This study presents an analytical, experimental, and numerical investigation of full-scale glulam trusses connected through birch plywood gusset plates. Three trusses were joined through mechanical connectors, while three were adhesively bonded. Failure was designed to occur in the plywood plates subjected to a multi-axial stress state due to the convergence of three glulam elements in the node. The thickness of these plates varied between 9, 12 and 21 mm. This research aimed to evaluate the structural behaviour and failure mechanisms of birch plywood plates in timber connections. Experimental results showed that glued trusses generally exhibited larger ultimate loads and greater global stiffness compared to mechanically connected trusses. Nevertheless, the global stiffness was independent of the thickness of the central plywood plates. Analytical estimations for specimens that failed in the plywood generally showed good agreement with the experimental capacity despite a slight overestimation, which was attributed to the size effect of the plywood's mechanical properties. Glued trusses with thicker plywood plates (12 and 21 mm) exhibited failure in the glue line between the central plywood plate and the glulam member. Despite the size of the glued area being the same for these two trusses, the experimental capacity differed significantly due to the varying values of wood failure percentage in the glued connections that failed. Furthermore, 2D planar numerical models demonstrated good prediction of the global experimental stiffness for both glued and dowelled trusses.

Earlier studies have demonstrated that birch plywood is a viable substitute for steel plates in timber connections, thanks to its excellent mechanical properties, workability, and sustainability. The mechanical properties of birch plywood have been thoroughly investigated concerning key parameters, such as face grain angle and moisture content. Furthermore, its potential use as gusset plates in timber connections has been studied. However, these investigations were mostly limited to uniaxial tension and rather small-scale compared to real applications. Therefore, additional research is needed to further understand the mechanical behaviour of birch plywood, thus ensuring a safe design of such connections. This thesis aims to gain new knowledge on connections with birch plywood plates. The results of this study will then be used to improve the design of such connections, both in terms of safety and structural efficiency.

The influence of the connector type (i.e. smooth dowel or fully threaded screw) was investigated concerning the shear capacity of connections using birch plywood with varying thickness. The results showed that, when utilizing screws, the analytical estimations according to EN1995-1 Eurocode 5 (EC5) significantly underestimate the shear capacity. This is mainly due to EC5’s conservative estimation of the so-called “rope effect” contribution to the connection’s shear capacity. Moreover, when the number of shear planes was increased from two to four, the discrepancy between EC5’s estimation and experimental results substantially increased regardless of the type of fastener. This substantial discrepancy was associated with the failure mode in the inner shear planes.

Furthermore, both glued and mechanical connections were investigated by performing tests on full-scale glulam trusses joined through birch plywood gusset plates. Failure was designed to occur in the plywood plates subjected to a complex stress state. In this investigation, the thickness of these plywood plates was varied between 9, 12 and 21 mm, testing only one specimen for each configuration. The experimental results showed that glued trusses possessed higher load-carrying capacity and almost twice the stiffness of those with mechanical fasteners. Failure occurred in the plywood plates for all trusses with mechanical fasteners and the glued truss with thinner plywood plates (9 mm). However, the trusses with thicker plates (i.e. 12 and 21 mm) showed glue line failure between the plywood plate and the glulam beam. The analytical estimations for specimens that failed in the plywood plate generally showed good agreement with the experimental capacity, although they exhibited a slight overestimation. This was expected and was attributed to the size effect, as the input strength values of plywood adopted in the analytical estimations were obtained from small-scale tests. 

Furthermore, based on the size of the glued area compared to previous studies, bonding strength values were estimated and used in the analytical calculations of the bonding capacity for glued connections. However, these strength values should be further investigated to calibrate the analytical model and obtain more realistic estimations.

In addition, two-dimensional numerical models demonstrated good agreement in predicting the global stiffness of the tested trusses.

For future work, it is suggested to investigate the size effect on the mechanical properties of birch plywood and bonding strength, respectively, to further calibrate and validate the analytical models. Furthermore, more advanced numerical simulations of the truss specimens are suggested.

