The National Science Foundation has awarded $4.6 million to support earthquake engineering research at the University of California, Davis. The grant will fund upgrades to the large centrifuge at the Center for Geotechnical Modeling (CGM), part of the Department of Civil and Environmental Engineering at UC Davis. It will also pay for development of a new robot to work on the centrifuge, new sensor equipment, and for a high-speed computer network that will make UC Davis a "host facility" in the national Network for Earthquake Engineering Simulation. With a nine-meter (30 foot) arm, the UC Davis shaking-table centrifuge is one of the largest machines of its kind in the world, capable of spinning a payload of up to five tons to 40 times normal gravity, or 40 g, according to Facility Manager Dan Wilson. The centrifuge has been used to investigate problems such as how shaking is transmitted from rock through soil, how soils liquefy during earthquakes, how different types of foundations stand up to earthquake shaking, and how much ground movement will occur during earthquakes. Under normal gravity, scale models do not behave in the same way as real soils and structures, explains civil engineering professor Bruce Kutter. "By subjecting scale models to large centrifugal accelerations we increase the pressures in the model to match those in deep prototype soil deposits. For example, one foot depth of soil at 40 g behaves in the same way as a 40 foot depth of soil at 1 g," says Kutter. "We can apply the same pressure in the model as in the field, which makes the model tests much more accurate." While the centrifuge is spinning, a shaking table mounted on the end of the arm can be vibrated, sending pressure and shear waves up through the model mounted on the shaking table. "Very much like a real earthquake, the waves propagate up through the soil," says Kutter. Engineers at UC Davis and around the world are developing sophisticated computer models to predict the effects of earthquakes, says Kutter, but different theories give different answers. "We need real experimental data to test the theories, and real earthquakes provide an insufficient database," he says. The data obtained from experiments with the centrifuge can be used to study the mechanisms at work and verify the ability of computer models to predict the results, according to Kutter. Upgrades to the centrifuge will increase its speed and capabilities. "We should be able to go to 80 g with the upgrades," says Wilson. A new two-dimensional shaker will shake samples vertically as well as horizontally. "This will be the only machine in the world to do this," says Wilson. The upgrades also include designing and building a robot that can actually work on the samples while the centrifuge is spinning. The robot will move on a gantry over the payload and be equipped with stereo video cameras and a tool interface, says Kutter. The robot will be used to place or move sensors, or carry out construction tasks such as pile driving. Observers in the control room, or anywhere else on the NEES network, will be able to control the robot and inspect the spinning model using 3-D goggles, says Kutter. The robot will be designed and constructed by a team led by Steven Velinsky, director of the Advanced Highway Maintenance and Construction Technology Research Center at UC Davis. "Many of the skills that we have developed with highway robots can be applied here," says Velinsky. The biggest challenge will be in the structural requirements, he says, as the robot will be built to withstand forces up to 100 g. "The UC Davis grant includes the development of hundreds of digital Micro Electro-Mechanical Systems (MEMS) sensors, to drastically increase the spatial resolution of data during a model earthquake," says Kutter. The mechanical sensor, digitization circuitry and memory will all be on one chip. The sensors will collect data on pressure, soil movements and strains. Each sensor will be a node in the data network, explains Wilson. Instead of feeding data to a central computer, each sensor will calculate and store its own data and pass it to the computer when requested, making data collection much more efficient, he says. The upgrades also include an expanded control room, with a 10-foot video wall to display video, 3-D images and sensor data simultaneously, says Wilson. The funding is part of $82 million committed by NSF to the national Network for Earthquake Engineering Simulation over the next five years. A key aim of the network is to allow researchers in different sites to carry out experiments together and to share data. "We want someone in North Dakota to be able to collaborate on an experiment with us," says Kutter. "For example, a graduate student could come and work at UC Davis for a couple of months to build and test the model, then the supervisor could participate in the same experiment from elsewhere on the network." As it can take one or two months to construct a model for an experiment that lasts one or two days, the national network will allow more efficient use and more shared use of these large, costly machines, says Kutter. "We have a good history of sharing our machines," he says. Results from experiments over the last five years have been posted on the CGM Web site http://cgm.engr.ucdavis.edu for anyone to download and use. The grant for the UC Davis large centrifuge is the largest NEES grant so far from NSF, and will make UC Davis prominent in the network, says Kutter. NSF has also recently awarded funds to the University of Texas, Austin, for a large three-dimensional mobile shaker; to Rensselaer Polytechnic Institute, N.Y., for upgrades to their existing centrifuge; and to the University of Illinois at Urbana-Champaign for a scoping study of the NEES high-performance network.