Seismic strategy of non-seismically designed reinforced concrete columns retrofitted with steel rods

Abstract

This paper presents experimental results of combined cyclic load testing on a reinforced concrete (RC) column that was retrofitted with newly designed steel rods. The steel rods were installed around the column longitudinally and then anchored. The proposed steel rods utilize simple components and installation to enhance both the strength and ductility of RC columns. Cyclic lateral load tests were conducted on three specimens: one unreinforced specimen as reference, one specimen with the entire length of the column retrofitted, and one specimen with only the plastic hinge region of the column retrofitted. All specimens were tested under eccentric constant axial load and incrementally increasing lateral loading cycles with eccentricity. The implementation of steel rods resulted in significant improvement in ductility and an up to 60% increase in ultimate loading capacity.

Introduction

Non-seismically designed reinforced concrete (RC) columns do not possess adequate lateral strength and ductility because of gravity loading, which is considered a primary concern in design. Retrofitting of structures has been carried out if the RC columns have been damaged or if the lateral strength or ductility need to be improved according to recently updated building codes. Many retrofitting techniques for RC column rehabilitation have been investigated and developed. Steel jacketing methods have been widely used to retrofit the structural performance of deficient RC columns. Steel jacketing enhances the strength, ductility, and energy dissipation and also can improve the installation process via smalls increases of the cross-sectional dimensions and absence of cast-in-situ concrete^1–5,Cai et al.⁶; Lai et al.⁷). A significant amount of research on the steel jacketing method to enhance the seismic performance of old RC columns has been carried out^8–11. Among the many studies on the steel jacketing method, few of them have been performed on the behavior of RC columns retrofitted with steel jacketing under combined loading.

Montouri and Piluso¹² fabricated thirteen RC column specimens and conducted eccentric loading tests to predict the moment-curvature behavior of RC columns confined by a steel angle and steel batten,eight specimens were retrofitted with steel battens and steel angles and eccentric loads were applied. The test results showed that the presence of a steel angle and steel batten strengthen the ultimate axial load and provided lateral restraint of longitudinal rebars of RC columns. They also suggested a confinement model for steel angles and steel battens, demonstrating that the ultimate axial load of the specimens and the moment-rotation curve are similar. Garzon et al.¹³ fabricated ten RC column specimens with two lengths using beam-column joints to investigate the behavior of steel-jacketed RC columns when bending and axial load are applied. Four of the specimens were retrofitted with steel tubes and other four were retrofitted with a steel angle. A combined bending and axial load were then applied to each column. Both retrofitted specimens increased their shear resistance and ductility. When the bending moment of steel-angle reinforcement was compared to that of the confinement model of Monturi and Piluso¹², this confinement model exhibited satisfactory behavior at higher axial loads. These studies indicated that steel jacketing on an RC column can enhance strength and ductility. When a seismic load is applied to an irregular building, torsion can be transferred to the columns. This degrades the flexural performance of the columns and can cause shear failure¹⁴. Huang et al.¹⁵ confirmed that the load-carrying capacity of test specimens can be reduced by 39.2% due to torsional load by applying a combined load on the RC columns and that the horizontal eccentricity reduced the load-carrying capacity of the column.

Due to the multi-directional characteristics of earthquakes, the RC columns can be subjected to combined loading, including axial force, bending and a torsional moment during seismic excitations. The behavior of retrofitted RC columns subjected to combined loading should be considered as essential for evaluating the seismic performance. This paper proposes a new seismic retrofitting method for RC columns that installs steel rods near the columns and applies pressure to the steel rods. In conventional steel jacketing method, site work is required to even out the surface of existing structural members or to fill the gap between the steel jacket and the existing member with bonding materials such as mortar. Additionally, due to the heavy weight of steel jackets made to fit the size of the members, extra lifting equipment may be necessary. Unlike the steel jacketing, the steel rod method does not require steel caging or steel encasements for on-site work because of the grout between the steel jackets and the old columns surface and welding for steel jacket binding. The proposed steel rod method involves simple on-site procedures by securing steel rods between beams, solely retrofitting column. This simplicity allows for quick deployment in urgent retrofitting scenarios. A total of three RC column specimens were fabricated and tested under sustained eccentric axial loading and cyclic lateral loading with eccentricity. The objective of the test is to evaluate the behavior of the RC columns retrofitted with steel rods subjected to combined loading. The length of the steel rods for retrofitting was considered as the test parameter. In this paper, the effectiveness of steel rods at providing ductility and strength in RC columns in a combined load is investigated.

