Flexural Performance of Built-Up Beams Made with Plantation Wood

Article Flexural Performance of Built-Up Beams Made with Plantation Wood

Tsai-Po Chien 1, Te-Hsin Yang 2,* and Feng-Cheng Chang 1,*

1 School of Forestry and Resource Conservation, National Taiwan University, #1, Sec. 4, Roosevelt Rd., Taipei 10617, Taiwan 2 Department of Forestry, National Chung Hsing University, 145 Xingda Rd., South Dist., Taichung City 402, Taiwan * Correspondence: [email protected] (T.-H.Y.); [email protected] (F.-C.C.); Tel.: +886-4-22840345 (ext. 145) (T.-H.Y.); +886-2-33664619 (F.-C.C.) Received: 30 May 2019; Accepted: 25 July 2019; Published: 1 August 2019

Abstract: In this study, Japanese cedar (Cryptomeria japonica (L. f.) D. Don) harvested from a plantation in Taiwan was used to develop built-up beams using self-tapping screws as metal connectors and resorcinol formaldehyde resin as glue to assemble components based on various assembly configurations. Results showed that adding glue provided flexural rigidity, whereas assembly using self-tapping screws resulted in built-up beams with high ductility but relatively low flexural bearing capacity. Beams used glue exhibited approximately linear behavior, whereas those using only screws exhibited some undulating and stepwise responses, implying that shear force between the flanges and the web may cause buckling as well as the dislocation of the self-tapping screws. When using components of similar grades, adding another web can improve the performance. Furthermore, the grades of flanges can strongly influence the flexural load-bearing capacity. In addition, a smaller spacing between the screws can improve the flexural load-bearing performance, but also cause wooden components to crack. Typical bending failure modes were observed in the developed built-up beams, indicating tension failure of the bottom flange as well as slippage between flanges and the web due to horizontal shear, which also caused buckling deformations in the screws.

Keywords: built-up beams; Japanese cedar; Cryptomeria japonica; self-tapping screws; plantation wood; resorcinol formaldehyde resin

1. Introduction Because of changing resource conditions, plantation forests are expected to replace natural forests as the main source of timber in the near future. Japanese cedar (Cryptomeria japonica (L. f.) D. Don) is one of the main plantation species in Taiwan and Japan. Attributing to its fast-growing cycle and favorable processing properties, the processed Japanese cedar products are often used for structural and decorative applications. Based on government information, the plantation area of Japanese cedar in Taiwan is approximately 30,555 ha and the total stock volume is approximately 7.69 million m3 [1], which could be a great resource of domestic timber supply. On the other hand, proper forest landscape restoration activities such as afforestation and reforestation of planted forests and woodlots would be beneficial for CO2 sequestration efficiency [2]. Consequently, it is important to study the potential use for Japanese cedar plantation wood in future applications. However, plantation wood often has a medium or small diameter, and the proportion of juvenile wood is higher than that of natural forests, thereby limiting direct structural application. Csoka et al. [3] mentioned that the lumber obtained from young Japanese cedar contains a high percentage of juvenile wood, which is unfavorable in structural applications. Therefore, engineered products using plantation

