BRITTLE FAILURE OF CROSS-LAMINATED TIMBER AT CONNECTIONS WITH SELF-TAPPING SCREWS by Amir Einipour Rashti B.Sc., University of Guilan, 2014 M.Sc., K. N. Toosi University of Technology, 2018 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE IN ENGINEERING UNIVERSITY OF NORTHERN BRITISH COLUMBIA November 2024 © Amir Einipour Rashti, 2024 Abstract Cross-laminated timber (CLT) is becoming increasingly popular in construction, but its crosswise layup structure leads to distinct brittle failure modes compared to solid timber or glulam. Self-tapping screws (STS) are widely used as fasteners in timber construction and typically exhibit ductile failure when loaded laterally. However, when screws are installed at an angle, brittle failure modes may become critical, influencing design. Although the 2024 edition of the Canadian Standard for Engineering Design in Wood (CSA O86) includes provisions for connections with STS, there is limited guidance for estimating the resistance of CLT connections under brittle failure conditions. In this study, experimental and analytical investigations were conducted to examine the brittle failure modes and load-carrying capacities of CLT connections with STS. In the experimental phase, uniaxial tension tests were conducted on CLT connections with STS installed at a 45º angle. A total of 18 test series were performed, encompassing various connection layouts that considered CLT lay-up, screw penetration length, edge distance of the connection, and screw arrangements. Following the experimental phase, an analytical investigation was carried out on available test data to evaluate the predictive ability and limitations of current models for brittle failure in CLT. Based on these findings, a new approach for predicting brittle failure in CLT connections with inclined STS was proposed. ii Table of contents Abstract ............................................................................................................................. ii Table of contents ............................................................................................................. iii List of tables .................................................................................................................... vi List of figures ................................................................................................................ viii Acknowledgements ........................................................................................................ xii 1 2 Introduction .............................................................................................................. 1 1.1 Background ........................................................................................................ 1 1.2 Research need .................................................................................................... 2 1.3 Objectives .......................................................................................................... 3 1.4 Thesis organization ............................................................................................ 3 1.5 Scope and limitations ......................................................................................... 3 Literature review ....................................................................................................... 5 2.1 Background ........................................................................................................ 5 2.1.1 Environmental concerns ............................................................................. 5 2.1.2 Mass timber construction ........................................................................... 5 2.1.3 Cross-laminated timber .............................................................................. 6 2.1.4 CLT connections with STS ........................................................................ 6 2.1.5 Brittle failure of timber connections ........................................................... 8 2.2 Design provisions for brittle failure modes of STS connection ....................... 11 2.2.1 CSA O86 .................................................................................................. 11 2.2.2 Eurocode 5 ................................................................................................ 12 2.2.3 NDS 2024 ................................................................................................. 13 2.3 Research on brittle failure in glulam and LVL connections ............................ 14 2.4 Research on brittle failure in CLT connections ............................................... 16 2.4.1 Zarnani and Quenneville (2015) ............................................................... 16 2.4.2 Azinović et al. (2022) ............................................................................... 17 2.4.3 Ni and Niederwestberg (2022) ................................................................. 19 iii 2.4.4 2.5 3 5 Summary of literature review .......................................................................... 21 Experimental investigations ................................................................................... 23 3.1 Overview .......................................................................................................... 23 3.2 Materials .......................................................................................................... 23 3.3 Specimen description ....................................................................................... 24 3.4 Methods ........................................................................................................... 27 3.5 Analyses ........................................................................................................... 29 3.6 Results .............................................................................................................. 31 3.6.1 Overview .................................................................................................. 31 3.6.2 Series S1 and S2 (105 mm 3-ply CLT with 120 mm screws) .................. 32 3.6.3 Series S3 and S4 (100 mm 5-ply CLT with 120 mm screws) .................. 34 3.6.4 Series S5- S8 (139 mm 5-ply CLT with 120 and 180 mm screws) .......... 36 3.6.5 Series S9-S14 (175 mm 5-ply CLT with 180 and 240 mm screws) ......... 39 3.6.6 Series S15 and S16 (191 mm 7-ply CLT with 180 and 240 mm screws) 45 3.6.7 Series S17 and S18 (245 mm 7-ply CLT with 280 mm screws) .............. 48 3.7 4 Previous work at UNBC (2022) ............................................................... 20 Evaluation of connection parameters ............................................................... 49 Analytical investigation of experimental campaigns.............................................. 55 4.1 Overview .......................................................................................................... 55 4.2 Previous experimental campaigns ................................................................... 55 4.2.1 Azinović et al. (2022) ............................................................................... 55 4.2.2 Ni and Niederwestberg (2022) ................................................................. 57 4.2.3 Previous work at UNBC (2022) ............................................................... 58 4.3 Application of prediction models .................................................................... 60 4.4 Prediction based on CSA O86-2024 ................................................................ 63 4.5 Prediction based on prEN1995 (2024) ............................................................. 66 4.6 Prediction based on stiffness model ................................................................. 69 4.7 Evaluation of prediction models ...................................................................... 71 Proposed model for brittle failure in CLT with inclined STS ................................ 75 iv 6 5.1 Overview .......................................................................................................... 75 5.2 Effective depth for brittle failure of CLT with inclined STS .......................... 75 5.3 Prediction based on the proposed model ......................................................... 77 5.4 Comparison of the proposed prediction model with CSA O86-2024 .............. 81 Conclusions ............................................................................................................ 84 6.1 Summary of experimental work....................................................................... 84 6.2 Summary of analytical work ............................................................................ 84 6.3 Future research ................................................................................................. 85 References ...................................................................................................................... 87 Appendix A: CSA O86-2024 ......................................................................................... 93 Appendix B: prEN1995 (2024) ...................................................................................... 98 Appendix C: Stiffness model for CLT ......................................................................... 102 Appendix D: Test series sketches ................................................................................. 105 v List of tables Table 3-1 CLT properties. .............................................................................................. 23 Table 3-2 Test series overview. ...................................................................................... 27 Table 3-3 Summary of experimental tests ...................................................................... 31 Table 4-1 Test series overview in the experiments by Azinović et al.[15]. ................... 56 Table 4-2 Results of test series in experiments by Azinović et al. [15]. ........................ 56 Table 4-3 Test series overview in the experiments by Ni and Niederwestberg [40]...... 57 Table 4-4 Results of test series in experiments by Ni and Niederwestberg [40]............ 58 Table 4-5 Test series overview in the previous experiments at UNBC [41]. ................. 59 Table 4-6 Results of test series in previous experiments at UNBC [41]. ....................... 60 Table 4-7 Different approaches for calculation of plug shear failure [10], [13], [29].... 61 Table 4-8 Mean properties of lumber used in CLT products [45] [46]. ......................... 63 Table 4-9 CSA O86 predictions for test series with STS at a 90º. ................................. 64 Table 4-10 CSA O86 predictions for test series with STS at a 45º. ............................... 65 Table 4-11 prEN 1995 predictions for test series with STS at a 90º. ............................. 66 Table 4-12 prEN 1995 predictions for test series with STS at a 45º. ............................. 68 Table 4-13 Stiffness model predictions for test series with STS at a 90º. ...................... 69 Table 4-14 Stiffness model predictions for test series with STS at a 45º. ...................... 70 Table 4-15 Metrics obtained for the different models. ................................................... 73 Table 5-1 Observed effective depth of the failure blocks for all test series. .................. 76 Table 5-2 Overview of the proposed models for calculation of plug shear failure. ....... 78 Table 5-3 Geometric details of resistance planes in proposed model for each test series. ............................................................................................................................ 79 Table 5-4 Predicted load-carrying capacity based on the proposed model for inclined test data. .................................................................................................................... 80 Table 5-5 Metrics obtained for proposed and CSA O86 models in centered test series. 82 vi Table 5-6 Metrics obtained for proposed and CSA O86 models in offset test series..... 82 vii List of figures Figure 1.1 Brittle failure modes of small dowel type fasteners: Row Shear (a), Group Tear-Out (b), Plug Shear (c), Step Shear (d), Net Tension (e). ............................ 2 Figure 2.1 CLT panel layout [3]. ...................................................................................... 6 Figure 2.2 Partially threaded STS (top) and fully threaded STS (bottom). ...................... 7 Figure 2.3 Laterally loaded screws (a) and at angle loaded screws (b). ........................... 7 Figure 2.4 Typical CLT building with various connections [3]. ...................................... 8 Figure 2.5 Brittle failure modes of fully penetrated dowel-type fasteners in solid wood and glulam loaded parallel-to-grain [26]. ............................................................. 9 Figure 2.6 Brittle failure modes of small dowel type fasteners: Row Shear (a), Group Tear-Out (b), Plug Shear (c), Step Shear (d), Net Tension (e). .......................... 10 Figure 2.7 Plug shear failure (a) and step shear failure (b) in CLT. ............................... 11 Figure 2.8 Effective depth of failure block [13]. ............................................................ 12 Figure 2.9 Stiffness-based model for wood block tear-out resistance in CLT [10]. ...... 16 Figure 2.10 Different possible failure modes of wood block tear-out in CLT [10]. ...... 17 Figure 2.11 Possible failure modes for CLT connections [15]. ..................................... 18 Figure 2.12 Test set-up and test specimen with denotations [11]. ................................. 19 Figure 2.13 Test setup [40]. ............................................................................................ 19 Figure 2.14 Test setup at UNBC (2022) [41]. ................................................................ 20 Figure 2.15 Common observed brittle failure: Tension break in first lamella (a), plug shear failure (b), step shear failure (c) [41]. ................................................................. 21 Figure 3.1 ASSY plus VG screws and 45º angle washer. .............................................. 24 Figure 3.2 Steel plates, (a) schematic, (b) photo. ........................................................... 24 Figure 3.3 Specimen configuration, (a) cross section, (b) top view. .............................. 25 Figure 3.4 Preparation of specimens, (a) MTC pre-drilling jig equipped with angle washer, (b) wooden boards used as stoppers. ..................................................... 25 Figure 3.5 Schematic of specimens in, a) series S1, (b) series S2.................................. 26 viii Figure 3.6 Full test set-up. .............................................................................................. 27 Figure 3.7 Fixtures in test set-up(a) schematic, (b) photo. ............................................. 28 Figure 3.8 Test setup and location of string potentiometers. ......................................... 28 Figure 3.9 Loading protocol. .......................................................................................... 29 Figure 3.10 Possible brittle failure modes in CLT. ........................................................ 30 Figure 3.11 Failure modes in series S1, (a) Mode D, (b) Mode F, (c) net tension. ........ 32 Figure 3.12 Failure modes in series S2, (a) tension failure of parallel layer, (b) plug shear Mode D, (c) rolling shear failure. ....................................................................... 33 Figure 3.13 Load-displacement curves for series S1. ..................................................... 34 Figure 3.14 Load-displacement curves for series S2. ..................................................... 34 Figure 3.15 Failure modes in series S3, (a) Mode B, (b) Mode F. ................................. 35 Figure 3.16 Failure modes in series S4, (a) Mode F, (b) Mode H. ................................. 35 Figure 3.17 Load-displacement curves for series S3. ..................................................... 36 Figure 3.18 Load-displacement curves for series S4. ..................................................... 36 Figure 3.19 Failure modes in series S5, (a) plug shear Mode F, (b) Mode J. ................ 37 Figure 3.20 Failure modes in series S8, (a) Mode H, (b) tensile failure due to bending.37 Figure 3.21 Load-displacement curves for series S5. ..................................................... 38 Figure 3.22 Load-displacement curves for series S6. ..................................................... 38 Figure 3.23 Load-displacement curves for series S7. ..................................................... 39 Figure 3.24 Load-displacement curves for series S8. ..................................................... 39 Figure 3.25 Failure modes in series S9, (a) Mode J, (b) layer separation and rolling shear. ............................................................................................................................ 40 Figure 3.26 Failure modes in series S10. ....................................................................... 40 Figure 3.27 Load-displacement curves for series S9. ..................................................... 41 Figure 3.28 Load-displacement curves for series S10. ................................................... 41 Figure 3.29 Failure modes in series S11, (a) Mode F, (b) Mode H. ............................... 42 Figure 3.30 Failure modes in series S12, (a) Mode F, (b) Mode H. ............................... 42 ix Figure 3.31 Load-displacement curves for series S11. ................................................... 43 Figure 3.32 Load-displacement curves for series S12. ................................................... 43 Figure 3.33 Failure modes in series S13, (a) Mode B, (b) Mode H, (c) Mode J. ........... 44 Figure 3.34 Tension failure and rolling shear in series S14. .......................................... 44 Figure 3.35 Load-displacement curves for series S13. ................................................... 45 Figure 3.36 Load-displacement curves for series S14. ................................................... 45 Figure 3.37 Failure modes in series S15, (a) tension failure due to bending, (b) step shear, (c) net tension effect. .......................................................................................... 46 Figure 3.38 Failure modes in series S16, (a) Mode F, (b) Mode H. ............................... 