Description of the steel rods system

Figure 1 presents the details of proposed steel rods method. The steel rods method involves steel rods with adjustable link, L-shaped angles, and supporting plates. The L-shaped angles are fixed using anchor bolts at the joint of the column and beam or column and slab. The supporting plate bonded to both ends of the steel rod by welding, and the supporting plate and the L-shaped angle are connected by anchor bolts. Finally, by adjusting the adjustable link of the steel rods, a compressive force is introduced to the steel rods as shown in Fig. 1b. An internally threaded sleeve is used for connecting steel rods on adjustable link as shown in Fig. 1c¹⁶. The stress transfers through the interlock of thread. Also, the length of the steel rods can be adjusted through the thread. The sleeve length is determined by the diameter of the steel load. The sleeve length is designed to be twice the diameter of the steel rod. The sleeve can be made of steel pipe with a yield strength of 275 MPa. The steel rods enhance the lateral resistance of the old RC column if it is attached externally to the column, reducing the lateral displacement as shown in Fig. 2. In addition, it is expected that the bending and torsion can be resisted because the steel rod can readily follow the behavior of the column by confining the upper and lower members of the column. With length-adjustable details, steel rods in this retrofit method can accommodate various column lengths and steel rods installed via bolts can be easily replaced after being damaged. The L-shape angle is installed to fix the supporting plate to which the steel rod is welded. If buckling or failure occurs in the steel rods, it can be replaced easily with new steel rod by disassembling the bolts between L-shape angle and supporting plate. Compared to conventional steel jacketing, that wraps and restrains the entire column section and length, the similar level of increase in both maximum load and ductility is expected. However, site work such as drilling a hole in the old RC columns for installation of the retrofitting components is not required and less damage and easy installation are exhibited; thus, steel rod method can be expected to play a role not only in permanent retrofitting but also emergency situations to delay collapse.

Figure 1.

Schematic of steel rods method (a) Details of steel rods method (b) Steel rods with adjustable link (c) Details of adjustable link (d) Thread of adjustable link.

Figure 2.

Mechanism of steel rod retrofit (a) Schematic installation of steel rod (b) Load-resisting mechanism of steel rods system.

Experimental program

Specimen details

Three RC column specimens, including one un-retrofitted RC column specimen (CB) and two RC column specimens retrofitted with steel rods, were fabricated. The retrofitted specimens were distinguished by the retrofitted length, one was a specimen (SFB) that retrofitted the entire length of the column, and the other was a specimen (SPB) that retrofitted only plastic hinge length. A new dissipation device can be used to replace the yielded longitudinal reinforcement to ensure proper stress distribution along the longitudinal reinforcement in the plastic hinge zone^17,18. It is expected that the SPB method can be used as a new dissipation device in the plastic hinge zone. Also, it can be used for the rapid repair damaged columns due to the minimal intervention required.

Figure 3 shows the section geometry and reinforcement details for the specimens. All specimens have a square column with a height of 1.8 m and a cross section of 250 $\times$ 250 mm. An upper beam for loading and a foundation for fixing the specimen were manufactured. The upper beam also had a cross section of 250 mm $\times$ 250 mm and a length of 800 mm, and the foundation had a cross section of 1400 mm $\times$ 1270 mm and a depth of 425 mm. The diameters of the longitudinal rebar and transverse rebar were 22 mm and 10 mm, respectively, for the column, upper beam and foundation. The yield strength of the steel used for the rod and angle was 275 MPa. All three specimens on which the experiment was performed were reinforced with stirrup with a 90 $_{}^{\circ }$ hook to apply non-seismically details. From compressive strength, concrete cylinders with the size of 100 mm diameter and 200 mm height were utilized per ASTM C39¹⁹. To retrofit the steel rods, three pipe-type steel rods were installed on both sides of the column. The L-shape angle is installed to fix the supporting plate to which the steel rod is welded. The L-shape angle was fixed to the foundation with anchor bolts. The pulling strength of anchor bolts was checked with the Hilti PROFIS Engineering Program²⁰ developed based on ACI 318–19²¹. Four anchor bolts with a diameter of 16 mm and an effective length of 100 mm were constructed.

Figure 3.

Details of specimens (mm) (a) Section geometry and reinforcement details of RC specimens (b) L-shaped angle (top view) (c) Supporting plate (top view).

Unlike the steel jacketing method, which enhances performance by increasing confinement around the entire column cross-section, steel rod method involved fixing steel rods longitudinally near the column surface to resist lateral forces along with the columns. It becomes challenging to apply the thickness criteria of the effective confinement, as proposed by AASHTO. Therefore, in this paper, due to the retrofitting of the steel rod along the length of the column, the thickness of the steel rod was determined based on the volumetric ratio proposed by Prestley and Park²⁷. The failure of the transverse reinforcement limits the useful ultimate longitudinal compression strain. For effective volumetric ratios of confining reinforcement, longitudinal compression reinforcement ratios 0.005 to 0.03 were suggested by Priestley and Park²². In this paper, the thickness of the steel rod was determined to be 2.025 mm, and the volumetric ratios for the SFB and SPB specimens were calculated to be 0.017 and 0.009, respectively.

