Completing and sinking the caissons was perhaps the most stressful operation of the whole breakwater project, says Paul Neal. It meant a complex float-out, detailed logistics, and a battle with the silt.
Eleven, 17m high concrete boxes weighing between 2,500t and 4,500t were sunk into place and filled with sand to make the breakwater. Linked together they form two curved arms around the seaward side lock entrances.
The boxes are open on one side with a slatted structure of screed arms. Waves dissipate on entry, reflect off the back wall and interfere with incoming waves, explains designer Gibb's Peter Hunter. 'There is a lot of turbulence which disperses wave energy' he says.
The unusual structure is more effective than a conventional cofferdam and less sprawling than an armoured embankment which for the high tides in Cardiff would have needed to be 70m across at the base.
Unfortunately the structure is also very difficult to float, since it is asymmetric in weight distribution with a tendency to turn over in the water. Flotation was made more awkward because every unit was different, adds Neal. Two of them with curved bullnose ends were particularly troublesome. But that came later.
The breakwater was formed in two halves. The first six boxes were floated out while the sand bund was still in place. The second batch waited until the dredgers had been in, clearing the space needed on the seabed for the remaining structure. The two part programme tallied with fabrication of the boxes in a dock drydock, where there was just enough space to make six units.
'The drydock was virtually a separate operation,' says Neal. The dock's curved granite base was fitted with a stone floor capped with concrete to create a working surface; a steel base plate on each reduced what the team called 'stiction' at float out. Mobile lattice boom cranes on the dockside serviced the construction activity. Pumps delivered concrete from a local ARC ready mixed plant. Formwork was mainly supplied by Peri, though specially made steel forms were needed for the screed beams.
Float out required some complex thinking. Making the units buoyant would have caused instability. BB/C involved nautical engineers in complex calculations of 'metrocentric' points, the pivot around which a floating object rotates. 'If it coincides with the centre of gravity the object will rotate in the water - if it is separate the object will find upright,' explains Neal, adding that the difference on the caissons was just 350mm.
The solution was to add tanks, secured with diagonal tie rods to the toe, but that was difficult to achieve in the narrow drydock. Two sets of re-usable tanks were lifted in and then the dock flooded and the flotation of the units checked for stability. The dock would be pumped out again before the next two units were fitted.
Three dock basins had to be negotiated to leave harbour and channels into the first two of these were narrow. Skillful work with tugs was called for. In the third basin further 'wing' pontoons could be added, which since they sat wider had much greater moment and greatly increased stability.
Taking the units out on to station required juggling a huge number of factors. Wind and wave levels had to be minimal, since temporary splash guards could not guarantee that the open sided box would not take on water. Spring tides were needed for sufficient draught and there could be no prospect of mist. Three tugs had to be available and clearance from the Port Authority which has ships up to 38,000t moving in and out.
And then the foundation had to be ready, which was at times a nightmare. The caissons sit on a rock filled trench foundation with a 200mm thick top layer of 20mm aggregate. 'This has to be dead level, not only for the caisson but between caissons,' says Neal.
The foundation was prepared by divers using a special frame with adjustable screw jacks at the corners. Stone dumped into it was leveled off with a screed bar. Divers had to work blind since the silt in the water made it impenetrable.
Silt was also deposited on to the bed at every tide and since that would cause the caissons to slip, it had to be cleaned off. 'You could get a 1.5m layer of silt dropping on a single tide,' recalls chief engineer Nick Platt. Bringing in a suction dredger where divers were working added complications.
The result was that a single caisson could, and did, sit for weeks in the third basin waiting for the right combination of factors to arise.
Once in place the tugs held the unit until the tide dropped it into place. Divers then unlatched the tanks which were hauled back to the basin. A crane lifted them to drain before being trucked round to the dock for the next pair of units.
The final breakwater received an insitu top deck and furniture.