Forests 2019, 10, 647; doi:10.3390/f10080647 www.mdpi.com/journal/forests Forests 2019, 10, 647 2 of 11 wood for structural applications should be developed to solve the shortage of high-diameter timber. Glued wood products such as glued-laminated beams constructed using plantation wood have been studiedForests using 2019, di10,ff x erentFOR PEER species REVIEW and types of wood adhesives. However, few studies have investigated2 of 11 the use of plantation wood for built-up structural components. Built-upplantation timber wood for beams structural can be applications assembled shou in dildff erentbe developed cross-sectional to solve shapes;the shortage the I-shapeof high- is the mostdiameter commonly timber. adopted Glued and wood features products three such components, as glued-laminated namely beams two constructed flanges on using the top plantation and bottom and awood web have component been studied in the using middle. different The species advantages and types of built-upof wood adhesives. beams include However, high few strength, studies high stiffness,have andinvestigated light weight; the use moreover,of plantation built-up wood for beams built-up using structural metal components. connectors for assembly require Built-up timber beams can be assembled in different cross-sectional shapes; the I-shape is the shorter processing time than conventional glued-laminated beams or glued I-beams because of the long most commonly adopted and features three components, namely two flanges on the top and bottom curingand time a web required component for woodin the middle. adhesives The [ 4advantages–7]. More of importantly, built-up beams built-up include beams high strength, can be producedhigh withstiffness, different and configurations light weight; asmoreover, per requirement. built-up beams using metal connectors for assembly require Inshorter contrast processing to adhesives, time than metal conventional connectors glued- suchlaminated as nails beams and screws or glued used I-beams for assembling because of built-upthe beamslong can curing be highly time erequiredfficient andfor wood time-saving adhesives because [4–7]. theMore curing importantly, time for built-up adhesives beams is not can required. be Becauseproduced metal with connectors different areconfigurations mainly responsible as per requirement. for connecting the wood members of the built-up beam, theIn specifications contrast to adhesives, of the metal metal connectors connectors such and as nails their and connection screws used capacity for assembling with wood built-up should be carefullybeams can considered. be highly efficient Studies and have time-saving revealed becaus thate the the curing diameters time for of ad connectorshesives is not exhibit required. a linear relationshipBecause metal with withdrawalconnectors are resistance mainly responsible [8–11] and for that connecting the screw-withdrawal the wood members resistance of the built-up is usually beam, the specifications of the metal connectors and their connection capacity with wood should be higher than the nail-withdrawal resistance [11]. Furthermore, previous research has indicated that carefully considered. Studies have revealed that the diameters of connectors exhibit a linear addingrelationship glue when with using withdrawal metal resistance connectors [8– to11] assemble and that the built-up screw-withdrawal beams can resistance improve is their usually flexural mechanicalhigher than performance the nail-withdrawal [12]. In particular, resistance resorcinol [11]. Furthermore, formaldehyde previous resin research can resulthas indicated in high that bonding strengthadding and glue increase when theusing shear metal resistance connectors capacity to assemble [13,14 built-up]. beams can improve their flexural Inmechanical addition performance to wood I-beams, [12]. In flangesparticular, can resorcin substantiallyol formaldehyde enhance resin the flexuralcan result rigidity, in high bonding and the type and sizestrength of flange and increase influence the theshear flexural resistance load-bearing capacity [13,14]. capacity [15–23]. The depth of the web of wood I-beams affectsIn addition the flexural to wood load-bearing I-beams, flanges capacity, can substant but theially width enhance of the the web flexural does not rigidity, considerably and the type influence the flexuraland size rigidityof flange [ influence24–28]. Moreover,the flexural theload-bearing length of capacity the flanges [15–23]. or The web depth can of be the increased web of wood by finger I-beams affects the flexural load-bearing capacity, but the width of the web does not considerably jointing [29–31]. influence the flexural rigidity [24–28]. Moreover, the length of the flanges or web can be increased by In this study, Japanese cedar (Cryptomeria japonica (L. f.) D. Don) harvested from a plantation forest finger jointing [29–31]. in TaiwanIn was this usedstudy, to Japanese develop cedar built-up (Cryptomeria beams usingjaponica self-tapping (L. f.) D. Don) screws harvested as connectors from a plantation to assemble componentsforest in basedTaiwan on was various used to assembly develop built-up configurations. beams using The self-tapping effects of thesescrews configurations as connectors to on the flexuralassemble performance components and based failure on modes various were assembly investigated. configurations. The effects of these configurations on the flexural performance and failure modes were investigated. 2. Materials and Methods 2. Materials and Methods Japanese cedar (Cryptomeria japonica (L. f.) D. Don) lumber was harvested in Hsinchu, Taiwan, the tree age measuredJapanese cedar approximately (Cryptomeria 35 japonica years old, (L. f.) and D. Don) processed lumber by was Jang harvested Chang in Lumber Hsinchu, Industry Taiwan, Ltd., Taiwan.the Beforetree age processing, measured approximately the lumber was 35 oven-driedyears old, and and processed conditioned by Jang to approximatelyChang Lumber 15%Industry moisture content,Ltd., and Taiwan. the overall Before average processing, density the lumber of the lumberwas oven-dried was 452 and53 conditioned kg/m3. Processed to approximately lumber dimensions 15% moisture content, and the overall average density of the lumber± was 452 ± 53 kg/m3. Processed lumber for the flanges were 38 140 3300 mm3 and those for the web were 38 89 3300 mm3. Furthermore, dimensions for the flanges× × were 38 × 140 × 3300 mm3 and those for the web× were× 38 × 89 × 3300 mm3. the lumbersFurthermore, were stress-gradedthe lumbers were based stress-graded on Chinese based National on Chinese Standard National (CNS)-14630. Standard (CNS)-14630. The nominal The grade of E50 implies 4 GPa flexural modulus of elasticity < 6 GPa; E70 implies 6 GPa flexural modulus of nominal grade ≤of E50 implies 4 GPa ≤ flexural modulus of elasticity < 6 GPa; E70≤ implies 6 GPa ≤ elasticityflexural< 8 modulus GPa; E90 of implies elasticity 8 GPa < 8 GPa;flexural E90 implies modulus 8 GPa of elasticity≤ flexural

FigureFigure 1. 1.Self-tapping Self-tapping screw screw specifications. speciﬁcations.

