- Email: [email protected]
Shear behaviour of glued structural fibre composite sandwich beams
Construction and Building Materials 47 (2013) 1317–1327
Contents lists available at SciVerse ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Shear behaviour of glued structural fibre composite sandwich beams A.C. Manalo, T. Aravinthan ⇑, W. Karunasena Centre of Excellence in Engineered Fibre Composites (CEEFC), Faculty of Engineering and Surveying, University of Southern Queensland, Toowoomba, Queensland 4350, Australia
h i g h l i g h t s Structural beams were made from glue-laminated fibre composite sandwich panels. Glued sandwich beams were tested under asymmetrical shear. Beams with edgewise layers are twice stronger than beams with flatwise layers. Beams with edgewise layers exhibited a progressive failure behaviour. Theoretical evaluations were in good agreement with the experimental results.
a r t i c l e
i n f o
Article history: Received 27 July 2010 Received in revised form 10 May 2013 Accepted 17 June 2013 Available online 12 July 2013 Keywords: Glued sandwich beams Shear Structural composite sandwich panels Fibre composites Flatwise Edgewise
a b s t r a c t The shear behaviour of structural sandwich beams made by gluing together fibre composite sandwich panels in the flatwise and in the edgewise positions was investigated with a view of using this material for construction and building applications. The effect of the number and orientation of sandwich laminations on the shear strength and failure behaviour of glued sandwich beams is examined using an asymmetrical beam shear test. The results showed that the behaviour of the glued sandwich beams in the flatwise position is governed by the shear strength of the core while in the edgewise position by the shear strength of the skin. In the edgewise position, the skin carries almost 60% of the load but only 20% in the flatwise position. With increasing sandwich laminations, the glued sandwich beams in the edgewise position achieved over 200% shear strength than beams in the flatwise position. The presence of vertical fibre composite skins has resulted in a more ductile failure behaviour for beams in the edgewise position. The results of the theoretical and numerical evaluations on the shear strength of the glued sandwich beams using the shear properties of the skin and the core were in good agreement with the experimental results. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The development of fibre composite sandwich structures has been very widespread for automotive, aerospace, marine and other industrial applications [1]. A great variety of core materials, such as balsa wood, foam core and honeycomb cores have been bonded to fibre composite skins to manufacture composite sandwich panels and structures. The benefits of sandwich construction are their performance, and better bending strength and stiffness [2]. The many advantages of composite sandwich materials have drawn a lot of attention in the construction industry and for civil engineering applications [3]. The flexibility of composite sandwich construction allows innovative structural developments from this material. Composite sandwich panels can be combined with traditional construction materials or structural fibre composite pultrusions and formed to carry loads that cannot be carried by individual ⇑ Corresponding author. Tel.: +61 7 4631 1385; fax: +61 7 4631 2110. E-mail addresses: [email protected] (A.C. Manalo), [email protected] (T. Aravinthan), [email protected] (W. Karunasena). 0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2013.06.025
sandwich structure [4,5]. A sandwich structure can also be designed to a desired stiffness and strength with no additional weight to suit various structural applications [6]. Recently, a new generation composite sandwich panel made up of glass fibre composite skins and modified phenolic core material has been developed in Australia [7]. After evaluating the favourable characteristics of the individual sandwich beams in flexure and shear as detailed in Manalo et al. [8,9], an innovative beam concept made completely from this sandwich structure has been developed by the authors to increase the use of composite sandwich construction in civil engineering applications. As these fibre composite sandwich panels are produced in limited thicknesses for reasons of cost effectiveness and efficiency, a structural beam can be manufactured by gluing several sandwich panels together either in the flatwise (horizontal) or edgewise (vertical) positions. The flexural behaviour of these glued structural fibre composite sandwich beams has shown that gluing these composite sandwich panels together resulted in a more stable and stronger section [10]. The study has also demonstrated that the concept of gluing a number of composite sandwich panels to form a structural beam is highly
1318
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
practical. Before this system can be used effectively for construction and building applications, a thorough understanding of the shear behaviour of this glue-laminated sandwich beam concept is necessary as the shear strength is an important consideration when designing composite sandwich structures. Previous studies in [9] focused only on the in-plane shear behaviour of individual sandwich beams while studies reported on [8,10] investigates the flexural behaviour of individual and glue-laminated sandwich structures. It is important therefore that the shear behaviour in the flatwise position and when these composite sandwich panels are glued together to form a structural beam be determined. In the design of glued-laminated (glulam) beam, the highquality materials are typically provided at the outer laminations for higher strength and stiffness [11]. Rammer [12] suggested that the strength of the glulam beams is governed by the weakest lamination within the beam section. Failure of the glulam beam can happen just after the strength of the weakest lamination is exceeded. Under service loads, it has been demonstrated that most sandwich construction failed due to shear failure of the core material [13]. The brittle nature of the core causes a sudden collapse of the sandwich structure and could become the limiting factor in designing such structures. In structural design of reinforced concrete, shear is accounted for by providing shear reinforcement such as stirrups in beams and dowels in slabs [14]. In the context of composite sandwich structures, there have been considerable attempts to increase the shear strength of composite sandwich construction, though a number of these developments entails complex and costly process. In the production glulam timber beams, the timber laminations are mostly glued in the flatwise position. Similarly, most sandwich constructions are tested in the flatwise position as it is commonly used as structural panels for roof, floor, walls and bridge deck. Thus, the shear behaviour of glued sandwich beams produced with flatwise and with edgewise laminations were investigated in an attempt to improve the structural performance of the glued composite sandwich beams without any material modifications but only by orienting the fibre composite skins to carry the shear that is usually carried by the core material. In this paper, the shear behaviour of the glued fibre composite sandwich beams was evaluated through experimental and analytical investigations. A number of composite sandwich beams were glued together in the flatwise and in the edgewise positions and subjected to asymmetrical beam shear test. The effects of the number and the orientation of sandwich laminations on the strength and failure behaviour of glued structural sandwich beams were discussed. Simplified calculation method to describe the approximate shear strength of the glued composite sandwich beams was presented. Numerical analyses of the shear behaviour of the glued sandwich beams using the properties of the constituent materials were also conducted to verify the experimental and analytical results.
listed in Table 1 were determined using test of coupon specimens following ISO and ASTM standards and are reported in the earlier studies conducted by Manalo et al. [8,9]. 2.2. Test specimen and preparation The structural composite sandwich panels used to produce the glued sandwich beams were provided by LOC Composites, Pty. Ltd., Australia. The specimens for the characterisation of individual sandwich beam behaviour were cut directly from the composite sandwich panels provided by the manufacturer. A number of composite sandwich panels were assembled and glued together in 2, 3 and 4 layers using Techniglue-HP R5 structural epoxy resin. The glued sandwich panels were then clamped for 24 h to initially cure the epoxy and were removed from clamping to post-cure at 90 °C for 8 h to attain good bonding between the composite sandwich laminations. After curing, the glue-laminated sandwich panels were cut to the required specimen width. The descriptions of the test specimens are listed in Table 2. In this table, the B and D represent the width and depth of the glued composite sandwich beam, respectively while L denotes the test span. 2.3. Test set-up and procedure The shear test of glued structural composite sandwich beams was performed using an asymmetrical beam shear test. This test set-up is proven effective in inducing shear failure and has provided a good estimation of the shear strength of sandwich structures with high strength core material [9]. In this test method, the specimen was eccentrically loaded at two trisected points and the supports were applied at the remaining two points. This loading configuration generates a high shear stress and a nearly zero moment at the centre of the specimen. The test set-up and instrumentation for the asymmetrical beam shear test is illustrated in Fig. 1. The load was applied through a 100 kN servo-hydraulic universal testing machine with a loading rate of 1.3 mm/min. A steel spreader beam was used to transfer the single load applied by the loading machine to the specimen asymmetrically. The loading pins and the supports had a diameter of 20 mm to prevent any localised failure of composite sandwich beams. In the edgewise position, 30 mm steel plates were provided under the loading points and at the supports to prevent local indentation failure. Resistance strain gauges oriented at ±45° (A2A-06-C-085C-500 type, supplied by Biolab Pty. Ltd., Australia) to the loading axis were attached on the surface along the middle line at mid-height of the sandwich beam specimen to evaluate the shear response during the entire loading regime. The applied load, cross-head displacement and strains were obtained using a System 5000 data logger. All of the specimens were tested up to failure to determine the strength and failure mechanisms.
3. Experimental results and observations The behaviour of the glued composite sandwich beams under asymmetrical beam shear test when loaded in the flatwise and edgewise positions are discussed in this section. 3.1. Failure load Table 3 summarises the load at first shear crack of the core material and the maximum shear load carried by the glued composite sandwich beams tested in the flatwise and the edgewise positions. For specimens tested in the flatwise position, the load at first crack corresponds to the maximum load as the specimen Table 1 Mechanical properties of the skin and core of the composite sandwich panel.
2. Experimental program The asymmetrical beam shear tests of the glued structural composite sandwich beams are discussed in this section.
2.1. Material properties The structural composite sandwich panel used in this study is made up of glass fibre composite skins co-cured onto the modified phenolic core material using a toughened phenol formaldehyde resin [7]. The fibre composite skin is made up of 2 plies of stitched bi-axial (0/90) E-CR glass fibre fabrics manufactured by Fiberex Corporation. The modified phenolic foam core material is made primarily from natural plant products with a proprietary formulation of LOC Composites Pty. Ltd., Australia. The fibre composite skin has a nominal thickness of 1.8 mm while the phenolic foam core has a nominal thickness of 16.4 mm. The mechanical properties of the skin and the core material of the innovative composite sandwich panel
Test
Property
Skin
Core
Flexure
Modulus (MPa) Peak stress (MPa) Strain at peak (%)
14,284.50 317.37 2.29
1,326.25 14.32 1.22
Tensile
Modulus (MPa) Peak stress (MPa) Strain at peak (%) Poisson’s ratio
15,380.00 246.80 1.61 0.25
1,032.51 5.97 0.61 0.29
Compression
Modulus (MPa) Peak stress (MPa) Strain at peak (%)
16,102.13 201.75 1.24
1,350.43 22.99 3.51
Shear
Modulus (MPa) Peak stress (MPa) Strain at peak (%)
2465.82 23.19 3.08
526.88 4.25 0.81
1319
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327 Table 2 Description of specimen for asymmetrical beam shear test. Specimen
Illustration
AS-1LSW-F
D
Number of specimens
D, mm
B, mm
L, mm
Orientation of testing
5
20
50
240
Flatwise
5
50
20
240
Edgewise
2
40
50
240
Flatwise
2
50
40
240
Edgewise
2
60
60
240
Flatwise
2
60
60
240
Edgewise
2
80
80
240
Flatwise
2
80
80
240
Edgewise
B AS-1LSW-E
D B AS-2LSW-F
D B AS-2LSW-E
D B AS-3LSW-F
D B AS-3LSW-E
D B AS-4LSW-F
D B AS-4LSW-E
D B failed immediately after the formation of the first shear crack in the core material. As expected, the shear capacity of the glued composite sandwich beams increases with increasing number of sandwich laminations. The results also show that the specimen with the same number of sandwich laminations and tested in the edgewise position failed at a higher load than in the flatwise position. Similarly, the load when the first shear crack in the core material was observed in the edgewise specimen is significantly higher than the failure load recorded for the flatwise specimen.
initiation of shear cracking in the core material. After which, a non-linear load and crosshead displacement relation was observed due to the progressive shear failure of the fibre composite skins. After the maximum load is reached, a decrease in the capacity of the specimens AS-1LSW-E and AS-2LSW-E was observed with increasing displacement of the crosshead. For specimens AS-3LSW-E and AS-4LSW-E, an immediate drop in the load was observed when the maximum load is reached indicating the final failure of the sandwich beam.
