Abstract
Using precast segmental concrete columns in structures improves the construction efficiency and site safety, leads to better construction quality, and reduces the construction cost, site disruption and environmental impact. The performance of segmental columns to resist earthquake and impact loads is not well studied yet. As a structure might be subjected to such loads during its service life, understanding its resistance capacities is essential for structural safety. This article reports the findings of our recent studies on the response of precast segmental columns with unbonded prestress tendons. Pendulum impact test and uniaxial cyclic test were conducted on quarter-scale segmental columns. The seismic performance and impact-resistant capacity were evaluated experimentally and compared with a reference conventional monolithic column. Test results showed that under cyclic loading, segmental columns exhibited better deformation ability and smaller residual drift; under impact loading, segmental columns also showed better self-centring capacity and less residual displacement. By introducing concrete shear key, the shear resistance at segmental joint could be improved; however, shear key would also result in more concrete segment damage owing to stress concentration.
Introduction
During the design life, a structure may be subjected to various hazards, such as earthquake load and impact load. Since the amplitudes, frequency and duration of these loads differ from each other significantly, their effects on structures are therefore also very different. For example, earthquake loading induces large structural displacement; therefore, structural damage is governed by displacement-related parameters such as inter-storey drift and ductility ratio, whereas under impact loading, structural damages could be localized and shear capacity governs the failure that is associated with the brittle failure mode. Therefore, it is important to study the fundamental response and damage mechanisms of structures under earthquake and impact loads in order to develop a structure design that is able to enhance both the earthquake and impact resistance capacities.
Segmental columns have been widely used in constructions of buildings and bridges because of their many advantages, such as increased construction efficiency, improved quality control, and minimized traffic disruption. Numerous constructions have successfully employed precast segmental concrete columns (Ou, 2007; Sprinkel, 1985). However, most constructions used precast segmental columns located in areas of low seismicity due to limited knowledge about the seismic behaviour of segmental column. Recently, intensive researches have been made to understand and improve the performance of segmental column under seismic loading for their application in high seismic regions (Dawood et al., 2011; ElGawady and Dawood, 2012; Kim et al., 2010; Mashal et al., 2013; Ou, 2007; Ou et al., 2007; Shim et al., 2008). Different laboratory tests including cyclic tests and shake table tests and numerical simulations have been performed on segmental column with different designs. It is found that under seismic loading, segmental column exhibits unique features, which undergoes much larger lateral displacement than monolithic column (Ou et al., 2007). Since plastic hinge is normally found to form around the base of segmental columns when subjected to seismic loading, failure of segmental columns is mostly flexural dominated, while shear dominated failure such as shear slip between two adjacent segments is not commonly observed (Kim et al., 2010; Shim et al., 2008). Some recent studies by Sideris et al. (2014a, 2014b) introduced segmental column with hybrid sliding-rocking joints. More energy was reported to be consumed through frictions between different segments. However, amplified moment would be experienced by the column because of P-delta effect. Study on response of segmental columns against impact loading is still limited (Chung et al., 2014; Zhang et al., 2016a, 2016b). Since the loading amplitude, duration, and action location on the column from impact load are all very different from that of seismic loading, the response and failure mode of a segmental column under impact loading is likely to be different from that under seismic loading. And the influences of column design specification and feature could also be different which requires further study. Therefore, the multi-hazard resistance capacity of a segmental column under seismic loading and lateral impact loading has not been investigated properly yet.
The influences of different column design specifications and features such as the number of segments/joints, prestress level, and shear key have also been investigated in previous seismic studies. It was found that the number of segmental joints has very limited effect on the ductile behaviour when the location of joint is far from column base where plastic hinge is formed (Shim et al., 2008). Therefore, focus is recommended for the base connection between the segment and column footing in seismic analysis. For the effect of prestress level, Nikbakht et al. (2015) derived analytical solution and concluded that increasing prestress level would lead to higher column stiffness, increased column strength and improved energy dissipation capability. Wang et al. (2014) also found that under seismic loading, increased prestress level would help to reduce column residual displacement.
