UC Berkeley Web Feature

'Dateline' brings down the house with shake table quake simulation

By Liese Greensfelder, Media Relations | 23 May 2006

BERKELEY — Television viewers watching NBC's "Dateline" show on Sunday evening got a good idea of the damage an 8.0 earthquake could wreak on the thousands of single-family houses built in the 1930s and 1940s in San Francisco. The popular news program filmed an experiment featuring such a house being shaken on the earthquake simulator at UC Berkeley's Earthquake Engineering Research Center in Richmond.

If you missed the show, you can find a transcript of it — along with video footage of some of the shaking and rolling — on the MSNBC website.
 
The footage also featured the Sunset district family who lived in the dwelling that was replicated for the test. They were shown viewing live video of the event, in which the contents of their living and dining rooms smashed into piles of rubble on the floor.

The experiment was commissioned in March by NBC, said Khalid Mosalam, UC Berkeley associate professor of civil and environmental engineering. The network was looking for a visual project for the anniversary of the 1906 quake that would also further earthquake research and education. The project that NBC and Mosalam came up with — testing a two-story, furnished house typical of those built in the 1930s and '40s in San Francisco's Richmond and Sunset districts — was also of interest to the city of  San Francisco.

A replica of a real house, scaled just a bit smaller, was built at the campus's Earthquake Simulator Laboratory in Richmond. The central feature of this recently upgraded research facility is a 20' x 20' "shake table" that can simulate many ways the earth can move during quakes. The table is capable of testing structures up to 40 feet tall and weighing up to 60 tons.

The two-story test house built over a garage was the first such structure tested on the table. "We had a 1970s three-story house on the table about five years ago," Mosalam said. "But that structure was more code compliant for earthquakes, so it was a lot stronger than the current one, which was a replica of houses built before the codes were written." 
 
Mosalam said that he was impressed by the amount of freedom the network gave him on the scientific end of the project. "We suggested two levels of study," he said. "One was to look at the building contents, its furnishings and furniture. For that one, the effect on the structure itself was not the goal."

The first two of the three tests Mosalam ran compared what would happen if the furniture was secured, then unsecured, during a 52-second, approximately 8.0 quake. During these two tests, the house was fit with additional diagonal bracing to ensure that it remained standing. The third test was run without the bracing, to study the effects of the quake on the house itself. To prevent the house from collapsing and damaging the shake table, it was secured with loose cables during the tests.

During the third test, "The structure did collapse," Mosalam said, "but the cables tightened and held it up." Viewers may not have understood what they were seeing at this point, however, because due to time constraints, this portion of the test was broadcast without any comment. 

Because the building's first story was a garage, it was open on one side. A "soft story" like this generally makes a structure more susceptible to earthquake damage, Mosalam said.  "The structure was very flexible," during the test, he explained. "At the end, it was like a box sitting on very weak legs, and the legs had moved laterally to significant deformation."

This was the portion of the test of greatest interest to San Francisco engineers. "They are looking at some very simple retrofit mechanisms that involve the use of garage doors," Mosalam said. "If we can make the garage doors work to strengthen the first story on houses like this, that would be great. We tried a couple of scenarios that were somewhat successful."

During the tests, more than 70 sensors affixed to the house were collecting data. Mosalam and his graduate students have already written several preliminary reports based on the results, but the research team will be putting in many more hours this summer before all the analyses are complete.