Introduction

Moment-resisting frame structures are most widely adopted for low-rise buildings in high tectonic regions. However, the serviceability of the frame structures gets compromised by the poor performance of beam-column joints during severe ground shaking [1]. Strong earthquakes impose a greater shear demand on beam-column joints, making them more delicate components in frame structures [2]. The recently observed earthquake evidence shows that the critical part of the structural system most often failed brittle with diagonal cracks in the joint panels [3]. For RC frame structures, building codes provide guidelines for appropriate reinforcement in beam-column joint regions. Under lateral load demand, the framing elements are expected to carry and maintain substantial plastic deformations without undergoing any severe damages. The recent field studies after Chile (2010) and New Zealand (2010) earthquakes demonstrated that the structures designed based on current seismic codes performed better in achieving the life safety level; however, the resultant economic impacts were enormous [4, 5].

A field study revealed that the main reasons for structural joint failure in many developing countries are poor construction practices like sub-standard concrete, a reduced ratio of reinforcement, lack of shear hoops in joint zones, and practicing 90° shear hooks [6]. The structural joints encounter multi-directional shear stresses during lateral loading, making the structure component more vulnerable [7].

The RC structure seismic capacity is improved by providing additional reinforcement in the critical region of the structure, which provides enough confinement to the core material and hence improves the structural joint shear capacity. The ACI-ASCE committee 352 highly recommends that the structural detailing be sufficient in terms of confinement and detailing to avoid severe diagonal crack formations at the joint cores [8]. The resultant flexibility of the structural system gets improved with sufficient confining of the joints. However, following these guidelines usually results in reinforcement congestion, which results in construction problems such as concrete pouring and compaction, making the joint regions more fragile during lateral loading to maintain the designed level deformation [9]. Furthermore, significant flexural cracking and joint zone damage have been observed due to lateral load less than design demand low-strength concrete frames, which can cause shear cracks at the local level and contribute to soft story formation globally [10].

In addition to the amount and arrangement of reinforcement used for strengthening the joint areas, other studies have focused on using enhanced quality materials such as fiber-reinforced composites and high-performance fiber reinforced concrete (HPFRC) [11,12,13,14] in the joint regions for the improvement of structural performance. The newly developed fiber-based composite materials offer many emerging research topics in civil engineering industries [15,16,17,18]. High-performance concrete is more expensive than ordinary concrete, but this drawback can be compensated for by its improved response and lower unit weight. Liu et al. [19] studied high-strength concrete (HSC); the results showed that HSC members failed in a very brittle manner in the anchorage zone and significantly weakened the beam reinforcement bond in the joint zones. However, this drawback was overpowered by introducing fibers into the matrix. With the introduction of the fibers in the concrete matrix, an enhancement in structural ductility was noted, which is crucial in preserving the overall performance of a structural system, especially in improving the energy dissipation and elasticity [20, 21]. Previous studies have shown that fiber-reinforced material exhibits ductile behavior in compression and tension with hardening or softening tensile strain when enough fibers are mixed in concrete [22]. Consequently, the overall response of structure gets improved during lateral dynamic loading. The capability to sustain and resist significant plastic deformation due to shear reversals during seismic activity is another considerable requirement that has been met by the fiber-reinforced materials [2].

ECC is a class of high-performance fiber-reinforced concrete (HPFRC) with inherent high tensile and shear straining capacities [23, 24]. Effective use of fiber-reinforced composites at the critical regions of a structural system may improve the seismic behavior of frame structures. Several studies reveal that incorporating fiber-reinforced composites can significantly reduce or remove the shear reinforcement at joints [25]. Shannag et al. [26] study concluded that using steel FRC to replace traditional concrete in beam-column joints would substantially improve seismic behavior. Likewise, some experimental research has shown that ECC use in beam-column joints can significantly enhance the seismic stability of lightly strengthened joints [27, 28].

The literature has explicitly reported that the implementation of code requirements is an issue for practicing engineers. The same has been addressed with the application of high-performance fiber-reinforced concrete (HPFRC) in structural components. However, very little work has been carried out on this topic, and further experimentation is required to check the reliability of HPFRC in the joint region. Within this context, the possible global consequences of using fiber-reinforced concrete in the joint and plastic regions of earthquake-resistant RC frames and comparing the potential behavior with ordinary concrete structures are of significant interest. Therefore, this study focuses on an ambitious investigation conducted through shake table experiments to describe the seismic response of RC and ECC-RC composite frame structures. Such type of structures is widely used as typical residential and commercial buildings.