Forests 2019, 10, x FOR PEER REVIEW 3 of 11 Forests 2019, 10, 647 3 of 11 ForestsTo 2019 study, 10, x FORthe PEEReffects REVIEW of factors such as cross sections and component grades on the flexural3 of 11 capacity,To study the theCryptomeria effects of factors japonica such built-up as cross sectionsbeams andwere component assembled grades using on lumbers the flexural of different capacity, To study the effects of factors such as cross sections and component grades on the flexural thegrades,Cryptomeria and the cross japonica sectionsbuilt-up are beamsillustrated were in assembled Figure 2a. Moreover, using lumbers to enhance of different the bonding grades, between and the capacity, the Cryptomeria japonica built-up beams were assembled using lumbers of different crossflanges sections and the are web, illustrated as well in as Figure for comparison2a. Moreover, with toenhance the conventional the bonding glue-bonding between flanges method, and thetwo grades, and the cross sections are illustrated in Figure 2a. Moreover, to enhance the bonding between web,typesas of wellbuilt-up as for beams comparison were pr withoduced the as conventional shown in Figure glue-bonding 2b. method, two types of built-up flanges and the web, as well as for comparison with the conventional glue-bonding method, two beamsThe were side produced views of asthe shown configurations in Figure2 ofb. the built-up beams are shown in Figure 3. The distance types of built-up beams were produced as shown in Figure 2b. of edgeThe “a” side was views at least of the 20 configurations cm, and various of thespacing built-up “b” beamsvalues are(20 shown–60 cm) in were Figure applied3. The between distance the of The side views of the configurations of the built-up beams are shown in Figure 3. The distance edgescrews “a” to wasstudy at the least effect 20 cm, of andspacing various on the spacing flexur “b”al performance values (20–60 of cm)the buil weret-up applied beams. between In Table the 1, of edge “a” was at least 20 cm, and various spacing “b” values (20–60 cm) were applied between the screwsthe group to study code therepresents effect of different spacing on configurations the flexural performance with respect of to the the built-up cross-sectional beams.In shape Table (I-,1, the II-, screws to study the effect of spacing on the flexural performance of the built-up beams. In Table 1, groupand box-beam), code represents top flange different grade, configurations web grade, with and respect bottom to theflange cross-sectional grade, assembly shape (I-,method II-, and (S the group code represents different configurations with respect to the cross-sectional shape (I-, II-, box-beam),represents self-tapping top flange grade, screws web and grade,G represen andts bottom glued), flange and the grade, distance assembly between method screws. (S represents and box-beam), top flange grade, web grade, and bottom flange grade, assembly method (S self-tappingTo compare screws the and built-up G represents beams with glued), conventional and the distance glued-laminated between screws. timber (glulam) beams, E90 represents self-tapping screws and G represents glued), and the distance between screws. CryptomeriaTo compare japonica the built-up lumber beams and withresorcinol conventional formaldehyde glued-laminated resin (RF, timberpH = (glulam)8.6, viscosity beams, = 1296 E90 To compare the built-up beams with conventional glued-laminated timber (glulam) beams, E90 CryptomeriamPa*s, Wood japonica Glue Industriallumber and Co., resorcinol Ltd., Tainan, formaldehyde Taiwan) were resin employed (RF, pH = 8.6,to fabricate viscosity built-up= 1296 mPa*s,beams Cryptomeria japonica lumber and resorcinol formaldehyde resin (RF, pH = 8.6, viscosity = 1296 Woodwith glue Glue and Industrial 4-layer Co.,glued-laminated Ltd., Tainan, Taiwan)beams as were a reference. employed The to fabricateresin to curing built-up agent beams weight with ratio glue mPa*s, Wood Glue Industrial Co., Ltd., Tainan, Taiwan) were employed to fabricate built-up beams andwas 4-layer100:15; glued-laminatedthe spread amount beams was as250 a g/m reference.2 and the The pressing resin to time curing was agent 1 day. weight ratio was 100:15; with glue and 4-layer glued-laminated beams as a reference. The resin to curing agent weight ratio the spread amount was 250 g/m2 and the pressing time was 1 day. was 100:15; the spread amount was 250 g/m2 and the pressing time was 1 day.

Figure 2. Cross-sectional shapes of built-up beams using (a) self-tapping screws and (b) glue and screws. Figure 2. Cross-sectional shapes of built-up beams using (a) self-tapping screws and (b) glue and screws. Figure 2. Cross-sectional shapes of built-up beams using (a) self-tapping screws and (b) glue and screws.

FigureFigure 3.3. Self-tapping screw drive-in positions.positions.

Figure 3. Self-tapping screw drive-in positions.

Forests 2019, 10, 647 4 of 11

Table 1. Flange and web level of the built-up I-beam.

Stress Grade of Stress Grade of Stress Grade of Spacing between Number of Group RF-Glued Top Flange Web Bottom Flange Screws (cm) Tested Beams I-777-S40 E60–E80 E60–E80 E60–E80 – 40 10 II-777-S40 E60–E80 E60–E80 E60–E80 – 40 10 B-777-S40 E60–E80 E60–E80 E60–E80 – 40 10 I-565-S40 E50 E60 E50 – 40 5 I-569-S40 E50 E50 E90 – 40 5 I-4610-S40 E40 E60 E100 – 40 5 I-4410-S40 E40 E40 E100 – 40 5 I-959-S40 E90 E50 E90 – 40 5 I-777-G E60–E80 E60–E80 E60–E80 Yes – 10 I-777-GS40 E60–E80 E60–E80 E60–E80 Yes 40 10 Glulam E90 E90 E90 Yes – 10 I-777-S20 E60–E80 E60–E80 E60–E80 – 20 5 I-777-S30 E60–E80 E60–E80 E60–E80 – 30 5 I-777-S50 E60–E80 E60–E80 E60–E80 – 50 5 I-777-S60 E60–E80 E60–E80 E60–E80 – 60 5 *The beams were conditioned at indoor ambient environment for at least 4 weeks before ﬂexural tests.

For the flexural test, based on the CNS 11031, a 4-point bending configuration was adopted, and the loading scheme is shown in Figure4. The indoor ambient testing environment was approximately 23–25 ◦C and 65%–70% RH, controlled by an air conditioner, and moisture contents of beams were measured before tests using a moisture meter (BD-2100, Delmhorst Instrument Co., Towaco, NJ, USA). On the basis of the load-deflection curve of the built-up beams, the coefficient K, indicating the initial rigidity, was be determined by passing a linear straight line through 0.1 maximum flexural load × (Pmax) and 0.4 Pmax (Equation (1)). The elastic deflection was expected to be 0.1 Pmax to 0.4 Pmax, × × × and the modulus of elasticity (MOE) was calculated as shown in Equation (2).

P 24EI K = = (1) 2∆ a(3L2 4a2) − Ka 3L2 4a2 MOE = − (2) 24I where K = initial stiﬀness (N/mm), L = Length of the beam (mm), E = Modulus of elasticity (N/mm2), I = Moment of inertia (mm4), P = Load on the beam (N), ∆ = Deﬂection (mm), a = L/3 (mm). Moreover, the modulus of rupture (MOR), which indicates the bending strength of the built-up beam, can be calculated using Equation (3).

