Experimental evaluation of precast concrete sandwich wall panels with steel–glass fiber

Abstract

A new shear connector is proposed in this article. The shear connector is made of steel–glass fiber–reinforced polymer material. Twelve full-scale precast insulated concrete sandwich panels were tested under flexure to analyze their flexural behavior subjected to pressure. The test program was composed of eight sandwich panels with steel–glass fiber–reinforced polymer connectors and four panels for comparison that were panels using stainless steel truss connectors, pure glass fiber–reinforced polymer pin connectors, and no connectors, respectively. Their load–deflection relationships, load–slip relationships, concrete strain profiles along the wythes cross section, as well as the strains in the steel–glass fiber–reinforced polymer W-shaped connectors were investigated in this article. The panels exhibited a composite action in terms of strength exceeding 85% with steel–glass fiber–reinforced polymer connectors and 40 mm insulation thickness. In addition, the other panels with more than 40 mm insulation layer and different diameter connectors only exhibited 26%–62% composite action. When evaluating the degree of the composite action in terms of stiffness, all sandwich panel values ranged from 6% to 26%. But the compared specimens with pure glass fiber–reinforced polymer connector and smaller diameter steel truss connector had lower level composite action less than 10%. Reasonable design of steel–glass fiber–reinforced polymer W-shaped connectors may provide high composite action for panels and prevent the strength from dropping rapidly due to the steel inner core in the connectors.

Keywords

connector degree of partial composite action sandwich wall panel steel–glass fiber–reinforced polymer stiffness method strength method W-shape

Introduction

Precast concrete wall panels are advantageous in practice due to their fast construction in speed. Typical precast concrete sandwich wall panels which consist of two concrete wythes and a layer of insulation between them are often severed as the exterior cladding of buildings to resist normal and transverse earthquake or lateral wind loads in high-rise buildings. The concrete wythes may be conventionally reinforced or prestressed and the insulation layer may be made of extruded polystyrene (XPS) or expanded polystyrene (EPS) rigid foam insulation to provide high thermal resistance (R-value) for the wall panels. Generally, both concrete wythes are of the same thickness ranging from 50 to 150 mm (PCI Committee on Precast Concrete Sandwich Panels, 1997), and the insulation layer has the thickness ranging from 25 to 100 mm (PCI Committee on Precast Concrete Sandwich Panels, 2011).

The two concrete wythes are separated by the insulation layer. Therefore, their structural performance or integrity is determined by the tie elements acting as shear connectors to transfer shear force. Connectors are commonly arranged as solid concrete regions, discrete ties, and continuous ties. The structural behavior of the connectors mainly depends on their geometry and material. The materials of the connectors include but not limit to steel, fiber-reinforced polymer (FRP), glass fiber–reinforced polymer (GFRP; Choi et al., 2015; Einea et al., 1994; Kim and You, 2015; Tomlinson and Fam, 2014), carbon fiber–reinforced polymer (CFRP; Frankl et al., 2011; Hassan and Rizkalla, 2010), basalt fiber–reinforced polymer (BFRP; Tomlinson and Fam, 2015), and plastics. The geometries of the connectors on the market include delta, pin, C-clip, M-clip (Naito et al., 2009), bent-bar truss (Maximos et al., 2007), welded wire truss (Benayoune et al., 2008), gird (Choi et al., 2015; Kim and You, 2015), and so on (see Figure 1). Each connector has its own advantages and disadvantages. For example, the panel with pin connectors may be constructed conveniently, but its shear deformation between the wythes caused by flexure is often large, which decreases its flexural capacity.

Figure 1.

Configuration of connectors.

The wall panels against flexural demands can be designed as fully composite, non-composite, and partially composite which are determined by the shear transfer capacity or the connector deformation in shear between the two wythes. Fully composite wall panels are stiffer and stronger than ordinal sandwich panels (e.g. the panels using solid concrete connectors or steel welded wire truss connectors), but their thermal performance is poor and may give rise to thermal bridges at the locations of the connectors. Therefore, the solid concrete connectors have not seen its large use in the sandwich wall panels in recent years (Lee and Pessiki, 2007; Pessiki and Mlynarczyk, 2003). Non-composite wall panels with smaller number of connectors substantially reduce the thermal deficiencies created by the connectors, but sacrifice the structural performance so that they can increase the thickness of the wall panels. In addition, fully composite and non-composite wall panels can be analyzed by concrete sectional analysis. But most of the panels are partial composite, and to design a partial composite wall panel is more complicated due to the discontinuity of its strain distribution on the cross section along the thickness of the wall panel. The degree of the composite action has been evaluated by various methods, such as initial stiffness (Choi et al., 2015), strength, deflection (Frankl et al., 2011), effective moment of inertial (Pessiki and Mlynarczyk, 2003), and curvature (Hassan and Rizkalla, 2010). Previous studies on the configuration of shear connectors have seen that continuous connectors can achieve high degree of composite action.