47 Figure 3.39 Load-displacement curves for series S15. ................................................... 47 Figure 3.40 Load-displacement curves for series S16. ................................................... 47 Figure 3.41 Step shear failure observed in series S17. ................................................... 48 Figure 3.42 Step shear failure observed in series S18. ................................................... 48 Figure 3.43 Load-displacement curves for series S17. ................................................... 49 Figure 3.44 Load-displacement curves for series S18. ................................................... 49 Figure 3.45 Effect of unloaded edge distance on load-carrying capacities. ................... 50 Figure 3.46 Effect of unloaded edge distance on elastic stiffness of connections. ........ 51 Figure 3.47 Effect of connection width on load-carrying capacities. ............................. 52 Figure 3.48 Effect of connection width on elastic stiffness of connections. .................. 52 Figure 3.49 Effect of penetration depth into the parallel layers on load-carrying capacities. ............................................................................................................................ 54 Figure 3.50 Effect of penetration depth into the parallel layers on elastic stiffness of connections. ........................................................................................................ 54 Figure 4.1 Definition of plug shear failure based on CSA O86-2024. ........................... 62 Figure 4.2 Scatter plot of the experimental results and the CSA O86 predicted values. 71 Figure 4.3 Scatter plot of the experimental results and the prEN 1995 predicted values. ............................................................................................................................ 72 x Figure 4.4 Scatter plot of the experimental results and stiffness model predicted values. ............................................................................................................................ 72 Figure 5.1 Effective depth and penetration depth of inclined screws. ........................... 75 Figure 5.2 Failure mode and observed effective depth in series S11. ............................ 76 Figure 5.3 Definition of plug shear failure in proposed models. .................................... 77 Figure 5.4 Scatter plot of the experimental results and prediction based on CSA O86. 81 Figure 5.5 Scatter plot of the experimental results and prediction based on the proposed model. ................................................................................................................. 82 xi Acknowledgements I would like to extend my deepest gratitude to my supervisor, Dr. Tannert, for his unwavering guidance, invaluable expertise, and steadfast dedication throughout my research journey. His insightful feedback and encouragement have been instrumental in shaping this work and helping me overcome challenges along the way. I am also sincerely thankful to my committee members, Dr. Salenikovich and Dr. Zhou, for their constructive feedback and thoughtful suggestions, which have greatly enhanced the quality of this research. I am additionally grateful to Dr. Chui for his review and valuable input. My appreciation goes to lab instructor Maik Gehloff and lab technicians James Andal, Nathan Downie, and Ryan Stern for their technical expertise and consistent support during the experimental phase at the Wood Innovation Research Laboratory. I am also grateful to students Sabari, Sanya, and Jerry for their invaluable help and dedication throughout the testing process, which will always be remembered. Finally, I am profoundly grateful to my parents and sisters for their boundless love, support, and encouragement. Their belief in me and constant motivation have been a source of strength not only during this research journey but throughout all the challenges in my life. xii 1 Introduction 1.1 Background In recent years, timber construction has attracted more attention due to increasing housing demand and environmental awareness [1],[2]. Engineered wood products, particularly cross-laminated timber (CLT), have been central to this trend. CLT is composed of boards laminated in alternating directions, enhancing dimensional stability [3]. As an alternative to reinforced concrete and steel, CLT panels offer lightweight, sustainable, and ecofriendly properties. In addition to engineered wood products, modern timber connectors, such as self-tapping screws (STS), have greatly influenced timber design, enabling the construction of mid- and high-rise timber buildings [4]. The brittle nature of timber under tension or shear loading presents a significant challenge for its use in the construction industry [5]. Different brittle failure modes can occur in connections with solid timber and glue-laminated timber (glulam), depending on member geometry, loading direction, penetration depth, diameter, and spacing of fasteners. For small dowel-type fasteners in solid timber, common possible brittle failure modes include row shear, group tear-out, plug shear, step shear, and net tension, as shown in Figure 1.1. STS connections with screws installed at 90º and loaded laterally typically exhibit ductile failure. However, when screws are installed at a 45º angle and loaded in withdrawal, they are more prone to brittle failure. Accurate estimation of brittle failure resistance is particularly important for CLT elements connected with steel plates and large group of STS to ensure connection reliability. Several models exist for predicting brittle failure modes in solid or uniformly layered timber members, such as those proposed by Stahl et al. (2004), Eurocode 5 (2014), Zarnani and Quenneville (2014), and Cabrero and Yurrita (2021), the latter specifically 1 addressing plug shear failure [6], [7], [8], [9]. However, for plug shear failure in CLT, the model developed by Zarnani and Quenneville (2015) stands as the only one [10]. Figure 1.1 Brittle failure modes of small dowel type fasteners: Row Shear (a), Group Tear-Out (b), Plug Shear (c), Step Shear (d), Net Tension (e). 1.2 Research need Brittle failure of connections limits the ductility of the structure; disregarding this aspect can result in a sudden structural collapse [11]. The 2019 version of the Canadian Standard for Engineering Design in Wood, CSA O86 [12], addressed various brittle failure modes for bolted connections, including row shear, group tear-out, and net tension. In the 2024 edition of CSA O86 [13], design provisions for STS were added. However, the guidance on the various brittle failure modes of CLT connections with STS is still limited. While many previous studies e.g. [8], [11], [14], [15], focused on investigating brittle failure in glulam and solid timber with dowel-type fasteners, limited research has been conducted on CLT ([10], [15], [16]). Unlike glulam, CLT has distinct major and minor axes, which complicates the determination of failure modes in STS connections compared to those observed in glulam and solid timber. Therefore, further research is needed to fill the gaps in designing standards and establish reliable design provisions for STS connections in CLT. 2 1.3 Objectives The main objective of this thesis was to investigate the brittle failure modes and associated load-carrying capacities of connections with inclined STS in CLT. To achieve this objective, a total of 18 connection layouts were designed and tested with six replicates each to capture the anticipated brittle failure in CLT. Subsequently, an analytical study was conducted to evaluate the Canadian design provisions and other existing models for predicting brittle failure modes in CLT using the available experimental results. Finally, a new approach for predicting brittle failure in CLT connections with inclined STS was proposed. 1.4 Thesis organization In Chapter 2, a literature review on CLT and STS in mass-timber construction is provided, covering failure modes of timber connections, design provisions for STS, and previous research on brittle failure in CLT with STS. Chapter 3 describes the experimental materials, methods, and results. Chapter 4 presents previous experimental campaigns and uses different analytical models to predict failure modes and connection load-carrying capacities based on the available experimental data. In Chapter 5, a modified model for predicting brittle failure in CLT connections with inclined screws is proposed, and these predictions are compared with CSA O86 predictions. Finally, Chapter 6 summarizes the research findings, contributions, and recommendations for future studies. 1.5 Scope and limitations This study focuses on the brittle failure of CLT connections with STS, using both analytical and experimental approaches. Several limitations were encountered, including the inability to vary CLT panel and steel plate lengths due to test setup constraints, which restricted screw arrangement options. Additionally, applying the proposed analytical model to 90º connections was not feasible due to insufficient data on the observed depth 3 of the failure block from previous experiments. Also, the lack of a high-speed camera limited the ability to capture the exact sequence of failure planes during brittle failure, hindering a detailed understanding of the failure mechanisms. 4 2 Literature review 2.1 Background 2.1.1 Environmental concerns Environmental concerns have triggered global efforts to mitigate the impact of human activities on the environment. The construction industry is one of the key targets for environmental impact reduction due to accounting for approximately one-third of global energy consumption and around 15% of direct carbon dioxide emissions [17]. Choosing sustainable materials like wood instead of energy-intensive options such as concrete and steel can significantly reduce greenhouse gas emissions from manufacturing [18]. For instance, a rise of 17% in the utilization of wood led to a corresponding decrease of 20% in carbon emissions from construction materials [19]. This encourages the growth of the timber industry to address global warming. In the last decades, the rise of mass timber has played a crucial role in promoting sustainable methods of construction [20]. 2.1.2 Mass timber construction Mass timber, which refers to a variety of large dimension engineered wood products, is increasingly being used due to its strength, durability, and dimensional stability, with common types including CLT, glulam, laminated veneer lumber (LVL), dowel laminated timber (DLT), nail-laminated timber (NLT), and mass plywood panels (MPP) [21]. Mass timber construction offers environmental benefits, as its manufacturing process is less energy-intensive compared to concrete and steel, resulting in a reduced carbon footprint for buildings. Furthermore, mass timber construction provides practical advantages, such as off-site prefabrication, which allows for customization and quick assembly on-site [3]. 5 2.1.3 Cross-laminated timber CLT, developed in Austria and Germany in the 1980s, is an engineered wood product composed of boards stacked crosswise, typically at 90 degrees, and glued on their wide and sometimes narrow faces [3]. The orthogonal arrangement of CLT layers, as shown in Figure 2.1, creates a panel that can be used in a variety of structural applications, including walls, floors, and roofs [3]. One of the significant advantages of CLT compared to other wood products is its dimensional stability. Additionally, the panels are prefabricated with openings for doors, windows, stairs, service channels, and ducts, allowing for quick and efficient installation. These panels can be shipped to the construction site with pre-installed lifting straps, facilitating easy handling and fewer skilled workers on-site. Also, another advantage of CLT is its light weight, which reduces the foundation demands and construction time [22]. Figure 2.1 CLT panel layout [3]. 2.1.4 CLT connections with STS In structural design, connections play a critical role, influencing a structure's performance and cost. Dowel-type fasteners such as nails, rivets, bolts, and STS have been established as means of connecting CLT panels [3]. STS are manufactured from hardened steel, by creating a thread along the shank through rolling or forging [4]. Unlike traditional lag 6 screws, STS use a self-tapping tip that eliminates the need for predrilling and reduces drive-in torque. STS can achieve high yielding moments, high withdrawal resistance, and high tensile and torsional strengths. They are available in diameters up to 14 mm and lengths up to 1.5 m, making them ideal for deep mass timber cross-sections [23]. STS can be classified into two categories: partially threaded and fully threaded screws, as shown in Figure 2.2, with the latter providing withdrawal resistance along their whole length. Figure 2.2 Partially threaded STS (top) and fully threaded STS (bottom). STS can be loaded laterally (perpendicular to its axis), in withdrawal (along its axis), or at an angle in combination of axial and lateral loads. Connections with laterally loaded screws (Figure 2.3 (a)) offer relatively low stiffness but exhibit ductile failure, whereas axially loaded screws provide higher stiffness but lower ductility. Practically, it is preferable to load screws at an angle, usually no less than 30° to the grain (Figure 2.3 (b)) to benefit from both lateral and withdrawal strengths [24]. (a) (b) Figure 2.3 Laterally loaded screws (a) and at angle loaded screws (b). STS are recognized as the leading choice for connections in CLT structures [4] and can be used in various types of connections, as shown in Figure 2.4 [3]. Panel-to-panel connections (Detail A) are commonly used to join individual CLT panels along their longitudinal edge. Wall-to-wall connections (Detail B) connect walls positioned at right 7 angles, either between partition walls and exterior walls or between exterior corner walls. In this case, STS can be utilized directly at a 90° angle or at an angle on the narrow face of the wall, sometimes accompanied by wooden profiles. Wall-to-floor connections (Detail C) are made between CLT floors or roofs and walls below, typically employing STS directly at a 90° angle or at an angle from the CLT floor into the narrow face of the wall edge. Wall-to-roof connections (Detail D) are established between sloped or flat roofs and walls using STS at a 90° angle from the roof to the wall or at an angle from the wall to the roof. Figure 2.4 Typical CLT building with various connections [3]. 2.1.5 Brittle failure of timber connections Timber connections can fail in two ways: ductile or brittle. In a ductile failure, either the fastener yields, or the wood is compressed; this is accompanied by larger deformation. Conversely, in a brittle failure, the wood typically fails without any plastic behavior. Ductile failure modes of STS connections are well-described by the European Yield Model (EYM), introduced by Johansen (1949), which considers plastic embedment failure of wood and fastener yielding [25]. This model has been used for decades in wood design standards, including CSA O86 [12], [13]. Traditionally, timber joint design 8 focused on ductile failure modes, assuming that a minimum fastener spacing will prevent brittle failures, and accounting for brittle failure modes implicitly by group effect factors. Brittle failures in timber connections are influenced by various factors, including the type of timber member, fastener penetration depth, diameter, and spacing. In connections with dowel-type fasteners that fully penetrate solid wood and glulam under tension parallel to the grain, brittle failure modes include splitting, row shear, group tear-out, and net tension, as shown in Figure 2.5 [26]. Figure 2.5 Brittle failure modes of fully penetrated dowel-type fasteners in solid wood and glulam loaded parallel-to-grain [26]. However, for connections with partial penetration, additional possible failure modes include plug shear and step shear (Figure 2.6) [9]. 9 Figure 2.6 Brittle failure modes of small dowel type fasteners: Row Shear (a), Group Tear-Out (b), Plug Shear (c), Step Shear (d), Net Tension (e). Splitting is characterized by a single crack near holes, caused by tension perpendicular to the grain. Row shear involves two longitudinal cracks along the fastener row. Group tearout occurs along the outer boundaries of a fastener group, affecting the timber's full thickness, while net tension results in tensile breakage across the last fastener row. Plug shear failure appears at a specific depth within the timber along the outer boundaries of a fastener group, and step shear typically occurs in narrow specimens under tensile load, leading to timber failure across the last row of fasteners at a certain depth. In CLT members, the presence of cross-layers influences the width and depth of the failure block, altering the possible failure modes [10]. Splitting is rare in CLT because the reinforcement effect from cross-lamination [27]. For connections with fully penetrated fasteners in CLT, group tear-out and net tension are common [28]. However, for partially penetrated fasteners, plug shear and step shear are frequently identified as failure modes, influenced by the specimen width [15], as shown in Figure 2.7. 10 (a) (b) Figure 2.7 Plug shear failure (a) and step shear failure (b) in CLT. 2.2 Design provisions for brittle failure modes of STS connection 2.2.1 CSA O86 The 2019 edition of CSA O86 [12] lacked specific design guidance for STS connections. Consequently, alternative methods were employed, which include the use of conservative design factors or using lag screw [36] provisions. Brittle failure modes were addressed only for bolted connections, covering row shear, group tear-out, and net tension. Row shear and group tear-out were not required to be considered for CLT, which was not conservative, as group tear-out can occur in CLT panels [28]. The 2024 edition of CSA O86 [13] introduced design provisions for STS connections, addressing both ductile and brittle failure modes. Alongside provisions for the axial and lateral resistance of screws, it includes equations for calculating withdrawal resistance. With respect to brittle failure modes, five types are categorized: row shear, group tearout, plug shear, step shear, and net tension. The provisions for calculating brittle failure resistance in STS connections are based on the penetration of the fasteners into the wood member. For fully penetrated members, resistance is determined by the minimum of net tension, row shear, and group tear-out resistances, whereas for partially penetrated 11 members, it involves the lowest value among net tension, row shear, plug shear, and step shear resistances. Brittle failure resistance for each type is determined by the resistance of the head tensile, side shear, and bottom planes, varying with the failure type. Evaluating these planes’ resistance involves assessing their effective depth and width. CSA O86-2024 provides equations and provisions for calculating the effective depth (‫ݐ‬௘௙௙ ) of the head tensile and side shear planes for both fully and partially penetrated wood members. Full details of the design provisions for brittle failure modes according to CSA O86 [13] are presented in Appendix A. Figure 2.8 Effective depth of failure block [13]. 