The flexural behaviour of composite sandwich beams with a lattice-web reinforced wood core

Abstract

In this study, the flexural behaviour of lattice-web reinforced composite sandwich beams was investigated. The innovative composite sandwich beams consisted of glass fibre reinforced composite materials as the surface layer and web, Paulownia wood and South Pine as the core material. While keeping the external size unchanged, 1, 2, 3, 4 and 5 wood cores were formed by the vacuum infusion process at one time, to develop a different number of lattice-webs. In addition, the wood beam was used as a control. A four-point bending test was carried out on all specimen beams. The equivalent bending stiffness, equivalent shear stiffness, mid-span deflection of the specimens and bearing capacity were analysed. There is a larger plateau stage in the latter part of the load-midspan deflection curve of the beams strengthened by the lattice-web, which means the beams had ductile failure characteristics. Furthermore, the type of wood core greatly influenced the flexural behaviour of the composite sandwich beam, and the optimum lattice-web number of the Paulownia specimen is 2, and that of the South Pine specimen is 3.

Keywords

Lattice-web reinforcement composite sandwich beam wood core flexural behaviour comparative results

Introduction

Composite sandwich structures have been widely employed in aviation, military, transport, and civil applications for their unique characteristics, including their high specific bending stiffness, weight efficiency, structural designability and excellent corrosion resistance (Hollaway and Teng, 2008; Hassan et al., 2020; Ou et al., 2019, 2020). Currently, it has been increasingly difficult to meet the important multifunctional requirements in civil, marine and military engineering projects. Therefore, research focusing on composite sandwich structures is necessary in the present situation.

Composite sandwich members have been used as structural beams for roofing and floor slabs, floors, wall beams and bridge decks (Fang et al., 2019; Fernando et al., 2018; Miao et al., 2019). Commonly used foam and Balsa light wood core materials are soft, and are prone to sag when the local pressure substantially increases (Hou et al., 2020; Tuwair et al., 2016; CoDyre and Fam, 2016). Although the honeycomb or truss sandwich structure can withstand a greater pressure, the interfacial properties between the core and the facing are weaker due to the presence of a cavity (Lanssens et al., 2014; Correia et al., 2012; Ayrilmis et al., 2014). Therefore, in this paper, the use of Paulownia (PAW) wood and South Pine (SOP) wood as the core materials of the composite sandwich beam is proposed; these materials have excellent mechanical properties, such as shear resistance and compression resistance.

Manalo et al. (2010a) studied the flexural behaviour of composite sandwich beams with a glass fibre reinforced composite (GFRP) as the face sheet and modified phenolic foam as the core material; this beam was applied to railway sleeper structures. Wang et al. (2015) developed a simple and innovative sandwich beam with GFRP face sheets and a foam-GFRP web core. Satasivam and Bai (2014) presented a mechanically bolted modular assembly system of GFRP web-flange sandwich structures used for beam and slab applications. Fang et al. (2015) developed and studied innovative GFRP-bamboo-wood sandwich beams experimentally and by modelling.

In the early stages of our research, our group conducted related studies on composite sandwich beams and beams. Li et al. (2020) performed a three-point bending creep experiment on GFRP-balsa sandwich beams, and investigated the flexural creep behaviour of one-web and two-web reinforced GFRP-balsa sandwich beams. Qi et al. (2017) investigated the flexural behaviour of novel composite sandwich beams that feature a PAW or SOP wood core and GFRP face-skins reinforced with lattice-webs. Chen et al. (2018) conducted experimental and numerical analysis of the nonlinear flexural behaviour of lattice-web reinforced foam core composite sandwich beams. Zhu et al. (2018) investigated the bending behaviour of an innovative fibre reinforced polymer sandwich beam for application in a multistory building.

The lattice-web reinforced composite sandwich beam proposed in this paper is made by juxtaposing several wooden cores of the same wrapped fibre cloth in the horizontal direction (Gerber and Crews, 2009). The vacuum introduction of the resin allows the sheet to be integrally formed in the mould at one time, ensuring the integrity of the member and effectively improving the load carrying capacity of the sandwich beam. The bending test is carried out to study the influence of the number of lattice-webs and the type of wood on the bending properties of the composite sandwich beams.

