An experimental study on the effect of frame-to-wall connection type on the seismic behavior of steel frames infilled with autoclave-cured aerated concrete blocks

Abstract

In this article, in order to find some reliable connection types for avoiding the infill failure, an experimental study is conducted on six half-scale single-story single-bay specimens, including one bare frame and five steel frames with autoclave-cured aerated concrete walls, each having different frame-to-wall connection types. Polyurethane adhesive was used in bed joints of the autoclave-cured aerated concrete walls as a mortar. The specimens were tested under displacement-controlled loading to examine their effect on the in-plane behavior of steel frames. The results suggested that V- and T-shaped connectors can be used as prequalified connectors between the wall and the frame. In specimens with these connectors, the interaction between infill and frame was ignorable and the infill remained intact up to life safety performance level (2.5% drift ratio). Furthermore, the application of these connectors resulted in a two-level performance of the infill whereby the frame stiffness degradation and strength deterioration after the drift ratio of 2.5% was compensated.

Keywords

autoclave-cured aerated concrete blocks connection type infilled steel frames prequalified connectors

Introduction

In many buildings, the use of internal and exterior partition walls as an infill is inevitable for architectural reasons. Due to the complexity of the seismic behavior of infill walls and the lack of a reasonable and simple model, the effect of such elements is neglected in the building analysis and design phases. However, recent earthquakes have shown that ignoring infill wall influences is not always safe and may lead to great casualties and damages. One of the main reasons for the failure of infill walls is the unreasonable connector type between the infill wall and the surrounding frame elements (Wang and Li, 2017). Different techniques have been proposed for improving the behavior of infill walls which include using reinforced shotcrete (Kahn, 1984), thin ferrocement overlay (Mander et al., 1993), externally applied fiber-reinforced polymers (FRPs; Albert et al., 2001; ElGawady et al., 2006), exterior welded wire (Bertero and Brokken, 1983), glass-fiber-reinforced polymer (GFRP) laminates (El-Dakhakhni et al., 2004), and adding mesh-reinforced plaster layer to one or both faces of the wall (Altın et al., 2010). Strengthening infill walls with prefabricated, prestressed elements also improves their structural behavior (Schwartz et al., 2011).

The interaction between infill walls and the frame has both positive and negative effects on the frame, infill wall, and overall behavior of the building. Reducing displacement and lateral load share of the frame and increasing resistance and energy dissipation are among the positive effects. Creation of soft first story (irregularity in height) or short column (nonuniform distribution of forces in the columns) and torsion (irregularity in plan) are its negative effects. In recent decades, scholars have focused mostly on the positive effects of walls but have neglected their detrimental effects (Crisafulli et al., 2000; Jung and Aref, 2005; Sahota and Riddington, 2001).Ju et al. (2012) proposed to separate the reinforced concrete (RC) infill wall from the steel moment frame by slits in order to eliminate the soft-story problem caused by the vertically irregular configuration of the walls. Markulak et al. (2013) proposed a new masonry infill type by combining perforated clay and lightweight autoclave-cured aerated concrete (AAC) materials which allowed partial separation of the masonry from the steel frame at certain drift ratios and somehow prevented the infill’s detrimental effects. Morandi et al. (2018b) showed that the detrimental effects of the wall–frame interaction can be reduced by the combined use of sliding joints inserted in the masonry and deformable joints at the wall–frame interface. Some studies have also been conducted on how to achieve an engineered infilled frame by applying sliding fuses in the infill wall (Mohammadi and Akrami, 2010; Mohammadi et al., 2011; Mohammadi and Mohammadi-Ghazi, 2012). Hashemi et al. (2018a) used the technique of separating the infill wall from the frame for reducing the interaction between the infill wall and the surrounding frame. The technique increased the energy dissipation capability of the specimen compared with the bare frame and also raised the drift capacity compared to the infilled frame with a full connection between the infill and the frame. More studies are still required on how to connect the infill wall to the frame, especially the AAC walls which are widely used in constructions in recent years.

Although there are many experimental studies on the behavior of AAC masonry (Costa et al., 2011; Miccoli, 2018; Penna et al., 2015; Rosti et al., 2016; Tomaževič and Gams, 2011), not many have undertaken the effect of the frame-to-wall connection type. In one study, Wang and Li (2017) investigated the seismic behavior and failure modes of square concrete-filled steel tube frames having AAC infills. They showed that the steel frame with AAC will exhibit good cooperative behavior when reliable connectors such as hooked bolt, rocking connector, U-typed connector, and angle steel are used. This study focuses on the effect of the connection type between AAC infill walls and the steel frame. It constitutes an effort in finding some kinds of prequalified connectors to be used between the infill wall and the frame. For this purpose, five AAC infilled steel frames with different infill–frame connection types were tested. The specimens were compared with regard to their failure modes, strength, stiffness degradation, infill or frame damage evolution, cracking pattern, hysteretic behavior, viscous damping, and energy-dissipating capacities. Finally, two connection types were introduced as prequalified connectors which resulted in a two-level performance of the infill.