The yielding sequence of structural components follows as retrofitting member, joint and structural members, which is the reasonable plastic development of the structure for seismic design²³. In order to induce the steel rod to yield before the column failure, the shear strength of the steel rods is designed to be smaller than the shear strength of the column. The shear strength of the RC column was calculated according to ACI 318–19 as shown in Eq. (1) to (3), and the shear strength of the steel rod was calculated by the Eq. (4) proposed by Aboutaha et al.²⁴.

$$V_{\mathit{sj}}\le V_n=V_c+V_s$$ 1

$$V_c=\frac{1}{6}\left(1,+,N_u,/,14,A_g\right)\lambda \sqrt{f_{\mathit{ck}}}{b}_{w}d$$ 2

$$V_s=\frac{A_v{f}_{\mathit{yt}}d}{s}$$ 3

$$V_{\mathit{sj}}=A_{\mathit{sj}}(\frac{F_{\mathit{sj}}}{2})(\frac{d_{\mathit{sj}}}{s_{\mathit{sj}}})$$ 4

where, $V_{\mathit{sj}}$ is the shear strength of the steel rod, $V_n$ is the shear strength of the column, $V_c$ is the shear strength of the concrete, $V_s$ is the shear strength of the shear reinforcement, $N_u$ is the axial load, $A_g$ is the gross area of the concrete section, $\mathrm{\lambda }$ is the modification factor to reflect to the reduce mechanical properties of lightweight concrete, $f_{\mathit{ck}}$ is the compressive strength of the concrete, $b_w$ is the width of the column, $\text{d}$ is the effective depth, $A_v$ is the area of the shear reinforcement, $f_{\mathit{yt}}$ is the yield strength of the shear reinforcement, $\text{s}$ is the spacing of the shear reinforcement, $A_{\mathit{sj}}$ is the area of the steel rod, $F_{\mathit{sj}}$ is the yield strength of the steel rod, $d_{\mathit{sj}}$ is the depth of the steel rod, $s_{\mathit{sj}}$ is the spacing of the steel rod. The steel rod had an outer diameter of 30 mm and was 2.025 mm thick based on the volume ratio. The variables included in the Eq. (4) are shown in Fig. 3. Table 1 shows description of all specimens.

Table 1.

Description of specimens.

Member	Cross-section (mm)	Reinforcements		Materials			Retrofitted length (mm)
		Longitudinal	Transverse	Concrete	Steel	Steel rod
CB	250 × 250	4-D22	D10@125	24 MPa	400 MPa	–	–
SFB						275 MPa	Full length (1800)
SPB							Plastic hinge length (500)

A sufficient deformation capacity for RC columns can be achieved by providing adequate confinement at a potential plastic hinge region. The plastic hinge length for determining the retrofitting length was calculated as shown in Eq. (5), which was proposed by Priestley et al.²⁵ and used in Applied Technology Council (ATC)-32²⁶. The plastic hinge length of the specimen was 387.2 mm, and the retrofitting length was conservatively designed to be 500 mm.

$$L_p=0.08L+0.022f_y{d}_b\ge 0.044f_y{d}_b$$ 5

where $L_p$ is the plastic hinge length, L is length of the column, $f_y$ is the expected yield strength, and $d_b$ is the diameter of longitudinal rebar.

Measurements and instrumentation

Figure 4 presents the test setup. A constant column axial load and cyclic lateral load were applied using a hydraulic jack and an actuator. The hydraulic jack for applying the axial load was installed to maintain the gravity direction of the specimen throughout the test. The upper part of the counter beam where the hydraulic jack is installed is not constrained, and the lower part is hinge as shown in Fig. 4a. In a real building, columns are restrained by beams at both the top and bottom, which results in double curvature. However, considering the limited experimental conditions, the test setup and loading method planned to exhibit single curvature at the contra flexure point. A constant axial load of 0.17 $A_g{f\prime }_c$ was applied by hydraulic jack, which was placed on the top of the upper beam to better analyze the flexure and flexure—shear behaviors due to lateral force^27–29. A high axial load ratio is more likely to fail in a brittle pattern than the low axial load ratio. The coupling effects of torsion and high axial load can make the column more vulnerable under seismic load. Therefore, it is needed to evaluate the seismic performance of RC columns with considering high axial load ratio. Both axial and lateral loads were applied at a distance of 65 mm from the center of the upper beam to account for the combined compression load, bending, shear, and torsion, as shown in Fig. 5. Considering laboratory conditions yet aiming to induce maximum bending and torsion in the column, the eccentricity value was determined to be the larger of the following: either a value greater than the core section of 42 mm or 65 mm, which corresponds to 25% of the column length^30,31 The loading method was referenced to the load protocol given in ACI 374.1³², which states that the initial drift ratio should be within a range that confirms the linear elastic behavior, the subsequent drift ratio should not exceed 0.25%, and the subsequent step should be neither too large nor too small. The loading protocol was repeated three times for each drift ratio, as shown in Fig. 6, and it was planned to repeat up to a drift ratio of 6%. The experiment was terminated when lateral load dropped approximately 85% of the maximum load.