Experimental program

Geometrical and structural description of considered models

The geometrical and reinforcement details of the considered prototype two-story frame structure are illustrated in Fig. 1. It has a symmetrical plan of two-bay by one bay having a bay length and story height of 18-ft (5487 mm) and 12-ft (3658 mm). The structural design was carried out for a high seismic zone with soil type B and PGA of 0.40 g, using a lateral force-based procedure. Four different models were considered for the current study. The classification was based on reinforcement detailing (code compliant and non-compliant) and the use of standard concrete or composite material (ECC) at the structural joints and plastic hinge zones of beams and columns. The structural detailing of the code-compliant models was carried out as per modern seismic building code, where the columns having dimensions of 12-inch × 12-inch (305 mm × 305 mm) were reinforced with 8, #6 (8 × 20 mm) rebars and longitudinal and transverse beams with dimensions 12-inch × 18-inch (305 mm × 457 mm) provided with 6, # 6 (6 × 20 mm) steel rebars and confined with #3 rebars (10 mm) having 135° hooks at 3-inch c/c (76.2 mm c/c). Concrete having 28 days strength of 3000 psi (21 MPa) was used in code-compliant models. For the non-compliant models, the optimization in reinforcement detailing and the material were considered, i.e., the use of lower strength concrete, lack of confining ties at the joint panels, a larger spacing of reinforcement in framing members, a reduced ratio of main reinforcement, and provision of non-seismic hooks in shear hoops. The geometric configuration of the non-compliant models was the same as that of a code-compliant model with 6, #6 (6 × 20 mm) steel rebars in the column, 4, #6 (4 × 20 mm) steel rebars in beams confined with #3 (10 mm) bars at 9-inch c/c (228.6 mm). Additionally, low-strength concrete of 2000 psi (14 MPa) was used throughout the model construction, and no shear hoops were provided in the joint panels. The structural specifications for the composite frame models are the same as the code-compliant and non-compliant RC models, respectively. Furthermore, the hybrid frame models were cast with ECC incorporation at the core joint and extended on either side of the panel in the connecting beam and columns plastic hinge length (Fig. 4).

Fig. 1
figure 1figure 1

Plan, geometrical and structural description of the considered prototype

Full size image

The structural description of the selected models is summarized in Table 1. For mass simulation, each model was loaded with 2400 kg additional mass in the form of steel plates secured with the floor slab (Table 2). The mass simulation and characterization were based on similitude requirements in similar previous studies [29,30,31].

Table 1 Characteristic features of the considered models
Full size table
Table 2 Mix proportioning for ECC material
Full size table

Preparation of reduced scale representative models

Simple model idealization was used to cast 1:3 scale down representative models without altering the mechanical and stress–strain properties of the materials [32]. All linear dimensions were scaled down using a scale factor of 3 (Table 3). The standard procedure for material mix design by ACI-211 was used to select the appropriate mix proportion. A mix ratio of 1:1.80:1.60 (C: FA: CA) with w/c of 0.48 yields a compressive strength of 3000 psi (21 MPa) at 28 days, whereas for 2000 psi (14 MPa) concrete, a mix proportion of 1:3.5:2.87 (C: FA: CA) was selected with w/c of 0.80. Similarly, for the hybrid frame models, the mix proportioning of ECC was carried out in several trials. The constituents were proportioned for 1m3 of the mixed batch. High volume water-reducing admixtures were incorporated at the rate of 2.5% by the cumulative weight of cement and flash. The water to binding material ratio employed was 0.41. The optimum amount of polyvinyl alcohol (PVA) fibers was incorporated at the rate of 1.5% by weight of cement plus fly ash. The stress–strain relationship of the normal concrete in comparison to ECC is shown in Fig. 2. The summary of mix proportioning for ECC is illustrated in Table 2.

Table 3 Conversion scale factors based on similitude requirements for the prototype models
Full size table
Fig. 2
figure 2

Stress–strain relationship for concrete and ECC

Full size image

Initially, to ensure fixed support for the test models, the base pads were cast with standard concrete with 3000 psi (21 MPa) strength, where beam-type reinforcement was provided in a double layer. After gaining sufficient strength, construction of the superstructure was carried out in series. The ECC was monolithically cast within the specified regions with normal concrete in the formwork. The models were covered with moist bags for enough time to ensure proper curing. The fabrication process for test models is summarized in Fig. 3.