Experimental study

Test specimens

Six 1/2 scaled single-story single-bay steel moment-resisting frames including five frames infilled with AAC blocks and one bare frame were tested. The beams and columns of all frames were HE-A120. The infills were made of AAC blocks with the dimensions of 300 × 120 × 100 mm³. Polyurethane adhesive was used (StoneFix; Zettex Company, The Netherlands) as a bed-joint mortar. It is an affordable and easy-to-use adhesive and can be a good alternative to conventional and traditional mortars. To attach the AAC blocks to each other, thin steel plates and iron pins were used in horizontal even courses (Figure 1).

Table 1 presents the information of the specimens. In specimen CINF-0, a regular AAC infill was bonded to the beams and columns of the infilled frame with the polyurethane adhesive (Figure 2(a)). In specimen CINF-1, in addition to the polyurethane adhesive, steel plates with the dimensions of 220 × 50 × 3 mm³ were welded to the columns as shear connectors as well (Figure 2(b)). The distance between them was 240 mm. In specimen CINF-2, the infill wall was “completely” separated from the frame, and there was no interaction between the frame and the wall up to a drift ratio of 1.5%. The infill wall was surrounded by 2L40 × 4 steel angle sections at a distance of 12 mm from the inner surface of columns and the upper beam. The out-of-plane stability of the infill wall was provided by some steel plates with the dimensions of 200 × 80 × 4 mm³ welded to the column and beam flanges from both sides and in contact with the confining angle (Figure 2(c)). In specimen CINF-3, the infill wall surrounded by steel angle sections was “partially” separated from the frame. For connecting the infill to the frame, a number of V-shaped steel connectors were welded to the columns (three at each corner) and the confining angle (Figure 2(d)). The distance of the angle section from the column was 24 mm and that from the upper beam was 12 mm. The out-of-plane stability in this specimen was also provided by 200 × 80 × 4 mm³ steel plates welded on the middle of the column flanges and in midpoints and also at a distance equal to one sixth of the total length from the corner of the beams. In specimen CINF-4, T-shaped steel connectors were used between the wall and the frame. These connectors were welded to the columns (three at each corner) as shown in Figure 2(e). One side of this connector was welded to the column and the other side was horizontally placed in the bed joints of the AAC wall. The length of the horizontal part was 220 mm and the distance between these connectors was 120 mm. In order to provide out-of-plane stability, some steel plates were welded to the columns and beam flanges similar to specimen CINF-3. In this specimen, the wall was not confined with angle sections. The wall was in contact with the upper and lower beam flanges, and the distance between the wall and the columns was 24 mm.

Table 1.

Details of the test specimens.

Specimen	Wall type	Wall thickness (mm)	Mortar and glue	Wall connection type	Connector type
BF	–	–	–	–	–
CINF-0	Block	100	Polyurethane	Infill	Type A
CINF-1	Block	100	Polyurethane	Infill	Type B
CINF-2	Block	100	Polyurethane	Infill	Type C
CINF-3	Block	100	Polyurethane	Infill	Type D
CINF-4	Block	100	Polyurethane	Infill	Type E

Material properties

To determine the mechanical properties of the used steel in the frame sections and the connectors, three samples from flanges, three from webs of HE-A120 sections, and three from connectors were tested according to ASTM A370-11 (2011). The related results are reported in Table 2.

Table 2.

Summary of the mechanical properties of steel materials.

Property	Frame section			Connectors
Property	No. of specimens	Mean	CV (%)	No. of specimens	Mean	CV (%)
Yield stress, F_y (MPa)	6	297.5	2.47	3	288	0.60
Ultimate stress, F_u (MPa)	6	435.5	2.66	3	376	0.67
Ultimate strain, ε_u (%)	6	21	2.91	3	21.5	2.58
Young’s modulus, E (GPa)	6	202	1.78	3	198	0.77

CV: coefficient of variation.

For determining the compressive strength of AAC units, six units were tested according to ASTM C1386-07 (2009). The results reported the average compressive strength and the dry density of 3.4 MPa and 5.20 kN/m³, respectively (Table 3).

Table 3.

Summary of compressive, tensile, and shear strength of the AAC block.

Property	Type of specimens	No. of specimens	Mean (MPa)	CV (%)
Compressive strength, f_m	Units	6	3.40	12.47
Compressive strength, f_b	Prisms	3	1.07	11.79
Elastic modulus, E_m	Prisms	3	1637	6.36
Tensile strength, f_t	Wallettes	4	0.22	16.17
Shear strength, f_v	Wallettes	4	0.35	15.78

AAC: autoclave-cured aerated concrete; CV: coefficient of variation.