3.2. Load and crosshead displacement behaviour
3.3. Load–strain behaviour
Fig. 2 shows the load and the displacement of the crosshead during the entire test regime for individual and glued sandwich beams with 3 sandwich laminations. As indicated in the figure, the failure in the sandwich beam specimens is represented with a load drop in the load-crosshead displacement relation curve. For specimens tested in the flatwise positions, the load increased linearly with the displacement of the crosshead until final failure. A sudden load drop was observed which indicated the final failure of the composite sandwich beam tested in the flatwise position. For composite sandwich beams tested in the edgewise position, the load increased linearly with the crosshead displacement until the first load drop was observed. This load drop is due to the
The diagonal tensile (+45°) and absolute value of the compressive (45°) strain gauge measurements were plotted against the applied load during the entire test regime to determine if the asymmetrical beam shear test creates a pure shear for glued composite sandwich beams at the location of maximum shear. Zhou et al. [15] suggested that if the shear strain is pure, both tensile and compressive strains should be equal and opposite in sign. Fig. 3 shows the applied load and the indicated normal strains of the ±45° strain gauge attached to the glued sandwich beams tested under asymmetrical beam shear. The figure shows that in all the tested specimens, both tensile and compressive strain values are very close throughout the test, suggesting the presence of a
1320
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
P Strain gauge
30
80
40
40
80
30
300 mm
(a) Schematic diagram
(b) actual test-set up
Fig. 1. Test set-up and instrumentation for asymmetrical beam shear test.
Table 3 Failure load of composite sandwich beams under asymmetrical beam shear test. Specimen
Flatwise (kN)
AS-1LSW AS-2LSW AS-3LSW AS-4LSW
Edgewise (kN)
Peak load
Std. deviation
Load at 1st crack
Std. deviation
Peak load
Std. deviation
10.52 19.72 37.45 70.62
0.24 0.21 0.36 1.76
16.57 33.17 57.79 99.33
0.35 0.73 0.47 0.35
17.61 40.35 78.35 138.36
0.29 2.29 1.73 4.68
50
25
AS-1LSW-F
AS-2LSW-F
AS-1LSW-E
Load (kN)
Load (kN)
AS-2LSW-E
40
20 15 10
30 20 10
5
0
0 0
2
4
6
8
0
10
2
4
6
8
10
Displacement (mm)
Displacement (mm)
(a) AS-1LSW
(b) AS-2LSW 180
100
AS-3LSW-F
AS-4LSW-F
AS-3LSW-E
AS-4LSW-E
150
80
Load (kN)
Load (kN)
120 60 40 20
90 60 30
0
0
0
2
4
6
8
10
0
2
4
6
Displacement (mm)
Displacement (mm)
(c) AS-3L
(d) AS-4LSW
8
10
Fig. 2. Load and crosshead displacement relation of sandwich beams.
reasonably pure field of shear strain at the mid-depth and midspan of the glued sandwich beam specimen. However, some imbalance between the tensile and the compressive strains were recorded when the shear failure on the specimens were observed especially on sandwich beams tested in the edgewise position. Fig. 3a, c, e, and g shows that the recorded ±45° strains on the specimen tested in the flatwise position increased linearly with
the applied load up to failure. In the edgewise position (Fig. 3b, d, f, and h), gluing the sandwich beams together resulted to an almost linear load and strain behaviour up to the initial shear cracking of the core material. A non-linearity on the load–strain curve and a decrease in stiffness were then observed after this load up to failure. However, the specimens continued to carry load even after the initial shear failure of the fibre composite skins. For
1321
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
strain gages attached to the glued sandwich beam specimen, thus the failure strain was not measured. The total loading regime was still recorded to determine the maximum shear strength.
20
20
15
15
Load (kN)
Load (kN)
individual sandwich beam, the strain gauge captured the non-linear behaviour of the specimen up to failure while the failure of the outermost fibre composite skin resulted to the immediate failure of the
10
10
5
5
AS-SW-E(+45)
AS-1LSW-F(+45)
AS-SW-E(-45)
AS-1LSW-F(-45)
0
0
1000
2000
3000
4000
0
5000
5000
10000
Strain (microstrain)
(a) AS-1LSW-F
(b) AS-1LSW-E
25
50
20
40
15 10 5
30 20
AS-2LSW-E(+45)
AS-2LSW-F(-45)
0
1000
2000
3000
4000
AS-2LSW-E(-45)
0
5000
0
Strain (microstrains)
2000
4000
40
80
Load (kN)
Load (kN)
100
30 20
AS-3LSW-F(-45)
2000
3000
4000
40
AS-3LSW-E(+45) AS-3LSW-E(-45)
0
5000
0
2000
4000
6000
Strain (microstrains)
Strain (microstrains)
(e) AS-3LSW-F
(f) AS-3LSW-E
100
150
80
120
Load (kN)
Load (kN)
1000
60
20
AS-3LSW-F(+45)
0
8000
(d) AS-2LSW-E
50
0
6000
Strain (microstrains)
(c) AS-2LSW-F
10
20000
10
AS-2LSW-F(+45)
0
15000
Strain (microstrain)
Load (kN)
Load (kN)
0
60 40
8000
90 60 30
20
AS-4LSW-E(+45)
AS-4LSW-F(+45) AS-4LSW-F(-45)
0 0
1000
2000
3000
4000
Strain (microstrains)
(g) AS-4LSW-F
AS-4LSW-E(-45)
0 5000
0
2000
4000
6000
Strain (microstrains)
(h) AS-4LSW-E
Fig. 3. Load and +45° strain relationship of sandwich beams.
8000
1322
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
3.4. Failure behaviour Fig. 4 shows the failure behaviour of the glued composite sandwich beams under asymmetrical beam shear test. The experimental results show that the sandwich beam specimens tested in the flatwise position failed after the formation of the first shear crack in the core. In this position, a diagonal shear crack of approximately 45° propagates through the core material at the location of the maximum shear (Fig. 4a, c, e, and g). This failure is brittle and sudden which is accompanied by a loud noise after the appearance of the first shear crack. In the edgewise position, the initial failure is the shear cracking of the core material as indicated by
the load drop. The vertical fibre composite skin however inhibits the further development of shear cracks in the core to cause immediate failure. The specimen then continued to carry the load before shear cracking of the skin was observed (Fig. 4b, d, f, and h). This failure behaviour showed a more progressive failure for the composite sandwich beams in the edgewise than in the flatwise position. Compared to individual sandwich beams, the glued composite sandwich beams exhibited a more brittle failure behaviour with increasing sandwich laminations. This could be due to the joining effect of the epoxy adhesives which strengthen the bonded fibre composites together thereby increasing its shear stiffness.
Fig. 4. Failure of glued structural sandwich beams in shear.
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
4. Discussions The effects of the number and the orientation of composite sandwich laminations on the shear strength and failure mechanisms of the glued composite sandwich beams are discussed here. 4.1. Effect of number of sandwich laminations in shear strength The apparent shear strength of the glued composite sandwich beams was determined by dividing the shear force on the maximum shear region with the area of the homogenised beam section assuming an elastic response until failure. Assumption that the glued composite sandwich beams acted as a solid section with perfect bonding was also made. Fig. 5 shows the apparent shear strength of the glued sandwich beams with different number of sandwich laminations. The results indicate that the number of sandwich laminations has no significant effect on the shear strength of the glued sandwich beams tested in the flatwise position. In this position, the shear strength of glued sandwich beams is almost constant or slightly decreases with increasing sandwich laminations. The calculated shear strength of the glued composite sandwich beams in the flatwise position is between 4.8 and 5.5 MPa. This shear strength is comparable to the shear strength of the modified phenolic core material suggesting that the shear strength of the glued sandwich beam in the flatwise position is governed by the shear strength of the core regardless of how many laminations are there in a beam section. The slightly higher shear strength of the glued sandwich beams at failure compared to the shear strength of the phenolic core material as determined by the coupon test could be due to the contribution of the mid-layer horizontal fibre composite skins in carrying the shear. In the edgewise position, increasing the number of sandwich laminations resulted in an increase in the shear strength for the glued composite sandwich beam. Test results showed that the shear strength of sandwich beams in the edgewise position increased by at least 20% when the sandwich beams were bonded together compared to that of the individual sandwich beam. As indicated in the figure, the edgewise specimen has a shear strength of almost 9 MPa for single sandwich beam but increased to around 11 MPa for glued sandwich beams. Noticeably, the shear strength of the glued sandwich beams in the edgewise position is almost double than that of the sandwich beams in the flatwise position with higher number of sandwich laminations. 4.2. Effect of sandwich beam orientation in shear strength The results of the experimental investigation showed that the shear strength of glued structural composite sandwich beams in the edgewise position was consistently higher compared to those of the flatwise specimens. In the flatwise position, the shear strength of the glued sandwich beams is governed by the shear strength of the core material. This lacks of fibre reinforcement through the thickness direction of the core material results in a lower shear strength for sandwich beams in the flatwise position. The higher shear strength of the glued composite sandwich beams tested in the edgewise position is due to the presence of the vertical fibre composite skins which impeded the propagation and growth of shear cracks of the phenolic core thereby increasing the shear strength of the sandwich beam. Furthermore, the increase in the failure strength of the sandwich beams in the edgewise position can be attributed to the presence of structural epoxy which increased the shear stiffness of the bonded fibre composite skins due to the laminate restraining effect. Thus, an improvement in the shear behaviour can be made by orienting the sandwich beams in the edgewise position.
