This is something that the profession has done a terrible job in communicating with the public.
But if you think about the nature of earthquakes, it starts to make a little more sense. There is basically a probabilistic distribution of earthquake frequency and energy. The more powerful the quake, the less frequently it occurs, and there is a long tail to this distribution. So we need to draw a line in the sand somewhere and say “we will design for X, and detail the structure to avoid collapse if X is exceeded.” There is no such thing as an earthquake proof building, because there is always a bigger earthquake than the one considered for design. All we can do is play with probabilities.
In the US the life safety objective corresponds to earthquake hazard of 10% probability of exceedance in 50yrs. The industry has long accepted this as a good X point that balances construction cost and longevity. But more recently, with ballooning recovery costs in recent earthquakes, some are starting to see its not enough. For instance, in the Canterbury NZ earthquake many structures performed “well” and met their design intents, but around 70% of the downtown ended up being demolished. Insurance deemed them too costly and risky to repair.
Now industry is starting to use “performance based design” more often for seismic design of tall buildings (including the Millennium tower). It allows you to pick multiple hazards and set different objectives for them. For instance, we might say no damage allowed at 30% in 50yr hazard, cracking and light damage at 10%/50yr, and no collapse at 2%/50yr hazards. Then we do a bunch more sophisticated modelling to demonstrate the design meets the objectives. I wrote a little bit more about performance based design here :
Unfortunately it didn’t help the millennium tower with this geotechnical issue. Which I suppose says something about design being only as good as the input assumptions.
It’s also worth noting that geotechnical engineering is a high risk and low tech field. Most of their design relations involve pounding a stick into the ground and counting how many hits it took to sink a unit depth (that’s a little bit jest, but also not far from the truth)
Imagine being tasked with identifying a famous painting, But you aren’t allowed to see the painting. You are provided with the five predominant colours in the painting, the type of paint used, and a 10mm x 10mm sample of the painting.
But if you think about the nature of earthquakes, it starts to make a little more sense. There is basically a probabilistic distribution of earthquake frequency and energy. The more powerful the quake, the less frequently it occurs, and there is a long tail to this distribution. So we need to draw a line in the sand somewhere and say “we will design for X, and detail the structure to avoid collapse if X is exceeded.” There is no such thing as an earthquake proof building, because there is always a bigger earthquake than the one considered for design. All we can do is play with probabilities.
In the US the life safety objective corresponds to earthquake hazard of 10% probability of exceedance in 50yrs. The industry has long accepted this as a good X point that balances construction cost and longevity. But more recently, with ballooning recovery costs in recent earthquakes, some are starting to see its not enough. For instance, in the Canterbury NZ earthquake many structures performed “well” and met their design intents, but around 70% of the downtown ended up being demolished. Insurance deemed them too costly and risky to repair.
Now industry is starting to use “performance based design” more often for seismic design of tall buildings (including the Millennium tower). It allows you to pick multiple hazards and set different objectives for them. For instance, we might say no damage allowed at 30% in 50yr hazard, cracking and light damage at 10%/50yr, and no collapse at 2%/50yr hazards. Then we do a bunch more sophisticated modelling to demonstrate the design meets the objectives. I wrote a little bit more about performance based design here :
https://kinson.io/post/what-is-performance-based-design/
Unfortunately it didn’t help the millennium tower with this geotechnical issue. Which I suppose says something about design being only as good as the input assumptions.