Pseudodynamic Testing of Buildings Retrofitted with External Steel Reinforced Concrete Frames to Increase Lateral Strength for Earthquake Damage Prevention

In this study, a new technique for seismic retrofitting via the attachment of an “external steel reinforced concrete frame” (ESRCF) system was developed to strengthen medium-to-low-rise reinforced concrete (RC) buildings. Two methods (bolting and welding) were developed to connect existing RC frames and external strengthening elements; these methods are technically and practically suited to various construction conditions. ,e retrofitting method developed in this study can be used to perform seismic strengthening construction, while residents continue to live within the building. ,e method is categorized as a “strength design approach” implemented via retrofitting, allowing the lateral ultimate load capacity of RC buildings, whose failure mode is shear, to be increased easily. Test specimens were designed based on an existing RC building in Korea lacking seismic data and then strengthened using the ESRCF system. Pseudodynamic and cyclic loading tests were conducted to verify the effects of seismic retrofitting. In total, four RC frame specimens were prepared: one nonstrengthened control specimen for the cyclic loading test, one nonstrengthened control specimen, one specimen strengthened with a welded ESRCF system, and one specimen strengthened with a bolted ESRCF system for the pseudodynamic test. ,e earthquake response behavior with use of the proposed method, in terms of the maximum response strength, response displacement, and degree of earthquake damage, is compared with a control RC frame. ,e test results indicated that both the bolting and weldingmethods used for connecting the existing RC frame to the ESRCF effectively increased the lateral ultimate strength, resulting in reduced response displacement of building structures under large-scale earthquake conditions.


