- Email: [email protected]
Analysis of the fracture behavior of Radiata Pine timber and Laminated Veneer Lumber
Engineering Fracture Mechanics 116 (2014) 1–12
Contents lists available at ScienceDirect
Engineering Fracture Mechanics journal homepage: www.elsevier.com/locate/engfracmech
Analysis of the fracture behavior of Radiata Pine timber and Laminated Veneer Lumber Bettina Franke a,⇑, Pierre Quenneville b a b
Bern University of Applied Sciences, Department of Architecture, Wood and Civil Engineering, Solothurnstrasse 102, PO Box 6096, 2500 Biel/Bienne 6, Switzerland The University of Auckland, Department of Civil and Environmental Engineering, Private Bag 92019, Auckland 1142, New Zealand
a r t i c l e
i n f o
Article history: Received 20 April 2012 Received in revised form 7 July 2013 Accepted 8 December 2013
Keywords: Fracture mechanics Mode I Mode II Timber Radiata Pine Laminated Veneer Lumber
a b s t r a c t The investigation of the failure mechanism of the fracture mode I and II and the mixed mode of New Zealand Radiata Pine timber and of Radiata Pine Laminated Veneer Lumber is presented. Different test setups were compared and used for the determination of the material parameters fracture toughness and fracture energy. Furthermore, the differences, advantages and disadvantages of Radiata Pine Laminated Veneer Lumber versus solid wood are presented. The investigations show that Radiata Pine Laminated Veneer Lumber is more ductile than Radiata Pine timber. Ó 2013 Elsevier Ltd. All rights reserved.
1. Background Typical timber constructions consist of members and also joints, notches or holes which can lead to high stress singularities in the structure. Fracture mechanic gives the possibility to characterize situations with stress singularities which can lead to a failure including crack initiations and crack propagations. For the prediction of failures or load capacities, the fracture mechanics methods are used in new design proposals. However, important material parameters such as the critical fracture energy or the fracture toughness have to be known. International research results are published for solid wood e.g. in Radiata Pine by King et al. [1], in European Spruce e.g. by Franke [2], Aicher et al. [3], Valentin et al. [4], Larsen and Gustafsson [5] or in Canadian Spruce by Smith et al. [6] or Vasic [7]. For engineered wood products, only a portion of the values are published, e.g. Stanzl-Tschegg and Navi [8], Niemz et al. [9,10]. The trend to large scale and multistory timber constructions requires high performance wood products with increased strengths and stiffness. In order to allow the analysis of failures in structures made of Radiata Pine timber (RP) or Laminated Veneer Lumber (LVL), a study to determine the fracture energy and fracture toughness was initiated. The research results presented include the investigation of the failure behavior of New Zealand Radiata Pine LVL compared to timber. The difference, possible advantages or maybe disadvantages of LVL versus timber are investigated and compared with published international experimental test results. When investigating the failure behavior, different test setups are available to study the different fracture modes. The fracture mechanic method distinguishes three different failure modes, as shown in Fig. 1. They are: failure mode I – cracking under transverse tension stress in relation to the crack plane, mode II – cracking under in plane shear stress and mode III – the failure under out of plane shear stress. The so-called mixed mode failure is a combination of the mode I and mode II. The fracture modes I and II and for some cases, the mixed mode failure were investigated in experimental test series. None ⇑ Corresponding author. Tel.: +41 32 344 0308. E-mail addresses: [email protected] (B. Franke), [email protected] (P. Quenneville). 0013-7944/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.engfracmech.2013.12.004
2
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Nomenclature 0
E F Fc G Gf Gc K Kc Y a ac d h u
a b
r q I II III
specific modulus of elasticity (MPa) load (N) critical load (N) shear modulus (MPa) energy release rate (Nm/m2) critical fracture energy (Nm/m2) stress intensity factor (SIF) (kN/m3/2) fracture toughness (kN/m3/2) form function of the geometry (–) crack length (mm) corresponding crack length at critical load (mm) thickness of the specimen (mm) depth of the specimen (mm) displacement (mm) loading angle of the system (°) annual growth ring (°) stress (MPa) density (kg/m3) index for fracture mode I (–) index for fracture mode II (–) index for fracture mode III (–)
of the test setups available for the different failure modes are approved or standardized for wood. Therefore different methods were used and are compared for the same stress situation. 2. Material and method 2.1. Material and specimens New Zealand Radiata Pine timber and LVL were used for the analysis of the failure behavior of solid wood and wood product respectively. New Zealand Radiata Pine is a relatively fast growing plantation pine species which is fully grown after approximately 20–30 years. The timber used grew in a relatively constant climate on the North Island of New Zealand. The growth ring width was 10 mm on average and the specimens were without major defects. The timber was conditioned in an environment with a temperature of 20° Celsius and a relative humidity of 65%. At testing, the average moisture content from the test specimens was 11.4% and the mean density was 486 kg/m3. LVL is a wood product manufactured from rotary peeled veneer layers glued together to form slabs. The veneers are 3–4 mm thick and in general laid up with parallel to grain direction. Natural defects such as knots are more homogenously distributed, which lead to higher strength and less variation in comparison to timber. Even if there are small derivations between the grain directions of the veneers due to the practical production, which might influences the properties and e.g. the response in the post peak-resistance behavior; LVL shows no behavior related to cross-laminations and the influences are respected in the variations of the properties. LVL products are used in houses and large span buildings as structural members. The hyspan90 product from the Carter Holt Harvey company from the North Island of New Zealand was used. The LVL specimens were also stored in the conditioned environment chamber. The average moisture content of the LVL test specimens was 8.9% and the mean density was 600 kg/m3. For the consideration and investigation of the influence of the anisotropic material behavior of solid wood on the fracture energy in the different material directions, three different crack systems, RL, TL and NL were tested, as shown in Fig. 2. The
Fig. 1. Different test setups for the fracture mode I, II and III and the mixed mode.
3
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Fig. 2. Specimens crack systems and orientations.
