NC State
BioResources
Zhang, X., Luo, L., Fu, H., Sun, Y., and Hui, X. (2019). "Experimental investigation into the flexural behavior of hollow, full, and intermittently stiffened (bamboo-like) glulam beams from larch wood," BioRes. 14(1), 2171-2185.

Abstract

Bamboo-like glulam beams with hollow section units and intermittent internal reinforcement pieces were produced with small-diameter larch-wood pieces and one-component polyurethane. To better understand the design reliability, the failure mode, ultimate bearing capacity, and application potential were evaluated. Three types of beams (solid glulam, hollow glulam, and bamboo-like rectangular glulam beams) were compared and analyzed in this work. Stiffener pieces glued inside the bamboo-like beam were found to increase the bearing capacity and improve the failure mode relative to the hollow glulam beam. Comparison of the hollow section with a similar outside diameter showed that the ultimate bearing capacity increased by approximately 12.3% when the spacing between the stiffeners was 270 mm, and the ultimate bearing capacity increased by approximately 18.0% when the spacing between the stiffeners was 135 mm. Compared with the solid timber beams, wood consumption was reduced by 26.4% and 25.7% for the hollow and bamboo-like glulam beams, respectively. Also, a parameter analysis of the reasonable spacing and thickness of the stiffener was proposed by the finite element method.


Download PDF

Full Article

Experimental Investigation into the Flexural Behavior of Hollow, Full, and Intermittently Stiffened (bamboo-like) Glulam Beams from Larch Wood

Xiaofeng Zhang,a Lisheng Luo,b Haiyan Fu,a Youfu Sun,a, * and Xiran Hui a

Bamboo-like glulam beams with hollow section units and intermittent internal reinforcement pieces were produced with small-diameter larch-wood pieces and one-component polyurethane. To better understand the design reliability, the failure mode, ultimate bearing capacity, and application potential were evaluated. Three types of beams (solid glulam, hollow glulam, and bamboo-like rectangular glulam beams) were compared and analyzed in this work. Stiffener pieces glued inside the bamboo-like beam were found to increase the bearing capacity and improve the failure mode relative to the hollow glulam beam. Comparison of the hollow section with a similar outside diameter showed that the ultimate bearing capacity increased by approximately 12.3% when the spacing between the stiffeners was 270 mm, and the ultimate bearing capacity increased by approximately 18.0% when the spacing between the stiffeners was 135 mm. Compared with the solid timber beams, wood consumption was reduced by 26.4% and 25.7% for the hollow and bamboo-like glulam beams, respectively. Also, a parameter analysis of the reasonable spacing and thickness of the stiffener was proposed by the finite element method.

Keywords: Glulam beam; Bamboo-like; Stiffener; Bearing capacity

Contact information: a: Department of Wooden Structure Construction, College of Materials Science and Engineering, Nanjing Forestry University, Nanjing 210037, China; b: College of Civil Engineering and Architecture, Hainan University, Haikou 570228,China;

* Corresponding author: yfsun@njfu.edu.cn

INTRODUCTION

Glued laminated timber (glulam), an engineered wood product, was developed during the 19th century in Europe and has been used in engineering fields because of its high strength to weight ratio, as well as its aesthetic properties, easy processing, strength, and dimensional stability (Anshari et al. 2010).

There are many advantages in common between the gluing and splicing processes. The splicing process is the method of making beams using traditional Chinese craftsmanship. It refers to the crushing of complex shapes with small pieces or simple shapes to replace big-diameter timber with small-diameter timber. The split beam has a long history, and it has been seen in buildings from the Tang Dynasty (Zhao et al. 2014).

Currently, research is underway on steel-timber composite beams and timber-concrete composite beams to determine their process and mechanical properties (Ghavami 1995; Capozucca 1998; Eisenhut et al. 2016; Bradford et al. 2017; Keipour et al. 2018). According to the status quo of forestry resources in China and inspired by bamboo, a new type glulam beam with a hollow section unit has been proposed, which is manufactured with small-diameter larch (Larix olgensis Henry) and one-component polyurethane. It is based on a hollow glulam beam, within which internal pieces called “stiffeners” are adhered at intervals to increase its ultimate bearing capacity. The spacing of the stiffeners can be adjusted according to load requirements and their own thickness. Because the structure has similarities with bamboo, except that it has a rectangular cross-section, it will be defined here as a bamboo-like glulam beam. Also, the section has potential to be used in various structures and applications, such as in building columns or as an alternative to steel street lighting and traffic light poles (Gilbert et al. 2014). Furthermore, this beam with a hollow section unit has developed a market for low-value, small-diameter timber in many countries. This is a reasonable measure to save forest resources and contribute to sustainable environmental development.

