|
...sides. GTM erected the ring sections
on mobile falsework and structurally tied them together.
 |
| COMPLEX
Most serious flaws were a result of inadequate three-dimensional
modeling of complex structural scheme. (Photo courtesy
of Eiffel/Laubeu) |
To stiffen the shallow vault, curved
steel girders embrace its two sides. The tensile girders are
held away from the vault by regularly spaced, 10-cm-dia steel
struts on 20 cm-dia-plates embedded in roughly 10-cm-deep
shell recesses.
At the 20-m-long collapse area,
on one side of the link building, lower sections of three
alternate shells were omitted for footbridge access to the
concourse. Steel connections transferred loads from the shortened
shells to the full ones on either side.
Two failure modes have emerged,
said investigator Denis Aubry, a civil engineering professor
at lEcole Central de Paris. At the footbridge opening,
on the north side, several external struts punctured the shell.
Retrospective calculations showed some struts would have been
overstressed. Immediately opposite, on the south side, the
shell edge beam fractured, falling off its bearing to the
ground. Either mechanism could have been the prime trigger,
said Aubry.
Concrete creep and fatigue caused
by cyclical loading accelerated the failure, Aubry said. The
north and south sides of the concourse experience marked temperature
variations in changing sunlight.
No one realized exactly how complex
the building was, so all of the procedures, even administrative
ones, were not suited, said Calgaro. "GTM was only in charge
of the shellnot the piers, not the bearings and, in
fact, this was one major problem," he adds.
"Nobody made a full...three-dimensional
model, including the shell, the piers, the bearings and the
foundations. That is the reason [to] say the process was not
good," said Calgaro. Click
here to view chart
Paul Andreu, who retired in 2003
as AdPs chief architect, led the design. In the normal
French way, detail design was done by the contractors. That
work was checked both by AdP and Bureau Veritas, a Paris-based
international certification agency.
The investigative team confined
its work to analyzing the failure. A separate probe by the
justice department, due to report within two or three months,
will apportion individual blame, if any.
Established last May, Berthiers
investigative team also includes Anne Froment-Meurice, a building
contracts specialist from the Court of Auditors.
To maintain independence, the investigators
drew necessary technical support only from the governments
central civil engineering laboratory and bridge design organization
SETRA. "The most brilliant civil engineers in France are involved
as experts for one company or another," explained Calgaro.
At the governments request,
the investigators also traced the projects procurement
process to test the adequacy of official procedures. "In all
the processes there were weaknesses," said Calgaro.
As a building, the concourse went
through less rigorous structural analysis than would similarly
complicated pieces of civil engineering, said Calgaro.
"In France, we have a distinction
between civil engineering works...and buildings," he said.
"But there is no distinction between complex...and common
buildings." While complicated bridges would undergo peer review,
no such procedure was applied here, he added.
Though procured as a building,
the concourse was in fact "very, very deformable [and] complex,"
said Calgaro. "The model to study this structure should have
been more complex perhaps than the models used." "The responsibility
of Paul Andreu is perhaps to think that this structure was
not so difficult to control, in particular the control of
deformations, or flex-ibility," Calgaro explained. He also
hits at increasingly tight budgets. Tough price bidding is
"false economy," he said. Though it is necessary to give architects
freedom, it has to be recognized that if structural behavior
is complex, there must be enough money and "very, very clever"
designers.
|