My σ = (3) I

P(L a) M = − (4) 4 where σ = Bending strength (N/mm2), M = Bending moment (N mm) based on Equation 4, y = Distance · of the center of the cross section to the surface of the ﬂange (mm), I = Moment of inertia (mm4), P = Maximum load (N), L = Span of the bending test (mm), a = L/3 (mm). Forests 2019, 10, x FOR PEER REVIEW 5 of 11

Forests 10 Forests 2019, 10, x 647 FOR PEER REVIEW 5 5of of 11 11

Figure 4. The 4-point bending test scheme. Figure 4. The 4-point bending test scheme. 3. Results and Discussion Figure 4. The 4-point bending test scheme. 3. Results and Discussion 3. ResultsThe overalland Discussion average moisture contents of the flange and the web were 13.5% and 13.8%, respectively,The overall measured average moisturebefore flexural contents tests. of the Fi flangegure 5 and shows the web the wererepresentative 13.5% and 13.8%,load-displacement respectively, measuredcurvesThe of overall beforevarious flexuralaverage groups. tests. moisture The Figure results contents5 showsshow thethatof th representative beamse flange using and load-displacementglue the exhibitedweb were approximately 13.5% curves and of various13.8%, linear respectively,groups.behavior, The whereas resultsmeasured those show before thatusing beamsflexural only usingscrews tests. glue Fiexhibigure exhibited ted5 shows some approximately the undulating representative linearand stepwise load-displacement behavior, responses, whereas curvesthoseimplying using of variousthat only shear screws groups. force exhibited The between results some the show undulating flanges that andbeams and the stepwise using web gluemay responses, exhibitedcause implyingbuckling, approximately thatas well shear linearas force the behavior,betweendislocation thewhereas of flanges the self-tappingthose and theusing web screwsonly may screws causebecause buckling,exhibi of crushingted assome well failure undulating as the of dislocation the andwood stepwise ofbody the self-tappingat responses, the screw implyingscrewspenetration because that locations. shear of crushing force In general, between failure groups of the the flanges wood in wh bodyandich gluethe at the webwas screw addedmay penetration cause exhibited buckling, locations. higher as load-bearingwell In general, as the dislocationgroupscapacity, in whereas which of the glue self-tappinggroups was that added usedscrews exhibited only because self-tapping higher of crushing load-bearing screws failure revealed capacity, of highthe whereaswood ductility, body groups enduring at the that screw larger used penetrationonlydeflection self-tapping before locations. screwsfailure. In revealedgeneral, highgroups ductility, in wh enduringich glue was larger added deflection exhibited before higher failure. load-bearing capacity, whereas groups that used only self-tapping screws revealed high ductility, enduring larger deflection before failure.

Figure 5. Load-deflection curves of various groups. Figure 5. Load-deflection curves of various groups. Furthermore, different load-displacement responses among the glued-I beams (e.g., I-777-G), glued-screwed-IFurthermore, beams different (e.g.,Figure load-displacement I-777-GS40) 5. Load-deflection and glulam respon curves beamsses of variousamong were groups.the observed. glued-I Thebeams results (e.g., indicated I-777-G), thatglued-screwed-I the glued-I beambeams and (e.g., the I-777-GS40) glulam beam and glulam exhibited beams a brittle were observed. failure mode, The results initiating indicated the direct that Furthermore, different load-displacement responses among the glued-I beams (e.g., I-777-G), collapsethe glued-I of thebeam beams and afterthe glulam reaching beam maximum exhibited load; a br however,ittle failure the mode, glued-screwed-I initiating the beam direct exhibited collapse glued-screwed-I beams (e.g., I-777-GS40) and glulam beams were observed. The results indicated that ductilityof the beams after reachingafter reaching the maximum maximum load load; and howeve maintainedr, the a glued-screwed-I level of load resistance beam exhibited capacity, althoughductility the glued-I beam and the glulam beam exhibited a brittle failure mode, initiating the direct collapse aafter stepwise reaching decrement the maximum was also load observed. and maintained These observations a level of implyload resistance that the high capacity, bending although rigidity a of the beams after reaching maximum load; however, the glued-screwed-I beam exhibited ductility maystepwise be attributed decrement to glue;was also however, observ theed. self-tappingThese observations screws mayimply enhance that the the high ductility bending of rigidity the built-up may after reaching the maximum load and maintained a level of load resistance capacity, although a beams,be attributed preventing to glue; a sudden however, collapse the self-tapping of the structure screws in practical may enhance applications. the ductility of the built-up stepwise decrement was also observed. These observations imply that the high bending rigidity may beams,The preventing results of the a sudden flexural collapse performance of the ofstructure each group in practical are shown applications. in Table2. Comparing I-777-S40, be attributed to glue; however, the self-tapping screws may enhance the ductility of the built-up II-777-S40,The results and B-777-S40 of the flexural revealed performance a substantial of improvementeach group are in theshown flexural in Table load-bearing 2. Comparing capacity I-777- by beams, preventing a sudden collapse of the structure in practical applications. addingS40, II-777-S40, another web and component B-777-S40 (e.g.,revealed II-beam a substa and box-beam).ntial improvement Moreover, in adding the flexural another webload-bearing member The results of the flexural performance of each group are shown in Table 2. Comparing I-777- S40, II-777-S40, and B-777-S40 revealed a substantial improvement in the flexural load-bearing

