Jwo0dSg(2002)48138-146 The Japan Wood Society ORIGINAL ARTICLE Kei Sawata.Motoi Yasumura Determination of embedding strength of wood for dowel-type fasteners Received:August 2100/Accepted:April 18001 Introduction bedding strength of wood for the design of dowel-type 16.and 20mm.Embe EN383.The embedding strength parallel to the grain ment of the fastener governs properties for determining evaluate d by the method eners.Numerous h close to n of the embedding tes heeaPdhcameaheeaor8eha prenth perpendicul hen the and end-distance on the bearing characteristic of glued hen the aminat d timber mber.These stud ated by a maximum load up to 5mm displacement ac cording to EN383,the ratio of embedding strength might include the effects of the fracture of wood. nhaddina s compreh harve inve increased. and tempered hardboard with nails and bolts:and Ehlbeck ·Compressive strength the basiso h to the grain. in dealt only with ultimate embedding strength;they did not look at the yield embedding strength pendicular to the relations be ter an g d l ureS Shizuoka nated timber and some engineered woods,respectively 428529,92 These studis were based ve tests with cube tests parallel and perpendicular to the grain according to EN3 to examine both yield and ultimate embedding f the houhumerous studies have been reported on the relations between the embedding strength of wood or
J Wood Sci (2002) 48:138-146 9 The Japan Wood Research Societ 3, 2002 Kei Sawata 9 Motoi Yasumura Determination of embedding strength of wood for dowel-type fasteners Received: August 21, 2000 / Accepted: April 18, 2001 Abstract Embedding tests parallel and perpendicular to the grain were conducted to produce a database of embedding strength of wood for the design of dowel-type joints. Dowel diameters were 8, 12, 16, and 20mm. Embedding strength was evaluated by the 5% off-set method and a maximum load up to 5mm displacement according to EN383. The embedding strength parallel to the grain evaluated by the former method showed values close to those obtained with the latter method, but they showed a significant difference in tests conducted perpendicular to the grain. The embedding strength parallel to the grain was 0.9 times as large as the compressive strength parallel to the grain regardless of the evaluation method. The embedding strength perpendicular to the grain evaluated by the 5% off-set method was four times as large as the compressive strength perpendicular to the grain. When the embedding strength perpendicular to the grain was evaluated by a maximum load up to 5mm displacement according to EN383, the ratio of embedding strength perpendicular to the grain to the compressive strength perpendicular to the grain decreased as the dowel diameter increased. Key words Embedding strength 9 Compressive strength 9 Density 9 5% Off-set method K. Sawata. M. Yasumura (EEl) Faculty of Agriculture, Shizuoka University, 836 Ohya, Shizuoka 422-8529, Japan Tel. +81-54-238-4863; Fax +81-54-237-3028 e-mail: afmyasu@agr.shizuoka.ac.jp Part of this paper was presented at the annual meeting of the Architectural Institute of Japan 1999, Hiroshima, September 1999; the 50 th annual meeting of the Japan Wood Research Society, Kyoto, April 2000; and the World Conference on Timber Engineering 2000, Whistler, Canada, July-August, 2000 Introduction Johansen's yield theory is widely used for estimating the yield strength of dowel-type joints. According to this theory, the embedding strength of wood and the yield moment of the fastener governs properties for determining the strength of joints with dowel-type fasteners. Numerous studies have been performed on embedding characteristics of wood and wood-based materials with dowel-type fasteners. Hirai 2'3 investigated the influence of the embedding test method and round bar diameter on embedding strength and stiffness. Fujita et al. 4 examined the effect of