Reconstructing the collapse

A six-year study by investigators at the National Institute of Standards and Technology about the destruction of World Trade Center buildings tested the limits of modeling programs used to simulate the collapse of the structures.

NIST used a suite of four commercial and in-house tools to determine why and how the iconic WTC towers collapsed in the 2001 terrorist attacks and why WTC Building 7, damaged in the initial attacks, collapsed hours later. Models created with each of the programs were more complex than earlier studies, and this complexity was multiplied by the task of transferring the output data of one program as input data for another, said lead investigator Shyam Sunder.

'There wasn't an elegant interface,' Sunder said. 'There was a lot of hard coding of algorithms to make the data transfers possible.'

NIST released findings last month on the destruction of Building 7, concluding that it was fire and not explosives or impact from debris that caused the 47-story building to fail.

'Our study found that the fires in WTC 7, which were uncontrolled but otherwise similar to fires experienced in other tall buildings, caused an extraordinary event,' Sunder said. 'Heating of floor beams and girders caused a critical support column to fail, initiating a fire-induced progressive collapse that brought the building down.'

NIST began investigating the attacks in August 2002 and released its findings on the collapse of towers 1 and 2 three years later. Investigators spent the last three years investigating Building 7.

'The software per se hasn't changed' during the span of the investigations, Sunder said, 'but what we ended up doing in the analysis changed.'

The process of digitally reproducing the collapses began with a reference model of the buildings created with a structural analysis program. The general-purpose finite element analysis tool was used to create a refined model of the upper third of towers 1 and 2 using LS-Dyna, an advanced multiphysics simulation software package developed by Livermore Software Technology. Within the path of the two airplanes that hit the buildings, the modeling produced a fine mesh on a scale of inches. The rest of the detailed portion was on a scale of feet.

The speed, orientation and weight of the planes also were determined, and the final LS-Dyna models required input of about 2.5 million elements.

After the impact had been modeled, investigators used NIST-developed Fire Dynamics Simulator (FDS) software developed to show the growth and spread of the flames. This modeling included the workspace layouts of each floor of the buildings, with the location and type of combustible materials, including stacks of paper. The simulations were backed with physical experiments on the NIST campus.
With heat release data from FDS software, investigators used programs developed by Ansys to predict the effect of that heat on structural elements up to the point of failure where collapse began.
The process was similar for Building 7, Sunder said. 'The biggest differences were that we did not have an airplane impact,' but the team went into more detail on the building's collapse. Although modeling on the two towers stopped with the initiation of collapse, the process for Building 7 continued with the propagation of the collapse vertically, horizontally from east to west, and the entire collapse of the fa'ade.
'This meant that the computational complexity was magnified by an order of magnitude' for Building 7, Sunder said. After it was programmed, a complete run of the entire simulation for the two towers took two months on high-performance computing clusters. Running the entire simulation for Building 7 took eight months.
The investigation's results are described as the probable collapse sequence, for each building, Sunder said. 'We have a very good confidence in the physics and science behind the results.'

About the Author

William Jackson is a Maryland-based freelance writer.


  • Records management: Look beyond the NARA mandates

    Pandemic tests electronic records management

    Between the rush enable more virtual collaboration, stalled digitization of archived records and managing records that reside in datasets, records management executives are sorting through new challenges.

  • boy learning at home (Travelpixs/

    Tucson’s community wireless bridges the digital divide

    The city built cell sites at government-owned facilities such as fire departments and libraries that were already connected to Tucson’s existing fiber backbone.

Stay Connected