Wednesday, 16 November 2016

Earthquake engineering is harder than you think


DESIGNING BUILDINGS THAT respond safely in an earthquake is much harder than you think.  I hear folk who think it’s simple; who wonder why even in a “moderate” quake Wellington’s buildings seem to be falling apart. The thing is that designing to respond properly to every earthquake means designing to respond to every possible earthquake – and there’s no way of knowing what earthquake you are going to get when even The Moderate One comes.

Just remember that the earthquake that proved most disastrous in Christchurch was the one in which the ground did something it wasn’t supposed to do at all, which was move violently up an down. Buildings aren’t designed for that: until then we’d always assumed they are thrown side to side, and designed for that.

So we learned something from that, just as seismic engineers learn something from every earthquake. This is what the best of them keep a suitcase ready beside the bed when they hear a Big One hits somewhere in the world: they want to see what happened, and how, and which design responses worked best this time.

Seismic design is hard. You don’t want a building that’s too strong (because that strength attracts more loading), instead you want it to be ductile. You don’t want a building too closely tied to ground movement; because then it moves with every little jolt. And there is no such thing as a perfectly safe building. Even a perfectly ductile building with total base isolation is going to have problems if a fault line opens up underneath and starts moving – like it did under this house (that’s the driveway up there on the right on the section of land in Kekerengu that shifted over ten metres during the quake, dragging parts of the house with it):


Earthquake engineering is harder than you think. Basically what you’re going is designing to avoid total collapse, to get people out safely, and that’s about all. In other words, to fail safely. So in that respect, the house above succeeded admirably in difficult circumstances.

But it does mean that, once the shaking has stopped, every building that’s been shaken needs to be checked to make sure any failure that has occurred has only been in secondary structures (such as ceilings) and not in the primary structure itself. So do bear in mind this is perfectly normal, and not a failure of designers. Every building remaining standing after a quake is a success, regardless of how damaged it appears.

CONSIDER HOWEVER THE Statistics Building on reclaimed land in Wellington’s downtown, in which a whole floor appears to have dislodged itself and dropped onto the one below:


The Government Statistician Liz MacPherson, asked the obvious question:

How is it that a building that is as new as Stats House, with the [earthquake] code rating it had, could suffer this sort of damage.

The simple answer is that even with all our knowledge, the code is still not any guarantee of success. It’s not a matter of negligence; it’s one about the nature of knowledge. Basically, there is no way to be omniscient about how the ground will move – particularly reclaimed ground, as here, which is always prone to localised softening – so that any code will only reflect what that last generation knew (or thought they knew) about earthquake design. And a lot of earthquake design is about designing that safe mechanism of failure.

So what may have happened here, where the failure is clearly unsafe? First, be aware that all multi-storey buildings since the early eighties have been designed with ‘weak-beams/strong columns.’ This means that in any shake, it’s your beams that fail first – and since these beams are generally tied at each end, your floor will generally sag rather than collapse. In other words, to fail safely. The alternative (designing for strong beams/weak columns) promotes the potentially disastrous alternative of having columns fail first, with the disastrous failure mechanism thereafter of one floor dropping onto another, then another, until a twelve-story building quickly becomes a ten-metre high pile of rubble, as happened (from memory) with Christchurch's Pyne Gould building. (And when you contemplate that designing with strong beams/weak columns was considered “best practice” only a few decades ago, you can see how little even the last generation truly knew about all this.)

So to hazard a guess as to what happened here, it may be that the expected beam weakening was accompanied by some localised ground softening which also weakened the column to which the floor was tied, causing one end of the floor to drop (which is how it seems from photographs). Or it may be as simple as a tie at one end of a beam that failed – with the catastrophic failure you see above.

The only thing about which to be thankful is that it wasn’t a top floor that failed, causing a pancake failure onto floors below – and that the calamity happened at midnight instead of midday, when this would have been an utter tragedy. But seismic engineering is still not an exact science, and unlikely to ever be so.

ONE LAST QUESTION a lot of folk have is the very reasonable one of wondering why there are so many high-rise buildings in an obvious seismic trouble spot that have dangerous panes of glass ready to be dislodged. That’s a very fair question, but not one easily remedied by any code. It is instead one that every designer should be asking themselves.



  1. Cheers for the post, very interesting and informative

  2. Richard McGrath16 Nov 2016, 18:19:00

    Good opportunity to abolish the Statistics Dept

  3. An interesting post but having been in the BNZ building after the Seddon EQ's a few years ago I was impressed by the supposed technology but also surprised that it did not shrug off these earlier events which were well within the design spec as I understand it to have been. I wonder if we are pushing fashion too hard and its at the expense of resilience. Big and open spaces look nice but maybe they present complications that make them less suitable for a place where EQ's are the norm. Being closed for months every time you get something around a 6 or 7 magnitude must be annoying for owners, tenants and insurers. Cosmetic repairs are reasonably expected after EQ's but structural stuff is a different kettle of fish. New stuff should stand out but in some cases it hasn't so the issue is why not.


    1. Pre-Christchurch the accepted design standard was simply to fail in a way that allowed people to get out, without there being a catastrophic collapse. However I think most rebuilds in Christchurch now set a higher criteria, which is not just about preventing death, but also allowing any damage to be repaired and for the building to be reoccupied quickly. I suspect this is something insurance companies generally now insist on for any new builds.

  4. Spot on Sheriff.

    Do you have a better pic of the Stats building?
    I can't see any evidence of dislodging, other than a few paving stones in the car park.



1. Commenters are welcome and invited.
2. All comments are moderated. Off-topic grandstanding, spam, and gibberish will be ignored. Tu quoque will be moderated.
3. Read the post before you comment. Challenge facts, but don't simply ignore them.
4. Use a name. If it's important enough to say, it's important enough to put a name to.
5. Above all: Act with honour. Say what you mean, and mean what you say.