01 in. (0.25 mrn) and the instrument reader's tolerance is considered to be ±0.01 in. (±0.25 mm). The deflections of the Type I boxes were no more than 0.01 in. (0.25 mm), which is the range of instrument reader error.

The deformation that is permanently set into the newly cast segment ap­pears to be that measured when the concrete begins its initial set. The set is usually indicated by a rapid rise in temperature. Because the concrete mix included a retarder, this rapid rise oc­curred 5 to 6 hours after casting was complete. Shortly after this, a crack would appear between the segments indicating that the match cast joint was opening (see Fig. 11).

Prescon also reported similar cracks that normally appeared 5 to 6 hours after casting. These cracks indicate that the newly cast segment had set but the match cast segment continued to bow away. The critical deformation, the deformation that is set into the new segment, is that which occurs approxi­mately 6 hours after casting.

The match cast segments were mea­sured for several days following cast­ing. Over the course of 3 days, the segments returned to their original shape .

ANALYSIS

Method of Calculating Deformation

The temperatures recorded in the match cast segment in the current test and in Pres con's report' indicate a temperature gradient similar to that shown in Fig. 12. The maximum de­flection can be calculated by ftrst de­termining the equivalent moment, M, induced by the thermal gradient:

M = Ea f T(Y)b(Y)YdY (1)

where T = temperature difference between a

PCI JOURNAL

point at a distance Y from the centroid of the section and the centroid, °F CCC)

b = width of the section at a distance Y fro m the centroid of the sec­tion, in. (mm) (see Fig. 13)

a= coefficient of thermal expansion of concrete (:o:6 x 1Q·6fCF)(l.l x w-s;oq

Y = distance from centroid of section, in. (mm)

E = modulus of elasticity of the con­crete, ksi (MPa)

The curvature of the segment can then be calculated as:

</> = MIEI

where M =moment calculated in Eq. (1) I = moment of inertia of the section,

in.4 (mm4) (bV/12)

Finally, the maximum deflection of the segment in in. (mm) is:

,1 = tf>w2/8

where w is the width of the segment (wing tip to wing tip), in. (mm).

Calculated Deformations

Using this method and the measured gradients, the values of maximum de­flection for the instrumented segments were calculated. Fig. 14 graphically il­lustrates the accuracy of the calcula­tion method for two of the segments of the current project.

Based on reduction of the data from all segments in the Prescon study and the current study (see Ref. 4 for com­plete data), the following observations can be made:

1. In comparing the measured gra­dients in the wing with those in the wing-web-top slab juncture, the de­formations calculated based on the wing gradients best match the actual deformations.

2. The temperature on the free face of the segment (Din Fig. 12) was:

a. Normally cooler than the center of the segment in Prescon's seg­ments, which were cast late in the day in mid-April.

b. Approximately the same as the center of the segment in the 11 C segments, which were cast in June at approximately noon with the free face facing south.

c. Normally a few degrees warmer

July-August 1995

Time of Reading

-12:00pm

--3:00pm

---5:00pm

---7:00pm

!----9:00pm

Cast Complete at 11 :00 am -0 02 -0.5

-27 -18 -9 0 9 18 27 (-8.2)(-5.5)(-2 .7) (2 .7) (5.5) (8 .2)

Distance from Center Line, ft (m)

Deformations of Segment 44A-5 Caused by Heat ol Hydration T:me of

0.1 2.5 E Reading

0.08 2.0 E - 2:00pm

c: 0.06 1.5 -~ -- 4:00pm

5 0.04 +----Jf7'1~-~~+:::~t--t 1 .0 ~ --- 6:00pm ~ (5 - - -8:00pm E 0.02 °·5 ~ ---10:00pm ~ o~--f==:......,.+---+--....P--+....,;;;;-..o 0

c:

-0 .02 -0.5 -27 -18 -9 0 9 18 27

(-8.2)(-5.5)(-2.7) (2 .7) (5.5) (8.2) Distance from Center Line, ft (m)

Cast Complete at 12:00 pm

Deformations of Segment 11 C-8 Caused by Heat ol Hydration RTimde. ~!