Experimental programme

Test specimens

The test specimens were composed of a GFRP surface layer, a lattice-web and a wood core. The lattice-webs and surface layer were respectively wrapped with two layers of biaxial glass fibre cloths, and the areal density of the glass fibre cloths was 800 g/m², as shown in Figure 1. The calculated thickness of each layer of fibre cloth and resin after curing was 1 mm. The height of the beam as the control group was 70 mm. The height of the other test specimens was set to 80 mm, the specimen width was 120 mm, and the length was 1400 mm. The details of each specimen are summarized in Table 1.

Figure 1.

Lattice-web reinforced composite sandwich beam.

Table 1.

Descriptions of the specimens.

Specimens	Illustration	Size of the specimen			Size of the core
Specimens	Illustration	L (mm)	B (mm)	H (mm)	b (mm)	h (mm)
PAW/SOP		1400	110	70	110	70
1PAW/1SOP		1400	120	80	110	70
2PAW/2SOP		1400	120	80	53	70
3PAW/3SOP		1400	120	80	34	70
4PAW/4SOP		1400	120	80	24.5	70
5PAW/5SOP		1400	120	80	18.8	70

Notes: The green frame represents the GFRP surface layer and the lattice-web. The test specimens “PAW” and “SOP” represents Paulownia wood and South Pine, respectively. The front numbers represent the number of core materials. “PAW” and “SOP” are all pure wood beams and are the control groups. “L" represents the length of the specimen. “B" represents the width of the specimen. “H" represents the height of the specimen.

Test material preparation

The wood core was cut into the required specifications, planed and slotted along the grain, as shown in Figure 2(a); the single core was wrapped with two layers of the (±45°) glass fibre cloth. After assembling multiple pieces of the core material, the whole periphery was wrapped with two layers of (0,90°) glass fibre cloths to form the surface layer (the 0-direction refers to the length direction along the test specimen), and the glass fibre cloths were as flat as possible. The starting point of the wrap was arbitrary, but was not located at the edge of the specimen. The finish point of the wrap was 5 cm above the starting point, which means that there was an overlap of 5 cm. After that, the demoulding cloth, diversion net and vacuum bag were arranged, and the resin was poured with the aid of a vacuum to solidify and form the beam, as shown in Figure 2(b)–(d).

Figure 2.

Preparation and manufacturing process of the sandwich composite beams: (a) sheet cutting and forming; (b) vacuum introduction operation; (c) arrangement of the fibre cloths; and (d) completion of specimen production.

Test setup and instrumentation

The beam was subjected to four-point bending static loading, the distribution beam was arranged at two loading points, as shown in Figure 3(a), and the strain gauge was placed at the mid-span, front, bottom and loading points of the beam, as shown in Figure 3(b). Displacement and strain acquisition were performed using the DH3816N static strain test system.

Figure 3.

Experimental setup of the flexural behaviour of the beam: (a) schematic plot and (b) strain gauge arrangement.

Experimental observation and results

Experimental observation

(1) Experimental observation of the Paulownia wood composite sandwich beam

For the pure PAW wood specimen, the stiffness of the beam was small due to the lack of fibre composite wrapping on the outside, and the deflection increased rapidly due to the increasing load. When the load reached 9.30 kN, a cracking sound was heard, and the crack in the specimen is shown in Figure 4(a). The fracturing of the specimen was the result of brittle failure.

Figure 4.

Bending failure modes of the wood core composite beam: (a) failure mode of PAW; (b) 1PAW failure mode; (c) 2PAW failure mode; (d) 3PAW failure mode; (e) 4PAW failure mode; and (f) 5PAW failure mode.

For specimen 1PAW, when the load increased to 19.7 kN, a slight tearing sound was heard. When the load climbed to 22.5 kN, the tearing sound of the glass fibre cloth was transmitted from the specimen. The fibre was whitened and cracked at the lower position of the loading point. The crack extended vertically downward with the deflection of the specimen as shown in Figure 4(b). The maximum bearing capacity of the beam was 24.7 kN. After that, although the deflection of the beam continued to increase, the bearing capacity of the beam decreased, and the final shear failure occurred at the loading points at both ends. The ultimate load and corresponding deflection of the specimens 2PAW, 3PAW, 4PAW and 5PAW were 34.3 kN, 46.7 mm; 31.65 kN, 39.84 mm; 29.3 kN, 43.79 mm and 20.18 kN, 27.9 mm, respectively. The experimental observation of the other four specimens are similar to those of 1PAW. The increase in load caused the mid-span deflection of the beam to gradually increase, and a small tearing sound was heard from time to time. When the load reached a certain value, a large amount of noise was heard. The fibre was whitened, cracked and extended downward, as shown in Figure 4(c)–(f). As the displacement of the specimen continued to increase, the load of the specimen began to decline slowly. Eventually, shear failure occurred at the loading point of both ends, and the load stopped at that moment.