To design a panel with both well structural and thermal performance, many researchers have studied various FRP connectors. The FRP materials generally have high thermal resistance and high tensile strength, but have low shear strength. And they have no obvious yield point which means their stiffness is always constant till damage without warning under axial load. The failure modes of the FRP materials are brittle failure with low ductility. Therefore, it can be found that although in some cases the sandwich wall panels with pure FRP connectors have high ultimate strength or high composite action, their flexural capacities drop rapidly once the external force exceeds the ultimate capacities of the wall panels. The sandwich wall panel with steel truss connectors can obtain full or high degree of composite action (Benayoune et al., 2008). But there is a tendency to create larger thermal bridges through the insulation layer. It has been found that sandwich wall panels with 0.08% of the surface area penetrated with steel connectors may reduce approximately 38% of the thermal efficiency (McCall, 1985; Tomlinson and Fam, 2014). To make full use of the advantages of the materials of GFRP and steel, a steel–glass fiber–reinforced polymer (SGFRP) connector is used in the panels. The SGFRP connector is composed of two different materials that are GFRP and steel bar. The steel bar with a diameter of 4 or 6 mm is in the inner core and is wrapped by a layer of 1- to 3-mm-thick GFRP to form integrity to bear load together. The geometry of the SGFRP connector is discrete W-shaped so as to increase the anchorage capacity compared with pin connector.

This article describes the behavior of 12 full-scale panels through four-point bending tests. The impact of varying number of SGFRP connectors, diameter, and thickness of the insulation layer is investigated in this article. Furthermore, the behavior of wall panels with SGFRP connectors is compared with that of the test specimens which have no foam insulation layer or no shear connectors to analyze the impact of the adhesion bond between the wythes and insulation and solid panel on the composite action. The performance of SGFRP panels is also compared with that of pure GFRP and steel. The level of composite action for all panels is calculated by stiffness and strength methods.

Experimental program

Description of test specimens

There are 12 precast and nonprestressed concrete sandwich wall panels divided into four groups to be designed and tested under flexure (Figure 2). All panels are 3200 mm long and 1700 mm wide. The total thicknesses for all panels range from 150 to 190 mm. The two concrete wythes of all panels have the same thickness of 55 mm with rectangular cross section. The thicknesses of the insulation layers range from 40 to 80 mm to test the partial composed action of the wall panels with different thicknesses and different lag angles in the W-shaped shear connectors. In this study, the insulation for all specimens was made of XPS foam. For flexure, all wall panels were reinforced with D10 steel bar grids (diameter of 10 mm) spaced at 200 mm longitudinally and transversely in the exterior and interior (upper and lower) concrete wythes. The thicknesses of the concrete cover in the two wythes are both 15 mm.

Figure 2.

Wall panel design with W-shaped connectors: (a) general wall panel design with W-shaped connectors except specimens F1-2 and F1-3, (b) specimen F1-2, and (c) specimen F1-3.

There were discrete W-shaped shear connectors spaced at 500 mm longitudinally and 1000 mm transversely in all SGFRP test specimens except specimens F1-2 and F1-3 to connect the exterior and interior wythes and the insulation layer. Only one panel (that is, specimen F1-3) was spaced at 750 mm longitudinally and 1500 mm transversely with merely six W-shaped SGFRP shear connectors, and specimen F1-2 was spaced at 500 mm longitudinally and 1500 mm transversely with eight W-shaped SGFRP shear connectors (Figure 2). In group 2, specimens F2-1, F2-2, and F2-3 have insulation layer thickness of 40, 60, and 80 mm, respectively. The shear connectors of specimens F3-1 and F3-2 are 8 mm diameter with pure GFRP material and 12 mm diameter with 6 mm steel inner core and 3 mm thickness of GFRP, respectively (Figure 3). The concrete cover thickness of the shear connectors is only 10 mm hoping to make maximum use of the anchorage length in the two concrete wythes because the steel bar inner core was wrapped by corrosion-resistant GFRP. Therefore, the W-shaped shear connectors were embedded into each wythe with 45 mm depth. All horizontal lengths of the W-shaped connectors’ legs are 200 mm so as to avoid the contact between the embedded connectors and the steel grids.

Figure 3.

W-shaped shear connectors with different heights and corresponding lag angles and pictures: (a) schematic diagram of SGFRP W-shaped connector, (b) manufacture process of SGFRP bar, and (c) SGFRP W-shaped connector.

There are other four specimens including F4-1, F4-2, F4-3, and F4-4 for comparison, where the shear connectors of F4-1 and F4-2 are commercial products in the market of China that are pure GFRP pin connector and steel truss connector (Figure 4), respectively. The two kinds of connectors were used widely on the market, and the materials and shapes of the connectors were also representative (one was pin connector with GFRP, and the other was truss connector with stainless steel). Specimens F4-3 and F4-4 have no shear connectors. But the difference between specimens F4-3 and F4-4 was that F4-3 was fabricated using solid concrete without the insulation layer to compare with theoretical fully composite action of the sandwich wall panel, and F4-4 only had insulation layer but without any shear connectors to simulate the non-composite action despite the composite behavior provided by adhesion and friction between the concrete wythes and insulation layer. Previous research has proved that the insulation bond and friction have a significant impact on the composite action of the sandwich wall panels (Oh et al., 2013; Woltman et al., 2010). Unfortunately, insulation bond is quite variable and may not be reliable over time due to the thermal and loading circles (Tomlinson and Fam, 2014). All wall panels with different parameters are listed in Table 1, where the data of the connector diameter in the parentheses represent the inner-core steel bar diameter and the wrapped GFRP thickness (e.g. “6 + 4” denotes the cross section of the connector including 6 mm inner-core steel bar and 4 mm wrapped GFRP). Specimen F1-1 was the control test with 12 SGFRP connectors and 50 mm insulation layer.

Figure 4.

Pin connector (GFRP material) and truss connector: (a) GFRP pin connector and (b) stainless steel truss connector.

Table 1.

Test matrix for flexure tests.