Concrete shear key is often used for prefabricated structures in design practice. Commonly used geometries include prism, circular truncated cone, and hemispherical block (Mashal et al., 2013; Wang et al., 2008), which provide in-plane interlocking mechanism and improve shear resistance between adjacent structural members. Interlocking blocks with more complex geometries such as plate-like assembles of tetrahedrons or osteomorphic elements have also been developed (Ali et al., 2013; Molotnikov et al., 2007; Ramamurthy and Kunhanandan Nambiar, 2004). Nevertheless, the effect of shear key is normally ignored for segmental column in previous seismic analysis because under earthquake excitation, the magnitude of lateral load is relatively small. The primary response mode of segmental column is flexural bending. Friction force between contacting segments is generally sufficient to transfer shear force. In engineering practice, small concrete shear keys are used only for easy alignment of segments during erection (Wang et al., 2008). When the segmental column is subjected to lateral impact or blast loading, friction between segments may not necessarily always be sufficient to resist the large shear force. The failure of segmental column could be resulted from segmental joint which consequentially leads to shear failure (Hulton et al., 2013). Therefore, an effective form of shear key would be necessary to improve the impact resistance capacity of segmental column.
This article presents experimental study to investigate the multi-hazard (earthquake and impact loading) performance of segmental columns. Segmental columns with precast reinforced concrete (RC) segments and unbonded prestress tendon were designed and tested under uniaxial cyclic loading and impact loading to examine their responses and failure modes, respectively. The performance was also compared with conventional monolithic RC column to evaluate their advantages and disadvantages. The influence of concrete segment/joint number and the effect of tower-shaped concrete shear key to the performance of segmental column were experimentally studied.
Column design
A total of 10 quarter-scale columns with 5 in each group for cyclic and impact tests were designed and fabricated. They included one monolithic RC column, two plain segmental columns (S5N and S7N), and two segmental columns with tower-shaped shear key (S5KT and S7KT) in each group as shown in Figure 1. The columns were 800 mm tall with 100 mm2 × 100 mm2cross-section. The segmental column S5N included five RC segments of 160 mm tall, and column S7N had seven pieces of 115 mm tall RC segments. They were designed to study the influence of segment numbers. S5KT and S7KT have similar RC segments but with tower-shaped concrete shear key as shown in Figure 1(b). These two columns were to examine the effect of concrete shear key; 6-mm-diameter deformed bars were used as longitudinal reinforcements, and 4 mm plain bars at 50 mm spacing were used as tie. The longitudinal rebar did not extend through adjacent segments in the two segmental columns. Grade 25 self-compacting concrete with super-plaster was used for casting concrete columns. Maximum aggregate size was 10 mm. The averaged compressive strength of concrete was about 34 MPa and the flexural tensile strength was 5 MPa. The yield strengths of longitudinal rebar and tie were 500 and 300 MPa, and the Young’s modulus was 200 GPa. A 9.3 mm seven-wire prestress tendon was utilized to tighten the segmental columns with a prestress load of 30 kN. The proof stress and Young’s modulus of the prestress tendon were 1860 MPa and 195 GPa, respectively. A 140-mm-deep and 400 mm × 400 mm cross-sectional area footing was built for each column to bolt them onto the strong floor. The monolithic column was cast with a 400 mm × 400 mm × 50 mm flanges at its top to support the added mass. A concrete block of 400 mm long by 400 mm wide by 450 mm high in dimension together with another five pieces of 23 kg steel plates was placed on top of the columns. A total weight of about 288 kg was placed on top of the column. The weight of the column itself is about 18 kg, which resulted in the ratio of supported weight over column self-weight to about 16. Due to the restrain of the testing system, there was no space for additional mass to be placed on top of the columns.
Details of test specimens: (a) details of columns and (b) illustration of shear key.