To determine the pure shear strength of the polyurethane adhesive, based on the DIN EN 1052-3 (2002) standard and without applying precompression, tests were conducted in two series, each having three samples made of scaled AAC blocks with the dimensions of 300 × 120 × 100 mm³ (Figure 3). In sample A-i (i = 1, 2, 3), AAC blocks were jointed to each other using the adhesive, while in sample B-i (i = 1, 2, 3), thin steel plates were employed between the blocks in addition to the adhesive (Figure 3(a) and (b)). They were tested 72 h after construction. This time was determined because the used adhesive (StoneFix) takes up to 72 h to be fully cured according to the manufacturer (Zettex Company). The average shear strength results for the two samples are presented in Table 4.

Table 4.

Summary of shear strength of the glue.

Property	Sample	No. of specimens	Mean	CV (%)
Shear strength, f_vo (MPa)	A-i	3	0.41	9.30
Shear strength, f_vo (MPa)	B-i	3	0.22	9.32

CV: coefficient of variation.

For determining the compressive strength of the AAC wall, three AAC prisms were prepared, each consisting of two courses of AAC blocks bonded to each other by the adhesive. After 72 h, they were subjected to vertical compression strength tests using a hydraulic actuator according to ASTM C1314-07 (2009). The results reported the average vertical compressive strength of 1.07 MPa and the average vertical elastic modulus of 1637 MPa (Table 3). Furthermore, tensile strength and shear strength of AAC masonry prisms were measured by preparing four block wallettes according to ASTM E519/E519M-10 (2009) which is shown in Figure 4. Samples were tested after 72 h. They were instrumented with two linear variable displacement transducers (LVDTs) to measure their horizontal and vertical deformations. The pressure load was applied in a diagonal direction using a hydraulic actuator. According to the results, the average tensile strength and shear strength of the AAC block were 0.22 and 0.35 MPa, respectively (Table 3).

In-plane test setup and instrumentation

The test setup including the reaction frame, strong floor, loading equipment, instrumentation, and specimen is shown schematically in Figure 5. The reaction frame consisted of a rigid reaction beam, bolted firmly on the strong floor. For in-plane testing, each column base of a specimen was connected to a strong floor by four M24-HS-10.9 bolts. The frame was laterally supported at two points to supply its out-of-plane displacements. Specimens were tested under cyclic loading. For this purpose, two hydraulic actuators (500 kN capacity and stroke ±150 mm) were used at the upper beam’s level. A load cell of the hydraulic actuator was installed to monitor the applied in-plane lateral load at each of the two loading points, and the displacement was measured by nine LVDTs. The measured data were stored in a data logger. During horizontal cyclic loading, no vertical loading was applied to the specimens. The loading protocol was according to FEMA 461 (2007) (Figure 6). The loading speed was slow enough for quasi-static cyclic loading.

Experimental results

Failure modes, damage evolution, and cracking patterns

During testing specimen BF, plastic deformation started at column flanges at 18 mm displacement (1.1% drift). The maximum lateral load recorded in this cycle was 44.7 kN. After 98 mm (6% drift), the amount of lateral load (100 kN) remained almost constant. With the increase in the amplitude of horizontal displacement, local buckling at column flange and plasticization of the panel zone at the beam-to-column joint was observed, while no welding fracture was observed at the beam-to-column connections (Figure 7(a)).

In the tested specimen CINF-0 where the wall was connected to the frame using the polyurethane glue, a detachment of the wall from the surrounding frame was observed in the upper corner (interface cracking) at 2.3 mm displacement (0.14% drift). An increase in the amplitude caused horizontal bed-joint sliding in the even courses of the block wall. Another observed failure mode was vertical cracking at a distance of 30 cm from the columns at 0.56% drift and the separation between AAC blocks in vertical bed joints at a drift ratio of 1.1%. After 2.2% drift, the wall subdivided into six horizontal stripes (Figure 7(b)) able to slide on each other in positive and negative loading directions in the horizontal bed joints. No noticeable diagonal cracks were observed on the wall. When the displacement reached 65 mm (4.02% drift ratio), the test was stopped due to the severe cracks and damage evolution.

In the tested specimen CINF-1 where the wall was connected to the frame using the polyurethane glue and shear connectors, interface cracking was observed at 2.4 mm horizontal displacement (0.15% drift). When it reached 6.51 mm (0.4% drift), thin diagonal tension cracks appeared along with the compression diagonal as well as vertical cracks at a distance of 30 cm from the columns. The diagonal cracks tended to propagate through blocks and inclined at 45°. After the 0.56% drift ratio, the crack propagation pattern changed where blocks tended to slide on one another in the horizontal bed joints of courses 12 (at 9.11 mm), 8 and 10 (at 25 mm), and 4 (at 35 mm). In this specimen, like specimen CINF-0, the tendency of blocks for horizontal bed-joint sliding was observed in the even courses where thin steel plates and iron pins were used for attaching blocks together (see Figure 1). This was because the use of thin steel plates in horizontal bed joints reduced the shear strength of the polyurethane glue to 46%. When the displacement reached 35 mm (2.2% drift), specimen CINF-1 subdivided into three horizontal stripes able to slide in the horizontal bed joints (Figure 7(c)). Falling and corner crushing of AAC blocks at 3.4% drift and damage evolution and vertical cracking at 300 mm from the columns at 4% drift were other observations in this specimen.