To use these novel sections in structural applications, experimental tests must be done to confirm the structural behavior and failure modes, therefore ensuring reliable design properties (Gilbert et al. 2014, 2018). According to the literature, various hollow section components have been proposed to utilize small-diameter logs or replace solid timber members in the building structure market. Hollow octagonal utility poles are manufactured with composite wood flake panels, and the poles are tapered with the wall thickness decreasing as the height decreases. Additionally, the feasibility, design strength, and cost have been analyzed. It has been found that the proposed poles can be cost-competitive in the industry (Adams et al. 1979). Since 1994, Kyoto University has been working on composite structures and found that low quality logs have a remarkable influence on the strength of the product when used for veneer tapes and between butt-jointed veneers (Hara et al. 1994; Sasaki and Kawai 1994). Sasaki and Kawai (1994) proposed a hollow cylindrical laminated veneer lumber (LVL); a series of studies have been carried out, and they have indicated that this LVL exhibits a high strength and can be used in construction. Zhang et al. (1997) studied a type of hollow bamboo-wood composite slab (consisting of bamboo-based panels and Masson pine strips) and explored the effect of the hollow ratio of this plate on the mechanical properties. The test results showed that the section can be designed and the hollow ratio has no remarkable effect on the stiffness, but has a noticeable effect on the strength (Zhang et al. 1997). Harries et al. (2000) performed axial and bending tests on hollow composite columns. When compared with solid sections, the hollow sections had advantages in the carrying capacity when its diameter was greater than 127 mm × 127 mm (Harries et al. 2000). Wei (2005) studied cylindrical veneer lumber, which is a spirally wound structure in which the veneers are processed into a staggered laminate. As a result, a decrease in the elastic modulus is suppressed. When cylindrical LVL is produced, the elastic modulus of the laminated material is reduced to a small extent in the range of an angle of no more than 10°, and glass fiber can enhance the dimensional stability of cylindrical LVL. Cylindrical LVL has good mechanical properties, excellent engineering properties, and can be used in construction and other fields (Wei 2005). Cai (2014) proposed a single-plate laminated hollow column. It was found that local stress and strain concentrate at the bottom during the stress process. According to the conclusion of an ANSYS simulation by Cai (2014), it is necessary to strengthen the end of the hollow column to ensure its bearing capacity in practical applications. Hollow veneer-based composite utility poles are manufactured from a hard plantation of thinned trees to find a solution and change the current shortage of traditional solid utility poles in Australia. Gilbert et al. (2014) showed that the proposed cross-section will have an external diameter approximately 1.2 times greater than the former and would be approximately two times lighter. Hartig et al. (2016) produced a molded wooden tube made of beech with a height of 3 m, diameter of 0.32 m, and wall thickness of 0.02 m. Experimental and theoretical investigations showed that the wood molding process did not impair the structure of the material; in contrast, it increased the stiffness (Hartig et al. 2016). Shang et al. (2018) studied the bending behavior of a hollow Xingan larch (L. dahurica) glulam beam. Additionally, they studied the influence of the wall thickness and hollow ratio on the flexural capacity and stiffness of the hollow beam (Shang et al. 2018). Aicher and Tapia (2018) used a novel composite of softwood glulam with internal beech LVL layers in their study. The experimental and theoretical investigations showed that the composite was damage-tolerant and excelled at non-brittle fracture. The mechanical properties of the novel glue-laminated timber-laminated veneer lumber (GLT-LVL) composite beam structure were superior, especially its suitability to accommodate large holes (Aicher and Tapia 2018).

This work introduced experimental materials, specimen design, and experimental methods for production of bamboo-like glulam beams. The performance of the studied sections was compared with that of solid and hollow glulam beams. Also, the ultimate bearing capacity and failure modes of the three different sections were analyzed and discussed. Finally, the reasonable spacing and thickness of the stiffeners were determined by the finite element method.

