“Real-time hybrid simulation” has proved to be a reliable and efficient method for testing a powerful new damping system designed to reduce the structural damage and injuries that an earthquake can inflict.
The simulations are referred to as "hybrid" because they combine computational models with data from physical tests.
Researchers are able to perform structural tests at slow speed, but testing in real-time – or at the actual speed of an earthquake – sheds new light on how dampers may perform in structures under actual conditions. The real-time ability has only recently become feasible due to technological advances in computing.
The magnetorheological fluid dampers (or MR dampers), which have been subjected to the hybrid simulations, are shock-absorbing devices containing a liquid that becomes far more viscous when a magnetic field is applied.
"It normally feels like a thick fluid, but when you apply a magnetic field it transforms into a peanut-butter consistency, which makes it generate larger forces when pushed through a small orifice," said Shirley Dyke, a professor of mechanical engineering and civil engineering at Purdue University who has led the research with her doctoral students Gaby Ou and Ali Ozdagli.
This dramatic increase in viscosity enables the devices to exert powerful forces and to modify a building's stiffness in response to motion during an earthquake. To date the MR dampers have only seen limited commercial use and are not yet being used routinely in structures.
Dyke and her students are working with researchers at the Harbin Institute of Technology in China, home to one of only a few large-scale shake-table facilities in the world.
"Sometimes real-time testing is necessary, and that's where we focus our efforts," she said. "This hybrid approach is taking off lately. People are getting very excited about it."
The simulations can be performed in conjunction with research using full-scale building tests. However, there are very few large-scale facilities in the world, and the testing is time-consuming and expensive.
"The real-time hybrid simulations allow you to do many tests to prepare for the one test using a full-scale facility," Dyke said.
"The nice thing is that you can change the numerical model any way you want. You can make it a four-story structure one day and the next day it's a 10-story structure. You can test an unlimited number of cases with a single physical setup."
To prove the reliability of the approach, the researchers are comparing pure computational models, pure physical shake-table tests and then the real-time hybrid simulation. Research results from this three-way comparison thus far have demonstrated that the hybrid simulations are accurate.
Ou has developed a mathematical approach to cancel out "noise" that makes it difficult to use testing data and combined mathematical tools for a new "integrated control strategy" for the hybrid simulation. By integrating several techniques in the right mix, Ou has found the tests are more accurate than prior methods.
"It's a viable method that can be used by other researchers for many different purposes and in many different laboratories," Dyke said.