Forests 2019, 10, 647 6 of 11 improved the flexural rigidity and bending strength. However, because there was no significant difference between II-777-S40 and B-777-S40 with respect to various properties, the locations of the web components may not considerably influence the flexural load-bearing capacity and performance. Moreover, for the same grade of flange, using a lower grade of web (I-4410-S40) resulted in a marginally lower average maximum flexural load, flexural rigidity, and strength compared with a higher-grade web (I-4610-S40), although no significant difference was observed between these two groups. Furthermore, for web components of similar grades but flanges of different grades, I-959-S40 resulted in a significantly higher maximum flexural load-bearing capacity, flexural rigidity, and strength than I-565-S40, implying that a higher grade of flange played a key role in improving the flexural capacity and mechanical performance of the built-up beams, which was in agreement with some previous studies [6,15,16,18,20–22]. Nevertheless, comparisons between I-959-S40 and I-559-S40 revealed no significant difference in the maximum flexural load, flexural rigidity, and strength, although an improvement in the flexural performance owing to the use of a higher-grade top flange was observed, indicating that the grade of the top flange may not considerably affect the flexural capacity and mechanical performance. In addition, comparing the groups that featured different spacing between the screws suggested that this may influence the flexural performance of the built-up beams. In general, increasing the spacing resulted in a lower maximum flexural load-bearing capacity, flexural rigidity, and strength. When the spacing between the screws was as low as 20 cm, the maximum flexural load-bearing capacity and flexural strength were comparable with those of the II-beam and box-beam with 40 cm spacing between the screws. In addition, the results implied that for further fabrication of the built-up beams, adopting a spacing of 30–60 cm between the screws may produce beams of similar flexural load-bearing capacity, which would influence the selection of the number of screws used in production, thereby affecting the production cost.

Table 2. Flexural properties of the developed built-up beams.

Max. Flexural Load MOE Max. Bending Moment MOR Groups (N) (N/mm2)( 106 N mm) (N/mm2) × · I-777-S40 14379 (10.62)ef 2099 (6.73)de 07.12ef 12.65ef II-777-S40 23919 (15.78)c 3743 (8.84)c 11.84c 20.08c B-777-S40 24045 (9.76)c 3675 (5.38)c 11.90c 20.18c I-565-S40 11780 (10.43)f 1664 (3.86)e 05.83f 10.36f I-569-S40 15536 (20.18)def 1769 (5.75)e 07.69def 15.60de I-4610-S40 14966 (10.42)def 1943 (2.24)de 07.41def 16.21cde I-4410-S40 13694 (9.04)ef 1646 (3.92)e 06.78ef 14.85def I-969-S40 17833 (10.63)de 2175 (13.62)de 08.83de 16.03cde I-777-G 37813 (28.27)b 6528 (10.96)c 18.72b 31.30b I-777-GS40 38913 (14.25)b 6757 (16.43)ab 19.26b 34.24b I-777-S20 20288 (8.76)cd 2587 (6.53)d 10.04cd 17.85cd I-777-S30 17791 (12.65)de 2273 (7.27)de 08.81de 15.65de I-777-S50 16375 (6.78)def 2075 (7.35)de 08.11def 14.41def I-777-S60 13578 (9.66)ef 1782 (11.82)e 06.72ef 11.95ef Glulam 67758 (13.06)a 7285 (17.29)a 33.54a 62.22a * Numbers in parentheses are the coefficients of variation (%). ** Means with the same letter are not significantly different between groups based on Fisher’s least significant difference test (α = 0.05).

The effect of adding RF resin and self-tapping screws together was investigated by comparing I-777-G and I-777-GS40. The results showed no significant improvement in the maximum flexural load and strength by adding screws for the glued group; however, a significant increase in the flexural rigidity for I-777-GS40 was observed, suggesting that the self-tapping screws may be applied as a reinforcement to enhance the flexural rigidity for conventional glued products in practical use. Forests 2019, 10, x FOR PEER REVIEW 7 of 11

The effect of adding RF resin and self-tapping screws together was investigated by comparing I-777-G and I-777-GS40. The results showed no significant improvement in the maximum flexural loadForests and2019 , 10strength, 647 by adding screws for the glued group; however, a significant increase in7 ofthe 11 flexural rigidity for I-777-GS40 was observed, suggesting that the self-tapping screws may be applied as a reinforcementTo use built-up to beams enhance in structural the flexural applications, rigidity for weak conventional spots and glued failure prod modesucts should in practical be carefully use. consideredTo use andbuilt-up evaluated. beams In in this struct study,ural a applications, 4-point bending weak mode spots was and used, failure and themodes span-to-depth should be carefullyratio was considered 18, suggesting and long evaluated. beam flexural In this study, behavior. a 4-point Moreover, bending failure mode modes was of used, the built-up and the beamsspan- wereto-depth discussed ratio was based 18, onsuggesting ASTM D5055 long (2008)beam flexur [32]. Theal behavior. shear resistance Moreover, of the failure built-up modes beams of the resulted built- upfrom beams the connections were discussed between based the flangeson ASTM and D5055 the web. (2008) The [32]. stress The distribution shear resistance of the screwed of the built-up built-up beams resulted may be unlikefrom the that connections of typical between homogeneous the flanges beams, and and the the web. performance The stress distribution of the connection of the screwedfeatured built-up slippage beams between may the be members unlike that due of to ty horizontalpical homogeneous shear. Consequently, beams, and thethe resistanceperformance to the of thehorizontal connection shear featured of built-up slippage beams between relies mainlythe memb oners the due flexural to horizontal rigidity of sh theear. self-tapping Consequently, screws the resistanceresponsible to for the the horizontal bonding betweenshear of thebuilt-up flanges beams and the relies web, mainly as well on as the withdrawal flexural rigidity resistance of betweenthe self- tappingthe self-tapping screws responsible screws and for the the wood. bonding between the flanges and the web, as well as withdrawal resistanceIn general, between failures the self-tapping on the web screws and bottom and the flange wood. were observed under the loading positions (FigureIn 6general,), which failures was in accordanceon the web withand flangebottom failure flange under were tensionobserved and under indicated the loading typical positions bending (Figurefailure of 6), long-span which was flexural in accordance members. with flange failure under tension and indicated typical bending failure of long-span flexural members.

