1. INTRODUCTION
Concrete structures built prior to the use of design guidelines of the 1970’s structural manual generally lack lateral reinforcements. Old building codes assumed less service loads, thereby resulting to poor detailing of lateral reinforcements, which made them posses nonductile behavior and inherently defenseless against seismic forces.
The functionality of the existing bridges are now in question due to number of different circumstances. Factors such as, natural ageing of the structure, use of outdated design codes, and hostile environment contributed to the deterioration of existing concrete bridge piers. The problem with the declining performance of old concrete bridge piers is that the column dictates the performance of the entirety of structure. In order words, the inability of the column to withstand moments and lateral seismic forces could lead into serious damages, or worse, total collapse of the whole system.
A straightforward solution might be simply the demolition of old bridges and the reconstruction of updated structure, however, the need to maintain the functionality of bridges and to construct quickly the new structures while limiting the disruption to the community, and conforming to other nonstructural requirements are also crucial factors. Similarly, this practice does not conserve construction resources such as, time, cost, materials, and the reduction of overall carbon footprint in the construction industry. Furthermore, the destruction of existing bridges are avoided in practice whenever possible. As a result, researchers and engineers developed alternatives and more sustainable options by restoration of the structures while at the same time, enhancing the performance of the structure to the level currently required by the structural design codes. In addition, strengthening techniques and various repair methods are also found advantageous, effective, and economical for rapid repair of bridges resulting to smooth operation of postearthquake disasters.
Bearing in mind that the column of the bridges are vital in the performance of the entire system, hence, it is logical to focus on the improvement of the column to increase the overall behavior of the structure. In order to make up for the lack of lateral reinforcements required by the latest code, scientists and structural engineers provided additional external reinforcement by wrapping the structure with retrofit material. In doing so, the reinforcement levels up to the current design code to have the capability to resist moments and lateral seismic forces while the confinement offers triaxial pressure to the column and diminishing the nonductile deficiency of the concrete. Consequently, the prevention of brittle failure and insuring the ductile behavior of the structure denotes the improvement in energy dissipation capacity which is important factor in seismic performance. Furthermore, the confining tube produces additional lateral reinforcements and increases the capacity of the structure to counter the shear forces generated by seismic events. In addition, supplementary lateral reinforcements also acts to impede the spalling of concrete, thereby, insuring ductility, increasing flexural strength, ultimate drift, and displacement.
According to the past studies (Saadatmanesh et al., 1996;Ye et al., 2003;Rashid and Mansur, 2005;Özcan et al., 2008), existing columns with inadequate lateral reinforcement must be provided by external confinement to enhance the ductile behavior of the structure. Various confinement techniques have been applied to retrofit the damaged concrete bridge piers and many researchers have conducted field tests on scaleddown models of bridges. Several studies (Priestley et al., 1984;Chai et al., 1991;Sun et al., 1992) stated desirable improvement on strengthening concrete bridge piers using steel jackets, while other researchers (Priestley et al., 1992;Yamasaki et al., 1993;Ehsani et al., 1993;Toutanji, 1999) extended the study to the largescale experiments with the use of FRP as the retrofit material and summarized general improvement in the behavior of structure as well. Teng et al. (2002) stated in their study that the use of FRPs are approximately 20% cheaper than steel retrofit considering the costs, time, installation, and manpower of construction methodology.
The increasing demand on the sustainability and restoration of existing bridges have pushed the researchers to optimize the utilization of external confinement. Numerous parametric studies were conducted to establish the relationship of concrete core and retrofit material. Taghia and Bakar (2013) investigated the correlation of varying crosssection of reinforced short column and the CFRP layers. Varying shape of confinement were also assessed such as the study of Zeng et al. (2018) which investigated the behavior of columns with partially wrapped reinforcements. In the recent parametric study, Catuira and Park (2020) suggested that the full external confinement of the structure revealed significant improvement in the overall seismic performance of concrete bridge piers.
Therefore, in this study, the seismic behavior of the concrete bridge pier was investigated considering the variation in retrofit materials, CFRP, GFRP, and steel. This study focuses on the response and behavior of concrete column with respect different retrofit materials applied. The investigation in this study gives emphasis to identify the appropriate and suitable retrofit material needed required by structural retrofit design. The chosen materials to be evaluated given that in the recent stateoftheartreview of Raza et al. (2019) indicates that CFRP, GFRP, and steel were considered to be the leading retrofit materials for rehabilitation of concrete columns. The results of this study were compared to the data of Catuira and Park (2020) in order to evaluate the values obtained. The progressive advancement in technology was utilized using finite element program, ABAQUS (2013). The simulated models were subjected to dynamic loading.