(2) Experimental observation of the South Pine composite sandwich beam

The experimental observation of the SOP sandwich beam were similar to those of the corresponding PAW sandwich beam. The sound accompanied by SOP failure was brittle failure; a tiny fibre cracking voice was heard when 1SOP, 2SOP, 3SOP, 4SOP and 5SOP were slowly loaded. As the load increased, the tearing sound became obvious, and then a loud noise was heard suddenly. There were cracks under the loading point, and then the cracks developed downward. Finally, the fibre at the lower edge of the beam at the loading point cracked.

Experimental results of Paulownia wood composite sandwich beam

The load-displacement curve of the PAW wood sandwich composite beam is shown in Figure 5. Among the PAW, 1PAW, 2PAW, 3PAW, 4PAW, and 5PAW specimens, the ultimate load of the specimen and the corresponding mid-span displacement are 9.3 kN, 22 mm; 24.7 kN, 37.1 mm; 34.3 kN, 46.7 mm; 31.65 kN, 39.84 mm; 29.3 kN, 43.79 mm; 26.15 kN, 27.9 mm respectively. Compared with the wooden beam, the ultimate bearing capacity of the wood core composite beam without the lattice and the wooden core composite beam with the lattice structure is obviously improved. The ultimate bearing capacity of the wood core composite beam without the lattice structure is 166.6% higher than that of the wooden beam. The maximum bearing capacity of the wood core composite beam with the lattice structure is increased by 38.9%, 28.3%, 18.6%, and 5.9%.

Figure 5.

Load-midspan deflection curve of the PAW wood core beam.

The load-strain curve of the PAW wood core composite beam is shown in Figure 6.

Figure 6.

Load-midspan strain curve of the PAW wood core beam.

Experimental results of the South Pine composite sandwich beam

The load-displacement curve of the SOP wood core composite beam is shown in Figure 7. The figure shows that the ultimate loads of SOP, 1SOP, 2SOP, 3SOP, 4SOP and 5SOP are 13.43 kN, 35.38 kN, 37.28 kN, 40.54 kN, 29.68 kN and 33.44 kN, respectively, and the corresponding mid-span displacements are 27.8 mm, 35.91 mm, 36.72 mm, 44.12 mm, 32.05 mm and 40.24 mm, respectively. The ultimate bearing capacity of the wood core composite beam without the lattice is greatly increased by 152.6% compared with that of the pure wood beam. The ultimate bearing capacity of 2SOP and 3SOP specimens with the lattice is increased by 5.4% and 14.6%, respectively, although the ultimate bearing capacity of 4SOP and 5SOP decreased by 16.1% and 5.5%, respectively, compared with that of 1SOP.

Figure 7.

Load-midspan strain curve of the SOP wood core beam.

The load-strain curve of the SOP wood core composite beam is shown in Figure 7.

Test comparison

From a comparison of Figures 5 and 8, it can be seen that the ultimate load and deflection of the SOP wood core composite beam corresponding to the cross-section form are larger than those of the PAW wood core beam, in which the ultimate bearing capacity of the SOP beam is 44.41% higher than that of the PAW beam. The ultimate bearing capacity of the wood core composite beam with the lattice structure is increased by the greatest extent by 1SOP, and its ultimate bearing capacity is higher than that of the setting. The stiffness of the latticed wood core composite beam 1PAW is increased by 43.88%, and the test results of the SOP wood core composite beam are more ideal. From the load-midspan strain curves of Figures 6 and 7, it can be learned that the stiffness of the SOP wood core composite beam is greater, and the stiffness of the unrelated wood core composite beam 1SOP increased 54.55% compared to that of the unrelated wood core composite beam 1PAW. The stiffness of 1SOP is 45.31% higher than that of 1PAW, and the change in the stiffness is more obvious with the change in the lattice number. The composite sandwich beam with the largest number of lattice-web does not have the optimal bending performance. Based on the analysis of the curves, 3PAW has the largest stiffness in the PAW specimen, and 2SOP has the largest stiffness in the SOP specimen, which is not the maximum number of lattice-web for this kind of core material.

