Joint Heating AllowsWinter Construction on

Linn Cove ViaductJean MullerTechnical DirectorFigg & Muller Engineers, Inc.Tallahassee, Florida

James M. BarkerChief Bridge Engineer

Figg & Muller Engineers, Inc.Tallahassee, Florida

T he Linn Cove Viaduct, located nearGrandfather Mountain in North

Carolina, is a 1243-ft (379 m) multispanprecast segmental bridge being erectedby the progressive placing method. Fig.1 shows the construction progressthrough February 1982.

The project is owned by the NationalPark Service. The Federal HighwayAdministration (FHWA) is doing all ofthe contract administration for the proj-ect and has retained the services of theauthors' firm to provide segmental con-struction expertise, Jasper ConstructionCompany is the general contractor.

The design of this project was com-plicated by the fact that the surround-ing terrain is among the most ruggedand environmentally sensitive in theeastern part of the United States. Theuse of precast segmental constructionfor both sub- and superstructure con-struction, with all erection being donefrom the top (Fig. 2), was the solutionchosen to the challenge of building the

SynopsisWinter erection of precast con-

crete segments is always difficultwhen epoxy is used in the joints,because this material will notharden when temperatures fallbelow a certain level.

For erection of the Linn CoveViaduct to continue through thewinter, the epoxy in the jointsneeded to be heated to at least40 F (4.4 G) to cure properly.

This paper gives results ofthree tests used to develop asystem for heating and insulatingjoints, describes the heatingmethods used, and makes rec-ommendations for developmentof such systems for other proj-ects.

120

Fig. 1. Linn Cove Viaduct constructed with precast segments erected by progressiveplacing. This photo shows progress through February 1982.

bridge around Grandfather Mountainwithout damaging the terrain. The onlywork being done at ground level is thedrilling of the microshaft piles.

All materials, including microshaftreinforcement, subfooting, footing con-crete, and pier segments, are placedwith the stiffleg crane mounted on theend of the completed superstructure.The stiffleg is also used to erect the su-perstructure segments. A general de-scription of progressive placing erec-tion is included in References 1 and 2.

All erection activities must be ac-complished from the completed portionof the superstructure. To keep fromfalling behind schedule, the contractor,in the early fall of 1981, decided to de-velop a method for erecting segmentsduring the cold winter months. Al-though the project is located in NorthCarolina, the 4500-ft (1373 m) elevationof the site brings in rather severe win-ter weather with snow, subfreezingtemperatures and very high winds.

Fig. 2. All substructure and superstructureerection is being done from the top. Here,a precast pier segment is being lifted bythe stiffleg crane.

PCI JOURNAL/September-October 1982 121

Fig. 3. Thermocouple locations in instrumented segments. The thermocouples wereused in the test program to evaluate joint heating.

Table 1. Test I Data, 10121181, Weather: sunny and calm.

Time Air Temp. Thermocouple Readings Comments

1 2 3 4 6

12:15 p.m. 50 48 47 46 45 43 Thermocouple#5 not working.

1.35 p.m. 56 49 49 48 52 45

2:25 p.m. 57 50 50 50 53 47

4:10 p.m. 56 51

52

51

52

51

52

58

56

49

526:15 p.m. 50

Note: Temperatures given in F. C = 519 (F — 32).

122

Table 2. Test II Data, 10/22/81, Weather: sunny and calm.

Time Air Temp. Thermocouple Readings Comments

1 2 3 4 6

8:15 a.m. 43 51 51 51 49 49 Blanket onovernight

9:00 a.m. 49 51 51 51 49 51No heat

Heat appliedapproximately

10:00 a.m, 50 51 51 51 49 49

11:30 a.m. 56 51 51 51 52 50

12:30 p.m.1:00 p.m. 56 53 53 52 55 51

Temperaturebetween blanket

1:55 p. m. 58 53 53 53 56 52

533:25 p.m. 56 56 55 56 56

and concrete = 65 F(18.9 C)

4:30 p.m. 55 58 58 59 56 54 Temperaturebetween blanketand concrete = 65 F(8.3 C)

Note: Temperatures given in F. C = 5/9 (F – 32).

The Linn Cove project, like all otherprecast segmental projects with epoxyjoints, is subject to an epoxy specifica-tion requiring the concrete temperatureto be above 40 F (4.4 C) for applyingand curing the epoxy. Therefore, a safe,economical method of heating andmaintaining the temperature of thejoints had to he developed.