Tidigare studier har visat att björkplywood är ett effektivt alternativ till stålplåtar i träförband, främst tack vare dess goda mekaniska egenskaper, bearbetbarhet och hållbarhet. Dessa egenskaper har studerats ingående med avseende på viktiga parametrar såsom fiberriktning och fuktkvot. Dess användning som laskar i träkonstruktioner har även undersökts. Dock har tidigare studier främst varit begränsade till enaxliga dragprov och mindre provkroppar, vilket innebär att ytterligare forskning krävs för att förstå björkplywoodens mekaniska beteende i mer realistiska och komplexa tillämpningar. Syftet med denna avhandling är därför att bidra med ny kunskap om förband med laskar av björkplywood. Resultaten syftar till att förbättra dimensioneringen av sådana förband, med avseende på både säkerhet och strukturell effektivitet.

Typen av fästdon (släta dymlingar eller skruvar) har undersökts i relation till skjuvhållfastheten i förband med varierande plywoodtjocklek. Resultaten visade att EN1995-1 Eurokod 5 (EC5) underskattar skjuvhållfastheten vid användning av skruvar, vilket främst beror på en konservativ uppskattning av den icke-linjära ökningen av motståndet vid glidning som bidrar positivt till förbandets bärförmåga. Vidare ökade avvikelsen mellan de analytiska beräkningarna och de experimentella resultaten avsevärt när antalet skjuvplan ökades från två till fyra, oberoende av typ av fästdon. Denna betydande avvikelse förknippades med brottmoden i de inre skjuvplanen.

Vidare har både limmade och mekaniska förband undersökts genom fullskaleförsök på limträfackverk sammanfogade med skivor av björkplywood. Brott var avsett att initieras i plywooden, vilka utsattes för komplexa spänningstillstånd. Plywoodskivornas tjocklek varierades mellan 9, 12 och 21 mm, och en provkropp testades för varje konfiguration. De experimentella resultaten visade att de limmade fackverken uppvisade högre bärförmåga och nästan dubbelt så hög styvhet som de med mekaniska förband. Vid samtliga fackverk med mekaniska förband, samt det limmade fackverket med de tunnaste plywoodskivorna (9 mm), uppstod brott i plywooden. För de limmade fackverken med tjockare plywood (12 och 21 mm) inträffade dock brott i limfogen mellan plywooden och limträelementet. De analytiska beräkningarna för de prov där brotten uppstod i plywooden överensstämde generellt väl med de experimentella resultaten, även om en viss överskattning noterades. Detta var förväntat och tillskrivs skalningseffekter, då materialegenskaperna som använts i beräkningarna härrör från småskaliga tester.

Baserat på jämförelser med tidigare studier uppskattades även limhållfasthetsvärden för att inkluderas i de analytiska beräkningarna av bärförmågan i limmade förband. Dessa värden bör dock undersökas vidare för att kalibrera och förbättra de analytiska modellerna.

Därtill visade tvådimensionella numeriska modeller god överensstämmelse med den globala styvheten uppmätt i försöken.

För framtida arbete föreslås en fördjupad undersökning av skalningseffekter på björkplywoodens mekaniska egenskaper samt limhållfasthet, i syfte att vidare kalibrera och validera de analytiska modellerna. Även mer avancerade numeriska simuleringar föreslås.

In timber connections, adhesively bonded connections are, in general, cheaper, stiffer, and stronger compared to typically adopted mechanical connections. However, bond line quality is sensitive to various process-related parameters; thus, it is generally preferable to do the gluing in the factory with these parameters controlled during assembly and then transport the entire structure to the construction site. This leads to a limitation on the size of the timber structure owing to the limited transportation capacities. Another concern regarding fully glued timber-to-timber connections falls in their non-ductile behavior prior to the ultimate failure, which is crucial, especially in seismic regions. This paper focuses on the design of two innovative connection systems aimed at overcoming these two key limitations. Both systems offer the advantages of indoor adhesive application and on-site adhesive-free assembly. The first connection system involves a hybrid solution connecting prefabricated elements by means of steel rods and special ‘‘wheel-geared’’ notches of birch plywood, while the second connection system employs pure plywood notched connections during the on-site assembly. These two novel connection systems have potential for use in both moment- and force-resisting applications. In this study, they were introduced and designed in the format of a frame corner, where bending moment, axial force, and shear force are present. Analytical models predicting the capacity for each possible failure mode were developed and then validated by the test results. It can be found that the first connection system exhibits moderate ductile behavior, and its load-bearing capacity is considered to be satisfactory. The capacity can be further improved to be as strong as the fully glued connection if thicker plywood plates are utilized. The second connection system possesses lower strength and stiffness. However, it could still be applied in non-critical connection regions where no substantial external load exists.