Figure 8.

Load-midspan deflection curve of the SOP wood core beam.

Theoretical analysis

Equivalent flexural stiffness

According to the plane section assumption, the flexural stiffness of the non-lattice wood core beam under compression is equal to the sum of the flexural stiffness of each part (Manalo et al., 2010b), as shown in Figure 9(a)

{(E I)}_{1 e d g e} = E_{f} \frac{b t_{f}^{3}}{6} + E_{f} \frac{b t_{f} {(t_{c} + t_{f})}^{2}}{2} + E_{w} \frac{t_{w} t_{c}^{3}}{6} + E_{c} \frac{b_{c} t_{c}^{3}}{12}

(1)

Figure 9.

Section of the specimens: (a) section of 1PAW; (b) section of 2PAW; and (c) section of 3PAW.

For the section of 2PAW (as shown in Figure 9(b))

{(E I)}_{2 e d g e} = E_{f} \frac{b t_{f}^{3}}{6} + 2 E_{f} b t_{f} {(\frac{t_{c} + t_{f}}{2})}^{2} + E_{w} \frac{t_{w} t_{c}^{3}}{6} + E_{c} \frac{b_{c} t_{c}^{3}}{6} + E_{g} \frac{t_{g} t_{c}^{3}}{12}

(2)

where t_f and t_w are the thicknesses of the surface layer and the web, respectively; b and b_c are the section widths of the sandwich structure and the single core material, respectively; E_f, E_c , E_g and E_w are the elastic moduli of the surface layer, the core material, the lattice-web and the web, respectively; and t_c and t_g are the thicknesses of the single core material and the lattice-web, respectively.

For the 3PAW section (as shown in Figure 9(c))

{(E I)}_{3 edge} = E_{f} \frac{b t_{f}^{3}}{6} + 2 E_{f} b t_{f} {(\frac{t_{c} + t_{f}}{2})}^{2} + E_{w} \frac{t_{w} t_{c}^{3}}{6} + E_{c} \frac{b_{c} t_{c}^{3}}{4} + E_{g} \frac{t_{g} t_{c}^{3}}{6}

(3)

The same derivation process can be used for the lattice wood core beams with 4 or 5 layers of core material. The theoretical flexural stiffness formula can be summarized as

{(E I)}_{e q} = \frac{n t_{c}^{3}}{12} (t_{w} E_{c} + b_{c} E_{w}) + \frac{(n - 1)}{12} t_{g} t_{c}^{3} + [\frac{b t_{f}^{3}}{6} + 2 b t_{f} {(\frac{t_{f} + t_{c}}{2})}^{2}] E_{f}

(4)

Table 2 compares the theoretical value of the flexural stiffness with the test value, and the maximum difference between the experimental value and the theoretical value is 19.70%. The

E_{p r e}, E_{t e s t}, E_{p r e} / E_{t e s t}

, the mean value and the coefficient of variation(C.V) have been listed in Table 3. It can be concluded that the theoretical stiffness and the experimental value are basically consistent.

Table 2.

Comparison between the theoretical and test values of the flexural stiffness.

Specimen	EI_ex (kN·m²)	EI_eq (kN·m²)	D-value (+%)
PAW	25.3	30.2	19.4
1PAW	26.1	30.4	16.3
2PAW	29.7	30.6	3.1
3PAW	29.6	30.8	4.1
4PAW	26.2	31.1	18.8
5PAW	39.0	40.2	3.0
SOP	27.8	25.1	9.7
1SOP	35.9	33.7	6.3
2SOP	38.0	40.7	7.1
3SOP	39.3	41.1	4.7
4SOP	37.1	41.7	12.3
5SOP	35.1	42.0	19.7

Table 3.

Comparison between the theoretical and test values of the bending modulus of elasticity.