No.	Specimen	Shear connector material	Connector shape	No. of connectors	Insulation thickness (mm)	Connector height (mm)	Angle α	Connector diameter (mm)
1	F1-1	SGFRP	W-shape	12	50	140	34°	10 (6 + 4)
2	F1-2	SGFRP	W-shape	8	50	140	34°	10 (6 + 4)
3	F1-3	SGFRP	W-shape	6	50	140	34°	10 (6 + 4)
4	F2-1	SGFRP	W-shape	12	40	130	32°	10 (6 + 4)
5	F2-2	SGFRP	W-shape	12	60	150	39°	10 (6 + 4)
6	F2-3	SGFRP	W-shape	12	80	170	41°	10 (6 + 4)
7	F3-1	SGFRP	W-shape	12	50	140	34°	8
8	F3-2	SGFRP	W-shape	12	50	140	34°	12 (6 + 6)
9	F4-1	Stainless steel	Truss	4	50	110	26°	5
10	F4-2	GFRP	Pin	28	50	125	90°	–
11	F4-3	– (solid panel)	–	–	0	–	–	–
12	F4-4	–	–	–	50	–	–	–

SGFRP: steel–glass fiber–reinforced polymer; GFRP: glass fiber–reinforced polymer.

The R-value of the wall panels was calculated according to Thermal Design Code for Civil Building (GB 50176-1993, 1993)

R_{o} = R_{i} + R + R_{e}

(1)

R = R_{1} + R_{2} + \dots + R_{n}

(2)

where $R_{o}$ is the thermal conductance resistance of the panel; $R_{i}$ and $R_{e}$ are the heat exchange resistance of the interior and exterior surface, respectively; $R$ is the thermal resistance; and $R_{1} + R_{2} + \dots + R_{n}$ is the thermal resistance of each layer in panels. If there are two or more different materials in one layer, $R$ can be calculated as

R = \frac{δ (F_{1} + F_{2} + \dots + F_{n})}{λ_{1} F_{1} + λ_{2} F_{2} + \dots + λ_{n} F_{n}} \times 100

(3)

where $δ$ is the thickness of the layer, $F_{1}, F_{2}, \dots, F_{n}$ is the heat transfer area of each material, and $λ_{1}, λ_{2}, \dots, λ_{n}$ is the thermal conductivity of each material.

The R-value is 1.13 m² k/W for specimen F1-1 with 12 SGFRP shear connectors and 1.23 m² k/W for specimen F1-2 with 8 SGFRP shear connectors. The R-value of the commercial specimens is 1.58 m² k/W for F4-1 with stainless steel connectors and 1.77 m² k/W for F4-2 with pure GFRP connectors, respectively. All R-values of the wall panels meet the requirement of Design Standard for Energy Efficiency of Residential Buildings in Hot Summer and Cold Winter Zone (JGJ 134-2010, 2010) for external walls.

Material properties

The panels were fabricated and cast on three separated dates with the same type of concrete (i.e. the concrete with nominal compressive strength of 30 MPa). The concrete strengths tested by the cylinder on the test date ranged from 31 to 39 MPa. The deformed steel bars for the wire mesh had an average yield tensile strength f_y of 480 MPa, ultimate tensile strength f_u of 610 MPa, and elastic modulus E_s of 200 GPa based on the tension tests. The reinforcement ratios ρ_s of all wall panels (except solid F4-3) were 0.714% without considering the insulation layer. Stainless steel bars for the F5-1 shear connectors with 5 mm diameter had a guaranteed ultimate tensile strength of 600 MPa based on the manufacture data.

The SGFRP bar for the shear connector was made of 6-mm deformed steel bar and then wrapped by 1- to 3-mm GFRP which were bonded together through polyamide resin to bear loads as a composite material (Figure 3). The load–displacement relationship was measured by a designed test system that in consideration of the low transverse strength and stiffness of SGFRP, the ends of the SGFRP bar were fixed by two bond-type anchorages through epoxy resin in two steel pipes. The test setup and load–displacement relationship of SGFRP are exhibited in Figure 5. In addition, an extensometer was used to measure SGFRP strain during the material test. The curve of 6 + 4 shows an obvious second stiffness when the inner-core steel bar was yielded. Subsequently, steel bar worked independently after rupture of the outer wrapped GFRP. The stress–strain relationship of SGFRP can be expressed as follows (Wu et al., 2010)

σ_{sg} = {\begin{matrix} ε_{sg} (E_{s} A_{s} + E_{g} A_{g}) / A 0 \leq ε_{sg} < ε_{sgy} \\ f_{sgy} + (ε_{sg} - ε_{sgy}) E_{g} A_{g} / A ε_{sgy} \leq ε_{sg} \leq ε_{sgu} \\ f_{y} A_{s} / A ε_{sgu} < ε_{sg} \end{matrix}

(4)

where $σ_{sg}$ and $ε_{sg}$ are the stress and strain of SGFRP bar, respectively; $E_{s}$ , $A_{s}$ , and $f_{y}$ are the elastic modulus, area, and yield stress of steel bar, respectively; $E_{g}$ and $A_{g}$ are the elastic modulus and the area of GFRP, respectively; $A$ is the total area of SGFRP bar; $f_{sgy}$ and $ε_{sgy}$ are the yield stress and yield strain of SGFRP bar, respectively; $ε_{sgu}$ is the ultimate strain of SGFRP bar when the rupture of GFRP appears. The other materials do not show the obvious second stiffness due to the small steel bar-to-GFRP area ratios.

