Geometry Control of PrecastSegmental Concrete Bridges

Brice F. Bender, PESenior ConsultantSpecial Concrete BridgesHenningson, Durham andRichardson, Inc.Omaha, Nebraska

H. Hubert Janssen, PESenior Consultant

Special Concrete BridgesHenningson, Durham and

Richardson, Inc.Omaha, Nebraska

G eometry control needs to be ac-counted for from conceptual de-

sign through production and final erec-tion of segmental concrete bridges, Ifthis control is not adequately provided,the project will be unsatisfactory. Con-versely, if suitable geometry control isestablished in the beginning andmaintained throughout the project, ex-cellent results will be achieved.

Gegmetry control is requiredwhether the segmental bridge is precastor cast in place and erected in can-tilever. It is required whether thebridge is on tangent or curved, flat oron vertical curve. Control is also re-quired in other types of erection, buteach varies in some way. This discus-sion will concern precast segmental

NOTE: This paper is based on a presentationgiven at the Segmental Concrete Bridge Confer-ence in Kansas City, Missouri, March 9-10, 1982.The Conference was sponsored by the AssociatedReinforcing Bar Producers—CBS!, Federal High-way Administration, Portland Cement Association,Post-Tensioning Institute, and Prestressed Con-crete Institute.

bridges erected in cantilever sincemany bridges have been and are beingbuilt in this manner.

Time of ControlContrary to some thinking, geometry

control should not wait until construc-tion starts. Waiting until production orerection begins is a serious oversightand will cause problems and delays.Rather than starting control in the con-struction phase, the process shouldbegin in the conceptual design phaseand be carried through design, detailing,the casting curve, production plan, shopdrawings, form and manufacturing plantdesign, production of segments, curing,storage, erection of segments, andpost-tensioning.

Of course, the actual geometry con-trol as described later is accomplishedin the production and erection phases.Nonetheless, the designer should keepgeometry control in mind when de-veloping the structural concept, se-

72

Very close and constant control is necessary toobtain a completely satisfactory segmental concretebridge. This control must be established andmaintained in production and during erection ofsegments. Methods and amount of control arediscussed.

lecting the most likely erection method,developing details to facilitate the con-trol, and preparing the casting curve.Then the contractor must keep thecontrol aspect in mind while preparingthe production plan and the shopdrawings.

The contractor must certainly discussthe amount of control and methods toachieve it with the form designer. Con-trol during production and erection areusually understood, but one must notoverlook the curing and storage of seg-ments.

Changes in geometric measurementscan occur if just the newly cast seg-ment is steam cured, so it is recom-mended that both the new and the oldsegment be so cured. This amounts totwo time periods of curing for eachsegment and while usually not neededfor strength, maintaining the highermoist temperature does aid in geometrycontrol. Likewise, if segments are notset level on a firm storage pad withbearings under the webs, warping canoccur.

The designer must maintain a con-cern for geometry control throughoutthe design and detailing phases. Also,the designer must select a likelymethod of erection, especially in caseswhere erection loads will occur on thecantilevers, and compute accurate

casting curves which will become partof the plans.

Some might say that this is not neces-sary because the contractor might wantto change the erection scheme, the sizeof segments, post-tensioning or someother item which would alter thesecasting curves. However, it is highlydesirable to include this information inthe plans because it will inform possi-ble bidders and will direct attention to,the importance and need of geometrycontrol. Moreover, most bidders do nothave the staff, capability or desire to re-vise the plans.

It has been noted that with complete,detailed plans, the contractor usuallyuses most of the design and sometimesall of it. Also, it must be rememberedthat these complete plans include ten-don layouts for all permanent post-tensioning, adequate temporary post-tensioning used during erection, andthe sequence of tensioning, as thiscould alter the geometry in certaincases.

