Experimental Seismic Evaluation of Reinforced Concrete Columns with Plain Bars under Cyclic Lateral Loading

1. Concrete buildings reinforced by plain (smooth) bars are one of the special types of old reinforced concrete buildings. They were generally built before 1970’s; mostly in Europe, Asia and Oceania. Some of the older cases of such buildings were probably designed just for gravity loads and do not have special seismic detailing for structural members (e.g. beams, columns, joints, etc.) because the old codes did not include special seismic provisions at that time. After recent earthquakes, seismic vulnerability of old reinforced concrete buildings has became more highlighted and subsequently, demand for seismic evaluation and rehabilitation of such buildings have experienced a remarkable growth in recent decade. The first stage for proposing a rehabilitation strategy for an existing building is the seismic vulnerability assessment. In this stage a structural engineer tries to understand the behavior of the structure under seismic excitations. This behavior is controlled mostly by the weak links of the structure. Recognition of the structurally weak links under earthquake ground motions helps the engineer to define a proper rehabilitation strategy for the improvement of structural seismic performance. To achieve a better understanding of structural seismic behavior, it is essential to perform a proper analysis considering the nonlinear behavior of structural members. As columns are generally the most important structural members of a framed structure, understanding their realistic seismic behavior is very helpful in estimating structural deformations, forces and energy dissipation capacities. Furthermore, in most of old framed building structures, columns play a key role in the final behavior because of strong beam-weak column conditions. 2. Methodology2.1. Experimental study: Four cantilever half-scale column specimens were tested in this study under monotonic and cyclic loading protocols. Specimens had square sections and were conneced to a strong foundation. Concrete casting was done in two steps in vertical position. All of the specimens’ reinforcements consisted of four longitudinal plain (smooth) bars of 12 mm diameter. To investigate the effect of various types of old splice detailing practices, three types of specimens with various longitudinal reinforcement splicing were considered: specimens Without Overlap Splice (WOS specimens) which were the representative for column detailing below the floor level, Specimen with Straight Overlap Splice (SOS specimen) of 40 times of the bar diameter length representing old American practice for column detailing over the floor level and finally specimen with Hooked Overlap Splice (HOS specimen) of 20 times of the bar diameter length which was introducing old European practice for column detailing over the floor level. WOS specimens were built twice, one for monotonic test (WOS-M) and one for cyclic test (WOS-C). Transverse reinforcement details were similar in all of the specimens; i.e. plain bars with the diameter of 8 mm in 200 mm spacing intervals. The distance was adopted to be equal to the column effective depth. Full details of specimens are presented in Fig. 1. All of the specimens were tested under constant axial load equal to 15% of concrete gross section axial capacity. Loading history protocol according to ACI T1.1 [1] recommendations was applied. The history was displacement (drift) control with preliminary target drifts. Target cylindrical compressive strength for concrete specimens at the age of 28 days was 22.5 MPa.3. Results and discussion3.1. Test observations: Similar crack patterns were observed in all the tested specimens up to drift ratio of 2.20%. Three flexural cracks were formed in three elevations; i.e. base elevation (base-crack), approximately half-depth from base elevation (h/2-cracks) and approximately one-depth from the base elevation (h -cracks). In WOS-M and WOS-C specimens base-cracks were the most active cracks with larger openings; meanwhile h/2-cracks were more active than h cracks and had an inclined trend like flexure-shear cracks. In SOS-C and HOS-C specimens (-C represents cyclic loading protocol), base-cracks were similarly the most active cracks same as WOS specimens; however, h-cracks were more active than h/2-cracks and presented an inclined trend. It was visually obvious that base-cracks play a key role in total deformations of specimens and dictate a rocking like mode for all of the specimens; even in small drift ratios.3.2. Force-drift response of specimens: The response curve of WOS-M specimen was approximately linear up to drift ratio of 0.63% and then turned to nonlinear phase. WOS-C specimen had an origin-oriented hysteresis response. In the unloading phase, after fast degradation of shear strength, the curve was tending to the origin. In other words, high pinching and low residual drifts were the main characteristics of hysteresis response. SOS-C specimen same as WOS-C specimen demonstrated an origin oriented hysteresis response but in different flag-shaped appearances. Similarly, reversal curves tend to zero drifts with a negative slope in the second step of unloading phase after fast shear strength degradation and finally, near the zero drifts, the curves were tending to the origin. Consequently, high pinching and low residual drifts existed same as WOS-C specimen. The hysteresis response curves of HOS-C specimen