This study reveals the magnitude of stress spread angles in the design of plywood gusset plates when subjected to uniaxial tension, with a specific focus on mechanical connections. Plywood plates with elevating widths at three different load-face grain angles were destructively tested. The test series continued with consecutively increased plate widths until the measured forces reached plateaus. Two models, namely, the classic and modified stress spread models, adopted from the Whitmore effective width theory, were investigated to account for the observed phenomenon. The classic stress spread model considers a rigid fastener array and an evenly distributed stress block. A closer-to-reality modified model considers the summation of stress blocks contributed from each fastener line. For both models mentioned, the magnitudes of corresponding spread angles were calibrated utilizing a fitting scheme considering maximized R-square values. The validity of both models was later examined and validated versus the previous experimental data reported in the literature. It was found that the classic model, despite some close predictions, gave over-estimations on the load-bearing capacities of several connection patterns. The modified model was found to be conservative for almost all investigated fastener patterns. Accordingly, a hybrid adoption of stress spread models was suggested.

This study investigates the possibility of utilizing steel rebars as alternative mechanical fasteners in timber structures. Adopting rebars rather than conventional smooth dowels as fasteners in timber-timber connections might have positive aspects in terms of environmental impact and structural capacity. As to the environmental aspect, it should be mentioned that rebars are typically from two primary sources, namely: a) reused rebar offcuts or b) recycled steel scraps. Regardless of different origins, rebars have reduced environmental impact, especially since rebar offcuts are the primary waste source in concrete construction. As to the strength aspect, which is the main focus of this paper, the authors aimed to check whether enhanced load capacity of dowel-type connections could be achieved due to the higher withdrawal capacity of rebars as compared to smooth dowels. To verify this, tension-shear specimens with either two or four shear planes, adopting smooth dowels or steel rebars as the mechanical fasteners, were experimentally investigated. Plywood with thicknesses of 15 mm or 21 mm was adopted to further verify the contribution from the friction, assuming that thicker plywood generates a higher withdrawal capacity. Analytical models in Eurocode 5 were also adopted to predict the shear plane capacity and the connection stiffness during loading phases. Thereafter, a numerical model was constructed to verify the structural benefits of rebars by adopting different friction models. The model aims to explain experimental observations both quantitively and qualitatively.

Previous studies have demonstrated the potential of birch plywood as a substitute for steel plates in multi-plane shear connections of timber structures due to, among other things, its low environmental impact, better workability, and relatively low cost. However, models in modern building codes, such as Eurocode EN1995–1 (EC5), can be used to determine the load-carrying capacity of timber connections with up to two shear planes. Furthermore, some studies have shown that EC5 design model tend to underestimate the actual load-carrying capacity of timber connections. Therefore, there are some uncertainties concerning suitable design models to assess the load-carrying capacity of such connections. This study was preliminary conducted to shed light on the reasons for such discrepancies, conducting several experiments on both doweled and screwed birch plywood-to-timber connections, with either two or four shear planes. The analytical load-carrying capacities estimated by EC5 showed underestimation of the experimental results, with greater underestimation when fully threaded screws are adopted as fasteners, rather than smooth dowels. Furthermore, regardless of the type of fastener, a substantial discrepancy between EC5’s prediction and experimental results was observed when the number of shear planes was increased from two to four. The results of the investigations indicate that the main cause of the discrepancies might be associated with the so-called “rope effect” which is taken into account by EC5’s design approach in an over-conservative manner.