1323
The better shear performance of the glued composite sandwich beams in the edgewise position can be attributed to a number of factors. First, the phenolic core has a lower shear strength than the glass-fibre composite skin, implying the possibility of earlier failure of the core due to shear than that of the skin. Lastly, the vertical skins in the edgewise sandwich beams are oriented across the shear plane which resist the crack propagation, thus resulting to a higher shear strength. It can be concluded therefore that in the edgewise position, the shear strength of the glued composite sandwich beam is governed by the shear strength of the fibre composite skin. In this position, the vertical fibre composite skins carries almost 60% of the applied load while the core material carries the remaining 40% of the load. In comparison with the glued sandwich beam in the flatwise position, the core carries 80% of the load and the skin, the remaining 20%. These results indicate that the fibre composite skins has a significant contribution in the shear strength of the glued composite sandwich beams and should be considered in the overall evaluation of the shear behaviour of the composite sandwich beams. 4.3. Effect of sandwich beam orientation in failure mechanism The failure of the glued composite sandwich beams tested in the flatwise and the edgewise positions occurred along the intended shear plane. As expected, the shear failure of specimen tested in the flatwise position was sudden and catastrophic. In this position, the glued sandwich beam failed due to shear failure of the phenolic core material. The sandwich beam lost its capacity to carry load instantly without any residual load-carrying capacity beyond the peak load, which was observed in specimen tested in the edgewise position. The results of the experimental investigation showed that the glued sandwich beams tested in the edgewise position failed progressively as indicated by the load–displacement and load–strain relation curves. The applied load increased steadily until an initial shear crack in the core developed and a slight reduction in load was observed. The presence of vertical fibre composite skins impeded the shear cracking propagation and growth in the phenolic core material to cause immediate failure. On continued loading, this load reduction was recovered and exceeded, but at a slower loading rate because of the loss of beam stiffness and the progressive failure of the fibre composite skins. Subsequently, additional load drops caused by shear failure of the remaining skins occurred until complete failure of the beam. Gupta and Siller [16] determined the shear strength of structural laminated veneer lumber in two different orientations; plank orientation (flatwise) and joist orientation (edgewise). Their investigation showed that the failure plane is rough for specimens tested in the joist orientation while smooth in the plank oriented specimens resulting to a higher shear strength for specimen in the edgewise than in the flatwise position. For glued composite sandwich beams tested in the flatwise position, the final failure of the beam is due to shear failure of the core. This creates a smoother failure surface (a diagonal shear crack on the phenolic core) than the failure surface of the sandwich beams in the edgewise position. In the edgewise position, final failure of the beam is due to shear failure of the vertical fibre composite skins across the shear plane at a higher applied load. More importantly, failure mechanism in shear of the glued sandwich beam in the edgewise position is progressive compared to a brittle failure mode in the flatwise position. 5. Evaluation of glued composite sandwich beam behaviour The shear equations for composite sandwich beams presented in Manalo et al. [9] were extended to calculate the shear strength
1324
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
and the modified phenolic core material can be calculated by (3) while the failure load when transformed into an equivalent skin material using the ratio of the shear moduli, PE_G by (4):
Shear strength (MPa)
12
9
PE 6
PE
Edgewise
1
2
3
Fig. 5. Shear strength of glued sandwich beams with different number of laminations.
of the glued structural composite sandwich beams. Numerical simulations were also conducted to verify the analytical and experimental results on the behaviour of the glued composite sandwich beams. 5.1. Approach to estimate shear strength The observed failure mode of all the glued composite sandwich beams tested in the flatwise position was a shear failure in the phenolic core material while the beams in the edgewise position was a shear failure of the vertical fibre composite skins. These failure behaviours are similar to the observed failure mode of individual sandwich beams as reported in Manalo et al. [9]. The shear resistance of the glued composite sandwich beams under asymmetrical beam shear test was estimated using the empirical equation suggested by Triantafillou [17] and the proposed theoretical prediction based on the shear strength of the fibre composite skins and the modified phenolic core material. The shear capacity of the glued composite sandwich beams tested in the flatwise position when the composite section is transformed into an equivalent core material using the ratio of the moduli of elasticity of the skin and the core, PF_E is estimated using Eq. (1) while the shear capacity when all the materials were transformed into an equivalent core using the ratio of the shear moduli of the skin and the core material, PF_G is estimated using Eq. (2):
G
Gc ¼ nss tc þ 2ts B Gs
ð4Þ
4
Number of sandwich laminations
PF
ð3Þ
5.2. Finite element modelling and verification of sandwich beam behaviour
0
E
G
Ec ¼ nss tc þ 2t s B Es
3 Flatwise
PF
E
Es B ¼ nsc t c þ 2ts Ec
ð1Þ
Gs B ¼ nsc t c þ 2t s Gc
ð2Þ
where n is the number of sandwich laminations, tc and ts are the thicknesses of the core and the skin, respectively, sc is the shear strength of the core, Es and Ec are the flexural moduli of elasticity of the skin and core, respectively, Gs and Gc are the shear moduli of the skin and core, respectively and B is the width of the individual sandwich beam. It was observed during the course of the experimental study that failure of the glued composite sandwich beams tested in the edgewise position generally occurs due to shear failure of the vertical fibre composite skins. Thus, the shear capacity of the glued sandwich beams in the edgewise position is calculated by transforming all the materials in the composite beam section into an equivalent fibre composite skin material and failure is assumed when the shear strength of the fibre composite skin, ss is reached. The theoretical prediction of the failure load of the glued composite sandwich beam, PE_E when transformed into an equivalent skin material based on the elastic moduli of the fibre composite skins
Simulation of the asymmetrical beam shear test of the glued composite sandwich beams using Strand7 finite element program [18] has been performed to determine if the behaviour and the ultimate capacity of the composite sandwich beams could be predicted using the material properties established from the test of coupons. The FE model is created with the nominal dimensions of the composite sandwich beams using brick elements to replicate the actual geometry of the specimens. The FE model for composite sandwich beams is analysed by modelling the skin and the phenolic core as 20-node hexahedron (Hexa20) brick elements. The brick elements have aspect ratios of 1.4–1.9. Only half of the glued composite sandwich beam is modelled in asymmetrical beam shear as the specimen is symmetric along the XY plane to reduce the computational time. The established constitutive behaviour of the skin and the core materials was used in the analysis and assumption that no debonding failure between the skin and the core was also made. Further, the load spreader beam is included in the model to simulate the asymmetrical loading condition on the sandwich beam specimen and modelled using brick elements. Table 4 lists the number of Hexa20 bricks, nodes and the computation time for the FE model of the glued composite sandwich beams in shear while Fig. 6 shows the numerical model to simulate the shear behaviour in the flatwise and edgewise positions. The numerical analysis was conducted using the static solver in Strand7 finite element software in the FCD-XPP-034 computer (CPU-Intel P4). Under asymmetrical beam shear test, a load of 1 kN was applied to the specimen through the spreader beam. After analysis, a linear load combination factor was applied to the initial linear static load case result to generate the shear response of the sandwich beams for higher load increments. The behaviour of the sandwich beam was examined at each load increment and the load case where the maximum shear stress and strain of the constituent materials were exceeded was determined. This load was chosen as the failure load of the glued composite sandwich beam in shear. In the flatwise position, the glued sandwich beam is assumed to fail when the phenolic core material reaches its failure shear stress and strain. In the edgewise position, the first shear cracking of the core is considered when the maximum shear stress and strain of the core material of the sandwich beam were reached while the final failure of specimen is when
Table 4 FE model for sandwich beam under asymmetrical beam shear test. Specimen
Hexa20 bricks
Nodes
CPU time (sec)
AS-1LSW-F AS-1LSW-E AS-2LSW-F AS-2LSW-E AS-3LSW-F AS-3LSW-E AS-4LSW-F AS-4LSW-E
3380 5528 8598 9032 14,616 15,534 25,248 26,472
21,703 29,426 39,156 41,071 66,564 70,643 111,876 117,329
1858 2073 2874 3207 3996 3502 7095 7922
1325
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
Load spreader beam
Load spreader beam Loading rollers
Sandwich beam specimen
Loading rollers Loading plates
Sandwich beam specimen
Roller supports
Support plates
(a) flatwise
(b) edgewise
Fig. 6. FE model for glued sandwich beam under asymmetrical beam shear test.
Table 5 Predicted and actual failure load of composite sandwich beams in shear. Specimen
AS-1LSW AS-2LSW AS-3LSW AS-4LSW
Flatwise (kN)
Edgewise (kN)
Actual
1
2
FEM
Actual
3
4
FEM
10.52 19.72 37.45 70.62
13.11 27.25 49.93 90.62
11.16 22.77 42.98 75.27
11.00 23.00 42.00 75.00
17.61 40.35 78.35 138.36
23.61 45.03 84.99 148.87
16.36 31.41 59.17 103.87
17.00 31.50 61.00 105.00
the shear stress and strain of the fibre composite skins were exceeded. 6
6. Predicted results and comparison with experiments
Stress (MPa)
5 4 3 Expt-2F
Expt-3F
2
Expt-4F
6.1. Failure load
FEM-2F
1
FEM-3F FEM-4F
0 0
3000
6000
9000
12000
Strain (microstrains) Fig. 7. Shear stress and strain relationship of sandwich beams in the flatwise positions.
40
30
Stress (MPa)
The results of the analytical prediction and numerical simulations on the behaviour of glued structural composite sandwich beams and the comparison with the results of the experimental investigation are discussed in this section.
20 Expt-2E Expt-3E Expt-4E
10
FEM-2E FEM-3E FEM-4E
0 0
5000
10000
15000
20000
Strain (microstrains) Fig. 8. Shear stress and strain relationship of sandwich beams in the edgewise positions.