Figure 4.

Test setup (a) Schematic of the test setup (b) Photo of the test setup.

Figure 5.

Loading method.

Figure 6.

Loading protocol.

The steel rods were installed between the beam and the foundation by the compressive force applied while turning the threads of the adjustable link. The purpose of the compressive force on the steel rods was to completely secure it, which was confirmed using pre-attached strain gages on the steel rods. Compressive force was applied to the steel rod until strain level, reaching 10% of the yield strain of the steel rod, were verified³³. Two LVDTs were placed horizontally at a distance of 200 mm on the side of the column to measure the twist angle and determine whether the retrofitted specimens can resist torsional loads as shown in Fig. 7. The twist angle is determined by dividing the difference of the displacement measured at the two LVDTs by the distance between the two LVDTs.

Figure 7.

Crack patterns (a) Definition of each side of column (b) CB (c) SFB (d) SPB.

Experimental results and analysis

Cracks and failure modes of specimens

Figure 7 shows the crack patterns of each specimen according to the drift ratio. The side of the column was labeled alphabetically in counterclockwise order, with the side of the actuator labeled B. The steel rods were retrofitted to the B and D sides of the specimen. The CB specimen exhibited flexural cracks from the bottom of the column on the B and D sides at a drift of about 0.75%. These cracks continued to develop and additional flexure-shear cracks appeared with increasing drift ratio. The shear cracks appeared in the middle of the column at a drift ratio of 1.4%. The concrete cover crushing began forming in the plastic hinge region at a drift of 2.75%. The width of the shear cracks increased significantly. Concrete spalling occurred at the bottom of the surface, leading to failure of the CB specimen. The progress of initial cracks of the SFB specimen was similar for the CB specimen until a drift ratio of 1.75%. Flexural cracks were observed near the bottom of the column at a drift ratio of 0.75%. The shear crack developed near the bottom at a drift ratio of 1.75% and then propagated upward as the drift increased. These cracks induced concrete spalling at a drift ratio of 3.5%. At a drift ratio of 5%, the column failed due to concrete spalling. The crack propagation between SFB and SPB was similar. The SPB specimen began exhibiting flexural cracks near the bottom of the column on sides A and C. These cracks continued to grow and develop into shear cracks on sides A and C. Then, some shear cracks were observed above the L-shaped angle on side B. Concrete spalling and crushing began in the plastic hinge region on side C. The concrete needs to overcome the deformations induced by the crack. Because the steel rods played their role effectively, spalling and failure occurred after a longer time after the initial shear crack in SFB and SPB specimens than in CB specimens. As shown in Fig. 7, the flexural cracks progressed along the entire column in the SFB and SPB specimens; on the other hand, shear cracks were formed locally at the bottom of the column, causing significant concrete spalling in the CB specimen. In both the SFB and SPB specimens, lateral deformation of the steel rods was observed at a drift ratio of 3.5%, then followed by concrete spalling. This is because the deformation of the column was large owing to the yielding of steel rods. The results demonstrated that the steel rod method was able to make full use of the capacity for plastic deformation afforded by the steel rods. In the SFB specimen, no concentrated damage was observed, whereas in the SPB specimen, locally concentrated damage was observed on the upper part of retrofitting with steel rods. Due to the test setup conditions, which simulated cantilever behavior, damage and cracks were concentrated at the lower part of the column. However, with the retrofitting of the steel rod at the lower part of the column, the occurrence of damage and cracks shifted towards the upper part of the column. It is concluded that only the steel rods retrofitted in the plastic hinge region induced the short column effect. The occurrence of concrete spalling is one of the ways to evaluate the location of plastic hinge zone³⁴. In the CB specimen, concrete spalling was concentrated within 350 mm from the bottom of the column, whereas in the SPB specimen, concrete spalling occurred around 450 mm from the bottom of the column. In the SFB specimen, locally concentrated concrete spalling was not observed. This shows that the retrofitting method using a steel rod can relocate the plastic hinge.

Figure 8 shows the deformation of the steel rods. In the SFB specimen, asymmetrical deformation of the steel rods of sides B and D was observed. This is caused by torsion due to the eccentricity under axial and lateral loads. The deformation of the steel rods of the SFB specimen was more clearly observed than that of the SPB specimen because it effectively resisted the torsion of the entire length of the column. The loosening of bolts and a detached L-shaped angle did not occur until the experiment was terminated. Therefore, retrofitting the entire length of the column is more effective in terms of crack control and resisting torsion.

Figure 8.

Deformation of steel rods (a) SFB (b) SPB.