With these well conceived and de-tailed plans, all contractors can prepareaccurate bids. They will be forewarnedof the need for geometry control inthe construction phase, and if theyshould choose to change some item,they will know what is required in theirsubmittal. In preparing bids, the em-

PCI JOURNALJJuIy-August 1982 73

QABUT PIER SPLICE PIER ABUTII I i I ^

SELFWEIGHT & ' CANTILEVER T^NDONS

2

ROTATION & EN SPAN CONTINUITY TENDONS

13 1

CORRECTIONS

CBRIDGE AS ERECTED

5

SUPERIMPOSED 'DEAD LOADS &I CREEP

A

1PROFILEGRADE

Fig. 1. Various steps in correcting erection deflections.

phasis on control in the plans and spe-cial provisions will inform the formmanufacturers of the care they musttake to provide an adequate castingmachine.

The successful contractor can im-mediately begin the work by orderingthe form, preparing the productionschedule, and preparing shop drawings.With the casting curve and other itemson the plans, shop drawings can bestarted with no other computations ifthe design plans are used. Thegeometry control functions of dead loaddeflection, post-tensioning camber,

creep and shrinkage are included in thecasting curve. These, combined withthe structure geometry, provide the in-formation for each segment shopdrawing.

With adequate forms installed in acomplete manufacturing area, and withapproved shop drawings in hand, theproduction of segments can begin.Anyone who is familiar with segmentalprecast concrete bridges knows thatthe segments must be match cast. Thisis elementary, but most important.Likewise, all prudent engineers willrealize that a box girder constructed on

74

OUTSIDEFORMWORK

SOFFIT

CARRIAGE —`I L— ADJUSTMENT JACK

Fig. 2. Formwork (Short-line method).

INSIDE FORMWORK

BULKHEAD

NEW OLDSEGMENT JSEGMENT

SOFFIT

CARRIAGE

Fig. 3. Step 1 (Short-line method).

Fig. 4. Step 2 (Short-line method).

PCI JOURNAL/July-August 1982 75

iGET

11141 1\VIII I\ 1

BASE ELEVATION

MARKER

Fig. 5. Alignment control (Setup).

the ground in a perfect geometric pat-tern to that desired on the piers cannotbe erected in cantilever without severalcorrections to compensate for dead loaddeflection, prestressing cambers, creep,and shrinkage. Fig. 1 indicates the needfor these corrections. Each segment isintentionally deformed during produc-tion to achieve the desired results.From this follows the importance ofgeometry control in production,

Production ControlLet us now follow through the casting

procedure to indicate how the required

LEVELBOLT

geometry control is achieved. This se-quence is predicated on the use of theshort-line method. Figs. 2, 3, and 4 de-pict the various steps in the case of theshort-line method of production.

F'ig. 5 shows the setup required inthe casting area. Especially note the in-strument on a permanent base and apermanent target (Fig. 6). It is very im-portant that neither instrument nortarget be disturbed throughout theproduction. In case of a disturbance,control must be reestablished so ade-quate bench marks must be provided.Note that in Fig. 5 a new segment hasbeen match cast against an old segment.

76

The form is always plumb, level andsquare so the geometry control is es-tablished by properly setting the oldsegment.

But to begin, the first segment mustbe cast. This might be a pier segment,but it could be any other element thatthe contractor's scheduling indicatesshould be first, Since there is no oldsegment, a second bulkhead is requiredand the segment is cast plumb, level,and square. As the top slab is beingfinished, the four elevation bolts, A, B,C and D, as well as the centerlinemarker are inserted. The followingmorning the levels on the bolts are re-corded and the centerline is scribedonto the centerline markers. Now thetemporary second bulkhead can he re-moved and the segment rolled forwardfor match casting (see Figs. 7, 8, and 9a,9b).

Here it is desirable to interject theidea of permanent record keeping. Theinstrument person should be on the job Fig. 6. Instrument set up (Zilwaukeedaily, have the complete confidence of Bridge, Michigan).

Fig. 7. Match-cast segment showing elevation bolt.

PCI JOURNAUJuIy-Augusd 1982 77

' ^fflr ^^st t ¢

.r, wR } !±

Fig. 8, Match-cast segment showing centerline marker.

G MroE2w

the superintendent, and keep accurate,daily records.

