Eastenders' rail link upgraded to meet 1990s workload
TWI Bulletin, May/June 1988
Martin Ogle, BSc (Eng), ACGI, PhD, CEng, MICE, MWeldI, is Principal Design Consultant in the Engineering and Materials Group.
Unforeseen expansion of London's rejuvenated Docklands has forced engineers to take a second look at the design of the inland port's new light railway. Dramatic development of the onetime deserted Canary Wharf in particular has required both frequency and length of commuter trains to be increased to handle tomorrow's potential demand. The designer of the already completed railway bridges contacted The Welding Institute about testing modifications proposed for the structures . . .
A specimen comprised a concrete block cast between two parallel sided 35mm thickness steel plates. The pins can be seen above points A and B
The new Docklands Light Railway, opened to the public last year, contains seven steel I girder bridges. These bridges have a common feature in that the concrete deck acts compositely with the steel top flanges, thus saving in weight of steel. An essential feature of this design is the thousands of stud connectors welded along the length of each top flange. It is these connectors that ensure that the concrete deck and the top flange of the plate girders strain together. If the shear connection was to be lost the plate girders would become overstressed. (See front cover.)
The most onerous loading on most of the stud connectors arises from the live loading. As trains pass over the bridge so the horizontal shear fluctuates in each connector, in many cases there is a reversal of stress. BS 5400 Part 10 gives rules for assessing the fatigue strength of welded connectors and these were used to proportion the number of 125mm high 19mm diameter headed studs.
Once construction was under way a reassessment of the future traffic loading on the DLR network was made. Due to development which had not been foreseen at the time of drawing up the original specifications for the line, it was found that the frequency and length of commuter trains would have to be increased to cope with the potential demand. This had the effect of seriously reducing the predicted fatigue life of many of the shear connectors.
Had this information come to light before construction started it would have been a relatively simple matter to increase the number of shear connectors on the girders. As things were it posed a major problem. Not only were the top flanges covered by the concrete deck and track, but also the line was in constant use. This meant that any work done from above the deck would have to be restricted to limited possession periods. This would have been the preferred solution had access been readily available from the top. Holes would have had to be cut through the concrete above the flanges, additional connectors welded on and then concreted up again.
Early destruction tests revealed that significant forces caused the flange plates to part under load
Another solution had to be found and W S Atkins and Partners, the designers, and Mowlem, the contractors, had to look at possible ways of solving the problem from underneath the deck. Although this would require the erection of staging it would have the advantage of continuous access. The solution considered was to drill holes from underneath the flange into the concrete and to insert steel pins into the holes. The problem was to devise a way of ensuring that the pin was a snug fit in both the steel and the concrete part of the hole.
It was decided to use a roll spring or 'tension' pin, which has been widely used in the mechanical engineering industry for connecting components such as pulley shafts. The tension pin is made from a flat strip of steel which is rolled into an almost complete circle of the required diameter. The pin is then given a heat treatment to increase its strength and more importantly its elastic range. The end is slightly tapered to ease entry into the hole which must be slightly smaller than the diameter of the pin in its free state. The pin is driven or jacked into the hole which makes it a tight fit.
A similar type of fastener used again by the mechanical engineering industry is the spiral spring pin which is made from a thinner but longer piece of flat steel. This is coiled into a spiral like a clock spring. The effect is similar to the tension pin in that it has a diametrical elasticity many times greater than that of a solid pin and hence does not require high precision machining tolerances in the hole for an interference fit to be achieved.
This was the first time that such an application had been proposed for these components, so it was clearly essential to demonstrate that they would do the job required of them. At this stage W S Atkins and Partners, who are long standing Research Members approached The Welding Institute to have the appropriate testing done. The plan was for Mowlem to make the test specimens including the drilling of the holes and the installation of the pins. The specimens would be sent to Abington for testing.
The pins are to be installed in the top flange of the light railway's existing main I girders
It was clearly impracticable to test a scaled down section of simulated girder in bending as the scale of the pins could not be reliably reduced. It was therefore decided to test a short length of concrete containing 4 full size pins, 2 on each side. The block would be sandwiched between two parallel steel plates 35mm thickness representative of the flange thickness on the actual bridges.
The method used by Mowlem to manufacture the test blocks was to weld the side plates to a heavy base plate. A piece of 50mm thickness polystyrene was placed on the base plate between the two side plates. A small cage of reinforcing bars was mounted between the side plates and the two open sides closed with temporary forms. Concrete of the same grade and mix as used in the bridge deck was poured into the resulting mould, vibrated and left to cure.
When the concrete had reached a reasonable strength the four 20mm diameter pin holes were drilled by diamond drilling through the 35mm of steel and 125mm of concrete. The four holes were symmetrical about the two axes of the block in plan to ensure that the load sharing was equal. The tension pins were then driven into the holes making sure that the slots were orientated in a prescribed way as the stiffness varied with orientation. The polystyrene was removed from beneath the block.
The pins used in the tests are rolled from flat steel plate into a spiral and have great diametrical elasticity
Twelve test blocks, each weighing almost half a tonne, were delivered to Abington for testing. Some of the blocks were tested statically using a 1000kN jack. Load deflection plots were obtained so that the stiffnesses of the various shear connector systems could be compared. This was particularly important as the new connectors had to have stiffnesses comparable with those of the welded studs, otherwise they would not take their fair share of load, the new connectors only coming into play after the old ones had failed.
It was found that the tension pin needed stiffening up by driving a smaller tension pin into its hole. This had the disadvantage of making it into a double operation, whereas the spiral pin was only a single operation. It was also necessary to grind down the burr on the outer pin before being able to enter the inner pin.
One important feature noticed in the first static test was the tendency for the side plates to separate. This was no doubt due to bending action on the pins. It was decided to tie the side plates together with welded straps to prevent this. This emphasised the important part played by the head on the welded connectors in holding the deck into firm contact with the girder. The new connectors on their own would be unable to do this.
The greater number of blocks were subjected to fatigue testing using a 250kN hydraulically controlled actuator. During the course of each loading programme the test was stopped and a load deflection plot obtained. This would give a measure of any change in stiffness during the fatigue life.
The results of the tests are currently being studied by W S Atkins and Partners prior to making final recommendations for strengthening the Dockland Bridges. This test programme is another instance of successful co-operation with Research Members in solving difficult technical problems beyond the scope of existing knowledge.
