Study on the Collapse Behavior of Nonductile Reinforced Concrete Frames Subjected to Earthquake Loadings

• Main Authors: Wu-Wei Kuo, 郭武威
• Other Authors: Ing-Jaung Lin
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/70490109558752097631

博士 === 國立臺灣科技大學 === 營建工程系 === 97 === During the 1999 Chi-Chi Taiwan earthquake, a large number of older buildings sustained severe damage or complete failure, and there were thousands of casualties and a great loss of property. A large majority of building collapse resulted from the loss of vertical-load carrying capacities of columns. Most of damaged columns were found with non-ductile detailing, such as widely spaced hoops with 90 degree end hooks. These columns are known to have poor seismic performance in terms of ductility and energy dissipation capacity. Therefore, shake table tests of four 1-story-3-bay reinforced concrete frames using 921 earthquake records were conducted to study the dynamic collapse behavior of non-ductile columns. Each specimen was composed of 2 ductile columns and 2 non-ductile columns. Test variables considered were lap splice on longitudinal bars at the bottom of non-ductile columns and retrofitting using wing walls. Test results showed that the onset of degradation of lateral resistance was about at the story drift ratio of 3.5% to 4.5% for non-ductile columns dominated by flexural strength, and shear failure took place at the drift ratio of around 4.5% to 5.5%. This indicated that nonductile reinforced concrete columns with low axial loads, dominated by flexural strength, provided acceptable drift capacity. After shear failure, the axial load carrying capacity of nonductile columns started to decrease, and the axial load could be redistributed by the frame system. No.3 rebars of 30db lap splice lengths in non-ductile columns could yield and the section reached the nominal flexural strength, although the hoops were widely spaced with non-seismic end hooks. Bond slip failure of columns caused different response in positive and negative direction. The large deformation along with sustained strength after bond slip failure of lap splice could occur only once in this study, which was not observed in the negative direction. Therefore, the drift capacity was less than that of columns without any lap splice. However, if larger rebars (no smaller than No. 7) with 30db lap splice lengths in real full scaled structures, the required lap splice lengths calculated from current ACI 318-05 code will exceed 30db; therefore, lap splice lengths of 30db is only allowable for small size rebars (no larger than No. 6). Retrofitted using wing walls not only increased the initial stiffness and lateral resistance of nonductile columns, but also obviously enlarged the drift capacity, slowed down the degradation of lateral resistance and increased the stability of the frame system. However, due to the separation of wing walls from columns at the early stage, the combination of these elements could not be developed effectively; therefore, in practice the interface shear capacity should be checked in advance to avoid failure occurred as in this study. On the other hand, force transfer mechanism of reinforced concrete beam column elements and methods on how to establish the force displacement curve of columns were also proposed in this study. The proposed force transfer mechanism, considering the size of members, transverse reinforcement and concrete strength to determine the domain of D and B regions of elements, has conceptual discrepancies with the current ACI 318-05 code, which assumes that the dimension of the D-region is only related to the height of the member. Comparing the solutions of the proposed method with the 198 test data from the literature, the ACI 318-05 code provides a simple and effective design procedure for the shear strength estimation, while the proposed approach is more complicated but bears more physical significance. The proposed force displacement curves, including flexural failure and flexural shear failure, were compared with experimental results. Two prototype specimens could be reasonably predicted, while the large displacement caused by lap splice failure was unable to be modeled accurately. Finally, four existing methods to predict the force displacement relationship were compared with the test results to verify the applicability of predictive models under dynamic loadings. Predications by Zhu et al. showed the closest trend with experimental curves, and the updated version of ASCE/SEI 41-06 code gave the conservative predictions on drift capacity, which improved the over conservatism of drift capacity and unrealistic initial stiffness predicted by FEMA 356.