More and more details of the catastrophic collapse of Terminal 2E at Charles de Gaulle airport in Paris are slowly emerging. But the full contents of the 30-page official report into the disaster are now sub judice, as it has been seized as evidence by the judicial investigation.
Earlier this month chief investigator Jean Berthier set out the basic findings of the team's report, expressed in its conclusions, at a press conference in Paris. For legal reasons his investigation had to be based mainly on visual inspections of the debris and the study of relevant drawings, specifications, minutes and correspondence.
The collapse site remains as it did in May 2004 - still a crime scene under French law. Berthier was not allowed to remove any samples or carry out detailed petrographic examination of fractured surfaces.
These restrictions make it difficult to draw a full final impression of what happened on 23 May 2004, when part of the building collapsed killing four people. In particular the precise collapse sequence is still unclear.
The 650m long 4m high terminal building is made up of a series of 4m wide panels forming a deformed tube resting on parallel longitudinal beams 26.2m apart.
Structurally it acts as a form of extreme portal frame, with the curving composite wall units made up of a 300mm thick precast concrete shell and a single steel tube external tension truss, with simple vertical struts connecting the elements.
A conventional 300mm thick precast concrete roof unit is linked to the two wall units by insitu concrete stitches. The tube is surrounded by a glazed canopy which feeds light into the structure through square voids cast into the shell.
Slim rectangular columns support the edge beams. Any tendency for the superstructure to spread under load is contained by steel 'stirrups' tied around the column heads and bolted to the longitudinal beam.
French sources suggest these were added after construction actually started.
At two points on either side of the shell's central section, three walkways cut into the structure and it was at one of these points that the structure failed.
Berthier points to cracking caused by long term concrete creep exacerbated by differential thermal movement between the two elements making up the curved composite wall panels and by flexing under load.
This weakened the highly stressed concrete near the walkway cutouts. Berthier comments that this cracking could 'lead one to think that the steel reinforcement was insufficient or wrongly located.' He also comments on the large service holes in the 800mm deep edge beam, the recessing of the strut connections and the lack of redundancy or alternative loadpaths in the structural form chosen.
By the time of the collapse, the cracking had progressed until the factor of safety was effectively zero. On the day of the accident, it would not have taken much to trigger the failure because the concrete shell was already significantly weakened.
The rst sign of trouble was when a long strip of concrete fell from the underside of the shell, close to the weak point where three walkways pierced the structure, along the line of the steel struts connecting concrete compression member to steel tension member.
Ninety minutes later two almost simultaneous events occurred. The northern wall unit buckled and the struts punched through. Then the edge beam collapsed.
Very low temperatures the night before may have been the final straw. The investigators specifically ruled out faulty construction or defects in the steel or concrete used. It also ruled out column head failure, pointing to the lack of damage to these after the collapse.
But they say that failure of one of the edge beam restraints could also have triggered the collapse.