crack systems are described by the natural material axes, radial R, tangential T and longitudinal L. The letter N defines the growth ring orientation of about 45°, meaning an orientation between the tangential and radial orientation. The first letter specifies the direction normal to the crack plane and the second letter the crack growth direction. Due to the constitution of the wood product LVL, only the two crack system RL and TL were considered, as shown in Fig. 2. Furthermore for each crack system, the fracture mode I and mode II as well as the mixed mode of mode I and mode II were investigated within the test series. The fracture modes can be also specified by the load angle a, which is the angle between the load applied and the crack plane defined. A load angle a of 90°, resulting in transverse stress, leads to pure mode I failure where a of 0°, resulting in shear stress, leads to pure mode II failure. Load angles a between 0° and 90° result in a mixed mode failure of mode I and II. An overview of the test groups, the cracking systems as well as the number of each test configuration is given in Table 1. The sizes of the specimen are defined by the depth of the member h and the thickness d at the crack plane. 2.2. Test setups For the investigation of the failure behavior of wood under tension perpendicular to grain as well as shear and for the determination of fracture parameters, different test setups are available and accepted as shown in e.g. Lo et al. [11], Aicher et al. [3], Xu et al. [12], Valentin et al. [4], Valentin and Caumes [13] and Cramer and Pugel [14]. The first published test setups for wood were adopted from test setups developed for isotropic material such as steel. These test methods and associated equations were modified to make them valid for wood which can be, at best, assumed as an orthotropic material. Three test setups published were chosen for the investigation of Radiata Pine timber and LVL; for the pure fracture mode I, the single end notched beam specimen (SENB) [5], and for the pure fracture mode II, the compact shear specimen (CSS) [15,16]. The compact tension shear test specimen (CTS) [13,17], was used for the investigation of the mixed mode of the fracture mode I and II. This test setup allows the investigation of the pure fracture modes I and II as well as the mixed mode using the same specimen geometry and load application. Therefore the size effect and the effect due to loading procedure is neglected in the comparison of the fracture behavior and energies. For the fracture mode I, the single end notched beam (SENB) specimen was used according to the draft standard from Larson and Gustafsson [5]. This test setup was agreed by the CIB-W18 working group for a joint project for the determination of the fracture energy and the evaluation of the test setup. The test setup is established for wood and for the fracture mode I. The tested specimen is glued to two wooden beams, as shown in Fig. 3. The grain direction of the specimen is perpendicular to the span of the beam. The orientation and the pre cut notch results in tension stresses normal to the crack plane (failure
Table 1 Test program characteristics. Series
Material
Number
Size – h/d (mm)
Load angle – a
Crack system
Test setup
01 02 03 04 05 06 07 08 09
LVL LVL Timber Timber Timber LVL LVL LVL LVL
4 10 4 10 4 10 4 10 4 10 20 20 20 20
80/20 80/20 80/20 80/20 80/20 100/63 63/63 63/63 80/63
0°/30°/60°/90° 0°/30°/60°/90° 0°/30°/60°/90° 0°/30°/60°/90° 0°/30°/60°/90° 0° 0° 0° 90°
RL TL RL TL NL TL TL RL TL
CTS-test CTS-test CTS-test CTS-test CTS-test CSS-test CSS-test CSS-test SENB-test
4
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Fig. 3. Sketch of test setup and sizes for fracture mode I for the SENB-specimen.
Fig. 4. Test setup for fracture mode I.
Fig. 5. Sketch of test setup and sizes for fracture mode II.
Fig. 6. Test setup for fracture mode II.
mode I). The beam is simply supported and loaded in a three point bending test as shown in Fig. 4. The displacement is controlled and the load measured. For the pure fracture mode II, a variation of the compact shear specimen (CSS) from Jones and Chisholm [15] or Prokopski [16] was used. The specimens were precut with two notches symmetrically to the center, as shown in Figs. 5 and 6. The
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
5
Fig. 7. Sketch and sizes of CTS specimen for mixed mode test.
Fig. 8. CTS specimen test setup for pure mode I.
specimen is loaded in a displacement controlled regime and the load is applied on top of the specimen over the full thickness. The loading and support situation of the specimen leads to shear of the middle part. The compact tension shear test specimen (CTS) was introduced by Richard and Benitz [17] as a simple method for the determination of the fracture toughness under pure failure mode I and mode II and also for different mixed mode failures. The test setup developed can be used to investigate superimposed normal and shear stress situations in planar specimens. While rotating the specimen with regards to the load direction, different mode I to mode II ratios can be investigated, as shown in Figs. 7 and 8. A load angle a of 90° leads to pure mode I failure where a of 0° leads to pure mode II failure. The specimen is notched at the center and loaded in a displacement controlled regime. For the determination of the fracture energy of wood, relative stable load displacement curves are required. Therefore all tests were done displacement controlled with a constant rate of 0.5 mm per second. 3. Theory and calculations The fracture mechanic is a method for the description of the material behavior influenced by cracks or defects [6]. For the characterization of the failure process, two approaches are used: the strain energy release rate which is based on the global energy balance of Griffith; and the stress intensity factor (SIF) introduced by Irwin which is based on the local stress distribution around a crack tip [6]. Furthermore one distinguishes between the linear elastic fracture mechanics (LEFM) and the nonlinear fracture mechanics (NLFM). The LEFM requires that material describes a linear elastic material behavior up to the point of rupture. Therefore the NLFM respects plastic deformations at the crack tip or fiber bridges which leads to a nonlinear material behavior before failure. In most cases for wooden materials, the LEFM can be applied if the effects mentioned are small [6]. As stated in [6], the stress intensity factor uses the local surrounding stress field at the crack tip for the assessment of the crack initiation or propagation. The SIF K is a function of the specimen geometry, load applied and crack length. The crack
6
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
grows unstable if the critical value of the material, often referred to as the fracture toughness Kc, is reached. The fracture toughness Kc is considered as a material property which defines the material resistance against the crack growth. The energy balance method for determining the strain energy release rate uses the energy equilibrium of the complete system for the determination of the energy needed for the cracking process. The energy released for a crack growth is the so called fracture energy G. The critical energy release rate Gc generally describes the state where the stable crack growth becomes unstable and the system fails. For linear elastic materials, the energy release rate Gf can be expressed with the SIF’s, the specific modulus of elasticity E0 and the shear modulus G, as given in Eq. (1) [18].
Gf ¼
1 I2 1 III2 II2 þ K 0 K þK 2G E
ð1Þ
where KI, KII and KIII are the SIF’s for the fracture mode I, II and III respectively. The fracture energy can be determined from the load displacement curve observed during the complete separation of the cracking area of the test specimen under a displacement controlled loading. The integration of the load displacement curve divided by the cracking area results in the fracture energy Gf (representing an average fracture energy), and is given in Eq. (2), compare [6].
Gf ¼
1 ad
Z
u0
F du
ð2Þ
0
where a is the crack length and d the specimen thickness, representing the fracture area. For the analysis and the determination of the fracture energy for the mixed mode test series, the load was split into the two components, i.e. the transverse tension and shear forces. These components for the pure fracture mode I and mode II are calculated with these values respectively. The SIF K is defined by the stress r, the crack length a and the form function of the geometry Y, as given in Eq. (3), [19].
K ¼r
pffiffiffiffiffiffi pa Y
ð3Þ
The fracture toughness Kc is then the corresponding SIF at failure of the specimen with the maximum load Fc and the corresponding crack length ac. For the CTS specimen, the resulting equations for the fracture toughness for modes I and II are given in Eqs. (4) and (5) respectively, as given by Binner [20], where a is the loading angle of the system as shown in Fig. 7, h the depth and d the thickness of the specimen. The CTS specimen was originally developed for isotropic material, but can also be used without any significant alteration for wood or for other orthotropic materials, as shown in [13].
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi a pffiffiffiffiffiffi u 0:26 þ 2:65 ha pa cos a u t K ¼ a a 2 hd 1 ha 1 þ 0:55 ha 0:08 ha
ð4Þ
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi a pffiffiffiffiffiffi u 0:23 þ 1:40 pa sin a u ha t K ¼ a a 2 hd 1 ha 1 0:67 ha þ 2:08 ha
ð5Þ
I
II
F
F
4. Results 4.1. Observed behavior at macro scale Engineered wood products are mainly developed to improve the general material parameters such as bending strength and or stiffness. What is unknown for these products is their fracture failure behavior under transverse tension or shear. In general, during the tests, all mode II specimens failed in a sudden brittle manner, whereas for the fracture mode I specimens, the crack grew in a stable manner. The visible difference after the failure between Radiata Pine timber and LVL are small, since both materials show even crack surfaces in the RL-crack system and uneven or stepwise crack surfaces in the TLcrack system, as shown in Figs. 9 and 10. For the failure mode I, with a load angle a of 90° degrees, small fiber bridges could be observed in the TL crack system and more often for LVL than in timber. Furthermore, for the RL-crack system, the crack surface lies between the annual growth rings for timber or within one veneer of LVL. By decreasing the load angle a for the fracture mode II, the crack surfaces increase due to uneven cracks for both materials. In addition, for LVL, the crack jumps between the different veneers, which leads to an interlocked crack surface. 4.2. Mechanical failure behavior The load displacement curves for all loading cases and for both crack systems for Radiata Pine timber and LVL are shown in Figs. 11 and 12. To facilitate the comparison, the load displacement curves are normalized with the assumed crack area. In
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
7
Fig. 9. Crack surfaces of Radiata Pine timber for different crack systems in failure mode I (top row) and mode II (bottom row).