Introduction
Earthquakes are among the most hazardous natural disasters and have frequently wrought devastation, both economically and on human life. Although the repeatability and macroscopic distribution of historic earthquakes can be predicted to a certain extent, it is impossible to predict precisely when and where an earthquake will occur. e destruction of building structures vividly demonstrates the severity of earthquake damage, as observed during a recent series of earthquakes. Long-term earthquake observation and research has improved our understanding of their behaviors, including patterns of damage to structures, and significant efforts have been made to mitigate earthquake damage to new buildings through seismic design. In particular, correlations between the degree of damage caused by earthquakes and the seismic resistance mechanisms of structures have been subjected to vigorous theoretical and experimental investigations. Procedures for calculating shear forces and dynamic responses based on the results of these studies have been adopted in seismic design codes and standards during recent decades, including the International Building Code [1], Building code requirements for structural concrete of the American Concrete Institute [2], Standard for structural calculation of reinforced concrete structures of the Architectural Institute of Japan [3], and the Korean Building Code [4].
Progress in seismic design has resulted in new buildings with improved prospects for satisfactory behavior during earthquakes. Innovations in seismic design methodologies have led designers to question the adequacy of the seismic behavior of existing buildings lacking seismic data, including those affected by recent destructive earthquakes such as the 2008 Sichuan earthquake in China, the 2010 Chile earthquake, the 2011 Van earthquake in Turkey, the 2011 Christchurch earthquake in New Zealand, the 2012 Great East Japan earthquake, the 2016 Central Italy earthquake, the 2016 Kumamoto earthquake in Japan, and the 2017 Pohang earthquake in South Korea. According to the results of previous investigations of earthquake damage, including those listed above, a large proportion of earthquake damage occurs in medium-to-low-rise buildings with fewer than six stories, especially reinforced concrete (RC) buildings lacking seismic data. erefore, it is necessary to retrofit existing medium-to-low-rise RC buildings lacking seismic data to be resistant against earthquakes, and important existing buildings are required to be fully operational even under severe earthquake conditions.
Over the past three decades, rehabilitation procedures for existing RC buildings have been promoted, leading to the development of several seismic strengthening techniques to improve their seismic performance [5][6][7][8][9]. ese methods include techniques to improve ultimate strength or deformation capability, as well as technology for vibration control. Most medium-to-low-rise RC buildings lacking seismic data feature column hoop spacing of approximately 30 cm or wider, resulting in a high risk of shear failure [5,[10][11][12][13][14]; [6]. Retrofitting methods that adopt independent construction techniques for improving ductility have been shown to be inefficient due to extremely inadequate ultimate horizontal strength. Retrofitting techniques that increase strength to improve the seismic performance of medium-to-low-rise RC buildings lacking seismic data are more efficient during shear. Such methods (e.g., use of external joints and frames) can be applied, while residents continue to live within the building, and are thus more economical [7,[15][16][17].
However, most conventional seismic retrofitting techniques used to increase the ultimate strength of RC buildings, such as adding new shear walls to existing frames, steel braces to existing frames, side walls (wing walls) to existing columns, and steel panels to existing frames, and involve internal connections and postinstalled anchors. Seismic strengthening based on internal connections requires expansion into additional areas, which limits building space during construction. With such strengthening work, the retrofitting period can increase indefinitely due to the difficulty of securing adequate work space. In addition, precise construction is required for postinstalled anchor systems.
us, to overcome the drawbacks of the conventional seismic retrofitting technology used to increase ultimate strength, there is an urgent need for a new external internal connection technique that will allow residents to live within the building during construction.
e main purpose of this study was to develop a new method based on external connections for seismic strengthening of existing medium-to-row rise RC frame systems lacking seismic data. e proposed technique uses an "external steel reinforced concrete frame" (ESRCF) system. e seismic performance of the new ESRCF strengthening system was investigated, with the goal of overcoming the drawbacks associated with conventional seismic retrofitting technology, to increase the ultimate strength of existing medium-to-row rise RC frame systems lacking seismic data. Bolting and welding methods for connecting existing RC frames to external strengthening elements were also tested in this study. ese connection techniques are technically and practically suitable for various construction conditions. e retrofitting method developed in this study allows for seismic retrofitting construction, while residents continue to live within the building. e proposed method represents a strength design approach implemented via retrofitting to easily increase the ultimate lateral load capacity of RC buildings, which have shear failure mode.
On the other hand, the optimal method for evaluating the seismic performance of a test structure, e.g., a framing system subjected to earthquake ground motions, is a fullscale structural test performed on a shaking table. Generally, such a test is not possible within current experimental facilities and will never be practical for most large structures. Several alternative approaches have been developed, including shaking table tests of reduced-scale structural models. e drawbacks of such reduced-scale tests are obvious; many structures cannot be adequately represented by reduced-scale structural models. e pseudodynamic test method was developed to conduct realistic experimental tests on full-scale structures subjected to earthquake ground motions [18]. is test uses an online computer and associated test instrumentation to monitor and control a structure, such that displacement thereof closely resembles the consequences of real seismic excitation. e pseudodynamic test is as realistic as shaking table-based testing, where discretization of the model is feasible; its advantages over shaking tables include (a) versatility, where it allows for detailed observation of the specimen during the test, (b) the ability to test full-and large-scale models, thus circumventing the problem of dynamic similitude, (c) lack of requirement for physical structures, because the method uses a numerical model, (d) greatly reduced equipment, installation, and operation costs, (e) circumventing the problems associated with the interactions of a shaking table with heavy specimens, and (f ) improved actuator control and data acquisition reliability due to the relatively slow rate of loading. In principle, the test can be performed in real time; however, physical limitations of the instrumentation dictate that the test must be conducted step-by-step, i.e., pseudodynamically. Shaking table tests may be more appropriate in cases where strain rate effects are significant and/or in distributed parameter systems. Experimental measurements of restoring forces are performed during the test. ese measured forces are then fed into the computer, together with a set of mathematical equations for inertial response characteristics, to determine the structural displacement that would occur as a consequence of a given ground acceleration. is procedure is superior to quasistatic testing because nonlinear structural characteristics are based on instantaneous experimental feedback rather than hypothetical mathematical models. Pseudodynamic testing differs from classical computerbased structural dynamic simulations in that the latter depend on experimentally measured restoring forces rather than on restoring forces computed using a mathematical model.
In this study, test specimens strengthened with the ESRCF system were designed and fabricated on the basis of typical existing RC buildings constructed prior to the seismic code revision in Korea. A pseudodynamic test, as previously stated, was conducted to evaluate the seismic performance of the proposed ESRCF system, and the effectiveness of bolting and welding methods for connecting existing RC frames with external strengthening systems. e test results were compared with those of a control RC frame in terms of maximum response strength, response displacement, and degree of earthquake damage, according to earthquake levels specified by the KBC [4]. Prior to the pseudodynamic test, a cyclic loading test was performed to estimate ultimate and yield shear load-carrying capacity, deformation, and hysteresis in the lateral load-drift relationship of the control RC frame. Hysteresis properties must be evaluated to determine test variables for pseudodynamic tests. In total, four RC frame specimens were prepared: one nonstrengthened control specimen for the cyclic loading test, one nonstrengthened control specimen, one specimen strengthened with a welded ESRCF system, and one specimen strengthened with a bolted ESRCF system for the pseudodynamic test.