Load–displacement relationships and strains

Figure 9 and Fig. 10 presents the hysteretic loops and skeleton curves for each specimen. The drift ratio is the value obtained by dividing the lateral displacement of the specimen by the effective height to the column. The effective height of the column defined as the distance from the top surface of the foundation to the center of the upper beam. From the curves, five critical characteristics points, namely, shear crack occurrence, yield of longitudinal rebar, yield of steel rod, concrete spalling occurrence, and maximum load can be obtained. A yield point, an ultimate point, and a failure point are shown in Table 2. A yield point is determined by yielding of the longitudinal rebar of the column. The ultimate and failure points are determined as the maximum load and 85% of the maximum load, respectively.

Figure 9.

Hysteresis loops and skeleton curves (a) CB (b)SFB (c) SPB.

Figure 10.

Skeleton curve.

Table 2.

Test results.

Specimen	Yield point		Ultimate point		Failure point
	$P_y$(kN)	${\mathrm{\Delta }}_y$(mm)	$P_u$(kN)	${\mathrm{\Delta }}_u$(mm)	$P_f$(kN)	${\mathrm{\Delta }}_f$(mm)
CB	23.14	31	30.59	61	26	78.58
SFB	40.76	38.50	48.93	79	41.59	80.22
SPB	25.4	31	37.20	61	31.62	83.4

Both the ultimate load and stiffness of the specimens retrofitted with steel rods exhibited significant enhancement compared to the un-retrofitted, CB specimen. The maximum load of the CB specimen was 33.32 kN at a drift ratio of 4.5% in the negative direction, and 30.59 kN at a drift ratio of 3.5% in the positive direction. Since concrete spalled off at a drift ratio of 3.5%, the load was decreased after reaching the maximum load. At a drift ratio of 5%, the load reached 29.06 kN in the negative direction, which is 74% of the maximum lateral load, and 23.67 kN in the positive direction, which is approximately 53% of the maximum lateral load. For SFB, the maximum load was 54.11 kN at a drift ratio of 3.5% in the negative direction and 48.93 kN at a drift ratio of 4.5% in the positive direction. For SPB, the maximum load was 39.5 kN at a drift ratio of 3.5% in the negative direction and 37.2 kN at a drift ratio of 3.5% in the positive direction. The load–displacement hysteresis curves of all specimens were asymmetrical in the positive and negative directions. The reason for this phenomenon was that when load is applied in the negative direction, the specimen is pulled. At this time, a gap occurs between the surface of the specimen and the actuator. Due to this, as a displacement smaller than the actual actuator displacement is applied to the specimen, the accumulated damage is also reduced. Therefore, the data obtained in positive direction loading were used for comparison in terms of maximum load and displacement of all specimens. The difference of parameters had no obvious effect on the skeleton curves in the elastic range, as shown in Fig. 9. After passing through the linear region, a more gentle slope was observed in the curve of the CB specimen than that of the other specimens. In the later stage, the skeleton curves were affected by the retrofitting, and different maximum loads were shown in the curves of each specimen. The maximum load in the positive direction of SFB and SPB, which were retrofitted using steel rods, was 1.6 times and 1.22 times that of CB, respectively.

In SPB specimen, the yield of the longitudinal rebars occurred at similar load level that of the CB specimen. However, partially retrofitting the plastic hinge zone increased the displacement at the final failure. The yielding of longitudinal rebars cannot be delayed, but the rotational capacity of the column can be improved after the yielding of the longitudinal rebars. This is because the deformation capacity was improved by applying appropriate confinement to the potential plastic hinge zone. On the other hand, in SFB specimen, both maximum load and displacement increased because the steel rods resisted in the entire range where bending and shear occurred. This result demonstrated that the steel rod method improved the strength of existing RC columns as well as retrofitting the plastic hinge length, but retrofitting the entire length of the column provides better load-carrying capacity and ductility.

Figure 11 represents the strain of the longitudinal rebar during cyclic lateral loading as measured by a strain gauge installed 100 mm from the bottom of the column. The strains of the longitudinal rebar in every specimen exceeded the yield strain, 0.002. In the case of CB, the strain of the rebar was 0.00293 at a drift ratio of 1.75%. In the case of SFB, the strain of the longitudinal rebar was 0.00246 at a drift ratio of 2.2%, and in the case of SPB, the strain of the longitudinal rebar was 0.00225 at a drift ratio of 1.75%. The SPB yielded at the same drift ratio as did CB, but the strain was approximately 30% lower because the steel rods were engaged to resist deformation. In SPB, it was delayed the yielding of longitudinal rebar than that of SFB. It can be concluded that steel rod retrofitting can minimize the tensile deformation of the longitudinal rebar of the column and delayed buckling of the longitudinal rebar.

Figure 11.

Strains of longitudinal reinforcements due to lateral force (a) CB (b) SFB (c) SPB.