Cyclic behaviour
Test setup
Figure 2(a) illustrates the setup of the cyclic test. To be different from previous uniaxial cyclic tests (Billington and Yoon, 2004; Shim et al., 2011), a specially designed ‘yield frame’ was introduced to restrain the rotational degree of freedom at the column top so as to study the behaviour of column with double curvature. An actuator was installed which pushed and pulled the top of the column slowly to apply the quasi-static cyclic loading. A load cell and a laser linear voltage displacement transducer (LVDT) were used to measure the applied lateral load and to control the lateral displacement. The applied lateral drift increased gradually from 0.25% to 7% which was repeated three cycles for each drift ratio (Figure 2(b)).
Setup of cyclic test: (a) setup of cyclic tests and (b) loading sequence.
Result
The above-mentioned two plain segmental columns S5N and S7N and the monolithic column were tested. Column damage patterns and hysteric curves are presented to show the responses of different columns. Segmental column S7KT was then tested to evaluate the effect of shear key. Column S5KT was not included in the cyclic test because the influences of shear key on cyclic behaviours of S5KT segmental columns are similar.
Column response and damage
Figure 3 shows the damage of the tested columns under cyclic loading. Flexural horizontal cracks were initiated at the bottom of the monolithic column and developed upwards. Because of the restrain at the top of the column, concrete crack and crushing were also formed near the top flange. The lateral strength of the monolithic column dropped dramatically at 5% drift indicating failure as a result of the fracture of longitudinal reinforcements. The response of the segmental columns was very different. With the gradually increased drift level, instead of forming concrete tensile cracks, segmental joints between the bottom two segments gradually opened. For segmental column S5N, when the drift ratio increased to 5%, a 2.5 mm joint opening between the bottom two segments was formed, but no concrete tensile crack was found on the column. As the applied drift gradually increased, concrete crushing appeared and extended on the base segment due to the large compressive stress from the flexural bending as well as the axial forces from added weight and prestress force. At 7% drift, more damage was observed on the base segment. The response and damage of column S7N was similar to that of S5N. But more severe damage was found on the base segment (Figure 3(c)). Relative displacement between the first and second segments was observed on the segmental column S7N indicating insufficient shear resistance. When introducing shear key to the concrete segments, the segmental shear resistance was effectively improved. As evidenced in column S7KT, no relative displacement was found at any segmental joint. Nevertheless, because of the introduced shear key, stress concentration was resulted which led to more severe damage to concrete segments (Figure 3(d)).
Damage of columns under cyclic loading: (a) monolithic, (b) S5N, (c) S7N, and (d) S7K.
Hysteric curve
Figure 4 shows the hysteretic curves of the tested columns. For the monolithic column, yielding occurred as drift ratio 0.75%. At 2% drift, the lateral strength of the column reached its maximum (average 5.5 kN), after which it quickly reduced. Large residual displacements can be found on the monolithic column because of the plastic deformation of the reinforcements. At the ultimate drift of 5%, the averaged lateral capacity of the column reduced to about 2.5 kN. The hysteretic curves of the segmental columns showed typical flag-shaped pattern before 5% drift. Comparing with the monolithic column, relatively small enclosed areas were resulted by the segmental columns. Because of the existence of the prestress tendon, the laterally deformed column was always pulled back to its original position after the removal of the applied lateral load. As a result, much smaller residual displacements were resulted on the segmental columns. For plain segmental column S5N, an averaged maximum lateral strength reached about 5 kN at 5% drift. The lateral loading capacity of the column gradually reduced as a combined effect of the column degradation and loss of prestress due to damage of concrete segments. At the ultimate 7% drift, the lateral capacity ended at around 3 kN with severe compressive damage to the base segment. The behaviour of S7N was very similar to that of S5N. At 2% drift, the average strength of S7N was 4.7 kN which is marginally larger than that of S5N (4.4 kN). But as S7N experienced more segment damage, its strength quickly degraded to 4.6 kN at 6% drift while that for S5N was 4.9 kN. Shear key showed limited influence to column lateral strength when drift ratio was small, but as segments experienced more concrete damage under increased drift, column lateral strength quickly reduced. For instance, an averaged lateral strength of 1.3 kN was measured on column S7KT at 0.2% drift which is very close to that on column S7N (1.2 kN); but as drift ratio increased to 6%, the strength of S7KT reduced to only 3.6 kN while that of S7N was 4.6 kN. Through comparison, it can be found that the tested segmental columns showed much better deformation capability as compared with the monolithic column. Shear key leads to segment damage when drift ratio is large, which consequentially degrades column lateral strength.
