Project Topics
Get Complete Project Material File(s) Now! »
Towards Adhesive-Free Timber Buildings (AFTB) project
As Glulam and CLT are obtained by gluing several timber layers, significant petrochemical adhesives are used and this is pointed out as a concern that affects their sustainability, recyclability and broader environmental impact. More specifically, the adhesives under high-temperature conditions can emit toxic gas, which could cause cancer for people who inhale it (International Agency for Research on Cancer (2004)). In addition, constant exposure with the adhesives could lead to allergic skin reactions (Frihart and Hunt (2010)). On the one hand, the amount of CLT panels produced in 2016 in Europe was estimated to 680,000 m3 and expected to increase to 1,250,000 m3 by 2020 (Unece (2017)). On the other hand, the European Commission (2001) has a specific policy aiming to develop a low-carbon economy by reducing greenhouse gas emissions to 80% by 2050. There is, therefore, a need to develop more “green” and sustainable joining techniques to meet all acceptable principles related to life cycle assessment (LCA) approach of buildings. In this context, a consortium with six partners from six European countries was formed through a joint research project funded by the North-West European Interreg Programme, called AFTB, to develop adhesive free engineered wood products (AFEWPs), including adhesive free laminated beam (AFLB) and adhesive free CLT (AFCLT) using densified wooden dowels as a joining element to substitute the traditional toxic adhesives (Figure 1.2) . Here, the AFEWPs make use of irreversible moisture-dependent swelling of thermo-mechanically compressed wood dowels to achieve tight fitting of connections.
History of using wood dowel in timber structure
The use of wood dowels as joint material is not new. Hardwood dowels have been successfully used to connect structural members in timber frames for thousands of years (Figure 1.3).
Unfortunately, hardwood dowels (not densified) suffer from creep/relaxation and loss of stiffness over time. Some works presented adhesive free laminated timber assembled with hardwood dowels or nails (Ramage et al. (2017); Structure Craft (2018)) and referred to as “Brettstapel”.
According to Henderson et al. (2018), the earliest design of the Brettstapel technology was developed in the 1970s and involved the use of nail fasteners; however, about two decades later, metallic fasteners were replaced with the use of hardwood dowels (Figure 1.4b), since 1999. These EWPs fabricated without the use of metal fasteners or adhesives, are more sustainable and environmentally-friendly.
In recent years, another concept of wood mass timber panel (Figure 1.5), called dowellam (Dowel Laminated Timber) was designed and developed by Structure Craft in North-America (Vancouver Canada) (Structure Craft (2018)).
More recently an European Technical Assessment (ETA) was obtained and published by the Thoma Holz100 company (Deutsches Institut für Bautechnik (2018)) on the mechanical performance of AFCLT panels assembled using beech dowels (uncompressed) (Figure 1.6). In that document only data regarding the flexural behavior under transverse loading was given, the vibrational serviceability comfort was not assessed.
Furthermore, Rombach developed a design for Nur Holz panels that use threaded beech screws to connect the timber laminae, as shown in Figure 1.7. There are more than 300 buildings made of Nur Holz panels have been built worldwide (Habitat Naturel (2009)).
Some research works have been undertaken to investigate the structural behaviour of laminated timber beams assembled using rotationally-welded beech dowels (uncompressed) (O’Loinsigh et al. (2012a), (b)). Jin et al. (2015) undertook an experimental program to investigate the feasibility of wood-based two-dimensional lattice truss core sandwich timber structures using glued birch dowels.
However, it is believed that connections made with normal hardwood dowels could suffer from low stiffness and strength as well as from creep/relaxation and a loss of stiffness over time. The use of compressed (or densified) wood dowels could overcome these weaknesses as compressed wood has greater mechanical properties by increasing its density during the compression process (Guan et al. (2010); Sotayo et al. (2020a)). Also, the shape-memory effect of compressed wood, also known as irreversible swelling (springback) makes permanent tight-fitting of connections (Figure 1.8).
Performance of timber structures assembled by compressed wood dowels
Hassel et al. (2008) proposed a wooden block shear wall system which uses compressed wood as connecting elements. The springback of the compressed wood elements causes a tighter fit of the wooden blocks, thereby improving the system stiffness. Moreover, compressed wood elements absorb most of the stress and damage, allowing the structure to be readjusted and reused after earthquakes.
Guan et al. (2010) studied the structural performance of beam-to-column connection using compressed wood dowels and plates. It can be seen a high performance of the connection from the moment-rotation relationship.
Mehra et al. (2018) assessed the structural performance of beam-to-beam connection system using compressed wood dowels and plates. The load carrying capacity, bending stiffness, maximum moment capacity and rotational stiffness of the connection system were compared to that obtained from a similar connection system using steel dowels and plates. The mean failure load for the compressed wood connection system is only 20.3% less than that achieved for the steel connection system because the mechanical properties of compressed wood are smaller than that of steel. More recently, the mechanical performance of AFLB and AFCLT panel was evaluated in El-Houjeyri et al. (2019); Sotayo et al. (2020b). However, at our best knowledge the vibrational performance of timber structure assembled by compressed wood dowels has not been published elsewhere.
Vibration performance of timber structures
Measurements of vibration properties of timber floors, including CLT floors, have been performed by many researchers (Jarnerö et al. (2015); Chúláin et al. (2016); Rijal et al. (2016); Weckendorf et al. (2016); Ussher et al. (2017a); Ebadi et al. (2018); Suárez-Riestra et al. (2019); Zhang et al. (2019)), among others. These studies include natural frequencies, modal shapes, damping ratios as well as the effect caused by type of floors and boundary conditions on those dynamic properties. Many studies (Glisovic and Stevanovic (2010); Hu and Gagnon (2012); Negreira et al. (2015); Ussher et al. (2017b); Casagrande et al. (2018); Huang et al. (2020)) have been performed to measure motion responses of CLT floor caused by human footsteps, thereby improving the understanding of vibration serviceability performance of CLT floor and getting better design guidelines for residential timber floors.