Fig. 10. Crack surfaces of Radiata Pine LVL for different load angle a in RL-crack system (top row) and TL-crack system (bottom row).
general, the load capacities increase with decreasing the loading angle a, which confirms the known differences between the strength for tension perpendicular to grain and shear. Furthermore, the crack propagation during the failure changes from a relatively stable one to a more unstable crack growth, as observed from the comparison of the displacement u needed for the complete separation of the crack area. The material behavior and the corresponding curves show a linear part before failure, thus the LEFM can be used. The load capacities are on average in the same range for both materials but there are differences in the stiffness between Radiata Pine timber and LVL. For the RL crack system, timber is stiffer than LVL. This is maybe caused by the fabrication of the veneers for LVL. With the peeling of the veneers from the log, a pre micro cracking between the single fibers could occur. Later on, these small pre cracks between the fibers can influence the failure behavior in specific stress situations. For timber, due to the natural growing process, the fibers are in a compact composite structure. Regardless of defects such as knots, there are no pre cracks which can lead to a higher resistance against the applied stress situations but also to a more sudden brittle failure. The layered cross section of LVL in the TL-crack system is a distinct advantage. More energy is needed for the creation of the interlocked crack surface observed, which thus leads to a more ductile failure behavior. Small, not intended, relative derivations between the grain directions of the different veneers might also contribute to this effect, but it was not specifically investigated. 4.3. Fracture parameters The mechanical failure behavior described above is also reflected in the fracture parameters values. The fracture energy and fracture toughness are calculated for the complete test series. The values are summarized in Table 2 and classified based on the test setup and the crack system. The number of tests, used to calculate the results shown in Table 2, is sometimes less the total number of tested specimen as given in Table 1, because in some cases other failure occurred and these test results were not used. The first value is for the fracture mode I and the second for the fracture mode II. For each of the fracture parameters, the coefficient of variation (COV) is also given. Larsen and Gustafsson [5] obtained a linear relationship between the fracture energy of wood in tension perpendicular to grain and density for all test results observed from a joint testing project. The testing project covered a variety of timber species, e.g. Redwood, Pinus silvestris and Picea Abies from different countries. The fracture energies for mode I for Radiata Pine timber show no clear dependency on the density as observed for solid wood by Larsen and Gustafsson [5], and shown in Fig. 13. The diagram shows the values based on the crack systems. The fracture energies of the RL and TL crack systems
8
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Fig. 11. Load displacement curves for CTS specimen, TL crack system.
are not statistically different. For the NL crack system, the density and fracture energies are higher compared to the other ones, but this is also influenced by the change of the crack system itself. The approach by Larsen and Gustafsson [5] for mode I for the LVL specimen is in agreement with the lower limit of the test results. For the test series in LVL, the density is in a close range and the values show only differences between the two crack systems RL and TL, as shown in Figs. 14 and 15. For the RL crack system, the average fracture energy for mode I and mode II is lower than the one of the TL crack system, which confirms the mechanical failure behavior observed. The orientation of the layered cross section with respect to the stress situation and the loading angle a has to be considered in construction details. 5. Discussion Different test setups could be used for the determination of the fracture parameters presented. The CTS-test setup should be used carefully because small inaccuracy can lead to unsymmetrical loading and thus the stress situation described is not
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
9
Fig. 12. Load displacement curves for CTS specimen, RL crack system.
the case suspected. The preparation of the test specimen is more complex, but this test setup is the only one which gives the possibility to observe the mixed mode failure between modes I and II. The other test setups are valid for pure fracture modes only. There are only minor differences in the average values for the fracture mode I parameters using different test setups, as shown in Table 2. For the fracture mode II, the differences between the test setups CTS and CSS are higher. The fracture mode II requires due to the higher strength in shear than in tension perpendicular to grain of wood a stiffer test setup to get a stable crack growth for the determination of the fracture energy. The test results reached with the CTS series are clearly reduced and limit the direct comparison. The failure behavior of Radiata Pine LVL and timber can be further summarized using the fracture toughness observed within the common failure criteria. The failure criterion describes the relation between the failure modes I and II and the mixed mode failure. The linear or quadratic failure criterion are common ones, the failure criterion from Wu [21] was determined on Balsa wood and is often used or compared with wood species or wood products. Fig. 16 shows the values for Radiata Pine LVL and timber for the two crack systems RL and TL. The parameters observed are not enough to classify the failure behavior of Radiata Pine LVL and timber using one failure criterion. The concentration of the results to the pure fracture
10
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Table 2 Fracture toughness and fracture energy results. Speci men
Test setup
Crack system
Number of tests (–)
Density (kg/m3)
Fracture toughness Kc
Fracture energy Gf Mode I
Mode II
Mode I
Mode II
Nm/m2
COV (%)
Nm/m2
COV (%)
kN/m3/2
COV (%)
kN/m3/2
COV (%)
Timber
CTS
RL TL NL
8/6 10/6 10/4
476 424 556
485 623 1219
28.5 17.0 18.4
6962 8087 12,722
22.6 25.9 8.3
239 147 296
19.4 19.8 10.5
829 786 1020
15.9 4.5 16.1
LVL
CTS
RL TL RL TL TL
8/3 9/5 –/20 –/40 19/–
595 596 602 607 613
673 1213 – – 1192
19.8 26.6 – – 25.4
4009 7607 3284 5192 –
17.3 24.4 29.6 26.1 –
127 170 – – –
29.0 12.9 – – –
385 528 316 402 –
17.7 28.5 21.0 21.2 –
CSS SENB
Fig. 13. Fracture energies for mode I in Radiata Pine timber.
Fig. 14. Fracture energies for mode I in Radiata Pine LVL.
mode I confirms only the mechanical behavior observed and the known difference between the tension strength perpendicular to grain and the shear strength. For a final analysis, a finer variation of the loading angle a in the CTS test setup should be investigated. The failure behavior in Radiata Pine LVL and timber shows a high crack resistance in both fracture modes I and II in comparison to the one for European wood species. For example, the fracture energy for Radiata Pine timber is approximately double the one for mode I or five times for mode II compared to the average values of the fracture energy of European spruce or pine or Canadian spruce found by e.g. Aicher et al. [3,22], Boström [23], Daudeville [24], Franke [2], Kretschmann [25], Riberholt et al. [26], Schatz [27], Smith et al. [6], Vasic [7]. This high crack resistance could be partly transferred to LVL from Radiata Pine timber. The fracture energies observed for LVL in the fracture mode I are higher for both crack systems RL and TL. However, the differences of the fracture energies for the fracture mode II are in opposite direction. Radiata Pine timber clearly shows a higher crack resistance than LVL.
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
11
Fig. 15. Fracture energies for mode II in Radiata Pine LVL.
Fig. 16. Failure behavior of Radiata Pine LVL and timber compared to the common failure criteria.
When compared to the results by King et al. [1], the observed fracture toughness for mode I and II of Radiata Pine timber shows similar differences between the crack systems RL and TL but the values obtained are lower. LVL results show a lower crack resistance when compared to other wood products such as particle board (FPY), medium density fiberboard (MDF) or oriented strand board (OSB) as observed by Niemz et al. [9,10]. Since the manufacturing of particle boards involves a large quality of glue and results in fibers that are loaded in all possible directions, fracture energies of more than twice as the LVL ones are possible.