Figure 6. Failure usually resulting from the bottom flange at the position directly below the Figure 6. Failure usually resulting from the bottom flange at the position directly below the loading loading position. position. Figure7 presents the common failure modes of the developed built-up beams. The failures usually Figure 7 presents the common failure modes of the developed built-up beams. The failures resulted from the locations of the self-tapping screws and natural defects such as knots. Moreover, usually resulted from the locations of the self-tapping screws and natural defects such as knots. the failure locations were sometimes observed at the web, particularly at the position of the screws, Moreover, the failure locations were sometimes observed at the web, particularly at the position of indicating wood crushing failures caused by side compression from the self-tapping screws. Cracks the screws, indicating wood crushing failures caused by side compression from the self-tapping then propagated along the wood grain, eventually resulting in web failure. Johnson et al. [33] also screws. Cracks then propagated along the wood grain, eventually resulting in web failure. Johnson mentioned the failures would begin near the position of metal fasteners in flanges of wood I-beams. et al. [33] also mentioned the failures would begin near the position of metal fasteners in flanges of Occasionally, surface compression failure was observed at the loading positions on the top flanges. wood I-beams. Occasionally, surface compression failure was observed at the loading positions on For the glued built-up beams and the glulam beams, the failure locations were usually observed at the top flanges. For the glued built-up beams and the glulam beams, the failure locations were usually the natural defects of lumber due to stress concentration, which initiated from wood and propagated observed at the natural defects of lumber due to stress concentration, which initiated from wood and along the wood grain. Furthermore, tension failures at the bottom flange were typical cases for glued propagated along the wood grain. Furthermore, tension failures at the bottom flange were typical built-up beams and glulam. On the other hand, web buckling used to be of major concern for wood cases for glued built-up beams and glulam. On the other hand, web buckling used to be of major I-beams [34]; however, in this study, thick lumber webs were used so that no typical web buckling concern for wood I-beams [34]; however, in this study, thick lumber webs were used so that no typical was observed. web buckling was observed. Horizontal shear failure in the web at the edge of the beams can be found in the built-up I-beams, as shown in Figure8. In addition, slippage between the flanges and the web can be commonly observed in the screwed built-up beams (Figure9); furthermore, a longer spacing between the self-tapping Forests 2019, 10, x FOR PEER REVIEW 8 of 11 Forests 2019, 10, x FOR PEER REVIEW 8 of 11

ForestsHorizontal2019, 10, 647 shear failure in the web at the edge of the beams can be found in the built-up I-beams,8 of 11 Horizontal shear failure in the web at the edge of the beams can be found in the built-up I-beams, as shown in Figure 8. In addition, slippage between the flanges and the web can be commonly as shown in Figure 8. In addition, slippage between the flanges and the web can be commonly observed in the screwed built-up beams (Figure 9); furthermore, a longer spacing between the self- screwsobserved resulted in the inscrewed more noticeable built-up beams slippage (Figure between 9); furthermore, the flanges and a longer the web. spacing The slippagebetween betweenthe self- tapping screws resulted in more noticeable slippage between the flanges and the web. The slippage thetapping flange screws and web resulted may resultedin more fromnoticeable incomplete slippa interactionge between between the flanges the flangeand the and web. web The and slippage should between the flange and web may resulted from incomplete interaction between the flange and web bebetween carefully the considered flange and in web design may [15 resulted]. However, from adding incomplete glue improvedinteraction the between slippage the resistance flange and capacity web and should be carefully considered in design [15]. However, adding glue improved the slippage byand enhancing should be the carefully bonding considered between the in flangesdesign and[15]. web. However, Some previousadding glue studies improved also concluded the slippage that resistance capacity by enhancing the bonding between the flanges and web. Some previous studies metalresistance fastener-glued capacity by built-up enhancing beams the showed bonding greater between mechanical the flanges performance and web. Some than those previous using studies either also concluded that metal fastener-glued built-up beams showed greater mechanical performance metalalso concluded fasteners orthat glue metal [34,35 fastener-glued]. In particular, built-up when the beams screw showed position greater was close mechanical to the beam performance edge, deep than those using either metal fasteners or glue [34,35]. In particular, when the screw position was penetrationthan those using into the either flange metal due fasteners to surface or crushing glue [34,35]. by the In screws particular, was observed.when the Owingscrew position to horizontal was close to the beam edge, deep penetration into the flange due to surface crushing by the screws was shearclose to force, the beam self-tapping edge, deep screws penetration may be withdrawn; into the flange however, due to because surface self-tapping crushing by screws the screws exhibit was a observed. Owing to horizontal shear force, self-tapping screws may be withdrawn; however, because highobserved. grip onOwing wood, to surfacehorizontal crushing shear force, on flanges self-ta canpping be observed. screws may be withdrawn; however, because self-tapping screws exhibit a high grip on wood, surface crushing on flanges can be observed. self-tapping screws exhibit a high grip on wood, surface crushing on flanges can be observed.

Figure 7. Failure modes of the developed built-up and glulam beams: (a) tensile failure at bottom of Figure 7. Failure modes of the developed built-up and glulam beams: ( a) tensile failure at bottom of glulam, (b) compression failure at top flange, (c) failure at wood on glued surface, (d) horizontal shear glulam, (b) compression failurefailure atat top ﬂange,flange, (c)) failure at wood on glued surface, ( d) horizontal shear failure at the web, (e) crushing on the surface using self-tapping screws, (f) failure at knot, and (g) failure atat thethe web, web, ( e()e crushing) crushing on on the the surface surface using using self-tapping self-tapping screws, screws, (f) failure (f) failure at knot, at knot, and ( gand) crack (g) crack at the self-tapping screw joint location. atcrack the at self-tapping the self-tapping screw screw joint location. joint location.

Figure 8. Horizontal shear failure at th thee web of the built-up I-beams. Figure 8. Horizontal shear failure at the web of the built-up I-beams.

