Repeated simulated earthquakes in a mechanical device called a centrifuge are second only to the real thing for developing reliable ways of analyzing seismic safety in the event of the destructive phenomenon known as "soil liquefaction," according to the organizers of a four-day conference ending today at the University of California, Davis. The earthquake engineering meeting marks the end of a four-year international collaboration. Funded by the National Science Foundation, the project culminated this week with the first systematic review of the strengths and weaknesses of the most popular and most advanced earthquake engineering methods currently used around the world. Organized jointly by UC Davis and the California Institute of Technology, the meeting also helped establish centrifuge testing -- which realistically mimics the action of earthquakes upon scale-model simulations of soils and structures -- as a standard-setting technique for geotechnical engineering in certain situations. Not only will centrifuge modeling show what happens, the technique will give researchers clues about why soil liquefaction behaves as it does. Finally, centrifuge models can be used as they were in this project -- to test the accuracy of the current methods used to design structures that will withstand liquefaction. Present analytical methods are based on laboratory tests of soil samples and post-mortem evaluations of major earthquake devastation around the world. With the help of the centrifuge, researchers hope to come closer to figuring out the fundamental mechanisms of liquefaction and take much of the guesswork out of the present methods. The most complicated and costly soil behavior that occurs during earthquakes, soil liquefaction happens when solid ground literally turns to liquid and boils with sand and water during earthquakes. During an earthquake, large structures like freeways or waterfront-retaining walls often experience the effects of soil liquefaction and either settle uniformly or tilt and collapse. Freeways and bridges seem to be particularly susceptible to cracking and collapsing. Predicting what will happen to these very large structures during liquefaction has become one of the most pressing geotechnical engineering problems. "The liquefaction problem affects the stability of all existing infrastructure systems -- transportation, dams, waterfront-retaining structures," said UC Davis engineering professor Kandiah "Arul" Arulanandan, head of the four-year project and an organizer of the conference. "Therefore we need to know how best to analyze the problem to assess the consequences of soil liquefaction. Then and then only can you implement remedial measures so liquefaction doesn't cause billions of dollars of damage." The results of the four-year project revealed a new direction for research aimed at developing more economic and reliable analytical methods to build more stable structures. Perhaps more important for the short term, the results also exposed significant limitations in most present analytical methods. Centrifuge tests and be used to verify and improve analytical methods that then can be used to evaluate remedial efforts to strengthen existing structures. Engineers now designing new power plants and ports and those trying to evaluate options for seismic upgrading of bridges, freeways and waterfront structures might want to check the conference results to see just how well their chosen method of analysis fared in centrifuge tests. "Most current methods cannot predict reliably the consequences of soil liquefaction in the detail necessary to design safe and economical structures," Arul said. The least expensive and simplest of the reviewed methods did the poorest job of predicting centrifuge liquefaction results, while the most expensive and some complex methods corresponded reasonably well with centrifuge results, says conference co-organizer Ronald F. Scott, Hayman professor of engineering at the California Institute of Technology. Engineers rely on these methods to help them judge what will happen to a structure during liquefaction, but not all methods proved dependable for all situations. Some of the complex methods that did well in predicting certain centrifuge liquefaction results failed on others, Scott says. Conversely, for several centrifuge simulations, the better models all predicted different results. "The conference will have a major impact on how the profession will analyze liquefaction problems in the future," Arul said. The conference brought together two halves of a double-blind study. One group of researchers predicted the centrifuge test results using engineering methods now in common use for analyzing the liquefaction behavior of large infrastructures, such as transportation systems, dams, pipelines, power plants and port facilities. Then, another group of researchers conducted repeated centrifuge tests of specified soil models. The two groups did not learn the full results until this week. The situations where the centrifuge results seemed clear and immediately applicable were scale models that represented flat land on which an office building might be located, on the type of embankment used to support roads and bridge approaches, and on a type of waterfront-retaining wall used in port facilities. Other centrifuge tests need to be repeated before reliable data about other sites can be added to this new large data base, the first collection of its kind and scope, organizers say. Otherwise, the data will be available in computer accessible form to test and verify new and improved engineering methods. "It is clear that this is a major advance in the art of centrifuge model-testing," said conference participant Robert Whitman, professor emeritus of civil and environmental engineering at the Massachusetts Institute of Technology. "Engineers who design cars and airplanes can experiment with prototypes to work out problems that can be corrected in subsequent attempts. Engineers who design and build large civil structures, such as freeways and waterfront retaining walls, must do it right the first time. The shaking tests aboard a centrifuge are the next best thing to an earthquake."