Figure 5.

Material property of SGFRP bar: (a) test setup and (b) load–displacement relationship of SGFRP bar.

Fabrication process

The panels were fabricated in a prefabricated component factory. The longitudinal and transverse reinforcement was tied with wire. The EPS foam boards were precut with designed sizes by a machine. The gaps between the cut foam insulation boards were 20 mm to insert the shear connector into the reserved place after casting the lower wythe and placing the upper wythe mesh. The shear connectors were produced in another composite factory. And then, all the gaps were filled with foaming agent; subsequently, the upper wythe was casted. Panels were stripped after 1 day and cured with a period of 30 days. Finally, the panels were shipped to the structural laboratory via truck.

Test setup and loading sequence

All wall panels used the same test setup (Figure 6). Each panel was tested in four-point bending till failure under force and displacement double control using two 500-kN hydraulic jacks in tandem to increase the displacement range of the setup. The loading procedure was divided into two stages: In the first stage, the load was increased by a tick (0.5 MPa) on the jack dial till reaching the ultimate load. In the second stage, the wall panels were under displacement control with 5 mm each time until failure. The wall panels had a span of 3000 mm, a width of 1700 mm, a middle length of 500 mm, and shear spans of 1250 mm. The wall panels were placed on four 150-mm-length, 80-mm-width, and 8-mm-thick steel plates at each panel corner to simulate a four-point supporting in practice. In order to ensure one-way behavior, the load at the central point was distributed by two I steels with sectional dimensions of 180 × 180 × 8 × 8 mm with stiffening ribs across the panel width.

Figure 6.

Experimental test setup.

The deflection values were recorded by various displacement gauges that were placed at mid-span and the ends of the middle region, and there were additional displacement gauges placed at the panel ends to measure the relative slip between the two concrete wythes. The total end slips of the panels were the sum of the absolute value of the two end slips. Half of the connectors in all wall panels with W-shaped connectors were mounted with dozens of 5 mm and 120 Ω electric resistance strain gages on each lag surface of the connectors. Four additional strain gauges were mounted on the longitudinal and transverse reinforcing bars in the concrete wythes at the mid-span. The concrete sectional strains at the 500-mm middle region were measured with hand-held strain gauges, and a total of 12 measuring points were spaced at 250 mm at the top and bottom of each concrete wythe.

Experiment results and discussion

The section presents the experimental results in terms of load–deflection relationship, load–slip relationship, strain distribution along the panel depth, and strain variation at the mid-lags of the W-shape connections. Several key test results of each specimen are summarized in Table 2, including cracking, service, and ultimate load points. Specimen FC50 represents theoretical fully composite panel with 50 mm thickness of insulation layer (FC40 means 40 mm thickness and so forth), the results of which were not taken from the test. Specimen NC represents theoretical non-composite panel with two separated concrete wythes and an insulation layer that works independently. The yielding point is not exhibited in the table due to the high reinforcement ratios in the panels resulting in that the connectors in most wall panels were failure before the reinforcement yield. The service load and deflection in the table were just determined by a constraint condition that the mid-span deflection was limited to L/360 from the live load where L is the span of the panel (Canadian Standards Association (CSA), 2004; Tomlinson and Fam, 2015). Although there was no insulation layer in specimen F4-3, the displacement gauges were still mounted on the corresponding locations of the sandwich panels that were also at the ends of the panel. The total end slips of specimen F2-2 are not shown in this table because the gauges were damaged during the test. The test was terminated at the time when the specimen was failure or the stroke of the two hydraulic jacks was reached, whichever occurred first.

Table 2.

Test results of each specimen at key points.

Specimen	Cracking			Service			Ultimate
Specimen	Load (kN)	Deflection (mm)	Total end slip (mm)	Load (kN)	Deflection L/360 (mm)	Total end slip (mm)	Load (kN)	Deflection (mm)	Total end slip (mm)
F1-1	12.7	2.4	0.08	32.68	8.33	0.46	63	32.84	2.772
F1-2	16	3.43	0.17	34.53	8.33	0.65	59.286	25.76	2.049
F1-3	12.7	3.21	0.91	27.85	8.33	0.75	45.714	19.06	0.998
F2-1	18.5	2.41	1.33	40.57	8.33	0.35	79	33.94	3.42
F2-2	29	3.58	–	48.21	8.33	–	64	23.82	–
F2-3	22	3.03	0.89	35.33	8.33	0.92	64.8	35.71	3.23
F3-1	16	3.74	0.5	30.96	8.33	0.28	69.175	50.58	2.7
F3-2	16	2.65	0.69	35.89	8.33	1.11	62.807	34.39	3
F4-1	16	5.6	0.21	25.01	8.33	0.19	52.143	82.77	1.975
F4-2	12.7	3.8	0.03	18.33	8.33	0.15	37.117	77.53	2.825
F4-3	36	2.73	1.25	57.75	8.33	0.77	118	82.87	9.43
F4-4	18.5	5.48	2.54	22.79	8.33	2.85	32.5	106.61	12.76
FC40	34.06	1.33	–	–	–	–	86.78	–	–
FC50	39.02	1.25	–	–	–	–	94.02	–	–
FC60	44.28	1.18	–	–	–	–	101.25	–	–
FC80	55.72	1.05	–	–	–	–	115.72	–	–
NC	9.51	7.24	–	–	–	–	28.87	–	–

Load–deflection relationship

The load–deflection curves for all wall panels at mid-span are shown in Figure 7. The calculated, theoretical fully composite response is also plotted in the figure using dash line as the upper bound, and the test specimen F4-4 is plotted as the lower bound. Although there are bond strengths in specimen F4-4 between the concrete wythes and insulation layer, the insulation bond is quite variable and may not be reliable over time due to the thermal and loading circles. In general, all the other panels except the solid specimen F4-3 fall in the upper and lower bounds.

