The Christchurch Earthquake: Some Half-Assed Analysis

First, if you have some familiarity with what happens to building during earthquakes, you might want to skip my ramblings and go straight to the pictures.

Second, the following is based on Another Kiwi’s kind referral of before and after pictures, and a site I stumbled on. Click on the links to show professional journalists some love.

Third, a disclaimer. The closest I’ve ever been to New Zealand is The Navigator.

Finally, I’m sorry that this is so clinical. Unfortunately, the only way structural design really moves forward is by analyzing failures and that needs to be done before the sites are cleaned and the evidence hidden or destroyed. It’s second nature for me to do this and no disrespect is intended.

Seismic loading is fundamentally different from other ordinary structural loads, as it depends on the design of the structure itself and not simply outside effects. Gravity loading is based on the building’s dead weight and the load of its occupancy, regardless of whether the building’s weight comes from concrete slabs or fill over wood joists. Wind loading depends on the wind speeds expected and the building’s shape, but is not affected by whether the wind’s sideways pressure is resisted by a frame or by masonry walls. Seismic loading, on the other had, is determined both by the local ground accelerations that are expected (measured as a percentage of gravity acceleration) and by various measures of how stiff the building is. Two buildings that look similar but have different structural types (one a steel braced frame, the other a concrete moment frame, for example) will have different design seismic load because they will react differently to ground acceleration. More flexible frames generally perform better because their flexure absorbs energy while stiff buildings are damaged by their own rebound.

Beyond the analysis of how buildings resist sideways acceleration, the greatest issue is quake design is continuity. Brittle materials (unreinforced concrete and masonry) develop cracks that break single elements (walls or slabs) into smaller, weaker elements. Poor connections between elements (wood joists pocketed in, but not tied, masonry walls, for example) lead to the elements moving differentially because of their different inherent frequencies.

On to the evidence:

The Anglican cathedral, first page of AK’s link. The stone spire has collapsed where it changes from a relatively rigid base to the more flexible portion where the large windows begin. Note that the small cross on the sanctuary gable end has also collapsed. In modern construction, the masonry would be a skin, not the structure, so its stiffness would be less important. There’s little that can be done to retrofit this type of structure short of base isolation (putting shock absorbers below the entire building) or inserting a steel frame within the tower.

The CTV Building, third page of AK’s link.

As far as I can tell, a reinforced-concrete frame that failed in the worst way possible: the columns gave out. That type of failure leads to pancaking (officially known s “progressive collapse”) and is a symptom of lateral loads that were simply too large for the frame capacity. This could be because of reduction in capacity from last fall’s earthquake, an older building designed under less stringent codes, codes that underestimated the local ground acceleration, or misdesign. The last is possible, but least likely.

Knox Presbyterian Church, page 4 of AK’s link: Classic failure of the connection between the masonry gable-end walls and the wood truss roof. The roof appears to be in good condition, but there was (a) not enough of a tie between the walls and the truss members and (b) no ductility in the unreinforced masonry.

The Catholic cathedral, fifth at AK’s link:

The fact that the damage appears to be confined to one side of the building is odd and may indicate differing soil conditions across the site. In any case, the relatively weak but heavy towers survived long enough before collapsing to transmit movement to and damage the walls below. Again, classic damage to unreinforced masonry.

The Pyne Gould building, ninth at AK’s link.

Note that the columns and beams at the edge of the collapse are straight. The failure was in the beam-to-column connections, which were not sufficiently ductile for the load. Otherwise similar to the CTV building.

21 thoughts on “The Christchurch Earthquake: Some Half-Assed Analysis

  1. FWIW, the Pyne Gould Guinness building was 1960s design and construction.

    The CTV building was more recent but I don’t know when. Among other offices, it housed a counselling service for people suffering distress from last year’s earthquake. I’m going to drink beer now.

    • There have been big advances in seismic design of concrete since the 60s, so I would expect buildings from that era to underperform.

      One of the weightings used to determine seismic load is the “importance factor.” Critical facilities (hospitals, schools, fire houses) get a factor greater than 1, barns get one less than 1, most buildings are 1. Unfortunately, counseling services are not taken into account…meaning that maybe they should rent space at schools…

  2. Finally, I’m sorry that this is so clinical. Unfortunately, the only way structural design really moves forward is by analyzing failures and that needs to be done before the sites are cleaned and the evidence hidden or destroyed. It’s second nature for me to do this and no disrespect is intended.

