SkyCity fire: What’s the damage?

As the SkyCity Convention Centre blaze is finally contained, questions will turn to the impact the inferno has had on the $700m complex and the time-consuming and costly rehabilitation job ahead.

The biggest expense will likely be cleaning up smoke and water damage to the roof steelwork and to the lower levels of the structure. This is typically the case following a blaze in such a structure. For example in a much more severe fire in a multi-storey steel-framed building in the UK, 93 percent of the rehabilitation costs went on cleaning smoke and fire damage to the claddings and linings and other non-structural elements like walls and ceilings. Only 7 percent of the cost was in replacing structural members.

I expect in this case the cost of replacing structural steel members due to fire damage could be close to nothing and the condition of the structure to be as follows:

• The main columns and trusses will have no structural fire damage or distortion. Some of the intumescent paint covering them - designed to expand in severe fire and provide thermal insulation to the steelwork - may have started to expand, but I expect not very much of this; the temperatures at that level won’t have been very high in fire terms. This must be cleaned of fire debris, repainted where necessary and then put back into service.
• Many to most of the bracing members and rafters will be similarly undamaged structurally and can be treated the same as the main members. It will be easier to replace any members with visible distortion than to straighten them.
• The video footage shows the lightweight purlins, which are exposed once the bitumen, ply and straw has burned away, look to have minimal damage. However, these have been part of the sandwich roof panel construction and I expect it will be easier to replace any that are still straight with new purlins in the new sandwich panels than to try and rebuild them into the new roof. Given many of these purlins look to be in place and largely undamaged, it is hoped that the fire-resisting bottom layer that is attached to them has also remained mostly in place. This would shield the space below, including structural members from the fire.

I agree with the decision by Fire and Emergency New Zealand to let the roof burn itself out on the level that was on fire - unfortunately the main roof level. This may have reduced the amount of water they needed to douse the fire and therefore the water damage through to the lower floors, reducing the cost and difficulty of cleaning up.

More importantly, it saved fire-fighters the very difficult and dangerous task of getting onto the roof itself to try and stop it spreading, which would have been near impossible. Trying to do this from underneath would have been completely impossible.

The roof system comprises panels with compressed straw insulation and acoustic insulation, embedded between cold formed steel purlins (like you see on most large single-storey buildings). In this case the purlins are within the panels and surrounded by the straw insulation and acoustic insulation. They have a fire-resisting bottom layer to suppress fire getting into the roof from underneath. Over the top is a plywood layer, all this forming a sandwich panel that's put into place and then covered with the bitumen membrane outer layer. The sandwich panel construction is unusual and has been chosen for its thermal and acoustic insulation. The bitumen membrane is a common roof membrane on near flat roofs of multi-storey buildings.

The panels are supported on structural steel rafters which are in turn supported on the main structural steel trusses. There are also bracing members in the roof to make it rigid for earthquake and wind resistance. The trusses are supported off steel columns back down to the lower levels and the ground.

I don’t know why the decision was made to use this combination of materials in the sandwich panels on the roof. The straw may have been chosen over other forms of insulation for environmental reasons and also because the material is less toxic in fire than PUR or PIR (rigid polyurethane foam), which is the traditionally used insulation material. Once the fire got into this material it was not possible to extinguish it. The fire burned for so long because the roof is so big; it took 20 hours to spread across the roof until it burned everything combustible in that roof plane.

As to what caused the fire in the first place, I don’t have enough details to comment and probably never will. That will have to be determined from the on-site inquiry. During the first half of this year I watched roofers laying a similar bitumen membrane over an atrium roof below my office at the University of Auckland.

It is a time-consuming process which involves using a blow torch to heat the joins between the bitumen strips to bond them together. The blow torches are very noisy when in operation and it is very unlikely that someone left a torch on at full power while they went on a break. Furthermore, they typically work in pairs laying this roof, further reducing the chance of someone leaving a torch burning unattended.

When I was watching the roofers at the university, I noticed when they went away for short periods to get new materials they would sometimes leave the torches on a box with the pilot flame on and pointed away from anything combustible, but this flame is low intensity. I expect we will find out what happened, but it is unlikely, from my observational experience, that the fire would have started when new strips of bitumen layer were being bonded to already-laid strips.