EVALUATION OF EFFECTIVE FLANGE WIDTH IN TIMBER COMPOSITE BEAMS

EVALUATION OF EFFECTIVE FLANGE WIDTH IN TIMBERCOMPOSITE BEAMSR. Masoudnia*, A. Hashemi, P. QuennevilleDepartment of Civil and Environmental Engineering, Faculty of Engineering,University of Auckland, Auckland, New Zealand*Email: rmas551@aucklanduni.ac.nzA part of this paper was originally published for the 2017 INTER Conference in Kyoto, Japan.ABSTRACTA timber composite beam consists of a Cross Laminated Timber (CLT) panel attached to a girder such as aLaminated Veneer Lumber (LVL) beam. Under positive bending moment, part of the CLT panel acts as theflange of the LVL girder and resists compression. When the spacing between the LVL girders becomes large,simple beam theory is not applicable because the compressive stresses in the flange vary with the distancefrom the LVL girder web and the flange area over the web is more highly stressed than the extremities; thisphenomenon is termed shear lag. For the design of steel-concrete composite sections, the effective flange widthconcept has been introduced into national and international design specifications. Despite the large number ofstudies regarding steel and concrete composite structures, comparative, comprehensive research has not beenconducted on timber composite structures. In this study, a numerical model was developed and experimentallyvalidated for analyzing different configurations of timber composite beams. Based on a parametric study, aformula is proposed for determining the effective flange width of timber composite beams.KEYWORDSCLT, Timber Composite Beams, Effective Flange Width, LVL, shear lag1 INTRODUCTION AND LITERATURE REVIEWShear lag is a confirmed phenomenon in T-sectionbeams and occurs when the in-plane shear strainin the flange of a girder under loading and bendingcauses smaller longitudinal displacements of therequirements for determining the effective flangewidth from various sources. For practical reasons, thecode provisions simplify these requirements. In mostcases, the provisions have remained unaltered for along time (Castro et al 2007; Salama et al 2011).areas of the flange far from the web compared withOnlyareas near the web. This phenomenon can lead tocomposite T-beams, including timber double skinimprecise estimates of the displacements and stressespanels and stress-laminated timber bridges. Theat the extreme fibers of composite sections usingresearch on the former is a case study on stress-Euler bending theory (Amadio et al 2002; Aref et al.laminated timber bridges consisting of laminated2007; Zou et al. 2011; Chiewanichakorn et al. 2004;deck sections combined with glued-laminated timberTimoshenko et al. 2009). For the design of compositebeams compressed transversely with high-strengthsteel-concrete sections, the concept of an effectivesteel rods. The research on the latter focuses onflange width has been introduced into national anddouble skin panel floors made of oriented strand boardinternational design specifications. Based on this(OSB) or plywood; this study presented a formula forconcept, various simplified effective flange widthpredicting the effective flange width for timber boxformulas have been derived from analytical andsections (Porteous et al 2013; Ozelton et al 2008;experimental results. For example, Table 1 shows theDavalos et al 1993). A basic analytical study on the14limitedstudieshaveinvestigatedtimberVOL 26· ISSUE 2 » NEW ZEALAND TIMBER DESIGN JOURNAL

Table 1. Effective flange width (be) formulas in variousdesign codes (AISC 2001; CSA 2001; DD ENV 1994-2 2001;Chiewanichakorn et al. 2007; ACI 2001).SourceFormulastherefore encouraging more economical buildingdesigns (Masoudnia et al 2013; Masoudnia et al 2016;Masoudnia et al 2017).be is the smallest of:(1) Beam span/4AISC-LRFD:13.1(2) bs(3) 2 times the distance tothe slab edgeCANADIANCSA: S16-01 &Euro-Code (2000)A timber composite beam was experimentally testedin the University of Auckland test hall, and thebe is the smallest of:obtained results were used to verify a numerical(1) Beam span/4model. As shown in Figure 1, the timber composite(2) bsbeam consisted of a seven meter long planed LVLbe is the smallest f:beam with a width of 300 mm and a depth of 600(1) Beam span/4ACI2 TEST SETUP AND NUMERICAL MODELLING(2) bw 16hf(3) Center-to-center beamspacingbe: effective flange width of concrete flange for compositebeambs: width of concrete flange for composite beamhf: flange thicknessbw: web breadtheffective flange width of timber composite beams witha CLT slab showed that the non-uniform distributionof normal stresses along the flange width in a timbercomposite T-beam is the result of shear deformationin the CLT panel (Thiel et al 2016).Despite the largemm and a top flange of a 5 layer CLT panel with awidth of 2000 mm, a depth of 200 mm and a lengthof 6000 mm. The panel was mechanically fastenedto the LVL beam by forty-eight 550 mm self-tappingscrews with a diameter of 11 mm. The screws wereset at a 45 angle to provide additional compositeaction between the panel and the beam. A bendingtest was conducted using a Material Testing Systems(MTS) actuator testing machine with sufficient loadcapacity to apply a service load on the test specimen.To monitor the vertical deflection, 3 linear variabledifferential transducers (LVDTs) were attached at themid-span of the beam and at the supports (Masoudnianumber of studies regarding steel and concreteet al 2017).composite structures, comparative, comprehensiveThe finite element package ABAQUS version 6.13-3research has not been conducted on timber structureswas used to determine the exact stress distributions(Salama et al 2011; Davalos et al 1993; Adekola etin the longitudinal layers of the CLT panels for theal 1968; Elkelish et al 1986; Timoshenko et al 2009;simply supported timber composite beam. TheNassif et al 2005; Adekola et al 1974). Recognizingaccuracy of the numerical technique used in this studythe lack of research on the effective flange widthhas been previously validated by comparing the mid-in timber structures, this study seeks to investigatespan deflection of the T-section and the associatedthe effective flange width in a CLT slab in timbercomponents (CLT panel and LVL beam), comparingcomposite beam under positive bending. The conceptthe slip between the panel and the LVL beam everyof effective flange width is important for a simplified1m along the span of the timber composite beamstructural analysis, especially for computing stressesand finally the effective flange width of the timberand displacements (Chen et al 2007). These datacomposite beam from the numerical models withassist more efficient and effective design of timberexperimental data (Table 2 to Table 4 and Figure 2).structural members, such as timber flooring systems,Table 2. Comparison of the experimental and numerical results for mid-span deflectionDeflectionTest SpecimenW T L(mm) (mm) (mm)Experimental (mm)Numerical (mm)CLT2030 200 600017.9 *17.9LVL300 605 60003.1 *3.1Timber Composite BeamCLT LVL(connected by screws)1.8 **1.7W: Width (mm), T: Thickness (mm), L: Length (mm)* Deflection for the 50-kN four-point loading test,** Deflection for the 50-kN single-point loading test.NEW ZEALAND TIMBER DESIGN » JOURNAL VOL 26· ISSUE 215