Overview of the ESRCF Strengthening System
e widely used chemical lap-splicing method, which employs steel reinforcing bars (rebar) to connect existing RC frames to external strengthening elements, inevitably leads to cracks in the concrete surface during an earthquake ( Figure 1) due to insufficient ductility capacity of the joint concrete. is cracking prevents transmission of the earthquake load from the existing structure to the strengthening system; thus, successful seismic retrofitting is unlikely [7].
e ESRCF system was proposed to improve structural integrity compared with the conventional chemical lap-splicing method.
is study also proposes bolting and welding methods for connecting existing RC frames to external strengthening elements. ese methods are technically and practically suited to various construction conditions, including in terms of work space and workability. Figure 2 depicts detailed diagram of a connection element and technical details of the welded ESRCF strengthening system. As shown in Figure 2, the welded ESRCF strengthening system is composed of (A) an existing structure, (B) an anchor plate to strengthen the existing structure, (C) an anchor bolt for jointing, (D) a steel frame, (E) a connection plate, (F) a plate to strengthen the connection, and (G) epoxy resin; the system is finished with concrete. e main features of the welded ESRCF strengthening system are the anchor plate (B), and the welding of the connection plate (E) to another plate to strengthen the connection (F). is welded connection integrates the existing structure with the reinforcing member. Table 1 lists the construction steps. e welded ESRCF system is superior in terms of workability because welding of connection elements can be conducted during construction work, but is disadvantageous in terms of constructability and work space. Figure 3 shows detailed diagram of a connection element and technical details of the bolted ESRCF strengthening system. As shown in Figure 3, the bolted ESRCF system consists of (A) an existing structure, (B) a steel plate to strengthen the existing structure, (C) an anchor for fixing the steel plate, (D) a steel frame, (E) a connecting device, (F) a bolt for fixing the slit plate, and (G) epoxy resin; the system is finished with concrete. Important features of the bolted ESRCF strengthening system include the bolting of the steel plate (B) to the connecting device, and (E) the fixing bolt of the slit plate (F), which is installed along with the steel frame (D).

e Bolted ESRCF Strengthening System.
is bolted connection results in integration of the existing structure with the reinforcing member. Table 2 lists the construction steps. e bolting method is superior in terms of constructability, but workability is reduced in terms of jointing, setting up, and other factors. Figure 4 shows a postconstruction image of the proposed ESRCF system. e retrofitting method developed in this study can be implemented for seismic retrofitting construction, while residents continue to live within the building. e method represents a "strength design approach" implemented via retrofitting, to easily increase the ultimate lateral load capacity of RC buildings lacking seismic data, which have shear failure mode.

Specimen Design and Fabrication.
e main objective of the cycling loading and pseudodynamic tests was to verify the seismic strengthening effectiveness of the welded and bolted ESRCF systems via connecting existing RC frames to external strengthening members. Specimens were designed and fabricated based on the frame of a typical three-story Korean RC building [17], designed in the 1980s and lacking seismic data. e specified compressive strength of the concrete used in this building was 21 MPa, and the story height was 3,300 mm. e structure consisted of RC frames with spandrel walls in the longitudinal direction, and infilled brick walls in the transverse direction. One story with one span of the exterior frame in the longitudinal direction, including the columns, beams, and spandrel walls of the structure, is shown in Figures 4 and 5. is building type was selected for testing because the seismic capacity of in-filled brick walls in the transverse direction is higher than that in the longitudinal direction [5,6]. e ratio of the dimension reduction of the specimens was approximately 60% (Table 3) when the size of the laboratory facility, actuator equipment, and capacity of the overhead crane were considered.    Figures 5 and 6. e purpose of the tests was to verify the seismic strengthening effectiveness of the ESRCF system, in terms of maximum response strength, response displacement, degree of earthquake damage, and hysteresis in the lateral load-drift relationship, compared with the control specimen. Specimens consisting of columns and beams were designed according to the guidelines for loadcarrying capacity of the JBDPA [6]. Table 3 lists the specifications of each specimen tested. In total, four test specimens were prepared: a control test specimen (nonstrengthened, C-RC) for the cyclic loading test, a control test specimen (nonstrengthened, PD-RC), test specimen strengthened with a welded ESRCF system (PD-WESRCF), and test specimen strengthened with a bolted ESRCF system (PD-BESRCF) for the pseudodynamic test. All specimens had identical dimensions and rebar arrangements. e column cross section was 300 × 300 mm, and the ratio of column clear height to depth (h 0 /D) was 4.7. Each column had an 8-D16-type SD400 main rebar, reinforced with D10 steel shear reinforcement bars at 250-mm intervals. Beams were fabricated for installation as external strengthening members; stiff stubs were installed at the tops of all specimens to secure the columns ( Figure 5). e spandrel wall was 480 mm high and constructed using B-type brick (190 mm long × 90 mm wide × 57 mm thick). e compressive strength of the brick was 8 MPa [17]. e ground motion acceleration in the pseudodynamic test was the north-south (NS) ground motion recorded at El Centro (Table 3) during the 1940 Imperial Valley earthquake,  which has been widely used to verify seismic performance in various building systems. e magnitude of ground motion was normalized to acceleration values of 250, 300, 375, and 450 gal. Figure 7 shows time history records of the normalized ground motion accelerations used in the pseudodynamic test, respectively, together with their acceleration response spectrum. e ground motion acceleration of 250 and 300 gal approximately corresponds to the seismic load in Zone-1 with S D -and S E soil conditions, under which two-thirds of the earthquake ground motion have a 2% probability of exceedance within 50 years (i.e., a 2,500-year return period), as specified by the KBC [4]. Ground motions of 375 and 450 gal were also used, to assess the seismic strengthening effectiveness of the ESRCF system in the context of a very strong earthquake, under which all earthquake ground motion has a 2% probability of exceedance within 50 years. Before the pseudodynamic test, a cyclic loading test was conducted to estimate the lateral load-drift relationship of an RC frame lacking seismic data (C-RC specimen).