Table 5 summarises the actual failure load based on the experimental investigations and the predicted failure load in shear of the glue-laminated composite sandwich beams using Eqs. (1)–(4) and based on the FEM simulations. In the flatwise position, the predicted failure load using Eq. (1) gives an almost 25–38% higher than the actual failure load while the difference between the predicted and the actual load using Eq. (2) is only 6–14%. The results suggest that the shear strength of the glued sandwich beams tested in the flatwise position can be predicted accurately when all the materials are transformed into an equivalent core using the ratio of the shear moduli. The higher predicted load when all the materials were converted into a homogenous section using the ratio of the elastic modulus results in an equivalent area of almost 75% provided by the skin while the core provides only 25%. Using the ratio of the shear moduli of the constituent materials resulted in a section with an equivalent area with only 45% contribution from the skin and 55% from the core. For specimens tested in the edgewise position, the predicted failure load using Eq. (3) gives a 34% higher predicted load for individual sandwich lamination but only 8–12% higher load compared than the actual failure load. On the other hand, the predicted failure load using Eq. (4) gives a 7% lower load than the actual for individual sandwich laminations but an almost 25% lower load than the actual failure load for glued sandwich beams. The higher actual than the predicted failure load of glued laminated sandwich beams is due to the laminate restraining effect of the bonded fibre composite skins as explained in Section 4.2. It can be observed however
1326
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
that this predicted load is comparable to the load when the first shear crack on the specimen was observed (Table 3). However, a closer predicted failure load to the actual failure load can be obtained using Eq. (4) if the 20% increase in shear strength of the glued composite sandwich beam in the edgewise position observed in the experimental investigation would be accounted in the theoretical prediction. This suggests that the maximum shear strength of the glued composite sandwich beam in the edgewise position could be predicted accurately when all the materials in the beam section is converted into an equivalent skin material using the shear moduli of the constituent materials. Using the maximum shear stress of the skin and the core of the composite sandwich beam, the FEM simulations was successful in the estimation of the failure load of the glued sandwich beams under asymmetrical beam shear test. Also, the failure load determined from the numerical simulation is almost the same with the predicted failure load using Eqs. (2) and (4) for glued sandwich beams in the flatwise and the edgewise positions, respectively. In all the FEM simulations, the failure of the glued composite sandwich beams occurred at the region of maximum shear. 6.2. Shear stress–strain behaviour The comparison of the predicted and the actual shear stress– strain relationship of the glued composite sandwich beams tested in the flatwise and in the edgewise positions are shown in Figs. 7 and 8, respectively. In these figures, the results of the experiment and the FEM simulations are designated with Expt and FEM, respectively. In the experimental shear stress–strain relation curve,
the shear stress of the glued composite sandwich beam is calculated by dividing the shear force with the transformed area of the composite section. For flatwise specimens, the cross section of the glued composite sandwich beam was transformed into an equivalent core material while the edgewise specimen into an equivalent skin material. This assumption was based on the observed failure behaviour of the glued composite sandwich beams in an asymmetrical beam shear test. The shear strain is determined from the indicated normal strains of the ±45° strain gauges attached to the specimen. The predicted shear stress and strain relationship using the FEM analysis showed a good agreement with the experiment. As illustrated in Fig. 7, the glued composite sandwich beam tested in the flatwise positions failed at a shear strain of around 8000 microstrains. This strain level is similar to the failure strain where the phenolic core material fails in shear. The glued sandwich beam failed at shear strength slightly higher than that of the core material due to the contribution of the fibre composite skin. In the edgewise position, gluing the sandwich beams together resulted to an almost linear shear stress–strain behaviour until failure (Fig. 8). In comparison to the experimental results, the sandwich beams in the edgewise position failed at a slightly lower level of shear strength and strains. This could be due to the reinforcing effect of the epoxy adhesive to the bonded vertical fibre composite skins which was not accounted for the FEM simulations. This complex behaviour exhibited by the composite sandwich beam in the edgewise position is beyond the capabilities of the Strand7 finite element software and was not incorporated into the finite element model.
Fig. 9. Predicted failure of glued sandwich beams under asymmetrical beam shear.
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
6.3. Failure mechanism Fig. 9 shows the failure behaviour of glued composite sandwich beams under asymmetrical beam shear test based on the FEM simulations. The failure mechanisms based on the results of the numerical simulation showed good agreement with the failure behaviour observed in the experimental investigation. In all of the specimens tested in the flatwise and the edgewise positions, the shear failure of the glued composite sandwich beams occurred in the region of the maximum shear. In the flatwise position, the composite sandwich beams failed due to shear failure of the core on the topmost and the bottom most sandwich laminations which progressed to the core of the inner sandwich laminations at higher load cases (Fig. 9a and c). In comparison with the actual failure behaviour, the shear failure of the core material for glued sandwich beams tested in the flatwise position occurred simultaneously in all sandwich laminations after the appearance of the first shear crack. This could be due to the large energy released when shear failure of the core material in the topmost and bottom most sandwich laminations occurred which caused the total collapse of the sandwich beam. This complex failure behaviour in shear of the sandwich beam in the flatwise position could not be modelled in Strand7 but the load when the core shear failure occurs can be predicted accurately. In glued composite sandwich beams tested in the edgewise position, the predicted failure mode is shear failure of the fibre composite skins. In the numerical simulation, the maximum shear stress and strain of the core material is exceeded at a load comparable to the applied load when the first shear crack of the core material was observed experimentally. However, this was not considered as the failure load of the sandwich beam tested in the edgewise position. As observed in the experimental investigation, the presence of the vertical fibre composite skins inhibits the development of shear cracks in the core material and prevented the immediate failure of the sandwich beams. Thus, the final failure of the glue-laminated sandwich beam in the edgewise position is due to shear failure of the vertical fibre composite skin (Fig. 9b and d). The load to reach this type of failure is comparable to the theoretically predicted failure load. 7. Conclusions The shear behaviour of glued structural composite sandwich beams was determined through experimental, analytical and numerical investigations. Structural sandwich beams with 1, 2, 3, and 4 composite sandwich panels glued together were subjected in asymmetrical beam shear test in the flatwise and in the edgewise positions. The results showed that the number of sandwich laminations has no significant effect on the shear strength of the glued composite sandwich beams tested in the flatwise position. In this position, the shear strength of the glued sandwich beam is governed by the shear strength of the core material regardless of how many laminations are there in a beam section. In the edgewise position, the shear behaviour of the glued sandwich beam is improved significantly with the presence of the vertical fibre composite skins. In this position, the fibre composite skins carries almost 60% of the applied load compared in the flatwise position where the skin carries only 20% of the applied load. With increasing sandwich laminations, the glued composite sandwich beams in the edgewise position achieved over 200% shear strength than beams in the flatwise position. The vertical fibre composite skins has also resulted in a progressive failure and a more ductile failure behaviour of the glued structural composite sandwich beams in the edgewise position. The prediction equation for shear strength when all the materials are transformed into an equivalent area based on the shear
1327
properties of the constituent materials predicted accurately the failure load of the glued composite sandwich beam in the flatwise position but gives a lower predicted load than the actual in the edgewise position. A closer predicted load to the actual failure load can be obtained if the percentage increase in shear strength in gluing the sandwich beams in the edgewise position would be accounted for. The prediction equation agreed very well with the experimental results and the FEM simulations, giving confidence in the validity of the assumptions adopted. The results of the study showed a significant improvement on the shear strength of the glued sandwich beams when the sandwich panels are positioned edgewise. This point towards the high feasibility of using glued composite sandwich beams for construction and building applications. A research and development project is now being implemented to develop full-size glued structural composite sandwich beams with the objective of exploring the practical application of this composite structure in replacement railway timber sleeper. Acknowledgements The authors gratefully acknowledge the LOC Composites Ltd., Pty., Australia for providing the composite sandwich panels used in this study. The support of Wayne Crowell, Atul Sakhiya and Christopher Pickford in the preparation and testing of the glued composite sandwich beams is greatly acknowledged. References [1] Bakis CE, Bank LC, Brown VL, Cosenza E, Davalos JF, Lesko JJ, et al. FRP composites in construction – state of the art review. ASCE J Compos Constr 2002;6(2):78–87. [2] Belouettar S, Abbadi A, Azari Z, Belouettar R, Freres P. Experimental investigation of static and fatigue behaviour of composite honeycomb materials using four point bending tests. Compos Struct 2008;87(3):265–73. [3] Keller T. Material tailored use of FRP composites in bridge and building construction. Switzerland: Swiss Federal Institute of Technology Lausanne; 2006. [4] Canning L, Hollaway L, Thorne AM. Manufacture, testing and numerical analysis of an innovative polymer composite/concrete structural unit’, Proceedings of Inst. Civil Eng Struct Build 1999;134:231–41. [5] Lopez-Anido R, Xu H. Structural characterization of hybrid fibre-reinforced polymer-glulam panels for bridge decks. ASCE J Compos Constr 2002;6(3):194–203. [6] Williams B, Shehata E, Rizkalla S. Filament-wound glass fibre reinforced polymer bridge deck modules. ASCE J Compos Constr 2003;7(3):266–76. [7] Van Erp G, Rogers D. A highly sustainable fibre composite building panel. In: Proceedings of the international workshop on fibre composites in civil infrastructure – past, present and future. University of Southern Queensland, Toowoomba, Queensland, Australia; 1–2 December 2008. [8] Manalo AC, Aravinthan T, Karunasena W, Islam MM. Flexural behaviour of structural fibre composite sandwich beams in flatwise and edgewise positions. Compos Struct 2010;92(4):984–95. [9] Manalo AC, Aravinthan T, Karunasena W. In-plane shear behaviour of fibre composite sandwich beams using asymmetrical beam shear test. J Constr Build Mater 2010;24(10):1952–60. [10] Manalo AC, Aravinthan T, Karunasena W. Flexural behaviour of glue-laminated fibre composite sandwich beams. Compos Struct 2010;92(11):2703–11. [11] Hernandez R, Davalos JF, Sonti SS, et al. Strength and stiffness of reinforced Yellow-Poplar glued-laminated beams. In: Res. Pap. FPL-RP-554. Madison (WI): US Department of Agriculture, Forest Service, Forest Products Laboratory; 1997. 28 p.. [12] Rammer DR. Shear strength of glued-laminated timber beams and panels. Forest Products Laboratory, USDA Forest Service; 1996. p. 192–201. [13] Kampner M, Grenestedt JL. On using corrugated skins to carry shear in sandwich beams. Compos Struct 2007;85:139–48. [14] Mirsayah AA, Banthia N. Shear strength of steel fibre reinforced concrete. ACI Mater J 2002;99(5):473–9. [15] Zhou G, Green ER, Morrison C. In-plane and interlaminar shear properties of carbon/epoxy laminates. J Compos Sci Technol 1995;55:187–93. [16] Gupta R, Siller TS. A comparison of the shear strength of structural composite lumber using torsion and shear block tests. Forest Prod J 2005;55(12):29–34. [17] Triantafillou TC. Composites: a new possibility for the shear strengthening of concrete, masonry and wood. J Compos Sci Technol 1998;58:1285–95. [18] Strand7. Strand7 Release 2.3.7 finite element analysis system. Sydney, Australia: Strand7 Software; 2005.
