Title:
Steel Coupling Beams in Low-Seismic and Wind Applications
Author(s):
Bahram Shahrooz
Publication:
CRC
Volume:
Issue:
Appears on pages(s):
319
Keywords:
DOI:
Date:
1/1/2019
Abstract:
Shahrooz Coupled structural (shear) walls (CSW) are a common structural system. This system is comprised of two or more structural walls that are linked, typically, at each floor by coupling beams. Based on the expected level of inelastic deformations, composite structural (shear) walls can be classified as Composite Ordinary Shear Wall (COSW) or Composite Special Shear Wall (CSSW). One common composite system involves linking reinforced concrete wall piers by steel (or steel-concrete composite) coupling beams that are embedded in the wall piers. Design and detailing of steel coupling beam-wall connection in COSW was the focus of the research reported herein. In the 2010 and earlier versions of AISC 341 Seismic Provisions, the coupling beam-wall connection was designed to develop the coupling beam's expected capacity. This provision in the 2016 version was replaced by the requirement that the connection in COSW be designed only to develop the demand from the coupling beam as calculated by linear-elastic analysis with no ductile detailing requirements. As a result, the design and detailing of the embedment region has been relaxed. This change leads to shorter embedment lengths and smaller reinforcement in the embedment region. Analytical studies conducted at the University of Cincinnati indicated the shorter embedment length could accelerate the loss of coupling beam-wall connection integrity, leading to a reduction in the level of coupling action between the wall piers. The loss of coupling action will affect the demands in the wall piers, and their capacities could be exceeded. Moreover, inter-story and overall drifts could surpass acceptable limits. Primarily to remedy these observations, AISC 341 Seismic Provisions was modified in 2022 by specifying a minimum embedment length of not being less than the coupling beam’s depth and requiring additional longitudinal reinforcement along the embedded region. A coordinated experimental (consisting of two half-scale and four three-quarter beam-wall subassemblies) and analytical study was conducted to examine the current design provisions for steel coupling beams in COSW outlined in AISC 341-2022. It is important to note that the current (and previous) AISC Seismic Provisions for coupling beams in COSW and CSSW are solely based on experimental research focused on coupling beam-wall connection details intended to resist high seismic loads. To the best of the authors’ knowledge, no experimental research had been conducted to understand the performance of COSW prior to the study presented in this report. The research data were used to evaluate the current AISC 341 Seismic Provisions and to develop new design and detailing provisions for COSW. Motter Structures are typically designed to yield and sustain damage in a controlled manner during design level earthquakes. While a similar approach has traditionally not been used for design-level windstorms, the recently-published ASCE/SEI Prestandard for Performance Based Wind Design (ASEC/SEI, 2019) describes design for modest nonlinear response of select structural members such as coupling beams. In this study, four steel reinforced concrete (SRC) coupling beams, with steel sections that embedded into a reinforced concrete wall, were tested quasi-statically under fully reversed cyclic wind demands with peak beam deformation of three times the yield rotation. The beams and walls were designed in accordance with seismic provisions in AISC 341-22 Section H5, and the walls were compliant with ACI 318-19 Section 18.10.6.5. The exception was the wall reinforcement for two of the four tests, in order to examine potential reductions to that prescribed. For one of these tests, the ratio of the strength of wall longitudinal reinforcement crossing the embedment length to that prescribed was 0.53. For the other of these tests, this value was 0.22 and the wall boundary transverse reinforcement at the embedment zone was also less than that prescribed. During each test, the wall was subjected to constant axial gravity load and fully reversed-cyclic lateral loading that was linearly proportional to the load in the test beam. The ratio of wall shear to beam shear was constant for the four tests, while the ratio of wall moment to beam shear was the same for three tests and was larger for one of the tests with wall reinforcement compliant with AISC 341-22 Section H5. For the test with the least wall reinforcement, significant damage was observed in the wall at the embedded connection. The load developed in the beam was limited by yielding in the wall. Significant pinching, characteristic of gapping, was observed in the load-deformation response. Significant stiffness degradation occurred for repeated loading cycles at 40% of the computed peak strength, and the beam was unable to develop 75% of the computed beam strength, despite being loaded to 6.0% chord rotation. The quantity of wall reinforcement was inadequate to promote favorable performance. Performance was more favorable for the other three tests, which were observed to have similarities in damage patterns and load-deformation responses. Damage concentrated at the beam-wall interface, with the majority of the coupling beam deformation at this location. Although the stiffness degradation for these three tests was much less than the test with wall yielding, stiffness degradation for repeated loading cycles at a given load level was found to be significant in these three tests, particularly for larger loading levels prior to yielding. However, significant strength degradation of initial cycles at new peak deformation demands was not observed in any tests, and significant pinching in the load-deformation response was not observed for the three tests with more favorable performance. Peak load resistance was reached at peak deformation demand, which was 5.70% chord rotation for the test with the largest wall demands, 4.80% chord rotation for two tests, and 6.0% for the test with wall yielding. The primary difference in load-deformation responses for the wind tests conducted in this study and previous seismic tests was the stiffness degradation with repeated loading cycles, noting that the number of cycles used in the wind tests was substantially higher than that used in typical seismic tests. Stiffness for the first loading cycle at 75% of the expected strength was examined using the results from the three test beams from this study that reached this level and three SRC coupling beams from other studies. The difference between stiffness in the positive and negative direction was more significant for larger cyclic wall demands, with higher stiffness in the positive direction due to wall demands producing compression at the embedment region. The average of the positive and negative stiffness was larger for walls with higher compression force in the wall on the positive excursion. If cyclic stiffness degradation for repeated cycles at a given increment is not explicitly modeled, it is recommended to use a backbone model based on average values of all cycles at each increment, as this would lead to equal area under the curve for the backbone model and test data. Parameters for a bilinear backbone model for nonlinear wind design are suggested, with effective stiffness of 75% of that prescribed in AISC 341-22 for seismic design, a yield force computed using moment-curvature analysis at full yielding of the tension flange using expected material properties, a computed expected strength from AISC 341-22, and a post-yield slope based on 4.0% chord rotation from yield to expected strength. It is recommended that the hysteretic model be determined by modeling the test beams and calibrating to dissipated energy test data for the three tests with favorable performance. Each of the four backbone parameters were determined based on fit to test data. This study did not include testing on SRC coupling beams that were designed using provisions in AISC 341-22 Section H4 and tested to peak deformation demands more consistent with ordinary walls. It is recommended that nonlinear wind design of steel reinforced concrete (SRC) coupling beams follow the seismic provisions in AISC 341-22 Section H5. It is recommended that the quantity of wall longitudinal reinforcement crossing the embedment length prescribed by AISC 341-22 Section H5 be reduced by 50% for cases in which wall demands do not exceed that applied for the test that supported this recommendation. These peak wall moment and tensile strain demands were 0.29My and 0.00019 tensile strain in outermost reinforcement at the coupling beam mid-height and an average of 0.04My and -0.00001 tensile strain (0.00001 compressive strain) in outermost reinforcement over one story height, taken as half a story above and below the coupling beam mid-height. These demands were determined from moment-curvature analysis for the moment and axial load in the wall determined by assuming transfer of coupling beam shear and moment to the wall at coupling beam mid-height. This recommendation applies for both seismic and wind design, due to favorable performance for this test under wind demands to a peak deformation of 4.65% chord rotation.