There is growing interest recently in reducing the usage of metals in timber structures. Birch plywood possesses satisfactory mechanical properties compared to other wood-based panels and is promising to be utilized in timber connections as a substitute for the more conventional slotted-in metal plate. There are essentially two possibilities to connect plywood plates and other timber elements by means of either mechanical connections or adhesively bonded connections. Despite the more commonly adopted mechanical connections in current timber structures, the adhesively bonded connections hold the distinct advantages of being more cost-effective, stiffer, and with a lower risk of moisture penetration in the timber elements. When employing birch plywood in timber structure applications such as trusses and frame corners, stresses from different directions need to be transmitted by the plywood gusset plate. However, it is still uncertain how the bonding strength is affected by different loading angles to the face grain. This research question, specifically concerning the bonding strength between birch plywood and spruce glulam, has been addressed in this paper. It was found that the bonding strength varies within a relatively small range when the load-to-plywood face grain angle varies from 0° to 90°, which is promising for the development of adhesively bonded joints. Failure mainly occurred in glulam at 0° and 15°; while at other angles, a mixture of cohesive failure in glulam and plywood face veneer was dominant. The weak angle-dependence of the bonding strength can be explained by further checking the shear strength of the weaker wood adherends between glulam and plywood. A strong positive correlation was observed between bonding strength and the wood shear strength.

This study reveals timber-timber composite joints consisting of glulam pieces and birch plywood plates with three different load-to-face grain angles. Utilizing a similar number of fasteners and arranging the fastener array from narrow to wide, uniaxial tension specimens were manufactured with four different fastener patterns. The thickness of birch plywood was intentionally under-designed so that the failure modes for all connections were net tension failure of birch plywood plates. Thereafter, the influence of the fastener pattern and face grain orientation on the load-bearing capacity and stiffness of the investigated composite joints was studied. The load capacity and nominal strength generally increased when the nail patterns varied from narrow to wide. This observation was associated with the Whitmore effective width theory (load spread angle) in steel gusset plate design. Moreover, to derive valid analytical methods to predict the net-tension capacity of birch plywood plates, the classic spread angle model that assumes rectangular stress blocks and the modified spread angle model that considers the summation of stresses from each fastener row were discussed. Both models were adopted to predict the net tension capacity of investigated specimens at 0°. In addition, the stiffness of joints was measured and compared with slip modulus formulas in Eurocode 5. The measured local stiffness values were found to be independent of the fastener pattern and load-face grain angles. The analytical slip modulus assuming the case without predrilling exhibited a satisfactory prediction, while formulas assuming the case with predrilling tend to give overestimations.

This study investigates frame corner joints built of birch plywood plates and glulam elements connected via self-tapping screws. Analytical calculations based on the fastener group's torsional moment resistance, the proposed fastener group's elastic and post-elastic load-bearing criteria, and the design formulas in Eurocode 5 were performed to predict the connection capacity in both elastic and post-elastic stages. A combined action check formula was adopted to predict the capacity of birch plywood plates and glulam elements. Frame corner specimens constructed with three different plywood thicknesses were planned to study the influence on global behavior and rotational stiffness. The specimens were intentionally designed so that failure occurred either in plywood or in glulam, in order to examine the robustness and validity of analytical calculation models. Another supplementary test group with 21 mm plywood and fewer fasteners was also designed and tested, in which the plastic yield of fasteners was expected. The test results of this supplementary group served to calibrate the analytical model that predicts the elastic and post-elastic capacity of the connection group. As a result of the comparison, the analytical calculations gave reasonable predictions on the failure of plywood, glulam, and the capacity of the fastener group. Only when the exposed moment exceeded the post-elastic limit of the fastener group did the plastic yielding of fasteners become observable. Moreover, numerical finite element models adopting the foundation zone-modeling scheme were constructed, which were proven to capture all test configurations' linear loading stiffness satisfactorily.