Analytical and Experimental Studies of the Seismic Performance of Reinforced Concrete Structural Wall Boundary Elements
- Author(s): Hilson, Christopher William
- Advisor(s): Wallace, John W
- et al.
Following the February 27, 2010 Mw 8.8 Maule earthquake, an international effort was undertaken to better understand reasons for observed damage to concrete structural walls in buildings located in the affected region of Chile and to address potential design implications. The Chilean building code for concrete structures is based on the U.S. ACI 318 building code; however, based on the observed performance of over 400 buildings in the March 1985 earthquake-impacted Viña del Mar, Chilean Code NCh433.Of96 included an exception that special boundary elements (SBEs)--which are commonly required for walls in U.S. buildings--need not be provided. By taking exception to the special boundary element detailing provisions, the Chilean code allowed thin wall boundary zones with relatively large (typically 20 cm) spacing of transverse reinforcement (essentially unconfined) to be constructed. Given these differences, the 2010 earthquake is an excellent opportunity to assess the performance of reinforced concrete buildings designed using modern codes similar to those used in the United States. Data from damaged and undamaged buildings, as well as from parametric and experimental studies, are used to provide recommendations to improve the efficacy of U.S. provisions designed to inhibit structural damage at wall boundaries.
Seven Chilean buildings were selected to investigate the performance of boundary elements during the 2010 earthquake. Several walls from each of the seven buildings were chosen to evaluate the ACI 318-11 Section 18.104.22.168 displacement-based trigger equation for determining if SBEs would have been required and if observed damage was consistent with the evaluation result (i.e., SBE required, no damage; SBE required, damage observed). The propensity of boundary longitudinal reinforcement to buckle was also investigated, taking into consideration the influence of boundary transverse reinforcement configuration and longitudinal reinforcement strain history. In conjunction with assessments of in-situ wall performance, parametric studies we conducted on wall sections with various attributes and laboratory tests were performed on prisms representative of wall boundary elements using NEES@UCLA facilities to further investigate structural wall boundary element performance under reverse-cyclic demands.
The evaluation indicated that Chilean wall sections where the neutral axis depth exceeded the limit imposed by ACI 318-11 equation 21-8 and SBE-level detailing was not provided typically suffered significant concrete crushing and longitudinal bar buckling. Rebar buckling studies showed that, under large magnitude strain reversals, longitudinal reinforcement in flanged, tension-controlled walls is susceptible to buckling at non-flanged wall boundaries. Test results demonstrated that spacing-to-bar-diameter (s/db) ratios of 10.7 (8" spacing) and 8.0 (6" spacing) led to concrete cover loss initiated by longitudinal bar buckling upon reloading into compression from tension strains of approximately 2% or larger, followed by substantial strength loss on the subsequent loading cycle into compression. Specimens that did not exhibit damage initiated by bar buckling eventually failed due to global out-of-plane buckling (lateral instability), or by sudden crushing failure after cover concrete spalling. Results from analyses and experimental studies suggest compression reloading strains greater than 2% may initiate buckling of longitudinal reinforcement if spacing of transverse reinforcement is extended to the maximum permitted by ACI 318-11 for non-SBE configurations, and bar buckling may be possible in SBEs with reloading strains greater than 4%.