In the tested specimen CINF-2 (full separation at the frame–wall interface), no damage was observed in the infill wall until 35 mm displacement (2.2% drift). The only failure was observed at 9.1 mm (0.56% drift) where a small part of the column flange yielded. At 35 mm (2.2% drift), a small crack was observed in the wall at a diagonal tensile corner with a length of about 80 mm. As the amplitude of horizontal displacement increased to 45.5 mm (2.8% drift), a diagonal tension crack with 1970 mm length appeared along with the compression diagonal. This crack propagated through blocks and inclined at 45°. With the increase of the displacement to 55 mm (3.4% drift), a diagonal tension crack with 3320 mm length appeared. This crack crossed over the diagonal crack which appeared at 2.8% drift and almost connected two compressive corners of the wall together. Another observation at this displacement was the slip of AAC blocks in the horizontal bed joints. The slip length was 780 mm observed in the eighth course and its starting point was from the end of the diagonal crack which appeared at 2.8% drift (Figure 7(d)). At 65 mm (4.02% drift), in addition to this slip, horizontal bed-joint sliding occurred in the 11th course. The detachment between AAC blocks in vertical bed joints at a distance of 300 mm from the columns was another failure mode observed at this displacement.

In the tested specimen CINF-3 (partial separation of the infill wall from the bonding frame using V-shaped connectors), no damage was observed in the wall up to 55 mm displacement (3.4% drift). At 9.1 mm displacement, plastic hinges were formed at the bottom ends of columns and then developed up to 65 mm (4.02% drift). In this specimen, there was a cooperative behavior between the frame and the wall due to the use of V-shaped connectors. Up to 45.5 mm displacement (2.8% drift), the connectors provided the possibility of sliding for the infill wall through in-plane bending (opening/closing). As a result, bending deformation in the frame did not have negative effects on the wall. At 45.5 mm, the frame entered into contact with the confining angle and caused a frame–wall interaction. The frame-to-angle contact at compressive corners and the space between them at tensile corners were clearly visible (Figure 8(a)). At 3.4% drift, two oblique cracks appeared in one third of the wall area on the loading side. These cracks propagated through blocks and inclined at 40° and 65°, and their lengths were 570 and 1190 mm, respectively. When the amplitude of horizontal displacement reached 65 mm (4.02% drift), in addition to the widening of diagonal cracks occurred in the previous drifts, two more diagonal cracks with 510 and 340 mm lengths appeared on the middle one third of the infill along with the compression diagonal with an inclination angle of 45°. Another failure mode at 65 mm displacement was the horizontal bed-joint sliding in the wall with a length of 960 mm observed in the sixth course which started from the end of the diagonal crack at the same displacement (Figure 7(e)). During the test, no damage was observed in the connectors.

In the tested specimen CINF-4 (partial separation of the infill wall from the frame using T-shaped connectors), sliding in the upper part of the wall along the bed joints was observed at 3.5 mm displacement (0.2% drift) where the slip length was equal to the length of the wall. At 12.8 mm (0.8% drift), the second shear slip occurred in the lower part of the wall. At this drift, at a distance of 420 mm from the column on the loading side, vertical bed joint of the one of the AAC blocks was separated changing the sliding location from the lower part of the wall to the bed joint of the first course which disturbed the performance of connectors in this row. Plastic hinge formation at the bottom ends of columns started when the displacement reached 9.1 mm (0.56% drift) and developed as the displacement increased. At 35 mm (2.2% drift), 45.5 mm (2.8% drift), and 55 mm (3.4% rift) displacements, some cracks appeared in the lower corner of the tensile diagonal. These cracks appeared in the lower corner on the loading side, and a part of tensile corners in the vertical and horizontal bed joints gradually disparted (Figure 7(f)). This failure mode started from 35 mm displacement and developed up to 55 mm displacement. At 2.8% drift, sliding occurred in the horizontal bed joints of the third course with a length of 300 mm near connectors which disturbed its performance on the loading sides (Figure 8(b)).

The cracking patterns observed on the loading side of the tested specimens in the drift ratios of 4.02% and 3% are illustrated in Figures 7 and 9, respectively. As can be seen, the used V- and T-shaped connectors considerably reduced the evolution of damage in specimens CINF-3 and CINF-4.