EXPERIMENTAL

Materials

The wood used was Northeast China larch (L. olgensis Henry), with diameters that ranged from 50 mm to 120 mm. This larch is a type of a domestic, fast-growing tree and was supplied by Xingyuan Wood Industry Co., Ltd. in Nanjing, China. The mechanical properties of larch can meet the production requirements for glulam (Wang et al. 2008). The dimensions were 4000 mm × 80 mm × 80 mm, and the grade was IIa, according to the visual stress grading method (GB 50005 2003). The properties of the materials were obtained via testing by referring to ASTM D4761-13 (2013). The final moisture content of the wood was 12.9% and the density was 0.63 g/cm3. The compressive strength parallel to the grain was 43.0 MPa, the tensile strength parallel to the grain was 71.8 MPa, the ultimate compression strain parallel to grain was 5040 με, and the ultimate tensile strain parallel to the grain was 5120 με. Additionally, the bending modulus was 14,000 MPa.

A single-component polyurethane was used, and was supplied by Shanghai Donghe Adhesive Co., Ltd (DH 106, Shanghai, China).

Specimen Design

All of the test specimens were manufactured in a laboratory at Nanjing Forestry University (Nanjing, China), with an ambient temperature of approximately 23 °C to 25 °C and a humidity of approximately 53% to 56%. The test material was naturally dried to a moisture content of less than 13%, and was then subjected to sawing, planing, sizing, cold pressing, and curing processes.

In total, 12 timber beams over four groups were tested to examine the bending performance. All of the specimens had the same external section size. The length along the wood grain direction was 1800 mm, the height was 88 mm, and the width was 56 mm.

Table 1. Dimensions of the Specimens

φ – Hollowness ratio, the percentage of the hollow volume vs. the apparent volume of the glulam beam

Fig. 1. Cross sections (mm) of the specimens

The specimens included three solid glulam beams labeled ZJA1 to ZJA3, three hollow glulam beams labeled ZJB1 to ZJB3, three bamboo-like (with intermitted insertion of stiffener pieces) glulam beams with a stiffener spacing of 270 mm labeled ZJC1 to ZJC3, and three bamboo-like glulam beams with a stiffener spacing of 135 mm labeled ZJD1 to ZJD3. According to ASTM D198-15 (2015) and CECS 254 (2009), the section hollow ratio of the hollow and bamboo-like glulam wood beams was defined. The specific size parameters of the specimens are shown in Table 1.

Experimental Methods

The tests were performed in the Wood Structure Laboratory of Nanjing Forestry University with an ambient temperature of approximately 22 °C to 25 °C and humidity of approximately 52% to 57%.

Figures 2 and 3 show the loading test method. For each test, the beam was positioned horizontally on the loading device in the direction of the grain. Five strain gauges were arranged equidistantly along the height direction of the beam span (BX120-20AA, Heshan District Guangce Electronics Co. Ltd., Yiyang, China). The length direction of the strain gauge was consistent with that of the wood along the grain direction, and the gauges were numbered 1 to 5 from top to bottom. A thimble displacement sensor (CPD-25, Shimadzu International Trading (Shanghai) Co. Ltd., Shanghai, China) was set at two fulcrum points at the end of the beam. There was an additional displacement sensor (YHD-100, Jiangsu Liyang Instrument Factory, Changzhou, China) at the mid-span on the bottom, which measured the settlement of the beam end supports and the change of the mid-span deflection, with an accuracy of 0.001 mm.

Fig. 2. Loading device and test point arrangement, LVDT- Displacement measurement (mm)

Fig. 3. Loading process of the test (a) before and (b) after loading

The test was done using an electronic universal testing machine (UTM5105, Shenzhen Sansi Vertical and Horizontal Technology Co. Ltd., Shenzhen, China) with a maximum load of 100 kN, and the measured data was collected by a static data acquisition instrument (TDS-530, Tokyo Measuring Instruments Lab., Tokyo, Japan). According to GB/T 50329 (2012), a support steel pad was placed on the lower surface of the specimen to prevent lateral compression of the wooden beam. The width of the support steel backing plate was larger than the width of the beam section, the length was 80 mm, and the thickness was 10 mm.

A four-point bending experiment was developed according to ASTM D198-15 (2015). Before formal loading, the F0 and F1 values of each specimen were estimated. The F0 value was taken as 0.5 kN to ensure that the loading plate and support pad were in close contact with the wooden beam, and the F1 value was taken as 5 kN, which was less than the proportional limit force value. After loading started (the force control mode was adopted at the beginning), the load started at the F0 value and increased to the F1 value, then the load was unloaded to the F0value, and the cycle was repeated five times in accordance with the loading mode to calculate the elastic modulus of the beam.