Fig. 3
figure 3

Fabrication of considered models

Full size image

The layout of ECC at critical regions

Various researchers worked out the performance assessments of ECC enhanced specimens from structural components to beam-column joints. Said et al. [33] applied ECC in the joint core and the plastic hinge zones of the connecting element. Qudah and Malej [34] worked out the reverse cyclic testing of ECC enhanced specimens, where the ECC was extended on either side of the joint core in the connecting frame elements. Conceptually, when the demand associated with the applied load exceeds the structural capacity, the plastic distortion gets localized at the plastic hinge zones. Plastic hinge zones are critical regions that control the structural performance in load-carrying capacity and deformation [35]. Therefore, the vital areas were cast with ECC for the ECC enhanced models, as shown in Fig. 4.

Fig. 4
figure 4

Schematic layout for ECC incorporation at joints and connecting element's plastic hinges zone

Full size image

Structural response requisition scheme

During the dynamic excitation, the structure's actual response was captured with the help of several sensors instrumented in the manner shown in Fig. 5. A total of three accelerometers were installed, one at each floor level and one at the base pad level with a capacity of ± 10 g. Additionally, each floor was provided with two linear displacement transducers and one at the base pad level with a capacity of 24 inches. The displacement sensors were mounted on a reference frame to record the relative motion of the frame model.

Fig. 5
figure 5

Instrumentation setup

Full size image

Base excitation and loading protocol

A single degree of freedom actuator was used to excite the test models. The selection of the input time history depends upon many parameters such as shake table capacity and tectonic similarity of the considered regions in the design. The horizontal component of natural accelerogram of 1994 Northridge record, shown in Fig. 6 (Retrieved from the 090 California Department of Mines and Geology (CDMG) Station 24 of 278—Pacific Earthquake Engineering Research (PEER) Center, at the University of California, Berkeley strong motion database) was selected after a thorough analysis of different time histories. The chosen time history can vibrate the structure with a PGA of 1.0 g, which is inside the seismic actuator's permissible limits and satisfies the regional seismicity of the selected region.

Fig. 6
figure 6

Northridge (1994) natural accelerogram applied as input excitation

Full size image

The selected accelerogram was applied in a series of runs to capture the test frame model's damaging and crack propagation behavior. Each model was subjected to dynamic loading ranging from 5 to 100% of the maximum amplitude of the selected record. The characteristic accelerogram has a PGA, PGV, and PGD of 0.57 g, 518 mm/sec, and 90 mm, respectively. The selected accelerogram was scaled to a PGA of 1.0 g and then applied to each model with 5% to 100% of the maximum amplitude through a seismic actuator; the time step of the selected record was reduced by \(1/\surd 3\) while keeping in view the similitude and scaling requirements [36,37,38]. After each test run, the models were thoroughly investigated for visual cracks and damages. All the models were loaded from elastic to the plastic regime till the incipient collapse state was reached.

Seismic response of the test models

Observed damage mechanism.

ECC/RC code compliant model

Initially, the model was subjected to a self-check run, where the model experienced a PGA of 0.71 g, and vigorous vibration was observed. Tiny flexural cracks appeared at the beam–column interface on both stories. The model underwent a peak roof displacement of 3.62 inches, corresponding to a 1.56% drift ratio. Similarly, during the second run, the model was subjected to a PGA of 0.78 g, yielding a roof drift of 3.25%. The preformed cracks further widened without new cracks formation. The selected accelerogram was applied gradually, scaled from 5 to 80% of actual record intensity. The damage pattern and crack path observed were noted for each test run. Up to 70% runs, no new crack pattern was observed except the degradation associated with the aggravation of existing cracks. In the last run, the model experienced a PGA of 0.63 g with a roof drift ratio of 7.23%. The failure at the ground story column base was observed, where cover spalling and concrete crushing were the primary reasons for the model tilting in out of the plane direction. No ECC spalling and crushing were observed at the structural joints, although the model was at the verge of collapse prevention level (Fig. 7a). It may be attributed to the bridging effect of PVA fibers and the integral nature of ECC material.

