An engineering lesson from 9/11

The two events that most severely impact on a multi-storey building structure are severe earthquake and severe fire. However, our understanding of their behaviour and design for the impact of these events could not be more different.

In the case of earthquake, the need to understand whole building behaviour and design has driven national and international research efforts since the 1970s. As a result, the severe earthquakes of 2010 and 2011 in Canterbury, and 2013 and 2016 in the north east of the South Island, caused no building collapses in well-designed and detailed buildings built since 1975.

In the case of severe fire, design of the structure is typically based on the performance of isolated structural elements (beams, columns and floors) under standardised fire exposure. The assumption is that the structural system response will be at least as good as the performance of these isolated structural elements.

But recent case histories of severe fires in multistorey buildings since 2000 has shown this assumption to be fundamentally wrong, with more than 35 percent of these events resulting in partial or total collapse of the buildings.

On September 11 we commemorate the 20th anniversary of the most devastating of these events, the attacks on the World Trade Center complex which caused the complete collapse of three major buildings, known in typical economical engineering terms as WTC1, WTC2 and WTC7.

Detailed studies of these collapses were undertaken and, in the case of the Twin Towers, WTC1 and WTC2, there is general agreement among fire and structural engineers on the mechanisms causing collapse and the sequences involved.

But WTC7 is a different story and much more relevant to New Zealand buildings. This was a 43-storey steel framed building with composite floor systems, comprising reinforced concrete floors on steel decking on steel beams, the same form of construction that is widely used in New Zealand.

WTC7 was situated adjacent to the World Trade Centre complex, just north of the Twin Towers. Burning debris falling from the Twin Towers ignited fires in WTC7. After more than eight hours of uncontrolled fires burning through the lower 13 storeys of the building, it suffered a complete, straight down collapse, as captured on various local webcams.

Official studies have pinpointed where the collapse started, on level 13 when a major internal floor support beam near one end of the building fell off its support to a very large internal column, triggering a cascading internal collapse. However, all these studies concluded that the collapse would have been confined to that end of the building, leaving the rest standing, contrary to what was observed.

This is where fire engineering developments from New Zealand can help explain the observed collapse sequence.

In 1991, a severe fire in a multi-storey steel framed building in London, UK, caused partial failures of internal columns and significant deformation of internal floors, but no collapse of the overall structure, which was repaired. This unexpectedly robust behaviour led to a series of landmark large building fire tests in the UK, which showed that composite floor systems have very much higher resistance to collapse in a severe fire than that of any of the individual parts of the floor system. From this a design model, called tensile membrane action (TMA) was developed in 2000 by the UK researchers to explain the behaviour.

A number of years earlier, in 1983, I had started the Structural Division of the New Zealand Heavy Engineering Research Association (HERA) where I was responsible for developing and disseminating technical design guidance for steel framed buildings with composite floors. The research into the enhanced fire resistance of composite floors from TMA was of great interest and I have led research into this behaviour and its implementation for New Zealand practice since 2000.

A design procedure called the Slab Panel Method (SPM) is the result of this research. It enables design of composite floor systems for dependable inelastic response to severe fires, typically resulting in a useful cost saving for the overall building compared with the traditional isolated element design approach. We have been continuously improving the SPM design procedure since it was first published in 2006, with the current procedure now the best desktop based design procedure for these systems worldwide and widely used in New Zealand.

So, how does this relate to the WTC7 collapse? One of the unanswered questions from the WTC7 collapse was why a plant room/penthouse, located near the centre of the building at roof level and visible in footage from several of the adjacent webcams, disappeared from view some 15 seconds before the final collapse, without anything further happening until a uniform collapse of the whole building occurred, with the roofline staying level until it disappeared from the webcams’ view.

I had always wondered if TMA might explain the way the top of the building remained stable while the building underwent an internal collapse from level 13 downwards, until so little of the inside was left to support the stable top of the building that it fell straight down into the void. In 2017 I supervised a pair of Part 4 Project CEE students to apply the SPM procedure to the entire floor area of a typical top floor, to determine if this could have provided support to the top floors and kept the top of the building stable until support was lost from the collapsing building.

The results showed this was possible, with more than three metres of sag generated near the centre of the floors under tensile membrane action. That was sufficient to explain the plant room/penthouse on the top disappearing from view, then no further movement occurring until the whole building collapsed, and matches the observed collapse sequence much better than the official studies do.

It also offers an alternative to the conspiracy theories that have flourished to fill the void between the observed collapse and the predictions of the official studies.