Material Properties.
As previously stated, the specified compressive strength of the concrete of the targeting building was 21 MPa. Cylindrical axial compression tests yielded a value of 21.5 ± 1.2 MPa. e test results compare well to the design compressive strength of concrete. e tensile strength of the steel rebar was 400 MPa. We used two rebar diameters: D10 for the shear reinforcement rebar and D16 for the main specimen rebar. e tensile strength of the steel rebar, measured using a universal testing machine (UTM), was 471.0 ± 1.33 MPa for D10 and 516.0 ± 1.55 MPa for D16.

Joint Connection Design of the ESRCF Strengthening
System. e type, embedding depth, and anchor space of the joint connections for the ESRCF strengthening system were based on the anchor design provided by the JBDPA [6]. Table 4 lists the calculation results, and Figures 3 and 4 show the diameter, spacing, and embedding depth of the anchors for the welded and bolted ESRCF strengthening systems, respectively. Figure 5 shows the configuration of a test specimen strengthened using the ESRCF system, based on the joint connection results shown in Table 4.
Based on these results, the joint connection anchors had a diameter (D) of 12 mm and spacing of 150 mm, in double arrangement. A total of 30 anchors were used in each column, with 12 anchors within each segment of 0.2 L C (L C : column lengths) from the top and bottom of the column, and 18 anchors within each segment of 0.6 L C . A total of 38 anchors were used within the beam: 16 anchors within each segment of 0.2 L B (beam lengths) from the left and right of the beam, and 22 anchors within each segment of 0.6 L B . e strength of the joint connections was approximately 4.0-fold higher for the 0.2 L C and 0.2 L B segments, and 1.2-fold higher for the 0.6 Lc and L B segments, versus the lateral resisting force on the ESRCF strengthening frame (Table 4).
e ESRCF system proposed in this study was able to resist the seismic load when used with the existing RC frame.