Table 3. Comparison of the experimental and the numerical slip resultsLVDTPosition of the LVDTSlip (mm) (Experimental)*Slip (mm) (Numerical)*1At mid-span0021 m from mid-span0.0560.05532 m from mid-span0.0840.08543 m from mid-span0.1210.122* Deflection for the 50-kN single-point loading testTable 4. Comparison of the experimental and numerical results of effective flange widthSpecimenEffective flange width(Experimental)Effective flange width(Numerical)(Experimental result)(Numerical result)CLT composite beam980 mm995 mm0.98The close agreement shown by these results confirmsflange width generally increases by approximately 5 %that the model is sufficiently precise to determine thewhen the modulus of elasticity of the boards increaseseffective flange width of the timber composite beamfrom 6 GPa to 8 GPa. For instance, when the modulus(Masoudnia et al 2017; ABAQUS 2014 ).of elasticity of the CLT panel increases from 6 GPa(specimen No. 8 in Table 5) to 8 GPa (specimen No. 4in Table 5), the effective flange width increases from3 NUMERICAL PARAMETRIC STUDY1880 mm to 1950 mm.This study focuses on the CLT panel configurations3.2 Effect of Wood Plank Characteristics on theand their effect on the effective flange width. TheEffective Flange Widthexperimentally validated numerical model was usedto study the effects of varying the characteristics ofthe longitudinal and transverse layers of the CLT panelin combination with the LVL beam on the effectiveflange width of timber composite beams. Variousthicknesses of CLT panels constructed of boards witha modulus of elasticity of 6 GPa, 8 GPa or 10 GPa wereinvestigated. Table 5 summarizes the specificationsand the obtained effective flange widths for thetimber composite beams.To consider the effect of the CLT panel wood plankwidths on the effective flange width of the timbercomposite beams, two beam configurations weremodeled using a 5-layer CLT with 90-mm or 180-mmwide boards with a modulus of elasticity of 6 GPa, 8GPa or 10 GPa. Various combinations of 40-mm-thickand 20-mm-thick layers were modeled to study theeffect of the two wood plank widths on the effectiveflange width. Figure 4 shows that the effective flangewidth increased when a CLT panel with 90-mm-wide3.1 The Effect of the Transverse Layer Thickness onboards was replaced by a CLT panel with 180-mm-the Effective Flange Widthwide boards.Figure 3 illustrates the effect of different thicknessesThe maximum difference in the effective flange widthof transverse layers on the effective flange widthsamong all twenty-four numerical analysis results wasin simply supported 6 m long beams under a singleobserved between configurations 14 and 20 (Tablevertical load. The thickness variation of the transverse5). This significant change was due to two factors:layers was considered for CLT panels with an elasticthe lower ratio of the transverse layer depth to themodulus of 6 GPa, 8 GPa and 10 GPa. This figure showslongitudinal layer depth in the CLT panel combinedthat increasing the depth of the transverse layerwith the higher modulus of elasticity in configurationto greater than the depth of the longitudinal layer20 and the use of smaller width boards in the CLTsignificantly increases the effective flange width. Forpanel in configuration 14. These changes led to a 560%instance, when 40 mm thick longitudinal layers aredifference in the effective flange width of these tworeplaced by 20 mm thick layers, the effective flangespecimens.width increases by approximately 2.5 times, from790 mm to 1880 mm (specimens No. 1 and No. 4 inTable 5). Furthermore, the effect of the modulus ofelasticity of the CLT timber boards is presented inFigure 3 and Table 5. Figure 3 shows that the effective16The presented results in Figure 3 and Figure 4 showthat increasing the CLT thickness has a positive effecton the effective flange width only when the increase isthe result of increasing the thickness of the transverselayers.VOL 26· ISSUE 2 » NEW ZEALAND TIMBER DESIGN JOURNAL

ure 1: General test setup of the timber composite beam during the bending test. (a) CLT panel; (b) LVDT at the supports;(c) LVDT at the mid-span; (d) Arrangement of the portal gauges on the surface of the CLT panel; (e) MTS machine; (f) Dataacquisition system; (g) LVL beam; (h) Roller support.Figure 2: Slip monument for every 1m along the span of the timber composite beam (Elevation).NEW ZEALAND TIMBER DESIGN » JOURNAL VOL 26· ISSUE 217