Hysteric curves of the tested columns.
Analysis and discussion
As observed above, the segmental column shows good self-centring capacity, but in the meanwhile, it develops less plastic deformation indicating less energy dissipation capacity. The residual displacements and the cumulative energy for the tested columns are summarized and used to analyse the performance of the tested columns.
Residual displacement
Figure 5 summarizes the residual displacements at the top of the columns after each applied drift ratio. As can be found, much smaller residual displacements were resulted on the segmental columns as compared with the conventional monolithic column. At 2% applied lateral drift, a residual displacement of 2.5 mm was measured on the monolithic column, while those on the segmental columns were less than 1 mm. The residual displacements on the monolithic column increased rapidly as the applied drift increased, but those on the segmental columns grew more gradually. At the 5% drift, a residual displacement of 23 mm was left on the monolithic column, while only 2 and 1.5 mm residual displacements were found for the columns S5N and S7N. Because of the excessive damage to concrete segments with shear key, the segmental column quickly degraded after the applied drift is over 5%. About 6 mm residual displacement (5% applied drift) was found on S7KT, while that on plain segmental column S7N without shear key was less than 2.5 mm. And at 6% lateral drift, the residual displacement on S7KT increased to about 21 mm which was much larger than that on column S7N (about 2 mm) because of the significant concrete crushing damage owing to stress concentrations associated with the shear keys.
Residual displacement versus drift ratio.
Cumulative energy
Similar level of cumulative energies was dissipated by the monolithic column (20 J) and the two plain segmental columns (14 J for S5N and 16 J for S7N) when the applied drift ratio was less than 0.5% because both concrete and reinforcement were mostly in elastic deformation. As the applied drift increased, more cumulative energy was consumed by the monolithic column than the segmental column since more tensile cracks were developed and the plastic deformations of reinforcements on the monolithic column were observed. For instance, at 2% drift, 214 J energy was dissipated by the monolithic column, while those by the segmental columns were only 113 J (S5N) and 162 J (S7N). And at 5% drift, about 1250 J energy was consumed by the monolithic column, while only 450 and 530 J were dissipated by S5N and S7N, respectively. When shear key was introduced, more concrete damage was resulted which led to more cumulative energy consumed. As shown when 5% drift was applied, about 670 J energy was dissipated by column S7KT which was 26% more than plain segmental column S7N. Nevertheless, the amount of energy dissipated was still much less than the monolithic column (1250 J). Without energy dissipation devices, segmental column consumes imposed seismic energy only through segmental friction, rocking and compressive damage of concrete. Therefore, energy dissipation device should be installed to improve the segmental column energy dissipation capacity (Figure 6).
Cumulative energy versus drift ratio.
Impact response
Test setup
To examine the impact response of the segmental column, a pendulum impact system as shown in Figure 7 was utilized. A 300 kg weight steel impactor was lifted to designed heights and then released to impact on the centre of the column. Each column was subjected to multiple impacts with the impactor released at gradually increased heights, which corresponded to 0.2 m/s (Impact 01), 0.6 m/s (Impact 02), 1.4 m/s (Impact 03), 2.7 m/s (Impact 04), and 3.6 m/s (Impact 05) impact velocities. It was designed to assess column responses with elastic response, minor damage, moderate damage, and major damage. A load cell was mounted on the front tip of the impactor to monitor the applied impact load onto the columns. Three LVDTs were placed at column top, mid-height, and near column base behind each tested columns. A high-speed camera was installed to monitor the deformation-to-failure process of the columns. A series of tracking dots were attached to the middle of each concrete segment (uniformly along the monolithic column) to assist recording the response of the tested specimens.
Illustration of pendulum testing system.