Description of the Pseudodynamic Testing System Used in
is Study. A simplified schematic diagram of the pseudodynamic testing system used in this study is shown in Figure 8, together with the test configuration. e system can   Shock and Vibration 7 be expressed as a single degree of freedom (SDF) structure. During the test, the computed displacement response is imposed on the specimen via hydraulic actuators. e actual restoring forces are physically measured during the test and used by the control computer to calculate the displacement. Data transformation is achieved using an analog-to-digital/ digital-to-analog converter (DA-16A) [19]. Filtering and amplification are performed to ensure reliable results and accurate closed-loop control. We used a pseudodynamic testing program [20] to calculate displacement based on the equation of motion of the SDF structure, as determined by the control computer: where y is the relative displacement of the mass with respect to its base, c is the damping coefficient, r is the hysteresis restoring force, and € u is the base acceleration. e time integration scheme adopted for solving the equation of motion was an adaptive time-stepping algorithm developed by Shing et al. [21], based on the α-method proposed by Hilber et al. [22]. e algorithm for numerical integration of the pseudodynamic test is expressed as follows: with where, y i , v i , and a i are the mass displacement, velocity, and acceleration at time i△t, respectively, △t is the integration time step, r i is the restoring force, and f i is the external force (i.e., −m € u i ). For a linearly elastic structure, r i � ky i , in which k is the elastic stiffness of the structure. α, β, and c are parameters that govern the numerical properties of the algorithm, which is unconditionally stable e displacement response for the next time step was calculated using equations (1)-(4), based on the stiffness (k), mass (m), and stiffness proportional damping coefficient ), in which the damping ratio ξ was assumed to be 0.03 (i.e., 3% of the critical damping value). e initial stiffness (elastic stiffness), as for the input data for the pseudodynamic testing program, was derived from the results of the cyclic loading test, as will be described. A 2,000-kN hydraulic actuator was used to apply the lateral load. e horizontal displacement used to calculate the displacement response was measured using a 300-mm linear variable differential transformer (LVDT). Each column was subjected to a constant vertical load of 219 kN, which is half of the total weight of one  span of the exterior frame in the longitudinal direction, using two 1,000-kN oil jacks. e ground motion acceleration for the pseudodynamic test was the NS ground motion recorded at El Centro during the 1940 Imperial Valley earthquake. As stated previously, the magnitude of the ground motion was normalized to acceleration values of 250, 300, 375, and 450 gal, based on the KBC [4]. Prior to the pseudodynamic test, a cyclic loading test was performed to estimate the ultimate and yield shear load-carrying capacity and deformation, as well as hysteresis in the lateral load-drift relationship of the C-RC specimen. A cyclic loading scheme with predetermined displacement was used. Load cycles were designed to have antisymmetric moments in the lateral direction and were repeated three times at lateral drift angles (R, %) of 0.04% (loading step: 1, 0.56 mm), 0.1% (loading step: 2, 1.4 mm), 0.2% (loading step: 3, 2.8 mm), 0.25% (loading step: 4, 3.5 mm), 0.33% (loading step: 5, 4.62 mm), 0.4% (loading step: 6, 5.6 mm), 0.5% (loading step: 7, 7 mm), 0.67% (loading step: 8, 9.38 mm), 1% (loading step: 9, 14 mm), 1.5% (loading step: 10, 21 mm), 2% (loading step: 11, 28 mm), 3.0% (loading step: 12, 42 mm), and 5% (loading step: 13, 70 mm). Figure 9 shows the experimental configuration for the pseudodynamic and cyclic loading tests. At the ultimate state, the failure mode was shear, which represents useful information for determining the failure mechanisms of existing Korean RC buildings lacking seismic data. Figure 10 shows load-displacement curves for the C-RC control specimen during the cyclic loading test. Table 5 lists its mechanical characteristics, i.e., shear strength and displacement, under positive and negative loads. Figure 11 shows the lateral response load-displacement relationship of the PD-RC control specimen during the pseudodynamic test under the 250 and 300 gal NS ground motion recorded at El Centro during the 1940 Imperial Valley earthquake; this approximately corresponds to the seismic load in Zone-1 under S D and S E soil conditions, under which two-thirds of the earthquake ground motion have a 2% probability of exceedance within 50 years (2,500year return period), as specified in the KBC [4]. e figure shows an envelope curve describing the lateral load-displacement relationship of the C-RC control specimen following the cyclic loading test. Figures 12-14 show the lateral response load-displacement relationships of the PD-WERCF and PD-BERCF specimens during pseudodynamic testing at 300, 375, and 450 gal, respectively. We also used ground motion of 375 and 450 gal to assess the seismic strengthening effectiveness of Shock and Vibration the ESRCF system under very strong earthquake conditions, where all earthquake ground motion has a 2% probability of exceedance within 50 years. e response load-displacement relationship of the PD-RC control specimen at 300 gal is characterized by an envelope curve for the C-RC control specimen following the cyclic loading test. Table 6 presents a comparison of maximum response strength, maximum response displacement, and degree of earthquake damage in the pseudodynamic test for the PD-RC specimen at 250 and 300 gal, and for the PD-WERCF and PD-BERCF specimens at 300, 375, and 450 gal. e degree of earthquake damage was estimated based on the technique for postearthquake damage evaluation of RC buildings proposed by the JBDPA [23] and Maeda et al. [11]. e maximum earthquake response of the PD-RC control specimen for an earthquake ground motion of 250 gal was observed under a load of 305 kN with lateral displacement of 11.9 mm, which was classified as heavy earthquake damage according to the JBDPA [23]. For ground motion of 300 gal, the same control specimen showed a maximum response strength of 310 kN with lateral displacement of 16.7 mm, which is practically equivalent to drift; this is compared with the displacement at the ultimate point under positive load of 19.5 mm for the C-RC specimen  e control specimen collapsed at the tops of both columns at 5.4 s, following the application of a positive load of 80 kN, with lateral displacement of 28.9 mm during an earthquake ground motion of 300 gal.

Results and Discussion
For an earthquake ground motion of 300 gal, the PD-WERCF specimen showed a maximum response strength of 362 kN with displacement of 9.0 mm, suffering slight damage in the form of minor flexural and shear cracks, whereas the PD-RC control specimen sustained shear collapse. e earthquake response behavior of, and degree of damage to, the PD-BERCF specimen strengthened with the bolted ESRCF system was similar to that of the PD-WERCF specimen. e PD-BERCF specimen showed a maximum response strength of 381 kN with lateral displacement of 9.1 mm at 300 gal. Among the strengthened specimens, the strength ratio (Table 6) at the maximum response to 300 gal was 1.2-fold higher on average than that of the control PD-RC specimen.
Under severe earthquake ground motion (375 gal), the PD-WERCF specimen showed a maximum earthquake response at a load of 472 kN with lateral displacement of 21.1 mm; this was similar to that of the PD-BERCF specimen, with a load of 435 kN and lateral displacement of 23.6 mm. e strength ratio of the reinforced specimens at 375 gal was 1.45-fold higher than that of the control specimen. For both specimens, earthquake damage was estimated as light. At the highest level of earthquake intensity (450 gal), the welded ESRCF strengthening system showed a maximum response strength of 489 kN with displacement of 26.1 mm, sustaining moderate damage according to the JBDPA [23] and Maeda et al. [11]. As observed in the welded ESRCF specimen, the bolted ESRCF specimen showed moderate damage, with shear cracks observed at a severe earthquake level of 450 gal and a maximum response strength of 493 kN with displacement of 23 mm. Both the welded and bolted ESRCF systems showed fewer and thinner cracks, and no major peeling failure of the concrete cover was observed at 450 gal (unlike the control PD-RC specimen, which collapsed at 300 gal).
ese results reveal that the external connection methodology using both welded and bolted ESRCF systems  is useful for seismic strengthening of building systems, especially in terms of increasing ultimate strength.