6. Conclusions A comprehensive test program to determine the fracture parameters of Radiata Pine LVL and timber was initiated. Characterization of the different test setups available for wood was possible and permitted the observation and analysis of the fracture failure behavior. Comparison of the various test setups shows no differences in the fracture parameters determined, only the preparation of the test specimens and the handling during the experiment is different. The failure behavior of Radiata Pine LVL shows a more ductile behavior in all fracture modes considered in comparison to timber. This leads to higher values of the fracture energy for mode I for LVL than for timber. On the other hand, the energy which was needed for the separation of the crack surface for mode II is smaller for LVL than for timber. The comparison of the mechanical behavior also shows influences of the orientation and the production method for LVL on the crack resistance. For both LVL and timber, smaller load capacities were observed in the RL crack system in comparison to the ones for the TL crack system. Radiata Pine timber shows on average a stiffer material behavior, but therefore leads almost to a quite sudden brittle failure. In structural details such as connections loaded perpendicular to grain or beams with holes or endnotched beams in complex constructions, the different failure behavior of Radiata Pine LVL and timber should be analyzed carefully and the orientation of the layers for LVL has to be considered for the design. The strength behavior values of timber and LVL depends strongly on the grain direction as well as the fracture behavior. In novel design approaches for the characterization of
12
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
notches [28,29], or connections with mechanical fasteners, [30] the fracture mechanic was applied. For their application the fracture material parameters are needed. The research results presented are useful in numerical simulation of complex wooden structures or in the development of new design approaches. The material parameters provided are important for a comprehensive material description of wood or wood products used in numerical simulations, such as in the splitting failures of connections as shown in [31,32]. Acknowledgments The research work was generously supported by the Structural Timber Innovation Company (STIC) from New Zealand. The students Armithab Athanari and Joseph Pearson worked on the specimen preparation and the experimental tests during a student project as well as Dmitry Volynkin as a summer student. References [1] King MJ, Sutherland IJ, Le-Ngoc L. Fracture toughness of wet and dry Pinus Radiata. Holz als Roh- und Werkstoff 1999;57:235–40. [2] Franke B. Zur Bewertung der Tragfähigkeit von Trägerausklinkungen in Nadelholz. Doctoral dissertation, Germany: Bauhaus University Weimar; 2008. [3] Aicher S, Boström L, Gierl M, Kretschmann D, Valentin G. Determination of fracture energy of wood in mode II. Rilem TC 133 Report. Sweden: Swedish National Testing and Research Institute; 1997. [4] Valentin HG, Boström L, Gustafsson PJ, Ranta Maunus A, Gowda S. Application of fracture mechanics to timber structures RILEM state-of-the-art report. Research Notes 1262. Finland: Technical Research Centre of Finland; 1991. [5] Larsen HJ, Gustafsson PJ. The fracture energy of wood in tension perpendicular to grain, results from a joint testing project. In: Conference proceedings CIB-W18A, paper 23–19-2, Portugal; 1990. [6] Smith I, Landis E, Gong M. Fracture and fatigue in wood. England: John Wiley & Sons Ltd.; 2003. [7] Vasic S. Applications of fracture mechanics to wood. Doctoral dissertation, Canada: The University of New-Brunswick; 2000. [8] Stanzl-Tschegg SE, Navi P. Fracture behaviour of wood and its composites. A review, Holzforschung 2009;63:139–49. [9] Niemz P, Diener M. Vergleichende Untersuchungen zur Ermittlung der Bruchzähigkeit an Holzwerkstoffen. Holz als Roh- und Werkstoff 1999;57:222–5. [10] Niemz P, Diener M, Pöhler E. Untersuchungen zur Ermittlung der Bruchzähigkeit an MDF-Platten. Holz als Roh- und Werkstoff 1997;55:327–30. [11] Lo KW, Gong YB, Tamilselvan T, Lai MO. A proposed specimen for KIIc testing. Int J Fract 2003;124:127–37. [12] Xu S, Reinhardt HW, Gappoev M. Mode II fracture testing method for highly orthotropic materials like wood. Int J Fract 1996;75:185–214. [13] Valentin HG, Caumes P. Crack propagation in mixed mode in wood: a new specimen. Wood Sci Technol 1989;23:43–53. [14] Cramer SM, Pugel AD. Compact shear specimen for wood mode II fracture investigations. Int J Fract 1987;35:163–74. [15] Jones DL, Chisholm DB. An investigation of the edge-sliding mode in fracture mechanics. Engng Fract Mech 1975;7:261–70. [16] Prokopski G. Investigation of wood fracture toughness using mode II fracture (shearing). J Mater Sci 1995;30:4745–50. [17] Richard HA, Benitz K. A loading device for the creation of mixed mode in fracture mechanics. J Fract Mech 1983;22:55–8. [18] Gross D. Bruchmechanik. Berlin: Springer Verlag; 1996. [19] Tan DM, Stanzl-Tschegg SE, Tschegg EK. Models of wood fracture in mode I and mode II. Holz als Roh- und Werkstoff 1995;53:159–64. [20] Binner SB. Fatigue crack growth studies under mixed mode loading. Int J Fatigue 2001;23:259–63. [21] Wu EM. Application of fracture mechanics to anisotropic plates. J Appl Mech 1967:974. [22] Aicher S. Bruchenergien, kritische Energiefreisetzungsraten und Bruchzähigkeit von Fichte bei Zugbeanspruchung senkrecht zur Faserrichtung. Holz als Roh- und Werkstoff 1994;51:361–70. [23] Boström L. Method for determination of the softening behaviour of wood and the applicability of nonlinear fracture mechanics model. Doctoral dissertation, Sweden: Lund; 1992. [24] Daudeville L. Fracture in spruce: experiment and numerical analysis by linear and non linear fracture mechanics. Holz als Roh- und Werkstoff 1999;57:425–32. [25] Kretschmann DE. Effect of varying dimensions on tapered end-notched flexure shear specimen. Wood Sci Technol 1995;29:287–93. [26] Riberholt H, Enquist B, Gustafsson PJ, Jensen RB. Timber beams notched at the support. Afdelingen for Bærende Konstruktioner, Lungby: Danmarks Tekniske Højskole, R280; 1991. [27] Schatz T. Zur bruchmechanischen Modellierung des Kurzeit-Bruchverhaltens von Holz im Rissöffnungsmodus I. Doctoral dissertation, Germany: University of Stuttgart; 1994. [28] Gustafsson PJ. A study of strength notched beams. In: Conference proceedings CIB-W18A, paper 21–10-1, Canada; 1988. [29] Jockwer R, Steiger R, Frangi A, Köhler J. Impact of material properties on the fracture mechanics design approach for notched beams in Eurocode 5. In: Conference proceedings CIB-W18A, paper 44–6-1, Italy; 2011. [30] Franke B, Quenneville P. Design approach for the splitting failure of dowel-type connections loaded perpendicular to grain. In: Conference proceedings CIB-W18A, paper 44–7-5, Italy; 2011. [31] Franke B, Quenneville P. Numerical modeling of the failure behavior of dowel connections in wood. ASCE, J Engng Mech 2010;137:186–95. [32] Franke B, Franke S, Quenneville P. Numerical modeling and analysis of the failure behaviour of dowel-type connections in wood. In: European congress on computational methods in applied sciences and engineering (ECCOMAS 2012), Austria; 2012.