Contents lists available at SciVerse ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Shear behaviour of glued structural fibre composite sandwich beams A.C. Manalo, T. Aravinthan ⇑, W. Karunasena Centre of Excellence in Engineered Fibre Composites (CEEFC), Faculty of Engineering and Surveying, University of Southern Queensland, Toowoomba, Queensland 4350, Australia
h i g h l i g h t s Structural beams were made from glue-laminated fibre composite sandwich panels. Glued sandwich beams were tested under asymmetrical shear. Beams with edgewise layers are twice stronger than beams with flatwise layers. Beams with edgewise layers exhibited a progressive failure behaviour. Theoretical evaluations were in good agreement with the experimental results.
a r t i c l e
i n f o
Article history: Received 27 July 2010 Received in revised form 10 May 2013 Accepted 17 June 2013 Available online 12 July 2013 Keywords: Glued sandwich beams Shear Structural composite sandwich panels Fibre composites Flatwise Edgewise
a b s t r a c t The shear behaviour of structural sandwich beams made by gluing together fibre composite sandwich panels in the flatwise and in the edgewise positions was investigated with a view of using this material for construction and building applications. The effect of the number and orientation of sandwich laminations on the shear strength and failure behaviour of glued sandwich beams is examined using an asymmetrical beam shear test. The results showed that the behaviour of the glued sandwich beams in the flatwise position is governed by the shear strength of the core while in the edgewise position by the shear strength of the skin. In the edgewise position, the skin carries almost 60% of the load but only 20% in the flatwise position. With increasing sandwich laminations, the glued sandwich beams in the edgewise position achieved over 200% shear strength than beams in the flatwise position. The presence of vertical fibre composite skins has resulted in a more ductile failure behaviour for beams in the edgewise position. The results of the theoretical and numerical evaluations on the shear strength of the glued sandwich beams using the shear properties of the skin and the core were in good agreement with the experimental results. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The development of fibre composite sandwich structures has been very widespread for automotive, aerospace, marine and other industrial applications [1]. A great variety of core materials, such as balsa wood, foam core and honeycomb cores have been bonded to fibre composite skins to manufacture composite sandwich panels and structures. The benefits of sandwich construction are their performance, and better bending strength and stiffness [2]. The many advantages of composite sandwich materials have drawn a lot of attention in the construction industry and for civil engineering applications [3]. The flexibility of composite sandwich construction allows innovative structural developments from this material. Composite sandwich panels can be combined with traditional construction materials or structural fibre composite pultrusions and formed to carry loads that cannot be carried by individual ⇑ Corresponding author. Tel.: +61 7 4631 1385; fax: +61 7 4631 2110. E-mail addresses: [email protected] (A.C. Manalo), [email protected] (T. Aravinthan), [email protected] (W. Karunasena). 0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2013.06.025
sandwich structure [4,5]. A sandwich structure can also be designed to a desired stiffness and strength with no additional weight to suit various structural applications [6]. Recently, a new generation composite sandwich panel made up of glass fibre composite skins and modified phenolic core material has been developed in Australia [7]. After evaluating the favourable characteristics of the individual sandwich beams in flexure and shear as detailed in Manalo et al. [8,9], an innovative beam concept made completely from this sandwich structure has been developed by the authors to increase the use of composite sandwich construction in civil engineering applications. As these fibre composite sandwich panels are produced in limited thicknesses for reasons of cost effectiveness and efficiency, a structural beam can be manufactured by gluing several sandwich panels together either in the flatwise (horizontal) or edgewise (vertical) positions. The flexural behaviour of these glued structural fibre composite sandwich beams has shown that gluing these composite sandwich panels together resulted in a more stable and stronger section [10]. The study has also demonstrated that the concept of gluing a number of composite sandwich panels to form a structural beam is highly
1318
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
practical. Before this system can be used effectively for construction and building applications, a thorough understanding of the shear behaviour of this glue-laminated sandwich beam concept is necessary as the shear strength is an important consideration when designing composite sandwich structures. Previous studies in [9] focused only on the in-plane shear behaviour of individual sandwich beams while studies reported on [8,10] investigates the flexural behaviour of individual and glue-laminated sandwich structures. It is important therefore that the shear behaviour in the flatwise position and when these composite sandwich panels are glued together to form a structural beam be determined. In the design of glued-laminated (glulam) beam, the highquality materials are typically provided at the outer laminations for higher strength and stiffness [11]. Rammer [12] suggested that the strength of the glulam beams is governed by the weakest lamination within the beam section. Failure of the glulam beam can happen just after the strength of the weakest lamination is exceeded. Under service loads, it has been demonstrated that most sandwich construction failed due to shear failure of the core material [13]. The brittle nature of the core causes a sudden collapse of the sandwich structure and could become the limiting factor in designing such structures. In structural design of reinforced concrete, shear is accounted for by providing shear reinforcement such as stirrups in beams and dowels in slabs [14]. In the context of composite sandwich structures, there have been considerable attempts to increase the shear strength of composite sandwich construction, though a number of these developments entails complex and costly process. In the production glulam timber beams, the timber laminations are mostly glued in the flatwise position. Similarly, most sandwich constructions are tested in the flatwise position as it is commonly used as structural panels for roof, floor, walls and bridge deck. Thus, the shear behaviour of glued sandwich beams produced with flatwise and with edgewise laminations were investigated in an attempt to improve the structural performance of the glued composite sandwich beams without any material modifications but only by orienting the fibre composite skins to carry the shear that is usually carried by the core material. In this paper, the shear behaviour of the glued fibre composite sandwich beams was evaluated through experimental and analytical investigations. A number of composite sandwich beams were glued together in the flatwise and in the edgewise positions and subjected to asymmetrical beam shear test. The effects of the number and the orientation of sandwich laminations on the strength and failure behaviour of glued structural sandwich beams were discussed. Simplified calculation method to describe the approximate shear strength of the glued composite sandwich beams was presented. Numerical analyses of the shear behaviour of the glued sandwich beams using the properties of the constituent materials were also conducted to verify the experimental and analytical results.
listed in Table 1 were determined using test of coupon specimens following ISO and ASTM standards and are reported in the earlier studies conducted by Manalo et al. [8,9]. 2.2. Test specimen and preparation The structural composite sandwich panels used to produce the glued sandwich beams were provided by LOC Composites, Pty. Ltd., Australia. The specimens for the characterisation of individual sandwich beam behaviour were cut directly from the composite sandwich panels provided by the manufacturer. A number of composite sandwich panels were assembled and glued together in 2, 3 and 4 layers using Techniglue-HP R5 structural epoxy resin. The glued sandwich panels were then clamped for 24 h to initially cure the epoxy and were removed from clamping to post-cure at 90 °C for 8 h to attain good bonding between the composite sandwich laminations. After curing, the glue-laminated sandwich panels were cut to the required specimen width. The descriptions of the test specimens are listed in Table 2. In this table, the B and D represent the width and depth of the glued composite sandwich beam, respectively while L denotes the test span. 2.3. Test set-up and procedure The shear test of glued structural composite sandwich beams was performed using an asymmetrical beam shear test. This test set-up is proven effective in inducing shear failure and has provided a good estimation of the shear strength of sandwich structures with high strength core material [9]. In this test method, the specimen was eccentrically loaded at two trisected points and the supports were applied at the remaining two points. This loading configuration generates a high shear stress and a nearly zero moment at the centre of the specimen. The test set-up and instrumentation for the asymmetrical beam shear test is illustrated in Fig. 1. The load was applied through a 100 kN servo-hydraulic universal testing machine with a loading rate of 1.3 mm/min. A steel spreader beam was used to transfer the single load applied by the loading machine to the specimen asymmetrically. The loading pins and the supports had a diameter of 20 mm to prevent any localised failure of composite sandwich beams. In the edgewise position, 30 mm steel plates were provided under the loading points and at the supports to prevent local indentation failure. Resistance strain gauges oriented at ±45° (A2A-06-C-085C-500 type, supplied by Biolab Pty. Ltd., Australia) to the loading axis were attached on the surface along the middle line at mid-height of the sandwich beam specimen to evaluate the shear response during the entire loading regime. The applied load, cross-head displacement and strains were obtained using a System 5000 data logger. All of the specimens were tested up to failure to determine the strength and failure mechanisms.
3. Experimental results and observations The behaviour of the glued composite sandwich beams under asymmetrical beam shear test when loaded in the flatwise and edgewise positions are discussed in this section. 3.1. Failure load Table 3 summarises the load at first shear crack of the core material and the maximum shear load carried by the glued composite sandwich beams tested in the flatwise and the edgewise positions. For specimens tested in the flatwise position, the load at first crack corresponds to the maximum load as the specimen Table 1 Mechanical properties of the skin and core of the composite sandwich panel.
2. Experimental program The asymmetrical beam shear tests of the glued structural composite sandwich beams are discussed in this section.
2.1. Material properties The structural composite sandwich panel used in this study is made up of glass fibre composite skins co-cured onto the modified phenolic core material using a toughened phenol formaldehyde resin [7]. The fibre composite skin is made up of 2 plies of stitched bi-axial (0/90) E-CR glass fibre fabrics manufactured by Fiberex Corporation. The modified phenolic foam core material is made primarily from natural plant products with a proprietary formulation of LOC Composites Pty. Ltd., Australia. The fibre composite skin has a nominal thickness of 1.8 mm while the phenolic foam core has a nominal thickness of 16.4 mm. The mechanical properties of the skin and the core material of the innovative composite sandwich panel
Test
Property
Skin
Core
Flexure
Modulus (MPa) Peak stress (MPa) Strain at peak (%)
14,284.50 317.37 2.29
1,326.25 14.32 1.22
Tensile
Modulus (MPa) Peak stress (MPa) Strain at peak (%) Poisson’s ratio
15,380.00 246.80 1.61 0.25
1,032.51 5.97 0.61 0.29
Compression
Modulus (MPa) Peak stress (MPa) Strain at peak (%)
16,102.13 201.75 1.24
1,350.43 22.99 3.51
Shear
Modulus (MPa) Peak stress (MPa) Strain at peak (%)
2465.82 23.19 3.08
526.88 4.25 0.81
1319
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327 Table 2 Description of specimen for asymmetrical beam shear test. Specimen
Illustration
AS-1LSW-F
D
Number of specimens
D, mm
B, mm
L, mm
Orientation of testing
5
20
50
240
Flatwise
5
50
20
240
Edgewise
2
40
50
240
Flatwise
2
50
40
240
Edgewise
2
60
60
240
Flatwise
2
60
60
240
Edgewise
2
80
80
240
Flatwise
2
80
80
240
Edgewise
B AS-1LSW-E
D B AS-2LSW-F
D B AS-2LSW-E
D B AS-3LSW-F
D B AS-3LSW-E
D B AS-4LSW-F
D B AS-4LSW-E
D B failed immediately after the formation of the first shear crack in the core material. As expected, the shear capacity of the glued composite sandwich beams increases with increasing number of sandwich laminations. The results also show that the specimen with the same number of sandwich laminations and tested in the edgewise position failed at a higher load than in the flatwise position. Similarly, the load when the first shear crack in the core material was observed in the edgewise specimen is significantly higher than the failure load recorded for the flatwise specimen.
initiation of shear cracking in the core material. After which, a non-linear load and crosshead displacement relation was observed due to the progressive shear failure of the fibre composite skins. After the maximum load is reached, a decrease in the capacity of the specimens AS-1LSW-E and AS-2LSW-E was observed with increasing displacement of the crosshead. For specimens AS-3LSW-E and AS-4LSW-E, an immediate drop in the load was observed when the maximum load is reached indicating the final failure of the sandwich beam.