When the first segment is rolled for-ward, it is reset to the information onthe shop drawings. The centerline willbe identical as to what it was the previ-ous day, unless the bridge is curvedand then an offset angle is used as

shown in Fig. 9a. Likewise, verticalcurvature is handled as shown in Fig.9b. But in addition to the vertical cur-vature adjustment, if any, it is necessaryto adjust for the segment deformationsas mentioned before and illustrated inFig. 1.

These deformations will be on the

HORIZONTAL CURVATURE

OLD SEGMENT

y i SEGMENTROTATED

_CENTERLINE ON

FORM ALWAYSAMOUNT OF

PLUMB,LEVELSAOUED ROTATIONDETERMINESHORIZONTAL

N CURVATURE.PLA

Fig. 9a. Match-casting operation showing horizontal curvature of segment.

78

shop drawings, and can be taken di-rectly from the plans unless changeswere made to these. In that case, thecontractor, or his designer, must re-compute them. After this first (the old)segment is properly reset and the setupfor the second segment is complete, in-cluding reinforcing bar cage, tendons,anchors, embedded items and the posi-tion of the inner mandrel, the secondsegment can be cast. This second seg-ment is called the new segment in Figs.5 and 10.

The following morning it is necessaryto mark the centerline and record thebolt elevations of segment number two.Also, before segment number one (theold segment) is moved, it is imperativeto recheck the elevations of its boltsand its centerline position to determineif its position has changed since it wasset so carefully the previous day. It isoften noted that position changes occurdue to settlement of the soffit rails bythe segment weight, or due to vibratingof fresh concrete against it, or clue totemperature changes.

There is no need to be concernedabout these small changes as long asthese are known and recorded everyday. From these, daily compensationscan he made. It is only from neglected

Fig. 10. Positioning the old segment(Zilwaukee Bridge, Michigan).

or accumulated errors that the geometrybecomes jeopardized. As compensationis made daily, the bridge superstructure

VERTICAL CURVATUREBOLTS A

AMOUNT OFTILT DETERMINESVERTICAL

URVATURE.

ELEVATION

Fig. 9b. Match-casting operation showing vertical curvature of segment.

PCI JOURNALJJulyAugust 1982 79

HORIZONTAL ALIGNMENT

5-0- + I 1 4 I

19 8`-0" 4-0

VERTICAL ALIGNMENTFig. 11. Graphical production chart showing horizontal and vertical alignment.

deviates both vertically and horizon-tally from the theoretical alignment byvery small amounts. However, if thesesmall deviations are left uncorrected,they will soon magnify and create anunsatisfactory condition.

The instrument person should main-tain a graph of these corrections fromthe records as shown in Fig. 11.

After the alignment of the "old seg-ment" is recorded, it can be removedfrom the soffit and moved to properstorage. Many times segments are keptone more day in the manufacturing areato complete transverse post-tensioningand clean -tip.

The setup to cast segment three, andall other segments, is similar to that ofcasting segment two with one excep-tion. When segment two is rolled for-ward, it is set to the proper position asper the drawings, but this time it mustalso have a small compensation for aposition change of segment one that oc-

cured during casting. If segment onehad moved '/s in. (3.2 mm) to the left,segment two will now he adjusted 4 in.to the right. The same procedure holdsfor any vertical adjustment. This showsthe importance of constant surveillance.

Superelevation ControlAnother geometric problem is han-

dling superelevation for horizontal cur-vature. If the superelevation is constantwithin a span, no transverse tilting ofthe old segment is required in setting itfor match casting. Rather, the superele-vation is achieved by setting the piersegment on the proper tilt as shown inFig. 12. However, as indicated in thebottom portion of Fig, 12, the tilting ofthe pier segment does create an unde-sirahle deflection which must be cor-rected by introducing counteractingvertical curvature. Likewise, if thestructure is on a strong vertical curve,

80

SUPERELEVATION (CONSTANT)

1J PER

FORM ALWAYSPLUMB, LEVEL Ak SQUARE

//CHANGE REFERENCE PLANE FOR^^\\GEOMETRY CALCULATIONS SURE RELEVATION

ACHIEVED BY TILT

eo^re. e. rna.0 11T

Fig. 12. Superelevation (constant).

an undesirable horizontal curve will betreated, which must be corrected byintroducing a compensating horizontalcurvature.