edge-distance and end-distance on the bearing characteristic of glued laminated timber and laminated veneer lumber. These studies were based on the embedding tests in tension, and the ultimate embedding strength obtained in these studies might include the effects of the fracture of wood. Whale et al. 5 conducted a comprehensive investigation of the embedding strength of softwood, hardwood, plywood. and tempered hardboard with nails and bolts: and Ehlbeck and Werner 6 carried out embedding tests on hardwood under various loading angles to the grain. These studies, being the basis of the design of dowel-type joints in Eurocode 5.- dealt only with ultimate embedding strength; they did not look at the yield embedding strength. Kawamoto et al. ~ carried out the embedding tests perpendicular to the grain of glued laminated timber: and Harada et al. 9 and Hwang and Komatsu ~ investigated the relations between the dowel diameter and the embedding properties obtained by the embedding tests of glued laminated timber and some engineered woods, respectively. These studies were based on compressive tests with cubeshaped specimens, and ultimate embedding strength was not considered. Therefore. we conducted the embedding tests parallel and perpendicular to the grain according to EN383 tl to examine both yield and ultimate embedding strengths. Although numerous studies 3-<s<~ have been reported on the relations between the embedding strength of wood or
139 wood-based materials and round bar diameter,few data R embedding strength have been one of the most efficien which was 1.5dand≤4d ing p The were cut near the embed was 32mm square in the loading section and 64mm in design.There height. to producein em design o ow Embedding tests type joint ing strength from the tests. of the wooden member and wer aralle Materials and methods th-0MP /mn tor the cEmbed. to the grain was to the grain were terminate hen the Specimens or when the load decreased to half the maximum load.The Embedding and compressive tests were conducted on embedding stress increment perpendicular to the grain was n to the elastic area.Em ding test 90.100.L110.L125) 1a9 grade tural Standard.Approximately 1000 specimens were cu crack reached the end of the wood. more than 300 Compressive tests nd cording to EN383 were 14d in length,6d in width,and Table 1.Number and density of embedding specimens Grade Parallel to the grain Perpendicular to the grain No.dnens Density Density Mean (kg/m) CV(% Mean (kg/m)CV(%) 8mm 744514 193 011 4514 110007 09 120660 1190023 000 dowel diameter CV. variation
wood-based materials and round bar diameter, few data on the variance of embedding strength have been sought. Reliability-based design is one of the most efficient methods for evaluating the mechanical performance of timber structures. A statistical approach when considering the variance of embedding strength might be required to adapt the yield theory to reliability-based design. Therefore, embedding tests were conducted on a thousand laminae with different grades, dowel diameters, and loading directions to produce a database for the design of doweltype j oints.~2'13 Estimating the embedding strength from the compressive strength of wood was also proposed by comparing the embedding test results with those of compressive tests. Materials and methods Specimens Embedding and compressive tests were conducted on ezomatsu (Picea jezoensis Carriere) and todomatsu (Abies sachalinensis Fr. Schmidt) laminae, which had four grades (L90, L100, Lll0, L125) according to the Japanese Agricultural Standard. TM Approximately 1000 specimens were cut from more than 300 laminae. Dowel diameters (d) were 8, 12, 16, and 20 mm. The number and densities of embedding specimens for each dowel diameter and grade are shown in Table 1. The dimensions of the embedding specimen according to EN383" were 14d in length, 6d in width, and 139 