0.1 2.5 E ea m"

0.08 2.0 ~ - 3:00pm

·- 006 c;· 5 --6:30pm

1.5:;::::; ~ ---8:30pm 0

~ 0.04

§ 0.02 -~ 0 ~

-0.02 -12

(-3.7)

r-~

"" ·-~ -6

(-1.8)

..::... --r-

0

.... -6

(1.8)

-1·0

J2 ---10:30pm 0·5 ~ Cast Complete 0 at 12:30 pm

-0.5 12

(3.7)

Distance frQm Center Line, ft (m)

Positive deformation in­dicates match cast seg­ment is bowing toward new cast segment

Fig. 10. Deformed shapes of Segments 44A-14, 44A-5, and 11 C-8 caused by heat of hydration.

than the center of the segment in the 44A segments, which were cast in August at approximately noon with the free face facing west.

3. Cracking occurred between 4 and

6 hours after casting was completed. Fig. 15 compares the measured de­

flection and the calculated deflection at the time of cracking for all of the segments in the study. The calculated values are based on the gradients mea-

67

Top View of Segments

Crack opens as the Match Cast Segment bows away from the New Cast Segment

Fig. 11. Cracks in segments indicating joint opening.

Match Cast Face

A

Fig. 12. Shape of thermal gradient in match cast segment.

sured in the wing tips. With the excep­tion of Segment 18E-4, the calculated values agree quite well with the mea­sured deflections.

ERECTION OBSERVATIONS

In the San Antonio "Y" Project, some of the effects of the bow shaped segment phenomenon were observed at the erection site. An indication of a

uniformly compressed joint is a bead of squeezed out epoxy along the entire width of the joint. Altogether, 22 spans were surveyed to assess the quality of the joints.

On average, the spans had a total of 17 joints. Of the 17 joints, an average of 10 joints per span showed signs of even squeeze out along the entire width. An average of four joints per span showed gapping in the joint of up to 'Is in. (3 mm) (see Fig. 16). The re-

maining joints were difficult to cate­gorize, usually showi ng neither squeeze out nor gapping.

On the majority of the gapped joints, the gap would appear for 8 to 10 ft (2.4 to 3.0 m) on each side of the centerline of the box. Near the wing tips, the joint would show signs of good squeeze out. This is consistent with the expected behavior of a ba­nana or bow shaped segment.

One detrimental aspect of the gapped joints was a reduced closure pour size. In most segmental bridges where two closure pours are placed in each span between the last typical seg­ment and the pier segment at each end, a slightly reduced closure pour would cause no problems. However, in the San Antonio "Y" Project TIC, the clo­sure pour was placed over the pier and dead end post-tensioning tendon an­chors were located in the joint. The closure pour was designed to be 12 in. (0.30 m) long, but the average mea­sured length of the closure pours was 9.2 in. (0.23 m). This reduced size caused problems in forming the joint and in installing the dead end tendon anchors in the joint.

The extra span length of 2.8 in. (0.07 m) per span equates to an aver­age joint thickness of around 'Is in. (3 mm). This indicates that even if a joint appeared to have good, even epoxy squeeze out, the joint was prob­ably thicker than expected. Where small closure pours over the piers are incorporated into design, the expected thickness of the epoxy joints should be considered during casting operations.

Both the Prescon project and the current project had some problems with the segments that had wiL > 9. They had no reported problems with the segments with wiL :=;; 3.

Table 1. Deformation calculations of segments for Prescon and San Antonio "Y" projects.

Prescon Project San Antonio "Y" wide box San Antonio "Y" narrow box (cold weather cast) (warm weather cast) (warm weather cast)

Ll = aT,axw' (7 .5L- I 06)/L' Ll = aTmaxw' (5.2L- 88)/L' Jl Ll = aTmaxw' (5.2L- 88)/L' "f ~ -

Ll = (6 X 10'6/°F)( l6°F)(708 in.)2 Ll = (6 X JO·'f°F)(l6°F)(672 in.)2 c Ll = (6 X J0·'/°F)(l6°F)(288 in.)2 "'

x [7.5(72 in.) - 106]/(72 in.)' x [5 .2(70.7 in.)- 88]/(70.7 in.)' x [5.2(1 00.5 in.)- 88]/(1 00.5 in.)' '~",

= 0.056 in. ( 1.4 mm) = 0.034 in. (0.86 mm) [D = 0.003 in. (0.08 mm) -~

100 ft span total = 17 segments per span

•. I I 00 ft span total = 17 segments per span 100ft span total= 12 segments per span

x 0.056 in. = 0.95 in. (24 mm) x 0.034 in. = 0.58 in. (I 5 mm) x 0.003 in.= 0.04 in. (I rnm)

68 PCI JOURNAL

w

New Cast Segment

The wide boxes on the San Antonio "Y" project (see Table 1) had a pre­dicted single segment deformation of 0.034 in . (0.86 mm) [measured aver­age of 0.035 in. (0.89 mm)] and a pre­dicted cumulative deformation for a 100 ft (30.5 m) span of 0.58 in. (15 mm). v?