Hysteric behavior and strength

The horizontal load–displacement curves obtained from all specimens are shown in Figure 10 and their envelope curves compared to specimen BF are presented in Figure 11. According to these figures, it can be said that the presence of AAC infill walls in the infilled frame increased the strength of all specimens. However, the increasing rate and its amount for the same drifts varied in specimens. There was no significant difference between envelope curves of CINF-0 and CINF-1 showing that the presence of shear connectors (CINF-1) does not noticeably increase the strength compared to their absence (CINF-0). The difference in specimens CINF-0 and CINF-1 did not show a degradation of the infill contribution for rigidly attached infills because of sliding mechanisms (Morandi et al., 2018b); however, their connection type evolved the damage in the infill (see Figure 9).

The strength of specimen CINF-2 was similar to that of specimen BF until the drift ratio of 1.5%. In this model, there was no interaction between the frame and the wall up to 1.5% drift due to the complete separation in the frame–wall interface. As the wall came in contact with the frame in this drift ratio, its strength increased and continued until 3% drift. Afterward, the strength decreased due to the evolution of damage in the wall and the frame. Envelope curves of CINF-3 and CINF-4 were almost identical and there was no significant difference between them. Due to the cooperation between the frame and the wall, the increase in the strength of these specimens was evident from the beginning of the cyclic loading. There was a slight increase in the strength of these two specimens up to 1.5% drift. Afterwards, their strength increased more until 4.02% drift.

The ratio of strength of the infilled frames to that of BF and CINF-0 was calculated in the drift ratios of 0.8%, 2.5%, and 3.5% in order to assess the effect of the frame-to-wall connection type and the infill wall on increasing lateral resistance of the infilled frame at performance levels of immediate occupancy (IO) and life safety (LS). The results are shown in Table 5. The strength of specimens CINF-0 and CINF-1 showed a higher increase (87%) at 0.8% drift compared to specimen BF. This rate declined at 2.5% drift due to the evolution of damage in the wall and reached 32% in CINF-0 and 34% in CINF-1. There was no increase in the strength of specimen CINF-2 at 0.8% drift; however, it increased by 40% at 2.5% drift. After 3% drift, no considerable increase was observed. In specimen CINF-3, the strength of the infilled frame increased by 23% at 0.8% drift compared to that of BF. This increase continued with increasing displacement such that at 3.5% drift, without any damage in the infill wall, the strength increase reached 40%. The increase rates in the strength of specimen CINF-4 compared to BF in 0.8%, 2.5%, and 3.5% drift ratios were 30%, 39% and 46%, respectively. Although failure modes of this specimen did not disturb its in- and out-of-plane stability, for an accurate statement about out-of-plane stability of this specimen and even other tested models after the evolution of in-plane damage in the wall, further study is required.

Table 5.

Comparison of the ratio of strength in the infilled frames to that of bare frame and CINF-0.

Specimen	Loading direction	0.8% drift			2.5% drift			3.5% drift
		P (kN)	Ratio		P (kN)	Ratio		P (kN)	Ratio
		P (kN)	(a)	(b)	P (kN)	(a)	(b)	P (kN)	(a)	(b)
BF	Average	31.4	1.00	0.53	78.3	1.00	0.76	87.9	1.00	0.74
CINF-0	Positive	58.8	1.87	1.00	104.6	1.32	1.00	120.2	1.35	1.00
	Negative	58.9			101.9			117.0
	Average	58.85			103.2			118.6
CINF-1	Positive	58.7	1.87	1.00	103.7	1.34	1.02	122.5	1.41	1.04
	Negative	58.8			106.8			125.2
	Average	58.8			105.3			123.9
CINF-2	Positive	31.5	1.01	0.54	112.3	1.40	1.06	124.7	1.40	1.04
	Negative	31.8			107.4			122.1
	Average	31.7			109.9			123.4
CINF-3	Positive	39.9	1.23	0.66	97.8	1.24	0.94	123.7	1.40	1.04
	Negative	37.2			97.1			122.6
	Average	38.6			97.5			123.2
CINF-4	Positive	39.5	1.30	0.69	111.6	1.39	1.06	129.4	1.46	1.08
	Negative	42.1			106.1			127.4
	Average	40.8			108.9			128.4

(a) represents the bare frame, whereas (b) represents CINF-0.

With respect to the connection type, among the tested models, only the strength of specimen CINF-0 can be calculated using Mainstone’s equations. In this regard, its ultimate strength was calculated using the following formulas according to FEMA 306 (1998)

F_{u} = a t_{\inf} F_{me} \cos θ

(1)

a = 0.175 {(λ_{1} h_{col})}^{- 0.4} r_{\inf}

(2)

λ_{1} = {[\frac{E_{me} t_{\inf} \sin 2 θ}{4 E_{fe} I_{col} h_{\inf}}]}^{\frac{1}{4}}

(3)

where h_col is the column height between centerlines of beams (mm), h_inf is the height of the infill panel (mm), E_fe is the expected modulus of elasticity of the frame material (N/mm²), E_me is the expected modulus of elasticity of the infill material (N/mm²), I_col is the moment of inertia of the column (mm⁴), r_inf is the diagonal length of the infill panel (mm), t_inf is the thickness of the infill panel and the equivalent strut (mm), and θ is the angle for which the tangent is the infill height-to-length aspect ratio (rad) which is defined by the following formula with L_inf representing the length of the infill panel (mm)