2.2.2 Eurocode 5 In the current Eurocode 5, EN1995 (2004) [7], CLT is not recognized, and brittle failures in connections under lateral load parallel to the grain are excluded. However, certain brittle failure modes are indirectly considered within the ductile mode calculation using the effective number of fasteners (݊௘௙ ), which adjusts the connection's capacity by reducing the number of fasteners. The informative Annex A of Eurocode 5 provides methods to determine block-shear and plug-shear failures, applicable to steel-to-timber connections, but excludes connections with more than two shear planes or timber-totimber connections. The characteristic load-carrying capacity for block and plug shear is determined by the resistance of the head tensile plane (H), lateral planes (L), and bottom 12 shear planes (B)—each multiplied by a numerical factor. For calculating the areas of the head tensile and lateral shear planes, the effective thickness of the failure mode is necessary. In EN1995 [7], the effective depth of failure planes (‫ݐ‬௘௙௙ ) depends on two parameters: the thickness of the steel plate and the fastener's yielding mode. In the 2024 draft of the second generation of Eurocode 5, prEN1995 [29], provisions are provided for both ductile and brittle failure modes in connections with dowel-type fasteners. For connections loaded parallel to the grain, row shear, block shear, net tensile failure, and plug shear are explicitly included. Like in EN1995 [7], splitting is indirectly considered within the ductile mode calculation using the effective number of fasteners (݊௘௙ ). The brittle failure capacity for other failure modes is determined by the resistance of the involved failure planes, including the head tensile, side shear, and bottom shear planes. In prEN1995 [29], two equations were introduced to establish the effective depth of the brittle failure block for connections with partially and fully penetrated fasteners. The effective depth for the head tensile plane is aligned with that of the shear planes. Detailed information for calculating the resistance of failure planes and the factored resistance of brittle modes is included in Appendix B. It is important to note that the draft explicitly mentions that applying these provisions for calculating brittle failure resistance in CLT members tends to yield conservative results [29]. 2.2.3 NDS 2024 Appendix E of the NDS 2024 [30], which is non-mandatory, includes design provisions for addressing brittle failure modes in connections including net tension, row shear, and group tear-out. These modes represent local wood failure mechanisms at fasteners, particularly when fasteners are closely spaced and subjected to parallel-to-grain loading. 13 In such cases, the capacity of the fastener group may be limited by net tension, row shear, or group tear-out due to localized stresses. NDS 2024 notes that connections with closely spaced, large-diameter bolts often face limitations due to the capacity of the surrounding wood, whereas connections using smaller-diameter fasteners, such as those in woodframe construction, are less likely to face these constraints. 2.3 Research on brittle failure in glulam and LVL connections Brittle failure in wood connections was first considered in the 1980s when wood fractures along nail rows were observed by Nozynski [31] and Smith and Steck [32] , leading to the suggestion of using an effective number of fasteners. In 2004, Stahl et al. [33] proposed a closed-form model for group tear out, plug shear, and step shear based on experimental data for rivet connections in glulam and solid timbers parallel to the grain. Quenneville and Mohammad [34] evaluated the strength of steel–wood–steel bolted connections under brittle failure modes, considering variables such as end distance, bolt spacing, row spacing, number of bolts per row, member thickness, and wood species (glulam or sawn lumber). Their tests aimed to identify brittle failure modes such as row shear, group tear-out, and splitting. The work revealed that longitudinal shear stress at failure was linked to a parameter based on the smaller of the end distance or bolt spacing and the specimen thickness. This relationship was used to develop design equations to predict row shear and group tear-out strengths for glulam specimens, using the specified strength values. Building on this, Mohammad and Quenneville [35] proposed design provisions, verified through tests on thirty groups of specimens, selected to represent fundamental brittle and ductile failure modes. Comparisons between experimental results and predictions for steel–wood–steel bolted connections showed that the predictions for row shear failure overestimated the resistance. Additionally, specimens that failed in row shear exhibited shear failure over a reduced thickness, leading to the proposal of an 14 adjustment using the reduced (effective) thickness concept. These findings led to the incorporation of brittle failure modes for bolted connections into CSA O86 in 2009 [36]. Hanhijärvi and Kevarinmäki (2008) [37] conducted analytical and experimental studies on the brittle failure of glulam connections with fully penetrated dowel-type fasteners. In 2014, Zarnani and Quenneville [8] tested timber rivet connections in glulam and LVL, proposing a closed-form, stiffness-based method to determine the load-carrying capacity under longitudinal loading. Their method considers the stiffness and strength of planes under non-uniform shear and tension stresses and includes an algorithm that allows designers to predict potential brittle, ductile, and mixed failure modes in these connections. In 2019, Yurrita and Cabrero [14] tested glulam and LVL connections using slotted-in steel plates to study brittle failure modes. These results were compared against existing models to evaluate their predictive accuracy, leading to an improved model for assessing brittle failure in multiple shear connections with slotted-in steel plates and dowel-type fasteners. In 2020, a new model for brittle failure in connections with large diameter, fully penetrated fasteners across glulam, LVL, and solid timber was introduced by Yurrita and Cabrero [26]. This model incorporated modifications such as effective timber thickness and lateral shear plane lengths, based on prior work by Quenneville and Zarnani [38]. In 2021, Yurrita and Cabrero [39] tested connections with small diameter fasteners, such as nails and screws, to study the plug shear brittle failure mode in glulam and LVL. Despite the overall tests failing in plug shear with a mix of brittle and mixed failures, it was found that parameters like steel plate thickness and edge distance had negligible impact, whereas timber type and fastener slenderness were critical. Then, a new model was proposed [39], which utilized parameters from their previous work on large dowel type fasteners [26] and adapted them for plug shear scenarios based on findings from 15 these tests. The new model was benchmarked against major existing models and demonstrated significant improvements in load-carrying prediction accuracy. 2.4 Research on brittle failure in CLT connections 2.4.1 Zarnani and Quenneville (2015) Zarnani and Quenneville [10] extended their stiffness-based design approach [8] to CLT connections. The arrangement of layers in CLT influences the stress distribution. In the earlier model [8], load transfer to failure planes was modeled based on the relative stiffness ratios of adjacent resisting volumes at each plane, including head tensile, bottom shear, and lateral shear planes. However, lateral shear planes were disregarded in CLT due to the uncertain alignment with parallel planks. Instead, the stiffness-based model for block tear-out resistance in CLT [10] features a spring system that distributes loads between the head tensile plane (‫ܭ‬௛ ) and the depth under the bottom shear plane (‫ܭ‬ௗ ). This distribution is calculated by summing the stiffnesses of the bottom shear plane at ݀௭ (‫ܭ‬௕ ), rolling shear (‫ܭ‬௥ ), and longitudinal shear (‫ܭ‬௔ ) in the cross-layer, as shown in Figure 2.9. Figure 2.9 Stiffness-based model for wood block tear-out resistance in CLT [10]. Six distinct failure modes in plug shear are considered, detailed in Figure 2.10: partial failure of the outer layer (A), complete failure of the outer layer (B), partial failure of the first and second layers (C), complete failure of the first and second layers (D), partial 16 failure of the third layer (E), and complete failure of the third layer (F). These modes are influenced by the effective penetration depth of the fastener ( ‫ݐ‬௘௙ ) in relation to the thickness and layout of the layers. Appendix C provides details on the calculation procedures for the load-carrying capacity of wood in a CLT connection, stiffness of the involved planes, and relevant equations for each failure mode. To validate the design approach, Zarnani and Quenneville [10] tested timber rivet joints. The comparison of experimental outcomes with the analytical approach showed that the proposed CLT stiffness-based model surpasses the previous glulam and LVL model. Figure 2.10 Different possible failure modes of wood block tear-out in CLT [10]. 2.4.2 Azinović et al. (2022) Azinović et al. [15] studied the brittle failure of STS connections in CLT under tensile loads using both analytical and experimental approaches. They employed the stiffness design approach by Zarnani and Quenneville [10], examining six plug shear failure modes from the original stiffness model [10] and adding two more—net tension and step shear— to address brittle failure in CLT with STS connections, as depicted in Figure 2.11. Connections between steel plates and CLT using 60 screws on both the top and bottom sides were tested, as shown in Figure 2.12. Test specimens varied in width, CLT layup, 17 board orientation, and screw length. Results showed a direct correlation between loadcarrying capacity and CLT specimen width and fastener penetration into parallel or transverse layers. Comparing experimental results with the analytical model, it was evident that stiffness model from Zarnani and Quenneville [10] generally provided nonconservative results, with low predictive capability for most metrics and only showing good correlation for parallel connections. The model performed poorly for connections loaded in the minor strength direction of CLT. The study also suggested that current assumptions about embedment strength and fiber orientation need reconsideration. Figure 2.11 Possible failure modes for CLT connections [15]. 18 Figure 2.12 Test set-up and test specimen with denotations [11]. 2.4.3 Ni and Niederwestberg (2022) Ni and Niederwestberg [40] explored brittle failure modes in CLT and glulam connections using STS. Their investigation focused on assessing the impact of STS penetration length and diameter on the load-carrying capacity and brittle failure modes. Nine configurations were designed with screws installed perpendicular to the panel surface, loaded parallel-to-grain, as depicted in Figure 2.13. Figure 2.13 Test setup [40]. 19 Upon examining specimens and the equations for determining effective thickness, it became apparent that the CSA O86-2024 design method underestimates the actual effective thickness. Therefore, there is a need for alternative equations that can more accurately determine the effective thickness. 2.4.4 Previous work at UNBC (2022) In a study conducted at UNBC [41], brittle failure modes in CLT connections with STS were examined. Steel plates were installed using STS at a 45-degree angle to the panel's surface at both ends of CLT panels, as shown in Figure 2.14. A total of 82 tests were conducted on 18 different screw arrangements with varying screw lengths and CLT thicknesses. Most panels measured 900 mm x 300 mm, while three setups used wider panels (600 mm). Figure 2.14 Test setup at UNBC (2022) [41]. The study showed differences between the expected failure modes outlined in CSA O862024 and those observed experimentally. In narrower specimens loaded in the major strength direction, failures typically started with tensile failure at the first lamella, followed by bending (due to the lack of lateral support) rather than breaking of the remaining lamellas (Figure 2.15 (a)). Another common failure resembled plug shear 20 where a group of lamellas parallel to the load pulled out after a tension failure along the last column of screws, accompanied by delamination (Figure 2.15 (b)). In specimens loaded in the minor strength direction, failures similar to step shear occurred, involving tensile failure perpendicular to the grain at the last row of screws and separation between the first two lamellas, see Figure 2.15 (c). Figure 2.15 Common observed brittle failure: Tension break in first lamella (a), plug shear failure (b), step shear failure (c) [41]. The findings indicated that the width of the specimens affected their failure modes and loads, with wider specimens able to support higher loads and show varied failure behaviors. Notably, the failure loads observed were consistently higher than those predicted by the CSA O86-2024 proposal, suggesting that the standard's equations may not accurately predict real failure loads but rather offer overly cautious estimates. 2.5 Summary of literature review Mass timber, particularly CLT, has become prominent in construction due to its environmental sustainability and structural properties. Timber connections, which can exhibit either ductile or brittle failure modes, are critical for the overall performance of timber-based systems. Consequently, building codes and design standards evolved to include provisions for CLT and STS. CSA O86-2024 [13] introduced design provisions 21 for STS connections in CLT, providing detailed guidelines for calculating resistance that cover both ductile and brittle failures. Also, the next version of Eurocode 5 [29], currently under development, will include a dedicated chapter on brittle connection failures. While much research has focused on brittle failures in glulam with STS connections, limited studies address CLT. Unlike glulam, CLT has distinct major and minor axes, with each layer supported by layers oriented in different directions. This structure affects the failure modes in CLT, making the identification of actual failure types in STS connections more complex compared to glulam. One of the few available analytical models for brittle failure in STS connections in CLT was developed by Zarnani and Quenneville [10], extending a stiffness-based design approach originally for glulam and LVL to CLT connections [8]. Given the gaps in design codes and research in this specific area, further investigation is crucial to develop reliable design provisions for connections with inclined STS in CLT. 22 3 Experimental investigations 3.1 Overview This chapter presents the experimental investigations on CLT connections with inclined STS under tensile loading. The tests were conducted at the UNBC Wood Innovation Research Laboratory (WIRL) in Prince George, BC. The connections were designed based on the CSA O86-2024 [13] provisions at mean level with the target to observe brittle failure. Specimens were made from CLT panels with various layups, using different STS lengths and placements to evaluate the impact of design parameters on the connection failure modes and load-carrying capacities. 3.2 Materials CLT panels of 3-ply (105 mm), 5-ply (100 mm, 139 mm, and 175 mm), and 7-ply (191 mm and 245 mm) were utilized. Most panels were V2 grade, produced by different manufacturers in accordance with ANSI/APA PRG-320 [42], and made of SPF No.1/No.2 lumber for both longitudinal and transverse layers. The exception was the 5ply 100 mm panel, which was European C24 grade. A summary of the CLT used in this study is presented in Table 3-1. Table 3-1 CLT properties. Thickness (mm) Number of layers Thicknesses of layers (mm) Strength grade Manufacturer Edge glued 105 3 35-35-35 V2 Structurlam ✖ 100 5 20-20-20-20-20 C24 Binderholz ✔ 139 5 35-17-35-17-35 V2 Structurlam ✖ 175 5 35-35-35-35-35 V2 Kalesnikoff ✔ 191 7 35-17-35-17-35-17-35 V2 Structurlam ✖ 245 7 35-35-35-35-35-35-35 V2 Kalesnikoff ✔ 23 The STS were ASSY plus VG carbon steel fully threaded STS with a countersunk (CSK) head and a diameter of 10 mm [43]. As depicted in Figure 3.1, 45º angle washers were utilized alongside four different screw lengths (120 mm, 180 mm, 240 mm, and 280 mm), which were selected during the design phase. Figure 3.1 ASSY plus VG screws and 45º angle washer. 3.3 Specimen description A total of 18 test series, each featuring distinct screw arrangements, lengths, and CLT thicknesses, were considered. The CLT specimen length varied between 1100 mm and 1250 mm, and the width from 400 mm to 600 mm. Identical steel plates were used as side members across all test series, as shown in Figure 3.2. a) b) Figure 3.2 Steel plates, (a) schematic, (b) photo. 24 The STS were installed at a 45º using angle washers on both ends of each CLT member. The specimens were designed to prevent any interaction between the screws installed in the top and bottom halves, as depicted in Figure 3.3. Figure 3.3 Specimen configuration, (a) cross section, (b) top view. Pilot holes were drilled using a jig equipped with a 45º angle washer (Figure 3.4-(a)). To prevent movement of the steel plates during assembly, boards were temporarily screwed on both sides (Figure 3.4-(b)). (a) (b) Figure 3.4 Preparation of specimens, (a) MTC pre-drilling jig equipped with angle washer, (b) wooden boards used as stoppers. 