Civil engineering professor Steven Glaser can’t see the Hayward Fault from his office in UC Berkeley’s historic Hearst Memorial Mining Building, but he knows it’s there. “Just on the other side of that structure over there is the fault,” he says, gesturing towards a window. Is he worried? Not really. Thanks to a recent $90.6-million renovation, the building now rests on 134 stainless steel and rubber “base isolators” which act as shock absorbers in the event of a quake. “As long as we move less than 37 or 39 inches, we’re OK,” he says with a shrug.
Of course, Glaser knows not everybody can be so cavalier. For people living in earthquake-prone places like California and Japan, the possibility of a major quake is an uncomfortable fact of life. And while seismic engineering, like the Berkeley-developed system used in Hearst Memorial, has greatly improved since the ’89 quake rocked the Bay Area, there’s still plenty of room for improvement.
Enter Terra-Scope™, a CITRIS-sponsored and NSF-funded project aimed at taking existing seismological technology, the vertical seismic array, and using MEMS (micro-electromechanical systems) to make it small, cheap, and easy enough for anyone building or retrofitting a site to use. After several years of work, Glaser believes the system is almost ready for its first trial run. He plans to place two miniature vertical seismic arrays on either side of the Hayward Fault that transects Berkeley’s Memorial Stadium, which will soon be undergoing seismic renovations of its own.
At its most basic, a seismic array is a network of seismographs, instruments that record ground movement, spread across a geographic area. First developed in the late 1950s to monitor underground nuclear explosions, they have become a key tool in understanding how terrain responds to earthquakes. In the case of vertical seismic arrays, they are arranged deep into the Earth rather than spread across the surface, providing detailed information about how the seismic waves change as they pass through the various layers of soil, information that’s especially useful if you are retrofitting or constructing a building.
“With a vertical seismic array, you would know what types of accelerations, what types of ground motions the structure you had there would be subjected to so you could retrofit it or design for it. And it helps us intellectually understand a bit better how sites behave in general,” says Glaser. There’s also a preventative use. The smaller, initial portion of the earthquake called a P-wave, when detected by VSAs, could activate automatic shutdowns of train systems seconds before the big one hit, preventing earthquake-triggered train derailments. Meanwhile, foreshocks can predict large-scale site behavior based on small foreshocks.
Unfortunately, current VSAs are too pricey to be used for construction projects. “Right now these cost hundreds of thousands to millions of dollars to install. We’re trying to make it so that each pod would cost in the $5000 to $10,000 range. If it’s cheap enough, every structure could have one,” Glaser explains.
In Glaser’s system, seismic waves traveling through the Earth would strike a series of networked pods arranged vertically along a 2-inch diameter hole. As the waves passed through each pod, it would record a range of readings: acceleration, tilt, and pore pressure. When the aftershocks followed, a built-in memory buffer would allow the pod to record those while still processing the main quake’s impact. The combined measurements would be transmitted to a base station to be interpreted and posted online.
Of course, nobody wants to wait until the “big one” to find out whether they’re standing on shaky ground. A key part of Glaser’s system is a computer model of the soil that uses the VSA’s data from smaller “microearthquakes” to predict what would happen in larger ones. To test their model, Glaser and his team analyzed data acquired by standard vertical seismic arrays during the massive 1995 earthquake in Kobe, Japan, and then ran the same analysis on much smaller foreshocks. What Glaser found is based on the small quakes the model “could predict, wiggle for wiggle, what the output was going to be” for a large quake.
“We’ve found that a very simple model works better than a complicated model,” he explains. The streamlined data means less for the device to process and the user to wade through. “Instead of having to send lots of data, we would like to send information,” Glaser says.
This is not the first time Glaser has used MEMS devices for seismological research. This current work springs out of a CITRIS project from 2001. He and several other CITRIS researchers instrumented a building with wireless sensors that would detect when it had suffered potentially hazardous structural damage in a quake. There are also plans to use sensors developed by Glaser to detect the movement of soil in the retaining walls CITRIS’s new headquarters.