Contents lists available at ScienceDirect
Engineering Fracture Mechanics journal homepage: www.elsevier.com/locate/engfracmech
Analysis of the fracture behavior of Radiata Pine timber and Laminated Veneer Lumber Bettina Franke a,⇑, Pierre Quenneville b a b
Bern University of Applied Sciences, Department of Architecture, Wood and Civil Engineering, Solothurnstrasse 102, PO Box 6096, 2500 Biel/Bienne 6, Switzerland The University of Auckland, Department of Civil and Environmental Engineering, Private Bag 92019, Auckland 1142, New Zealand
a r t i c l e
i n f o
Article history: Received 20 April 2012 Received in revised form 7 July 2013 Accepted 8 December 2013
Keywords: Fracture mechanics Mode I Mode II Timber Radiata Pine Laminated Veneer Lumber
a b s t r a c t The investigation of the failure mechanism of the fracture mode I and II and the mixed mode of New Zealand Radiata Pine timber and of Radiata Pine Laminated Veneer Lumber is presented. Different test setups were compared and used for the determination of the material parameters fracture toughness and fracture energy. Furthermore, the differences, advantages and disadvantages of Radiata Pine Laminated Veneer Lumber versus solid wood are presented. The investigations show that Radiata Pine Laminated Veneer Lumber is more ductile than Radiata Pine timber. Ó 2013 Elsevier Ltd. All rights reserved.
1. Background Typical timber constructions consist of members and also joints, notches or holes which can lead to high stress singularities in the structure. Fracture mechanic gives the possibility to characterize situations with stress singularities which can lead to a failure including crack initiations and crack propagations. For the prediction of failures or load capacities, the fracture mechanics methods are used in new design proposals. However, important material parameters such as the critical fracture energy or the fracture toughness have to be known. International research results are published for solid wood e.g. in Radiata Pine by King et al. [1], in European Spruce e.g. by Franke [2], Aicher et al. [3], Valentin et al. [4], Larsen and Gustafsson [5] or in Canadian Spruce by Smith et al. [6] or Vasic [7]. For engineered wood products, only a portion of the values are published, e.g. Stanzl-Tschegg and Navi [8], Niemz et al. [9,10]. The trend to large scale and multistory timber constructions requires high performance wood products with increased strengths and stiffness. In order to allow the analysis of failures in structures made of Radiata Pine timber (RP) or Laminated Veneer Lumber (LVL), a study to determine the fracture energy and fracture toughness was initiated. The research results presented include the investigation of the failure behavior of New Zealand Radiata Pine LVL compared to timber. The difference, possible advantages or maybe disadvantages of LVL versus timber are investigated and compared with published international experimental test results. When investigating the failure behavior, different test setups are available to study the different fracture modes. The fracture mechanic method distinguishes three different failure modes, as shown in Fig. 1. They are: failure mode I – cracking under transverse tension stress in relation to the crack plane, mode II – cracking under in plane shear stress and mode III – the failure under out of plane shear stress. The so-called mixed mode failure is a combination of the mode I and mode II. The fracture modes I and II and for some cases, the mixed mode failure were investigated in experimental test series. None ⇑ Corresponding author. Tel.: +41 32 344 0308. E-mail addresses: [email protected] (B. Franke), [email protected] (P. Quenneville). 0013-7944/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.engfracmech.2013.12.004
2
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Nomenclature 0
E F Fc G Gf Gc K Kc Y a ac d h u
a b
r q I II III
specific modulus of elasticity (MPa) load (N) critical load (N) shear modulus (MPa) energy release rate (Nm/m2) critical fracture energy (Nm/m2) stress intensity factor (SIF) (kN/m3/2) fracture toughness (kN/m3/2) form function of the geometry (–) crack length (mm) corresponding crack length at critical load (mm) thickness of the specimen (mm) depth of the specimen (mm) displacement (mm) loading angle of the system (°) annual growth ring (°) stress (MPa) density (kg/m3) index for fracture mode I (–) index for fracture mode II (–) index for fracture mode III (–)
of the test setups available for the different failure modes are approved or standardized for wood. Therefore different methods were used and are compared for the same stress situation. 2. Material and method 2.1. Material and specimens New Zealand Radiata Pine timber and LVL were used for the analysis of the failure behavior of solid wood and wood product respectively. New Zealand Radiata Pine is a relatively fast growing plantation pine species which is fully grown after approximately 20–30 years. The timber used grew in a relatively constant climate on the North Island of New Zealand. The growth ring width was 10 mm on average and the specimens were without major defects. The timber was conditioned in an environment with a temperature of 20° Celsius and a relative humidity of 65%. At testing, the average moisture content from the test specimens was 11.4% and the mean density was 486 kg/m3. LVL is a wood product manufactured from rotary peeled veneer layers glued together to form slabs. The veneers are 3–4 mm thick and in general laid up with parallel to grain direction. Natural defects such as knots are more homogenously distributed, which lead to higher strength and less variation in comparison to timber. Even if there are small derivations between the grain directions of the veneers due to the practical production, which might influences the properties and e.g. the response in the post peak-resistance behavior; LVL shows no behavior related to cross-laminations and the influences are respected in the variations of the properties. LVL products are used in houses and large span buildings as structural members. The hyspan90 product from the Carter Holt Harvey company from the North Island of New Zealand was used. The LVL specimens were also stored in the conditioned environment chamber. The average moisture content of the LVL test specimens was 8.9% and the mean density was 600 kg/m3. For the consideration and investigation of the influence of the anisotropic material behavior of solid wood on the fracture energy in the different material directions, three different crack systems, RL, TL and NL were tested, as shown in Fig. 2. The
Fig. 1. Different test setups for the fracture mode I, II and III and the mixed mode.
3
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Fig. 2. Specimens crack systems and orientations.
crack systems are described by the natural material axes, radial R, tangential T and longitudinal L. The letter N defines the growth ring orientation of about 45°, meaning an orientation between the tangential and radial orientation. The first letter specifies the direction normal to the crack plane and the second letter the crack growth direction. Due to the constitution of the wood product LVL, only the two crack system RL and TL were considered, as shown in Fig. 2. Furthermore for each crack system, the fracture mode I and mode II as well as the mixed mode of mode I and mode II were investigated within the test series. The fracture modes can be also specified by the load angle a, which is the angle between the load applied and the crack plane defined. A load angle a of 90°, resulting in transverse stress, leads to pure mode I failure where a of 0°, resulting in shear stress, leads to pure mode II failure. Load angles a between 0° and 90° result in a mixed mode failure of mode I and II. An overview of the test groups, the cracking systems as well as the number of each test configuration is given in Table 1. The sizes of the specimen are defined by the depth of the member h and the thickness d at the crack plane. 2.2. Test setups For the investigation of the failure behavior of wood under tension perpendicular to grain as well as shear and for the determination of fracture parameters, different test setups are available and accepted as shown in e.g. Lo et al. [11], Aicher et al. [3], Xu et al. [12], Valentin et al. [4], Valentin and Caumes [13] and Cramer and Pugel [14]. The first published test setups for wood were adopted from test setups developed for isotropic material such as steel. These test methods and associated equations were modified to make them valid for wood which can be, at best, assumed as an orthotropic material. Three test setups published were chosen for the investigation of Radiata Pine timber and LVL; for the pure fracture mode I, the single end notched beam specimen (SENB) [5], and for the pure fracture mode II, the compact shear specimen (CSS) [15,16]. The compact tension shear test specimen (CTS) [13,17], was used for the investigation of the mixed mode of the fracture mode I and II. This test setup allows the investigation of the pure fracture modes I and II as well as the mixed mode using the same specimen geometry and load application. Therefore the size effect and the effect due to loading procedure is neglected in the comparison of the fracture behavior and energies. For the fracture mode I, the single end notched beam (SENB) specimen was used according to the draft standard from Larson and Gustafsson [5]. This test setup was agreed by the CIB-W18 working group for a joint project for the determination of the fracture energy and the evaluation of the test setup. The test setup is established for wood and for the fracture mode I. The tested specimen is glued to two wooden beams, as shown in Fig. 3. The grain direction of the specimen is perpendicular to the span of the beam. The orientation and the pre cut notch results in tension stresses normal to the crack plane (failure
Table 1 Test program characteristics. Series
Material
Number
Size – h/d (mm)
Load angle – a
Crack system
Test setup
01 02 03 04 05 06 07 08 09
LVL LVL Timber Timber Timber LVL LVL LVL LVL
4 10 4 10 4 10 4 10 4 10 20 20 20 20
80/20 80/20 80/20 80/20 80/20 100/63 63/63 63/63 80/63
0°/30°/60°/90° 0°/30°/60°/90° 0°/30°/60°/90° 0°/30°/60°/90° 0°/30°/60°/90° 0° 0° 0° 90°
RL TL RL TL NL TL TL RL TL
CTS-test CTS-test CTS-test CTS-test CTS-test CSS-test CSS-test CSS-test SENB-test
4
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Fig. 3. Sketch of test setup and sizes for fracture mode I for the SENB-specimen.