More complicated is handling asuperelevation transition that occurswhen the superelevation is increasingor decreasing in a span. Now the oldsegment needs to be tilted with respectto the new segment as shown in Fig.13. While the permanent bulkhead re-mains plumb, level and square, the restof the form must assume the transition

to the tilted position of the old segment.While this is usually a very small twistin any one segment, it is a problem forthe form manufacturer to design a formwhich can accomplish the twist andremain structurally rigid for the con-crete casting and still maintain a mortartight seal at the point of joining the oldsegment. The practical point of rotationis the centerline of the bottom slab.

A consequence of transverse tilting ofthe old segment with respect to thenew segment is that the centerline of

PCI JOURNAUJuIy-August 198251

'OLD" SEGMENT TILTSIN RESPECT OF "NEW"

'A^^

"B"

BULKHEAD OF FORM I "REMAINSPLUMB, LEVEL a SQUARE FORMWORK IS

FORCED TO ASSUMETILTED SHAPE OF

FORM—^i OLD SEGMENT

WN

PREDETEF

Fig. 13. Superelevation Transition I during casting.

the top and bottom slabs are no longerin a vertical plane. It is ideal if the in-strument is in a position to establishboth the bottom slab centerline and thetop slab centerline. Since the centerlineof the top slab is offset a predeterminedamount, it is essential that this be mea-sured in the setup.

Fig. 13 shows that the bottom slab

centerline remains constant duringproduction. Fig. 14 indicates the rota-tion from no superelevation to fullsuperelevation. According to the pic-ture in the middle of Fig. 14, the seg-ments in the transition area will havesome small irregularities which arecaused by warping of the form.

In order to line up top slab center-

82

6aWJ

IrLU

DV7.J \\

z0

I-

zzQ

za

ROTATED

BOTTOM

ROT,

DEC

_ BOTTO

IN CASE SUPERELEVATION TRANSITION EXTENDS OVER SEVERALSPANS: DETERMINATION OF CORRECTIONS LEADS EASILYTO ERRORS BECAUSE OF REVERSAL OF PIER SEGMENT.

CONCLUSION: MAKE PROPER DRAWINGS OF r. BOTTOM ANDTOP SLAB IN THESE AREAS.

Fig, 14. Superelevation Transition II during erection.

lines instead of bottom slab centerlines,segments at the beginning and end ofthe superelevation transitions need tobe rotated.

This subject can be more complicatedthan shown in Figs. 13 and 14 becauseusually the transition will start within aspan and continue over several piers.This is complicating in itself, but the

complexity increases because the pro-duction often calls for a reversal of thepier segment after completion of half acantilever. In order to avoid errors,therefore, it is mandatory that propershop drawings be made clearly showingbottom and top slab centerlines.

It follows from the above that align-ment control is a matter of combining

PCI JOURNALJJuIy-August 1982 83

Fig. 75. Temporary support of pier segment group (Zilwaukee Bridge, Michigan).

the calculated deformations with thestructure geometry on the shop draw-ings and:

L Implementing these at the castingsite.

2. Recording each morning the cor-rections achieved.

3. Comparing the corrections withthe requirements.

4. Correcting the corrections.

Erection ControlIt is necessary to consider alignment

control during erection of segments.First, an erection table, chart, or graphmust be prepared after the segmentsare produced. This provides thetheoretical position of each segmentbased on actual production mea-surements, The same leveling bolts areused and offsets from an establishedcenterline provide horizontal control.Transverse or horizontal tilt or rotationis also on the theoretical chart.

The designer of the structure should

provide information regarding the ele-vation of each segment after erection ofthat segment. If this information is pro-vided, the actual position of the seg-ment can be checked against thetheoretical. Readings should be takenas the segments are erected in order tonotice discrepancies, if any, However,the only true readings that can be ob-tained are those taken early in themorning before the sun's heat has hadtime to affect the girder.

