Asia-Pacific Conference on FRP in Structures (APFIS 2007) S.T. Smith (ed) © 2007 International Institute for FRP in Construction

APPLICATIONS OF FRP IN CLADDING PANELS OF BUILDINGS T. Sharaf 1, A. Fam 1*, W. Shawkat 1 and B. Taylor 2 1

Department of Civil Engineering, Queen’s University, Kingston, On., Canada. Email: [email protected] 2 Cladding Panels Manufacturing Consultant, On., Canada

ABSTRACT The structural performance of a new cladding panel consisting of a polyurethane foam core sandwiched between two layers of glass fibre-reinforced polymer (GFRP) skins has been examied. Ten panels with two types of foam core of different densities were tested in one-way bending, using different loading configurations, namely: threepoint bending, four-point bending, a uniform load, and a reversed cyclic load. The study also examined the consistency and repeatability of test results of similar panels of the same foam type. The flexural strengths, strains, relative slip between the two GFRP skins, and various failure modes have been evaluated. It was shown that the foam core density has a significant effect on flexural strength and stiffness. Also, the failure mode was dependent on the loading configuration. KEYWORDS Cladding, Composites, Sandwich, FRP, Polyurethane Foam, Walls. INTRODUCTION Conventional cladding walls are typically made of two layers of precast reinforced concrete with an insolating foam layer in between. These cladding panels are an essential component of any low rise or high rise building. They serve many functions, including protection against the elements and insulation and are mainly designed for wind loading. Sandwich composite panels on the other hand are typically composed of a high density, lightweight foam core sandwiched between two thin layers of fiber reinforced polymer (FRP) laminates, as shown in Figure 1, and the two materials are bonded using adhesives such as epoxy. The FRP skin carries the inplane compressive and tensile stresses resulting from bending moment, while the main function of the foam core is to keep the two FRP skins apart at the desired distance and to resist shear. Sandwich panels were introduced in the 20th century and were first used in aircraft industry (Allen, 1969). Since then, the main use of sandwich panels has been in the aerospace, automotive and marine industries, especially after introducing FRP materials. Allen (1969) assumed that there is no contribution of the foam core to the overall bending stiffness of the sandwich panel cross section, which was described before as an antiplane core by Filon (1937) Ojalvo (1977) ignored the peeling stresses between the skins and the core, while assumed different deflections of the face skins. Others made the assumption that sandwich panels with foam core act like an ordinary beam with equivalent properties (Ogorkiewicz and Sayigh, 1973). Frosting and Baruch (1990) investigated a sandwich beam, taking into consideration the flexibility of the core in the transverse direction, particularly the compressibility under the applied loads. Equations for predicting deflection, normal stresses in skins, and core shear stresses were developed by Allen (1969), without considering the core flexibility. Frostig and Baruch (1990) developed the governing differential equations for these engineering properties, but without giving a closed form equations. A high order bending theory was developed by Frostig et al (1992) to evaluate the governing deferential equation to predict the bending behavior of a sandwich beam with flexible core. Niu and Talreja (1999) presented a wrinkling model for sandwich panels in compression with the assumption of a continous isotropic linear elastic core. A closed form high order theory for the analysis of sandwich panels with a core made of a material characterized by a nonlinear conistitutive relation was introduced by Schwarts-Givli and Frostig (2001). It is proposed to use FRP-Foam sandwich panels for cladding applications in buildings as replacement of the precast concrete panels, due to their substantially smaller dead weight, and excellent durability and insulation characteristics. This paper presents an experimental flexural investigation of these panels.

431

FRP Top Skin

Foam Core FRP Bottom Skin

y

x z

y

x z

y

z

x

Figure 1. Typical sandwich panel configuration used in this study EXPERIMENTAL PROGRAM Ten panels, 1400 x 300 x 76 mm each, were tested to study the overall behaviour of sandwich panels in one way bending, using different loading configurations. The following sections provide details of the experimental program. Material Properties The sandwich panels consisted of polyurethane foam core sandwiched between two GFRP face skins. The face skins were fabricated from 54 oz 2 weave E-glass and CoPoxy 4281A resin, and had a thickness of 1.6 mm. The tensile modulii in both longitudinal and transverse directions of the GFRP skin were both 20.7 GPa and the shear modulus was 2.07 GPa. The ultimate tensile strength and strain in both directions were 290 MPa and 0.014, respectively. The foam core was fabricated from Corafoam U020. Two densities of the closed-cell polyurethane foam were considered in this study, namely: 0.31 kN/m3 (2 pcf) and 0.63 kN/m3 (4 pcf). The polyurethane foam material properties provided by manufacturer are summarized in Table 1.