Fig. 4. Test setup for fracture mode I.
Fig. 5. Sketch of test setup and sizes for fracture mode II.
Fig. 6. Test setup for fracture mode II.
mode I). The beam is simply supported and loaded in a three point bending test as shown in Fig. 4. The displacement is controlled and the load measured. For the pure fracture mode II, a variation of the compact shear specimen (CSS) from Jones and Chisholm [15] or Prokopski [16] was used. The specimens were precut with two notches symmetrically to the center, as shown in Figs. 5 and 6. The
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
5
Fig. 7. Sketch and sizes of CTS specimen for mixed mode test.
Fig. 8. CTS specimen test setup for pure mode I.
specimen is loaded in a displacement controlled regime and the load is applied on top of the specimen over the full thickness. The loading and support situation of the specimen leads to shear of the middle part. The compact tension shear test specimen (CTS) was introduced by Richard and Benitz [17] as a simple method for the determination of the fracture toughness under pure failure mode I and mode II and also for different mixed mode failures. The test setup developed can be used to investigate superimposed normal and shear stress situations in planar specimens. While rotating the specimen with regards to the load direction, different mode I to mode II ratios can be investigated, as shown in Figs. 7 and 8. A load angle a of 90° leads to pure mode I failure where a of 0° leads to pure mode II failure. The specimen is notched at the center and loaded in a displacement controlled regime. For the determination of the fracture energy of wood, relative stable load displacement curves are required. Therefore all tests were done displacement controlled with a constant rate of 0.5 mm per second. 3. Theory and calculations The fracture mechanic is a method for the description of the material behavior influenced by cracks or defects [6]. For the characterization of the failure process, two approaches are used: the strain energy release rate which is based on the global energy balance of Griffith; and the stress intensity factor (SIF) introduced by Irwin which is based on the local stress distribution around a crack tip [6]. Furthermore one distinguishes between the linear elastic fracture mechanics (LEFM) and the nonlinear fracture mechanics (NLFM). The LEFM requires that material describes a linear elastic material behavior up to the point of rupture. Therefore the NLFM respects plastic deformations at the crack tip or fiber bridges which leads to a nonlinear material behavior before failure. In most cases for wooden materials, the LEFM can be applied if the effects mentioned are small [6]. As stated in [6], the stress intensity factor uses the local surrounding stress field at the crack tip for the assessment of the crack initiation or propagation. The SIF K is a function of the specimen geometry, load applied and crack length. The crack
6
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
grows unstable if the critical value of the material, often referred to as the fracture toughness Kc, is reached. The fracture toughness Kc is considered as a material property which defines the material resistance against the crack growth. The energy balance method for determining the strain energy release rate uses the energy equilibrium of the complete system for the determination of the energy needed for the cracking process. The energy released for a crack growth is the so called fracture energy G. The critical energy release rate Gc generally describes the state where the stable crack growth becomes unstable and the system fails. For linear elastic materials, the energy release rate Gf can be expressed with the SIF’s, the specific modulus of elasticity E0 and the shear modulus G, as given in Eq. (1) [18].
Gf ¼
1 I2 1 III2 II2 þ K 0 K þK 2G E
ð1Þ
where KI, KII and KIII are the SIF’s for the fracture mode I, II and III respectively. The fracture energy can be determined from the load displacement curve observed during the complete separation of the cracking area of the test specimen under a displacement controlled loading. The integration of the load displacement curve divided by the cracking area results in the fracture energy Gf (representing an average fracture energy), and is given in Eq. (2), compare [6].
Gf ¼
1 ad
Z
u0
F du
ð2Þ
0
where a is the crack length and d the specimen thickness, representing the fracture area. For the analysis and the determination of the fracture energy for the mixed mode test series, the load was split into the two components, i.e. the transverse tension and shear forces. These components for the pure fracture mode I and mode II are calculated with these values respectively. The SIF K is defined by the stress r, the crack length a and the form function of the geometry Y, as given in Eq. (3), [19].
K ¼r
pffiffiffiffiffiffi pa Y
ð3Þ
The fracture toughness Kc is then the corresponding SIF at failure of the specimen with the maximum load Fc and the corresponding crack length ac. For the CTS specimen, the resulting equations for the fracture toughness for modes I and II are given in Eqs. (4) and (5) respectively, as given by Binner [20], where a is the loading angle of the system as shown in Fig. 7, h the depth and d the thickness of the specimen. The CTS specimen was originally developed for isotropic material, but can also be used without any significant alteration for wood or for other orthotropic materials, as shown in [13].
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi a pffiffiffiffiffiffi u 0:26 þ 2:65 ha pa cos a u t K ¼ a a 2 hd 1 ha 1 þ 0:55 ha 0:08 ha
ð4Þ
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi a pffiffiffiffiffiffi u 0:23 þ 1:40 pa sin a u ha t K ¼ a a 2 hd 1 ha 1 0:67 ha þ 2:08 ha
ð5Þ
I
II
F
F
4. Results 4.1. Observed behavior at macro scale Engineered wood products are mainly developed to improve the general material parameters such as bending strength and or stiffness. What is unknown for these products is their fracture failure behavior under transverse tension or shear. In general, during the tests, all mode II specimens failed in a sudden brittle manner, whereas for the fracture mode I specimens, the crack grew in a stable manner. The visible difference after the failure between Radiata Pine timber and LVL are small, since both materials show even crack surfaces in the RL-crack system and uneven or stepwise crack surfaces in the TLcrack system, as shown in Figs. 9 and 10. For the failure mode I, with a load angle a of 90° degrees, small fiber bridges could be observed in the TL crack system and more often for LVL than in timber. Furthermore, for the RL-crack system, the crack surface lies between the annual growth rings for timber or within one veneer of LVL. By decreasing the load angle a for the fracture mode II, the crack surfaces increase due to uneven cracks for both materials. In addition, for LVL, the crack jumps between the different veneers, which leads to an interlocked crack surface. 4.2. Mechanical failure behavior The load displacement curves for all loading cases and for both crack systems for Radiata Pine timber and LVL are shown in Figs. 11 and 12. To facilitate the comparison, the load displacement curves are normalized with the assumed crack area. In
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
7
Fig. 9. Crack surfaces of Radiata Pine timber for different crack systems in failure mode I (top row) and mode II (bottom row).