2. REVIEW OF RELATED STUDIES
2.1 Confined Concrete Background
Early publications on concrete confinement began by conducting smallscaled experimental tests of confined concrete. Roy and Sozen (1964) conducted rectangular steel confinement on concrete based on downscaled field tests and observed no increase in concrete strength but substantial increase in ductility. These test results were later on used in the study of Kent and Park (1971) and revealed remarkable increase in the strength capacity and ductile behavior of confined members. Afterwards, the widely used model of Kent and Park sparked numerous studies and branched out to circular models such as the study of analytical models (Leslie, 1974;Desayi et al., 1978) and experimental investigations (Iyengar et al., 1970), which also reported notable increase in strength and ductility of confined members.
Popovics (1973) proposed a detailed equation for a complete stressstrain relationship for both confined and unconfined concrete and researchers expanded the study to a more realistic sizes on scaled versions of actual buildings. Results of previous investigations (Vallenas et al., 1977;Sheikh and Uzumeri, 1980;Scott et al., 1982) published the experimental results and indicated improvement in the behavior of confined square sections in terms of strength and ductility, but not in good agreement with the analytical calculations. As more and more literatures are becoming available, Mander et al. (1988) adopted the study of Popovics in 1973 and illustrated the widely used compressive stressstrain model for confined and unconfined concrete as shown in Fig. 1. The model is based on a constant confining pressure, f_{l}, where f′_{cc} is the peak strength of confined concrete, ∈_{cc} is the peak strain of confined concrete and f′_{c} is the unconfined strength of the concrete. For concretefilled tubes (CFT), the lateral confining pressure, f_{l}, could be calculated based on the freebody diagram as shown in Fig. 2. Based from the freebody diagram of CFT, the lateral confining pressure of a concrete filled tube manifests continuous confinement offered by the outer steel tube.
The behavior of CFT has been studied by many researchers not only by using steel as outer tube but also FiberReinforced Polymer (FRP) as confining material. Several publications (Fardis and Khalili, 1981;Samaan et al., 1998;Saafi et al., 1999;Pessiki et al., 2001) analyzed the behavior of CFT confined by steel while numerous literatures (O’Shea and Bridge, 1997a;1997b;Sakino and Tomii, 1981) conducted investigation on CFT using FRP.
Both materials, steel and FRP, presented enhancement in the behavior of concrete and in the dissertation of Han (2006), the compressive stressstrain behavior of the confined concrete behaves differently in comparison with unconfined concrete. It is explained that the outer tube provides continuous confinement to the concrete core, as a result, the filledin concrete acts as triaxial than uniaxial until the outer yields.
2.2 Stressstate Mechanism of Confined Concrete
According to the various publications (Richart et al., 1928;De Lorenzis and Tepfers, 2001;2003;Teng et al., 2002;Guo, 2014), the confinement action exerted by the outer tube to the inner concrete is passive type. When axial pressure is applied to the inner concrete, the filledin core horizontally expands and this enlargement is restrained by the outer tube. The confining tube acts against to the lateral expansion of the core and corresponding confining pressure induces triaxial state of stress in the concrete. Fig. 3 illustrates the mechanism of confined concrete.
2.3 Ductility
Ductile concrete column denotes proper dissipation of seismic energy during earthquake events. Ductility of the structure also implies adequate transverse reinforcement which opposes the moments and lateral forces propagated by seismic energy. Therefore, ductility is an important criterion to consider in analyzing the seismic performance of a structure.
In this paper, the ductility of the evaluated structures was defined using the concept of energy. Jeong (1994) formulated an energybased method to calculate the ductility of the structure by the ratio of any two areas energy, inelastic, elastic, and total energy, as displayed in the ductility indices on Fig. 4. Herein, the ratio of inelastic energy to total energy was adopted and obtained through numerical integration. Moreover, Grace et al. (1998) proposed the classification of energy ratio listed on Table 1.
The slope, S, that separates the inelastic and elastic energy is represented by the following equation:
whereas, the slopes, S_{1} , S_{2} , and S_{3}, are calculated through analytical computation. Loads, P_{1} and P_{2}, are the points of intersection attributed to the extended slopes and P_{3} is the ultimate load.
