Tuesday, September 17, 2013

Techy Tuesday 2 - Making Bridges Earthquake Resistant

Like much of America, I heard of the Loma Prieta earthquake on October 17, 1989, when the news bulletins hit that the World Series had been interrupted by a strong quake.  An image that will never quite leave my mind is the collapsed dual-deck highway, I-880, the Cypress Viaduct. Thinking of what the people trapped between those spans went through as they died has always made my blood run cold.

But I'm an engineer and while friends say we "think different", we are always about fixing things and making them better.  My homies in the world of Civil Engineering haven't been idle about quake damage and have been researching how to build bridges that can survive really big quakes.  We read in a Design News update that current research is into nitinol, the nickle-titanium alloy used in lightweight, flexible eyeglass frames.  DN links to original work by Cal Tech undergrad Misha Raffiee published at the National Science Foundation
A large majority of bridges are made of steel and concrete. While this combination is convenient and economical, steel-concrete bridges don't hold up as well in strong earthquakes (7.0 magnitude or higher). Conventional reinforced columns rely on the steel and concrete to dissipate energy during strong earthquakes, potentially creating permanent deformation and damage in the column and making the column unusable.
The idea is to replace the steel reinforcing bar ("rebar" in common terms) with superelastic Nitinol.  Nitinol is a member of a family of "shape memory alloys"
While the majority of SMAs are only temperature-sensitive, meaning that they require a heat source to return to their original shape, Nitinol is also superelastic. This means that it can absorb the stress imposed by an earthquake and return to its original shape, which makes nitinol a particularly advantageous alternative to steel. In fact, the superelasticity of nickel titanium is between 10 to 30 times the elasticity of normal metals like steel. 
The new designs were modeled with earthquake simulation software developed at the Peoples' Republic of Berkley, and prototype scale models of bridge structures built.  There's a video on another page at the NSF of them testing a prototype to magnitude 8 earthquake levels.  The models and experiment are in agreement, a sure sign they're progressing well.  

They estimate the safer bridge will cost 3% more, but I find that hard to believe.  Civil projects always seem to come in over budget (as well as late) but maybe they mean the materials cost 3% more.  Materials are probably a small portion of the expense in building it.  

Progress is a good thing.  Not getting people crushed in another bridge like the I-880 is a great thing. 


  1. current price of Ni is ~ $6/lb, Ti ~ $7/lb. Steel ~$0.3/lb. NiTi is ~ 50-50, call it $10/lb.
    World production of Ni is ~1.4 MT, steel is 800 MT.
    What's the mass of steel / mile of overpass?
    I really doubt they'd only see a +3% rise in the cost.

    1. Yes, but...

      We don't know anything about the quantities of each metal needed. Thinks can be made stiffer by changing the cross section shape. There's just not enough info about their design to know.

      Like I said, though, I find it hard to believe cost will be 3% higher, but I don't think it'll be 30 times higher - your $10/lb vs .30/lb. The consensus on that Design News page was "it will be much more".

  2. Dang it, this is one of those times I wish there was a Email follow up button at the bottom. I will check back for your reply.

    I can see the Nitinol being flexible but to my layman's understanding, the concrete is still the load bearing material here and it isn't flexible by nature.

    The steel rebar just reinforces the concrete and by trying to make the support column flex with the wave motion, the concrete will still buckle and crack, no mater what is used inside for reinforcement. Or am I missing something here?

    1. There isn't a button, but there is an email address over there => in the right sidebar, right under the "About Me" box.

      It's been a long time, like 25 years or so, since I studied up on reinforced concrete for a test, but as I recall, it's a bit more involved. Concrete is a composite (cement and gravel/sand) and reinforcing it makes it doubly so. Concrete is strong in compression: if you poured a large thin slab (on a good, solid ground), you could put a lot of load on it. If you put a shear load on it, though, like lifting one corner of that slab and trying to lift the whole slab with it, it'll break. The idea is that the composite transfers the shear stresses into the rebar. They can handle those stresses better than the concrete can handle it. So the concrete isn't bearing the entire load; it's transferring some of the stresses into the reinforcing metal.

      Concrete undergoes brittle failure. If you tried to stretch it, it will just snap. The alloys are very ductile and won't just suddenly snap, so while you won't stretch reinforced concrete much, the idea is to transfer the stresses into the metal.

      Composites behave differently than either material by itself. Take fiberglass - properly called "FRP" or fiberglass reinforced plastic. Graphite composites are an even better example. The cloth is very flexible, way too floppy, but hard to cut. If you could isolate strands and just pull on them, they can be as strong as steel, if not stronger. The epoxy is more rigid but brittle; nowhere near as strong as the cloth. The composite is much less flexible than cloth and much less brittle than epoxy.

  3. OK, makes sense enough.
    Thanks for your reply.

  4. Your post immediately reminded me of Rearden Metal from "Atlas Shrugged"!