Specimen	$E_{p r e}$ (N/mm²)	$E_{t e s t}$ (N/mm²)	$E_{p r e} / E_{t e s t}$
PAW	8046.6	9603.5	0.84
1PAW	8301.3	9667.1	0.86
2PAW	9446.3	9732.5	0.97
3PAW	9414.5	9794.3	0.96
4PAW	8333.1	9891.5	0.84
5PAW	12,404.2	12,785.9	0.97
SOP	8842.0	7983.2	1.11
1SOP	11,418.2	10,718.4	1.07
2SOP	12,086.1	12,944.9	0.93
3SOP	12,499.6	13,072.1	0.96
4SOP	11,799.9	13,262.9	0.89
5SOP	11,163.8	13,358.4	0.84
M-value	—	—	0.94
C.V	—	—	31%

Equivalent shear stiffness

Assuming that the sandwich acts like a Timoshenko beam, according to the first-order shear deformation theory, the basic assumptions are as follows: (1) the elastic analysis of small deformation is conducted; (2) the longitudinal and vertical compression deformation of core materials is not considered; (3) no relative slip between the core material and the beam is assumed; (4) the shear deformation of the beam is not considered; and (5) the deformation of core material meets the assumption of plane section.

The core material of the sandwich structure involved in this paper is semi-continuous in space, so it will be very tedious to solve it directly by numerical methods. At present, it is common to adopt the equivalent method to deal with the calculation of the sandwich structure with wood core material. The core material is considered to be equivalent according to certain principles to obtain a homogeneous and continuous core material (Manalo et al., 2010c). It is convenient to use the classical sandwich beam theory to solve the questions in this way.

According to the principle of the shear stiffness equivalence, the wood core material can be equivalent to a web. The section form of 1PAW is taken as the theoretical analysis model, as shown in Figure 10, which is the model comparison diagram before and after the equivalence. The parameters t_f, t_c and t_w^’ are the thickness of the surface layer, the height of the web and the width of the web after equivalence, respectively; G_w is the shear modulus of the lattice-web; and b is the total width of the section. Therefore, the equivalent shear stiffness is

{(A G)}_{e q} = \frac{2 t_{w}^{'} b (t_{c} + t_{f})}{b} G_{w} = 2 t_{w}^{'} (t_{c} + t_{f}) G_{w}

(5)

Figure 10.

Equivalent models of 1PAW: (a) pre-equivalent model and (b) post-equivalent model.

Likewise, a similar equivalent method can be used for other forms of lattice-web wood core beams. According to the shear stiffness equivalence principle, the wood core material is equivalent to two sides of the lattice-web, as shown in Figure 11, where t_m is the width of the lattice-web after equivalence, and the distanceΔ_x is an element taken from the equivalent sandwich beam for analysis. Therefore, the equivalent shear stiffness of the sandwich layer is

{(A G)}_{e q} = \frac{t_{m} b (t_{c} + t_{f})}{Δ x} G_{w}

(6)

Figure 11.

Initial and equivalent model: (a) pre-equivalent model and (b) post-equivalent model.

Deflection

According to first-order shear deformation theory, it is basically assumed that the composite surface layer and the wood core material are well bonded, no interface delamination failure occurs, and the strain distribution along the cross-sectional height under the bending moment is a continuous straight line, which meets the flat-section assumption. According to Timoshenko beam theory (Allen, 1969), when the thickness of the core material is large and the load p is applied to the beam, the deflection is composed of the bending deformation w₁ and the shear deformation w₂.

Δ = w_{1} + w_{2} = \frac{9 P L^{3}}{512 {(E I)}_{e q}} + \frac{3 P L}{16 {(A G)}_{e q}}

(7)

where (EI)_eq and (AG)_eq are the equivalent flexural stiffness and the equivalent shear stiffness, respectively; L is the net span of the sandwich beam; and A is the section area of the core material.

The deflection of each specimen was calculated by formula (7), and compared with the test value, as shown in Table 4.

Table 4.

Comparison between the experimental and theoretical values of the deflection.

