Dem olition of post-tensione post-tensione d concr ete Safe release of stored energy requires a reverse sequence of the original construction procedures
L
ittle has been published about the demolition of posttensioned concrete but there is enough such concrete now in service to become a problem for demolition contractors. When a d emoli tion sequence sequ ence is being planned it is important to h ave a knowled know led ge of the ere cti on p ro c e d u re and a set of the as-built as-bu ilt record dra wings. This Thi s fac f actt of o f lif l ife e was conf irmed recen tly when 9 prestressed beams had to be demolished and extensive repairs made to other beams within the decks of bridges under construction for a highway in England. Exact identification of prestressing cables along each beam was of great import impo rtance in the demolition process and fundamenta dame ntall to the remedial remed ial stressing. The bridges concerned we re built with 3-spa n continu co ntinuous ous post-t p ost-tenensioned segmental l-beams, with a compo site pre stressed deck slab. Each beam had been pre s t ressed to a load of about 750 75 0 tons. (1)* This article describes how the beams we re s e vered a nd re moved from the adjaadj acent decks deck s and discusses dis cusses rep la cement ment pro pro ce du res bri br i ef ly.
Construction details An attra c ti ve design desi gn had been produce rod uced d for f or ten te n 3-span 3-sp an ove rpasses. With one exception, exception, each had a central span of o f 121 12 1 feet(2) and end spans of 42 feet.(3) The main beams we re precast in five f ive segments segme nts and erected on their permanent bearings at the * Numbers in parenthes es refer to metric metric eq uivalents uivalents listed w ith this this a rticle. rticle.
The crack in this C segment extended in length and width after the string course was complete. The crack follows the position of the displaced upper upper cable duct. Bridge 35 west parapet parapet beam.
bankseats and piers, with two temporary trestles trestle s supportin supp orting g three thre e segseg ments in the main span. All segments men ts we re cast cas t to suit sui t a ve rti c a l curve of 7,000 -foot(4) ra dius, but bu t all al l (5) straight grades gra des and an d 3,000-foot 3 ,000-foot radii were accommodated accommodated by adjustment of the relative joint joi nt leve lev els. A large range of deck d eck widths was obtained in the va rious bri dges by using 3, 5, 6 or 8 beams at 6- to 7-
foot(6) centers with a composite prestressed deck d eck slab 10 inches i nches(7) thick. The two main cables, each containing 19 high-strength strands of 0.6-inch diameter(8) we re winched winc hed through the 4-inch-diameter (9) c orrugated ducts, which we re designed to follow a continuously curved prof i l e. After connecting and sealing the cable ducts at the joints between segments, the gaps, nominalnomin al-
ly 4 inches,(10) were concreted at the same time as the intermediate, pier and end diaphragms. Pre s t ressing the cables was accomplished by pulling all 19 strands simultaneously using double-end stressing with identical pre stressing jacks of 500-ton(11) c a pa ci t y. Fi r st stage prestressing consisted of applying a load of about 390 tons(12) to the lower cables in each beam. Cable extensions we re checked at each jack; for the standard length bridge a total extension of about 14 to 15 inches(13) was recorded. The applied load caused the beams to lift approximately 1 inch(14) at midspan and made it possible to rem ove t e m p o rary supports immediately. The beams were fixed at one bankseat and piers we re hinged at the base and pinned into the beams so that the total elastic deform ation was allowed for at the slidi ng bearings on the other bankseat. Shortly after first-stage pre stressing, the deck slab was cast either in 3 bays or in a single operation, depending on the width of the bri dge. The upper cables we re subsequently stressed in a similar manner to the lower gro u p, but of course this second stage of prestressing acted upon the fully composite section. The projecting lengths of highstrength strands we re trimmed back close to the anchorage blocks, the end recesses concreted and then both cable ducts grouted with ord inary portland cement, water and an expanding agent to countera ct shrinkage. The superstru cture was completed by casting in place re i n f o rc e d c onc rete string courses along the edge of the parapet beams.