are very similar to SOS-C specimen but with more pinching effects because of lower initial strength degradation at the first steps of unloading phase.3.3. General mode of behavior of specimens: The main characteristic of behavioral mode of specimens was rocking movement around a point near the toe. The distance between the rocking rotation point and toe (contact depth) had an approximately constant value of 50 mm after 1% drift ratio (based on experimental observations). Contact depth was spalled at drift ratio of 2.20% and then it was crushed at drift ratio of 3.5%. After the crushing, rocking continued around a point near the previous point. One may conclude here that the main governing phenomenon for specimen deformations is rocking action which is restrained by two rows of plain bars at both sides.3.4. Restrained-rocking model: As mentioned before, restrained rocking mechanism was the governing behavior mode of specimens, such as rocking of a rigid body block around its toe. However, the block was not rigid itself in this case and it was deforming in flexure and shear modes. The formation of some cracks (1 or 2) at the bottom height of specimens was evident. The cracks were active until the end of test period and contributed in total deformation of the specimens. To achieve a better understanding of load-displacement hysteresis response of the specimens, it was initially assumed that the specimens did not have longitudinal reinforcement continuity between the column and foundation. In other words, a concrete block was assumed under an axial load. The block was loaded horizontally as well. Roh and Reinhorn [2] studied the behavior of such an element and called it rocking element. Next, it is assumed that plain bars are added to the rocking element. It was obvious that before the total bond deterioration of dowel bars, yielding strength of bars at tension and compression was added to the rocking strength; thereafter, bond resistance was added. Hysteresis response curve was located approximately between two linear upper and lower bound lines which were nearly parallel to the apparent negative stiffness part of the rocking curve. The upper bound line was parallel but apart from the rocking curve negative stiffness line for a constant added strength above it, while the lower bound line was similarly parallel the rocking curve but apart for another constant value below. The upper bound line passed through maximum displacement points of hysteresis curve for cycles with peak drift ratio more than 2.2%. In other words, it had a common part with the negative stiffness slope line of backbone curve which was initiated at Negative Stiffness Initiation (NSI) point that coincided with spalling occurrence onset. The lower bound line passed through points of slope changing in unloading branch for cycles with peak drift ratio more than 2.2%. Better understanding can be achieved if we consider two upper and lower bound curves instead of upper and lower bound lines. The upper bound added strength before the NSI point can be referred to the yield properties of plain bars. After NSI point, it can be referred to the bond properties of dowel bars, especially in SOS-C and HOS-C specimens; while in WOS-C specimen, it can be referred to a combination of yield and bond properties. It’s worth noting that shear strength dropped between upper bound and lower bound curves (at the first step of unloading stage), passed through the rocking action curve. It may be assumed that the strength difference between upper bound and rocking curve arises from full tension yielding release or full tension bond deterioration. On the other hand, the distance between rocking curve and lower bound curve arises from compression yielding development or compression bonding development in the opposite direction of concrete block movement. In fact, the lower bound curve is the rocking action curve minus an opposite yield or bond resistance. The described sequences are demonstrated in Fig. 2. The main concept here, is the fact that the shear strength difference between upper and lower bounds is the summation of two strength values; one is the released strength from upper bound to rocking curve (f1 or f3) and the other one is developed from rocking curve to the lower bound (f2 or f4); totally form an origin oriented flag-shaped hysteresis response curve.4. According to the monotonic and cyclic tests that were performed on four half-scale cantilever column specimens under low axial loads, the following main conclusions are developed;•Damage pattern of concrete columns with plain bars consists of limited numbers of flexural cracks that are formed at the bottom of the column specimen, in which the base crack (crack at the interface between column and foundation) has large opening. •Hysteresis force-drift curves of plain bar specimens are origin oriented in the specimen without overlap splice and are flag-shaped in specimens with overlap splices. High pinching effects and low residual displacements are the main characteristics of all the hysteresis curves. •General mode of behavior of all specimens seems to be restrained rocking action, independent from types of splice detailing. This is proved through experimental calculated rotations which are derived from LVDT readings.•A simple theory for the explanation of hysteresis force-displacement response of specimens is proposed. The theory assumes a concrete block rocking element which is restrained with plain bars at both ends.