Specimen	Test value (mm)	Theoretical value (mm)	Difference (%)
PAW	22.4	20.2	8.2
1PAW	37.1	33.4	10.0
2PAW	46.7	42.9	8.1
3PAW	39.8	37.6	5.5
4PAW	43.8	40.2	8.2
5PAW	42.1	38.9	7.6
SOP	27.8	25.1	9.7
1SOP	35.9	33.7	6.3
2SOP	36.7	33.9	7.6
3SOP	44.1	41.4	6.3
4SOP	32.1	28.9	10.0
5SOP	31.2	27.8	10.9

It can be seen from Table 4 that the experimental value and the theoretical value are in good agreement. Due to the defects of the test piece and other uncertain factors, the value of the mid-span deflection obtained by the test is smaller than the theoretical value, so the theoretical design value is slightly conservative.

Failure mode and ultimate bearing capacity

Under the bending load, the composite sandwich structure may have different failure modes because of the different component materials and size arrangement. The failure modes can be summarized as follows: 1) compression yield or tension fracture of the beam; 2) buckling of the compression beam; 3) shear of the core material; and 4) delamination failure. Considering that many factors may affect the shear and delamination failure of the core materials, such as different processes, interface treatment forms and resin properties, this paper focuses on the ultimate bearing capacity corresponding to the first two failure modes according to the observed failure modes.

(1) The beam is subjected to tensile fracture or compression yield critical load

p = B_{3} σ_{y f} b c (\frac{t}{l})

(9)

(2) Buckling critical load of the compression beam

σ_{c r} = C_{1} E_{f}^{\frac{1}{3}} E_{c}^{\frac{1}{3}} G_{c}^{\frac{1}{3}}

C_{1} = 3 {[12 {(3 - ν_{c})}^{2} {(1 + ν_{c})}^{2}]}^{\frac{- 1}{3}}

(10)

where B₃=4; b is the width of the specimen; c is the distance from the centre point of the upper and lower layers; t is the thickness of the surface layer; l is the span of the specimen; σ_yf is the compressive fracture strength of the beam; ν_c is the core material Poisson’s ratio; E_f is the beam elastic modulus; E_c is the elastic modulus of the sandwich layer; and G_c is the shear modulus of the sandwich layer. E_f, E_c and G_c can be calculated by the method adopted in reference (Jin and Deng, 2009).

Prediction of the ultimate bearing capacity

According to the theory of classical sandwich beams, the theoretical value of the ultimate bearing capacity of the PAW wood sandwich composite beam with the lattice-web can be obtained and compared with the experimental value (see Table 5).

Table 5.

Comparison between the theorical and experimental ultimate flexural bearing capacities.

Specimen	Test value (kN)	Theoretical value (kN)	Difference (%)
PAW	9.3	8.2	11.8
1PAW	24.7	22.4	9.3
2PAW	34.3	30.2	12.0
3PAW	31.7	28.8	9.0
4PAW	29.3	26.4	9.9
5PAW	20.2	22.8	13.1
SOP	34.4	30.5	11.4
1SOP	35.4	32.2	9.0
2SOP	37.3	34.2	8.3
3SOP	40.5	35.8	11.7
4SOP	29.7	26.5	10.7
5SOP	33.4	36.4	8.8

It can be seen from Table 5 that the measured values of the ultimate bearing capacity of each test piece are in good agreement with the theoretical values, indicating that the ultimate failure load of each test piece predicted by the above theoretical model is more accurate.

Influence of the lattice-web on the ductility

The ductility index reflects the ductility of a component by its displacement ductility coefficient. The displacement ductility coefficient of a component is defined as the ratio of the ultimate displacement to the yield displacement of a structure after yielding.

μ_{Δ} = \frac{Δ_{μ}}{Δ_{y}}

(11)

whereΔ_μ is the displacement of the specimen in the limit state andΔ_y is the maximum displacement of the structural member under the elastic yielding load. The elastic yielding point can be determined using the method from Feng et al. (2015) ; i.e. the yielding point is the point on the curve with a tangent slope that is identical to the slope of the straight line connecting the origin point to the peak point. If there are two or more points, then the yielding point can be determined from the projection of the average load values. The calculated values of the displacement ductility coefficient of the sandwich beam are shown in Table 6.

Table 6.

Displacement ductility coefficient of the PAW wood core beam.