The cracking that led to demolition Du ring pre stressing operati ons on the first three bridges completed, a series of cracks was detected in the central spans. Most cracks seemed to start at the intermediate diaphragms and follow the line of the cable ducts. At first these cra c ks we re not wide enough to warra nt much concern, but as a safety mea-
Cutt ing cables in the web at the north intermediate joint of an internal beam of Bridge 34 using a t hermal lance.
sure the next two bridges we re pres t ressed in 50-percent and 25-percent increments of final load to even out any local effects. Initially there was little cracking on these two b ridges but as time passed cra cks began to appear and grow in length. The signif icance of these cra ck s was reassessed when a part icularly bad one on a parapet opened to 1 ⁄ 8 inch(15) in width shortly after the parapet was cast. X-ray ra diog rap hs taken through the 10-inch-thick (16) web re vealed the upper cable duct had floated some 18 inches(17) up ward from its design position. This discovery led to a systematic check of all beam segments.
Demolition of fully grouted beams A major concern was t o re move the condemned beams without significantly damagi ng or alteri n g s t ress conditions in adjacent beams. When main cables were being cut the sizes and lengths of all existing cracks on the remainder of the bridge we re noted before, during and after the demolition proced u re s.
South half of an A segment being lift ed out by a 40-ton (3 6-met ric-ton) crane. One internal beam and both parapet beams were replaced in t his deck, leaving only two original beams in the final structure. Bridge 24.
The demolition method adopted was to re verse the construction sequence by isolating the affected beam and cutting it into the ori ginal simply supported segments. These segments we re split furt he r into half-lengths to ensure complete loss of prestress in the cables and to p rovide convenient 15-ton(18) p ortions for removal by crane. From the first three bri d g e s, each of which had been fully stressed and grouted, five parapet beams and three internal beams had to be demolished. One parapet beam was removed from the fourth bri dge, but it had only been constructed to the point where the deck slab had been cast after first-stage prestress. Tem p o rary supports we re ere cted under the condemned beams and positioned at each end of the 10 half-segments. The intermediate piers and areas below were protected from falling pieces of concrete by means of scaffolding and timber decking just below beam soffit level. As a safety precaution, the immediate bridge area was fenced off at ground level to channel the tra ff ic b e l ow the brid ge at we l l - d e f i ne d positions.
Demolition of parapet beam The first beam to be demolished was the west parapet beam of Bridge Number 35. Grouting operations had proved rather tro ub lesome on this beam, so it was feare d that there might be a number of voids in the mid-span region which could lead to a violent release of the energy stored in the cables. The va rious positions at which the deck slab, beams and main cables we re cut and the typical manner of removal of a parapet beam are shown in the figure. First the cast-in-place re inforce d concrete string course was remove d by breaking out a longitudinal slot in the deck with heavy pneumatic hammers and burning through the t ran s verse rei nf orcement with acety lene torc h es. Pieces of stri ng course about 10 feet(19) long we re gently removed by holding them with a crane while the last few transverse bars we re cut. The next step was to cut slots in the deck slab about 6 feet long (20) by 12 to 18 inches(21) wide over the position of the intermediate dia p h ra g m s. The slots we re situated close to the edge of the upper flange
so that the tra nsverse deck steel could be cut with the maximum lap length available for connecting into the replacement beam. By bending this deck steel up into a ve rtical position access was given to the intermediate diaphragms from above. The diaphragms were then bro ken out with pneumatic hammers, leaving the bottom diaphragm steel to act as a flexible tie into the remaining deck and be reused with the replacement beam. The longitudinal slot in the deck was extended adjacent to the beam for its full length except for bri dging pieces 30 inches(22) wid e. These pieces we re left at about 8-foot(23) and 12-foot(24) centers alternately, so that there we re two re straining deck slab ties to each half segment. Pier and end diaphragms were b roken out at virtually the same ti me, with an effort again being made to leave the main diaphra g m steel for reuse with the new beam. Then came the critical opera tion of cutting the main cables; an acetylene torch was used on the first beam. The cables we re carefully exposed with small pneumatic hammers, the intention being to cut out only a minimum amount of concrete to gain access. Pre stressing cables were seve red in a strict sequence as detailed on Section A—A, which shows a halflength of a typical 3-span beam. At pier diaphragms. As each cable was being burned the last few strands snapped with a sudden release of energy and the cable ends m oved apart as the bond with the g rout failed. The joints between A and B segments opened out at the top as expected and the segments rotated downward as continuity was lost over each pier. Longitudinal cracking appeared along the junction of the web and upper flanges on both sides of each pier but no spalling of the concrete occurred. At intermediate diaphra gms. It was o bvious at th ese points that a great deal of energy remained in the cab l es, since substantial shocks were heard and felt throughout the remaining deck. Si mul ta ne ousl y,
