Building codes for designing structures in earthquake zones have until now focused on the superstructure, with little regard for how the foundations perform. This is because codes are geared up to prevent loss of life and very few deaths in earthquakes have been attributed to foundation failures.
Consequently it is quite possible to design and build structures that can safely withstand any feasible earthquake.
However the bigger issue in countries with mature earthquake engineering, such as Japan and US, has recently switched on to how to avoid huge repair costs.
In the California Northridge earthquake of 1994, loss of life was minimal, and in this respect the US seismic codes could be considered to have fulfilled their purpose. Yet the cost of damage was huge, in the order of $200bn, with much of the cost coming from work to repair or replace damaged foundations. From this has come the recognition by building owners that the existing codes are not acceptable. And for insurance companies the high sums are not acceptable.
Suddenly building codes are changing rapidly, and European code writers are leading the way. The draft of Euorcode 8 Design provision for earthquake resistance of structures is the first to give more explicit reference to designing more earthquake resistant foundations.
'Seismic engineers have in the past not thought too deeply about foundations. Most seismic codes are based on US practice, in which foundations were ignored or were a non-issue,' says Ted Piepenbrock of consultant Ove Arup.
Conventionally, when it comes to foundation design, seismic engineers have looked only at inertia forces, these being the forces the superstructure, in moving, exerts on the foundation.
Eurocode 8, codifies for the first time, the concept of kinematic forces in the piles resulting from the soil pile interaction during dynamic ground movements.
'What we are seeing is a move towards performance based approach to seismic engineering,' says Piepenbrock.
Previously seismic codes assumed that as far as piles were concerned, the maximum shear on the piles resulted from inertia forces and occurred at the underside of the pile cap. But from recent efforts to monitor yield and cracking in piles, it is clear the problem has been under thought.
Now developments in computing power mean that what was, until recently, painstaking university research is now within the capabilities of the best equipped consultants.
According to Piepenbrock 'guessing the answer is no longer acceptable'.
Consultants using methods such as non-linear dynamic soil-structure analyses can now get a reasonable idea of how the foundations behave during an earthquake. Such approaches have shown that greatest pile damage is unlikely to occur at the top of the pile, but at the interface between ground horizons of different stiffness. This observation is with hindsight hardly startling.
On a practical level the implications for pile design are surprisingly straightforward. A pile may include increased longitudinal steel or contain steel hooping to confine the concrete in the area at levels in the pile where the greatest bending moments are predicted.
Alternatively, steel piles are more ductile than concrete and so provide another option.