Figure 12 shows the variation of strains in the longitudinal rebars along the column height. The plastic hinge of RC column exhibits a physical behavior in which curvature is concentrated. This can be predicted by the strain of the longitudinal rebar³⁴. Figure 12 shows the theoretical plastic hinge length calculated by Eq. 5 and the physical plastic length based on strains. In the CB specimen, stress concentration at the plastic hinge zone led to increase deformations in the column. However, in the SFB and SPB specimens retrofitted by steel rod, retrofitting the plastic hinge zone resulted in redistributing the plastic hinge from the lower to upper portion of the column. Two plastic hinges were observed in the SPB specimen. Due to the partial retrofitting with steel rods at the upper and lower part of the column, stress concentrated respectively at the upper and lower part of the column, resulting in significant deformation in the column.

Figure 12.

Variation of maximum strain of the longitudinal rebar (a) CB (b) SFB (c) SPB.

The moment distribution of the CB and SFB specimens were similar, but the SPB specimens were different. The maximum strains of the longitudinal rebars took place 325 mm from the bottom of the column in the CB and SFB specimens. On the other hand, the longitudinal rebar of the SPB specimen showed a maximum strain at a 325 mm and 1475 mm from the bottom of the column. In the CB specimen where the top of the column was not restrained, no deformation was observed at the upper part of the longitudinal rebar, but in the SPB and SFB specimens in which the lateral displacement of the top of the column was restrained by the steel rods, deformation was observed at the top of the longitudinal rebar. Since the lateral displacement of the upper part of the specimen was constrained by the steel rods, a high moment was generated and the strains were observed. However, in the SFB specimen, relatively small deformation occurred in the longitudinal rebar than in the SPB specimen because the stress generated in the column was distributed across the steel rod over the entire length of the column. In the SPB specimen, since the 500 mm lengths of the upper and lower parts of the column were partially retrofitted with a steel rod, the stress was concentrated on the short length steel rod. In addition, since the upper part of the column was partially constrained by the steel rods, it was observed that moment generated large strain in the longitudinal rebar.

In Figures 13, the deformation of the steel rods in the SFB and SPB specimens is represented. It was observed that regardless of the steel rod length, their deformation increased gradually with the increasing lateral load. In the SFP specimen, significant deformation occurred in the lower part of the steel rods, closer to the location of the plastic hinge zone. Conversely, in the SPB specimen, while the lower steel rods displayed a pattern similar to the deformation of the longitudinal reinforcement, substantial deformation was not observed in the upper part of the steel rods. The simulating behavior akin to a cantilever column due to the installation method, resulting in lesser moment generation at the top of the column compare to the bottom.

Figure 13.

Deformation of the steel rods (a) Steel rod in SFB (b) Comparing steel rod and longitudinal rebar in SFB (c) Steel rod in SPB (d) Comparing steel rod and longitudinal rebar in SPB.

Twist response

To measure the angle of twist of the column, two LVDTs were installed on the side of the upper beam at an interval of 100 mm, as shown in Fig. 14. Using the data measured from the two LVDTs, the angle of twist is calculated using Eqs. (6) and (7). The torsional moment was calculating using the Eq. (8), which multiplies the applied load by the eccentricity.

$${\mathrm{\Delta }}_{\mathit{avg}}=\left[\frac{{\mathrm{\Delta }}_{LVDT1}+{\mathrm{\Delta }}_{LVDT2}}{2}\right]$$ 6

$${\theta }_{\mathit{twist}}=\text{arctan}\left[\frac{{\mathrm{\Delta }}_{LVDT1}-{\mathrm{\Delta }}_{LVDT2}}{d}\right]$$ 7

$$M_t=P\times e$$ 8

where ${\mathrm{\Delta }}_{\mathit{avg}}$ is the average displacement obtained from two parallel LVDTs in the lateral direction, ${\theta }_{\mathit{twist}}$ is the twist angle occurring at a point, ${\mathrm{\Delta }}_{LVDT1}$ and ${\mathrm{\Delta }}_{LVDT2}$ are the displacement, and $d$ is the distance between ${\mathrm{\Delta }}_{LVDT1}$ and ${\mathrm{\Delta }}_{LVDT2}$, $M_t$ is the torsional moment, $P$ is the lateral load applied by the actuator, $e$ is the eccentricity of the lateral load.

Figure 14.

Methods for measuring torsional moment and twist angle.