θ = \tan^{- 1} (\frac{h_{\inf}}{L_{\inf}})

(4)

Using the above equations, the ultimate strength of specimen CINF-0 was obtained as 87.73 kN, while its experimental result was 124.3 kN. This indicates that Mainstone’s equations do not provide a precise estimate of the ultimate strength of CINF-0. In a recent study on the infills constructed with tongue and groove clay, Morandi et al. (2018a) showed that formulations had better fit with the test outcomes. This discrepancy shows the need for further research in this area.

Stiffness degradation

Figure 12 shows the stiffness degradation curves of the specimens. To describe the stiffness degradation, the stiffness degradation factor (K_j) was obtained as (Wang and Li, 2017)

K_{j} = \frac{\sum_{i = 1}^{n} P_{j}^{i}}{\sum_{i = 1}^{n} u_{j}^{i}}

(5)

where $P_{j}^{i}$ shows the maximum horizontal load at the ith loading cycle, $u_{j}^{i}$ is the corresponding displacement at the ith cycle when the column end displacement is equal to j, and n represents the number of loading cycles for each cycle. To quantificationally evaluate the effect of the connection type on increasing the stiffness of specimens and their stiffness degradation, three typical characteristic points of specimens were introduced to stiffness degradation curves corresponding to the drift ratios of 0.8%, 2.5%, and 3.5%. These points are summarized in Table 6. Results showed that the presence of AAC infill walls highly increased the stiffness of specimens CINF-0 and CINF-1 from the very first stages of loading, while in CINF-2, CINF-3, and CINF-4 a slight increase was observed. This is because of using different frame-to-wall connection types in these specimens. Quantitatively, the initial stiffness of CINF-0 and CINF-1 was approximately 3.07 times the initial stiffness of BF. The stiffness values for CINF-2, CINF-3, and CINF-4 were 1.01, 1.15, and 1.28 times, respectively. Overall, with the increase of displacement amplitude, stiffness of all specimens was degraded where the speed of degradation in CINF-0 and CINF-1 was higher.

Table 6.

Comparison of the ratio of stiffness degradation in the infilled frames to the initial stiffness of the bare frame.

Specimen	Loading direction	K_ini		0.8% drift		2.5% drift		3.5% drift
Specimen	Loading direction	K_ini (kN/mm)	Ratio	K_j (kN/mm)	Ratio	K_j (kN/mm)	Ratio	K_j (kN/mm)	Ratio
BF	Average	2.84	1.00	2.62	0.92	2.09	0.74	1.67	0.59
CINF-0	Positive	8.02	3.02	4.90	1.72	2.79	0.97	2.29	0.80
	Negative	9.15		4.91		2.72		2.23
	Average	8.59		4.90		2.75		2.26
CINF-1	Positive	8.11	3.02	4.89	1.72	2.77	0.99	2.33	0.83
	Negative	9.07		4.90		2.85		2.38
	Average	8.60		4.90		2.81		2.36
CINF-2	Positive	2.78	1.01	2.63	0.93	2.99	1.03	2.38	0.83
	Negative	2.67		2.65		2.86		2.33
	Average	2.73		2.64		2.93		2.35
CINF-3	Positive	3.37	1.15	3.33	1.13	2.61	0.92	2.36	0.83
	Negative	3.19		3.10		2.59		2.34
	Average	3.28		3.21		2.60		2.35
CINF-4	Positive	3.68	1.28	3.29	1.20	2.98	1.02	2.46	0.86
	Negative	3.61		3.51		2.83		2.43
	Average	3.65		3.4		2.9		2.45

Energy dissipation capacity

The energy dissipation capacity can be estimated from the areas enclosed by the hysteretic load–displacement loops in each cycle. In this study, for comparing the energy dissipation capability of different specimens at each displacement level, we used the approach applied in previous studies (Hashemi et al., 2018b). The dissipated hysteretic energy at certain displacements normalized by the peak-to-peak displacement (2Δ) and the dissipated energy ratio of the infilled frames to that of the BF are shown in Figure 13. The high energy dissipation capacity of specimens CINF-0 and CINF-1 was evident compared to other tested models, but they were seriously damaged during testing until 2.5% drift. The energy dissipation capacity of CINF-2 was close to that of BF. The greater dissipation capacity of CINF-3 and CINF-4 compared to BF was another notable result. No damage was recorded for CINF-3 until 3% drift, while for CINF-4 a small local damage was observed. The energy dissipation ratio of the specimens compared to that of the BF varies from 5.1 to 1.51 for CINF-0, from 4.7 to 1.79 for CINF-1, from 1 to 1.20 for CINF-2, from 1.27 to 1.59 for CINF-3, and from 1.48 to 1.95 for CINF-4 up to a drift ratio of 2.5%.