Result
Column response and damages
Figure 8 shows the responses of the tested columns recorded by high-speed camera when ultimate column failure was resulted in Impact 04. For the monolithic column, upon the impact, flexural cracks at mid-span were developed and grew into major cracks which quickly extended across the column. Concrete crush occurred on the other side of the column. The column collapsed with a flexural bending failure and also a thorough diagonal shear crack failure near the base of the column. The segmental column responded very differently. For segmental column S5N, large joint opening was formed as a result of significant flexural bending under the substantial impact loading. Concrete in the compressive surface of this joint was crushed because of the flexural compression and also the direct impact force on the segment. It can also be observed in Figure 8(b) that an apparent relative displacement was formed between the third and the fourth segments when the impactor struck on the column. The response of column S7N was very similar to column S5N. Severe concrete crushing failure occurred on the central segments due to large flexural bending–induced compressive failure. Compressive damage on the base segment was also found as the column rotated against its base. The segmental columns with shear keys exhibited similar response under impact loading but with more severe segment damage. As the impactor struck onto segmental column (S5KT), a large flexural bending was formed and concrete segment damages were observed in the central segment. As the response of the column transferred from local vibration mode into global mode, joint opening between the two central segments gradually closed, but a large opening was formed at the joint between the two upper segments. Because the existence of shear key weakened the section bending capacity of the concrete segments, severe damage was resulted to the top segment during column flexural deformation, which consequentially led to the total collapse of the column. The response of column S7KT was slightly different from S5KT. With more segments, column S7KT was more flexible. A significant joint opening between the two central segments was formed which led to more damage as a result of excessive flexural compression. Concrete crushes in the two bottom segments were also developed as the column bent. The localized flexural response mode gradually changed to global flexural response mode as the inertia effect on the top of the column diminished. The top of the column swayed sideway. Due to excessive lateral displacement and severe damage in concrete segments around column central joint and base joint, the column lost stability at these two joints and eventually collapsed under the axial force from the added mass.
High-speed camera images of columns in Impact 04 (30°): (a) monolithic column, (b) S5N, (c) S7N, (d) S5KT, and (e) S7KT.
Comparing the damage and failure of the tested columns, it can be found that the lateral impact resulted in large flexural deformation at the centre of the columns, which was very different from that under cyclic loading. Flexural tensile cracks were formed in the concrete and fracture of longitudinal reinforcement was resulted on the monolithic column. In addition, under the significant lateral impact force, a diagonal shear crack was also formed near the base of the monolithic column. In comparison, with a series of segmental joints, the segmental columns were very flexible. Under impact loading, segmental joints opened which avoided concrete tensile cracks and eliminated plastic deformation of longitudinal reinforcement. The presence of prestress tendon pulled the deformed column back to its original position, showing excellent self-centring capacity. Under impact loading, the damage of the segmental columns was mainly the compressive damage of concrete segments due to excessive flexural bending. Relative displacement between segments could be developed when the segmental column is subjected to large lateral impact. The existence of concrete shear key helped to reduce this relative displacement. However, similar to the case under cyclic loading, the introduced concrete shear key also lead to stress concentration at segmental joint which consequentially result in more severe concrete segment damages.
Impact load and displacement time histories
Figures 9 and 10 show the impact load and column central deflection time histories. As shown, because of the large stiffness and integrity of the monolithic column, the load increased sharply upon the impactor struck onto the column, which quickly reduced to zero. The load time histories on the segmental columns differed significantly. Because of the interaction between the impactor and the flexible segmental columns, after the initial peak load resulted from the impact, the load dropped but then formed a second peak load on the loading time history. Comparing with monolithic column, the peak impact loads on the segmental columns were always smaller but associated with longer duration. For instance, when the impactor struck at 0.2 m/s, a peak load of 7 kN lasting 50 ms was measured on S5N, while 12 kN peak load with duration of 24 ms was measured on the monolithic column.
Load time histories under repeated impacts.
Central displacement time histories under repeated impacts.