Fig. 10. Crack surfaces of Radiata Pine LVL for different load angle a in RL-crack system (top row) and TL-crack system (bottom row).
general, the load capacities increase with decreasing the loading angle a, which confirms the known differences between the strength for tension perpendicular to grain and shear. Furthermore, the crack propagation during the failure changes from a relatively stable one to a more unstable crack growth, as observed from the comparison of the displacement u needed for the complete separation of the crack area. The material behavior and the corresponding curves show a linear part before failure, thus the LEFM can be used. The load capacities are on average in the same range for both materials but there are differences in the stiffness between Radiata Pine timber and LVL. For the RL crack system, timber is stiffer than LVL. This is maybe caused by the fabrication of the veneers for LVL. With the peeling of the veneers from the log, a pre micro cracking between the single fibers could occur. Later on, these small pre cracks between the fibers can influence the failure behavior in specific stress situations. For timber, due to the natural growing process, the fibers are in a compact composite structure. Regardless of defects such as knots, there are no pre cracks which can lead to a higher resistance against the applied stress situations but also to a more sudden brittle failure. The layered cross section of LVL in the TL-crack system is a distinct advantage. More energy is needed for the creation of the interlocked crack surface observed, which thus leads to a more ductile failure behavior. Small, not intended, relative derivations between the grain directions of the different veneers might also contribute to this effect, but it was not specifically investigated. 4.3. Fracture parameters The mechanical failure behavior described above is also reflected in the fracture parameters values. The fracture energy and fracture toughness are calculated for the complete test series. The values are summarized in Table 2 and classified based on the test setup and the crack system. The number of tests, used to calculate the results shown in Table 2, is sometimes less the total number of tested specimen as given in Table 1, because in some cases other failure occurred and these test results were not used. The first value is for the fracture mode I and the second for the fracture mode II. For each of the fracture parameters, the coefficient of variation (COV) is also given. Larsen and Gustafsson [5] obtained a linear relationship between the fracture energy of wood in tension perpendicular to grain and density for all test results observed from a joint testing project. The testing project covered a variety of timber species, e.g. Redwood, Pinus silvestris and Picea Abies from different countries. The fracture energies for mode I for Radiata Pine timber show no clear dependency on the density as observed for solid wood by Larsen and Gustafsson [5], and shown in Fig. 13. The diagram shows the values based on the crack systems. The fracture energies of the RL and TL crack systems
8
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Fig. 11. Load displacement curves for CTS specimen, TL crack system.
are not statistically different. For the NL crack system, the density and fracture energies are higher compared to the other ones, but this is also influenced by the change of the crack system itself. The approach by Larsen and Gustafsson [5] for mode I for the LVL specimen is in agreement with the lower limit of the test results. For the test series in LVL, the density is in a close range and the values show only differences between the two crack systems RL and TL, as shown in Figs. 14 and 15. For the RL crack system, the average fracture energy for mode I and mode II is lower than the one of the TL crack system, which confirms the mechanical failure behavior observed. The orientation of the layered cross section with respect to the stress situation and the loading angle a has to be considered in construction details. 5. Discussion Different test setups could be used for the determination of the fracture parameters presented. The CTS-test setup should be used carefully because small inaccuracy can lead to unsymmetrical loading and thus the stress situation described is not
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
9
Fig. 12. Load displacement curves for CTS specimen, RL crack system.
the case suspected. The preparation of the test specimen is more complex, but this test setup is the only one which gives the possibility to observe the mixed mode failure between modes I and II. The other test setups are valid for pure fracture modes only. There are only minor differences in the average values for the fracture mode I parameters using different test setups, as shown in Table 2. For the fracture mode II, the differences between the test setups CTS and CSS are higher. The fracture mode II requires due to the higher strength in shear than in tension perpendicular to grain of wood a stiffer test setup to get a stable crack growth for the determination of the fracture energy. The test results reached with the CTS series are clearly reduced and limit the direct comparison. The failure behavior of Radiata Pine LVL and timber can be further summarized using the fracture toughness observed within the common failure criteria. The failure criterion describes the relation between the failure modes I and II and the mixed mode failure. The linear or quadratic failure criterion are common ones, the failure criterion from Wu [21] was determined on Balsa wood and is often used or compared with wood species or wood products. Fig. 16 shows the values for Radiata Pine LVL and timber for the two crack systems RL and TL. The parameters observed are not enough to classify the failure behavior of Radiata Pine LVL and timber using one failure criterion. The concentration of the results to the pure fracture
10
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
Table 2 Fracture toughness and fracture energy results. Speci men
Test setup
Crack system
Number of tests (–)
Density (kg/m3)
Fracture toughness Kc
Fracture energy Gf Mode I
Mode II
Mode I
Mode II
Nm/m2
COV (%)
Nm/m2
COV (%)
kN/m3/2
COV (%)
kN/m3/2
COV (%)
Timber
CTS
RL TL NL
8/6 10/6 10/4
476 424 556
485 623 1219
28.5 17.0 18.4
6962 8087 12,722
22.6 25.9 8.3
239 147 296
19.4 19.8 10.5
829 786 1020
15.9 4.5 16.1
LVL
CTS
RL TL RL TL TL
8/3 9/5 –/20 –/40 19/–
595 596 602 607 613
673 1213 – – 1192
19.8 26.6 – – 25.4
4009 7607 3284 5192 –
17.3 24.4 29.6 26.1 –
127 170 – – –
29.0 12.9 – – –
385 528 316 402 –
17.7 28.5 21.0 21.2 –
CSS SENB
Fig. 13. Fracture energies for mode I in Radiata Pine timber.
Fig. 14. Fracture energies for mode I in Radiata Pine LVL.
mode I confirms only the mechanical behavior observed and the known difference between the tension strength perpendicular to grain and the shear strength. For a final analysis, a finer variation of the loading angle a in the CTS test setup should be investigated. The failure behavior in Radiata Pine LVL and timber shows a high crack resistance in both fracture modes I and II in comparison to the one for European wood species. For example, the fracture energy for Radiata Pine timber is approximately double the one for mode I or five times for mode II compared to the average values of the fracture energy of European spruce or pine or Canadian spruce found by e.g. Aicher et al. [3,22], Boström [23], Daudeville [24], Franke [2], Kretschmann [25], Riberholt et al. [26], Schatz [27], Smith et al. [6], Vasic [7]. This high crack resistance could be partly transferred to LVL from Radiata Pine timber. The fracture energies observed for LVL in the fracture mode I are higher for both crack systems RL and TL. However, the differences of the fracture energies for the fracture mode II are in opposite direction. Radiata Pine timber clearly shows a higher crack resistance than LVL.
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
11
Fig. 15. Fracture energies for mode II in Radiata Pine LVL.
Fig. 16. Failure behavior of Radiata Pine LVL and timber compared to the common failure criteria.
When compared to the results by King et al. [1], the observed fracture toughness for mode I and II of Radiata Pine timber shows similar differences between the crack systems RL and TL but the values obtained are lower. LVL results show a lower crack resistance when compared to other wood products such as particle board (FPY), medium density fiberboard (MDF) or oriented strand board (OSB) as observed by Niemz et al. [9,10]. Since the manufacturing of particle boards involves a large quality of glue and results in fibers that are loaded in all possible directions, fracture energies of more than twice as the LVL ones are possible.