2.4 Damage Index
Damage Index (DI) is a tool to quantitatively measure the extent of damages induced by earthquakes. Damage Index is developed to accurately predict seismic damage by incorporating parameters such as ductility, displacement, stiffness degradation, and energy absorption capacity. Powell and Allahabadi (1988) established a technique using Eq. (3), where Δ_{max} is defined as the maximum deflection, Δ_{yld} is the yield displacement and Δ_{ult} is identified as the ultimate deformation. Calculated DI having values greater than 1 indicates the outset of structural disintegration.
3. FINITE ELEMENT MODELING
The increasing growth and development of technology has provided a great number of powerful software for quick and accurate numerical simulations of engineering problems. Computer modeling has become widely utilized to complement or predict the behavior of experimental tests that can be expensive or difficult to execute in reallife scenarios. As stated, this study takes advantage of the progressive technology and used a finite element software, ABAQUS, to simulate and predict the seismic performance of different retrofit materials applied to concrete bridge piers. The dimensions of the simulated model were taken from the reallife experimental tests of hydraulic actuator. However, in order to avoid generating nonessential elements that slows down the finite element analysis, insignificant elements to the evaluation of the structure were eliminated. Hence, the footing of the simulated specimen was replaced to encastre boundary condition as shown in the schematic diagram of Fig. 5. To evaluate the parametric study, the simulated structure was investigated under four (4) different conditions, namely; (a) Initial, (b) CFRP, (c) GFRP, and (d) steel as displayed in Fig. 6. A threedimensional (3D) finite element was modeled using C3D8R hexahedral elements for the concrete column and loading cap while S4R shell elements were used for the confining tube. The discretized mesh of the finite element model is presented in Fig. 7. The whole structure was subjected to gravity loading and dynamic loading was positioned in the middle of the loading cap until failure. Identical to reallife simulations, the lateral force applied was loaded repeatedly for 3 cycles as graphed in Fig. 8.
The mechanical properties of the external reinforcement incorporated in the model were listed in Table 2. To account for the seismic performance of the retrofit materials, the properties of the retrofit materials were taken from the same standard provision specified by American Concrete Institute (2008). The crosssection of the circular column was made constant and the thickness of confining materials was set to fixed dimension of 3 mm. To accurately mimic the behavior of concrete, concrete damage plasticity (CDP) option was adopted in the study. A concrete with a compressive strength of 30 MPa was incorporated and values taken were calibrated in the published literature of Senturk and Pul (2017) as presented Table 3 and Fig. 9. CDP option includes the tabulated parameters where, f_{b0}/f_{c0}, is the ratio of strength in biaxial state (f_{b0}) to strength in uniaxial state (f_{c0}) and K_{c}, is the ratio of the distances between the hydrostatic axis and respectively the compression meridian and the tension meridian in the deviatoric crosssection.
4. ANALYSIS RESULTS
To evaluate the seismic performance of each retrofit material, the compressive stressstrain curve, loaddeflection hysteresis and skeleton curve, ductility, and damage index of the structure were analyzed and compared. The interpreted results were extracted from an element within the critical section of the simulated model, shown in Fig. 10, using finite element analyses. The obtained data of this study considering the change in retrofit materials, GFRP and steel cases, were compared to the study of Catuira and Park (2020), initial and CFRP case.
4.1 Stressstrain Results
Fig. 11 shows the comparison of stressstrain results considering the change in retrofit materials. Based on the results, finite element analyses revealed general improvement in the compressive strength of the structure with external confinement compared to the initial case, or column not reinforced by any retrofit materials. The stressstrain curve reveals that both CFRP and GFRP displayed remarkable descending motion which started after the ultimate stress, compared to the curve formed by steel. Another approach to examine the behavior of the confined structures is to observe the area under the stressstrain curve. According to the study of Samaan (1998), the area under the stressstrain curve denotes the energy absorption capacity of the structure. The larger the area under the stressstrain curve provides greater energy absorption capacity during seismic events. In addition, data shows that the column externally confined by CFRP significantly improved the capacity of the structure, followed by GFRP, then steel.