3.2. Load and crosshead displacement behaviour
3.3. Load–strain behaviour
Fig. 2 shows the load and the displacement of the crosshead during the entire test regime for individual and glued sandwich beams with 3 sandwich laminations. As indicated in the figure, the failure in the sandwich beam specimens is represented with a load drop in the load-crosshead displacement relation curve. For specimens tested in the flatwise positions, the load increased linearly with the displacement of the crosshead until final failure. A sudden load drop was observed which indicated the final failure of the composite sandwich beam tested in the flatwise position. For composite sandwich beams tested in the edgewise position, the load increased linearly with the crosshead displacement until the first load drop was observed. This load drop is due to the
The diagonal tensile (+45°) and absolute value of the compressive (45°) strain gauge measurements were plotted against the applied load during the entire test regime to determine if the asymmetrical beam shear test creates a pure shear for glued composite sandwich beams at the location of maximum shear. Zhou et al. [15] suggested that if the shear strain is pure, both tensile and compressive strains should be equal and opposite in sign. Fig. 3 shows the applied load and the indicated normal strains of the ±45° strain gauge attached to the glued sandwich beams tested under asymmetrical beam shear. The figure shows that in all the tested specimens, both tensile and compressive strain values are very close throughout the test, suggesting the presence of a
1320
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
P Strain gauge
30
80
40
40
80
30
300 mm
(a) Schematic diagram
(b) actual test-set up
Fig. 1. Test set-up and instrumentation for asymmetrical beam shear test.
Table 3 Failure load of composite sandwich beams under asymmetrical beam shear test. Specimen
Flatwise (kN)
AS-1LSW AS-2LSW AS-3LSW AS-4LSW
Edgewise (kN)
Peak load
Std. deviation
Load at 1st crack
Std. deviation
Peak load
Std. deviation
10.52 19.72 37.45 70.62
0.24 0.21 0.36 1.76
16.57 33.17 57.79 99.33
0.35 0.73 0.47 0.35
17.61 40.35 78.35 138.36
0.29 2.29 1.73 4.68
50
25
AS-1LSW-F
AS-2LSW-F
AS-1LSW-E
Load (kN)
Load (kN)
AS-2LSW-E
40
20 15 10
30 20 10
5
0
0 0
2
4
6
8
0
10
2
4
6
8
10
Displacement (mm)
Displacement (mm)
(a) AS-1LSW
(b) AS-2LSW 180
100
AS-3LSW-F
AS-4LSW-F
AS-3LSW-E
AS-4LSW-E
150
80
Load (kN)
Load (kN)
120 60 40 20
90 60 30
0
0
0
2
4
6
8
10
0
2
4
6
Displacement (mm)
Displacement (mm)
(c) AS-3L
(d) AS-4LSW
8
10
Fig. 2. Load and crosshead displacement relation of sandwich beams.
reasonably pure field of shear strain at the mid-depth and midspan of the glued sandwich beam specimen. However, some imbalance between the tensile and the compressive strains were recorded when the shear failure on the specimens were observed especially on sandwich beams tested in the edgewise position. Fig. 3a, c, e, and g shows that the recorded ±45° strains on the specimen tested in the flatwise position increased linearly with
the applied load up to failure. In the edgewise position (Fig. 3b, d, f, and h), gluing the sandwich beams together resulted to an almost linear load and strain behaviour up to the initial shear cracking of the core material. A non-linearity on the load–strain curve and a decrease in stiffness were then observed after this load up to failure. However, the specimens continued to carry load even after the initial shear failure of the fibre composite skins. For
1321
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
strain gages attached to the glued sandwich beam specimen, thus the failure strain was not measured. The total loading regime was still recorded to determine the maximum shear strength.
20
20
15
15
Load (kN)
Load (kN)
individual sandwich beam, the strain gauge captured the non-linear behaviour of the specimen up to failure while the failure of the outermost fibre composite skin resulted to the immediate failure of the
10
10
5
5
AS-SW-E(+45)
AS-1LSW-F(+45)
AS-SW-E(-45)
AS-1LSW-F(-45)
0
0
1000
2000
3000
4000
0
5000
5000
10000
Strain (microstrain)
(a) AS-1LSW-F
(b) AS-1LSW-E
25
50
20
40
15 10 5
30 20
AS-2LSW-E(+45)
AS-2LSW-F(-45)
0
1000
2000
3000
4000
AS-2LSW-E(-45)
0
5000
0
Strain (microstrains)
2000
4000
40
80
Load (kN)
Load (kN)
100
30 20
AS-3LSW-F(-45)
2000
3000
4000
40
AS-3LSW-E(+45) AS-3LSW-E(-45)
0
5000
0
2000
4000
6000
Strain (microstrains)
Strain (microstrains)
(e) AS-3LSW-F
(f) AS-3LSW-E
100
150
80
120
Load (kN)
Load (kN)
1000
60
20
AS-3LSW-F(+45)
0
8000
(d) AS-2LSW-E
50
0
6000
Strain (microstrains)
(c) AS-2LSW-F
10
20000
10
AS-2LSW-F(+45)
0
15000
Strain (microstrain)
Load (kN)
Load (kN)
0
60 40
8000
90 60 30
20
AS-4LSW-E(+45)
AS-4LSW-F(+45) AS-4LSW-F(-45)
0 0
1000
2000
3000
4000
Strain (microstrains)
(g) AS-4LSW-F
AS-4LSW-E(-45)
0 5000
0
2000
4000
6000
Strain (microstrains)
(h) AS-4LSW-E
Fig. 3. Load and +45° strain relationship of sandwich beams.
8000
1322
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
3.4. Failure behaviour Fig. 4 shows the failure behaviour of the glued composite sandwich beams under asymmetrical beam shear test. The experimental results show that the sandwich beam specimens tested in the flatwise position failed after the formation of the first shear crack in the core. In this position, a diagonal shear crack of approximately 45° propagates through the core material at the location of the maximum shear (Fig. 4a, c, e, and g). This failure is brittle and sudden which is accompanied by a loud noise after the appearance of the first shear crack. In the edgewise position, the initial failure is the shear cracking of the core material as indicated by
the load drop. The vertical fibre composite skin however inhibits the further development of shear cracks in the core to cause immediate failure. The specimen then continued to carry the load before shear cracking of the skin was observed (Fig. 4b, d, f, and h). This failure behaviour showed a more progressive failure for the composite sandwich beams in the edgewise than in the flatwise position. Compared to individual sandwich beams, the glued composite sandwich beams exhibited a more brittle failure behaviour with increasing sandwich laminations. This could be due to the joining effect of the epoxy adhesives which strengthen the bonded fibre composites together thereby increasing its shear stiffness.
Fig. 4. Failure of glued structural sandwich beams in shear.
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
4. Discussions The effects of the number and the orientation of composite sandwich laminations on the shear strength and failure mechanisms of the glued composite sandwich beams are discussed here. 4.1. Effect of number of sandwich laminations in shear strength The apparent shear strength of the glued composite sandwich beams was determined by dividing the shear force on the maximum shear region with the area of the homogenised beam section assuming an elastic response until failure. Assumption that the glued composite sandwich beams acted as a solid section with perfect bonding was also made. Fig. 5 shows the apparent shear strength of the glued sandwich beams with different number of sandwich laminations. The results indicate that the number of sandwich laminations has no significant effect on the shear strength of the glued sandwich beams tested in the flatwise position. In this position, the shear strength of glued sandwich beams is almost constant or slightly decreases with increasing sandwich laminations. The calculated shear strength of the glued composite sandwich beams in the flatwise position is between 4.8 and 5.5 MPa. This shear strength is comparable to the shear strength of the modified phenolic core material suggesting that the shear strength of the glued sandwich beam in the flatwise position is governed by the shear strength of the core regardless of how many laminations are there in a beam section. The slightly higher shear strength of the glued sandwich beams at failure compared to the shear strength of the phenolic core material as determined by the coupon test could be due to the contribution of the mid-layer horizontal fibre composite skins in carrying the shear. In the edgewise position, increasing the number of sandwich laminations resulted in an increase in the shear strength for the glued composite sandwich beam. Test results showed that the shear strength of sandwich beams in the edgewise position increased by at least 20% when the sandwich beams were bonded together compared to that of the individual sandwich beam. As indicated in the figure, the edgewise specimen has a shear strength of almost 9 MPa for single sandwich beam but increased to around 11 MPa for glued sandwich beams. Noticeably, the shear strength of the glued sandwich beams in the edgewise position is almost double than that of the sandwich beams in the flatwise position with higher number of sandwich laminations. 4.2. Effect of sandwich beam orientation in shear strength The results of the experimental investigation showed that the shear strength of glued structural composite sandwich beams in the edgewise position was consistently higher compared to those of the flatwise specimens. In the flatwise position, the shear strength of the glued sandwich beams is governed by the shear strength of the core material. This lacks of fibre reinforcement through the thickness direction of the core material results in a lower shear strength for sandwich beams in the flatwise position. The higher shear strength of the glued composite sandwich beams tested in the edgewise position is due to the presence of the vertical fibre composite skins which impeded the propagation and growth of shear cracks of the phenolic core thereby increasing the shear strength of the sandwich beam. Furthermore, the increase in the failure strength of the sandwich beams in the edgewise position can be attributed to the presence of structural epoxy which increased the shear stiffness of the bonded fibre composite skins due to the laminate restraining effect. Thus, an improvement in the shear behaviour can be made by orienting the sandwich beams in the edgewise position.