Fig. 7
figure 7figure 7

Observed damages at the incipient collapse stage, Refer to Table 3 for damages evaluation

Full size image

ECC/RC non-compliant model

The failure behavior for the optimized model was considerably like the ECC/RC code-compliant model. During self-check, the model experienced a PGA of 0.27 g and underwent a drift ratio of 1.04%. Tiny flexural cracks developed at the end of columns at the ground story (both top and bottom). In the following runs, the existing damages aggravated without new pattern formation. In the final run, the model experienced a PGA of 0.48 g and roof drift of 5.93%, where tiny hairline dense cracks in the joint were observed at the ground story. The rest of the model joints were safe without any crack formation in the joint core. At the ground story column base, slight column toe crushing and spalling of concrete were observed (Fig. 7b), due to which the model tilted in the out-of-plane direction and hence reached the incipient collapse state.

RC code-compliant model

During self-check excitation, the model experienced a peak roof drift of 0.87%. At the ground story, beams flexure cracks appeared, which can be attributed to the yielding of longitudinal reinforcement in flexure and plastic hinging at the beam's plastic zone. The beam–column interface underwent slight cracking at the ground story due to the main rebars slip mechanism. Flexural cracks were also observed in the ground story column base. In the final run, the model experienced a PGA of 1.06 g with a peak roof drift ratio of 5.26%. Due to great demand on the external concrete cover in terms of compressive strain during this run, the column base and top at the ground story underwent concrete crushing and cover splitting. Slight cover spalling was also noted at the top story column base (Fig. 7c). Severe crisscross-type cracking of the joint cores at the ground story was noted due to higher moment demand concentration, leading the model to collapse level. It illustrates that the beam plastic hinge length possesses a higher moment-resisting capacity than the joint yielding capacity.

RC non-compliant model

During the self-check run, the model experienced a peaked roof drift ratio of 0.36%, where slight flexural cracking at the ground and top story columns was observed at both ends of the ground story. Hairline cracking was also observed at the beam ends. In the following runs, no new cracks developed, and the existing cracks were further aggravated. In the final run having a PGA of 0.47 g, the model underwent a peak roof drift ratio of 3.92%. Cover spalling and crushing of concrete at column ends were observed in both stories. Significant vertical cracks appeared at the beam–column interface on the ground story. Additionally, both stories' joint panels exhibit brittle diagonal cracking and a bat-like concrete chunk detached from the top joint (Fig. 7d). The comparative damage evaluations are presented in Table 4 for all the considered models.

Table 4 Models' comparison and damages evaluation
Full size table

Assessment of structural dynamic response

Lateral load-deformation capacity envelops the frame models.

The response data were scaled and transformed for prototype structures according to the similitude requirements. The peak response in each exciting run and the corresponding roof drift were computed. For base shear forces, the cumulative inertial forces associated with each floor mass were calculated and summed. The lateral force–displacement curve was obtained for each model by joining the points corresponding to maximum base shear and the associated roof drift for each test run, as shown in Fig. 8.

Fig. 8
figure 8

Lateral load-deformation envelops for the considered prototype structures

Full size image

The actual response curve was idealized as a bilinear curve using the equal energy principle to calculate different response parameters. The initial slope of the idealized curve is kept similar to the actual curve, and the actual peak displacement observed was taken as the ultimate displacement value. The bilinearization was carried out using iterative procedures, i.e., to equalize the area under the curves. With the incorporation of ECC, the structural yield stiffness and ductility improved. The change in yield strength, yield stiffness, and ductility factor are illustrated in Fig. 9.

Fig. 9
figure 9

Comparison of different dynamic response parameters of RC and ECC/RC specimens

Full size image

Response modification factor (R-factor)

The elastoplastic behavior of frame models can be described in terms of response modification factor (R-Factor). Once the bilinear capacity curves of the structures are formulated, the R-factor can be calculated from the following relation,

$$R = \frac{{V_{{\text{e}}} }}{{V_{{\text{s}}} }} = \frac{{V_{{\text{e}}} }}{{V_{{\text{y}}} }} \times \frac{{V_{{\text{y}}} }}{{V_{{\text{s}}} }} = R_{\mu } \times R_{{\text{s}}}$$
(1)

where \(V_{{\text{e}}}\),\(V_{{\text{s}}}\) and \(V_{{\text{y}}}\) represent the elastic force in the corresponding flexible system, design base shear force, and yielding force in the elastoplastic system. \(R_{{\text{s}}}\) is known as the overstrength-dependent factor, obtained experimentally from dividing the yield force by design base shear force. For ductility dependent factor \(R_{\mu }\), empirical equations are proposed by Newmark and Hall [39, 40] as a function of the structural time period, given as:

$$R_{\mu } = 1,\;{\text{ for}} \;T < 0.20\; {\text{sec}}$$
(2)
$${\text{or}} \;0.20 \;{\text{sec}} < T < 0.50 \;{\text{sec}}$$
(3)
$$R_{\mu } = \mu , \;{\text{for}} \;T > 0.50\; {\text{sec}}$$
(4)

where T represents the elastic time period of a structural system. The summary of the calculation of parameters required for the response modification factor is reported in Table 5. RC frame models underwent a sway type failure mechanism with plastic hinges at the beam ends, with a crack extending to the joint panels. Contrary, the ECC enhanced model showed a column mechanism, where the plastic hinging was observed at the top ends of the column. Furthermore, hairline cracks extended toward the long beam spread in the plastic hinge zone without cracking the joint panels. Compared with the code recommended value of R = 5, the R-factor obtained was 50% greater for the RC code-compliant model, which increased by 140% for the ECC counterpart model. This relatively high difference is attributed to the elasticity and confinement of ECC material.

Table 5 Response modification factor (R-Factor) calculations
Full size table

Similarly, for the non-compliant model, the R-factor of ECC-RC differs by 24%, showing a relatively higher energy absorbance capacity of the ECC enhanced model. The high drop in R-factor value for the non-compliant models is due to low standard concrete, reduced shear, main reinforcement ratio, and lack of confining reinforcement at the joint panels. Overall, the ECC enhanced model response was entirely satisfactory in comparison to standard concrete models.

Seismic response curves for the prototype frames

The seismic response curves are helpful for structural response analysis and fragility functions calculation. For each considered model, the seismic response envelops plotted by correlating the model base pad acceleration as PGA with the peak roof displacement of structures. Unlike the lateral load-deformation envelops, a clear trend in response cannot be noted due to the high degree of sensitivity of input PGA to the high-frequency noise. From Fig. 9, at the nascent collapse level, the RC code-compliant model can resist the high demand of input PGA (1.06 g). A distinct increase in roof displacement can be observed at the ECC enhanced code-compliant model at the collapse stage. For the non-compliant model, the response of the ECC enhanced model is comparatively better than the RC counterpart in terms of input PGA and peak roof displacement. Incorporating the ECC in the joints and beam/column hinge length had a ductile impact on the displacement ductility of the structure. The ECC structure shows a linear trend up to a PGA of 0.47 g, after which a more significant displacement can be observed (Fig. 10).

Fig. 10
figure 10

Seismic response curves for the considered prototype structures

Full size image

Conclusions

This study's following conclusions have been drawn to evaluate the seismic performance of RC and ECC-RC moment-resisting frames.

  1. 1.

    The efficient utilization of ECC in frame structures at the critical regions significantly improves the stiffness and post-yielding behavior of structures located in high seismic zones.

  2. 2.

    In the RC code-deficient models, the cracking mechanism shift from beam failure to column and joint failure. Still, the confinement and bridging effect at the joint panels can significantly improve compliant and non-compliant structures with ECC.

  3. 3.

    The ill effect associated with the brittle failure of concrete in joint cores can be mitigated effectively by incorporating ECC in critical and plastic hinge zones. It was observed that the ECC region suffered minor damages compared to the RC counterpart due to microcracking produced in it.

  4. 4.

    Compared with the RC models, the yield stiffness of the code-compliant and code-deficient models increased approximately 60% and 20% in the ECC models, respectively.

  5. 5.

    The computed R-Factor for the code-compliant RC and ECC-RC was approximately 7.5 and 12, respectively. These R-factor values are more significant than the code-specified value of 5, showing an increase of 50% and 140%, respectively. This increase in the R-factor of ECC-RC frames may be attributed to the dissipating energy capacity of ECC in tension, and the maximum roof drifts displacement due to ductility higher values which caused the main effect on the R-factor.

  6. 6.

    The R-factor for code-deficient RC and ECC-RC models was approximately 2.9 and 3.5, respectively. However, compared with code-specified values, these R-factors, with all those structural deficiencies in RC and ECC-RC models, showed only a decrease of 42% and 30%, respectively. Therefore, a slight increase in the R-factor of the ECC-RC deficient model may indicate the energy dissipation and confining effect of the ECC fiber matrix.