U020 (0.3142 kN/m3) (2 pcf) U020 (0.6283 kN/m3) (4 pcf)

Table 1. Polyurethane foam mechanical properties Compressive Strength (kPa) Tensile Strength (kPa) Parallel to Perpendicular Parallel to Perpendicular rise to rise rise to rise

Shear Strength (kPa)

Shear modulus (kPa)

283

145

441

352

159

230

640

483

896

752

600

408

Test Specimens Table 2 provides a summary of the test specimens and parameters. Specimens P1 to P6 were fabricated from the 0.31 kN/m3 (2 pcf) density foam while specimens P7 to P10 were fabricated from the 0.63 kN/m3 (4 pcf) density foam. Specimens P1 and P2 were tested in 3-point and 4-point bending, respectively, while all other specimens were tested under uniform distributed load simulated by eight concentrated loads. Figure 2 shows the different loading configurations. All specimens, except P6 and P10 were tested monotonically at a rate of 1 mm/min. And for the cyclic loading, four loading levels were considered, four cycles were applied at each loading level. The specimen was then flipped over and then the process was repeated. At the end the specimen was monotonically loaded to failure. Specimens (P3 to P6) and (P7 to P9) were identical in each group, to confirm for repeatability of performance and test results. Test Setup and Instrumentation All specimens had a span of 1400 mm and loads were applied across the full width of the panel, using steel rollers resting on rigid steel plates.

APFIS 2007

432

Teflon sheets were placed under the steel plates in order to protect the panels from stress concentrations. The specimens were also supported along the full width at both ends. Longitudinal strains in the GFRP skins were monitored at mid span using four electric resistance strain gages, two on the top face (compression) and two on the bottom face (tension). Mid-span deflections were monitored using two linear potentiometers. The relative slip between the upper and lower skins was measured using two linear potentiometers, placed at both ends. The specimens were tested using a 900 kN Riehle machine. Table 2. Summary of test specimen results Specimen Density ID (pcf) P1 2 P2 2 P3 2 P4 2 P5 2 P6 2 P7 4 P8 4 P9 4 P10 4 R = Repeatability

Loading type

Parameters

3-Point Load 4-Point Load Uniform R Uniform R Uniform R Uniform (Cyclic) C Uniform R Uniform R Uniform R Uniform (Cyclic) C C = Cyclic Loading

L L D D D D D D

L L

Mu (N.m)

Pu (KN)

889.7 2.542 1064 4.56 1195.95 6.834 1295.175 7.401 1275.4 7.288 1413.125 8.075 3156.125 18.035 3019.975 17.257 3401.65 19.438 3068.8 17.536 D = Density Effect

δu (mm)

Failure Mode

27.78 Wrinkling(Inward) 31 Wrinkling(Inward) 61.03 Shear 65.96 Shear 64.91 Shear 68.74 Shear 61.5 Shear 58.72 Shear 65.94 Shear 40 Wrinkling+Shear L = Loading Configuration

EXPERMENTAL RESULTS AND DISCUSSION Table 2 summarize the results for the tested panels of different foam and loading types, and the failure modes observed. Figure 3 shows the flexural behaviour of the panels in terms of load-deflection responses for the effects of both; the foam density and the loading type, namely static and cyclic. Only the last loading path to failure is shown for the cyclic case. The load-longitudinal strain responses are presented for both the 2 pcf and 4 pcf foam types, in Figure 4. Moment-curvature responses are presented for panels with different loading configurations in Figure 5. The load-slip responses are presented in Figure 6. Finally, the different failure modes are shown in Figure 7. The effect of various parameters will be discussed in light of these figures.

(b)

(a)

(c) Figure 2. Loading configurations: (a) Three-point bending, (b) Four-point bending, and (c) Uniform and cyclic load. As shown in Figure 3(a), the ultimate load carrying capacity of the sandwich panel increases as the foam core density increases. The figure clearly shows the significant effect of foam density on strength and stiffness of sandwich panels and that the effect of foam core on flexural behaviour can’t be neglected. Figure 3(b) shows that

