A series of pretensioned strands and traditional rebar were added to concrete columns that were then bookended by steel caps. The arrangement, which is meant to help bridges withstand strong earthquakes, is currently being tested on shake tables at the University of Nevada. © John Stanton
Shake tables will test a precast bent system that can be constructed using accelerated bridge construction (ABC) techniques and offers better seismic performance.
July 15, 2014—A research team is completing a series of tests on the large shake tables at the University of Nevada that are focusing on a one-quarter scale version of a new bridge bent system that can be constructed with the aid of accelerated bridge construction (ABC) techniques and also offers better seismic performance. The final test will replicate the magnitude 6.9 earthquake that struck Kobe, Japan, in 1995.
Led by John Stanton, Ph.D., P.E., a professor in the civil and environmental engineering department at the University of Washington, the team includes Marc Eberhard, Ph.D., a professor at the University of Washington, and David Sanders, Ph.D., F.ASCE, a professor at the University of Nevada. Two graduate research assistants at the University of Washington, Travis Thonstad, S.M.ASCE, and Olafur Haraldsson, S.M.ASCE, also are members, along with Islam Mantawy, a graduate research assistant at the University of Nevada.
Stanton says that traditional ABC methods that rely on precast concrete pose a problem in seismic areas because the on-site connections between beams and columns are usually weaker than the members themselves. “The bridge is most easily constructed if the individual precast pieces are straight, like traditional beams and columns, and are connected at their intersections, but unfortunately those intersections are exactly where the earthquake forces get to be the worst,” Stanton says. “So by making the bridge easy to construct, you are then making your earthquake headache much bigger. It’s a real challenge—how can you make these connections work for both constructability and earthquake resistance? We have beaten our heads against brick walls for a long time to try to work that one out.”
And then, Stanton says, one day several years ago he had an epiphany while preparing for a seismic engineering conference. After sketching out the concept, he developed more detailed drawings and sought the opinions of colleagues and contactors, he recalls. All indicated the concept was feasible, and subsequent testing has borne this out.
The team conducted pseudostatic tests of the bents that progressively increased the horizontal forces. In those tests, the team first pushed the tops of the columns by the approximately 2 percent drift expected in a design earthquake and then by the 3 to 4 percent drift expected in a maximum credible earthquake, which has a return period of 2,500 years. The testing then went beyond that.
“We have, in the lab so far, had this system up to a 10 percent drift, compared to the 3 to 4 percent maximum credible,” Stanton says. “And basically we had no concrete damage at all.”
How is that possible? Rather than bend to accommodate the horizontal movements of the bridge deck, the columns in the bents rock as rigid bodies. Because the columns do not deform, they do not crack. And to stop them from falling over when they rock, the columns are reinforced vertically with a series of pretensioned strands as well as traditional rebar. The system effectively functions like the springs and shock absorbers in an automobile, Stanton explains. The rebar is fully bonded to the concrete and dissipates energy by yielding during an earthquake, whereas the pretensioned strands remain elastic, bringing the structure back to its starting position.
Stanton’s approach differs from a traditional pretensioned system in that the strands are partially debonded from the concrete, which is thus isolated from the deformation of the strands during an earthquake. The strands are debonded by being wrapped in plastic piping, and in this way they are able to move through the central region of the column during an earthquake without disturbing the concrete, acting like giant rubber bands to snap the column back upright when the ground motion stops. The strands are anchored into the concrete at the tops and bottoms of the columns, which in turn are embedded in respectively cap beams and cast-in-place footings. Smaller portions of the rebar also are debonded to prevent fracturing as a result of excessive strain.
During an earthquake, the bents rock at the top of the footings and at the bottom of the cap beams. Preventing the concrete at these points from being crushed was the “last headache” of the design, Stanton says.
“When the column rocks, you get some monstrous compression stresses where it is still in contact, because a tiny area of contact has to carry both the weight of the bridge and the force from the prestressing strands,” Stanton says. “In order to stop the concrete getting smashed up, you have to do something to protect it. We’ve been through several stages of development for that. We tried using regular concrete, [but] we had damage at the rocking interface. Then we tried using fiber-reinforced concrete and it did better, but it still didn’t completely eliminate the crushing damage.”
The eventual solution that the team developed is a confining steel tube, one placed at the top and another at the bottom of each column in the bent. The tubes have the same diameter as the column and extend up the column a distance of approximately half the diameter, forming a protective cap to the end of the column. In the pseudostatic testing that has been conducted so far, the caps have been successful in eliminating the damage.
By extending only a short distance from the top and bottom of the column, the caps provide the confinement needed for the concrete while enabling the precast fabricators to cast the columns in the more efficient horizontal position.
Stanton says the team is optimistic about the test it is about to conduct on a 70 ft long span built atop three shake tables in Reno. This round of examinations involves a more realistic test of a system of bents, in contrast to individual columns, in continuous motion.
“In Reno, they are shaking it on a shaking table, so the whole thing is dynamic,” Stanton says. “You don’t get the luxury of slowing it down and looking at cracks halfway through the loading. The earthquake hits, and the bridge just has to go through it.
“There may be something that is different between the static and the dynamic that would give us some difference in performance,” Stanton says, adding that, if so, he would be “very surprised.” As he puts it, “I’m actually hoping it will be a really boring test and we will learn nothing that we don’t already know.”
Stanton says the team expects to collect a great deal of data from the testing and plans to analyze those data for several months. The data will eventually be archived and be made available to the public via the George E. Brown, Jr. Network for Earthquake Engineering Simulation (NEES). The University of Nevada is part of the NEES.