A series of Precast/Prestressed Concrete Institute-sponsored torsion tests on precast L-shaped spandrel beams has yielded new design procedures which, compared to those in current code requirements, will economize production through lower material costs (approximately 30 percent less web steel is required) and significantly decreased labor (about 50 percent less time to tie the steel cage). Investigators at North Carolina State University, Raleigh, recently concluded the tests on 12 full-size specimens and will prepare a final report and formal design procedures by mid-2009.
The tests were conducted to investigate the behavior of end regions of such beams subjected to eccentric loading, according to Paul Johal, P.E., director of PCI Research and Development. The objective was to develop appropriate design procedures that simplify reinforcement detailing requirements for precast, L-shaped spandrel beams subjected to eccentric ledge loading. The appropriateness of the current design procedure has been questioned, said Johal, because it is based on data collected from tests of conventionally reinforced and prestressed concrete members having compact cross sections.
Precast specimens, software, engineering support and workers for the testing were donated by PCI members Metromont Corp., Tindall Corp., Finfrock Industries Inc., High Concrete Group LLC, Stresscon Corp., JVI Inc., and Leap Software (now Bentley), Wiss, Janney, Elstner Associates, and NCSU. A final report containing the new design procedures should be completed by the end of June. The new design guidelines also will be included in the next edition of the PCI Design Handbook. The new design procedures will maintain the required high levels of safety while reducing internal steel.
The first phase of the experimental program included four 8-in.-wide _ 60-in.-deep _ 45-ft.-long L-shaped spandrel beams. Two of these beams were prestressed and two utilized conventional flexural reinforcing steel. Within each group, one specimen was reinforced at its ends with closed stirrups designed and detailed per current code requirements to resist shear and torsion. The other specimen in each group was reinforced at its ends with simple, open end-region reinforcement, designed and detailed to resist shear and plate bending only, ignoring torsion.
All four beams were subjected to several loading cycles to the full factored load level (calculated failure load) and sustaining that load for 24 hours. After unloading, each beam was loaded gradually until actual failure occurred. The failure mode in all four specimens appeared to be a combination of plate bending, shear and twisting at a load significantly higher than the calculated failure load. There was no evidence of the classic torsional distress Û that is, spiral cracking and face shell spalling.
The second experimental phase consisted of six 8 in. _ 60 in. _ 45-ft.-long specimens, three with 8- _ 8-in. ledges, and three with individual spot corbels. All were reinforced with open reinforcement at the ends, that is, no closed stirrups. The results of the tests to failure were similar to those of the first four-beam phase. All 10 specimens were 8 in. wide _ 60 in. deep, with an aspect ratio of 7.5.
The final phase consisted of two 10-in.-wide _ 46-in.-deep _ 45-ft.-long prestressed L-spandrels. At an aspect ratio of 4.6, they were more compact than the first 10 specimens. Both were reinforced with open reinforcement at their end regions. The results of the tests to failure were similar to those of comparable specimens among the first 10 deeper spandrels.
The experimental results have been used to calibrate an analytical model. Analytical and experimental results will be used together to develop a refined and improved design procedure. According to Johal, the tests on the more compact L-spandrels were quite valuable since they showed that the basic failure mode doesn't change significantly as the slender spandrel (aspect ratio 7.5) becomes more compact (4.6).