Response Displacement-Time History Relationships and
Maximum Response Displacement. At 250 gal, the maximum response displacement of the PD-RC control specimen occurred at 4.8 s, with a lateral drift of 14.7 mm; heavy damage was observed. At 300 gal, the control PD-RC specimen showed a maximum response after 5.4 s with lateral displacement of 29.1 mm, resulting in shear collapse of both columns. Figure 15 shows the response displacement-time history relationships of the PD-RC control during pseudodynamic testing under the 250 and 300 gal NS ground motion recorded at El Centro during the 1940 Imperial Valley earthquake. e strengthened PD-WERCF specimen, on the other hand, suffered only slight damage, with a maximum response displacement of 8.9 mm at 2.6 s. e response displacement-time history of the PD-BERCF specimen strengthened using bolted ESRCF was very similar to that of the PD-WERCF specimen under the same earthquake intensity level. At 375 gal, the PD-WERCF specimen showed a maximum response displacement at 4.5 s, with a lateral drift of 23.5 mm, resulting in light damage; this result was similar to that of the PD-BERCF specimen, which showed a lateral drift of 20.9 mm after 4.6 s. At the highest level of earthquake intensity (450 gal), the PD-WERCF specimen strengthened with the welded ESRCF system showed a maximum response displacement of 43.0 mm at 5.3 s and suffered moderate damage. Similar to the response displacement-time history of the welded ESRCF specimen, the bolted ESRCF specimen showed moderate damage, with shear cracks occurring at 450 gal and a maximum response displacement of 38.6 mm after 5.2 s. Figures 16-18 compare response displacement-time history relationships between the PD-WERCF and PD-BERCF specimens during pseudodynamic testing at 300, 375, and 450 gal, respectively, as well as that of the PD-RC control specimen at 300 gal. e earthquake response parameter of primary interest is the decrease in maximum response displacement of the specimens strengthened by the ESRCF system relative to the control. Table 7 lists these values as percentages of ground acceleration at 300 gal, allowing comparison of maximum response displacement between the PD-RC control and PD-ERCF-reinforced specimens. Specimens reinforced using welded and bolted ESRCF systems (PD-WERCF and PS-BERCF) had displacement responses approximately 70% lower than those of the nonstrengthened control specimen.

Crack and Failure Patterns.
Crack and failure patterns that appeared in the PD-RC, PD-WESRCF, and PD-BESRCF specimens during pseudodynamic testing were investigated, as well as in the C-RC specimen during the cyclic loading test. Following the pseudodynamic and cyclic loading tests, both specimens failed in response to shear strain. However, structural behavior, including the failure pattern in the final stage for the control (PD-RC and C-RC) and reinforced (PD-WESRCF and PD-BESRCF) specimens, differed significantly. e welded and bolted ESRCF specimens exhibited similar evidence of failure, i.e., similar crack appearances, load-drift relations, and earthquake responses. In the following discussion, we focus on crack and failure patterns in terms of lateral displacement.

Cyclic Loading Test of the Nonstrengthened Control Specimen (C-RC).
e control specimen was designed based on the frame of a typical Korean RC building [17] lacking seismic data, under the expectation that it would exhibit shear failure. e cyclic loading test was conducted on the C-RC specimen to investigate the hysteresis characteristics of the Korean RC frame lacking seismic data and to determine the variables for the pseudodynamic test. Figure 19 shows the crack patterns observed on the C-RC specimen following the final stages of the cyclic loading test. e first crack occurred at a positive load of 39 kN, and a small flexural crack appeared in both the bottom and upper Moderate damage * Strength ratio is the maximum response strength of a specimen strengthened with ESRCF at each earthquake intensity level relative to that of the control specimen at 300 gal. * * e degree of earthquake damage was estimated based on results reported by JBDPA [23] and Maeda et al. [11]. column faces after three cycles during the third loading step (R � 0.2%, 2.8 mm). Cracking was not observed in the center of the column. Flexural cracks extended into the middle of the column after step four. Following the seventh loading step (R � 0.5%, 7 mm), at both positive (234 kN) and negative (-211 kN) loads, shear cracks were observed at the top faces of the columns, and a number of diagonal shear cracks appeared, some of which were more than 3 mm wide.
When the applied load reached 271 kN, at the ninth positive loading step (R � 1.0%, 14 mm), additional shear   14 Shock and Vibration cracks were observed, with increasingly large widths. We observed peeling failure due to shear forces emanating from the concrete cover, likely due to insufficient shear confinement. Shear collapse occurred at the top of both columns following the application of a load of 302 kN, with a lateral drift of 21 mm (R � 1.5%). e maximum load capacity of the frame of the C-RC specimen was a positive load of 311 kN, with a lateral drift of 19.5 mm (R � 1.4%). e maximum positive load capacity (281 kN) was similar to the maximum negative load capacity, with a lateral drift of 18.2 mm.  Figure 20 shows the crack pattern observed in the PD-RC specimen during the pseudodynamic test, performed by applying earthquake ground motions of 250 gal (El Centro, NS).