1323
The better shear performance of the glued composite sandwich beams in the edgewise position can be attributed to a number of factors. First, the phenolic core has a lower shear strength than the glass-fibre composite skin, implying the possibility of earlier failure of the core due to shear than that of the skin. Lastly, the vertical skins in the edgewise sandwich beams are oriented across the shear plane which resist the crack propagation, thus resulting to a higher shear strength. It can be concluded therefore that in the edgewise position, the shear strength of the glued composite sandwich beam is governed by the shear strength of the fibre composite skin. In this position, the vertical fibre composite skins carries almost 60% of the applied load while the core material carries the remaining 40% of the load. In comparison with the glued sandwich beam in the flatwise position, the core carries 80% of the load and the skin, the remaining 20%. These results indicate that the fibre composite skins has a significant contribution in the shear strength of the glued composite sandwich beams and should be considered in the overall evaluation of the shear behaviour of the composite sandwich beams. 4.3. Effect of sandwich beam orientation in failure mechanism The failure of the glued composite sandwich beams tested in the flatwise and the edgewise positions occurred along the intended shear plane. As expected, the shear failure of specimen tested in the flatwise position was sudden and catastrophic. In this position, the glued sandwich beam failed due to shear failure of the phenolic core material. The sandwich beam lost its capacity to carry load instantly without any residual load-carrying capacity beyond the peak load, which was observed in specimen tested in the edgewise position. The results of the experimental investigation showed that the glued sandwich beams tested in the edgewise position failed progressively as indicated by the load–displacement and load–strain relation curves. The applied load increased steadily until an initial shear crack in the core developed and a slight reduction in load was observed. The presence of vertical fibre composite skins impeded the shear cracking propagation and growth in the phenolic core material to cause immediate failure. On continued loading, this load reduction was recovered and exceeded, but at a slower loading rate because of the loss of beam stiffness and the progressive failure of the fibre composite skins. Subsequently, additional load drops caused by shear failure of the remaining skins occurred until complete failure of the beam. Gupta and Siller [16] determined the shear strength of structural laminated veneer lumber in two different orientations; plank orientation (flatwise) and joist orientation (edgewise). Their investigation showed that the failure plane is rough for specimens tested in the joist orientation while smooth in the plank oriented specimens resulting to a higher shear strength for specimen in the edgewise than in the flatwise position. For glued composite sandwich beams tested in the flatwise position, the final failure of the beam is due to shear failure of the core. This creates a smoother failure surface (a diagonal shear crack on the phenolic core) than the failure surface of the sandwich beams in the edgewise position. In the edgewise position, final failure of the beam is due to shear failure of the vertical fibre composite skins across the shear plane at a higher applied load. More importantly, failure mechanism in shear of the glued sandwich beam in the edgewise position is progressive compared to a brittle failure mode in the flatwise position. 5. Evaluation of glued composite sandwich beam behaviour The shear equations for composite sandwich beams presented in Manalo et al. [9] were extended to calculate the shear strength
1324
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
and the modified phenolic core material can be calculated by (3) while the failure load when transformed into an equivalent skin material using the ratio of the shear moduli, PE_G by (4):
Shear strength (MPa)
12
9
PE 6
PE
Edgewise
1
2
3
Fig. 5. Shear strength of glued sandwich beams with different number of laminations.
of the glued structural composite sandwich beams. Numerical simulations were also conducted to verify the analytical and experimental results on the behaviour of the glued composite sandwich beams. 5.1. Approach to estimate shear strength The observed failure mode of all the glued composite sandwich beams tested in the flatwise position was a shear failure in the phenolic core material while the beams in the edgewise position was a shear failure of the vertical fibre composite skins. These failure behaviours are similar to the observed failure mode of individual sandwich beams as reported in Manalo et al. [9]. The shear resistance of the glued composite sandwich beams under asymmetrical beam shear test was estimated using the empirical equation suggested by Triantafillou [17] and the proposed theoretical prediction based on the shear strength of the fibre composite skins and the modified phenolic core material. The shear capacity of the glued composite sandwich beams tested in the flatwise position when the composite section is transformed into an equivalent core material using the ratio of the moduli of elasticity of the skin and the core, PF_E is estimated using Eq. (1) while the shear capacity when all the materials were transformed into an equivalent core using the ratio of the shear moduli of the skin and the core material, PF_G is estimated using Eq. (2):
G
Gc ¼ nss tc þ 2ts B Gs
ð4Þ
4
Number of sandwich laminations
PF
ð3Þ
5.2. Finite element modelling and verification of sandwich beam behaviour
0
E
G
Ec ¼ nss tc þ 2t s B Es
3 Flatwise
PF
E
Es B ¼ nsc t c þ 2ts Ec
ð1Þ
Gs B ¼ nsc t c þ 2t s Gc
ð2Þ
where n is the number of sandwich laminations, tc and ts are the thicknesses of the core and the skin, respectively, sc is the shear strength of the core, Es and Ec are the flexural moduli of elasticity of the skin and core, respectively, Gs and Gc are the shear moduli of the skin and core, respectively and B is the width of the individual sandwich beam. It was observed during the course of the experimental study that failure of the glued composite sandwich beams tested in the edgewise position generally occurs due to shear failure of the vertical fibre composite skins. Thus, the shear capacity of the glued sandwich beams in the edgewise position is calculated by transforming all the materials in the composite beam section into an equivalent fibre composite skin material and failure is assumed when the shear strength of the fibre composite skin, ss is reached. The theoretical prediction of the failure load of the glued composite sandwich beam, PE_E when transformed into an equivalent skin material based on the elastic moduli of the fibre composite skins
Simulation of the asymmetrical beam shear test of the glued composite sandwich beams using Strand7 finite element program [18] has been performed to determine if the behaviour and the ultimate capacity of the composite sandwich beams could be predicted using the material properties established from the test of coupons. The FE model is created with the nominal dimensions of the composite sandwich beams using brick elements to replicate the actual geometry of the specimens. The FE model for composite sandwich beams is analysed by modelling the skin and the phenolic core as 20-node hexahedron (Hexa20) brick elements. The brick elements have aspect ratios of 1.4–1.9. Only half of the glued composite sandwich beam is modelled in asymmetrical beam shear as the specimen is symmetric along the XY plane to reduce the computational time. The established constitutive behaviour of the skin and the core materials was used in the analysis and assumption that no debonding failure between the skin and the core was also made. Further, the load spreader beam is included in the model to simulate the asymmetrical loading condition on the sandwich beam specimen and modelled using brick elements. Table 4 lists the number of Hexa20 bricks, nodes and the computation time for the FE model of the glued composite sandwich beams in shear while Fig. 6 shows the numerical model to simulate the shear behaviour in the flatwise and edgewise positions. The numerical analysis was conducted using the static solver in Strand7 finite element software in the FCD-XPP-034 computer (CPU-Intel P4). Under asymmetrical beam shear test, a load of 1 kN was applied to the specimen through the spreader beam. After analysis, a linear load combination factor was applied to the initial linear static load case result to generate the shear response of the sandwich beams for higher load increments. The behaviour of the sandwich beam was examined at each load increment and the load case where the maximum shear stress and strain of the constituent materials were exceeded was determined. This load was chosen as the failure load of the glued composite sandwich beam in shear. In the flatwise position, the glued sandwich beam is assumed to fail when the phenolic core material reaches its failure shear stress and strain. In the edgewise position, the first shear cracking of the core is considered when the maximum shear stress and strain of the core material of the sandwich beam were reached while the final failure of specimen is when
Table 4 FE model for sandwich beam under asymmetrical beam shear test. Specimen
Hexa20 bricks
Nodes
CPU time (sec)
AS-1LSW-F AS-1LSW-E AS-2LSW-F AS-2LSW-E AS-3LSW-F AS-3LSW-E AS-4LSW-F AS-4LSW-E
3380 5528 8598 9032 14,616 15,534 25,248 26,472
21,703 29,426 39,156 41,071 66,564 70,643 111,876 117,329
1858 2073 2874 3207 3996 3502 7095 7922
1325
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
Load spreader beam
Load spreader beam Loading rollers
Sandwich beam specimen
Loading rollers Loading plates
Sandwich beam specimen
Roller supports
Support plates
(a) flatwise
(b) edgewise
Fig. 6. FE model for glued sandwich beam under asymmetrical beam shear test.
Table 5 Predicted and actual failure load of composite sandwich beams in shear. Specimen
AS-1LSW AS-2LSW AS-3LSW AS-4LSW
Flatwise (kN)
Edgewise (kN)
Actual
1
2
FEM
Actual
3
4
FEM
10.52 19.72 37.45 70.62
13.11 27.25 49.93 90.62
11.16 22.77 42.98 75.27
11.00 23.00 42.00 75.00
17.61 40.35 78.35 138.36
23.61 45.03 84.99 148.87
16.36 31.41 59.17 103.87
17.00 31.50 61.00 105.00
the shear stress and strain of the fibre composite skins were exceeded. 6
6. Predicted results and comparison with experiments
Stress (MPa)
5 4 3 Expt-2F
Expt-3F
2
Expt-4F
6.1. Failure load
FEM-2F
1
FEM-3F FEM-4F
0 0
3000
6000
9000
12000
Strain (microstrains) Fig. 7. Shear stress and strain relationship of sandwich beams in the flatwise positions.
40
30
Stress (MPa)
The results of the analytical prediction and numerical simulations on the behaviour of glued structural composite sandwich beams and the comparison with the results of the experimental investigation are discussed in this section.
20 Expt-2E Expt-3E Expt-4E
10
FEM-2E FEM-3E FEM-4E
0 0
5000
10000
15000
20000
Strain (microstrains) Fig. 8. Shear stress and strain relationship of sandwich beams in the edgewise positions.