Figure 7.

Load–deflection relationships for all test specimens at mid-span: (a) varying number of connectors, (b) varying insulation layer thickness of panels, (c) varying connector diameter, and (d) steel and pin connectors.

The sandwich wall panels with GFRP W-shaped shear connectors had high initial stiffness and strength except specimen F1-3 with merely six shear connectors. The layout of shear connectors in the panel had a great effect on the strength for the panels under flexure. The requirement for the shear transfer in the vicinity of panel ends is higher than that at the mid-span region with constant moment. Therefore, rational arrangement of shear connectors was necessary for panels under flexure. Yielding generally occurred after the failure of the outmost shear connectors at the end of the panels accompanied with noise, and the ultimate load was reached around the same time. Figure 7 indicates that the partial composite behavior contributes considerable flexure strength to the wall panels with SGFRP W-shaped shear connectors. In group 1, the ultimate loads increased 1.94, 1.82, and 1.41 times (specimens F1-1–3 versus F4-4, respectively); in group 2, the ultimate loads increased 2.43, 1.97, and 1.99 times (specimens F2-1–3 versus F4-4, respectively; although the insulation thicknesses of the panels in group 2 were different, the non-composite action behavior was usually simulated by two single-concrete wythes); in group 3, the ultimate loads increased 2.13 and 1.93 times (specimens F3-1 and 2 versus F4-4, respectively).

For panels with different insulation layer thicknesses, specimen F2-1 with 40 mm thickness of insulation layer had a fantastic ultimate load. This may be explained by the fact that smaller insulation thickness and angle α reduced shear deformation between the two concrete wythes and then enhanced the composite behavior. But a fine bond behavior between the connector and concrete was required to ensure that connector strength was fully utilized, which was greatly influenced by the construction quality with uneven wythe thickness. For panels with different connector diameters and materials in group 3, specimen F3-1 with pure GFRP connectors and diameter of 8 mm had higher ultimate load than F3-2 but lower stiffness, and then the load decreased rapidly. The small stiffness was due to the small elastic modulus of GFRP material. Figure 7(d) shows the load–deflection responses of the control specimens F4-1 and F4-2 with 5-mm-diameter steel truss and pure GFRP pin connectors. Both the specimens had low stiffness and ultimate load relative to F1-1. F4-2 with GFRP pin connectors almost had no partial composite action and its load–deflection shape was similar to F4-4. The calculated maximum load of fully composite panel with 50 mm thickness insulation was approximately 80% of the solid specimen F4-3.

Load–slip relationship

The load–slip responses of all test wall panels are shown in Figure 8. The slips were measured between the exterior and interior concrete wythes at the east and west ends of the panels and were divided into dominant and non-dominant slips representing higher and lower slips, respectively. In most panels, end slips were generally increased with the cumulative load and deflection at mid-span. The increased dominant slip reflects the degree of the partial composite action. The cracking of the concrete wythes, loss of the bond strength between the wythes and insulation, and failure of shear connectors will lead to excessive increase in end slip. The yielding or failure of shear connectors made the flexural bearing capacity of the panels decrease and prevent fully composite behavior of the panels from being developed.

Figure 8.

Load–slip relationships for all test specimens at the end: (a) varying number of connectors, (b) varying insulation layer thickness of panels, (c) varying connector diameter, and (d) steel and pin connectors.

For the ultimate load being reached, the connectors were failure at only one side resulting in great different slips between the two sides of the panels. The steel truss connectors in specimen F4-1 did not fail during the test because the connectors were welded with ordinal steel bars before placing into the concrete wythes longitudinally at the mid-thickness. Therefore, the bond behavior is pretty good and the bond failure did not happen during the test. Its end slip was also lower than panels with GFRP connectors. But its flexural capacity was lower due to the small cross-sectional area in the shear connectors leading to lower shear stiffness in transferring the shear caused by flexure. Some pictures of the wall panels taken during and after the test are shown in Figure 9.

Figure 9.

Specimen photographs taken during and after test: (a) end slip after the test of F1-1, (b) delaminating between wythes and insulation of F1-2, (c) F1-3 during the test, (d) failed connector in F2-1, (e) cracking layout at the bottom of F2-2, (f) side view of F2-3 after the test, (g) F3-1 during the test, (h) F3-2 after the test, and (i) deflection shape of F4-1.

Strain distribution along the cross section at the middle regions of the panels

Strain distributions for specimens F2-1–3, F3-1 and 2, and F4-4 along the cross section at the middle region are shown in Figures 10 and 11 at different stages, including cracking, service, ultimate load, and ultimate displacement stages. The strain distribution reflects the degree of the partial composite action intuitively. The middle region was divided into east and west middle regions placed at symmetrical position, and their boundary was at mid-span. One was dominant region, and the other one was non-dominant region like the end slips. Due to the localized effects caused by nonuniform cracking distribution in the concrete wythes of the panels, the strain distribution had large discontinuities between concrete wythes. In addition, the force redistribution caused by the failure of shear connectors on one side will also result in different stiffness and strains between the east and west middle regions. The two plane sections did not remain planes at the ultimate load stage in most panels.