TEST PROGRAMTo obtain information which would

assist the FHWA in evaluating con-tractor proposals for winter erection,the consulting engineers and the con-tractor cooperated in instituting a testprogram to determine the thermalproperties of the segments.

The FHWA is instrumenting thebridge, and had placed thermocouplesin three segments. One of the threesegments was used for the heating testswhile in storage. Fig. 3 shows the loca-tion of the thermocouples in a typicalinstrumented segment. Three testswere run on the instrumented segmentto determine whether the joints couldhe heated, and if so, at what rate.

TestTest I was run without any artificial

heat applied to determine the naturalrelationship between the concrete andambient air temperatures. Table I givesthe data taken during this test.

This test showed, as expected, thatthe concrete temperature followed the

PCI JOURNAL/September-October 1982 123

ambient temperature, but with somelag, The thin flanges of the box sectionallow a fairly rapid temperature in-crease in the concrete, with little gra-dient in the top slab. Because the seg-ment was sitting on the ground, thebottom slab showed a gradient. Thisimplied that the joints could probablybe heated by insulating and heating thesurrounding air.

The initial idea for heating the can-tilevered flanges of the segments was touse electric blankets. The contractorobtained two electric blankets whichwere 9 x 15 ft (2.7 x 4.6 in). The manu-facturer claimed the blankets wouldmaintain a substrate temperature of50 F (10 C) for an ambient temperatureof as low as 20 F (6.7 C).

Test IITest II was run with an electric blan-

ket wrapped over the edge of the seg-ment so that the concrete surrounding

the top slab thermocouples was heatedfrom both directions. No insulation wasprovided, and the thermocouples in thebottom slab received no artificial heat.Fig. 4 shows the test segment with theblanket as placed for Tests II and III.Table 2 shows the data for Test II.

This test showed that the blanketcould heat the concrete, but it was dif-ficult to determine how much of theheat was from the blanket and howmuch was from the sun. The tempera-ture rise in the bottom slab was equal tothat in the top slab. To eliminate theeffects of the sun, it was decided to runanother test at night.

Since Test II was started the morningafter Test I and the blanket was placedimmediately after the end of Test I, theinsulating properties of the blanketcould be generally evaluated. For anambient temperature drop of 7 deg F(3.9 C), the temperature of the concreteunder the blanket only dropped 1 deg F

. µ r Yj E OVA.:

I1 1 ? t E r.a^M"► 0

Fig. 4. The test segment being heated with an electric blanket. The blanket waswrapped around the edge of the segment and fastened.

124

(0.6 C). On the other hand, the temper-ature change of the unprotected bottomslab was 7 deg F (3.9 C), the same asthe ambient temperature change. It wastherefore concluded that the electricblankets would provide some insulatingvalue.

Test IIITest III was run with the same blan-

ket arrangement as Test II (see Fig. 4).As mentioned before, this test was runat night and somewhat later in the yearto test the blankets in the absence ofsun and with colder weather. Table 3presents the data taken for Test III.

The results of Test III were by far themost interesting. The data showed thatthe concrete could be heated evenunder severe conditions. The windconditions were ideal for observation ofheat loss, as demonstrated by the 12deg F (6.7 C) temperature drop in theuninsulated bottom slab. However, the

temperature rise was steady in the topslab and only dropped 6 deg F (3.3 C)in 7V2 hours after the heat was shut off.At this point in time the joints could beheated.

APPLICATIONThe contractor evaluated several

ideas for heating the joints before sub-mitting the final scheme for approval byFHWA. An insulating frame was builtto fit tightly around the bottom slab,webs, and the bottom of the wings. Fig.5, a photo of this insulating frame,shows two bar joists supporting a ply-wood form.

Insulation was provided by fasteningfiberglass batts to the outside of theplywood and attaching a foam strip onthe inside. The foam strip assured atight fit around the bottom of the boxsection and created a dead air spacearound the concrete.

Table 3. Test III Data, 11/6-7/81, Weather: cold with wind gusts to 60 mph.

Time Air Temp. Thermocouple Readings Comments

1 2 3 6 Gauge 4 notworking.

Heat appliedat 5:00 p.m.

Heat stopped.Blanket remained.

5:00 p.m. 47 54 53 54 —

6:05 43 54 53 54 46

7:15 p.m. 38 57 57 58 45

8:05 p.m. 35 59 58 60_- 44

9:20 p.m. 34 61 61 62 43

11:00 p.m. 34 65 64 65 41

6:30 a.m. 31 59 59 59 34

10:00 a. m. 41 57 57 47 34

Note: Temperatures given in F. C = 5/9 (F – 32).

