The supports were designed as rigid reinforced concrete frames and, except at the wide section of the bridge, consist of four circular columns and a capping beam at each grid. Twin supports were provided at the wide section to minimise thermal and shrink- age stresses.
Short spans are supported on 3ft. 3in. diameter columns, the diameter being increased to 4ft. for longer spans. At each grid the middle two columns were built slightly higher than the outer ones to provide for a 1:40 crossfall on the bridge deck.
All capping beams are in the form of an inverted T-beam, the flange of the beam providing seating for longitu- dinal beams. For aesthetic reasons the soffits of the capping beams were made flush with the longitudinal beams.
The flanges of the capping beams are the most highly stressed sections of the structure and enough reinforce- ment was provided to take the whole of the shear and torsional stresses.
Columns were cast in steel moulds and, since no construction joints were permitted, each was concreted in one operation using elephant trunking. A very satisfactory finish was obtained throughout and no other treatment was required except for a final washing down.
Plywood shuttering was used for capping beams which were also cast without construction joints.
Prestressed beams
For economical and practical rea- sons, beams longer than about 60ft. for a beam-and-slab type of bridge deck are usually stressed by employing
a post-tensioning system of prestressing. Economically, there was little dif- ference between a number of the better-known systems considered, and for the particular set of beams, Cable Covers Ltd.'s "Cabco" multi-strand system for seven 1⁄21⁄2in. diameter strands was selected mainly because of its simplicity and ease of operation in the field.
Two types of beams were needed edge beams of box section and internal ones of l-section. Altogether 358 beams were precast using 6500/%in. grade concrete.
The layout and number of casting beds was determined by the anticipated speed of handling of various beams and, on examination, it appeared that the turnover on 90ft. long 1-beam beds would be critical. A programme was evolved which called for a production of one I-beam per day and, although there were doubts in the early stages that this could be met, in practice the rate was achieved without difficulty.
Casting beds were laid down in 6in. concrete on a prepared crushed stone base. An 1/8in. thick steel plate on top of the concrete formed the working soffit. Special fittings were provided at the end of each bed to allow free rotation of the beams when cambering under stress.
Fabricated steel panels were used as side shutters and all joints between the panels were sealed with rubber gaskets to prevent leakage of grout. When casting, concrete was compacted using internal and external vibrators.
Sheathing for the 2in. dia. ducts, normally obtained off site, was manu- factured on the site. A machine pro-
vided for the purpose formed it from lin. wide strip metal supplied in reels.
Tendons were made up by running in. diameter strand on to benches placed alongside a casting bed. It was cut to length with an electric carborun- dum wheel and the ends marked up in distinctive colours for ease of identi- fication when stressing. Each group of seven strands was bound together at intervals by light gauge wire and sheat- ing was then passed over it, great care being taken that there was no twisting of strand. No spacers were used inside the ducts because all tendons were of simple and gentle curves.
Stressing
The specification called for stress- ing each strand from both ends simul- taneously in order to reduce frictional losses and achieve a prestressing force profile symmetrical about mid-span of the beams. Early in the contract, and in order to effect economies, it was thought that the same result would be achieved by stressing each strand from one end only, provided that alternate tendons of similar form in each beam were stressed from opposite ends with enough overstress to offset the in- creased friction.
In practice, serious slipping of strand occurred at the dead-end an- chorages and it was decided to revert to double-end stressing. The reason for slipping was concluded to be the very rapid formation of rust between grips and strand preventing adequate hold on the latter. By stressing from both ends, the grips were inserted whilst the strand was under highest stress and brought firmly into contact
A
Far East BUILDER, April 1969
Edge beam being lowered; bogies on capping beam
35