25 The screw spacing parallel to the grain (Sp) and perpendicular to the grain (Sq) followed the minimum spacing requirements outlined in CSA O86-2024, set at 70 mm and 40 mm, respectively. The loaded end distances (aL) varied based on the screw lengths in each series. For series with connections centered on the CLT panel, the unloaded edge distance (ep) depended on the width of the specimens. However, in test series where the connection was offset to a side, this distance was 30 mm to meet the minimum edge distance specified in CSA O86-2024. Figure 3.5 presents a schematic of the specimen configurations in S1 (without offset) and S2 (with offset), illustrating how the STS were placed within the CLT. The schematic of all test series is presented in Appendix D. Also, an overview of the test series is provided in Table 3-2. It should be noted that the moisture content (MC) of each specimen was measured at three points using a moisture meter, and the average value was reported for each test series. (a) (b) Figure 3.5 Schematic of specimens in, a) series S1, (b) series S2. 26 Table 3-2 Test series overview. 3.4 ID # of Tests Position tCLT (mm) LCLT (mm) wCLT (mm) LSTS (mm) nRow nCol aL (mm) eP (mm) MC [%] S1 6 Center 105 1200 400 120 5 5 89 120 8.5 S2 6 Offset 105 1200 400 120 5 5 89 30 8.2 S3 6 Center 100 1200 600 120 5 5 89 220 7.8 S4 6 Offset 100 1200 600 120 5 5 89 30 8.5 S5 6 Center 139 1250 500 180 5 5 112 170 10.1 S6 2 Center 139 1200 500 120 5 5 89 170 8.7 S7 2 Center 139 1200 500 180 3 5 112 170 8.7 S8 6 Offset 139 1200 500 180 5 5 112 30 10.1 S9 6 Center 175 1250 550 240 5 5 128 195 8.8 S10 6 Offset 175 1250 550 240 5 5 128 30 9.4 S11 6 Center 175 1250 550 180 5 5 112 195 9.5 S12 6 Offset 175 1250 550 180 5 5 112 30 10.6 S13 6 Center 175 1090 550 240 3 5 128 195 10.3 S14 6 Offset 175 1080 550 240 3 5 128 30 8.7 S15 6 Center 191 1250 425 240 5 5 128 133 10.6 S16 6 Center 191 1250 425 180 5 5 112 133 9.6 S17 6 Center 245 1100 550 280 5 5 144 195 11.3 S18 6 Offset 245 1100 550 280 5 5 144 30 11.5 Methods Testing was carried out using a load frame and two hydraulic actuators with a combined capacity of 1000 kN, illustrated in Figure 3.6. Figure 3.6 Full test set-up. 27 As shown in Figure 3.7, steel double T brackets served as fixtures, with the bottom bracket anchored to the strong floor and the upper bracket connected to the actuators. (a) (b) Figure 3.7 Fixtures in test set-up(a) schematic, (b) photo. To measure the displacements of the STS connection on both ends of the specimens, four string potentiometers were utilized. These were mounted on a wooden board placed in the center of the CLT, with the other ends attached to the corners of the steel plates, as illustrated in Figure 3.8. Figure 3.8 Test setup and location of string potentiometers. 28 A quasi-static uniaxial tension load was applied according to a modified EN-26891 [44] at a constant displacement rate of 4 mm/min until failure. The tests were stopped when the post-peak load dropped below 80% of the maximum load (Fmax). The monotonic loading protocol included an initial pre-loading phase up to 40% of Fmax followed by loading up to failure, as shown in Figure 3.9. The first test in each series was conducted to validate the estimated Fmax, and the protocol was then adjusted if necessary. Figure 3.9 Loading protocol. 3.5 Analyses In each test, the load carrying capacity (Fmax), and corresponding displacement (d@Fmax) were recorded for the connections on both ends of the specimens. The elastic stiffness (k) was calculated as the secant stiffness between 10% and 40% of Fmax, in accordance with EN 26891 [44], for both connections. The characteristic load-bearing capacities (Fmax,ch) were calculated in accordance with accordance with EN 14358:2016 [45], assuming a normal distribution, using the statistical factor for the 5% quantile with a one-sided 75% confidence level and a sample size twice the number of tests for each test series. Load-displacement curves were extracted for each test series. Additionally, the observed brittle failure modes were documented for each specimen. The possible brittle failure 29 modes are shown in Figure 3.10, including various plug shear modes, as well as step shear and net tension failures. The plug shear failure modes include partial failure of the outer first layer (A), total failure of the first layer (B), failure of the first layer and partial second layer (C), total failure of the first two layers (D), failure of the first and second layers with partial third layer (E), total failure of the first three layers (F), failure of the first three layers and partial fourth layer (G), total failure of the first four layers (H), failure of the first four layers and partial fifth layer (I), and total failure of the first five layers (J). Figure 3.10 Possible brittle failure modes in CLT. 30 3.6 Results 3.6.1 Overview The summary of the experimental results is provided in Table 3-3. Table 3-3 Summary of experimental tests Test ID Fmax (kN) CoV (%) d@Fmax (mm) Fmax,ch (kN) k (kN/mm) Failure mode S1 258 8 4.1 217 88.9 Plug shear Mode D and F, net tension S2 204 20 3.6 121 60.1 S3 300 6 4.8 261 84.7 S4 267 16 5.2 182 85.3 Plug shear Mode F and Mode H. S5 394 12 5.3 301 115.7 Plug shear failure Mode F and Mode J. S6 250 8 5.0 198 117.2 Ductile (withdrawal of STS). S7 339 2 5.2 320 103.2 Ductile (withdrawal of STS). S8 379 15 5.5 267 98.8 S9 458 7 5.1 395 139.3 S10 305 12 4.8 228 120.7 S11 405 12 4.0 304 120.8 S12 328 6 4.6 289 107.9 S13 420 10 6.1 339 114.5 S14 239 15 6.0 168 74.6 S15 485 12 5.3 365 140.5 Slight bending, step shear, net tension S16 475 20 6.5 282 115.7 Plug shear Mode F and Mode H S17 540 4 4.9 495 140.9 Step shear S18 480 14 5.3 346 132.5 Tensile failure of parallel layer and rolling shear, step shear 31 Tensile failure of parallel layer, rolling shear, plug shear Mode D Plug shear Mode B and F, rolling shear, layer separation Plug shear Mode H, tension failure and rolling shear Plug shear Mode J, rolling shear, tension failure of parallel layers, separation Slight bending, tension failure in first layers, rolling shear Plug shear failure Mode F and H, layer separation and rolling shear Plug shear Mode F, Mode H, tensile failure and rolling shear, Plug shear Mode J, Mode H and Mode B Slight bending, tension failure of parallel layers under connection, rolling shear 3.6.2 Series S1 and S2 (105 mm 3-ply CLT with 120 mm screws) Two series were tested on 105 mm 3-ply CLT, using a 5x5 arrangement of 120 mm screws that penetrated the second lamella. In series S1, the connections were centrally placed, while in series S2, they were positioned off-centre. Series S1 exhibited an average loadcarrying capacity of 258 kN and an average elastic stiffness of 89 kN/mm. Brittle failures in series S1 included plug shear Mode D (2 of 6), plug shear Mode F (3 of 6), and net tension failure (1 of 6), as shown in Figure 3.11. In contrast, series S2 exhibited both a lower average load-carrying capacity of 204 kN and a reduced average elastic stiffness of 60.1 kN/mm compared to series S1. The brittle failures observed in series S2 were more complex, involving tensile failure of the parallel layers (2 of 6), rolling shear in most specimens (2 of 6), and plug shear in Mode D (2 of 6), as depicted in Figure 3.12. (a) (b) (c) Figure 3.11 Failure modes in series S1, (a) Mode D, (b) Mode F, (c) net tension. 32 (a) (b) (c) Figure 3.12 Failure modes in series S2, (a) tension failure of parallel layer, (b) plug shear Mode D, (c) rolling shear failure. The load-displacement curves for both test series, measured on both sides, are presented in Figure 3.13 and Figure 3.14. These curves illustrate the applied load and corresponding displacement, representing the movement of the connections at the bottom (b) and top (t), measured between the steel plates and the center of the CLT panel. The sudden drops in the curves clearly indicate the occurrence of brittle failure. 33 S1 - CLT105_120_5x5_center 400 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 300 200 100 0 0 5 10 Displacement (mm) Figure 3.13 Load-displacement curves for series S1. S2 - CLT105_120_5x5_offset 400 1-b 2-b 3-b 300 4-b Load (kN) 5-b 6-b 200 1-t 2-t 3-t 100 4-t 5-t 6-t 0 0 5 10 Displacement (mm) Figure 3.14 Load-displacement curves for series S2. 3.6.3 Series S3 and S4 (100 mm 5-ply CLT with 120 mm screws) Two test series were conducted on 100 mm 5-ply CLT using a 5x5 arrangement of 120 mm screws that extended through the third lamella. In series S3, where the screws were centrally configured, the tests revealed an average load-carrying capacity of 299 kN and an elastic stiffness of 84.7 kN/mm. The observed brittle failures were plug shear in Mode 34 F (3 of 6) and Mode B (2 of 6), along with rolling shear and layer separation, as depicted in Figure 3.15. Series S4 showed a slightly lower average load-carrying capacity of 267 kN and a similar elastic stiffness of 85.3 kN/mm compared to Series S3. The observed brittle failure modes included plug shear in Mode F for most specimens (3 of 6), with some extending through four layers in Mode H (2 of 6), along with rolling shear, as shown in Figure 3.16. The load-displacement curves for these test series are detailed in Figure 3.17 and Figure 3.18. (a) (b) Figure 3.15 Failure modes in series S3, (a) Mode B, (b) Mode F. (a) (b) Figure 3.16 Failure modes in series S4, (a) Mode F, (b) Mode H. 35 S3 - CLT100_120_5x5_center 400 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 300 200 100 0 0 5 10 Displacement (mm) Figure 3.17 Load-displacement curves for series S3. S4 - CLT100_120_5x5_offset 400 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 300 200 100 0 0 5 10 Displacement (mm) Figure 3.18 Load-displacement curves for series S4. 3.6.4 Series S5- S8 (139 mm 5-ply CLT with 120 and 180 mm screws) Four test series were conducted on 139 mm 5-ply CLT, including S5 and S8 with 5x5 centered and offset arrangements of 180 mm screws, S7 with a 3x5 centered arrangement of 180 mm screws, and S6 with a 5x5 centered layout of 120 mm screws. In S6, screws penetrated the third lamella, while in the other series, they reached the fourth. Series S5 36 showed an average load-carrying capacity of 394 kN and an average elastic stiffness of 115.7 kN/mm. The predominant brittle failure observed in most specimens was Mode F (5 of 6). In one test, the plug shear failure extended through the 5th layer of the CLT, resulting in a brittle failure classified as Mode J, as shown in Figure 3.19. (a) (b) Figure 3.19 Failure modes in series S5, (a) plug shear Mode F, (b) Mode J. In series S6 and S7, the shorter screws and fewer fasteners led to withdrawal failure before brittle failure of CLT, resulting in the tests being stopped after two replicates showed consistent withdrawal failure. For series S8, all replicates exhibited brittle failure modes, with an average load-carrying capacity of 379 kN and an average elastic stiffness of 98.8 kN/mm. In this series, plug shear failure, predominantly Mode H (4 of 6), was observed across the specimens. Additionally, due to slight bending near the top connection, tensile failure of layers and rolling shear were also noted, as depicted in Figure 3.20. (a) (b) Figure 3.20 Failure modes in series S8, (a) Mode H, (b) tensile failure due to bending. 37 The load-displacement curves of series S5 to S8 are shown in Figure 3.21 to Figure 3.24. For S6 and S7, which did not exhibit brittle failure, the ductile nature of the failure is evident in their respective load-displacement curves. S5 - CLT139_180_5x5_center 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 500 Load (kN) 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.21 Load-displacement curves for series S5. S6 - CLT139_120_5x5_center 400 1-b Load (kN) 300 2-b 200 1-t 100 2-t 0 0 5 Displacement (mm) Figure 3.22 Load-displacement curves for series S6. 38 10 S7 - CLT139_180_3x5_center 400 1-b Load (kN) 300 2-b 200 1-t 100 2-t 0 0 5 10 Displacement (mm) Figure 3.23 Load-displacement curves for series S7. S8 - CLT139_180_5x5_offset 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 500 Load (kN) 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.24 Load-displacement curves for series S8. 3.6.5 Series S9-S14 (175 mm 5-ply CLT with 180 and 240 mm screws) Six test series were conducted on 175 mm 5-ply CLT. These included S9 and S10 with 5x5 centered and offset arrangements of 240 mm screws, S11 and S12 with similar arrangements of 180 mm screws, and S13 and S14 with 3x5 layouts of 240 mm screws, 39 both centered and offset. It should be noted that the 240 mm screws penetrated the fourth layer of CLT, while the 180 mm screws penetrated the third layer. In series S9, the average load-carrying capacity was 458 kN, with an average elastic stiffness of 139.3 kN/mm. Brittle failure occurred through the entire thickness of the CLT for most specimens (3 of 6), indicating Mode J of plug shear failure. Additionally, rolling shear, layer separation, and tension failure of the first layer were observed in the tested specimens, as shown in Figure 3.25. In series S10, which used the same connection arrangement as S9 but with an offset, the load-carrying capacity decreased to 305 kN, as expected, with a corresponding reduction in elastic stiffness to 120.7 kN/mm. Slight bending was observed in most tests, resulting in tension failure and rolling shear being the predominant failure modes in the tested specimens, depicted in Figure 3.26. (a) (b) Figure 3.25 Failure modes in series S9, (a) Mode J, (b) layer separation and rolling shear. Figure 3.26 Failure modes in series S10. 40 The load-displacement curves for the connections on both sides in series S9 and S10 are presented in Figure 3.27 and Figure 3.28, respectively. S9 - CLT175_240_5x5_center 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 500 Load (kN) 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.27 Load-displacement curves for series S9. S10 - CLT175_240_5x5_offset 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 500 Load (kN) 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.28 Load-displacement curves for series S10. In series S11, the average load-carrying capacity was 405 kN, with an elastic stiffness of 120.8 kN/mm. In most replicates, clear plug shear failure in Mode F was observed (4 of 41 6). In one replicate, the failure extended through the fourth layer, indicating Mode H, as shown in Figure 3.29. Additionally, layer separation and rolling shear were also observed. In series S12, which used the same connection arrangement as S11 but with an offset, the load-carrying capacity and elastic stiffness were lower than in S11, at 328 kN and 107.9 kN/mm, respectively. As depicted in Figure 3.30, the primary brittle failure mode observed was Mode F (3 of 6), indicating plug shear failure in the first three layers of CLT. However, one test exhibited brittle failure extending through four layers, resulting in failure Mode H. Additionally, layer separation and rolling shear were also observed in Series 12. (a) (b) Figure 3.29 Failure modes in series S11, (a) Mode F, (b) Mode H. (a) (b) Figure 3.30 Failure modes in series S12, (a) Mode F, (b) Mode H. The load-displacement curves for the connections on both sides in series S11 and S12 are presented in Figure 3.31 and Figure 3.32, respectively. 42 S11 - CLT175_180_5x5_center 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 500 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.31 Load-displacement curves for series S11. S12 - CLT175_180_5x5_offset 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 500 Load (kN) 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.32 Load-displacement curves for series S12. For series S13, which featured a narrower centered arrangement of STS, the average loadcarrying capacity was 420 kN, with an elastic stiffness of 114.5 kN/mm. Meanwhile, series S14, which had an offset arrangement, exhibited a significantly lower load-carrying capacity of 239 kN and an elastic stiffness of 74.6 kN/mm. In Series S13, the predominant brittle failure mode was plug shear, propagating through five layers of CLT, indicating 43 Mode J in most specimens (3 of 6). Mode B, Mode H, and rolling shear failure were each observed once in this series, as shown in Figure 3.33. In series S14, despite the use of lateral support during offset tests, slight bending of the specimens was still observed. This resulted in tension failure in parallel layers near the connection (Figure 3.34), rolling shear, and layer separation being the most representative failure modes, rather than a clear plug shear failure. (a) (b) (c) Figure 3.33 Failure modes in series S13, (a) Mode B, (b) Mode H, (c) Mode J. Figure 3.34 Tension failure and rolling shear in series S14. 44 S13 - CLT175_240_3x5_center 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 500 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.35 Load-displacement curves for series S13. S14 - CLT175_240_3x5_offset 400 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 300 200 100 0 0 5 10 Displacement (mm) Figure 3.36 Load-displacement curves for series S14. 3.6.6 Series S15 and S16 (191 mm 7-ply CLT with 180 and 240 mm screws) Two test series were conducted on 191 mm 7-ply CLT: S15 and S16, featuring 5x5 centered arrangements of 240 mm and 180 mm screws, penetrating the 5th and 4th layers of the CLT, respectively. Series S15 demonstrated an average load-carrying capacity of 485 kN and a elastic stiffness of 140.5 kN/mm. Despite lateral support, slight bending of 45 the specimens was observed in S15, leading to tensile failure in the parallel layers and rolling shear. Due to the narrow width of the CLT panels, brittle failure mostly occurred across the entire width of the specimens. In tests where failure penetrated only a portion of the panel's depth, step shear was observed (2 of 6). However, in tests where failure extended through the entire thickness of the panel, signs of net tension failure were noted (2 of 6). Figure 3.37 illustrates the brittle failure observed in series S15. In Series S16, the average load-carrying capacity was 475 kN, slightly lower than that of S15, with a reduced elastic stiffness of 115.7 kN/mm. The observed failure modes included plug shear failure in Mode F (3 of 6) and plug shear failure through four layers in Mode H (3 of 6), as shown in Figure 3.38. The load-displacement curves for the connections on both sides in series S15 and S16 are illustrated in Figure 3.39 and Figure 3.40, respectively. (a) (b) (c) Figure 3.37 Failure modes in series S15, (a) tension failure due to bending, (b) step shear, (c) net tension effect. 