This effect is very pronounced, and iscaused by the fact that the top slab ofthe box girder heats rapidly while thebottom remains cool. This conditionwill cause substantial deflection of thegirder. Each morning after accuratereadings are taken, a determinationmust be made whether any correctionsare necessary in erecting the next seg-ment.

Usually only two segments are placedat each end of the cantilevers in a workday. Hence, there is an accurate read-ing after every two segments. From this

84

Fig. 16. Vertical sag (Turkey Run Bridge, Indiana).

it can he decided whether any field cor-rections are required.

If small erection corrections are de-sirable, shimming is recommendedwith stainless steel screen wire meshabout'/ in. (1.6 mm) thick, This allowsthe epoxy to flow through and preventsstress concentration points, yet it willeffectively turn the next segment tip,down or laterally. It is recommendedthat shimming not be started too soon.Only after a number of segments havebeen erected in a cantilever, and afterthe temporary bearing assemblies arechecked for any discrepancy, can ajudgment of the alignment be made.And if measurements indicate that theerected structure will be out of toler-ance, careful use of shims will regulatethe problem.

It is advisable to start erection of thepier segment or group of segments ontemporary bearings. These temporarybearings are designed to allow the can-tilevers to be adjusted after the seg-ments are erected and then let down

onto permanent bearings. It is very im-portant that the pier group of segmentsbe precisely set because an error madein this will be multiplied upon com-pletion of the cantilever (Fig. 15).

Other items affecting geometry dur-ing erection are provisional post-tensioning, post-tensioning of perma-nent tendons, and midspan splices.Both provisional and permanent ten-dons are a design function, and the useof correct number and correct placingare a production responsibility. How-ever, the sequence of tensioning can beimportant and should be established inthe work check list for the erectingcrew. Before the splices are cast, thecantilevers must he aligned. Structural.hardware should he provided to holdthe cantilevers in line during midspansplice casting and continuity post-tensioning (Figs. 16 and 17).

Further information on the subject ofgeometry control in segmental con-struction can be obtained from Refer-ences 1 through 6.

PCI JOURNALJJuIy-August 1982 85

Fig. 17. Horizontal curve (Zilwaukee Bridge, Michigan).

CONCLUSIONSFrom the preceding discussion the

following conclusions can be stated:I. Geometry control is most defi-

nitely needed for a satisfactory seg-mental bridge.

2. Control is essential in all stages ofthe work from design through castingthe final midspan splice.

3. Control functions are relativelysimple, but it is more important that

control be started as each phase startsand does not cease until that phaseends.

4. Cooperation must exist betweengroups responsible for various tasks. Noone group has more or less responsibil-ity than another. Teamwork is essential.

5. Careful, complete control will in-sure that the bridge will meet the seg-ment criteria.

REFERENCESI. Precast Segmental Box Girder Manual,

published jointly by the Prestressed Con-crete Institute and the Post-TensioningInstitute, 1978, pp. 66-68.

2. Lovell, J. A. B., "The Islington AvenueBridge," PCI JOURNAL, V. 25, No, 3,May-June 1980, pp. 32-66.

3. Bridges, Conrad P., and Coulter, CliffordS., "Geometry Control for IntercityBridge," PCI JOURNAL, V. 25, No. 3,May-June 1979, pp. 122-125.

4. Kelly, John B„ Janssen, H. Hubert, and

Walker, Homer M., "The Kentucky RiverBridge," PCI JOURNAL, V. 26, No. 4,July-August 1981, pp. 60-85.

5. Bender, Brice F.,"Economics and Use forLightweight Concrete in PrestressedStructures," PCI JOURNAL, V. 25, No, 6,November-December 1980, pp. 62-67.

6. Joint PCI-PTI Committee on SegmentalConstruction, "Recommended Practicefor Precast Post-Tensioned SegmentalConstruction," PCI JOURNAL, V. 27, No.1, January-February 1982, pp. 14-61.

86