Table 5 summarises the actual failure load based on the experimental investigations and the predicted failure load in shear of the glue-laminated composite sandwich beams using Eqs. (1)–(4) and based on the FEM simulations. In the flatwise position, the predicted failure load using Eq. (1) gives an almost 25–38% higher than the actual failure load while the difference between the predicted and the actual load using Eq. (2) is only 6–14%. The results suggest that the shear strength of the glued sandwich beams tested in the flatwise position can be predicted accurately when all the materials are transformed into an equivalent core using the ratio of the shear moduli. The higher predicted load when all the materials were converted into a homogenous section using the ratio of the elastic modulus results in an equivalent area of almost 75% provided by the skin while the core provides only 25%. Using the ratio of the shear moduli of the constituent materials resulted in a section with an equivalent area with only 45% contribution from the skin and 55% from the core. For specimens tested in the edgewise position, the predicted failure load using Eq. (3) gives a 34% higher predicted load for individual sandwich lamination but only 8–12% higher load compared than the actual failure load. On the other hand, the predicted failure load using Eq. (4) gives a 7% lower load than the actual for individual sandwich laminations but an almost 25% lower load than the actual failure load for glued sandwich beams. The higher actual than the predicted failure load of glued laminated sandwich beams is due to the laminate restraining effect of the bonded fibre composite skins as explained in Section 4.2. It can be observed however
1326
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
that this predicted load is comparable to the load when the first shear crack on the specimen was observed (Table 3). However, a closer predicted failure load to the actual failure load can be obtained using Eq. (4) if the 20% increase in shear strength of the glued composite sandwich beam in the edgewise position observed in the experimental investigation would be accounted in the theoretical prediction. This suggests that the maximum shear strength of the glued composite sandwich beam in the edgewise position could be predicted accurately when all the materials in the beam section is converted into an equivalent skin material using the shear moduli of the constituent materials. Using the maximum shear stress of the skin and the core of the composite sandwich beam, the FEM simulations was successful in the estimation of the failure load of the glued sandwich beams under asymmetrical beam shear test. Also, the failure load determined from the numerical simulation is almost the same with the predicted failure load using Eqs. (2) and (4) for glued sandwich beams in the flatwise and the edgewise positions, respectively. In all the FEM simulations, the failure of the glued composite sandwich beams occurred at the region of maximum shear. 6.2. Shear stress–strain behaviour The comparison of the predicted and the actual shear stress– strain relationship of the glued composite sandwich beams tested in the flatwise and in the edgewise positions are shown in Figs. 7 and 8, respectively. In these figures, the results of the experiment and the FEM simulations are designated with Expt and FEM, respectively. In the experimental shear stress–strain relation curve,
the shear stress of the glued composite sandwich beam is calculated by dividing the shear force with the transformed area of the composite section. For flatwise specimens, the cross section of the glued composite sandwich beam was transformed into an equivalent core material while the edgewise specimen into an equivalent skin material. This assumption was based on the observed failure behaviour of the glued composite sandwich beams in an asymmetrical beam shear test. The shear strain is determined from the indicated normal strains of the ±45° strain gauges attached to the specimen. The predicted shear stress and strain relationship using the FEM analysis showed a good agreement with the experiment. As illustrated in Fig. 7, the glued composite sandwich beam tested in the flatwise positions failed at a shear strain of around 8000 microstrains. This strain level is similar to the failure strain where the phenolic core material fails in shear. The glued sandwich beam failed at shear strength slightly higher than that of the core material due to the contribution of the fibre composite skin. In the edgewise position, gluing the sandwich beams together resulted to an almost linear shear stress–strain behaviour until failure (Fig. 8). In comparison to the experimental results, the sandwich beams in the edgewise position failed at a slightly lower level of shear strength and strains. This could be due to the reinforcing effect of the epoxy adhesive to the bonded vertical fibre composite skins which was not accounted for the FEM simulations. This complex behaviour exhibited by the composite sandwich beam in the edgewise position is beyond the capabilities of the Strand7 finite element software and was not incorporated into the finite element model.
Fig. 9. Predicted failure of glued sandwich beams under asymmetrical beam shear.
A.C. Manalo et al. / Construction and Building Materials 47 (2013) 1317–1327
6.3. Failure mechanism Fig. 9 shows the failure behaviour of glued composite sandwich beams under asymmetrical beam shear test based on the FEM simulations. The failure mechanisms based on the results of the numerical simulation showed good agreement with the failure behaviour observed in the experimental investigation. In all of the specimens tested in the flatwise and the edgewise positions, the shear failure of the glued composite sandwich beams occurred in the region of the maximum shear. In the flatwise position, the composite sandwich beams failed due to shear failure of the core on the topmost and the bottom most sandwich laminations which progressed to the core of the inner sandwich laminations at higher load cases (Fig. 9a and c). In comparison with the actual failure behaviour, the shear failure of the core material for glued sandwich beams tested in the flatwise position occurred simultaneously in all sandwich laminations after the appearance of the first shear crack. This could be due to the large energy released when shear failure of the core material in the topmost and bottom most sandwich laminations occurred which caused the total collapse of the sandwich beam. This complex failure behaviour in shear of the sandwich beam in the flatwise position could not be modelled in Strand7 but the load when the core shear failure occurs can be predicted accurately. In glued composite sandwich beams tested in the edgewise position, the predicted failure mode is shear failure of the fibre composite skins. In the numerical simulation, the maximum shear stress and strain of the core material is exceeded at a load comparable to the applied load when the first shear crack of the core material was observed experimentally. However, this was not considered as the failure load of the sandwich beam tested in the edgewise position. As observed in the experimental investigation, the presence of the vertical fibre composite skins inhibits the development of shear cracks in the core material and prevented the immediate failure of the sandwich beams. Thus, the final failure of the glue-laminated sandwich beam in the edgewise position is due to shear failure of the vertical fibre composite skin (Fig. 9b and d). The load to reach this type of failure is comparable to the theoretically predicted failure load. 7. Conclusions The shear behaviour of glued structural composite sandwich beams was determined through experimental, analytical and numerical investigations. Structural sandwich beams with 1, 2, 3, and 4 composite sandwich panels glued together were subjected in asymmetrical beam shear test in the flatwise and in the edgewise positions. The results showed that the number of sandwich laminations has no significant effect on the shear strength of the glued composite sandwich beams tested in the flatwise position. In this position, the shear strength of the glued sandwich beam is governed by the shear strength of the core material regardless of how many laminations are there in a beam section. In the edgewise position, the shear behaviour of the glued sandwich beam is improved significantly with the presence of the vertical fibre composite skins. In this position, the fibre composite skins carries almost 60% of the applied load compared in the flatwise position where the skin carries only 20% of the applied load. With increasing sandwich laminations, the glued composite sandwich beams in the edgewise position achieved over 200% shear strength than beams in the flatwise position. The vertical fibre composite skins has also resulted in a progressive failure and a more ductile failure behaviour of the glued structural composite sandwich beams in the edgewise position. The prediction equation for shear strength when all the materials are transformed into an equivalent area based on the shear
1327
properties of the constituent materials predicted accurately the failure load of the glued composite sandwich beam in the flatwise position but gives a lower predicted load than the actual in the edgewise position. A closer predicted load to the actual failure load can be obtained if the percentage increase in shear strength in gluing the sandwich beams in the edgewise position would be accounted for. The prediction equation agreed very well with the experimental results and the FEM simulations, giving confidence in the validity of the assumptions adopted. The results of the study showed a significant improvement on the shear strength of the glued sandwich beams when the sandwich panels are positioned edgewise. This point towards the high feasibility of using glued composite sandwich beams for construction and building applications. A research and development project is now being implemented to develop full-size glued structural composite sandwich beams with the objective of exploring the practical application of this composite structure in replacement railway timber sleeper. Acknowledgements The authors gratefully acknowledge the LOC Composites Ltd., Pty., Australia for providing the composite sandwich panels used in this study. The support of Wayne Crowell, Atul Sakhiya and Christopher Pickford in the preparation and testing of the glued composite sandwich beams is greatly acknowledged. References [1] Bakis CE, Bank LC, Brown VL, Cosenza E, Davalos JF, Lesko JJ, et al. FRP composites in construction – state of the art review. ASCE J Compos Constr 2002;6(2):78–87. [2] Belouettar S, Abbadi A, Azari Z, Belouettar R, Freres P. Experimental investigation of static and fatigue behaviour of composite honeycomb materials using four point bending tests. Compos Struct 2008;87(3):265–73. [3] Keller T. Material tailored use of FRP composites in bridge and building construction. Switzerland: Swiss Federal Institute of Technology Lausanne; 2006. [4] Canning L, Hollaway L, Thorne AM. Manufacture, testing and numerical analysis of an innovative polymer composite/concrete structural unit’, Proceedings of Inst. Civil Eng Struct Build 1999;134:231–41. [5] Lopez-Anido R, Xu H. Structural characterization of hybrid fibre-reinforced polymer-glulam panels for bridge decks. ASCE J Compos Constr 2002;6(3):194–203. [6] Williams B, Shehata E, Rizkalla S. Filament-wound glass fibre reinforced polymer bridge deck modules. ASCE J Compos Constr 2003;7(3):266–76. [7] Van Erp G, Rogers D. A highly sustainable fibre composite building panel. In: Proceedings of the international workshop on fibre composites in civil infrastructure – past, present and future. University of Southern Queensland, Toowoomba, Queensland, Australia; 1–2 December 2008. [8] Manalo AC, Aravinthan T, Karunasena W, Islam MM. Flexural behaviour of structural fibre composite sandwich beams in flatwise and edgewise positions. Compos Struct 2010;92(4):984–95. [9] Manalo AC, Aravinthan T, Karunasena W. In-plane shear behaviour of fibre composite sandwich beams using asymmetrical beam shear test. J Constr Build Mater 2010;24(10):1952–60. [10] Manalo AC, Aravinthan T, Karunasena W. Flexural behaviour of glue-laminated fibre composite sandwich beams. Compos Struct 2010;92(11):2703–11. [11] Hernandez R, Davalos JF, Sonti SS, et al. Strength and stiffness of reinforced Yellow-Poplar glued-laminated beams. In: Res. Pap. FPL-RP-554. Madison (WI): US Department of Agriculture, Forest Service, Forest Products Laboratory; 1997. 28 p.. [12] Rammer DR. Shear strength of glued-laminated timber beams and panels. Forest Products Laboratory, USDA Forest Service; 1996. p. 192–201. [13] Kampner M, Grenestedt JL. On using corrugated skins to carry shear in sandwich beams. Compos Struct 2007;85:139–48. [14] Mirsayah AA, Banthia N. Shear strength of steel fibre reinforced concrete. ACI Mater J 2002;99(5):473–9. [15] Zhou G, Green ER, Morrison C. In-plane and interlaminar shear properties of carbon/epoxy laminates. J Compos Sci Technol 1995;55:187–93. [16] Gupta R, Siller TS. A comparison of the shear strength of structural composite lumber using torsion and shear block tests. Forest Prod J 2005;55(12):29–34. [17] Triantafillou TC. Composites: a new possibility for the shear strengthening of concrete, masonry and wood. J Compos Sci Technol 1998;58:1285–95. [18] Strand7. Strand7 Release 2.3.7 finite element analysis system. Sydney, Australia: Strand7 Software; 2005.