4.2 Loaddeflection Hysteresis Curve
Based on Figs. 1213, the base shear capacity and the allowable lateral deformation of the structure with retrofit material increased in comparison to the column without external reinforcements. It could be clearly seen on the envelope curve of Fig. 12, that CFRP as a retrofit material produced the highest capacity of base shear and lateral deformation. GFRP and steel yielded nearly identical results of ultimate base shear capacity and allowable lateral displacement, however, the skeleton curve formed by GFRP shows advantage over steel reinforcement. According to the study of Wang and Li (2012), the spindle shape formed by the loaddeflection hysteresis curve of the structure explains the good plastic deformation ability, seismic performance, and energy dissipation capacity. Therefore, in reference to Fig. 13, both FRPs, CFRP and GFRP, generated steeper spindle shape than the hysteresis curve formed by steel. Thus, GFRP has more superiority to steel as an external cover in terms of base shear and horizontal deformation capacity.
4.3 Ductility
Fig. 14 presents the comparison of the calculated ductility of the retrofitted columns compared to the structure without external confinement. Ductility was obtained after performing series of analytical integration using energybased method from the publication of Jeong (1994). Results show improvement in ductility of concrete bridge pier confined by retrofit materials. The solid line in the graph indicates the ductility of the structure without external confinement. The bars of CFRP, GFRP, and steel exceeding the solid line demonstrates the considerable enhancement in the ductility when confined by external reinforcements. It was observed that confining the structures generally improved the ductility from 78% up to 9092%. Furthermore, the column confined by CFRP responded with the highest ductility of 92%, which suggests the largest ability to dissipate seismic energy under earthquake excitations.
4.4 Damage Index
Fig. 15 displays the relationship of damage index and drift ratio obtained through numerical calculations. All of the evaluated structures exceeded the value of 1, which indicates the beginning of structural collapse. The slope formed with gradual inclination denotes better seismic performance than the slope with steeper course of action. Concrete column confined by CFRP material responded with lowest damage index of 1.003, which signifies lowest predicted seismic damage amongst the rest of the cases.
4.5 Summary of Seismic Performance Evaluation
Table 4 summarizes the results of the seismic performance of retrofit materials. Results showed that considering equal external reinforcement ratio, CFRP is superior in seismic performance considering the investigation in terms of stress, base shear, allowable deformation, ductility, and damage index. In addition, numerical integration displays that the use of GFRP as retrofit material exhibited similar behavior with concrete column confined by steel. However, analytical investigation reveals that GFRP surpassed the seismic performance of the concrete bridge pier confined by steel.
5. CONCLUSION
The principal purpose of this research is to analyze the seismic performance of concrete bridge piers considering the influence of different retrofit materials, CFRP, GFRP and steel, and establish the optimum material for rehabilitation of bridges.

1. It was confirmed that confining the concrete bridge column by any retrofit material generally improved the overall seismic performance of the column. The comparison of retrofit materials indicated that CFRP material provided the most significant and desirable results under seismic excitation. CFRP generated the largest increase in the capacity of yield and ultimate stress, base shear, deformation. In addition to that, the ductility of the structure using CFRP increased from 78% up to 92%. The concrete column confined by CFRP also generated the lowest damage index, 1.003, which indicates the lowest predicted seismic damage to the column.

2. The compressive stressstrain of both concrete columns confined by FRPs, CFRP and GFRP, generated a peculiar declining trend that lies after the ultimate stress. The resulting downward trend indicates that FRPs has higher slenderness ratio that signifies the active confinement of the material and denotes the energy absorption capacity of the structure. Considering equal amount of external reinforcement ratio, using FRPs as external reinforcement to concrete columns exhibited superior enhancement of the structure than the use of steel.

3. Loaddisplacement hysteresis curve illustrates that both of the FRPs, CFRP and GFRP, produced steeper spindle hysteresis curve than the spindle shape generated by steel. In other words, the envelope curve of FRPs indicate better plastic deformation capacity and greater dissipation of energy than the use of steel as external confining material.

4. GFRP and steel retrofit materials produced similar outcomes especially in terms of ultimate base shear and ultimate displacement, however, the skeleton shape of the hysteresis curve played an important deciding factor to verdict that GFRP material surpassed the seismic performance of steel retrofit material. Furthermore, the analytical integration evaluated under the area of the skeleton curve of loaddeflection hysteresis proves that the ductility and damage index of GFRP is better than steel. Hence, the seismic performance of the column retrofitted by GFRP is superior than the concrete columns externally confined by steel.

5. The desirable property of retrofit material is critical in continuous confinement of concrete bridge pier. The high strengthtoweight ratio or specific strength of the retrofit reinforcement governed the significant improvement of seismic performance of the structure. Hence, the promising mechanical properties of CFRP immensely contributed to the desirable seismic performance of rehabilitated concrete bridge pier.