Forests 2019, 10, 647 9 of 11 Forests 2019, 10, x FOR PEER REVIEW 9 of 11 Forests 2019, 10, x FOR PEER REVIEW 9 of 11

Figure 9. End slip of the built-up beams. In addition, Figure 10 displays disassembled built-up beams, showing the representative results In addition, Figure 10 displays disassembled built-up beams, showing the representative results of distortion deformation of the screws used for assembling beams, indicating that screws used on of distortion deformation of the screws used for assembling beams, indicating that screws used on the bottom flanges caused more severe deformation than those used on the top flanges because of the bottom flanges caused more severe deformation than those used on the top flanges because of higher shear stress between the bottom flange and web than that between the top flange and the higher shear stress between the bottom flange and web than that between the top flange and the web. web. Moreover, increasing the spacing between the screws caused more apparent deformations. Moreover, increasing the spacing between the screws caused more apparent deformations. Nevertheless, because adding glue improved the horizontal shear resistance, smaller deformations Nevertheless, because adding glue improved the horizontal shear resistance, smaller deformations were observed on screws used in groups with glue, and the surface crushing on the flanges was were observed on screws used in groups with glue, and the surface crushing on the flanges was not not severe. severe.

Figure 10. Distortional deformation of the screws of the built-up I-beam (a) the top flange of I-777-S40; Figure 10. Distortional deformation of the screws of the built-up I-beam (a) the top flange of I-777- (b) the bottom flange of I-777-S40. S40; (b) the bottom flange of I-777-S40. 4. Conclusions 4. Conclusions In this study, built-up beams were developed using Japanese cedar (Cryptomeria japonica (L. f.) In this study, built-up beams were developed using Japanese cedar (Cryptomeria japonica (L. D. Don)In this and study, self-tapping built-up screws, beams and were the developed effects of diusingfferent Japanese assembly cedar configurations (Cryptomeria on japonica the flexural (L. f.) D. Don) and self-tapping screws, and the effects of different assembly configurations on the load-bearingf.) D. Don) and capacity self-tapping and failure screws, modes and of the the products effects wereof different investigated. assembly Results configurations showed that addingon the flexural load-bearing capacity and failure modes of the products were investigated. Results showed that adding glue provided flexural rigidity, whereas using self-tapping screws resulted in built-up

Forests 2019, 10, 647 10 of 11 glue provided flexural rigidity, whereas using self-tapping screws resulted in built-up beams with high ductility but relatively low flexural bearing capacity. However, enhancing the flexural ductility is important for practical applications. When using components of similar grades, adding more web components can improve the performance; however, the location of the web components may not significantly influence the results. Furthermore, the grades of the web have negligible effect on the flexural load-bearing capacity, whereas the grades of the flanges have substantial influence on the flexural load-bearing capacity. In addition, a smaller spacing between the screws improves the flexural load-bearing performance, but also causes cracks in wooden components. Typical bending failure modes were observed in the built-up beams, indicating tension failure on the bottom flange and slippage between the flanges and the web due to horizontal shear, which also caused buckling deformation in the screws. On the other hand, horizontal shear failure in the web at the edge of the beams can be found in the built-up I-beams, particularly when longer spacing between the self-tapping screws was applied, and adding glue would improve the horizontal shear resistance. As well, surface failure on flanges due to compression and wood crushing failures at webs caused by side compression from the self-tapping screws and could also be found. Moreover, failures commonly result from natural defects on the bottom flange or the place on the bottom flange directly below the loading position.

Author Contributions: Conceptualization, T.-H.Y. and F.-C.C.; Data curation, T.-P.C.; Formal analysis, T.-P.C.; Funding acquisition, F.-C.C.; Investigation, T.-H.Y.; Methodology, F.-C.C.; Project administration, F.-C.C.; Resources, T.-H.Y.; Supervision, T.-H.Y. and F.-C.C.; Visualization, T.-P.C.; Writing – original draft, T.-P.C.; Writing – review & editing, T.-H.Y. and F.-C.C. Funding: This research was funded by National Taiwan University Experimental Forest. Project number [105B03 and 106B02]. This research was also partially funded by the Ministry of Science and Technology, Taiwan [MOST 107-2313-B-005 -019 -MY3]. Acknowledgments: The authors are grateful for the technical support provided by Min-Jay Chung, National Taiwan University Experimental Forest, and Meng-Ting Tsai, The Department of Architecture, National Taiwan University of Science and Technology. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Chiu, L.W.; Huang, C.H.; Wu, C.C.; Hsieh, H.T. The fourth national forest resources inventory investigation results summary. Taiwan For. J. 2015, 41, 3–13. (In Chinese) 2. Bernal, B.; Murray, L.T.; Pearson, T.R.H. Global carbon dioxide removal rates from forest landscape restoration activities. Carbon Balance Manag. 2018, 13, 22–34. [CrossRef][PubMed] 3. Csoka, L.; Zhu, J.J.; Takata, K. Application of Fourier analysis to determine the demarcation between juvenile and mature wood. J. Wood Sci. 2005, 51, 309–311. [CrossRef] 4. Leichti, R.J.; Tang, R.C. Comparative performance of long-term loaded wood composite I–beam and swan lumber. Wood Fiber Sci. 1989, 21, 142–154. 5. Yeh, M.-C.; Lee, W.-H.; Lin, Y.-L. Study of structural glulam manufacturing from domestic Japanese cedar plantation wood. Taiwan J. For. Sci. 2006, 21, 531–546. (In Chinese) 6. Yeh, M.-C.; Xie, Z.-X.; Chen, M.-J.; Wu, M.-Y. Evaluation on the flexural properties of flooring I-joist using commercial wood species. For. Prod. Ind. 2003, 22, 33–44. (In Chinese) 7. Neksin, S.A. Prefabricated Wood Joists in Wood Structures: A Design Guide and Commentary; ASCE Structural Division: New York, NY, USA, 1975; pp. 213–215. 8. Broker, F.W.; Krause, H.A. Preliminary investigations on the holding power of dynamically loaded wood-screws. Holz Roh Werkst. 1991, 49, 381–384. 9. Meghlat, E.M.; Oudjene, M.; Ait-Aider, H.; Batoz, J.L. A new approach to model nailed and screwed timber joints using the finite element method. Constr. Build. Mater. 2013, 41, 263–269. [CrossRef] 10. Yeh, M.-C.; Hsieh, C.-H. An evaluation on withdrawal resistance of nail joints for domestic Cryptomeria japonica sawn lumber. For. Prod. Ind. 1995, 14, 551–564. (In Chinese) Forests 2019, 10, 647 11 of 11