Figure 10.

Test strain distributions of selected specimens at dominant region: (a) specimen F2-1, (b) specimen F2-2, (c) specimen F2-3, (d) specimen F3-1, (e) specimen F3-2, and (f) specimen F4-4.

Figure 11.

Test strain distributions of selected specimens at non-dominant region: (a) specimen F2-1 and (b) specimen F3-1.

In group 2, specimen F2-1 with 40 mm thickness insulation indicated a well strain distribution that the concrete at the bottom of the upper wythe and the whole lower wythe was under tension during the test. The plane section remained almost plane before ultimate meaning that the effectiveness of connectors with smaller insulation thickness and lag angle of the connectors in transferring the shear was increased relative to specimens F2-2 and F2-3 at least in this experiment. In most specimens, the two concrete wythes worked independently at the ultimate displacement stage. For specimen F4-4 with no connectors, the upper and lower concrete wythes had almost identical strain distribution. The adhesion bond between the wythes and insulation layer may have little contribution to the partial composite action for wall panels without shear connectors. The concrete strain at the top of the upper wythe was lower than the ultimate compressive strain of the concrete. Therefore, no concrete was crushed except for the concrete at the loading points. The longitudinal reinforcement in the mid-thickness of the lower wythe in most specimens just yield at the ultimate displacement, which had little effect on the structure behavior.

Strain in SGFRP connectors

The strain gauge layouts and selected load–strain profiles in the SGFRP W-shaped connectors are shown in Figures 12 and 13, respectively. The electric resistance strain gauges were mounted on the middle of the connector lags axially. The strains in most connectors varied greatly and irregularly during the test especially at the ultimate stage. The failure of SGFRP connectors lead to force redistribution in the connectors. The connector strains in specimens F2-1 and F3-2 were obviously lower than that in specimen F3-1 due to the connectors with larger diameter and elastic modulus. Therefore, most of the connectors were damaged by being pulled out of the concrete wythes before reaching the connector ultimate strength except the small diameter connectors. It is worth mentioning that not all the outmost lags (gauge S1) of the connectors were subjected to tension during the test (Figure 13(e)).

Figure 12.

Strain gauge layout in the W-shaped connectors: (a) strain gauge layout of all specimens with SGFRP connectors except F1-2 and F1-3 and (b) strain gauge layout of F1-2 and F1-3.

Figure 13.

Typical strain in W-shaped shear connectors in selected specimens: (a) F1-2 (gauges S5–S8), (b) F2-1 (gauges S1–S4), (c) F3-1 (gauges S1–S4), (d) F3-1 (gauges S5–S8), and (e) F3-2 (gauges S1–S4).

Assessment of degree of partial composite action

The degree of partial composite action for each panel was evaluated using stiffness method at cracking load and strength method at ultimate load. The theoretical stiffness of the panel can be calculated using equation (5)

D = E_{c} I = \frac{Fa}{48 Δ} (3 L^{2} - 4 a^{2})

(5)

where $E_{c}$ is the elastic modulus of the concrete, $I$ is the moment of inertia of the cross section, $F$ is the total acting load on the panels at the load point at panel mid-span, $Δ$ is the mid-span deflection, $a$ is the shear span length, and $L$ is the panel span length.

To calculate the theoretical deflection beyond cracking accurately, the effective moment of inertia $I_{eff}$ was used and calculated as follows ( Building Code Requirements for Structural Concrete, ACI 318-08 and Commentary, ACI 318R-08 )

I_{eff} = {(\frac{M_{cr}}{M_{a}})}^{3} I_{g} + [1 - {(\frac{M_{cr}}{M_{a}})}^{3}] I_{cr}

(6)

where $M_{cr}$ is the cracking moment, $M_{a}$ is the acting moment, $I_{cr}$ is the cracked transformed moment of inertia of the concrete section, and $I_{g}$ is the gross moment of inertia of the concrete section. $I_{g}$ and $M_{cr}$ can be expressed as follows

I_{g} = \frac{1}{12} b (h_{g}^{3} - h_{ins}^{3})

(7)

M_{cr} = \frac{f_{r} I_{g}}{y_{t}}

(8)

where $b$ is the panel width, $h_{g}$ is the total thickness of the panel, $h_{ins}$ is the insulation layer thickness, $f_{r}$ is the modulus of rupture of concrete, and $y_{t}$ is the distance from the centroidal axis of the gross section to the tension face.

For each specimen, the degree of the composite action can be evaluated in terms of stiffness of the panel at the cracking load $k_{D}$ expressed as

k_{D} = \frac{D_{\exp} - D_{NC}}{D_{FC} - D_{NC}} \times 100

(9)

where $D_{\exp}$ is the experimental cracking or service stiffness using measured deflection at mid-span, $D_{FC}$ is the theoretical uncracked fully composite stiffness, and $D_{NC}$ is the uncracked theoretical non-composite stiffness.

The degree of the composite action can also be evaluated based on the strength $k$ ( $k_{s}$ for service strength and $k_{u}$ for ultimate strength) for each specimen at the two load points as follows

k = \frac{F_{\exp} - F_{NC}}{F_{FC} - F_{NC}} \times 100

(10)

where $F_{\exp}$ is the measured load of the specimens, $F_{FC}$ is the theoretical composite ultimate strength, and $F_{NC}$ is the theoretical non-composite ultimate strength that is calculated by adding the two separated wythes’ strength. The degree of the composite action in terms of the stiffness and the ultimate strength is summarized in Table 3.