46 (a) (b) Figure 3.38 Failure modes in series S16, (a) Mode F, (b) Mode H. S15 - CLT191_240_5x5_center 600 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t Load (kN) 500 400 300 200 100 0 0 5 Displacement (mm) 10 Figure 3.39 Load-displacement curves for series S15. S16 - CLT191_180_5x5_center 700 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 600 Load (kN) 500 400 300 200 100 0 0 5 Displacement (mm) Figure 3.40 Load-displacement curves for series S16. 47 10 3.6.7 Series S17 and S18 (245 mm 7-ply CLT with 280 mm screws) Two test series were conducted on 245 mm 7-ply CLT: S17 and S18, featuring 5x5 centered and offset arrangements of 280 mm and 180 mm screws, respectively, which penetrated the 5th layer of the CLT specimens. For series S17, the average load-carrying capacity was 540 kN with an average elastic stiffness of 140.9 kN/mm. Aside from rolling shear failure in one specimen, brittle failure in most specimens (5 of 6) was characterized by failure extending across the entire width to a certain depth, indicating clear step shear, as shown in Figure 3.41. For series S18, as expected, the average load-carrying capacity and elastic stiffness were slightly lower than those of S17, measuring 480 kN and 132 kN/mm, respectively. The most common failure modes observed in the replicates included tension failure of the CLT layers, rolling shear, and delamination, leading to failure patterns like step shear, as depicted in Figure 3.42. The load-displacement curves for the connections on both sides in series S17 and S18 are illustrated in Figure 3.43 and Figure 3.44, respectively. Figure 3.41 Step shear failure observed in series S17. Figure 3.42 Step shear failure observed in series S18. 48 S17 - CLT245_280_5x5_center 700 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 600 Load (kN) 500 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.43 Load-displacement curves for series S17. S18 - CLT245_280_5x5_offset 700 1-b 2-b 3-b 4-b 5-b 6-b 1-t 2-t 3-t 4-t 5-t 6-t 600 Load (kN) 500 400 300 200 100 0 0 5 10 Displacement (mm) Figure 3.44 Load-displacement curves for series S18. 3.7 Evaluation of connection parameters The effects of connection design parameters on the brittle failure mode and resistance of CLT were examined, with particular attention to: i) unloaded edge distance, ii) connection width, and iii) penetration depth of screws in parallel layers. 49 CSA O86-2024 [13] specifies a minimum unloaded edge distance (ep) but does not address the impact of exceeding this requirement. In this testing campaign, the effect of changing ep was investigated by shifting the connection from the center to the edge of the CLT panel. For test series with centered connections, ep values varied due to differing specimen widths and screw arrangements, while all offset test series had a fixed ep of 30 mm, consistent with the CSA O86 minimum requirement (three times the fastener diameter). Despite having identical connection and equal resistance plane areas, as shown in Figure 3.45, specimens with offset connections (depicted in red) exhibited lower loadcarrying capacities compared to those with centered connections (depicted in gray). On average, shifting the connection from the center to the edge reduced load-carrying capacities by 20%, with specific reductions of 21%, 11%, 4%, 33%, 19%, 43%, and 11% for Series S1 to S2, S3 to S4, S5 to S8, S9 to S10, S11 to S12, S13 to S14, and S17 to S18, respectively. 600 S17 S9 Load-carrying capacity (kN) 500 S5 S12 S10 S3 S13 S11 S8 400 300 S18 S4 S1 S14 S2 200 100 0 120 30 220 30 170 30 195 30 195 30 195 30 195 Unloaded edge distance (mm) Figure 3.45 Effect of unloaded edge distance on load-carrying capacities. 50 30 Similarly, the elastic stiffness of the connections decreased by an average of 16% from the center to the offset test series. Except for Series S3 to S4, which maintained the same stiffness (Figure 3.46), reductions were observed across Series S1 to S2, S5 to S8, S9 to S10, S11 to S12, S13 to S14, and S17 to S18, with decreases of 32%, 15%, 13%, 11%, 35%, and 6%, respectively. 175 S17 S9 S10 150 S5 Elastic stiffness (kN/mm) S18 S11 S13 S12 125 S1 S3 100 S8 S4 S14 S2 75 50 25 0 120 30 220 30 170 30 195 30 195 30 195 30 195 30 Unloaded edge distance (mm) Figure 3.46 Effect of unloaded edge distance on elastic stiffness of connections. CSA O86-2024 [13] provides minimum spacing requirements and accounts for connection width based on the area of the bottom shear and head tensile failure planes. In this testing campaign, the effect of reducing these failure planes was investigated by using a narrower screw arrangement. To assess the impact, 5x5 and 3x5 layouts with identical screw lengths and positions were compared, resulting in a 50% reduction of head tensile and shear plane areas. The effect on load-carrying capacity is illustrated in Figure 3.47. The results showed an average decrease in load-carrying capacity of 15%, with reductions of 14% from S5 to S7 (centered 180 mm screws), 8% from S9 to S13 (centered 240 mm screws), and 22% from S10 to S14 (offset 240 mm screws). A similar trend observed for 51 elastic stiffness, as shown in Figure 3.48, with an average reduction of 22% and specific decreases of 11%, 18%, and 38% across these comparisons. This decrease was expected, as narrower screw arrangements have fewer fasteners, leading to reduced withdrawal resistance. Additionally, the smaller area of the head tensile resistance plane and plug shear resistance contributed to the reduced capacity in narrower configurations. Load-carrying capacity (kN) 600 S9 500 S13 S5 400 S7 S10 300 S14 200 100 0 160 80 160 80 Connection width (mm) 160 80 Figure 3.47 Effect of connection width on load-carrying capacities. 175 S9 Elastic stiffness (kN/mm) 150 S10 S5 125 S13 S7 100 S14 75 50 25 0 160 80 160 80 Connection width (mm) 160 80 Figure 3.48 Effect of connection width on elastic stiffness of connections. 52 In the testing campaign, different CLT layups resulted in varying portions of parallel and transverse layers being engaged for the same screw penetration depths. Therefore, the effect of screw penetration depth into the stronger parallel layers of CLT on the loadcarrying capacity and elastic stiffness of the connections was examined. Figure 3.49 illustrates the load-carrying capacities as a function of penetration depth in the parallel layers. The overall trend shows that increasing penetration depth generally leads to an increase in load-carrying capacity, although some inconsistencies are noted. For instance, from S1 to S11 and S3 to S17, increasing the penetration depth from 35 mm to 59 mm and 32 mm to 95 mm resulted in load-carrying capacity increases of 57% and 80%, respectively. However, for S11 to S5 and S17 to S15, increasing the penetration depth from 59 mm to 70 mm and 95 mm to 105 mm led to a slight reduction in loadcarrying capacity by approximately 3% and 10%, respectively. These variations could be attributed to factors such as different penetration depth in transverse layers. Despite these outliers, the general trend indicates that deeper penetration typically enhances loadcarrying capacity. Figure 3.50 shows the elastic stiffness of the connections relative to penetration depth. Similar to the load-carrying capacity results, elastic stiffness tends to increase with penetration depth, though again with some inconsistencies. For example, from S1 to S11 and S3 to S17, elastic stiffness increased by 36% and 66%, respectively. Conversely, a minor reduction of approximately 4% was observed between S11 and S5, despite the increased penetration depth. These fluctuations in stiffness could similarly be related to variations in penetration depth in transverse layers. Overall, deeper penetration into parallel layers generally results in higher elastic stiffness, consistent with the trends observed for load-carrying capacity. 53 600 S17 S15 70 95 103 S16 500 Load-carrying capacity (kN) S9 S11 S5 400 S3 300 S1 S6 200 100 0 32 35 35 59 70 70 Penetration depth in parallel layers (mm) Figure 3.49 Effect of penetration depth into the parallel layers on load-carrying capacities. 175 S17 S16 S6 150 S11 Elastic stiffness (kN/mm) S15 S9 S5 125 100 S3 S1 75 50 25 0 32 35 35 59 70 70 70 95 103 Penetration depth in parallel layers (mm) Figure 3.50 Effect of penetration depth into the parallel layers on elastic stiffness of connections. 54 4 Analytical investigation of experimental campaigns 4.1 Overview This chapter presents an analytical investigation of both previous and current experimental tests on the brittle failure of CLT with STS connections. The existing design provisions for predicting brittle failure in CLT with STS connections were originally developed for glulam and solid lumber, and their application to CLT remains unclear, often requiring significant simplifications that may reduce accuracy. After summarizing the previous experimental campaign, this chapter examines the predictive ability of the analytical approaches from CSA O86-2024 [13], prEN 1995 (2024) [29], and Zarnani and Quenneville [10] in detail to identify their limitations in predicting brittle failure in CLT. 4.2 Previous experimental campaigns 4.2.1 Azinović et al. (2022) The study by Azinović et al. [15] consisted of connections with a single steel plate attached with 60 screws, installed at a 90º to both ends of a CLT member. The number of screw rows, columns, and their spacing were consistently applied, using Ø8 mm fully threaded STS with lengths ranging from 40 mm to 100 mm. The CLT was made from softwood boards of strength class C24, consisting of 3 or 5 layers, with layer thicknesses ranging from 20 mm to 40 mm. The details are provided in Table 4-1. Table 4-2 summarizes the experimental results of tests conducted by Azinović et al. [15]. All specimens experienced brittle failure, primarily through plug shear. Some cases revealed a secondary failure mode, such as plug shear combined with row shear or step shear, the latter for specimens with a perpendicular grain orientation. Characteristic loadcarrying capacities were determined following 14358:2016 [45], assuming a normal 55 distribution. Although plastic deformation of screws was observed in most test series, the specimens ultimately failed in a brittle manner. Table 4-1 Test series overview in the experiments by Azinović et al.[15]. Edge glued Test ID n ‫ݓ‬େ୐୘ (mm) ‫ݐ‬େ୐୘ (mm) ‫ܮ‬େ୐୘ (mm) ܵ୯ ݁୮ ‫ܮ‬ୗୡ୰ୣ୵ ݊ୖ୭୵ ݊ୡ୭୪ ܵ୮ ܽ୐ (mm) (mm) (mm) (mm) (mm) (mm) (mm) AB1 6 245 142 1228 40 6 10 32 16 48 34 40-20-20-20-40 Yes AB2 6 250 142 1498 40 6 10 32 16 48 37 33-20-33-20-33 No CD1 7 250 101 988 40 6 10 32 16 48 37 20-20-20-20-20 No CD2 7 251 101 1201 60 6 10 32 16 48 37 30-40-30 Yes CD2-90 6 248 101 1200 60 6 10 32 16 48 37 30-40-30 Yes CD3 6 500 101 1199 60 6 10 32 16 48 162 30-40-30 Yes CD3-90 3 500 101 1200 60 6 10 32 16 48 162 30-40-30 Yes 16 48 283 30-40-30 Yes CLT layup CD4 6 743 101 1198 60 6 10 32 CD5 6 250 140 1200 60 6 10 32 16 48 35 33-20-33-20-33 No CD5-90 8 250 140 1200 60 6 10 32 16 48 35 33-20-33-20-33 No EF1 4 250 100 1000 100 6 10 32 16 48 35 20-20-20-20-20 No EF2 6 250 140 1200 100 6 10 32 16 48 35 33-20-33-20-33 No EF3 3 500 140 1200 100 6 10 32 16 48 165 33-20-33-20-33 No Table 4-2 Results of test series in experiments by Azinović et al. [15]. Test ID ‫ݓ‬େ୐୘ (mm) ‫ݐ‬େ୐୘ (mm) ‫ܮ‬ୗୡ୰ୣ୵ (mm) ‫୲ܨ‬,୫ୣୟ୬ (kN) CoV ‫୲ܨ‬,୩ (kN) Failure mode AB1 245 142 40 261 9% 201 Row/Plug shear AB2 250 142 40 237 9% 182 Row/Plug shear CD1 250 101 40 190 4% 104 Plug shear CD2 251 101 60 247 7% 174 Plug shear CD2-90 248 101 60 143 11% 117 Plug/Step shear CD3 500 101 60 305 8% 230 Row/Plug shear CD3-90 500 101 60 197 29% 181 Plug/Step shear CD4 743 101 60 358 9% 274 Row/Plug shear CD5 250 140 60 224 6% 143 Row/Plug shear CD5-90 250 140 60 146 9% 113 Plug/Step shear EF1 250 100 100 225 5% 124 Tension failure EF2 250 140 100 269 9% 205 Row/Plug shear EF3 500 140 100 287 12% 230 Row/Plug shear 56 4.2.2 Ni and Niederwestberg (2022) The study conducted by Ni and Niederwestberg [40] included nine test series of CLT loaded parallel to the grain with steel plates on both sides at one end and a large, bolted connection at the opposite end. The specimens were made of non-edge-glued, V2M1.1 graded 5-ply CLT. Test series were conducted using Ø8 mm and Ø10 mm partially threaded STS ranging from 80 mm to 120 mm in length, installed at a 90-degree angle. The CLT panels varied in width (140 mm and 420 mm). The main difference between the test series with identical configurations and screw arrangements was whether the CLT members beneath the connection had full lamination or a central gap in the connection area. Detailed information about these tested series is provided in Table 4-3. Table 4-3 Test series overview in the experiments by Ni and Niederwestberg [40]. Test ID n ‫ݓ‬େ୐୘ (mm) ‫ݐ‬େ୐୘ (mm) ܵ୯ ݁୮ ‫ܮ‬େ୐୘ ݀୤ ‫ܮ‬ୗୡ୰ୣ୵ ݊ୖ୭୵ ݊େ୭୪ ܵ୮ ܽ୐ (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) CA4 3 420 175 588 8 80 5 3 40 24 96 20.5 35-35-35-35-35 CA5 3 420 175 588 8 80 5 3 40 24 96 20.5 35-35-35-35-35 CA6 3 140 175 588 8 80 5 3 40 24 96 20.5 35-35-35-35-35 CA10 3 420 175 588 8 120 5 3 40 24 96 20.5 35-35-35-35-35 CA11 3 420 175 588 8 120 5 3 40 24 96 20.5 35-35-35-35-35 CA12 3 140 175 588 8 120 5 3 40 24 96 20.5 35-35-35-35-35 CA16 3 420 175 588 12 100 3 3 60 36 144 34 35-35-35-35-35 CA17 3 420 175 588 12 100 3 3 60 36 144 34 35-35-35-35-35 CA18 3 140 175 588 12 100 3 3 60 36 144 34 35-35-35-35-35 CLT layup The results by Ni and Niederwestberg [40] are summarized in Table 4-4. Most specimens exhibited yielding of fasteners before experiencing brittle failure. The specimens which showed significant deformation due to fastener yielding, were categorized as ductile failure. As a result, the brittle failure modes were considered secondary failure modes. Narrower specimens primarily experienced step shear as the brittle failure mode. 57 Characteristic load-carrying capacities were determined following EN 14358 [45], assuming a normal distribution. However, due to the low number of replicates (3 replicates per series), these characteristic values are highly penalized. Table 4-4 Results of test series in experiments by Ni and Niederwestberg [40]. Test ID ‫ݓ‬େ୐୘ (mm) ‫ܮ‬ୗୡ୰ୣ୵ (mm) ݊ୖ୭୵ (mm) ݊େ୭୪ (mm) ‫୲ܨ‬,୫ୣୟ୬ (kN) CoV ‫୲ܨ‬,୩ (kN) Failure mode CA4 420 80 5 3 213 4% 185 Yielding CA5 420 80 5 3 220 5% 185 Plug Shear CA6 140 80 5 3 203 5% 173 Step Shear CA10 420 120 5 3 309 1% 301 Yielding CA11 420 120 5 3 297 11% 195 Plug Shear CA12 140 120 5 3 228 16% 111 Step Shear CA16 420 100 3 3 236 4% 205 Yielding CA17 420 100 3 3 254 8% 192 Yielding CA18 140 100 3 3 253 9% 184 Yielding 4.2.3 Previous work at UNBC (2022) The dataset from previous work at UNBC [41] includes 18 arrangements of screws, installed at a 45º angle, with a single steel plate (single shear plane) attached to both ends of a CLT member. Detailed information is provided in Table 4-5. In total, 82 tests were performed using Ø10 mm fully threaded STS ranging from 100 mm to 160 mm in length installed in 3-layer and 5-layer V2 graded CLT. Most specimens exhibited brittle failure, except for series S1, S2 and S7, where screw withdrawal occurred. The lack of lateral support during testing caused bending in narrower specimens, possibly leading to premature failures. The results and observed brittle failure details from the previous tests conducted at UNBC [41] are presented in Table 4-6. 58 Table 4-5 Test series overview in the previous experiments at UNBC [41]. Test ID n ‫ݓ‬େ୐୘ (mm) ‫ݐ‬େ୐୘ (mm) ‫ܮ‬େ୐୘ Axis (mm) ݀୤ (mm) ܵ୯ ܽ୐ ‫ܮ‬ୗୡ୰ୣ୵ ݊ୖ୭୵ ݊େ୭୪ ܵ୮ CLT layup (mm) (mm) (mm) (mm) (mm) (mm) Edge glued S-1 2 300 100 300 Major 10 100 3 4 40 30 50 20-20-20-20-20 Yes S-2 2 300 100 300 Major 10 100 6 3 40 30 50 20-20-20-20-20 Yes S-3 2 300 100 300 Major 10 100 6 6 40 30 50 20-20-20-20-20 Yes S-4 2 300 100 300 Major 10 140 3 4 40 30 50 20-20-20-20-20 Yes S-5 2 300 100 300 Major 10 140 6 3 40 30 50 20-20-20-20-20 Yes S-6 7 300 100 300 Major 10 140 6 6 40 30 50 20-20-20-20-20 Yes S-6W 6 600 100 600 Major 10 140 6 6 40 30 50 20-20-20-20-20 Yes S-7 2 300 100 300 Minor 10 100 3 3 30 40 50 20-20-20-20-20 Yes S-8 3 300 100 300 Minor 10 100 6 3 30 40 50 20-20-20-20-20 Yes S-9 7 300 100 300 Minor 10 120 6 3 30 40 50 20-20-20-20-20 Yes S-9W 6 600 100 600 Minor 10 120 6 3 30 40 50 20-20-20-20-20 Yes S-10 7 300 139 300 Major 10 140 6 6 40 30 50 35-17-35-17-35 No S-10W 6 600 139 600 Major 10 140 6 6 40 30 50 35-17-35-17-35 No S-11 2 300 139 300 Major 10 160 6 5 40 30 50 35-17-35-17-35 No S-12 2 300 139 300 Minor 10 140 2 3 30 40 50 35-17-35-17-35 No S-13 2 300 139 300 Minor 10 140 3 3 30 40 50 35-17-35-17-35 No S-14 6 300 139 300 Minor 10 140 6 3 30 40 50 35-17-35-17-35 No S-15 2 300 139 300 Minor 10 160 6 2 30 40 50 35-17-35-17-35 No S-16 1 300 105 300 Major 10 120 3 5 40 30 50 35-35-35 No S-17 6 300 105 300 Major 10 120 6 3 40 30 50 35-35-35 No S-18 7 300 105 300 Major 10 120 6 6 40 30 50 35-35-35 No 59 Table 4-6 Results of test series in previous experiments at UNBC [41]. Test ID ‫୲ܨ‬,୫ୣୟ୬ (kN) CoV ‫୲ܨ‬,୩ (kN) Details of observed failure modes S-1 - - - Ductile failure (withdrawal of STS). S-2 - - - Ductile failure (withdrawal of STS). S-3 218 4% 202 Rolling shear, tension failure in first lamella. S-4 202 6% 169 Rolling shear, tension failure in first lamella, delamination between lamellas. Row shear effect in just first lamella. S-5 241 8% 189 Some plug shear effects, bending at end of panel. S-6 226 24% 116 Some plug shear effect, tension failure in first lamella, bending at end of panel. S-6W 336 10% 264 Clear plug shear. S-7 - - - Ductile failure (withdrawal of STS). S-8 148 3% 137 Clear step shear. S-9 142 20% 85 Partial step shear, bending at end of panel. S-9W 144 6% 129 Rolling shear, tension crack. S-10 385 16% 263 Some plug shear. Tension failure in first and third lamella. S-10W 468 7% 390 Clear plug shear. S-11 359 3% 335 Clear plug shear. S-12 68 19% 33 S-13 74 3% 69 S-14 105 16% 71 S-15 85 21% 36 S-16 74 - - Rolling shear, delamination between lamellas. S-17 127 23% 68 Clear plug shear, bending at end of panel. S-18 188 15% 129 Clear plug shear, bending at end of panel 4.3 Step shear effect, rolling shear, delamination between lamellas, bending at end of panel. Step shear effect, rolling shear, delamination between lamellas, bending at end of panel. Step shear effect, rolling shear, delamination between lamellas, bending at end of panel. Rolling shear, tension in longitudinal lamella, bending at end of panel. Application of prediction models The differences of the design approaches from CSA O86 (2024) [13], prEN 1995 (2024) [29], and Zarnani and Quenneville [10] for calculating plug shear failure in CLT are briefly explained in this section and summarized in Table 4-7, with further details provided in Appendices A to C. 60 Table 4-7 Different approaches for calculation of plug shear failure [10], [13], [29] Strength of plane Analytical model Failure Head tensile (H) Lateral shear (L) Bottom shear (B) CSA O86 (2024) L+H+B 1.25݂௧ ܾ௧ ‫ݐ‬௘௙ 1.5݂௩ ‫ܮ‬௦,௜ ‫ݐ‬௘௙ 0.75 ݂௩ ‫ܣ‬௉,௦௕ or 0.75 ݂௩,௥ ‫ܣ‬௉,௦௕ ݇௖௟ ቌ prEN 1995 (2024) Maximum of L or H+B ݇୲ ܾ୬ୣ୲ ‫ݐ‬௘௙ ݂௧,଴,ୢ ݇୴ ‫ݐ‬ୣ୤ ݈ୡ୭୬ ݂୴,ୢ ݇୴ ݈ୡ୭୬ ܾ௡௘௧ ݂୴,ୢ 7‫ݐ‬௛ ߙ௖௟ ቌ ‫ ݐ‬ቍ ≤ ‫ݐ‬௛ 3+ ௛ ݀ Head tensile (H) Adjacent plane (A) Bottom shear (B) ∑‫ܭ‬ ݂௩ ‫ܣ‬௩௕ ‫ܭ‬ௗ or ∑‫ܭ‬ ݂௩,௥ ‫ܣ‬௩௕ ‫ܭ‬ௗ Stiffness model Minimum of H, A or B ݂௧ ‫ܣ‬௧௛ ∑‫ܭ‬ ‫ܭ‬௛ ݂௩,௥ ‫ܣ‬௩௔ ∑‫ܭ‬ ‫ܭ‬௔ Effective depth 7‫ݐ‬௛ ‫ ݐ‬ቍ ≤ ‫ݐ‬௛ 3+ ௛ ݀ Elastic or post-yield effective depth The CSA O86-2024 provisions consider plug shear failure and the corresponding resistance planes based on work done on glulam, as shown in Figure 4.1. CSA O86 provisions exclude the transverse layer from the effective depth of failure block when calculating head plane resistance, considering only the parallel layers. However, the brittle failure in CLT is more complex as described in Figure 3.10. For side shear resistance, explicit guidance is not provided. Therefore, herein, the effective depth is divided into two parts: the section of layers oriented parallel to the load, with their resistance calculated using the parallel to grain shear strength, and the transverse layers, with their resistance calculated using the rolling shear strength. This method accounts for the contribution of both layers in resistance of side shear plane. To calculate bottom plane resistance according to CSA O86-2024, shear resistance should be determined based on the orientation of the CLT layers in the bottom plane—either parallel or perpendicular to the grain. 