32 mm in thickness, as shown in Fig. 1. The thickness of our specimen was within the range of that defined in EN383, which was ->l.5d and -<4d. Compressive specimens were cut near the embedding specimens. The dimension of compressive specimens was 32mm square in the loading section and 64ram in height. Embedding tests Embedding tests according to EN383 were conducted as shown in Fig. 1. Steel side plates 12mm thick were placed on both sides of the wooden member and were connected with a dowel. There was no clearance between the steel plates and the specimen. The embedding stress increment parallel to the grain was 10-30 MPa/min for the elastic area. Embedding tests parallel to the grain were terminated when the embedding displacement was equal to the dowel diameter or when the load decreased to half the maximum load. The embedding stress increment perpendicular to the grain was 3-10 MPa/min for the elastic area. Embedding tests perpendicular to the grain were stopped when the embedding displacement was equal to the dowel diameter or when the crack reached the end of the wood. Compressive tests Compressive tests parallel and perpendicular to the grain were conducted on the specimens cut from the wood near Table 1. Number and density of embedding specimens Grade Parallel to the grain No. of Density specimens Mean (kg/m 3) CV (%) Perpendicular to the grain No. of Density specimens Mean (kg/m 3) CV (%) d = 8mm Total 57 391 11.4 L90 14 381 11.0 L100 14 369 8.36 Lll0 15 399 9.58 L125 14 414 12.3 d = 12mm Total 117 394 10.2 Lg0 30 359 6.47 L100 30 393 12.0 Lll0 30 394 6.05 L125 27 422 8.26 d = 16mm Total 212 399 11.1 Lg0 50 350 10.2 L100 56 389 4.99 Lll0 56 411 7.52 L125 50 448 5.33 d = 20mm Total 117 403 10.7 Lg0 29 372 7.05 L100 30 381 9.65 Ll10 30 419 5.63 L125 28 441 9.67 57 387 10.9 14 373 9.57 14 364 7.10 15 398 9.13 14 412 12.0 119 389 10.3 30 351 5.05 30 383 9.36 30 401 9.87 29 420 6.89 212 403 11.1 50 357 7.64 56 392 4.45 56 404 7.73 50 460 6.08 118 401 12.2 30 361 6.26 30 372 8.50 30 418 5.68 28 450 11.1 d, dowel diameter; CV coefficient of variation
140 Parallel to the grain Perpendicular to the grain Load cell 5●》● Dowe 3d "3d 32m 40 d=8mm 8mm d=12mm d=12mm =16mm (edW)ssans 3uppac d-20mm 30 d=20mm 30 20 d=16mm 10 Parallel to the grain Perpendicular to the grain 0 10 15 20 25 10 15 25 Displacement(mm) Displacement(mm) Fig.2.Relations bet displacemen the embedding specimen according to the Japanese Indus- almost constant after vielding regardless of the increase trial Standard. Approximately 500 specimens with a den- siy生 to that of the pto te Ylcldins of terminated after the maximum load was attained and thos ment ratio after yielding was smaller as the dowel diameter as shown i the yield embedding strength.With the former method the Results and discussion ine that goes through the points on the curves correspond Evaluation methods of embedding strength o the direction and embedding strength is defined as he tion o this the edding showed a linear increase up to the yielding of wood and was 5%embedding strength and 5mm embedding strength in
140 Fig. 1. Configuration of embedding test. d, dowel diameter (millimeters) 50 [ d=amm ] 40 [ d=amm d=lZmm ~" [ / d=12mm [ ~" / ~ /.~ d=16mm ,40t I i3~176 ~' 30 20 I 20 m 0 0 5 1 0 15 20 25 0 5 10 15 20 25 Displacement(ram) Displacement(ram) Fig. 2. Relations between embedding stress and displacement the embedding specimen according to the Japanese Industrial Standard. 