Experimental Methods in Earthquake Engineering - Dynamic Behavior of Concrete and Seismic Engineering - page 169

and of 2 test stands that can withstand loads about a few thousand kN, i.e. the loads
necessary to strain and seriously damage real size structures. Although the wall is by
far the longest in Europe, there is one more or less similar in the USA, in San Diego,
but both are surpassed in size by the Japanese Tsukuba 21 m high wall.
Figure 5.2. Reaction wall at the ELSA
The implementation carried out at the ELSA was the first to use fully numerical
control algorithms for the motions of the jacks [MAG 91], [MAG 93]. These allow
very accurate control and implementation of the different algorithms available for
the numerical integration of motion equations.
It is not usually possible to test structures as big as bridges or oilrigs.
Nevertheless, the seismic loading will often only damage some parts of them. For a
lot of bridges, the damage will mostly concentrate at the base of the piers. In these
cases, we can estimate that the remaining structure can be numerically modeled
using the finite element method. It is then easy to combine the PSD test for one part
of the tested structural portion with the time integration of the motion equations of
the model of the remaining structure or numerical structure. For a bridge, the piers
constitute the tested structure, whereas the bridge deck is the numerical structure.
This kind of sub-structuration test forms the essence of the PSD method [DER 85],
but had only been used in practice for small components in the early 1990s. At the
ELSA large-scale tests were carried out for the first time [PIN 95a, PIN 96].