PCI JOURNALJSeptember-October 1982 125

Fig. 5. The insulating frame was built to the general configuration of the outside of thebox girder. The out-to-out width is 4 ft (1.2 m).

Since the width of the frame was 4 ft(1.2 m), only 2 ft (0.6 m) of the segmentson either side of the joint would be di-rectly heated. The consulting engineerswere concerned that the resulting tem-perature gradient might cause a warp-ing of the segment to be erected. Toavoid this problem, the range of heatingwas limited to the allowable range ofthe Type III epoxy, which was 40 to25 F (4.4 to 18.3 C) for this project.

The contractor also decided, for thesake of the laborers, that erection wouldcontinue only when the weather was atleast 20 F (-6.7 C). Therefore, thesegment was analyzed for a possibledistortion due to a 45 deg F (25 C) tem-perature gradient. The amount of dis-tortion was within acceptable limits forthis project because the calculated dis-tortion was very small; experience hadshown the segments on this project tobe cast very accurately.

Propane heaters were used to heatthe bottom slab and webs. The rate ofconcrete temperature change was reg-ulated by the precast specification foraccelerated heat curing. The maximumconcrete temperature was 65 F (18.3 C),and the heaters could not blow directlyon any part of the concrete.

Baffles were placed on the inside ofthe segments so that only 2 1/2 segmentswere heated. This minimized the fuelconsumption and allowed better controlof the air temperature.

The top wing joints were heated withthe electric blankets extending thelength of wings. The portion of the topslab between the webs was coveredwith tarps. Since heat rises, we did notfeel this area needed insulating. In fact,we were concerned about it getting toohot, The tarps allowed enough heat toescape to keep the top slab within rea-sonable temperature limits.

126

The heating system was used in oneof two ways, depending on the concretetemperature when the erection processcommenced. If the substrate tempera-ture was above 40 F (4.4 C) but theweather forecast indicated that 40 F(4.4 C) would not be maintained, theepoxy was applied and the segment at-tached to the cantilever in the normalmanner. Then the insulating frame wasplaced and heat applied for a timeequal to twice that which was specifiedfor the epoxy to harden.

If the substrate temperature of theconcrete started below 40 F (4.4 C), thesegment to be erected was suspendedabout 6 in. (152 mm) from the end ofthe cantilever. Support was provided bythe stiffleg crane, and the temporarypost-tensioning thread bars were in-stalled and hand tightened as a safetyprecaution.

The insulating frame was thenbrought up and heat was appled untilthe concrete temperature was above40 F (4.4 C). Then the frame wasdropped, the epoxy applied, and thetemporary bars stressed. Finally, thejoint was again sealed by the frame andheated to assure hardening of theepoxy.

PROTECTION OF TENDONSThe minimum specified temperature

for grouting the post-tensioning ten-dons was 35 F (16.7 C). Since this tem-perature was not reached for a longenough period for grout to cure, a ten-don corrosion protection system had tobe used for the tendons installed andstressed during the winter months.Also, blockouts and open ducts had tobe sealed to prevent water from enter-ing and freezing.

A vapor phase inhibitor (VPI) wasused to protect against tendon corro-sion. VPI crystals were periodicallyblown with compressed air throughgrout tubes into the stressed tendons.The grout tubes were then bent over

and wired to seal the tendons. SinceVPI crystals have a definite effective-ness time limit, the engineers kept rec-ords on all tendons and required reap-plication before the VPI effectivenessexpired.

QUALITY CONTROLThe consulting engineering firm

handled all quality control for the erec-tion and heating of the segments. Ahand-held thermocouple was used todetermine the concrete temperature ofboth the segment to be erected and theend of the cantilever. Thermocoupleswere also placed in all joints heated.Joints were monitored throughout theheating period with a digital thermom-eter which had jacks for attaching thethermocouple lines.

The digital thermometer was alsoused to monitor the inside air temper-ature. In addition, a recording ther-mometer was used to record the airtemperature for the heating period.

For the first few joints, a close visualinspection was performed to detect anycracking caused by a temperature gra-dient through the webs or . slabs. Thisinspection confirmed that the designtemperature gradient was sufficient.