the beam webs cracked heavily along the line of each duct but again there was no instance of any beam concrete bursting away. At midspan of C segment. Little energy remained in the cables after the first four major cuts so no obv ious change occurred in the existing cracks in this segment. At midspan of B segments. Cu tting the tendons at these positions caused some local increase in cracking; the segments were cra cke d along their entire length. At midspan of A segm ents. The end anc horage s had re m a i n e d untouched so some energy still remained in these segments. Cracks developed throughout the A segm ents al on g the li ne of the cable ducts. Once cable cutting was completed, each beam segment was divided at the cut position and lifting gear was attached to the half segments. The remaining bridging pieces of deck we re then broken out so that the segment could be lifted clear of the support s. Although the effect of cutting the cables was undoubtedly tra nsmitted through the adjacent deck in the form of an impact force, no obvious signs of damage or worsening of cracks were detected in other be am s. The above proc ed ure was t h e re f o re repeated for subsequent parapet and internal beams, except that cable and concrete cutting was much improved with the use of a t h e rmal lance. This eliminated tedious exposure of the cables with pneumatic hammers, speeded the cutting sequence and allowed the operator to stand well clear. Cables became hot only around the cutting position since it did not take long to cut a cable. Thus the cable tension did not appear to relax significantly due to expansion and a similar mode of cracking was observed after each sudden release of energy.
Demolition of ungrouted beam The ungrouted parapet beam was isolated from the adjacent deck in a manner similar to the method used
on the fully grouted beams. Eve n though the lower cable was ung routed and the cable ends had not been cropped off close to the end b lo cks, destressing was not completely stra i gh tf orwa rd . The multistrand system could not be detensioned with the 500-ton(11) jack used for tensioning, so individual strands were destressed by a small jack that could release the wedges and allow their withdrawal. The additional friction betwe e n strands required that forces up to 93 percent of the ultimate strength of a st rand had to be applied before some wedges could be re l e ase d. This level of ove r s t ress is well above the allowable limit of 77 percent of the ultimate tensile strength applied during stressing operations. Consequently the cable was no longer safe for use and could not be incorporated into the new beam.
Construction of replacement beams The deck slab adjacent to the demolished beams was trimmed back to provide a suitable 2-foot 6-inch (25) minimum lap for the main transverse deck steel. Since pier positions were now fixed by the existing deck, new sliding beari ngs were necessary at the piers to accommodate the rela tive movements of the new beams during stressing but no change in the b ea ring type was re q ui red at the bankseats. The general re construction pro c e d u re followed a pattern similar to that used for the original beams except for the following modifications. Joints between the beam segments we re cast without completing the d i ap hragm connections with the adjacent deck; the new deck slab was cast leaving a 6-inch (26) gap between the old and new slabs, so that ve rtical and hori zontal move ments could freely occur in the new beam during second-stage stressing; after waiting for a minimum period of 7 days to permit some further creep and shrinkage movements to occur, the gaps in the deck and tra nsverse diaphragms we re finally concreted.
The results of tests perf o rmed after re co nstruction had been completed indicate that both the demolition and remedial stre ssi ng p roc e d u res used on these bri dge s had been successful.
6. 1.8- to 2.1-metre
20. 1.8 metres
7. 250 millime tres
21. 300 to 450 millimetres
8. 15.2 millimetres
22. 760 millime tres
9. 100-millimetre-diame ter
23. 2.4-metre
10. 100 millime tres
24. 3.6-metre
11. 455-metric-to n
25. 0.76-metre
12. 350 metric tons
26. 150-millimetre
13. 355 to 380 millimetres
Metric equivalents
14. 25 millime tres
1. 680 metric tons
15. 3-millimetre
2. 36.9 metres
16. 250-millimetre-thick
3. 12.9 metres
17. 460 millime tres
P UB LI CATI ON #C760256
4. 2100-metre
18. 13.6-metric-ton
5. 910-metre
19. 3 metres
Copyright © 1976, The Aberdeen Group All rights re served