61 Figure 4.1 Definition of plug shear failure based on CSA O86-2024. In the calculation according to prEN 1995 (2024), which was developed for glulam and solid lumber as well, no guidance is provided for CLT. As a result, similar to the approach for CSA O86, the portion of the effective depth within transverse layers was disregarded when calculating the head tensile plane for CLT. For side shear plane resistance, shear strength parallel to the grain was applied to the portion in parallel layers and rolling shear strength to the portion in transverse layers. Similarly, the bottom shear layer was calculated using the same approach as CSA O86, based on the orientation of the bottom plane. Zarnani and Quenneville's model uses a stiffness-based approach to assess the contribution of each failure plane in resisting applied loads, substituting lateral plane resistance with adjacent plane resistance in its calculations. CSA O86-2024 and prEN 1995 (2024) utilize the same formula to calculate the effective depth of the failure block. This contrasts with the stiffness model, where the effective depth is based on the elastic and plastic behavior of fasteners and their yielding modes. To ensure consistency between different models, the same effective depth equation was applied to the stiffness model. 62 4.4 Prediction based on CSA O86-2024 CSA O86-2024 [13] provisions for STS connections were evaluated by comparing the load-carrying capacity from previous [15], [40], [41] and current experimental campaigns with the predicted load-carrying capacity based on brittle failure resistance of connection at the mean level. This was done using the mean material properties (Table 4-8), and the non-factored resistance for each test series. Table 4-8 Mean properties of lumber used in CLT products [45] [46]. Strength class ߩ (kg/m3) ‫ܧ‬଴ (MPa) ‫ܩ‬଴ (MPa) ‫ܩ‬୰ (MPa) C24 420 11500 650 V2 445 9500 594 ݂୲,ଽ଴ (MPa) ݂୴ (MPa) 65.0 ݂୲,଴ (MPa) 32.5 1.1 5.5 ݂୴,୰ (MPa) 1.9 59.4 21.4 2.6 5.87 2.0 The predictions based on CSA O86-2024 provisions are presented in Table 4-9 for test series with STS installed at 90º, and in Table 4-10 for test series installed with STS at 45º. For STS installed at 90º, the predicted lateral yielding resistance was lower than the brittle failure resistance for all series and brittle failure modes occurred after fastener yielding. A comparison of the predicted and experimental data for series with STS connections at 90º showed that, for test series exhibiting brittle failure, the predicted failure modes matched the observed modes in all cases except for Series EF1, which failed in net tension. The ratio of experimental to predicted load-carrying capacity at the mean level ranged from 0.8 to 1.9, with an average ratio of 1.2. For STS installed at 45º, more than half of the predicted failure modes matched the observed ones. Discrepancies were primarily found in earlier test data [41], where less than half of the test series matched the predicted failure modes. However, in the new experimental tests conducted in this thesis, 16 out of 18 test series showed alignment between the predicted and observed failure modes. The comparison of experimental and predicted load-carrying capacities for each 63 test series that failed in a brittle manner at the mean level showed the ratio of experimental to predicted results ranging from 0.5 to 1.8, with an average ratio of 1.0. Table 4-9 CSA O86 predictions for test series with STS at a 90º. ID Observed Prediction Fexp/Fpre Fexp (kN) Failure mode Fpre (kN) Failure mode teff (mm) AB1 AB2 261 237 RS/PS RS/PS 220 220 PS PS 24 24 1.2 1.1 CD1 CD2 CD2-90 190 247 143 PS PS PS/SS 137 187 164 PS PS PS 24 36 36 1.4 1.3 0.9 CD3 CD3-90 CD4 CD5 CD5-90 305 197 358 224 146 RS/PS PS/SS RS/PS RS/PS PS/SS 187 164 187 198 152 PS PS PS PS PS 36 36 36 36 36 1.6 1.2 1.9 1.1 1.0 EF1 EF2 EF3 CA4 CA5 CA6 CA10 CA11 CA12 CA16 CA17 CA18 225 269 287 213 220 203 309 297 228 236 254 253 NT RS/PS RS/PS EYM PS SS EYM PS SS EYM EYM EYM 237 205 205 251 251 222 255 255 222 222 222 222 PS PS PS PS PS SS PS PS SS PS PS PS 44 44 44 40 40 40 45 45 45 57 57 57 0.9 1.3 1.4 0.8 0.9 0.9 1.2 1.2 1.0 1.1 1.1 1.1 Notation: EYM=European yield model; NT=Net tension; SS=Step shear; RS=Row shear; PS=Plug shear After comparing the CSA O86 predictions with the experimental results, it was found that the load-carrying capacities for test series CD3, EF2, S-9, and S-10 were lower than their counterparts CD4, EF3, S-9W, and S-10W, which had identical layouts but were conducted on wider CLT panels. The wider panels exhibited higher load-carrying capacities, yet the CSA O86 predictions remained identical for both narrower and wider panels, indicating that the provisions should account for CLT panel width in similar layouts. 64 Table 4-10 CSA O86 predictions for test series with STS at a 45º. ID Observed Prediction Fexp/Fpre Fexp (kN) Failure mode Fpre (kN) Failure mode teff (mm) S-3 218 TF 216 PS 38 1.0 S-4 202 TF 165 PS 48 1.2 S-5 241 PS 254 PS 48 0.9 S-6 226 PS 308 NT 48 0.7 S-6W 336 PS 418 PS 48 0.8 S-8 148 SS 205 PS 38 0.7 S-9 142 SS 143 PS 44 1.0 S-9W 144 TF 143 PS 44 1.0 S-10 385 PS 304 PS 48 1.3 S-10W 468 PS 304 PS 48 1.5 S-11 359 PS 274 PS 51 1.3 S-12 68 SS 68 PS 48 1.0 S-13 S-14 S-15 74 105 85 SS SS TF 99 174 153 PS NT PS 48 48 51 0.7 0.6 0.6 S-16 S-17 S-18 S1 S2 S3 S4 S5 S6 S7 S8 S9 74 127 188 258 204 300 267 394 250 339 379 458 RS/PS PS PS PS PS PS PS PS PS PS PS PS 164 206 300 326 326 431 431 506 326 317 506 339 PS PS PS PS PS PS PS PS PS PS PS PS 44 44 44 44 44 44 44 53 44 53 53 57 0.5 0.6 0.6 0.8 0.6 0.7 0.6 0.8 0.8 1.1 0.7 1.3 S10 S11 S12 305 405 328 PS PS PS 339 335 335 PS PS PS 57 53 53 0.9 1.2 1.0 S13 S14 S15 S16 S17 420 239 485 475 540 PS PS PS PS SS 235 235 535 506 341 PS PS PS PS PS 57 57 57 53 59 1.8 1.0 0.9 0.9 1.6 S18 480 SS 341 PS 59 1.4 Additionally, in the test series with offset connections (S2, S4, S8, S10, S12, S14, and S18), load-carrying capacities were lower compared to their counterparts with identical 65 but centered connections (S1, S3, S5, S9, S11, S13, and S17). However, the predicted load-carrying capacities based on CSA O86 remained unchanged, because the CSA O86 provisions currently do not account for such offset connection configurations. 4.5 Prediction based on prEN1995 (2024) The predictions based on prEN1995 (2024) [29] are summarized in Table 4-11 for test series with STS installed at 90º, and in Table 4-12 for STS installed at 45º. These predictions were calculated at the mean level using mean material properties. Table 4-11 prEN 1995 predictions for test series with STS at a 90º. ID Observed Prediction Fexp/Fpre Fexp (kN) Failure mode Fpre (kN) Failure mode teff (mm) AB1 261 RS/PS 105 PS 24 2.5 AB2 237 RS/PS 105 PS 24 2.3 CD1 190 PS 79 PS 24 2.4 CD2 247 PS 119 PS 36 2.1 CD2-90 143 PS/SS 82 PS 36 1.7 CD3 305 RS/PS 119 PS 36 2.6 CD3-90 197 PS/SS 82 PS 36 2.4 CD4 358 RS/PS 119 PS 36 3.0 CD5 224 RS/PS 126 PS 36 1.8 CD5-90 146 PS/SS 78 PS 36 1.9 EF1 225 TF 113 PS 44 2.0 EF2 269 RS/PS 136 PS 44 2.0 EF3 287 RS/PS 136 PS 44 2.1 CA4 213 EYM 151 PS 40 1.4 CA5 220 PS 151 PS 40 1.5 CA6 203 SS 151 PS 40 1.3 CA10 309 EYM 159 PS 45 1.9 CA11 297 PS 159 PS 45 1.9 CA12 228 SS 159 PS 45 1.4 CA16 236 EYM 222 PS 57 1.1 CA17 254 EYM 222 PS 57 1.1 CA18 253 EYM 222 PS 57 1.1 66 The comparison for 90º installation revealed discrepancies; while prEN 1995 does not account for step shear failure and plug shear failure was predicted, the experiments exhibited various brittle failure modes. The ratio of experimental to predicted loadcarrying capacity ranged from 1.1 to 3.0, with an average ratio of 1.9, indicating that the predicted values were significantly lower than the observed results. For the dataset with STS connections installed at 45º, prEN 1995 consistently predicted plug shear failure due to the absence of provisions for step shear failure. However, the observed failure modes included both step shear and tensile failure. A comparison of the experimental and predicted load-carrying capacities for each test series that exhibited brittle failure showed ratios ranging from 0.5 to 2.7, with an average of 1.4. By comparing test series with identical STS layouts but conducted on wider CLT panels (CD4, EF3, S9W, and S10W), it was observed that prEN 1995 predictions, like those from CSA O86-2024, do not account for the effect of CLT panel width, while the tests showed the impact of specimen width. It should be noted that the design equations in prEN 1995 was developed for glulam and lacks specific guidance for CLT. However, herein, predictions incorporated the effect of transverse layers similar to the approach in CSA O86-2024. As a result, when comparing the predicted values for CD2, CD3, and CD5 with CD2-90, CD3-90, and CD5-90, respectively, the reduction in predicted load-carrying capacity in the minor direction matched the experimental results. Additionally, in the test series with offset connections (S2, S4, S8, S10, S12, S14, and S18), the load-carrying capacities were lower than those of the corresponding series with centered connections (S1, S3, S5, S9, S11, S13, and S17), despite having identical layouts. However, the predicted load-carrying capacities based on prEN 1995 (2024) 67 remained the same, highlighting that, like CSA O86, the effect of offset configurations is not considered in the prEN 1995 provisions. Table 4-12 prEN 1995 predictions for test series with STS at a 45º. ID Observed Prediction Fexp/Fpre Fexp (kN) Failure mode Fpre (kN) Failure mode teff (mm) S3 218 TF 123 PS 38 1.8 S4 202 TF 119 PS 48 1.7 S5 241 PS 197 PS 48 1.2 S6 226 PS 297 PS 48 0.8 S6W 336 PS 297 PS 48 1.1 S8 148 SS 175 PS 38 0.8 S9 142 SS 88 PS 44 1.6 S9W 144 TF 88 PS 44 1.6 S10 385 PS 186 PS 48 2.1 S10W 468 PS 186 PS 48 2.5 S11 359 PS 164 PS 51 2.2 S12 68 SS 68 PS 48 1.0 S13 74 SS 68 PS 48 1.1 S14 S15 105 85 SS TF 165 138 PS PS 48 51 0.6 0.6 S16 S17 74 127 RS/PS PS 155 120 PS PS 44 44 0.5 1.1 S18 188 PS 180 PS 44 1.0 S1 S2 258 204 PS PS 184 184 PS PS 44 44 1.4 1.1 S3 300 PS 344 PS 44 0.9 S4 267 PS 344 PS 44 0.8 S5 394 PS 374 PS 53 1.1 S6 S7 250 339 PS PS 184 196 PS PS 44 53 1.4 1.7 S8 379 PS 374 PS 53 1.0 S9 458 PS 199 PS 57 2.3 S10 S11 305 405 PS PS 199 193 PS PS 57 53 1.5 2.1 S12 328 PS 193 PS 53 1.7 S13 420 PS 199 PS 57 2.1 S14 239 PS 199 PS 57 1.2 S15 S16 485 475 PS PS 385 374 PS PS 57 53 1.3 1.3 S17 540 SS 202 PS 59 2.7 S18 480 SS 202 PS 59 2.4 68 4.6 Prediction based on stiffness model The predicted results based on the stiffness model [10] is presented in Table 4-13 for STS installed at 90º and in Table 4-14 for STS installed at 45º. It should be noted that the stiffness model only accounts for plug shear failure modes in CLT, however differentiated into six different types of plug shear. For STS installed at 90º, the ratio of experimental to predicted load-carrying capacity at the mean level ranged from 0.8 to 3.3, with an average ratio of 1.6. For STS installed at 45º, the comparison between experimental and predicted load-carrying capacities for each test series that failed in a brittle manner at the mean level showed ratios ranging from 0.5 to 3.2, with an average ratio of 1.6. Table 4-13 Stiffness model predictions for test series with STS at a 90º. ID Observed Prediction Fexp/Fpre Fexp (kN) Failure mode Fpre (kN) Failure mode teff (mm) AB1 261 RS/PS 185 Mode A 24 1.4 AB2 237 RS/PS 191 Mode A 24 1.2 CD1 190 PS 103 Mode C 24 1.8 CD2 247 PS 177 Mode C 36 1.4 CD2-90 143 PS/SS 177 Mode C 36 0.8 CD3 305 RS/PS 145 Mode C 36 2.1 CD3-90 197 PS/SS 145 Mode C 36 1.4 CD4 358 RS/PS 117 Mode C 36 3.0 CD5 224 RS/PS 121 Mode C 36 1.8 CD5-90 146 PS/SS 121 Mode C 36 1.2 EF1 225 TF 271 Mode E 44 0.8 EF2 269 RS/PS 113 Mode C 44 2.4 EF3 287 RS/PS 87 Mode C 44 3.3 CA4 213 EYM 167 Mode C 40 1.3 CA5 220 PS 167 Mode C 40 1.3 CA6 203 SS 159 Mode C 40 1.3 CA10 309 EYM 179 Mode C 45 1.7 CA11 297 PS 179 Mode C 45 1.7 CA12 228 SS 168 Mode C 45 1.4 CA16 236 EYM 211 Mode C 57 1.1 CA17 254 EYM 211 Mode C 57 1.2 CA18 253 EYM 185 Mode C 57 1.4 69 Table 4-14 Stiffness model predictions for test series with STS at a 45º. ID Observed Prediction Fexp/Fpre Fexp (kN) Failure mode Fpre (kN) Failure mode teff (mm) S3 218 TF 138 Mode C 38 1.6 S4 202 TF 130 Mode E 48 1.6 S5 241 PS 119 Mode E 48 2.0 S6 226 PS 202 Mode E 48 1.1 S6W 336 PS 202 Mode E 48 1.7 S8 148 SS 97 Mode C 38 1.5 S9 142 SS 111 Mode E 44 1.3 S9W 144 TF 111 Mode E 44 1.3 S10 385 PS 162 Mode C 48 2.4 S10W 468 PS 145 Mode C 48 3.2 S11 359 PS 116 Mode C 51 3.1 S12 68 SS 29 Mode C 48 2.4 S13 74 SS 52 Mode C 48 1.4 S14 105 SS 151 Mode C 48 0.7 S15 85 TF 94 Mode C 51 0.9 S16 74 RS/PS 147 Mode C 44 0.5 S17 127 PS 137 Mode C 44 0.9 S18 188 PS 217 Mode C 44 0.9 S1 258 PS 237 Mode C 44 1.1 S2 204 PS 237 Mode C 44 0.9 S3 300 PS 206 Mode E 44 1.5 S4 267 PS 206 Mode E 44 1.3 S5 394 PS 257 Mode E 53 1.5 S6 250 PS 185 Mode C 44 1.4 S7 339 PS 224 Mode E 53 1.5 S8 379 PS 257 Mode E 53 1.5 S9 458 PS 246 Mode C 57 1.9 S10 305 PS 246 Mode C 57 1.2 S11 405 PS 274 Mode C 53 1.5 S12 328 PS 274 Mode C 53 1.2 S13 420 PS 139 Mode C 57 3.0 S14 239 PS 139 Mode C 57 1.7 S15 485 PS 257 Mode E 57 1.9 S16 475 PS 257 Mode E 53 1.8 S17 540 SS 232 Mode C 59 2.3 S18 480 SS 232 Mode C 59 2.1 70 Similar to CSA O86 and prEN 1995, the stiffness model does not account for CLT panel width, or variations in unloaded edge distance for comparable layouts. Additionally, it lacks specific guidance for CLT in the minor strength direction. Consequently, the same procedure used for the major strength direction was applied to the minor direction, resulting in identical predicted values for connection layouts in both orientations. 4.7 Evaluation of prediction models To evaluate the analytical models for brittle failure of CLT with STS connections, scatter plots comparing predicted and observed load-carrying capacities from experimental tests were generated (Figure 4.2 to Figure 4.4). The accuracy of each model is indicated by the closeness of its data points to the ideal line, which represents a 1:1 correlation between predicted and experimental values. Predicted load-carrying capacity (kN) 600 Azinović et al Ni and Niederwestberg UNBC 2022 UNBC 2024 (Centered) UNBC 2024 (Offset) 500 400 300 200 100 0 0 100 200 300 400 500 600 Experimental load-carrying capacity (kN) Figure 4.2 Scatter plot of the experimental results and the CSA O86 predicted values. 71 Predicted load-carrying capacity (kN) 600 Azinović et al Ni and Niederwestberg UNBC 2022 UNBC 2024 (Centered) UNBC 2024 (Offset) 500 400 300 200 100 0 0 100 200 300 400 500 600 Experimental load-carrying capacity (kN) Figure 4.3 Scatter plot of the experimental results and the prEN 1995 predicted values. Predicted load-carrying capacity (kN) 600 Azinović et al Ni and Niederwestberg UNBC 2022 UNBC 2024 (Centered) UNBC 2024 (Offset) 500 400 300 200 100 0 0 100 200 300 400 500 600 Experimental load-carrying capacity (kN) Figure 4.4 Scatter plot of the experimental results and stiffness model predicted values. Three metrics were utilized to assess model performance, similar to the work done in previous studies [9], [15], [47]: i) the mean relative error (MRE); ii) the slope of a linear fit through the origin (m); and iii) the concordance correlation coefficient (CCC). MRE offers a metric for assessing model error, with lower MRE values indicating better model 72 performance. CCC evaluates both precision (deviation between predictions and experimental values) and accuracy (how well the model aligns with the ideal 1:1 line through the origin) [47]. A higher CCC value reflects a more reliable model. Finally, a slope m of close to 1.0 indicates a linear fit through the origin and suggests a strong correlation and good model performance. Equations 4-1 to 4-3 were used to calculate MRE and the CCC respectively, where ‫ݕ‬୧ represents the observed experimental values, ݂௜ denotes the predicted values from the models, ‫ݕ‬ത and ݂ ̅ are the means of the experimental and predicted values. The metrics calculated for the prediction models are shown in Table 4-15. ௡ 1 ෌௜ୀଵ(‫ݕ‬୧ − ݂௜ ) ‫= ܧܴܯ‬ ݊ ‫ݕ‬ത ݉ = 4-1 ∑௡௜ୀଵ(‫ݕ‬௜ − ‫ݕ‬ത) ൫݂௜ − ݂൯̅ ௡ ෌௜ୀଵ(‫ݕ‬୧ − ‫ݕ‬ത) ‫= ܥܥܥ‬ ௡ 4-2 ଶ ௡ ̅ ௜ − ‫ݕ‬ത) 2 ෌௜ୀଵ(݂௜ − ݂)(‫ݕ‬ 4-3 ௡ ෌௜ୀଵ(݂௜ − ݂)̅ ଶ + ෌௜ୀଵ(‫ݕ‬௜ − ‫ݕ‬ത)ଶ + ݊(݂ ̅ − ‫ݕ‬ത) Table 4-15 Metrics obtained for the different models. Prediction model MRE m CCC CSA O86-2024 0.24 0.94 0.71 prEN 1995 (2024) 0.38 0.63 0.39 Stiffness model 0.38 0.60 0.30 Upon evaluating the scatter plots and calculated metrics, it is clear that the CSA O862024 model provides more accurate predictions compared to the others. The mean relative error (MRE) was the lowest with 0.24, the slope m was the highest with 0.94, and the concordance correlation coefficient (CCC) was the highest with 0.71. The prEN 1995 and 73 stiffness model predictions were identical in terms of MRE and m; however, CCC was higher for prEN 1995 than for the stiffness model. Overall, all three metrics indicated a lower predictive power of all three prediction models for brittle failure in CLT connections with STS compared to similar studies on brittle failure in glulam and solid lumber [9]. This aligns with the findings of Azinović et al. [15], where CCC values for predicting brittle failure in CLT using the stiffness model were around 0.29 at the characteristic level. These results suggest that the current prediction models are not appropriate for predicting brittle failure in CLT with STS connections, indicating the need for an improved model, which is presented in the next chapter. 