RESULTSFig. 6 shows the insulating frame in

place under the cantilever. This bridgehas a very complicated geometry, withS curves, transitions, and super-eleva-tions, but the foam gasket on the insideprovided sufficient adjustment to sea]around the joints as the curve shown inthe photo was negotiated. The framewas suspended by and traveled alongsteel rails which were temporarily at-tached to the underside of the curbareas.

The frame incurred only one mishap.A windstorm blew the frame off the endof the bridge, and it was damaged quiteextensively when it hit the rocks. How-

PCi JOURNAL/September-October 1982 127

Fig. 6. The foam on the edges of the heating frame provided a good seal as the framewas moved from joint to joint around the curve.

ever, by the end of the next day, it washack together and ready for service.

The heating blankets did not fare aswell. They were too fragile for thefolding and unfolding. Also, the foottraffic of the workmen, along withdropped tools and equipment, soon hadthe blankets in rough shape.

As an alternative, the contractor thenbuilt four wooden boxes, 4 ft (1.2 m)wide, which when placed end-to-endwould extend across the entire width ofthe deck. The boxes were insulatedwith styrofoam on the inside and eachwas wired for four electric heatinglamps (Fig. 7), This system was veryeffective for heating the top slab joint.Also, as long as enough replacementheat lamps were kept on hand, thissystem was much more durable_

All in all, the system was very effec-tive. From the middle of Novemberuntil the middle of March, over 20segments were erected. Without the

heating system, the project would havehad to shut down part time, and sub-stantially fewer than 10 segmentswould have been erected.

The heating system cost approxi-mately $4000, not including the fuelcost, which is impossible to determinebecause fuel was also used for otherpurposes. But in retrospect, the deci-sion by the contractor to develop theheating system proved to be beneficialand economical.

CONCLUSIONSWinter ection of precast concrete

segments is always difficult whenepoxy is used in the joints because ofthis material's inability to harden attemperatures below 40 F (4.4 C). How-ever, as demonstrated by this project,this inconvenience can be overcomeeasily and economically. This can be animportant consideration when contrac-

128

Fig. 7. Heating box, built with plywood and insulated with styrofoam, for heating the topof the joint.

tors are bidding precast segmental proj-ects in the northern part of the UnitedStates and Canada.

For those contractors and engineerswho may consider heating joints, thefollowing advice is offered:

1. Analyze the section to determinethat temperature gradients will notcause cracking in the concrete or sig-nificant distortion of the segments. Thiswas not a problem on Linn Cove, but itcould be a problem on other projects.The cross-sectional properties andlength of the segments are important.

2. Since the concrete temperature inareas other than the heated zones willmost likely be below 35 F (1.7 C) dur-ing winter erection, the post-tensioningtendons cannot be grouted. Plan on de-vising a corrosion protection system ac-ceptable to the owner. The post-ten-sioning supplier can help in this matter.

3. Use the project specifications asmuch as possible to control the heating.

Precast curing specifications work well.4. Exercise very rigid quality control

procedures. During cold weather,epoxy will appear to harden but willsoften as the temperature rises. Be verycareful about satisfying the minimumrequired heating time to insure that theepoxy in each joint is completely hard-ened.

5. Develop a system with durablematerials. When it is very cold outside,workmen usually take less time to carefor equipment.

6. Limit the concrete temperaturerange to one epoxy formulation temper-ature range. Overlapping epoxy formu-lations could cause the epoxy to go offtoo fast due to high temperatures orharden too slowly due to low tempera-tures.

7. Make the system simple and easyto follow. Both the initial cost and thecost of correcting mistakes will belower.

PCI JOURNAUSeptember-October 1982 129

making superstructure 77 ft (23.5 m) in one directional cantilever.

REFERENCES1. Muller, Jean, "Ten Years of Experi-

ence in Precast Segmental Con-struction," PCI JOURNAL, V. 20,No. 1, January-February 1975, pp.28-61.

2. Barker, James M., "ConstructionTechniques for Segmental ConcreteBridges," PCI JOURNAL, V. 25, No.4, July-August 1980, pp. 66-86.

CREDITSOwner: National Park Service.Contract Administrator: Federal

Highway Administration, Region 15,Sevierville, Tennessee,

Segmental Construction Expertise andDesigner: Figg & Muller Engineers,Inc., Tallahassee, Florida.

General Contractor: JasperConstruction Co., Plymouth,Minnesota.

NOTE: Discussion of this paper is invited. Please submityour discussion to PCI Headquarters by May 1, 1983.

130