15 Approximately 500 specimens with a density (+_10%) close to that of the embedding specimens were tested. Compressive tests parallel to the grain were terminated after the maximum load was attained, and those perpendicular to the grain were stopped when the displacement was >10% of the height of the specimen. Results and discussion Evaluation methods of embedding strength The typical embedding stress-displacement curves are shown in Fig. 2. The embedding stress parallel to the grain showed a linear increase up to the yielding of wood and was almost constant after yielding regardless of the increase in displacement. The embedding stress perpendicular to the grain showed a liner a increase up to the yielding of wood and a continuous increase after yielding. The increment ratio after yielding was smaller as the dowel diameter was increased. Embedding strengths were evaluated by the 5% off-set method and according to EN383, as shown in Fig. 3. The 5% off-set method was adopted to evaluate the yield embedding strength. With the former method the line that goes through the points on the curves corresponding to 10% and 40% of the maximum load up to 5mm displacement was moved 5% of the dowel diameter parallel to the X-direction, and embedding strength is defined as the intersection of this line and the load-displacement curve. The latter is defined as the maximum load up to 5mm displacement. The former and the latter are expressed as 5% embedding strength and 5mm embedding strength in
14 Embedding stres Parallel to the grain 0.05d f尼.5m 尼.5 20 0.4 fe.sm 0.1尼5n Perpendicular to the grain 5mm Displacement d=8mm d产l2mmd-16mmdk-20mm Fig3.Method for evaluating embedding strength (fe) r 10:1a红d 60 60 5% 5mm fe,0,5%=-0.188+36.64 e,0,5mm=-0.12836.85 (edW) 40 40 五 10 e,90,5%=-0.095d+15.52 f尼，90,5mm=-0.508d+25.66 8 12 16 20 24 Y 16 20 24 Dowel diameter(mm) Dowel diameter(mm) this paper,respectively.Embedding strength was calculated strength had a positive correlation with the density.The as follows rent with dowels ofs and 12mm diameter sho slight increase.The increase in the 5mm embedding the load,is the thickness of the specimen,and d is the endcnewasoeheahgetendeai dowel diameter. This might be caused by the fact that there wasitteffer ence in average Embedding strength average density was larger with higher lamina rade for dowels 16 or 20mm in diameter.The data from the various amina grades were c in this study to simplify the y inereased the faninm grade hhigher.as show The embedding strengths parallel and perpendicular to di
Embedding stress f~), 5mm ,fi?.5% 0.4 ,/i,, 5~m~ O. 1 ./b,Sm~ 0 5mm Displacement Fig. 3. Method for evaluating embedding strength (re) 141 tt~ ,< 50 40 3O ._,_-1 20 10 0 d=8mm Parallel t t the grain 'erpendicule d=121nm to the gra t d = 16mm d--20mm Fig. 4. Relations between 5 mm embedding strength and lamina grade. Diamonds, L90; squares, L100; triangles, Lll0; circles, L125 60 50 .q" 40 "" 30 Lz~ 20 10 5% .1~,0,5% = -0.188d+36< 64 60 50 4O 30 ,~ 20 .1e,90,5O/o = -0.095d+15.52 0 i i i i 0 4 8 12 16 20 24 4 Dowel diameter(mm) 10 5mm ./e,0, 5mm = -0.128d+36.85 .fe,9~ = -0 508d 25.66 I I I I 8 12 16 20 Dowel diameter(ram) 24 Fig. S. Embedding strength as a function of dowel diameter. Circles, parallel to the grain; squares, perpendicular to the grain; symbols and vertical bars denote the mean value and standard deviation, respectively this paper, respectively. Embedding strength was calculated as follows. P fe~,~ - td (1) where fe is the embedding strength, a is the loading angle to the grain, fl is the evaluation method (i.e., 5% or 5mm), P is the load, t is the thickness of the specimen, and d is the dowel diameter. Embedding strength Table 2 shows the mean values and the coefficient of variation of the embedding strength parallel and perpendicular to the grain for each dowel diameter and lamina grade. The density increased as the lamina grade was higher, as shown in Table 1. Some studies 3'8 reported that the embedding strength had a positive correlation with the density. The 5 mm embedding strength for each lamina grade are shown in Fig. 4. As the grade increased, the 5mm embedding strength with dowels of 8 and 12ram diameter showed a slight increase. The increase in the 5mm embedding strength due to lamina grade was more significant in the specimens