In the UK there are few alternatives to de-icing salts to stop roads freezing over in winter. But the chloride ions from the salt can penetrate concrete structures and attack the steel reinforcement, leaving engineers with the task of dealing with the consequent spalling through expensive maintenance programmes. Canadian winters are even worse, but three eminent Canadian bridge engineers believe they may have found the solution by omitting the reinforcement altogether.
Reinforcement is unnecessary, Dr Baidr Bakht, Dr Aftab Mufti and Dr Leslie Jaeger believe, if the compressive membrane or 'arching' action of the deck slab is utilised.
The theory is not new. In fact, 40 years ago engineers carried out tests which showed that interior floor slabs could carry up to four times the predicted load before they failed. But, until now, engineers have been reluctant to rely on arching action as the main load-carrying mechanism in a structure.
Arching action develops because of the inability of concrete to cope with tensile strain. Under load, a simply supported concrete beam or slab will develop tension cracks as its soffit tries to expand longitudinally (see diagram). But if it is fully restrained it cannot expand, and a corresponding force, or arching thrust, develops.
If the beam or slab is acting as a two-way structural element, the arching will take place in two directions, causing a dome effect known as compressive membrane action. The extent of arching action depends on how well the beam or slab is restrained.
The three engineers have been studying arching action for 20 years. They now claim that, with effective restraint, an unreinforced bridge deck slab will generate compressive membrane action which easily copes with standard traffic loads. They also say that flexural strength is no longer the critical mode of failure, but restrained slabs with enhanced arching action will fail at very high loads under highly localised punching shear.
Utilising arching or compressive membrane action depends on the level of restraint. Bakht, Mufti and Jaeger say 'a deck slab supported by parallel longitudinal beams does not require any reinforcement, provided the slab is suitable confined in both the longitudinal and transverse directions'.
In the longitudinal direction, the deck slab can be restrained by making it composite with the beams and by providing edge beams with suitably high flexural rigidity in the plane of the slab. Transverse restraint can be provided by making sure the supports of the slabs (that is, the top flanges of steel girders or concrete beams) cannot move laterally by tying them together with, for example, steel straps.
Additional requirements are that the deck is of uniform thickness, with haunches at the beam locations.
Design provisions for steel-free deck slabs have been incorporated in the next edition of the Canadian Highway Bridge Design Code, due out in the autumn. The new code will stipulate that a steel-free deck slab must have a minimum thickness of either 175mm or one fifteenth of the beam spacing, whichever is the greater.
There is also a formula for calculating the cross-sectional area of the transverse confining system (for example, steel straps) which takes account of: the spacing of the straps; thickness of the deck slab; spacing of the supporting beams; and the modulus of elasticity of the strap material. Larger straps are required for outer panels in the deck.
Straps can be made out of a range of materials including various types of steel or carbon fibre. 'In reinforced concrete design you put reinforcement in for strength,' Mufti says. 'But here the straps are for stiffness - that's the big difference.'
During initial tests, the Canadian trio chose to weld the weathering steel straps to the underside of the flanges of steel girders. To avoid overhead welding on site, they adapted the system by fixing the straps to the tops of the girders. In the latest development, the straps are manufactured with shear studs at the girder locations, so they can be laid out over the girders in between the main beam/deck shear studs.
This obviates the welding and ensures that, as well as the deck and beams acting compositely, the straps provide full transverse restraint.
In the UK, square grids of reinforcement are often placed around the shear studs in composite decks to prevent failure due to separating stresses. The Canadian engineers argue this is not necessary. 'There's no way to calculate what that force is,' says Bakht. 'Our tests show that this force is independent of the span length, and it can easily be taken by the straps.'
Early cracking in the concrete is controlled by randomly distributed low modulus fibres in the concrete mix.
Additional measures have to be taken to cope with negative moments generated by cantilever overhangs and bridge barriers.
Jaeger says: 'In the early stages we wanted to avoid cantilever overhangs, but we still have to provide a barrier, and that induces negative moments. We worked hard to get that down and not put negative moments into the deck.'
They developed a new form of bridge barrier to reduce the moment, and also adopted the use of glass fibre reinforced plastic reinforcement in the top of the slab.
Bakht says the entirely steel-free bridge might not be the solution to everyone's problems. 'There is a whole spectrum of possibilities, and the entirely steel-free bridge is one end of the spectrum.'
The three engineers say their system can cut costs in a variety of ways. In addition to reducing the amount of steel from, say, 32kg/m2 to as little as 6kg/m2, the deck thickness is also reduced with resulting savings throughout the structure. They also claim savings on maintenance make 'steel-free' solutions ideal for new build and refurbishment of existing bridges.