Scale Effect on the Shear Strength of Reinforced Concrete Columns with Termination of Main Reinforcements

A number of RC bridge columns which are similar with the columns which suffered extensive damage during the 1995 Kobe, Japan earthquake due to premature shear failure resulting from the termination of longitudinal reinforcements with insufficient development and the lack of shear reinforcements still exist. To improve the seismic performance, some of them were retrofitted. Evaluation of shear strength is important to determine the effectiveness of retrofit. To clarify the premature shear failure mechanism, a shake table experiment on a 7.5 m tall 1.8 m diameter full-scaled RC column with termination of main reinforcements was conducted. This paper presents evaluation of shear strength of concrete in full-scaled RC column with termination of main reinforcements. From the experimental results, the maximum shear stress of concrete was ′ c f 155 . 0 and it occurred before shear cracks initiated and tie bars yielded. After shear cracks initiated and tie bars yielded, the peak shear stress deteriorated to ′ c f 100 . 0 . Even when the shear cracks do not initiated or tie bars do not yield, the shear strength deteriorated due to the cyclic loadings. By comparing with the results of the scaled model column experiments in the past studies, the scale effect on shear strength of concrete is evaluated. By defining the shear strength as the maximum shear stress, the shear strength is proportional to 6 / 1 − d . On the other hand, by defining the shear strength as the shear stress when a tie bar yield, the shear strength is proportional to 3 / 1 − d . INTRODUCTION To reduce amount of longitudinal bars at the section where number of bars required at the maximum moment could be reduced based on the moment distribution, it was common practice until the mid 1980s for terminating longitudinal bars at mid-heights (cut-off) in reinforced concrete columns. Prior to the introduction of ductility design in 1990, bridges were designed based on a combination of the static lateral force method and the working stress design approach (seismic coefficient method). Main problems of design requirements at those days were insufficient development of terminated longitudinal reinforcements, overestimated concrete shear strength and insufficient amount of ties. In spite of those problems, shear dominant failure was not developed until the 1970s because large-section piers with enough redundancy for shear failure were constructed. However as number of columns with smaller section increased since the end of 1970s due to restrictions resulted from space limitation and smooth river flow, shear dominant failure was gradually noticed. Several bridges suffered damage due to this problem in the 1978 Miyagi-ken-oki earthquake, and the failure occurred more apparently at Urakawa Bridge during the 1982 1 Dept. of Civil Engineering, Tokyo Institute of Technology, Japan 2 Dept. of Civil Engineering, Tokyo Institute of Technology, Japan 2 Urakawa-oki earthquake. The shear failure due to cut-off was the major sources of contribution to the extensive damage of bridges during the 1995 Kobe earthquake. Various researches on the failure mechanism and the effect of seismic retrofit were initiated after the 1982 Urakawa-oki earthquake. Evaluation on the vulnerability of shear failure due to cut-off was studied by Yamamoto and Ishibashi et al(1984), Ozaka and Suzuki et al(1986), Kawashima, Hoshikuma and Unjoh(1995). They developed their own method for vulnerability evaluation and verified them based on cyclic loading experiments. Ikehata, Adachi, Yamaguchi and Ikeda (2001) studied failure process of model columns with cut-off based on quasi-static loading. Park (1996) studied the failure of a 18-span viaduct which collapsed during the 1995 Kobe earthquake based on New Zealand code. Whereas those studies contributed to the progress of seismic retrofit, reliability of the experiments using small-size models was always the arguments for simple extension to prototype piers under strong earthquake excitation. It should be noted that several large-scale experiments on the shear failure of bride columns were conducted. For example, Kosa et al. (1996) found based on in-situ lateral loading experiment for an 11 m tall 2 m diameter on-ramp bridge column on Hanshin Expressway that it failed in shear after flexural damage. Iwata, Otaki and Iemura (2001) found that a 5.5 m tall 2 m wide square section as-built column which had failed in shear failed in flexure after it was retrofitted by steel jacketing. Ohtaki, Benzoni and Priestley (1996) found based on cyclic loading experiments on a 3.66 m tall 1.83 m diameter pre-1971 column and its 1/3 scaled model that shear strength capacity decreased by 23% in the larger model than the 1/3 scaled model. It is important to note that the scale effect on the shear capacity to columns still remains unsolved problem. Shear capacity and its dependence on the size has been studied essentially for beams based on unilateral quasi-static loading, and only few investigations were conducted for bridges columns under shake table experiment. Based on the above background, a 3D shake table experiment was conducted for a 7.5 m tall 1.8 m diameter circular reinforced concrete column (called hereinafter as C1-2 column) using EDefense of the National Research Institute for Earth Science and Disaster Prevention, Japan (Katayama 2005) that is the largest in the world. This was conducted as a part of NEES/EDefense collaboration program on the seismic performance of bridges (Nakashima et al. 2008). This paper shows the evaluation of scale effect on shear strength based on C1-2 experiment. SHAKE TABLE EXPERIMENT ON A FULL-SCALE RC COLUMN WITH TERMINATION OF MAIN BARS Model and Experimental Setup Photo 1 shows the experimental setup of C1-2 column on E-Defense. The model setup was the same with C1-1 and C1-5 excitations (Kawashima et al. 2009A and Kawashima et al. 2009B). A catch frame was set under the lateral beam of C1-2 so that it could prevent collapse of C1-2 when it was seriously damaged. Tributary mass to the column by two decks, four weights, two fixed bearings including four side sliders on the column and two movable bearings including four side sliders on two end supports, and thirty two load cells was 307 t and 215t in the longitudinal and transverse directions, respectively. Because two end supports carried a part of the lateral force in the transverse direction, the tributary mass in this direction was smaller than that in the longitudinal direction.