Table 3.

Degree of composite action by stiffness and strength methods for all panels.

Specimen	Composite action by stiffness method				Composite action by strength method
Specimen	Cracking load (kN)	Deflection (mm)	D (10⁹ N mm²)	k_D (%)	Ultimate load (kN)	k_u (%)
F1-1	12.7	2.4	2.86	13	63	52
F1-2	16	3.43	2.52	11	59.29	47
F1-3	12.7	3.21	2.14	9	45.71	26
F2-1	18.5	2.41	4.15	26	79	86
F2-2	29	3.58	4.38	23	64	49
F2-3	22	3.03	3.92	16	64.8	41
F3-1	16	3.74	2.31	6	69.18	62
F3-2	16	2.65	3.26	16	62.81	52
F4-1	16	5.6	1.54	5	52.14	36
F4-2	12.7	3.8	1.81	7	37.12	13
F4-3	35	1.2	15.76	93	118	137
F4-4	18.5	5.48	1.82	7	32.5	6
FC40	34.06	1.33	13.81	100	86.78	100
FC50	39.02	1.25	16.88	100	94.02	100
FC60	44.28	1.18	20.35	100	101.25	100
FC80	55.72	1.05	28.62	100	115.72	100
NC	9.51	7.24	0.71	0	28.87	0

In Table 3, it can be found that the degree of composite action in terms of stiffness was generally lower than that seen by the strength method for the sandwich specimens whether using SGFRP W-shaped connectors, steel truss, or pin connectors. But it can still reflect the level of the composite action to certain extent. For specimens F1-3, F3-1, F4-1, F4-2, and F4-4, their values of $k_{D}$ were all less than 10%. Therefore, smaller quantity of connectors, pure GFRP connector, and smaller diameter steel truss connector will give rise to low initial rigidity for the sandwich panels in transforming the shear caused by flexure. The adhesion bond between wythes and insulation also gave approximately 7% of the composite action for the panel (specimen F4-4). There was a difference of approximately 7% in stiffness between the test solid specimen F4-3 and FC50 although their inner layer materials were different. This was also, in part, attributed to the variance between the theoretical value and test value. For evaluating the degree of composite action by the strength method, F2-1 exhibited the highest degree in all test sandwich panels, and then followed by F3-1, F1-1, F3-1, and so on.

From the test results and theoretical composite and non-composite prediction, high level of composite action may be greatly affected by the stiffness and strength of the shear connectors, and panel thickness especially the insulation layer thickness in which the connectors are embedded into it, accompanied with appropriate longitudinal reinforcement ratio. Ideally, the shear connectors failed after the yielding of the reinforcement bar. Although the inner core of the connector with steel bar (SGFRP) will not improve the ultimate load of the panels under flexure, it can enhance the stiffness of the panel and prevent the bearing capacity of the panels from dropping rapidly after the ultimate.

Conclusion

In this study, the flexural behavior of the precast insulated concrete sandwich panel with W-shaped SGFRP connectors was investigated under flexure by analyzing the impact of varying the number of connectors, insulation layer thickness, the connector’s lag angle, and the connector diameter. The load–deflection relationships, load–slip relationships, concrete strain profiles along the wythes cross section, and the strains in the SGFRP W-shaped connector were measured and analyzed in this article. Four other panels were tested for comparison that were panels using stainless steel truss connectors, pure GFRP pin connectors, and no connectors, respectively. In order to obtain the level of composite action, the test results were compared with the values of theoretical fully composite and non-composite panels in terms of the strength and the stiffness. The conclusions were as follows:

For evaluating the degree of composite action by the strength method, specimen F2-1 showed the highest degree of 86% in all test sandwich panels. The other sandwich panel values ranged from 26% to 62%. The ultimate loads increased 1.94, 1.82, and 1.41 times in group 1; 2.43, 1.97, and 1.99 times in group 2; and 2.13 and 1.93 times in group 3 and all of them were compared with specimen F4-4. Smaller insulation thickness and lag angle α reduced shear deformation between the two concrete wythes and then enhanced the composite behavior. This method can be used for ultimate limit state design of the sandwich panels.

When evaluating by stiffness method, the degrees of specimens F1-3, F3-1, F4-1, and F4-4 were all less than 10%, while other panels ranged from 11% to 26%. Lower number of connectors, pure GFRP W-shaped connector, and smaller diameter steel truss connector will lead to low initial rigidity for the sandwich panels in transforming the shear produced by flexure. This method can be used for checking serviceability limit state of the sandwich panels.

In group 3 with different diameter connectors, the increasing W-shaped connector diameter did not bring too much bearing capacity for the panels. Inversely, F3-1 with pure GFRP connectors and diameter of 8 mm had higher ultimate load than F3-2 with 12 mm diameter, but had lower stiffness due to the small elastic modulus of GFRP material, and then decreased rapidly after failure of the connectors. Most of the connectors with large diameter were damaged by pulling out from the concrete wythes without reaching the connector ultimate strength. Therefore, reasonable diameter in the connectors was required to fully utilize its strength. However, SGFRP connectors with inner core of steel bar enhance the panel stiffness and make the load not drop rapidly after the ultimate that can give warnings of failure.

For the compared specimens, the adhesion bond between wythes and insulation in specimen F4-4 without the connectors attributed to approximately 7% and 6% of the composite action by stiffness and strength methods, respectively. The adhesion bond failed after the concrete cracking at the bottom of the lower wythe. Therefore, it is suggested not to consider the adhesion bond strength in design due to its unreliability.