The different responses of the segmental columns and the monolithic column can also be observed through central deflection time histories. As shown in Figure 10, for the segmental columns, a large central deflection was resulted due to the impact. But because of the restoring force from the prestress tendon, the segmental column quickly rebounded. As the local deformation mode transferred to the global mode, a second peak deflection was resulted and the column continued to vibrate around its original position until it came to rest. The segmental joints opened in response to the column flexural deformation. In comparison, for the monolithic column, a large central deflection was formed under the lateral impact, which rebounded but because of the plastic deformation of the longitudinal reinforcement, much larger residual displacements were formed on the monolithic column.
Analysis and discussion
To evaluate the performance of the tested columns under impact loading, the impact load, column maximum and residual deflections, and dissipated energy are used for analysis and comparison. Relative displacement between segments is utilized to discuss the effect of shear key.
Impact load
The impact loads are summarized in Table 1. Despite the impactor was of the same weight and released at the same height in each group of test, the measured impact loads varied between the segmental columns and the monolithic column because the action of the impactor was coupled with the response of the columns. The peak impact forces are plotted against the integrated impulses in Figure 11. As shown, when the impactor struck with the same impact velocity, higher peak forces were always observed on the monolithic column than those on the segmental columns. But similar or even larger impulses were always found on the segmental columns because of longer impact duration. The peak forces on column S5N were slightly higher than that on column S7N because the stiffness of the former column was relatively higher than that of S7N with more joints. Shear key increased the stiffness of the undamaged column. As evidenced in Impact 02, larger peak load (14 kN) and impulse (339 kN ms) were recorded on column S5KT, while those on column S5N without shear key were 13.4 kN and 301 kN ms. As discussed above, the existence of shear key caused more concrete damage which degraded column stiffness; in Impacts 03 and 04, lower peak impact forces were found on column S5KT than those on column S5N. But as the impact duration increased, higher impulses were still measured on column S5KT than on S5N. For instance, in Impact 03, about 10% smaller peak impact load was measured on column S5KT than that measured on S5N, but the impulse on S5K was 16% higher. Despite the existence of shear key in S7KT which was supposed to increase the stiffness of the column, even in Impact 01, this column suffered minor damage to some segments, which led to the degraded column stiffness. A peak impact load of 7.9 kN was resulted on S7KT, which was 3% lower than that on S7N. Smaller peak impact forces were always resulted on column S7KT as more damages accumulated in the subsequent impacts. Nevertheless, because the interaction between the impactor and the column was longer on column S7KT, larger impulses were resulted. In Impact 02, 364 kN ms impulse was measured on column S7KT indicating a 13% larger impulse than that on S7N.
Summary of impact loads.
Recorded impulses and peak impact loads.
Maximum and residual deflections
To evaluate the response of the segmental columns and compare their performance with monolithic column, the maximum column central deflections are summarized and plotted against the impulse. As shown in Figure 12, in general, the maximum deflection increases with the applied impulse for both monolithic and segmental columns. When the impulse was low, a larger maximum deflection was found on the monolithic column than on the segmental columns. For instance, when the impulse was about 134 kN ms, a maximum deflection of 5.2 mm was measured on the monolithic column, while under a similar impulse, the maximum deflection of the segmental column S5N was 3.9 mm. This was mainly because the prestress prevented joint opening on the segmental columns when the impact load was relatively small, which minimized the column deflection. As the impulse increased, joint openings occurred. Larger maximum central deflections were measured on segmental columns than on the monolithic column. For example, a maximum central deflection of 14.6 mm was measured on the monolithic column when subjected to 478 kN ms impulse. In comparison, the maximum deflections of 32.8 and 24.3 mm were recorded on columns S5N and S7N with respect to about 560 kN ms impulse. The influence of shear key to column maximum deflection varied when the columns were subjected to different levels of impact. Under small impact loading in Impact 01, the columns with shear key (S5K and S7K) both exhibited smaller central deflections comparing with those without shear key. This is because of the increased stiffness of column due to shear key. A maximum central deflection of over 5 mm was measured on column S7N while that on column S7KT was 4.4 mm (–12%); and the maximum deflection on column S5K was only 2.9 mm, which was 26% smaller than that on column S5N (3.9 mm). Despite smaller peak deflection of the column S7KT in Impact 01, concrete crushes were resulted due to stress concentration near shear keys. As a result, larger central deflections were always found on this column in the subsequent impact tests comparing with the one without shear key (S7N). The column also responded relatively slower as its stiffness degraded (Figure 10). Similarly, in Impact 02, concrete damages in certain segments of column S5K were also generated, which resulted in degraded column stiffness and larger central deflection consequentially (Table 2).