Figure 14 illustrates the equivalent damping factor versus relative horizontal displacement of the tested specimens. The equivalent viscous damping ratio $(ξ_{eq})$ was obtained using equation (6) where S_ABC and S_CDA are the areas of the hysteresis curve of ABC and CDA, respectively, and S_OBE and S_ODF are the triangle areas of OBE and ODF, respectively (Figure 15). The figure shows that $ξ_{eq}$ of all specimens reduced until the 13th cycle (12.7 mm displacement), and after this loading cycle their $ξ_{eq}$ increases. This cycle can be considered as the yielding point of all specimens. Up to 2.5% drift, the total average values of $ξ_{eq}$ for specimens CINF-0 to CINF-4 are 2.18, 2.18, 1.57, 1.70, and 1.76 times the average damping value of specimen BF, respectively

ξ_{eq} = \frac{1}{2 π} \frac{S_{ABC} + S_{CDA}}{S_{OBE} + S_{ODF}}

(6)

Conclusion

In this research, the behavior of steel frames infilled with AAC blocks and five different frame-to-wall connection types under cyclic lateral loading was studied. Experimental results showed that the following:

The increase in the strength of specimens CINF-3 and CINF-4 at 3.5% drift in comparison to the BF was 40% and 46%, respectively. This increase in the strength of infilled frames, where no damage was observed in their infill wall until the 2.5% drift ratio (equal to LS performance level based on FEMA 356, 2000), can be considered as an additional safety factor for the seismic resistance of buildings during an earthquake attack even if the wall is constructed with fragile materials such as AAC blocks.

The use of V- and T-shaped connectors can be a promising method for providing a reliable frame-to-wall connection in infilled steel frames for a two-level performance of the infill wall. In these infilled frames, connectors changed the interaction between the wall and the frame into a cooperative behavior, and the wall was not damaged until the drift ratio equal to the LS performance level. Therefore, during an earthquake, these infill walls can help the major frame by increasing its strength in drift ratios larger than the LS performance level.

With respect to the connection type, among the tested models, only the strength of specimen CINF-0 could be calculated using Mainstone’s equation for the purpose of comparison with the experimental results. Based on the comparison results, it was found that Mainstone’s equations were unable to provide a precise estimate of the ultimate strength of CINF-0. Although based on the experimental results of only one sample, we cannot rule out the inefficiency of Mainstone’s equation in estimating the ultimate strength of steel frames infilled with AAC blocks, the apparent contradiction between the calculated value and the experimental results of this specimen indicates the need for further research in this area.

With increasing horizontal displacement, the stiffness of all specimens was degraded where the speed of degradation in CINF-0 and CINF-1 was higher. At 2.5% drift, the initial stiffness values of CINF-2, CINF-3, and CINF-4 were 1.03, 0.93, and 1.02 times the initial stiffness of the BF, respectively, while for the BF it was dropped by 36% indicating that in these specimens the infills compensated for the stiffness degradation of the BF without being damaged.

Results of specimens CINF-2, CINF-3, and CINF-4 revealed that the negative effects caused by the increase of initial tangent stiffness in the infilled frame could be reduced by employing some connectors for the frame-to-wall connection.

This experimental study suggested that the existence of a reliable wall connection type can be reasonable and compromise construction measure for reducing the negative effects of the wall–frame interaction. The V- and T-shaped connectors exhibited good cooperative behavior in terms of safety. The results of this article can be used as a reference for the application of these connectors as prequalified connectors for the frame-to-wall connection in steel structures. Further experimental studies are needed to investigate the in- and out-of-plane seismic behavior of steel frames with the AAC walls using different frame-to-wall connectors. Moreover, the thermal and the acoustic insulation should be studied in order to make the system a viable solution in real construction practice.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Saheb Ali Asadzadeh

Seyed Alireza Zareei

References

Albert

Elwi

Cheng

(2001) Strengthening of unreinforced masonry walls using FRPs. Journal of Composites for Construction 5(2): 76–84.

Altın

Anıl

Kopraman

, et al. (2010) Strengthening masonry infill walls with reinforced plaster. Proceedings of the Institution of Civil Engineers-Structures and Buildings 163: 331–342.

ASTM A370-11 (2011) Standard test methods and definitions for mechanical testing of steel products.

ASTM C1314-07 (2009) Standard test methods for compressive strength of masonry prisms.

ASTM C1386-07 (2009) Standard specification for precast autoclaved aerated concrete (AAC) wall construction units.

ASTM E519/E519M-10 (2009) Standard test method for diagonal tension (shear) in masonry assemblages.

Bertero

Brokken

(1983) Infills in seismic resistant building. Journal of Structural Engineering 109(6): 1337–1361.