Fig. 13. Variables in segment dimensions.

RECOMMENDATIONS

Recommended Design Gradient

Based on this study and Prescon's study, a design gradient can be pro­posed for climates similar to that of San Antonio, Texas. Fig. 17 shows the wing gradients in seven segments at the time of the appearance of the wing tip crack.

The seven curves are similar in shape on the side closest to the match cast joint. On the free edge, the Prescon segments were considerably cooler than those of the present study due to seasonal climate variations. Fig. 18 shows a design gradient based on these segments.

This gradient is based on concrete mixes using Type III cement (high early strength) in a 550 to 650 lbs per cu yd mix (260 to 310 kg/m3

). In both projects, high range water reducers and retarders were used in the mixes. Different batch designs could have a significant effect on the gradient.

Based on a design gradient of this shape, the following simplified equa-

tions for the maximum segment defor­mation can be developed: For cold weather casting:

..1 = a Tmax W 2 (7 .5L- 106)/V

In metric units:

..1 = a Tmax W2 (191L- 68340)/U

For warm weather casting:

..1 = a Tmax w2 (5.2L- 88)/U

In metric units:

..1 = a Tmax w 2 (133L- 56720)/V

where

w = width of segment (wing tip to wing tip), in. (mm)

L = length of segment, in. (mm) Tmax = maximum temperature differ­

ence [l6°F (8.9°C)] recom­mended (see Fig. 18)

Based on these simplified equations (see Table 1), the Prescon project had a predicted single segment deformation of 0.056 in. (1.4 mm) [measured average of 0.048 in. (1.2 mm)] and a predicted cumulative deformation for a 100 ft (30.5 m) span of 0.95 in. (24 mm).

The narrow boxes of the San Anto­nio "Y" project (see Table 1) had a predicted single segment deformation of 0.003 in. (0.08 mm) [measured av­erage of 0.0 in. (0.0 mm)] and a cumu­lative deformation for a 100ft (30.5 m) span of 0.04 in. (1.0 mm).

From construction observations, the banana shape of the wide segments was detrimental to the construction process , while the narrow boxes caused no difficulties.

There is some discrepancy between the predicted excess span length for the wide boxes of 0.58 in. (15 mm) and the measured average excess span length of 2.8 in. (71 mm). This could be due to additional deformation of the segments after the initial set of the new concrete, or it could be due to other problems during temporary post­tensioning.

The San Antonio "Y" Project IIC segments had a very large surface area on the faces between the segments . During temporary post-tensioning, by the time the epoxy had been smeared on both segment faces of a joint and the three post-tensioning bars used for the temporary post-tensioning had been stressed, at least 30 to 40 minutes would hilVe elapsed after mixing of the epoxy.

It is possible that the very thick

Measured vs. Calculated Deformations of Segment 44A-5

0.1 ~-----r----------,

Measured vs. Calculated Deformations of Segment 44A-14

3mm ,._ Calrulated

.E 0.08 c 0

i 0.06 e .e ., 0.04 0

0.02

0 0 4 8 12 16 20

Time since casting, hours

0.12~-----.----------,

2mm .E 0.09 c 0

i e o.06 1mm .e

Q) 0 0.03

9 0

.,._ Calrulated

....... Measured

2 4 6 8 Time since casting, hours

10

Fig. 14. Comparison of measured and calculated deformations in Segments 44A-5 and 44A-14.

July-August 1995

2mm

1mm

69

.e r:: 0 :p tel

§ .E Cl)

0

0.1 T""""-----.--------------~2.5mm 1!111!1 Calcu lated • Measured

0.08~=====-----------

Calculations and Measurements At time of cracking at joint

44A-5 44A-1 4 11C-4 11C-8 18E-1 18E-2 18E-4 18E-5 18E-7 18E-8

Curren~roject I Pri:con Project

Segment Designation

2.0mm

1.5mm

1.0mm

0.5mm

Fig. 15. Calculated vs. measured deformations due to heat of hydration at time of wing tip crack opening.

30

Fig. 16. Gapped joint in segment.

epoxy was beginning to react and was, therefore , extremely difficu lt to squeeze from the joints; thus, thicker than expected joints resulted. (17) - 11C-4 - · 44A-5 -- 18E-1 1~

Recommended Design and Construction Approach

A possible design and construction approach would be as follows:

1. Check the w/L ratio. If it is Jess than 6, casting temperature gradients should not be a problem. If it is greater than 6, continue to Step 2.