After the 5th cycle the test was then conducted monotonically with a loading rate of 2 mm/min until complete failure .

The work performance of the glulam beam was numerically simulated and verified by Workbench 15.0 with ANSYS finite element software (ANSYS Company, Pittsburgh, Carlisle, PA, USA).

Main Results of Testing

The specific test results of all 12 glulam beam specimens are shown in Table 2. For the flexural glulam beams, the apparent elastic modulus (MOE) and equivalent flexural strength (MOR) were converted according to the test results. The conversion formula was as follows,

 (1)

 (2)

where a is the distance from the loading point to the support (mm), ΔP is the load increment of the elastic stage (N), l is the calculated span of the beam specimen (mm), I is the moment of inertia of the beam section, W is the section modulus of the beam section, and Pu is the ultimate load (kN).

RESULTS AND DISCUSSION

Failure Mode Analysis

There were three typical failure modes in the specimens, and the specific failure modes are shown in Figure 4. First was brittle tensile failure. The glulam beam showed obvious plastic deformation in the late stage of loading, and the stiffness of the beam decreased. As the load continued to increase, a crack appeared in the beam span near the bottom plate, which was accompanied by a tearing sound from the wood fiber. The crack continued to increase, the layered tear at the bottom of the glulam beam was destroyed, and the bearing capacity decreased rapidly. After unloading, the deflection of the wooden beam was restored to the initial stage, but there was partial plastic deformation. The second failure mode was web shear failure. In the ZJB and ZJC groups, the hollow glulam beam webs were subjected to large shear stress. This led to the first crack at the loading point to the bearing area of the hollow glulam beam. The crack developed obliquely until shear failure occurred. Because of insufficient development of plastic deformation during shear failure, the equivalent bending strength (MOR) of the ZJB group was smaller than that of the other groups. Also, it was evident that the stiffeners improved the failure mode of the glulam beams. The third failure was tensile ductility destruction. In the ZJC and ZJD specimen groups, the compressive side of the wood fibers first reached the yield strength during the loading process and the stress remained unchanged. The strain increased and the beam was plasticized, which was accompanied by the sound of the internal wood fibers breaking. Deformation occurred until the tensile side of the wood fiber reached the ultimate tensile strain and exhibited ductility.

It was found that stiffeners improved the failure mode of the hollow timber beams and increased their ultimate bearing capacity and bending strength.

Fig. 4. Failure mode of the specimens: (a) ZJA1, (b) ZJB1, (c) ZJC2, and (d) ZJD2

All of the specimens showed elastic characteristics during the initial stage of loading, and the force-controlled cyclic loading was performed in the elastic stage. Because of the presence of a small amount of knots in the glulam beam, the bending stiffness of the beam was reduced and the MOE of the glulam beam was reduced. This resulted in the MOE of the glulam beam in Table 2 being less than the MOE obtained from the experimental material. Compared with the hollow glulam beam, the average ultimate bearing capacity of group C increased by approximately 12.3%. Also, when the distance between group D was 135 mm, the average ultimate bearing capacity increased by approximately 18.0%.

Table 2. Main Bending Test Results

Notes: εc is the ultimate compressive strain at failure (με), εt is the tensile strain at failure (με), and ω is the deflection corresponding to the ultimate load (mm).

NUMERICAL ANALYSIS

Unit Selection and Meshing

The solid model of the glulam wood beam was established with a Solid186 solid element. The element was a three-dimensional 20-node high-order hexahedral element with a quadratic displacement mode. Each node had three translational degrees of freedom, and the material nonlinearity and geometric nonlinearity were considered. The finite element model after meshing is shown in Fig. 5.

Fig. 5. Finite element model

Material Parameters and Constitutive Model

Wood is an anisotropic material, which means that the establishment of a triaxial constitutive model can be complicated. It has been studied to establish a constitutive model with longitudinal axial mechanical properties. Bechtel and Norris (1952) proposed a bilinear tensile and compressive constitutive model based on many other experiments (Fig. 6). The model assumes that the tension zone is linearly elastic and the compression zone is also elastic. After the strength reaches the yield, the stress value remains unchanged, which is the ideal elastoplastic model. The mathematical expression of the model is as follows,

 (3)

where E is the flexural modulus of the wood (14016.3 MPa), εt is the ultimate tensile strain (5119 με), εc is the ultimate compressive strain (5037 με), and σ0 is the compressive yield strength of the wood (43.05 MPa).