	Wood core beam without the lattice-web	Wood core beam with the lattice-web
Specimen	1PAW	2PAW	3PAW	4PAW	5PAW
Δ_μ(mm)	36.29	46.67	39.84	43.79	27.90
Δ_y (mm)	31.55	38.57	28.67	26.91	22.73
$μ_{Δ}$	1.15	1.21	1.39	1.62	1.23
Specimen	1SOP	2SOP	3SOP	4SOP	5SOP
Δ_μ(mm)	35.91	50.10	44.12	32.05	40.24
Δ_y (mm)	23.46	29.81	29.18	23.66	25.65
$μ_{Δ}$	1.53	1.68	1.51	1.35	1.56

It can be seen from Table 6 that when PAW is used as the core material, the ductility coefficient of the lattice-web wood core beam is higher than that of the wood core beam without the lattice-web. The ductility coefficient of 4PAW is the largest, reaching 1.62, and an increase in the lattice-web in the wood core beam without the lattice-web greatly improves the ductility of the specimen. When SOP is used as the core material, only the ductility coefficients of 2SOP and 5SOP are higher than those of the wood core beam without the lattice-web, and the ductility coefficient of 2SOP is the largest, reaching 1.68, but the overall improvement effect is not as obvious as that of the PAW wood core beam.

Conclusion

In this paper, focusing on the flexural behaviour of composite sandwich beams strengthened by a lattice-web, the number of lattice-webs and the type of core material are selected as the variable parameters. According to the four-point bending test, the failure modes of the structure are observed, and the theoretical analysis of its ultimate strength, stiffness and deflection is carried out. The following conclusions are drawn:

(1) The strength and deflection of the SOP wood composite beam with the same section are larger than those of the PAW wood composite beam because the local compressive strength of SOP wood is greater than that of the PAW wood. The failure mode is that the upper face sheet of the beam is crushed first, then the core material is destroyed by extrusion, and finally the lower surface layer is broken. It can be summarized that the local compressive strength of the core material has a great impact on the failure mechanism.

(2) The ultimate bearing capacity of 2PAW, 3PAW, 4PAW and 5PAW is 39.49%, 28.71%, 19.15% and −18.30% higher than that of 1PAW, respectively. The result shows that the optimal lattice number is 2 because the ultimate bearing capacity of 2PAW has the highest percentage improvement compared to 1PAW. While the ultimate bearing capacity of 2SOP, 3SOP, 4SOP and 5SOP is 7.43%, 14.58%, −16.10% and −5.65% higher than that of 1SOP, respectively. In the same way, the result indicates that the optimal lattice number is 3. Therefore, for different core materials, the optimal lattice number is different. The wood core composite beam with the largest number of lattice-webs does not have the optimal bending performance. Based on the comprehensive test curve and calculation result analysis, 3PAW has the largest rigidity in the PAW wood specimens and 2PAW has the largest ultimate bearing capacity; 2SOP has the largest rigidity in the Southern pine specimens and the maximum ultimate bearing capacity is 3SOP. None of them demonstrates the maximum lattice-web number of the corresponding core material.

Footnotes

Declaration of conflicting interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (Grant No. 52078248), the Natural Science Foundation for Distinguished Young Scholars of Jiangsu Province (Grant No. BK20190034) and the Funds for Youth Creative Research Groups of Nanjing Tech University.

Data availability statement

The data used to support the findings of this study are available from the corresponding author upon request.

ORCID iD

Yuntiam Wang

References

Allen

(1969) Analysis and Design of Structural Sandwich Beams. London: Pergamon Press.

Ayrilmis

Kwon

Han

(2014) Improving bending and tensile properties of lignocellulosic filled polypropylene composite panels using aramid fabric. Fibers and Polymers 15(11): 2410–2415.

Chen

Fang

Liu

, et al. (2018) Experimental and numerical analysis of nonlinear flexural behaviour of lattice-web reinforced foam core composite sandwich beams. Advances in Civil Engineering: 1–11.

CoDyre

Fam

(2016) The effect of foam core density at various slenderness ratios on axial strength of sandwich panels with glass-FRP skins. Composites Part B: Engineering 106: 129–138.

Correia

Garrido

Gonilha

, et al. (2012) GFRP sandwich beams with PU foam and PP honeycomb cores for civil engineering structural applications. International Journal of Structural Integrity(3): 127–147.