Figure 15 shows the torsional moment and the angle of twist. Prakash³⁵ conducted cyclic loading tests on the reinforced concrete column with the torsion-to-moment ratio (T/M ratio) as a variable to analyze the influence of torsion on columns under combined loads. It was noted that specimens with higher T/M ratios exhibited torsional effects, rendering them vulnerable in terms of strength, stiffness, and energy dissipation capacity compared to those with the T/M ratio of 0. In this study, the T/M ratio of 0.036, calculated based on an eccentricity length of 65 mm relative to a column length of 1800 mm, was set to induce torsion. The maximum torsional moment of SFB was 3.15 kN·m in a positive direction and 3.5 kN·m in a negative direction, which represent increases of approximately 68.5% and 64%, respectively, compared with CB. The maximum torsional moment of SPB was 2.38 kN·m in a positive direction and 2.5 kN·m in a negative direction, which represents increases of approximately 27.3% and 17.4%, respectively, compared with CB. The twist angle at the drift ratio of the maximum load of CB was 0.34° in a positive direction and 0.18° in a negative direction, that of SFB was 0.16° in a positive direction and 0.22° in a negative direction, and that of SPB was 0.17° in a positive direction and 0.24° in a negative direction. When the experiment was terminated, the twist angles of SFB and SPB were 1.48 and 1.41 greater than the twist angle at the maximum load, but in CB, the twist angle increased approximately 2.7 times after reaching the maximum load. This result indicated that the twist angle was not amplified by retrofitting with the steel rod even if the load decreased after reaching the maximum load. The retrofitting method using steel rods can prevent abrupt failure due to the twist of the column. While the twist angle of SFB and SPB specimens were similar, the torsional moment of SFB specimen was approximately 40% greater than that of SPB specimen. This is because the twist angle was measured at the top of the column, and the SPB specimen was locally constrained with steel rods at the top of the column. The torsion of the entire member was not considered. Therefore, retrofitting the entire length of the column can resist torsion effectively in terms of the torsional moment.

Figure 15.

Torque/twist curve.

Effective stiffness

With the increases of loading cycles, concrete cracking and steel deformation increase. Also, the stiffness of the column will degrade. The effective stiffness was calculated as shown in Fig. 16 which represented in Tsonos³⁶. Effective stiffness is defined as the slope obtained by dividing the maximum load of each cycle of the specimen by the displacement. The stiffness degradation was slow during the middle stage of the test. The SFB specimen showed stiffness degradation of a gentle curve compared to the other specimens. The initial effective stiffness of CB, SFB, and SPB was 1.64 kN/mm, 2.48 kN/mm, and 2.08 kN/mm, respectively. When flexural cracks were formed initially, effective stiffness was decreased to 1.21 kN/mm in CB, 1.72 kN/mm in SFB, and 1.41 kN/mm in SPB. The degradation of effective stiffness when the load reached maximum load than initial effective stiffness was 77% in CB, 80% in SFB, and SPB. In addition, when the experiment terminated, the effective stiffness of SFB and SPB were 30% and 10.2% greater at a drift ratio of 5% compared to that of CB.

Figure 16.

Influence of test parameters on the effective stiffness degradation.

Ductility and energy dissipation

Ductility is an important factor for evaluating the deformation capacity of specimens when a seismic load is applied. In this study, the displacement ductility index was calculated using Eq. (9), as described by Sheikh and Khoury³⁷, and the torsional ductility index was also calculated using Eq. (10), as described by He et al.³⁸. The ultimate displacement was defined as the displacement at the point at which the load decreased to 85% of the maximum load after the specimen reached the maximum load. The yield displacement was defined as the displacement at which the extension of initial stiffness reached the maximum load based on an equivalent elastoplastic system using initial stiffness in the elastic region.

$${\mu }_{\mathrm{\Delta }}={\mathrm{\Delta }}_u/{\mathrm{\Delta }}_y$$ 9

where ${\mu }_{\mathrm{\Delta }}$ is displacement ductility index, ${\mathrm{\Delta }}_u$ is ultimate displacement, and ${\mathrm{\Delta }}_y$ is yield displacement

$${\mu }_{\theta }={\theta }_u/{\theta }_y$$ 10

where ${\mu }_{\theta }$ is torsional ductility index, ${\theta }_u$ is the twist at maximum torsional moment, and ${\mathrm{\theta }}_y$ is the twist at yield torsional moment.

Figure 17 shows comparison of displacement and torsional ductility index of each specimen. The SFB and SPB showed an increased displacement ductility index of approximately 18% and 15%, respectively, compared with CB. In terms of torsional ductility, the index of the SFB specimen increased by about 40% and the SPB specimen by 33% more than that of the CB specimen. The difference due to the retrofitting length of the steel rod in the displacement and torsional ductility index was not significant. This is because the steel rod retrofitting method can minimize the concentrated local damage of the concrete columns in the plastic hinge zone induced by flexural and shear cracks, buckling of the longitudinal rebar could be minimized.

Figure 17.

Ductility index.

The steel rod method increased the maximum load more than 60% while the ductility increased by about 18%. It is advantageous for strength improvement rather than ductility. This is because the steel rod is designed to yield before the concrete column, and it resist the lateral load effectively before lateral deformation of the steel rods. The steel rod retrofitting is more effective in terms of torsion control than displacement.