6. Conclusions A comprehensive test program to determine the fracture parameters of Radiata Pine LVL and timber was initiated. Characterization of the different test setups available for wood was possible and permitted the observation and analysis of the fracture failure behavior. Comparison of the various test setups shows no differences in the fracture parameters determined, only the preparation of the test specimens and the handling during the experiment is different. The failure behavior of Radiata Pine LVL shows a more ductile behavior in all fracture modes considered in comparison to timber. This leads to higher values of the fracture energy for mode I for LVL than for timber. On the other hand, the energy which was needed for the separation of the crack surface for mode II is smaller for LVL than for timber. The comparison of the mechanical behavior also shows influences of the orientation and the production method for LVL on the crack resistance. For both LVL and timber, smaller load capacities were observed in the RL crack system in comparison to the ones for the TL crack system. Radiata Pine timber shows on average a stiffer material behavior, but therefore leads almost to a quite sudden brittle failure. In structural details such as connections loaded perpendicular to grain or beams with holes or endnotched beams in complex constructions, the different failure behavior of Radiata Pine LVL and timber should be analyzed carefully and the orientation of the layers for LVL has to be considered for the design. The strength behavior values of timber and LVL depends strongly on the grain direction as well as the fracture behavior. In novel design approaches for the characterization of
12
B. Franke, P. Quenneville / Engineering Fracture Mechanics 116 (2014) 1–12
notches [28,29], or connections with mechanical fasteners, [30] the fracture mechanic was applied. For their application the fracture material parameters are needed. The research results presented are useful in numerical simulation of complex wooden structures or in the development of new design approaches. The material parameters provided are important for a comprehensive material description of wood or wood products used in numerical simulations, such as in the splitting failures of connections as shown in [31,32]. Acknowledgments The research work was generously supported by the Structural Timber Innovation Company (STIC) from New Zealand. The students Armithab Athanari and Joseph Pearson worked on the specimen preparation and the experimental tests during a student project as well as Dmitry Volynkin as a summer student. References [1] King MJ, Sutherland IJ, Le-Ngoc L. Fracture toughness of wet and dry Pinus Radiata. Holz als Roh- und Werkstoff 1999;57:235–40. [2] Franke B. Zur Bewertung der Tragfähigkeit von Trägerausklinkungen in Nadelholz. Doctoral dissertation, Germany: Bauhaus University Weimar; 2008. [3] Aicher S, Boström L, Gierl M, Kretschmann D, Valentin G. Determination of fracture energy of wood in mode II. Rilem TC 133 Report. Sweden: Swedish National Testing and Research Institute; 1997. [4] Valentin HG, Boström L, Gustafsson PJ, Ranta Maunus A, Gowda S. Application of fracture mechanics to timber structures RILEM state-of-the-art report. Research Notes 1262. Finland: Technical Research Centre of Finland; 1991. [5] Larsen HJ, Gustafsson PJ. The fracture energy of wood in tension perpendicular to grain, results from a joint testing project. In: Conference proceedings CIB-W18A, paper 23–19-2, Portugal; 1990. [6] Smith I, Landis E, Gong M. Fracture and fatigue in wood. England: John Wiley & Sons Ltd.; 2003. [7] Vasic S. Applications of fracture mechanics to wood. Doctoral dissertation, Canada: The University of New-Brunswick; 2000. [8] Stanzl-Tschegg SE, Navi P. Fracture behaviour of wood and its composites. A review, Holzforschung 2009;63:139–49. [9] Niemz P, Diener M. Vergleichende Untersuchungen zur Ermittlung der Bruchzähigkeit an Holzwerkstoffen. Holz als Roh- und Werkstoff 1999;57:222–5. [10] Niemz P, Diener M, Pöhler E. Untersuchungen zur Ermittlung der Bruchzähigkeit an MDF-Platten. Holz als Roh- und Werkstoff 1997;55:327–30. [11] Lo KW, Gong YB, Tamilselvan T, Lai MO. A proposed specimen for KIIc testing. Int J Fract 2003;124:127–37. [12] Xu S, Reinhardt HW, Gappoev M. Mode II fracture testing method for highly orthotropic materials like wood. Int J Fract 1996;75:185–214. [13] Valentin HG, Caumes P. Crack propagation in mixed mode in wood: a new specimen. Wood Sci Technol 1989;23:43–53. [14] Cramer SM, Pugel AD. Compact shear specimen for wood mode II fracture investigations. Int J Fract 1987;35:163–74. [15] Jones DL, Chisholm DB. An investigation of the edge-sliding mode in fracture mechanics. Engng Fract Mech 1975;7:261–70. [16] Prokopski G. Investigation of wood fracture toughness using mode II fracture (shearing). J Mater Sci 1995;30:4745–50. [17] Richard HA, Benitz K. A loading device for the creation of mixed mode in fracture mechanics. J Fract Mech 1983;22:55–8. [18] Gross D. Bruchmechanik. Berlin: Springer Verlag; 1996. [19] Tan DM, Stanzl-Tschegg SE, Tschegg EK. Models of wood fracture in mode I and mode II. Holz als Roh- und Werkstoff 1995;53:159–64. [20] Binner SB. Fatigue crack growth studies under mixed mode loading. Int J Fatigue 2001;23:259–63. [21] Wu EM. Application of fracture mechanics to anisotropic plates. J Appl Mech 1967:974. [22] Aicher S. Bruchenergien, kritische Energiefreisetzungsraten und Bruchzähigkeit von Fichte bei Zugbeanspruchung senkrecht zur Faserrichtung. Holz als Roh- und Werkstoff 1994;51:361–70. [23] Boström L. Method for determination of the softening behaviour of wood and the applicability of nonlinear fracture mechanics model. Doctoral dissertation, Sweden: Lund; 1992. [24] Daudeville L. Fracture in spruce: experiment and numerical analysis by linear and non linear fracture mechanics. Holz als Roh- und Werkstoff 1999;57:425–32. [25] Kretschmann DE. Effect of varying dimensions on tapered end-notched flexure shear specimen. Wood Sci Technol 1995;29:287–93. [26] Riberholt H, Enquist B, Gustafsson PJ, Jensen RB. Timber beams notched at the support. Afdelingen for Bærende Konstruktioner, Lungby: Danmarks Tekniske Højskole, R280; 1991. [27] Schatz T. Zur bruchmechanischen Modellierung des Kurzeit-Bruchverhaltens von Holz im Rissöffnungsmodus I. Doctoral dissertation, Germany: University of Stuttgart; 1994. [28] Gustafsson PJ. A study of strength notched beams. In: Conference proceedings CIB-W18A, paper 21–10-1, Canada; 1988. [29] Jockwer R, Steiger R, Frangi A, Köhler J. Impact of material properties on the fracture mechanics design approach for notched beams in Eurocode 5. In: Conference proceedings CIB-W18A, paper 44–6-1, Italy; 2011. [30] Franke B, Quenneville P. Design approach for the splitting failure of dowel-type connections loaded perpendicular to grain. In: Conference proceedings CIB-W18A, paper 44–7-5, Italy; 2011. [31] Franke B, Quenneville P. Numerical modeling of the failure behavior of dowel connections in wood. ASCE, J Engng Mech 2010;137:186–95. [32] Franke B, Franke S, Quenneville P. Numerical modeling and analysis of the failure behaviour of dowel-type connections in wood. In: European congress on computational methods in applied sciences and engineering (ECCOMAS 2012), Austria; 2012.