Pseudodynamic Test of the Nonstrengthened Control
In the 300-gal earthquake ground motion test, after approximately 1.52 s, under a negative load of 204 kN with lateral displacement of 10.0 mm, large, visible shear cracks were observed on all faces of the columns. Shear cracks increased in number and extended in width after 1.8 s, with lateral displacement of 16.6 mm. After 2.18 s, with lateral displacement of 24.9 mm, we observed peeling failure due to earthquake shear forces from the concrete cover at the top faces of the columns. Shear collapse occurred at the tops of both columns within 5.4 s, following the application of a positive load of 80 kN, with lateral displacement of 28.9 mm. Figure 21 shows the crack pattern observed in the PD-RC specimen during the pseudodynamic test, performed by applying earthquake ground motions of 300 gal (El Centro, NS).

Pseudodynamic Testing of Specimen Strengthened with the Welded ESRCF System (PD-WESRCF).
e PD-WESRCF specimen featured an RC frame strengthened by the welded ESRCF system (Figure 2). Figures 22-24 show the sequence of cracks observed in the PD-WESRCF specimen during pseudodynamic testing, using normalized acceleration values of 300, 375, and 450 gal, respectively, based on NS ground motion recorded at El Centro. As stated above, ground motion of 375 and 450 gal was additionally applied to determine the seismic strengthening effectiveness of the  ESRCF system in the context of a very strong earthquake, under which all earthquake ground motion has a 2% probability of exceedance within 50 years. For the 300-gal earthquake ground motion, a small flexural crack was observed in the upper faces of the columns within approximately 1. With earthquake ground motion of 375 gal, after 1.77 s, under a positive load of 368 kN with lateral displacement of 8.6 mm, diagonal shear cracks were observed at the top and bottom faces of the columns. Shear cracks increased in number and extended to the middle of both columns after 4.5 s; however, the cracks remained small, with widths <1 mm. In contrast, under severe earthquake ground motion (450 gal), under which all earthquake ground motion has with a 2% probability of exceedance within 50 years, visible shear cracks were observed on all faces of the columns, increasing in number and width (>2 mm) after 2.16 s with lateral displacement of 33 mm (641 kN). Compared with the control PD-RC specimen, which collapsed at 300 gal, the cracks were fewer and thinner, and major peeling failure of the concrete cover was not observed. us, the reinforced PD-WESRCF specimen showed moderate damage (i.e., shear cracks) under a severe earthquake level of 450 gal; greater strength was observed in comparison with a nonstrengthened control PD-RC specimen, in which major earthquake damage and shear failure were observed in both upper columns.

Pseudodynamic Testing of the Specimen Strengthened with the Bolted ESRCF System (PD-BESRCF).
e PD-BESRCF specimen was strengthened by the bolted ESRCF system ( Figure 3). Generally, crack appearance, load-drift relationships, and earthquake responses of the bolted ESRCF specimen were very similar to those of the welded ESRCF specimen at all earthquake intensity levels (300, 375, and 450 gal). Figures 25-27 show the sequences of cracks observed in the PD-BESRCF specimen during pseudodynamic testing using normalized acceleration values of 300, 375, and 450 gal, respectively, based on the NS ground motion recorded at El Centro.
At 300 gal, a small flexural crack was observed in the upper faces of the columns with lateral displacement of 3.9 mm at approximately 1.76 s. e flexural cracks extended to the middle of each column after 2.28 s, with lateral positive displacement of 6.6 mm. However, the flexural cracks remained small, and no shear cracks were observed. After approximately 2.55 s, under a positive load of 381 kN with lateral displacement of 8.9 mm, very slight diagonal shear cracks were observed at the top faces of the columns (Figure 25(a)).
With earthquake ground motion of 375 gal, after 2.17 s, under a negative load of 372 kN with lateral displacement of 5.2 mm, diagonal shear cracks were observed at the top and bottom faces of the columns. Shear cracks increased in number, extended to the middle of both columns after 2.65 s, and remained small (<1 mm wide). e bolted ESRCF strengthening system proposed in this study showed good reinforcement effectiveness when subjected to the maximum earthquake ground motion considered in the KBC. Under severe earthquake ground motion (450 gal, Figure 27), visible shear cracks were observed on all faces of the columns, increasing in number and width (>2 mm) after 2.2 s with lateral displacement of 26 mm. Compared with the control PD-RC specimen, which collapsed at 300 gal, the cracks were fewer and thinner, and no major peeling failure of the concrete cover was observed.
As observed for the welded ESRCF specimen, the reinforced PD-WESRCF specimen showed moderate damage, with shear cracks observed under severe earthquake ground motion (450 gal). Greater strength was noted compared with the nonstrengthened control PD-RC specimen, which showed major earthquake damage with shear failure in both upper columns.