APFIS 2007

433

the low cycle fatigue had an insignificant effect on the behaviour of the panel. However, high cycle fatigue still needs to be studied. Figure 4 suggests that the strains in the GFRP skins at failure are substantially lower than the ultimate strain of GFRP (0.014). This is attributed to the failure modes. Specimens P1 and P2 failed due to inward wrinkling of the compression skin, as shown in Figure 7(a) while specimens P3 to P9 failed in shear within the foam core, as shown in Figure 7(b). Figure 6 shows that the panel has the highest moment capacity when it is subjected to a uniform load. It is also noticed that the effect of inward wrinkling of specimen P1 forced the panel curvature to increase suddenly in a nonlinear manner, instead of being linear as in specimens P2 and P5. Figure 6 shows the slip at both ends of specimens P5 and P7 of the two different foam densities. It is noticed that the average slip of specimen P5 is higher than that of P7 due to the lower foam density. 25.0 4 pcf foam

20.0 15.0

P8 2 pcf foam

10.0

P4,P5

5.0

40

60

2 pcf foam

P6 (cyclic) P5 (static)

5.0

0.0 20

P7 (static)

10.0

P3

0

P10 (cyclic)

15.0

P7 Load (KN)

Load (KN)

4 pcf foam

20.0 P9

0.0

80

0

20

Deflection (mm)

(a) Foam density

40 Deflection (mm)

60

80

(b) Static vs. Cyclic loading

Figure 3. Effects of Foam Type and Loading Type

P3

6 5

P2

4 3

P1

-4.E-03

-3.E-03

P9

P3

P7 P8

15

P2

10

2 0 -2.E-03 -1.E-03 0.E+00 Strain (mm/mm)

1.E-03

P9 P7

5

P1

1

P8

20

P4

Load (KN)

P5

25

P5

7

Load (KN)

8

P4

2.E-03

-6.E-03

-4.E-03

0 -2.E-03 0.E+00 2.E-03 Strain (mm/mm)

4.E-03

(b) 4 pcf Foam Type

(a) 2 pcf Foam Type

Figure 4. Top and Bottom Longitudinal Strains 1400

Moment (N.m)

1200

P5

1000 800

P1

P2

600 400 200 0 0.00000

0.00005

0.00010 0.00015 Curvature (mm-1)

0.00020

0.00025

Figure 5. Effect of Loading Configuration on Moment-Curvature Relationship

APFIS 2007

434

6.E-03

20.0 P7 - right end

Load (KN)

15.0

P7 - left end

10.0

P5 - right end

P5 - left end

5.0

0.0 0

1

2

3

4

5

Slip (mm)

Figure 6. Effect of Foam Type on Panels’ End Slip

(a)

(b)

Figure 7. Failure Modes (a) Inward Wrinkling and (b) Foam Core Shear CONCLUSIONS This study has shown that composite panels manufactured using foam core sandwiched between two adhesively bonded GFRP skins have a great potential as flexural members in structural applications. It was noticed that the structural performance is quite sensitive to both the loading configuration and density of the foam core. The following conclusions are drawn: 1. Flexural strength and stiffness of the panels increase substantially as the density of the foam core increases. 2. Significant relative slip occurs between the upper and lower GFRP skins as a result of shear deformation of the foam core. This slip reduces, however, as the foam density increases. 3. The failure mode under concentrated loads is inward wrinkling of the compression GFRP skin, while under uniform loads, shear failure occurs in the foam core. 4. The strains measured in the GFRP skins at failure are well below the ultimate values. ACKNOWLEDGMENTS The authors wish to aknowledge financial support provided by Res-Precast Inc and Materials and Manufacturing Ontario (MMO). REFERENCES Allen, H.G. (1969). Analysis and design of structural sandwich panels. Pergamon Press, Oxford, London, England. Frostig, Y., Baurch, M. (1990). “Bending of sandwich beams with transversely flexible core.” AIAA J., Vol. 28 (3) pp. 523-31. Frostig, Y., Baurch, M., Vilnay, O. & Sheinman, I. (1992). “A high order theory for the behavior of sandwich beams and a flexible core.” ASCE J., EM Div., Vol. 118, No. 5, pp. 1-16

APFIS 2007

435

Filon L.N.G. (1937). “On antiplan stress in an elastic solid.” Proc. Roy. Soc. A, 160, pp. 137-54. Niu, K., and Talreja, R. (1999).“Modeling of wrinkling in sandwich panels under compression.” Journal of Engineering Mechanics Vol. 125, No. 8, 875-883. Ogorkiewicz, R.M. and Sayigh, A.A.M. (1973). “Deflection of carbon fiber/acrylic foam sandwich beams.” Composites, 254-257. Ojavlo, I.V. (1977). “Departure from Classical beam theory in laminated sandwich and short beams.” AIAA J., 15(10), 1518-1521. Schwarts-Givli, H. and Frostig, Y. (2001). “High-order behavior of sandwich panels with a bilinear transverselyflixible core.” Journal of Composite Structures 53, 87-106.

APFIS 2007

436