with a 16 or 20mm diameter dowel. The same tendency was observed with the 5% embedding strength. This might be caused by the fact that there was little difference in average density among each lamina grade used for specimens with dowels 8 or 12 mm in diameter, whereas the average density was larger with higher lamina grade for dowels 16 or 20mm in diameter. The data from the various lamina grades were combined in this study to simplify the analysis. The embedding strengths parallel and perpendicular to the grain for each dowel diameter are shown in Fig. 5. Although yielding of the dowel does not occur with a slen-
142 Table 2.Results of embedding tes Evaluation method Parallel to the grain Perpendicular to the grain Mean (MPa) CV(%) Mean (MPa) Cv(%) 5%Embedding strength 飞级助1110 6500 289T9 edding strength 6068 防纸级k10 87889 was obtained from the experimental an ular to the grain regardless of dowel diameter for practical purposes.It was that obvious that the 5mm embedding strength perpendicular to he er ng strength m constant me the strengths parallel to the grain were almost constant regar grainpcpCndiCularoienrooeachdoNeldia less of dowel diameter.which agreed with the latter study. eter.The ratio for the 5%embedding strength was almost
142 Table 2. Results of embedding test Evaluation method Parallel to the grain Mean (MPa) CV (%) Perpendicular to the grain Mean (MPa) CV (%) 5% Embedding strength d = 8mm Total 26.7 17.1 14.6 17.7 L90 26.8 8.42 15.0 10.8 L100 25.1 11.6 13.5 14.7 Lll0 28.8 23.3 15.3 20.8 L125 25.9 15.1 14.4 19.4 d = 12mm Total 35.4 10.5 14.2 15.5 L90 36.4 9.81 14.1 8.84 L100 35.6 12.0 14.2 19.5 Lll0 33.6 7.43 14.0 15.1 L125 36.2 11.0 14.7 15.8 d = 16mm Total 32.6 13.2 14.3 17.3 L90 28.4 10.4 12.7 16.8 L100 31.3 7.81 14.5 13.6 Ll10 33.9 9.61 14.2 20.1 L125 36.7 9.74 15.9 10.7 d - 20mm Total 33.9 11.2 13.3 20.3 Lg0 31.9 8.18 12.5 14.6 L100 31.9 9.51 11.8 17.1 Lll0 35.1 7.18 13.3 12.3 L125 36.6 11.9 15.4 21.8 5 mm Embedding strength d = 8mm Total 34.2 10.8 22.4 16.9 L90 33.8 6.26 21.8 10.4 LI00 32.9 6.11 21.6 16.7 Lll0 35.3 14.1 23.6 18.7 L125 34.7 11.9 22.5 18.2 d = 12mm Total 37.1 10.6 18.6 17.4 L90 37.3 9.83 17.6 il.9 L100 37.0 12.8 18.7 19~0 Ll10 36.0 7.43 18.5 I6.9 L125 38.3 11.0 19.7 18.1 d = 16mm Total 33.8 12.9 17.9 16.1 L90 29.6 11.6 15.4 16.3 L100 32.4 6.98 18.1 10.1 Lll0 34.7 9.42 !7.9 116.1 L125 38.5 7.84 20.3 10.1 d = 20mm Total 34.3 11.1 15,3 20.9 Lg0 32.5 7.63 14.5 14.2 L100 32.5 9.61 13.7 17.7 Lll0 35.6 7.64 15.5 14.6 L125 37.0 12.0 17.7 22.9 derness ratio of 4, the effects of bending the dowel on the load-displacement curve cannot be ignored. The regression equation of the 5 % embedding strength parallel to the grain was obtained from the experimental results with dowels 12, 16, and 20mm in diameter, excluding those with dowels 8mm in diameter. Whale et al. 5 reported that the embedding strength parallel to the grain decreased as the dowel diameter increased. Hirai 3 and Harada et al. 9 reported that the embedding strength was almost constant regardless of dowel diameter. In our study, the 5 % and 5 mm embedding strengths parallel to the grain were almost constant regardless of dowel diameter, which agreed with the latter study. For conditions perpendicular to the grain, numerous studies 3'4'6'9 reported that embedding strength decreased as the dowel diameter increased, which may be caused by the effects of crack propagation. 3 The 5% embedding strength perpendicular to the grain decreased slightty as the dowel diameter increased but could be considered almost constant regardless of dowel diameter for practical purposes. It was obvious that the 5 mm embedding strength perpendicular to the grain decreased as the dowel diameter increased. Figure 6 shows the embedding strength parallel to the grain/perpendicular to the grain ratio for each dowel diameter. The ratio for the 5% embedding strength was almost