74 5 Proposed model for brittle failure in CLT with inclined STS 5.1 Overview This chapter presents an analytical approach for predicting brittle failure in CLT connection with inclined STS. The effective depth of the failure block in CLT was investigated based on the observed failure modes from the experimental tests in Chapter 3. A modified model for evaluating the contribution of active resistance planes was applied to predict the load-carrying capacities of CLT connections with STS at a 45º angle. These predictions were then compared with experimental results and those based on CSA O86-2024 provisions. 5.2 Effective depth for brittle failure of CLT with inclined STS A key factor in calculating brittle failure resistance is the effective depth (teff) of the failure block, as shown in Figure 5.1. For inclined screws, CSA O86 recommends using the projected bearing length of the screws, perpendicular to the applied load, as the penetration depth (Lp) to determine the effective depth. However, the equation for teff was developed for laterally loaded screws, where fasteners typically start yielding before brittle failure. For connections with inclined STS, no such yielding was observed in the experiments and brittle failure was the first and the only failure mode while the fasteners remained straight. Figure 5.1 Effective depth and penetration depth of inclined screws. 75 To shed light on this issue, the actual depth of failure block for brittle failure of CLT with inclined STS in the conducted tests was measured based on the observed failure modes. To illustrate, Figure 5.2 shows the failure mode in series S11, which exhibited plug shear failure involving the first three layers and the corresponding effective depth of the failure block. The observed effective depths for all test series are reported in Table 5.1. Figure 5.2 Failure mode and observed effective depth in series S11. Table 5-1 Observed effective depth of the failure blocks for all test series. Test ID tCLT (mm) Penetrated layer Lp (mm) CSA teff (mm) Observed Failure mode Observed teff (mm) S1 105 2nd 52 44 Mode D/F 70 S2 105 2nd 52 44 Mode D 70 S3 100 3rd 52 44 Mode F 60 S4 100 3rd 52 44 Mode F 60 S5 139 4th 94 53 Mode F 87 S6 139 3rd 52 44 _ _ S7 139 4th 94 53 _ _ S8 139 4th 94 53 Mode H 104 S9 175 4th 137 57 Mode J 140 S10 175 4th 137 57 TF/RS _ S11 175 3rd 94 53 Mode F 105 S12 175 3rd 94 53 Mode F 105 S13 175 4th 137 57 Mode H/J 145 S14 175 4th 137 57 TF/RS _ S15 191 5th 137 57 SS 139 S16 191 4th 94 53 Mode F/H 104 S17 245 5th 165 59 SS 140 S18 245 5th 165 59 SS 175 76 As detailed in Table 5-1, the experimental results reveal a significant discrepancy between the effective depth recommended by CSA O86 and the observed values. On average, the effective depth calculated using CSA O86 is 49% lower than the depth observed in the tests. This substantial difference suggests that the current CSA O86 provisions underestimate the actual failure depth in CLT connections with inclined STS. In contrast, the observed failure block depth closely corresponds to the projected penetration length (penetration depth) of the inclined screws (see Figure 5.1), with only a 10% average relative difference across all test series. Therefore, using the penetration depth of screws for brittle failure of CLT connection with inclined STS is more consistent with the experimental observations. 5.3 Prediction based on the proposed model In this section, the contribution of the three resistance planes is evaluated: 1) head tensile (blue), 2) side shear (green), and 3) bottom shear (red), as illustrated in Figure 5.3. Figure 5.3 Definition of plug shear failure in proposed models. The proposed model, based on the observed depth of the failure block in experimental tests on CLT connections with inclined STS, uses the screw penetration depth (Lp) as the effective depth of failure block. For brittle failure calculations, the model incorporates 77 plane strength values and the actual areas of resistance planes—head tensile (AH), side shear (AS), and bottom shear (AB)—based on the actual length of failure block (L). Additionally, the screw length within the transverse CLT layers is excluded from the calculations for head tensile and side shear resistance, and the load-carrying capacity is determined by the maximum resistance across all layers. An overview of the proposed model and its differences from the CSA O86-2024 model are provided in Table 5-2. Table 5-2 Overview of the proposed models for calculation of plug shear failure. Strength of plane Prediction models Failure Proposed model Maximum of H, L, B CSA O86 (2024) L+H+B Head tensile (H) Lateral shear (L) ݂௧ ܾு ‫ݐ‬௘௙௙ ݂௩ ‫ݐ ܮ‬௘௙௙ 1.25݂௧ ܾு ‫ݐ‬௘௙௙ 1.5݂௩ ‫ܮ‬௦,௜ ‫ݐ‬௘௙௙ Bottom shear (B) Effective depth (‫ݐ‬௘௙௙ ) ݂௩ ‫ܾ ܮ‬஻ or ݂௩,௥ ‫ܾ ܮ‬஻ 0.75 ݂௩ ‫ܮ‬௦,௜ ܾ஻ or 0.75 ݂௩,௥ ‫ܮ‬௦,௜ ܾ஻ ‫ܮ‬௣ ݇௖௟ ቌ 7‫ܮ‬௣ ቍ ≤ ‫ܮ‬௣ ‫ܮ‬௣ 3+ ݀ It is worth mentioning that, in the CSA O86-2024 provisions, the length of the failure block ( ‫ܮ‬௦,௜ ) is calculated using the number of fasteners in row and the minimum of the spacing parallel to the grain (Sp) and the loaded edge distance (aL), as shown in equation 5-1. 5-1 ‫ܮ‬௦,௜ = ݊஼ ∗ ݉݅݊(ܽ௅ , ܵ௣ ) This calculation may slightly differ from the actual length of the connection. The geometric details of the resistance planes used in predicting brittle failure based on the proposed model for all test series are presented in Table 5-3. In this table, bH, bB, and L represent the width of the head plane, the width of the bottom plane, and the length of the failure block, respectively. 78 Table 5-3 Geometric details of resistance planes in proposed model for each test series. ID Lp (mm) Penetrated layer bH (mm) bB (mm) L (mm) AH (mm2) AB (mm2) AS (mm2) S-3 38 2nd 119 150 403 2380 60426 8057 S-4 66 4th 48 60 290 1904 17382 11588 S-5 66 4th 119 150 233 4760 34971 9325 S-6 66 4th 119 150 403 4760 60426 16114 S-6W 66 4th 119 150 403 4760 60426 16114 S-8 38 2nd 119 150 233 2134 34971 4180 S-9 52 3rd 119 150 233 2380 34971 4663 S-9W 52 3rd 119 150 233 2380 34971 4663 S-10 66 3rd 119 150 403 5856 60426 19825 S-10W 66 3rd 119 150 403 5856 60426 19825 S-11 80 3rd 119 150 346 7539 51941 21938 S-12 66 3rd 24 30 233 405 6994 3963 S-13 66 3rd 48 60 233 809 13988 3963 S-14 66 3rd 119 150 233 2023 34971 3963 S-15 80 3rd 119 150 177 2023 26485 3002 S-16 52 2nd 48 60 346 1666 20776 12120 S-17 52 2nd 119 150 233 4165 34971 8160 S-18 52 2nd 119 150 403 4165 60426 14099 S1 52 2nd 135 160 370 4732 59200 12950 S2 52 2nd 135 160 370 4732 59200 12950 S3 52 3rd 135 160 370 4336 59200 11866 S4 52 3rd 135 160 370 4336 59200 11866 S5 94 4th 135 160 392 9464 62720 27440 S6 52 3rd 135 160 370 4742 59200 12976 S7 94 4th 68 80 392 4732 31360 27440 S8 94 4th 135 160 392 9464 62720 27440 S9 137 4th 135 160 408 9464 65280 28560 S10 137 4th 135 160 408 9464 65280 28560 S11 94 3rd 135 160 392 8044 62720 23323 S12 94 3rd 135 160 392 8044 62720 23323 S13 137 4th 68 80 408 4732 32640 28560 S14 137 4th 68 80 408 4732 32640 28560 S15 137 5th 135 160 408 13915 65280 41993 S16 94 4th 135 160 392 9464 62720 27440 S17 165 5th 135 160 424 12872 67840 40368 S18 165 5th 135 160 424 12872 67840 40368 79 The predicted load-carrying capacity of the proposed model for the experimental data of CLT connections with STS at a 45º angle, along with the contribution of head tensile plane (RH), bottom shear plane (RB), and side shear plane (RS), are presented in Table 5-4. Table 5-4 Predicted load-carrying capacity based on the proposed model for inclined test data. ID teff (mm) RH (kN) RB (kN) RS (kN) Rplug (kN) Fexp Fexp/Fpre S-3 S-4 S-5 S-6 S-6W S-8 S-9 S-9W S-10 S-10W S-11 S-12 S-13 S-14 S-15 S-16 S-17 S-18 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 38 66 66 66 66 38 52 52 66 66 80 66 66 66 80 52 52 52 52 52 52 52 94 52 94 94 137 137 94 94 137 137 137 94 165 165 51 41 102 102 102 46 51 51 125 125 161 9 17 43 43 36 89 89 101 101 93 93 202 101 101 202 202 202 172 172 101 101 297 202 275 275 118 34 68 118 118 205 68 68 355 355 305 14 27 68 52 41 68 118 116 116 348 348 123 348 61 123 128 128 368 368 64 64 383 123 398 398 95 136 109 189 189 49 55 55 233 233 258 47 47 47 35 142 96 166 152 152 139 139 322 152 322 322 335 335 274 274 335 335 493 322 474 474 118 136 109 189 189 205 68 68 355 355 305 47 47 68 52 142 96 166 152 152 348 348 322 348 322 322 335 335 368 368 335 335 493 322 474 474 218 202 241 226 336 148 142 144 385 468 359 68 74 105 85 74 127 188 258 204 300 267 394 250 339 379 458 305 405 328 420 239 485 475 540 480 1.8 1.5 2.2 1.2 0.9 0.7 2.1 0.7 1.1 1.3 1.2 0.6 0.7 1.5 1.6 0.5 1.3 1.1 1.5 1.2 0.9 0.8 1.2 0.7 1.1 1.2 1.4 0.9 1.1 0.9 1.1 0.6 1.0 1.5 1.1 1.0 80 It should be noted that the proposed model cannot capture the effects of CLT panel width or variations in unloaded edge distance for comparable layouts. 5.4 Comparison of the proposed prediction model with CSA O86-2024 The predictive ability of the proposed model is compared to that of the CSA O86-2024 model for test series with CLT connections at a 45º angle. The predicted load-carrying capacities for the proposed model and CSA O86-2024 for the test series with inclined STS were previously presented in Table 5-4 and Table 4-10, respectively. Herein, scatter plots, as shown in Figure 5.4 and Figure 5.5, are used to illustrate the comparison between the experimental, including both centered and offset test series, and predicted results for both models. 600 Predicted load-carrying capacity (kN) Centered Offset 400 200 0 0 200 400 600 Experimental load-carrying capacity (kN) Figure 5.4 Scatter plot of the experimental results and prediction based on CSA O86. 81 600 Predicted load-carrying capacity (kN) Centered Offset 400 200 0 0 200 400 600 Experimental load-carrying capacity (kN) Figure 5.5 Scatter plot of the experimental results and prediction based on the proposed model. To further compare the two scatter plots, the metrics introduced in Chapter 4 were applied to the predictions from the proposed model and CSA O86, and the results for the centered and offset test series are separately presented in Table 5-5 and Table 5-6. Table 5-5 Metrics obtained for proposed and CSA O86 models in centered test series. Prediction model MRE m CCC Proposed model 0.19 0.87 0.86 CSA O86-2024 0.22 0.92 0.77 Table 5-6 Metrics obtained for proposed and CSA O86 models in offset test series. Prediction model MRE m CCC Proposed model 0.16 1.04 0.77 CSA O86-2024 0.27 1.08 0.29 82 The metrics for the proposed model and CSA O86 predictions were evaluated separately for centered and offset test series. For the centered test series, the proposed model showed a lower mean relative error (MRE) of 0.19 compared to 0.22 for CSA O86. While the CSA O86 model achieved a slightly better slope from the origin of 0.92, closer to the ideal 1:1 line than the 0.87 of the proposed model, the proposed model demonstrated superior overall performance with a CCC of 0.86 compared to 0.77 for CSA O86. In the offset test series, the proposed model outperformed CSA O86 in terms of MRE, achieving 0.16 compared to 0.27 for CSA O86. For the slope, the proposed model also performed better, with a value of 1.04, closer to the ideal slope of 1 for the 1:1 line than the 1.08 of CSA O86. Additionally, the proposed model demonstrated superior overall performance with a much higher CCC value of 0.77 compared to 0.29 for CSA O86. Overall, the proposed model provided more accurate predictions of brittle failure in CLT connections with inclined STS, aligning more closely with experimental observations such as the effective depth of the failure block. 83 6 Conclusions 6.1 Summary of experimental work This research focused on experimental tests of CLT connections with inclined STS at a 45º angle under tensile loading, aimed at observing brittle failure. The connections were designed according to CSA O86-2024 provisions at the mean level, with specimens made from CLT panels of various layups, utilizing different STS lengths and placements to assess the impact of design parameters on failure modes and load-carrying capacities. The experimental findings led to several key conclusions: I. The observed brittle failure modes across test series indicated that brittle failure in CLT with STS connections is more complex than in glulam and differs from the failure modes outlined in CSA O86-2024. Specifically, plug shear failure occurred differently than suggested in the CSA provisions, with failure modes varying based on the contribution of different CLT layers. II. Changing the unloaded edge distance of identical STS connections in CLT by offsetting them from the center to the edge of the panel, despite having equal resistance plane areas, resulted in an average 20% reduction in load-carrying capacities for offset connections compared to centered ones. III. For test series that exhibited plug shear and step shear with distinguishable failure blocks, the observed failure block depth closely matched the penetration depth of the inclined screws, with an average relative difference of only 10% between the observed effective depth and the screw penetration depth. 6.2 Summary of analytical work An analytical investigation was conducted on available test data to examine the predictive ability of the approaches from CSA O86-2024, prEN 1995 (2024), and Zarnani and 84 Quenneville, identifying their limitations in predicting brittle failure in CLT. Following this, a modified approach for predicting brittle failure in CLT connections with inclined STS was proposed. The analytical findings led to the following key conclusions: I. The evaluation of the prediction models—CSA O86-2024, prEN 1995 (2024), and the stiffness model by Zarnani and Quenneville—demonstrates that the CSA O86-2024 model provides more accurate predictions than the others. The CSA O86-2024 model showed better overall performance, with a CCC of 0.71 across all test series, compared to 0.39 for prEN 1995 and 0.30 for the stiffness model. II. The proposed model provides more accurate predictions of brittle failure in CLT connections with inclined STS than the CSA O86 model, while also aligning more closely with experimentally observed effective depth of the failure block. It demonstrated better overall performance, with a CCC value of 0.85, compared to 0.73 for the CSA O86 model. III. None of the existing prediction models—CSA O86, prEN 1995, the stiffness model, and the proposed model—account for the effects of CLT panel width or variations in unloaded edge distance for similar layouts. All models produced identical loadcarrying capacity predictions for test series with both narrower and wider CLT panels, as well as for offset connections near the panel edge and centered connections. 6.3 Future research Aside from the findings of this research, further experimental and analytical investigations are needed to address the gaps in current provisions for predicting brittle failure in CLT connections with STS. I. Conducting further tests on offset connections with varying unloaded edge distances to assess the direct impact of edge distance on the load-carrying capacity of the connection. 85 II. Performing additional tests on brittle failure of CLT connections with inclined screws in the minor strength direction to develop a comprehensive dataset for evaluating the efficiency of available prediction models in both minor and major directions. III. Carrying out further tests on brittle failure in CLT connections with STS at 90-degree angles to refine equations for determining the effective depth of the failure block in laterally loaded screw connections. IV. Proposing a new prediction model that accounts for the effects of unloaded edge distance, specimen width, and adjacent resistance planes in predicting brittle failure modes and load-carrying capacities in CLT. V. Conducting an analytical study using characteristic values from experimental tests, instead of mean values, to evaluate the accuracy of prediction models for brittle failure in CLT at the design level. 86 References [1] H. E. Ilgın and M. Karjalainen, “Perceptions, Attitudes, and Interests of Architects in the Use of Engineered Wood Products for Construction: A Review.” [Online]. Available: www.intechopen.com [2] M. H. Ramage et al., “The wood from the trees: The use of timber in construction,” Feb. 01, 2017, Elsevier Ltd. doi: 10.1016/j.rser.2016.09.107. [3] Karacabeyli and Gagnon, “clt-handbook-complete-version-en-low,” 2019. [4] P. Dietsch and R. Brandner, “Self-tapping screws and threaded rods as reinforcement for structural timber elements-A state-of-the-art report,” Constr Build Mater, vol. 97, pp. 78–89, Oct. 2015, doi: 10.1016/j.conbuildmat.2015.04.028. [5] M. Popovski and I. Gavric, “Performance of a 2-Story CLT House Subjected to Lateral Loads,” Journal of Structural Engineering, vol. 142, no. 4, Apr. 2016, doi: 10.1061/(asce)st.1943-541x.0001315. [6] D. C. Stahl, Begel, Marshall, and R. W. Wolfe, “Simplified analysis of timber rivet connections,” 2000. [7] “EN 1995-1-1:2004+A2:2014 - Eurocode 5: Design of timber structures - Part 11: General - Common rules and rules for buildings,” 2014. [8] P. Zarnani and P. Quenneville, “Wood Block Tear-Out Resistance and Failure Modes of Timber Rivet Connections: A Stiffness-Based Approach,” Journal of Structural Engineering, vol. 140, no. 2, Feb. 2014, doi: 10.1061/(asce)st.1943541x.0000840. 