Fang

Bai

Liu

, et al. (2019) Connections and structural applications of fibre reinforced polymer composites for civil infrastructure in aggressive environments. Composites Part B: Engineering 164: 129–143.

Fang

Sun

Liu

, et al. (2015) Mechanical performance of innovative GFRP-bamboo-wood sandwich beams: experimental and modelling investigation. Composites Part B: Engineering 79: 182–196.

Feng

Cheng

Bai

, et al. (2015) Mechanical behavior of concrete-filled square steel tube with FRP-confined concrete core subjected to axial compression. Composite Structures 123: 312–324.

Fernando

Teng

Gattas

, et al. (2018) Hybrid fibre-reinforced polymer-timber thin-walled structural members. Advances in Structural Engineering 21(9): 1409–1417.

10.

Gerber

Crews

(2009) Timber stressed-skin panels: design guidelines for Australian practice. Australian Journal of Structural Engineering 9(3): 207–216.

11.

Hassan

Usman

Hanif

, et al. (2020) Improving structural performance of timber wall panels by inexpensive FRP retrofitting techniques. Journal of Building Engineering 27: 101004.

12.

Hollaway

Teng

(2008) Strengthening and Rehabilitation of Civil Infrastructures Using Fibre-Reinforced Polymer (FRP) Composites. London: Woodhead Publishing and Maney Publishing.

13.

Hou

Wang

, et al. (2020) Testing of insulated sandwich panels with GFRP shear connectors. Engineering Structures 209: 109954.

14.

Jin

Deng

(2009) Research on equivalent modulus of sandwich core plane for rectangular fill materials use energy method. Chinese Quarterly of Mechanics(2): 330–336.

15.

Lanssens

Tanghe

Rahbar

, et al. (2014) Mechanical behavior of a glass fiber-reinforced polymer sandwich panel with through-thickness fiber insertions. Construction and Building Materials 64: 473–479.

16.

Liu

Fang

, et al. (2020) Flexural creep behavior of web reinforced GFRP-balsa sandwich beams: experimental investigation and modeling. Composites Part B: Engineering 196: 108150.

17.

Manalo

Aravinthan

Karunasena

, et al. (2010a) Flexural behaviour of structural fibre composite sandwich beams in flatwise and edgewise positions. Composite Structures 92(4): 984–995.

18.

Manalo

Aravinthan

Karunasena

, et al. (2010b) Flexural behaviour of glue-laminated fibre composite sandwich beams. Composite Structures 92(11): 2703–2711.

19.

Manalo

Aravinthan

Karunasena

, et al. (2010c) In-plane shear behaviour of fibre composite sandwich beams using asymmetrical beam shear test. Construction and Building Materials 24(10): 1952–1960.

20.

Miao

Fernando

Heitzmann

, et al. (2019) GFRP-to-timber bonded joints: adhesive selection. International Journal of Adhesion and Adhesives 94: 29–39.

21.

Fernando

Gattas

(2019) Experimental investigation of a novel concrete-timber floor panel system with digitally fabricated FRP-timber hollow core component. Construction and Building Materials 227: 116667.

22.

Gattas

Fernando

, et al. (2020) Experimental investigation of a timber-concrete floor beam system with a hybrid glass fibre reinforced polymer-timber corrugated core. Engineering Structures 203.

23.

Fang

Shi

, et al. (2017) Bending performance of GFRP-wood sandwich beams with lattice-web reinforcement in flatwise and sidewise directions. Construction and Building Materials 156: 532–545.

24.

Satasivam

Bai

(2014) Mechanical performance of bolted modular GFRP composite sandwich structures using standard and blind bolts. Composite Structures 117: 59–70.

25.

Tuwair

Volz

ElGawady

, et al. (2016) Testing and evaluation of polyurethane-based GFRP sandwich bridge deck panels with polyurethane foam core. Journal of Bridge Engineering 21(1): 04015033.

26.

Wang

Liu

Fang

, et al. (2015) Behavior of sandwich wall panels with GFRP face sheets and a foam-GFRP web core loaded under four-point bending. Journal of Composite Materials 49(22): 2765–2778.

27.

Zhu

Shi

Fang

, et al. (2018) Fiber reinforced composites sandwich panels with web reinforced wood core for building floor applications. Composites Part B: Engineering 150: 196–211.