Summaries of the maximum and residual deflections at column centre.
Summary of the maximum and residual deflections.
The segmental columns show more advantages over monolithic column in terms of smaller residual deflections. As shown (Figure 12), for all the impacts, smaller residual deflections were always measured on the segmental columns. For instance, under low-level impacts (Impacts 01 and 02), residual deflections of 1.5 and 2.9 mm were measured on the monolithic column. In comparison, less than 1 mm residual deflections (merely 1/3 of that on monolithic column) were measured on the segmental columns. As the impulse increased to 478 kN ms (Impact03), the monolithic column experienced severe flexural and shear damage with a residual deflection of about 11 mm. In contrast, under even larger impulse (about 560 kN ms), a residual deflection of 2.6 mm was found on the segmental column S5N. It can therefore be concluded that under lateral impact, segmental column shows much better self-centring capacity than conventional monolithic column. This was because no plastic deformation was formed in its longitudinal reinforcement and the prestress tendon inside the segmental column pulled the swayed column back. Similar level of residual deflections was found on the two segmental columns S5N and S7N under low-level impact loading because only minimum joint opening and concrete crushing damage on either column were found. Since the footing of column S5N was damaged in Impact 03, the footing could not provide the same restraint to the column. As a result, larger residual deflection of 28.0 mm was measured on column S5N in Impact 04. Under the similar magnitudes of impulse, the residual deflection of column S7N was 7.3 mm. Smaller residual deflections were also found in segmental columns with shear key before column was damaged. For instance, after Impact 01, the central residual deflection was only 0.3 mm on column S5KT, while that on column S5N was 0.5 mm. It can also be seen clearly that after the forced vibration, the column S5KT vibrated faster with smaller amplitude (Figure 10). Again, this was because of the higher stiffness of S5KT due to the concrete shear key. Despite the shear key improved column stiffness, more concrete damages were developed due to stress concentration especially when the column is subjected to large-scale impact. As a result, larger residual deflections were resulted on the segmental columns with shear key. For example, a residual central deflection of 7.9 mm was measured on column S7KT in Impact 03. In comparison, that of column S7N was only 1.1 mm.
Relative displacement
Despite shear key did not contribute significantly to reducing the column maximum and residual deflections, the existence of the shear key could help to minimize relative displacement between adjacent segments. To evaluate the effect of shear key, relative displacements between adjacent segments on the tested segmental columns were measured and plotted against the corresponding impulse in Figure 13. As shown, for the two columns with five segments, no slip between segments was observed under small impact loads in Impacts 01 and 02. When column S5N was subject to about 560 kN ms impulse in Impact 03, a relative displacement of about 4 mm was resulted between the two central segments. In contrast, only 1.2 mm relative displacement (70% less) was found on column S5KT with shear key. The relative displacement on column S5N further enlarged to about 8.7 mm in Impact 04, while that on the segmental column with shear key was only 2.2 mm (75% less). For segmental columns composed of more segments, the effect of shear key in controlling the relative displacement was more significant. A relative displacement of about 2 mm was found between the two central segments when column S7N was tested in Impact 03. This was increased to about 2.3 mm in the subsequent impact. No slip was observed between the central segments on the column with shear key (S7KT). Nevertheless, because of stress concentration on concrete near the shear key, concrete crushing damages were observed and developed on the segments near column base. As a result, a small relative displacement (about 0.7 mm) occurred between Segments 02 and 03 after Impact 01. The segmental drift grew to about 1.3 and 1.5 mm under the subsequent impacts. As damage extended into the base segment (Segment 01) when the column was under 624 kN ms impulse (Impact 04), very small relative displacement (about 0.5 mm) between Segments 01 and 02 was also found. Overall, it can therefore be easily concluded that the shear key was effective in reducing relative movement between segments especially when the column was subjected to large impact loading. However, the drawback of the current shear key design is also obvious that it results in stress concentration in the segment.