Energy dissipation capacity is an important factor to evaluate seismic performance as it measures the ability of a structure to absorb seismic energy. It was calculated as the ar0065ea within the load–displacement curve of each specimen. The cumulative energy dissipation capacity was calculated up to the drift ratio that decreased the lateral load to 85% of the maximum load. The energy dissipation capacities of all the specimens are shown in Fig. 18. Below 1.7% drift ratio, all specimens showed a similar energy dissipation capacity. Whereas, after the shear cracks were observed, above 1.7% drift ratio, the specimens retrofitted with steel rods experienced higher levels of dissipated energy. The SFB specimen absorbed more energy before the steel rod yielded than CB specimen. The energy dissipation capacity of SPB specimen was also slightly higher than that of the CB specimen. After the 3.5% drift ratio, the increase in energy dissipation observed similarly. The cumulative energy dissipation capacity of CB was approximately 5900 kN·mm, while those of SFB and SPB were 6572 kN·mm and 6176 kN·mm, respectively. The RC columns retrofitted with steel rod had a 5 to 11% increase of energy dissipate capacity. This is attributed to the fact that steel rods dissipate energy as applied lateral displacement increases, the specimen retrofitted with steel rods lost their resistance to the lateral displacement load after the steel rod yields.

Figure 18.

Cumulative energy dissipation capacity.

The retrofitting method using steel rods was the effective in improving the ductility and energy dissipation of the un-retrofitted column. In addition, SFB specimen was shown to be more effective than SPB specimen regard to torsional ductility and energy dissipation. This is the because the steel rods of the SFB specimen effectively resisted the torsional moment and shear force occurring in the entire range of the column.

Conclusions

A new seismic retrofitting method on RC columns is presented and evaluated by cyclic loading test. Three specimens were manufactured including one RC column without seismic details and two retrofitted specimens with steel rods. The main conclusions obtained from the study are as follows:

1. The flexural cracks progressed along the entire column in retrofitted specimens, SFB and SPB, on the other hand, the shear cracks were formed at the bottom of the column locally causing significant spalling of concrete in non-retrofitted RC column, CB. In the CB specimen, concrete spalling occurred concentrated on the bottom of the column, while in the SPB specimen, some shear cracks and spalling were observed in the upper part of the retrofitted with steel rods. Concrete spalling and crushing were not observed in the SFB specimen. It was demonstrated that the steel rod method was able to make use of the capacity for plastic deformation afforded by the steel rods.

2. The maximum loads in the positive direction of SFB and SPB were 1.6 times and 1.22 times that of CB, respectively. This result demonstrated that the steel rod method can enhance load-carrying capacity of existing RC column and also retrofitting entire length of the column provides better load-carrying capacity than retrofit only plastic hinge length, relatively. This is because the torsional moment and shear forces generated in the entire range of column are effectively resisted by the steel rods installed along the entire length of the column.

3. The torsional moment of SFB and SPB increased about 67% and 22.35%, compared with CB. When the experiment was terminated, the twist angles of SFB and SPB were 1.48 and 1.41 greater than the twist angle at the maximum load, but in CB, the twist angle increased approximately 2.7 times after reaching the maximum load. This result indicated that the twist angle was not amplified by retrofitting with the steel rod even if the load decreased after reaching the maximum load. Although the twist angles of the SFB and SPB specimens were similar, a large difference was observed in the torsional moment. This is because the twist angle was measured at the upper beam, and the upper beam was partially constrained by the steel rod. It is recommended that the entire length of the column is retrofitted with steel rods in order to effectively resist the torsional moment.

4. The degradation of effective stiffness when the load reached maximum load than initial effective stiffness was 77% in CB, 80% in both SFB and SPB. In addition, the retrofitted specimens improved the ductility and energy dissipation capacity of CB about 20% and 10%, respectively. These results indicated that the steel rod method increased ductility and energy dissipation capacity of the existing column.

5. The experimental results indicated that the steel rod method improved the performance of the RC column in terms of cracking, maximum load, effective stiffness, ductility and energy dissipation. It is recommended to retrofit the entire column length rather than only in the plastic region for better maximum load resistance It can also be easily applied due to simple details and easy installation.

6. In the experiment of this study, the upper part of the specimen (column) was not fixed and free to move from left to right. This setup made the difference between the behaviors of the test specimen and the actual column, but it made easy to apply pushing or pulling the force onto the specimen. Further studies are needed to consider the real behavior of the column through the experiments. In addition, it is required to evaluate the effects of the diameter, number of steel rods, and strengths of the steel rods on the seismic performance in order to present a practical design guideline for seismic retrofitting using steel rods.

Author contributions

Young Hak Lee devised the main conceptual ideas and proof outline. Min Sook Kim planned the and carried out the experiments. Min Sook Kim worte the manuscript with support from Young Hak Lee. All authors discussed the results and commented on the manuscript.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (No. 2020R1A2C2009351 and No. RS-2023–00217322).

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.