The flexural capacity of sandwich panel was governed by stiffness and strength of the shear connectors, and panel thickness especially the insulation layer thickness, and appropriate longitudinal reinforcement ratio. The panels with SGFRP W-shaped connectors could provide sufficient strength under flexure with reasonable design.

Footnotes

Appendix 1 Acknowledgements

The authors would like to thank the research group members of Jiang Huanzhi and Liu Hui who helped to conduct the experiments.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National “Twelfth Five-Year” Plan for Science & Technology (no. 2011BAJ10B03), Fundamental Research Funds for the Central Universities and Graduate Student Scientific Research Innovation Projects of Jiangsu Province University (KYLX15_0079), and Guide Funds for the Modernization of Science & Technology Building Industry.

References

American Concrete Institute (ACI) 318-08 (2008) Building Code Requirements for Structural Concrete (ACI 318-08) and Commentary. Farmington Hills, MI: American Concrete Institute.

Benayoune

Samad

AAA

Trikha

et al . (2008) Flexural behaviour of pre-cast concrete sandwich composite panel—experimental and theoretical investigations. Construction and Building Materials 22(4): 580–592.

Canadian Standards Association (CSA) (2004) Design of Concrete Structures (CAN/CSA-A23.3-04). Toronto, ON, Canada: CSA.

Choi

Kim

H-R

(2015) Composite behavior of insulated concrete sandwich wall panels subjected to wind pressure and suction. Materials 8(3): 1264–1282.

Einea

Salmon

Tadros

et al . (1994) A new structurally and thermally efficient precast sandwich panel system. PCI Journal 39: 90–101.

Frankl

Lucier

Hassan

et al . (2011) Behavior of precast, prestressed concrete sandwich wall panels reinforced with CFRP shear grid. PCI Journal 56: 42–54.

GB 50176-1993 (1993) Thermal Design Code for Civil Building. Beijing, China: China Standard Publishing House (in Chinese).

Hassan

Rizkalla

(2010) Analysis and design guidelines of precast, prestressed concrete, composite load-bearing sandwich wall panels reinforced with CFRP grid. PCI Journal 55(2): 147–162.

JGJ 134-2010 (2010) Design Standard for Energy Efficiency of Residential Buildings in Hot Summer and Cold Winter Zone. Beijing, China: China Architecture & Building Press (in Chinese).

10.

Kim

You

Y-C

(2015) Composite behavior of a novel insulated concrete sandwich wall panel reinforced with GFRP shear grids: effects of insulation types. Materials 8(3): 899–913.

11.

Lee

Pessiki

(2007) Design and analysis of precast, prestressed concrete, three—wythe sandwich wall panels. PCI Journal 52(4): 70–83.

12.

McCall

(1985) Thermal properties of sandwich panels. Concrete International 7(1): 35–41.

13.

Maximos

Pong

Tadros

et al . (2007) Behavior and Design of Composite Precast Prestressed Concrete Sandwich Panels with NU-Tie. Lincoln, NE: University of Nebraska–Lincoln.

14.

Naito

Hoemann

Bewick

et al . (2009) Evaluation of shear tie connectors for use in insulated concrete sandwich panels. Air Force Research Laboratory report, AFRL-RX-TY-TR-2009-4600. Available at: http://www.dtic.mil/dtic/tr/fulltext/u2/a516054.pdf

15.

Jang

Lee

et al . (2013) Insulation type effect on the direct shear behavior of concrete sandwich panel (CSP) with non-shear connectors. Advanced Materials Research 663: 154–158.

16.

PCI Committee on Precast Concrete Sandwich Panels (1997) State-of-the-art of precast/prestressed sandwich wall panels. PCI Journal 42(2): 1–61.

17.

PCI Committee on Precast Concrete Sandwich Panels (2011) State of the art of precast/prestressed concrete sandwich wall panels. PCI Journal 56(2): 131–176.

18.

Pessiki

Mlynarczyk

(2003) Experimental evaluation of the composite behavior of precast concrete sandwich wall panels. PCI Journal 48(2): 54–71.

19.

Tomlinson

Fam

(2014) Experimental investigation of precast concrete insulated sandwich panels with glass fiber-reinforced polymer shear connectors. ACI Structural Journal 111(3): 595–606.

20.

Tomlinson

Fam

(2015) Flexural behavior of precast concrete sandwich wall panels with basalt FRP and steel reinforcement. PCI Journal 60: 51–71.

21.

Woltman

Tomlinson

Fam

(2010) A comparative study of various FRP shear connectors for sandwich concrete walls. In: Proceedings of the FRP composite in civil engineering, Beijing, China, 27–29 September, pp. 237–240. Berlin: Springer.

22.

Luo

et al . (2010) Experimental and theoretical studies on the mechanical properties of steel-FRP composite bars properties of steel-FRP composite bars. China Civil Engineering Journal 43(2): 53–61 (in Chinese).

Experimental evaluation of precast concrete sandwich wall panels with steel–glass fiber–reinforced polymer shear connectors

Abstract

Keywords

Introduction

Experimental program

Description of test specimens

Material properties

Fabrication process

Test setup and loading sequence

Experiment results and discussion

Load–deflection relationship

Load–slip relationship

Strain distribution along the cross section at the middle regions of the panels

Strain in SGFRP connectors

Assessment of degree of partial composite action

Conclusion

Footnotes

Appendix 1

Acknowledgements

Declaration of Conflicting Interests

Funding

References