11. Bues, C.T.; Schulz, H.; Eichenseer, F. Investigation of the pull-out resistance of nails and screws in pine wood. Holz Roh Werkst. 1987, 45, 514. 12. Chang, K.-Y.; Yeh, M.-C. Flexural performance of I-beam fabricated from Cunninghamia lanceolata var. lanceolata. For. Prod. Ind. 2001, 20, 195–205. (In Chinese) 13. Liu, C.-T. Flour in adhesives. Q. J. For. Res. 1979, 1, 16–29. (In Chinese) 14. Liu, C.-T.; Lee, W.-J.; Hsia, I.-C. Chemically modified wood and its gluing properties (IV)—Gluing and coating properties of acetylated wood. Q. J. For. Res. 2000, 22, 27–36. (In Chinese) 15. Booth, L.G. The effect of flange-web joint design on the design of plywood I–beams. J. Inst. Wood Sci. 1974, 6, 13–24. 16. Fergus, D.A. Effect of web voids and stiffeners on structural performance of composite I-beams. Ph.D. Thesis, Purdue University, West Lafayette, IN, USA, 1979. 17. Geimer, R.L.; Lehmann, W.F. Product and process variables associated with a shaped particle beam. For. Prod. J. 1975, 25, 72–80. 18. Hilson, B.O.; Rodd, P.D. The ultimate shearing strength of timber I-beams with hardboard webs. Struct. Eng. B 1979, 57, 25–36. 19. Laufenberg, T.L. Exposure effects upon performance of laminated veneer lumber and glulam materials. For. Prod. J. 1982, 32, 42–48. 20. Leichti, R.J. Assessing the Reliability of Wood Composite I-Beams. Ph.D. Thesis, Auburn University, Auburn, AL, USA, 1986. 21. Samson, M. Influence of flange quality on load capacity of composite webbed I–beams in flexure. For. Prod. J. 1983, 33, 38–42. 22. Superfesky, M.J.; Ramaker, T.J. Hardboard-Webbed I-Beams Subjected to Short-Term Loading; USDA Forest Service Research Paper FPL 264; Forest Products Laboratory: Madison, WI, USA, 1976; p. 12. 23. Ohashi, Y.; Toda, M.; Fujiwara, T.; Sato, T.; Hirai, T. Mechanical properties of wooden I-beams with plantation timber materials in Hokkaido (I) bending, shear and partial bearing properties. Mokuzai Gakkaishi 2008, 54, 24–32. (In Japanese) [CrossRef] 24. Chen, G.H.; Tang, R.C.; Price, E.W. Environmental Effects on the Engineering Performance of Wood Composite I-Beams; Technical Forum Abstract, International Meeting; Forest Products Research Society: Madison, WI, USA, 1987; p. 2. 25. Lai, C.-Y.; Yeh, M.-C. Effects of web materials on flexural properties of China fir composite I-beams. Taiwan J. For. Sci. 2002, 17, 219–230. (In Chinese) 26. Tsai, H.-J. Application of Japanese Cedar Structural Cross Panel on the Composite I-Beam Manufacturing and Performance Evaluation. Master’s Thesis, National Pingtung University of Science and Technology, Neipu, Taiwan, 2010. (In Chinese). 27. Hunt, M.O. Structural particleboard for webs of composite beams. For. Prod. J. 1975, 25, 55–57. 28. Percival, D.H.; Hunt, M.O.; Comus, Q.B.; Suddarth, S.K. Pilot test of four 16-foot, wood-base composite garage headers. For. Prod. J. 1977, 27, 45–48. 29. Bullen, J.C.; Van Der Straeten, E. The design and construction of glued joints. J. Inst. Wood Sci. 1986, 10, 220–228. 30. Jokerst, R.W. Finger-Jointed Wood Products; USDA Forest Service Research Paper FPL 382; Forest Products Laboratory: Madison, WI, USA, 1981; p. 23. 31. Leichti, R.J.; Tang, R.C. Analysis of Wood Composite I-Beams with Glued Flange-Web Joints; Proceedings, Spring Meeting; Society for Experimental Stress Analysis, Brookfield Center: Brookfield, MA, USA, 1983; pp. 45–50. 32. ASTM D5055. Standard Specification for Establishing and Monitoring Structural Capacities of Prefabricated Wood I-Joist; ASTM International: West Conshohocken, PA, USA, 2008. 33. Johnson, J.A.; Ifju, G.; Rogers, H.W. The performance of composite wood/particleboard beams under two-point loading. Wood Fiber Sci. 1975, 8, 85–97. 34. Leichti, R.J.; Falk, R.H.; Laufenberg, T.L. Prefabricated wood composite I-beams: A literature review. Wood Fiber Sci. 1990, 22, 62–79. 35. Van Wyk, W.J. The strength, stiffness and durability of glued, nail-glued and screw-glued timber joints. S. Afr. For. J. 1986, 138, 41–44. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).