Relative displacements versus impulses.
Energy dissipation
Energy absorption capacity is an important parameter. The impact load recorded by the load cell was integrated against the column centre deflection to derive the total energy dissipated by the columns. Figure 14 summaries the derived energy of the tested columns under different impulses. From Figure 14(a), it can be found that the dissipated energy increases with the applied impulse. When the applied impulses were relatively low (Impacts 01 and 02), similar level of energy was dissipated by the monolithic column and the segmental columns. Slightly more energy was consumed by the two segmental columns S7N and S7KT because concrete segment cracks were developed even under small-level strike in Impact 01, while no cracks were found on other columns. When large-scale impacts were applied to the columns, more energy was dissipated by the segmental columns than that of the monolithic column. For instance, in Impact 03 only 130 J was dissipated by the monolithic column, while that by the segmental column S5N was over 230 J. This was because the monolithic column dissipated the impact energy mainly by transforming it into internal energy, that is, plastic deformation of reinforcement and damage of concrete. The impactor forced the column to develop flexural deformation. When it reached a maximum deflection, the column suffered concrete tensile and compressive damage as well as reinforcement plastic deformation leading to damage of the column, and the damage accumulated with noticeable residual displacement. Segmental column was capable of dissipating the impact energy through more versatile means, such as concrete compressive failure, friction between adjacent segments, and localized plastic deformation of prestress tendon. Since the concrete blocks are segmented and not linked with conventional reinforcements, under lateral impact loading, segmental joints opened in lieu of concrete tensile failure and reinforcement plastic deformation. The segmental column would then recover its flexural deformation under the restoring force of the prestress tendon and vibrated which transforms the impact energy to kinetic energy. Since internal energy of the column is correlated to centre deflection when subjected to mid-span impact, the ratio of the total dissipated energy over column maximum central deflection can be utilized to assist the evaluation of column energy dissipation capacity. It can be observed from Figure 14(b) that when the columns were subjected to lateral impacts with considerable column damage and vibration (Impacts 02 and 03), higher ratio of energy over deflection was achieved by the monolithic column than those segmental columns indicating that the segmental column shows less superior energy dissipation capacity through internal energy. Energy dissipation device could be introduced to segmental column so as to improve their energy dissipation capacity.
(a) Energy dissipation and (b) ratio of energy over maximum displacement versus imposed impulse.
Comparing the segmental columns tested, it can be found that in general, the columns with shear key consumed more impact energy as a result of more severe damage to concrete segments. For example, about 15% more energy was dissipated by column S5KT with shear key than S5N in Impact 02. The dissipated energy was also 9% and 22% more on column S5KT than S5N in Impacts 03 and 04. Similarly for segmental column composed of more segments, in Impact 04, about 938 J energy was consumed by column S7KT which was nearly 50% higher than that on column S7N (632 J).
Conclusion
In this study, we performed cyclic test and impact test to investigate the multi-hazard resistance capacity of segmental column under earthquake and impact loading. Through comparison with monolithic column, it was found that under cyclic loading, the tested segmental column showed similar level of lateral strength compared with monolithic column but it had better deformation capacity and better self-centring capacity with much less residual displacement. Under impact loading, the tested segmental column was more flexible than conventional monolithic column, which resulted in lower peak impact force but larger impulse under the same impact velocity. The segmental columns also showed good self-centring capacity under impact loading with minimized residual displacement. Good energy dissipation performance was also found on segmental columns under impact loading but not on cyclic loading. In general, concrete shear key could improve the shear resistance at segmental joints. Relative displacement between adjacent segments could be effectively reduced by introducing shear key for both cyclic loading and impact loading. Nevertheless, the introduced shear key led to stress concentration which weakens section bending capacity. More severe damage could be resulted on segmental columns with shear key.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was financially supported by Australian Research Council.