87 [9] M. Yurrita, J. M. Cabrero, and E. Moreno-Zapata, “Brittle failure in the parallelto-grain direction of timber connections with small diameter dowel-type fasteners: A new design model for plug shear,” Eng Struct, vol. 241, Aug. 2021, doi: 10.1016/j.engstruct.2021.112450. [10] P. Zarnani and P. Quenneville, “New design approach for controlling brittle failure modes of small-dowel-type connections in Cross-laminated Timber (CLT),” Constr Build Mater, vol. 100, pp. 172–182, Dec. 2015, doi: 10.1016/j.conbuildmat.2015.09.049. [11] M. Yurrita and J. M. Cabrero, “On the need of distinguishing ductile and brittle failure modes in timber connections with dowel-type fasteners,” Eng Struct, vol. 242, Sep. 2021, doi: 10.1016/j.engstruct.2021.112496. [12] Canadian Standards Association (CSA), “CSA O86-19,” Engineering design in wood, 2019. [13] Canadian Standards Association (CSA Group), “CSA O86 ‘Engineering Design in Wood,’” 2024. [14] M. Yurrita, J. M. Cabrero, and P. Quenneville, “Brittle failure in the parallel-tograin direction of multiple shear softwood timber connections with slotted-in steel plates and dowel-type fasteners,” Constr Build Mater, vol. 216, pp. 296–313, Aug. 2019, doi: 10.1016/j.conbuildmat.2019.04.100. [15] B. Azinović, J. M. Cabrero, H. Danielsson, and T. Pazlar, “Brittle failure of laterally loaded self-tapping screw connections for cross-laminated timber structures,” Eng Struct, vol. 266, Sep. 2022, doi: 10.1016/j.engstruct.2022.114556. [16] J. M. Cabrero et al., “Brittle failures of connections with self-tapping screws on CLT plates.” 88 [17] Z. Duan, Q. Huang, and Q. Zhang, “Life cycle assessment of mass timber construction: A review,” Aug. 01, 2022, Elsevier Ltd. doi: 10.1016/j.buildenv.2022.109320. [18] N. Heeren, C. L. Mutel, B. Steubing, Y. Ostermeyer, H. Wallbaum, and S. Hellweg, “Environmental Impact of Buildings - What Matters?,” Environ Sci Technol, vol. 49, no. 16, pp. 9832–9841, Jul. 2015, doi: 10.1021/acs.est.5b01735. [19] A. H. Buchanan and S. B. Levine, “Wood-based building materials and atmospheric carbon emissions,” Environ Sci Policy, vol. 2, no. 6, pp. 427–437, 1999, doi: https://doi.org/10.1016/S1462-9011(99)00038-6. [20] M. Kitek Kuzman, S. Klarić, A. Pirc Barčić, R. P. Vlosky, M. M. Janakieska, and P. Grošelj, “Architect perceptions of engineered wood products: An exploratory study of selected countries in Central and Southeast Europe,” Constr Build Mater, vol. 179, pp. 360–370, 2018, doi: https://doi.org/10.1016/j.conbuildmat.2018.05.164. [21] Canadian Wood Council, “Mass timber.” Accessed: Mar. 10, 2024. [Online]. Available: https://cwc.ca/en/how-to-build-with-wood/wood-products/mass- timber/ [22] L. Evans, “Cross Laminated Timber: Taking wood buildings to the next level,” Architectural Records, vol. 41, pp. 7427–7443, 2013. [23] A. Hossain, CONNECTIONS “EXPERIMENTAL WITH INVESTIGATIONS SELF-TAPPING-SCREWS OF SHEAR FOR CROSS- LAMINATED-TIMBER PANELS,” 2019. [24] M. P. Gutknecht, “PARALLEL-TO-GRAIN WITHDRAWAL RESISTANCE OF STRUCTURAL SELF-TAPPING SCREWS IN CANADIAN TIMBER,” 2017. 89 [25] K. W. Johansen, “Theory of timber connections,” International Assoc. of Bridge and Structural Engineering Publication, vol. 9, pp. 249–262, 1949. [26] M. Yurrita and J. M. Cabrero, “New design model for brittle failure in the parallelto-grain direction of timber connections with large diameter fasteners,” Eng Struct, vol. 217, Aug. 2020, doi: 10.1016/j.engstruct.2020.110557. [27] M. Mohammad, S. Gagnon, B. K. Douglas, and L. P. Podesto, “WOOD DESIGN FOCUS V. 22, N. 2 3 Introduction to Cross Laminated Timber.” [28] H. Asgari, T. Tannert, and C. Loss, “Block tear-out resistance of CLT panels with single large-diameter connectors,” European Journal of Wood and Wood Products, vol. 80, no. 4, pp. 805–816, Aug. 2022, doi: 10.1007/s00107-022-018093. [29] “CEN/TC250/SC5/WG5 (ed.) (2024) (unpublished) N619 prEN 1995-1-1 v0 to v1A. Con-nections, in Eurocode 5: Design of timber structures – Part 1-1: General rules – General rules and rules for buildings. Comité Européen de Normalisation, Brussels, Belgium. Status: 18-7-2024.” [30] American Wood Council, “NDS 2024 National Design Specification® for Wood Construction.,” 2024. [Online]. Available: www.awc.org. [31] W. Nozynski, “Investigation of the effect of number of nails in a joint on its load carrying ability,” CIB W-18 timber structures, 1980. [32] I. Smith and G. Steck, “Influence of number of rows of fasteners or connectors upon the ultimate capacity of axially loaded timber joints,” CIB-W18: Beit Oren, Israel, 1985. [33] D. C. Stahl, M. Asce, R. W. Wolfe, and M. Begel, “Improved Analysis of Timber Rivet Connections,” 2004, doi: 10.1061/ASCE0733-94452004130:81272. 90 [34] J. Quenneville and M. Mohammad, “On the failure modes and strength of steelwood-steel bolted timber connections loaded parallel-to-grain,” 2000. [35] M. Mohammad and J. H. P. Quenneville, “Bolted wood-steel and wood-steel-wood connections: Verification of a new design approach,” Canadian Journal of Civil Engineering, vol. 28, no. 2, pp. 254–263, 2001, doi: 10.1139/cjce-28-2-254. [36] Canadian Standards Association (CSA), “CSA O86-09,” Engineering design in wood, 2009. [37] Antti. Hanhijarvie, Ari. Kevarinmaki, and Valtion teknillinen tutkimuskeskus., Timber failure mechanisms in high-capacity dowelled connections of timber to steel : experimatnal results and design. VTT Technical Research Centre of Finland, 2008. [38] P. Quenneville and P. Zarnani, “Proposal for the connection chapter of the New Zealand design of timber structures,” Unpublished, 2017. [39] M. Yurrita and J. M. Cabrero, “Experimental analysis of plug shear failure in timber connections with small diameter fasteners loaded parallel-to-grain,” Eng Struct, vol. 238, Jul. 2021, doi: 10.1016/j.engstruct.2020.111766. [40] C. Ni and J. Niederwestberg, “INVESTIGATION OF BRITTLE FAILURE MODES IN SELF-TAPPING SCREW STEEL TO WOOD CONNECTIONS PARALLEL TO THE GRAIN,” 2022. [41] G. Armstrong and T. Tannert, “Brittle Failure Modes of Connections in CrossLaminated Timber with Self-Tapping Screws [Unpublished],” 2022. [42] A. P. A. PRG, “320 Standard for Performance-Rated Cross-Laminated Timber,” American National Standards Institute: Washington, DC, USA, 2018. 91 [43] M. T. C. Solutions, “Structural screw design guide,” Surrey, BC, Canada, 2020. [44] B. S. En, “26891; Timber Structures-Joints Made with Mechanical Fasteners– General Principles for the Determination of Strength and Deformation Characteristics,” British Standards Institution: London, UK, 1991. [45] CEN European Committee for Standardization, “EN 14358; Timber structures— Calculation and verification of characteristic values,” CEN: Brussels, Belgium, 2016. [46] Canadian Wood Council, Canadian Lumber Properties. Canadian Wood Council, 1994. [Online]. Available: https://webstore.cwc.ca/product/canadian-lumberproperties/ [47] J. M. Cabrero and M. Yurrita, “Performance assessment of existing models to predict brittle failure modes of steel-to-timber connections loaded parallel-to-grain with dowel-type fasteners,” Eng Struct, vol. 171, pp. 895–910, Sep. 2018, doi: 10.1016/j.engstruct.2018.03.037. 92 Appendix A: CSA O86-2024 The CSA O86-2024 offers various equations for calculating the factored brittle resistance of different failure modes in wood members with STS connections. For net tension failure mode, the total factored net tension resistance of a connection, ܶே௥் , is calculated using Equation A-1: (A-1) ܶே௥் = ∑ ܶ௥,௜ where ܶ௥,௜ = the factored tensile resistance parallel to grain of wood member i. For row shear failure mode, the total factored row shear resistance of a connection, PRrT, is calculated using the Equation A-2 and Equation A-3: ܴܲ௥் = ∑ ܴܲ௥,௜ (A-2) ܴܲ௥,௜ = ߶௪ ܲ‫ܤ‬௦,௝ ௠௜௡ ݊ோ (A-3) where ܴܲ௥,௜ = row shear resistances of wood members resisting the load ߶௪ = 0.7 ܲ‫ܤ‬௦,௝ ௠௜௡ = minimum row shear resistance of any row in the connection from ܴܲ௦,ଵ ‫ܴܲ ݋ݐ‬௦,௡ோ ݊ோ = number of fastener rows ܲ‫ܤ‬௦,௝ = side shear plane resistance of fastener row j For the group tear-out failure mode, the factored group tear-out resistance of a connection, PGrT, is calculated using the Equation A-4 and Equation A-5: (A-4) ܲ‫ܩ‬௥் = ∑ ܲ‫ܩ‬௥் 93 (A-5) ܲ‫ܩ‬௥,௜ = ߶௪ ൣܲ‫ܤ‬௧ + ൫ܲ‫ܤ‬௦,ଵ + ܲ‫ܤ‬௦,௡ோ ൯⁄2൧ where ܲ‫ܩ‬௥,௜ = group tear-out resistances of fully penetrated wood members resisting the load ߶௪ = 0.7 ܲ‫ܤ‬௧ = head tensile plane resistance of the critical net area between rows 1 and ݊ோ of wood member ܲ‫ܤ‬௦,ଵ = side shear plane resistance of wood member along row 1 bounding the fastener group ܲ‫ܤ‬௦,௡ோ = side shear plane resistance of wood member along row ݊ோ bounding the fastener group For plug shear failure in a two-member connection involving a partially penetrated wood member, the factored plug shear resistance, ܲܲ௥் , is calculated using Equation A-6 for lumber, glulam, and MLT, and Equation A-7 for CLT. (A-6) ߶௪ ( ܲ‫ܤ‬௧ + ܲ‫ܤ‬௦௕ ) ܲܲ௥் = ݉ܽ‫ ݔ‬ቐ ߶௪ ൫ܲ‫ܤ‬௦,ଵ + ܲ‫ܤ‬௦,௡ோ ൯⁄2 ܲܲ௥் = ߶௪ ( ܲ‫ܤ‬௧ + ܲ‫ܤ‬௦௕ + ൫ܲ‫ܤ‬௦,ଵ + ܲ‫ܤ‬௦,௡ோ ൯⁄2) (A-7) where ߶௪ = 0.7 ܲ‫ܤ‬௧ = head tensile plane resistance of the critical net area between rows 1 and ݊ோ of wood member ܲ‫ܤ‬௦,ଵ = side shear plane resistance of wood member along row 1 bounding the fastener group 94 ܲ‫ܤ‬௦,௡ோ = side shear plane resistance of wood member along row ݊ோ bounding the fastener group ܲ‫ܤ‬௦௕ = bottom shear plane resistance of wood member For step shear failure in a two-member connection involving a partially penetrated wood member, the factored step shear resistance, PSrT, is calculated using Equation A-8: (A-8) ܲܵ௥் = ߶௪ ൫ܲ‫ܤ‬௧ + ܲ‫ܤ‬௦,௕ ൯ where ߶௪ = 0.7 ܲ‫ܤ‬௧ = head tensile plane resistance of net area of the step in wood member ܲ‫ܤ‬௦,௕ = bottom shear plane resistance of the bottom shear plane of the step in wood member As shown in the equations above, the resistance of failure planes is required to calculate the brittle failure resistance of a wood member with an STS connection. The head tensile plane resistance (ܲ‫ܤ‬௧ ) of a wood member, is calculated using Equation A-9: (A-9) ܲ‫ܤ‬௧ = 1.25݂௧ (‫ܭ‬஽ ‫ܭ‬ௌ௧ ‫ܾ) ்ܭ‬௧ ‫ݐ‬௘௙ where ݂௧ = specified strength in tension parallel to grain of wood member, MPa ܾ௧ = critical width of head tensile plane, mm ‫ܭ‬ௌ௧ = service-condition factor for tension parallel to grain ‫ݐ‬௘௙ = effective depth of head tensile plane, mm The side shear resistance (ܲ‫ܤ‬௦,௝ ) of a wood member, is calculated using Equation A-10: 95 (A-10) ܲ‫ܤ‬௦,௝ = 1.5݂௩ (‫ܭ‬஽ ‫ܭ‬ௌ௩ ‫ܮ ) ்ܭ‬௦,௜ ‫ݐ‬௘௙ where ݂௩ = specified longitudinal shear strength of wood member, MPa ‫ܮ‬௦,௜ = critical width of head tensile plane, mm ‫ܭ‬ௌ௩ = service-condition factor for longitudinal shear ‫ݐ‬௘௙ = effective depth of side shear plane, mm The bottom shear resistance (ܲ‫ܤ‬௦௕ ) of a wood member, is calculated based on Equation A-11: (A-11) ܲ‫ܤ‬௦௕ = 0.75 ݂௩ (‫ܭ‬஽ ‫ܭ‬ௌ௩ ‫ܣ) ்ܭ‬௉,௦௕ where ݂௩ = specified longitudinal shear strength of wood member, MPa ‫ܣ‬௉,௦௕ = critical area of bottom shear plane, mm ‫ܭ‬ௌ௩ = service-condition factor for longitudinal shear The effective depth of the head tensile and side shear planes for fully penetrated members is equal to the member thickness, while for partially penetrated members, it is calculated using Equation A12: (A-12) 7‫ݐ‬௜ ݇௖௟ ቌ ‫ ݐ‬ቍ ≤ ‫ݐ‬௜ 3+ ௜ ݀ி where ‫ݐ‬௜ = bearing length of screws in partially penetrated wood member i, mm ݀ = nominal diameter, mm 96 ݇௖௟ = Clamping condition factor The critical length of the side shear plane for net tension, step shear, group tear-out, and plug shear ( ‫ܮ‬௦,௜ ) is calculated using Equation A-13: (A-13) ‫ܮ‬௦,௜ = ݊஼ ∗ ݉݅݊(ܽ௅ , ܵ௣ ) where ݊஼ = number of fasteners in row ܽ௅ = loaded end distance, mm ܵ௣ = minimum of the spacing parallel to the grain, mm 97 Appendix B: prEN1995 (2024) The draft of Eurocode 5 (2024) offers equations for calculating the design brittle failure resistance of connections with dowel-type fasteners. For row shear failure, the design row shear resistance of a timber member, ‫ܨ‬୰ୱ,ୖୢ, is calculated using Equation B-1: (B-1) ‫ܨ‬୰ୱ,ୖୢ = 2 ݊ଽ଴ ‫୴ܨ‬,୪ୟ,ୢ where ݊ଽ଴ = the number of fasteners in a row perpendicular to grain ‫୴ܨ‬,୪ୟ,ୢ = the design shear resistance per side shear plane in the timber member For block shear failure, in cases where the fastener penetrates the full timber member thickness, the design block shear capacity of a timber member, ‫ܨ‬ୠୱ,ୖୢ, is calculated using Equation B-2: (B-2) ‫ܨ‬ୠୱ,ୖୢ = ݉ܽ‫(ݔ‬2 ‫୴ܨ‬,୪ୟ,ୢ ; ‫ܨ‬௧,ௗ ) where ‫୴ܨ‬,୪ୟ,ୢ = the design shear resistance per side shear plane in the timber member ‫ܨ‬௧,ௗ = the design tensile resistance parallel to grain of the head tensile plane For plug shear failure, when the fasteners partially penetrate the timber member thickness, the design plug shear capacity of a timber member, ‫ܨ‬୮ୱ,ୖୢ, is calculated using Equation B-3: 2‫୴ܨ‬,୪ୟ,ୢ ‫ܨ‬୮ୱ,ୖୢ = ݉ܽ‫ ݔ‬൜ ‫୲ܨ‬,ୢ + ‫୴ܨ‬,ୠ,ୢ (B-3) where ‫୴ܨ‬,୪ୟ,ୢ = the design shear resistance per side shear plane in the timber member ‫ܨ‬௧,ௗ = the design tensile resistance of the head plane in the timber member 98 ‫୴ܨ‬,ୠ,ୢ = is the design shear plane resistance of the bottom shear plane in the timber member In the case of net tension failure, the design net tensile failure is determined according to Section 8.1.2 of prEN 1995, with consideration given to the reduction of the cross-section due to the pre-drilled holes for the fasteners and potential slots for metal plates. The net section shall be determined at the location with the highest number of fasteners perpendicular to the load direction. For the side shear planes, the design shear resistance per lateral shear plane, ‫୴ܨ‬,୪ୟ,ୢ , is calculated using Equation B-4: (B-4) ‫୴ܨ‬,୪ୟ,ୢ = ݇୴ ‫ݐ‬ୣ୤ ݈ୡ୭୬ ݂୴,ୢ with ‫ܩ‬௠௘௔௡ ‫ܧ‬଴,௠௘௔௡ (B-5) ݈௖௢௡ = ܽଵ (݊଴ − 1) + ܽଷ,௧ (B-6) ݇௩ = 0.4 + 1.4 ඨ where ݇୴ = the reduction factor for shear ‫ݐ‬ୣ୤ = the effective thickness of the plane ݈ୡ୭୬ = the length parallel to grain of the connection ݂୴,ୢ = the design shear strength of timber ܽଵ = the spacing of dowel-type fasteners parallel to grain ݊଴ = the number of fasteners in a row parallel to grain ܽଷ,௧ = the loaded end distance parallel to grain ‫ܩ‬୫ୣୟ୬ = the mean shear modulus of the timber ‫ܧ‬଴,୫ୣୟ୬ = the mean modulus of elasticity of the timber parallel to grain 99 For the bottom shear plane, the design bottom shear failure resistance, ‫୴ܨ‬,ୠ,ୢ , is calculated using Equation B-7: (B-7) ‫୴ܨ‬,ୠ,ୢ = ݇୴ ݈ୡ୭୬ ܾ௡௘௧ ݂୴,ୢ with (B-8) ܾ௡௘௧ = (ܽଶ − ݀௛௢௟௘,௠௔௫ )(݊ଽ଴ − 1) where ݇୴ = the reduction factor for shear ‫ݐ‬ୣ୤ = the effective thickness of the plane ݈ୡ୭୬ = the length parallel to grain of the connection ܾ௡௘௧ = the net width between the fasteners over the whole cross-section ݂୴,ୢ = the design shear strength of timber ܽଵ = the spacing of dowel-type fasteners parallel to grain ݊଴ = the number of fasteners in a row parallel to grain ݊ଽ଴ = the number of fasteners in a row perpendicular to grain ܽଷ,௧ = the loaded end distance parallel to grain ܽଶ = the spacing of dowel-type fasteners perpendicular to grain ݀௛௢௟௘,௠௔௫ = the largest of the diameter of the predrilled hole or fastener diameter. For the head tensile plane in block shear failure, the design tensile failure resistance parallel to grain of the head tensile plane, ‫୲ܨ‬,ୢ , is calculated using Equation B-9: (B-9) ‫୲ܨ‬,ୢ = ݇୲ ܾ୬ୣ୲ ‫ݐ‬௘௙ ݂௧,଴,ୢ with 100 (B-10) ‫ܩ‬௠௘௔௡ ݇௩ = 0.9 + 1.4 ඨ ‫ܧ‬଴,௠௘௔௡ (B-11) ܾ௡௘௧ = ൫ܽଶ − ݀௛௢௟௘,௠௔௫ ൯ + (݊ଽ଴ − 1) where ݇୲ = the reduction factor for shear ܾ௡௘௧ = the effective thickness of the plane ‫ݐ‬௘௙ = the length parallel to grain of the connection ݂௧,଴,ୢ = the net width between the fasteners over the whole cross-section ‫ܩ‬௠௘௔௡ = the design shear strength of timber ‫ܧ‬଴,௠௘௔௡ = the spacing of dowel-type fasteners parallel to grain ܽଶ = the spacing of dowel-type fasteners perpendicular to grain ݀௛௢௟௘,௠௔௫ = the largest of the diameter of the predrilled hole or fastener diameter. ݊ଽ଴ = the number of fasteners in a row perpendicular to grain. 101 Appendix C: Stiffness model for CLT In Zarnani and Quenneville stiffness model for CLT, the average tensile stiffness for the head plane (‫ܭ‬௛ ) and the average bottom shear plane stiffness (‫ܭ‬௕ ) can be calculated using the equations given for LVL or glulam in the previous model. The average bottom shear plane stiffness can be defined by summation of the average pure shear stiffness (‫ܭ‬௦௕ ) and the additional average tensile stiffness for the bottom block cross section, (‫ܭ‬௧௕ ) multiply by a reduction factor. Also, ‫ܭ‬௥ and ‫ܭ‬௔ can be expressed using Equation C-1 and C-2: ‫ܭ‬௥ = ‫ܩ‬௥ ‫ݓ‬௠ (‫ܮ‬௖ + ݀௔ ) ‫ݓ‬௖ + ‫ݓ‬௠ ‫ܩ‬௥ (‫ܮ‬௖ + ݀௔ )(‫ݓ‬௖ + ‫ݓ‬௠ ) × = ‫ݐ‬௣௘௥ 2‫ݓ‬௠ 2‫ݐ‬௣௘௥ (C-1) ‫ܭ‬௔ = 2‫ݐܩ‬௣௘௥ (‫ܮ‬௖ + ݀௔ ) 1 4‫ݐܩ‬௣௘௥ (‫ܮ‬௖ + ݀௔ ) × ×2= ‫ݓ‬௖ /2 2 ‫ݓ‬௖ (C-2) The load-carrying capacity of wood in a CLT connection, ܲ௪ , is calculated using the equation below, which accounts for potential block tear-out failures in. ܲ௪ = ݊௣ . ⎧ ܲ௪,஺ ⎪ ⎪ ܲ௪,஻ ⎪ ⎪ ⎪ܲ௪,஼ if ‫ݐ‬௘௙௣ < ‫ݐ‬௣௔௥ ⎨ ܲ௪,஽ ⎪ ⎪ܲ ⎪ ௪,ா ⎪ ⎪ ܲ௪,ி ⎩ if ‫ݐ‬௘௙௣ = ‫ݐ‬௣௔௥ + ‫ݐ‬௣௘௥ if ‫ݐ‬௘௙௣ < ‫ݐ‬௣௔௥ if ‫ݐ‬௣௔௥ < ‫ݐ‬௘௙௣ < ‫ݐ‬௣௔௥ + ‫ݐ‬௣௘௥ (C-3) if ‫ݐ‬௣௔௥ + ‫ݐ‬௣௘௥ < ‫ݐ‬௘௙௣ < 2‫ݐ‬௣௔௥ + ‫ݐ‬௣௘௥ if ‫ݐ‬௘௙௣ = 2‫ݐ‬௣௔௥ + ‫ݐ‬௣௘௥ In this equation, ݊௣ represents the number of plates, and the variables ܲ௪,஺ , ܲ௪,஻ , ܲ௪,஼ , ܲ௪,஽ , ܲ௪,ா , and ܲ௪,ி represent wood resistance corresponding to various failure modes. Wood resistance for failure mode A is determined as the minimum between the wood resistance in head tensile plane failure (‫݌‬௪,௛ ) and the wood resistance in bottom shear plane failure (‫݌‬௪,ௗ ). For failure modes B, C, D, and E, the wood resistance is calculated 102 as the lesser value among head tensile plane, bottom shear plane, and adjacent shear plane failures (‫݌‬௪,௔ ). Additionally, for failure Mode F, the wood resistance is equal to the lesser value between head tensile plane failure and adjacent shear plane failure. The resistance to joint failure induced by head tensile plane failure (‫݌‬௪,௛ ) is calculated using Equation C-4: ‫݌‬௪,௛ = ݂௧ ‫ܣ‬௧௛ ∑‫ܭ‬ ‫ܭ‬௛ (C-4) where ݂௧ represents wood strength in tension parallel-to-grain, determined by the CLT lumber grade; ‫ܣ‬௧௛ denotes the area of the head plane under tensile stress, with variations ‫ݓ‬௖ ‫ݐ‬௘௙ for modes A and B, ‫ݓ‬௖ ‫ݐ‬௣௔௥ for modes C and D, and ‫ݓ‬௖ (‫ݐ‬௘௙ − ‫ݐ‬௣௘௥ ) for modes E and F. The total stiffness of the resisting planes (∑ ‫)ܭ‬, is equal to ‫ܭ‬௛ + ‫ܭ‬௕ + ‫ܭ‬௥ + ‫ܭ‬௔ for mode A, ‫ܭ‬௛ + ‫ܭ‬௥ + ‫ܭ‬௔ for modes B and C, ‫ܭ‬௛ + ‫ܭ‬௕ + ‫ܭ‬௔ for modes D and E, and ‫ܭ‬௛ + ‫ܭ‬௔ for Mode F. The resistance to joint failure triggered by bottom shear plane failure( ‫݌‬௪,௕ ), is calculated using Equation C-5: ∑௄ ‫݌‬௪,௛ = ቐ ݂௩ ‫ܣ‬௩௕ ௄ ೏ ∑௄ , ݂௩,௥ ‫ܣ‬௩௕ ௄ , For modes A and E (C-5) For modes B, C and D ೏ where ݂௩ represents the longitudinal shear strength, ݂௩,௥ is the rolling shear strength, ‫ܣ‬௩௕ is the area under the bottom plane subjected to shear stress. ‫ܭ‬ௗ is the overall stiffness corresponding to the depth below the bottom shear plane equating to ‫ܭ‬௕ + ‫ܭ‬௥ + ‫ܭ‬௔ for mode A, ‫ܭ‬௥ + ‫ܭ‬௔ for Mode H, ‫ܭ‬௥ for mode C, and ‫ܭ‬௕ for modes D and E. The resistance to joint failure induced by adjacent shear planes (‫݌‬௪,௔ ) is calculated using Equation C-6: 103 ‫݌‬௪,௔ = ݂௩,௥ ‫ܣ‬௩௔ ∑‫ܭ‬ ‫ܭ‬௔ (C-6) where ‫ܣ‬௩௔ represents the interface area between the cross layer and the parallel planks adjacent to the loaded planks. This area is equal to (‫ݓ‬௠ − ‫ݓ‬௖ )(‫ܮ‬௖ + ݀௔ ) for mode (B) and 2 (‫ݓ‬௠ − ‫ݓ‬௖ )(‫ܮ‬௖ + ݀௔ ) for other applicable modes. 104 Appendix D: Test series sketches 105 106 107 108 109 110 111 112 113