    No apology needed, old chum. We know that you’re a kindly, humane man, and that a dispassionate look at the evidence is crucial to preventing future tragedies.

    • I fully agree with studying the evidence. However, I have seen the unreinforced masonry and nonductile concrete buildings fail miserably in earthquakes (e.g. Mexico City and Adapazari, Turkey). Evidence from those earthquakes have been thoroughly studied and the findings published. However, the problem of existing hazardous buildings have not been fully addressed where these disasters have occurred (people have short memories). In other places, the people tend to think that such disasters only occur elsewhere. It saddens me to see all the casualties knowing that we know which buildings will fail and kill people and why, but we don’t do anything about it.

      • Non-engineers tend to focus (rightly) on the human issues, leaving engineers to sort through the forensic analysis. It is unfortunately rare that the engineering conclusions – which may take a few years to solidify – almost never make it into the popular press.

  3. CTV House also contained a language school.
    It was designed as generic office space — defs not an importance factor — and came to house anyone who could afford the rent.

  4. Thanks for posting this. As an interested layman, I find it fascinating. Your explanations make the whole thing much clearer.

    The Seattle Times had an article up the other day pointing out that the Puget Sound region’s soil (glacial till, for the most part, where it isn’t sawdust around Elliott Bay in Seattle and Commencement Bay in Tacoma) is similar to that around Christchurch and with our active geology (thanks, Ring of Fire!) we’re at a high level of risk for a similar event someday. Whee, can’t wait until the north and south halves of Seattle are split by collapsed bridges.

    Anyway, this is really fascinating.

  5. The failure was in the beam-to-column connections, which were not sufficiently ductile for the load.

    Does not sufficiently ductile mean not enough rebar was used with the concrete?

    • That’s one possibility. The bar being detailed badly is another.

      For example: shear ties in beams used to be “stirrups”: U-shaped bars extending across the bottom and up the sides but open at the top. They are no closed loops, which confines the concrete better and therefore provides more ductile behavior.

  6. The Catholic cathedral, fifth at AK’s link:
    The fact that the damage appears to be confined to one side of the building is odd

    The appearance of asymmetry is exaggerated by the building itself being asymmetrical. Towards the far end — the working end of the cathedral with the altar and all — it rises to a large hemispherical dome, resting on a cylinder, and that part will be double-shell construction, and well-integrated into the rest of the building.

    At the other end, the ‘front’ of the building is flanked by two smaller cupolas (copper-sheathed wood) sitting on square towers. It’s those two cupolas that are now scattered across the landscape.

    One of the towers was at the corner of the building closest to the photographer, and now looks like someone stepped on it; the other tower (at the corner on the left) has lost its cupola but otherwise has most of its bricks in place, but I suspect it could easily have gone the other way.

  7. I really appreciate this NB. I went into my town’s centre today and looked for stuff such as you point out. I need 10 or 15 minutes warning to get out of there if something’s coming. It all looks bad.

  8. And we have nice canyons of olde tyme buildings.
    I got my first look at the spiffy new office block “On the Square” today.
    Is it supposed to be a joke about East Germany? Or contractors securing future work for themselves

  9. You did a fine job of explaining for a non-specialist audience, both in the clarity of the text and in the illustrating / diagnosing photos you selected.

    I am a not an engineer, but I once encounted a similar problem of explaining the basic principles of seismic-resistant buildings to local public officials and school boards…a FEMA video to persuade them to build seismic-resistant schools in the future. Stills from Long Beach. Animated graphics of buildings under stress, types of construction & collapse. The hardest part was to convince people that it wasn’t just a ‘California’ problem. (Had a map graphic….added pulsing dots for New Madrid, Charleston. Tried to make the economic case–only added @ 1.5 % to cost of building. Miniscule cost, especially because schools serve as community shelters in disasters. FEMA handed them out like aspirin. I hope they did some good.

    Sorry for the windy and self-referential comment, but I enjoyed this article from a Subject Matter Expert (training development jargon).

    • The short answer is that engineers design buildings to whatever acceleration the codes specify. Since the design loads are based, in part, on the response characteristics of the building, a 2g vertical ground acceleration does not translate to a 300%g vertical design load.
      It’s also possible that any given quake may exceed the code design values. I don’t know the NZ codes.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s