Conclusion
In this study, we introduced a new seismic retrofitting approach involving the external attachment of steel RC frames. We applied two methods (bolting and welding) to connect existing RC frames to external strengthening elements, with the goal of overcoming the drawbacks associated with conventional seismic retrofitting technology and thus increasing the ultimate strength of existing medium-to-lowrise RC frame systems lacking seismic data. A typical RC frame building constructed prior to the seismic code revision in Korea was selected to assess the seismic strengthening effectiveness of the ESRCF system proposed in this study. We tested specimens designed and fabricated based on the RC frame and strengthened with the ESRCF system. Pseudodynamic and cyclic loading tests were performed to verify the seismic strengthening effectiveness of the proposed ESRCF method, compared with an original control frame, in terms of lateral load-displacement capacity, maximum response strength, response displacement, and degree of earthquake damage under earthquake ground motions specified by the KBC [4]. e test results revealed that when externally installed on the RC frame, the proposed ESRCF retrofitting system effectively increased lateral ultimate strength, leading to marked resistance under a large-scale earthquake. e major results of this work are summarized as follows (1) e ultimate load capacity of the nonstrengthened control C-RC specimen during the cyclic loading test was 311 kN under a positive load with a lateral drift of 19.5 mm (R � 1.4%). Shear collapse finally occurred at the tops of both columns following the application of a load of 302 kN, with a lateral drift of 21 mm (R � 1.5%). e failure mode at ultimate state was shear, which represents useful information for identifying the failure mechanisms of existing Korean RC buildings without seismic details. (2) e maximum earthquake response of the PD-RC control specimen at 250 gal was observed at a load of 305 kN with lateral displacement of 11.9 mm, which corresponds to heavy earthquake damage. However, at 300 gal, the PD-RC specimen showed a maximum response strength of 310 kN with lateral displacement of 16.7 mm, similar to the C-RC specimen during the cyclic loading test. e top of both columns of the control specimen collapsed after 5.4 s, following the application of a positive load of 80 kN, with lateral displacement of 28.9 mm.
(3) For the PD-WERCF specimen, reinforced using the welded ESRCF system, the pseudodynamic test at 300 gal showed a maximum earthquake response at a load of 362 kN, with lateral displacement of 9.0 mm, corresponding to slight earthquake damage. e earthquake response behavior of, and degree of damage to, the PD-BERCF specimen strengthened using the bolted ESRCF system was similar to that of the PD-WERCF specimen at the same earthquake intensity level. However, the nonstrengthened PD-RC specimen sustained severe damage (i.e., collapse) at 300 gal.
(4) Under a severe earthquake ground motion of 375 gal, the PD-WERCF specimen showed a maximum earthquake response at a load of 472 kN with lateral displacement of 21.1 mm; this was similar to the PD-BERCF specimen, which showed a maximum response at a load of 435 kN and lateral displacement of 23.6 mm. Both specimens exhibited light earthquake damage. At the highest earthquake intensity level (450 gal), the welded ESRCF strengthening system showed a maximum response strength of 489 kN with lateral displacement of 26.1 mm, similar to the bolted ESRCF specimen, which showed a maximum response strength of 493 kN and displacement of 23 mm. Both structures exhibited moderate earthquake damage.
(5) e maximum response displacement of the PD-RC control specimen was 14.7 mm (heavy damage) at 250 gal, and 29.1 mm at 300 gal (shear collapse). Both of the strengthened specimens (PD-WERCF and PD-BERCF) suffered slight damage, with a maximum response displacement of approximately 9 mm. Under a severe earthquake ground motion of 375 gal, the PD-WERCF specimen experienced maximum response displacement at 4.5 s, with a lateral drift of 23.5 mm and light damage. ese results were similar to those of the PD-BERCF specimen, which exhibited a lateral drift of 20.9 mm after 4.6 s. At the highest earthquake intensity level (450 gal), the PD-WERCF specimen showed a maximum response displacement of 43.0 mm (moderate damage). e PD-BESRCF specimen also showed moderate damage, with shear cracks occurring at 450 gal, with a maximum response displacement of 38.6 mm. (6) e decrease in maximum response displacement of the specimens strengthened by the ESRCF system relative to the control during pseudodynamic testing at 300 gal was 69.5 and 69.3% for the welded and bolted ESRCF systems, respectively. (7) ese test results demonstrate that the bolting and welding methods proposed in this study for ESRCF strengthening effectively increased the lateral ultimate strength of the structures, resulting in a reduction in response displacement under large-scale, intense earthquake conditions. e retrofit method developed in this study can be applied for seismic retrofitting construction, while residents continue to live within the building. e new methods also increase the ultimate lateral load capacity of existing RC buildings with shear failure mode.

Data Availability
All datasets generated during this study are available from the corresponding author on reasonable request.

Conflicts of Interest
e authors declare no conflicts of interest.