Costa

Penna

Magenes

(2011) Seismic performance of autoclaved aerated concrete (AAC) masonry: from experimental testing of the in-plane capacity of walls to building response simulation. Journal of Earthquake Engineering 15(1): 1–31.

Crisafulli

Carr

Park

(2000) Analytical modelling of infilled frame structures—a general review. Bulletin of the New Zealand Society for Earthquake Engineering 33(1): 30–47.

10.

DIN EN 1052-3 (2002) Methods of test for masonry. Part 3: determination of initial shear strength.

11.

El-Dakhakhni

Hamid

Elgaaly

(2004) Seismic retrofit of concrete-masonry-infilled steel frame with glass fiber-reinforced polymer laminates. Journal of Structural Engineering 130(9): 1343–1352.

12.

ElGawady

Lestuzzi

Badoux

(2006) Shear strength of URM walls retrofitted using FRP. Engineering Structures 28(12): 1658–1670.

13.

FEMA 306 (1998) Evaluation of earthquake damaged concrete and masonry wall buildings.

14.

FEMA 356 (2000) Prestandard and commentary for the seismic rehabilitation of buildings.

15.

FEMA 461 (2007) Interim testing protocols for determining the seismic performance characteristics of structural and nonstructural components.

16.

Hashemi

Razzaghi

Moghadam

(2018a) Behaviour of sandwich panel infilled steel frames with different interface conditions. Proceedings of the Institution of Civil Engineers-Structures and Buildings 171(2): 166–177.

17.

Hashemi

Razzaghi

Moghadam

, et al. (2018b) Cyclic testing of steel frames infilled with concrete sandwich panels. Archives of Civil and Mechanical Engineering 18: 557–572.

18.

Lee

Chen

, et al. (2012) Experimental study on separating reinforced concrete infill walls from steel moment frames. Journal of Constructional Steel Research 71: 119–128.

19.

Jung

Aref

(2005) Analytical and numerical studies of polymer matrix composite sandwich infill panels. Composite Structures 68(3): 359–370.

20.

Kahn

(1984) Shotcrete strengthening of brick masonry walls. Concrete International: Design and Construction 6(7): 34–40.

21.

Mander

Nair

Wojtkowski

, et al. (1993) An experimental study on the seismic performance of brick-infilled steel frames with and without retrofit. Technical Report NCEER-93-0001. Buffalo, NY: National Center for Earthquake Engineering Research (NCEER).

22.

Markulak

Radic

Sigmund

(2013) Cyclic testing of single bay steel frames with various types of masonry infill. Engineering Structures 51: 267–277.

23.

Miccoli

(2018) Seismic performances of AAC masonry: a review of experimental and numerical approaches. Mauerwerk 22(5): 314–328.

24.

Mohammadi

Akrami

Mohammadi-Ghazi

(2011) Methods to improve infilled frame ductility. Journal of Structural Engineering 137(69). Available at: https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29ST.1943-541X.0000322

25.

Mohammadi

Akrami

(2010) An engineered infilled frame: behaviour and calibration. Journal of Constructional Steel Research 66(6): 842–849.

26.

Mohammadi

Mohammadi Ghazi

(2012) A new infilled steel frame with engineering properties. Proceedings of the Institution of Civil Engineers-Structures and Buildings 165(1): 15–25.

27.

Morandi

Hak

Magenes

(2018a) Performance-based interpretation of in-plane cyclic tests on RC frames with strong masonry infills. Engineering Structures 156: 503–521.

28.

Morandi

Milanesi

Magenes

(2018b) Innovative solution for seismic-resistant masonry infills with sliding joints: in-plane experimental performance. Engineering Structures 176: 719–733.

29.

Penna

Mandirola

Rota

, et al. (2015) Experimental assessment of the in-plane lateral capacity of autoclaved aerated concrete (AAC) masonry walls with flat-truss bed-joint reinforcement. Construction and Building Materials 82: 155–166.

30.

Rosti

Penna

Rota

, et al. (2016) In-plane cyclic response of low density AAC URM walls. Materials and Structures 49: 4785–4798.

31.

Sahota

Riddington

(2001) Experimental investigation into using lead to reduce vertical load transfer in infilled frames. Engineering Structures 23(1): 94–101.

32.

Schwartz

Mojsilović

Becker

, et al. (2011) Cyclic shear load tests on seismically strengthened masonry walls. In: Proceedings of the 11th North American masonry conference (NAMC), Minneapolis, MN, 5–8 June, pp. 1331–1342. Available at: http://www.schwartz.arch.ethz.ch/Publikationen/Dokumente/mauerwerk.pdf

33.

Tomaževič

Gams

(2011) Shaking table study and modelling of seismic behavior of confined AAC masonry buildings. Bulletin of Earthquake Engineering 10(3): 863–893.

34.

Wang

(2017) Cyclic testing of square CFST frames with ALC panel or block walls. Journal of Constructional Steel Research 130: 264–279.