2. Determine the worst case design gradient.

3. Calculate the segment deforma­tion at the time of concrete set.

4. Calculate the cumulative defor­mation for all segments of a span.

5. If the calculated maximum de­formation for one segment exceeds 0.03 in. (0.8 mm) or the cumulative de­formation for one span exceeds 0.50 in. (12 mm), require that measures be taken during construction to reduce the thermal gradient.

Measures to Reduce Thermal Gradients

The most obvious means of elimi­nating excessive deformations is by keeping the match cast segment warm. An isothermal enclosure, as advocated by Podolny2 is one possi­bility. Curing blankets and plastic sheeting is sufficient in warmer eli-

70

§ - 11Co8 - ·· 44A-1~ - 18E~ &L 20 g (11)

! 10

0 (5.6)

~ 0 ! G) Q.

~ -10 (-5.6)

-~

~--............. ·· -·--~ ~

' ' '\

-20 (-11) 0 2 (.81) 4 (1.2) 6 (1.8) 8(2A)

Distance from New Cast Segment, ft (m)

Fig. 17. Thermal gradient in match cast segments at time of initial set of newly cast segment.

mates, but continued steam curing may be necessary in colder climates. Any means of warming the match cast segment should help in reducing the problems associated with the bow shaped segments induced by thermal gradients.

CONCLUSIONS Thermal gradients causing bow

shaped segments in segmental post­tensioned concrete box girder bridges

have caused problems in the past. The variables that affect the magnitude of the problem are:

1. The width-to-length ratio of the segment: the higher the ratio, the worse the deformation (ratios over 9 have caused problems in the past; ra­tios over 6 could cause problems).

2. The concrete mix design: Type III cement, used for high early strength, heats to higher temperatures sooner than Type I mixes.

3. Ambient temperature: cooler air

PCI JOURNAL

Match Cast Face

9" (229 mm)

F 2oF (4.4oC)

16°F (T,.,) (8.9°C)

(1.1°C)

-8°F (-4.4°C)- for cold weather casting 0° F - for warm weather casting

Fig. 18. Design thermal gradient of segment.

temperatures create more severe gradi­ents and deformations.

4. Orientation of the casting bed: free face exposure to the sun reduces gradients by keeping the free face warm.

5. Time of casting: morning cast­ings, when ambient temperatures are on the rise, produce smaller gradients than evening castings when the air temperature is on the decline.

The approach presented herein should indicate when thermal gradi­ents and resulting segment deforma­tions could cause problems in segmen­tal bridge projects. Construction measures can be taken to reduce the

July-August 1995

magnitude of the problem. Elimination of the permanent segment deforma­tions will produce more trouble free erection operations.

ACKNOWLEDGMENT The findings reported in this paper

were collected as part of a compre­hensive field study of the San Antonio "Y" Project IIC. This project was sponsored by the Texas Department of Transportation and the Federal Highway Administration.

The views presented in this article are those of the authors and not neces-

sarily those of the sponsors. The project was conducted at the

Phil M. Ferguson Structural Engineer­ing Laboratory at the University of Texas and at the job site. The project was supervised by Drs. John E. Breen and Michael E. Kreger. The success of the project can be attributed to the great cooperation and assistance the research team received from the con­tractor, Austin Bridge and Road, Dal­las, Texas, and from the Texas DOT personnel both in the field and in the Bridge Design Division.

Finally, the authors would like to express their appreciation to Alan R. Phipps, regional director, Western Re­gional Office, Figg Engineers, Inc., Denver, Colorado, for his very articu­late comments on the initial manu­script and his interest in the project.

REFERENCES 1. Prescon Corporation, "Segment Moni­

toring Test Report," Unpublished Re­port, San Antonio, TX, June 22, 1988.

2. Podolny, W. Jr., "The Cause of Crack­ing in Post-Tensioned Concrete Box Girder Bridges and Retrofit Proce­dures," PCI JOURNAL, V. 30, No. 2, March-April 1985, pp. 82-139.

3. Figg and Muller Engineers, Prestressed Concrete Segmental Bridge Construc­tion Manual, Bridge Contractors Semi­nar, Asheville, NC, June 1981.

4. Roberts, C. L., "Measurement Based Revisions for Segmental Bridge Design and Construction Criteria," Doctoral Dissertation, University of Texas at Austin, Austin, TX, December 1993.

71