Recently, the structural dynamics in low-frequency range of dowellam floor was assessed by Filippoupolitis et al. (2017). The results show that the dowellam floor meets the vibration serviceability requirements of Eurocode 5.
However, on the one hand, up to now there is no published works dealing with the vibrational serviceability comfort of adhesive free engineering wood products (AFEWPs) assembled through compressed wood dowels. On the other hand, the vibrational serviceability comfort of AFEWPs using compressed wood dowels is not covered by any building standards. There is, therefore, a need to assess the dynamic characteristics of such new products towards their utilization in construction.
Production of adhesive free laminated beams and adhesive free CLT panels
In adhesive free laminated beams and CLT panels, dowels are important parts. They connect all laminae and resist the separation of laminae vertically and horizontally by bending (O’Loinsigh et al. (2012a)). They are subjected to greater equivalent stress compared with laminae when the laminated beam is bent. When the mechanical properties of dowels are improved, the stiffness of the beam and panel may be improved (O’Loinsigh et al. (2012b)). Therefore, compressed spruce with higher mechanical properties compared with normal timber is the most suitable material for dowels.
Spruce timber blocks with average density value of 437 kg/m3, with 10±2% moisture content, were used to manufacture compressed spruce dowels. The compression of the spruce blocks was done, in the radial direction, using a hot-pressing machine (Figure 2.1a) with a maximum pressure force of 1500kN.
Adhesive free laminated beams (AFLB) and adhesive free CLT (AFCLT) panels
The manufacture and testing of AFLB is necessary as a first stage to access the vibrational characteristics of the adhesive free multilayered timber structures. On one hand, the study of AFLB gives us a good understanding before working on the adhesive free timber floor. Indeed, the deflection of AFLB is mainly by bending in the long direction while the flooring system may be bent in both long and wide directions. On the other hand, AFLB can be produced in large numbers to compare the influence of parameters such as the number of dowels, material of dowels and layers. The smaller scale of AFLB compared with the floor can help us save material and manufacture time.
The AFLB were manufactured by stacking three timber layers having the following dimensions: 70 mm x 22.5 mm x 1450 mm (Figure 2.4). Two timber species have been considered for the laminates, namely spruce timber and oak timber. All the laminae were weighed and moisture content measured prior to testing to obtain information about the density. The dowels were either made of 68% compressed spruce or normal oak timber with 16 mm diameter (d =16 mm). Two different dowel spacings were considered, namely 7 times and 3 times the dowel diameter (7d and 3d), involving 13 dowels and 27 dowels per beam, respectively. In addition, several three-oak-layer and three-spruce-layer glued beams with similar dimensions compared with AFLB were manufactured and tested for comparison.
Moisture content condition
Moisture content has significant effect on mechanical behavior of wood (Gülzow et al. (2011); Ido et al. (2013); Silva et al. (2012)). Therefore, all single layer beams, AFLB and AFCLT panels were tested at 9% moisture content corresponding to the ambient environment (temperature and relative humidity) at the LERMAB laboratory at time of testing. However, taking into account the effect of moisture content on vibrational characteristic of AFLB seems relevant and necessary. The influence of moisture content on frequencies of a single layer beam, a glued beam and an AFLB is presented in Section 2.5.3.
Repeatability test
Repeatability tests were conducted on beams and AFCLT panel systems prior to the modal analysis in order to check the measurement uncertainty. Figures 2.9 and 2.10 show examples of FRF for both AFLB and AFCLT panel which were subjected to two series of hammer impacts. Each series of test contains five hammer impacts for which a mean FRF is calculated.
Table of contents :
Chapter 1: Introduction
1.1 Background and motivations
1.1.1 Multilayered timber structures
1.1.2 Towards Adhesive-Free Timber Buildings (AFTB) project
1.1.3 History of using wood dowel in timber structure
1.1.4 Performance of timber structures assembled by compressed wood dowels
1.1.5 Vibration performance of timber structures
1.2 Objective of the thesis
1.3 Outline
Chapter 2: Experimental modal analysis
2.1 Production of adhesive free laminated beams and adhesive free CLT panels
2.1.1 Compressed spruce dowels
2.1.2 Adhesive free laminated beams (AFLB) and adhesive free CLT (AFCLT) panels
2.2 Experimental set-up
2.2.1 Hammer impact excitation
2.2.2 Moisture content condition
2.2.3 Repeatability test
2.3 Characteristic of wood material
2.4 Single layer beams
2.5 Three-layer AFLBs
2.5.1 Effect of compressed wood dowels
2.5.2 Effect of the number of dowels
2.5.3 Effect of moisture content
2.6 AFCLT panels
2.7 Academic contribution
2.8 Conclusion
Chapter 3: Finite element models
3.1 Introduction
3.1.1 Existing finite element (FE) models
3.1.2 Verification and Validation methodology
3.2 FE model for single layer beam
3.2.1 Sensitivity analysis
3.2.2 Verification of the model
3.2.3 Validation of the model
3.2.4 Effect of homogeneous material assumption
3.3 FE model for three-layer AFLB
3.3.1 Sensitivity analysis
3.3.2 Verification of the model
3.3.3 Validation of the model
3.3.4 Effect of stiffness of CSD on frequencies of AFLB
3.4 FE model for AFCLT panel
3.4.1 Verification of the model
3.4.2 Validation of the model
3.5 FE model for full-size AFCLT panel
3.5.1 Prediction of frequencies and modal shapes
3.5.2 Comparison with EC5 requirements
3.5.3 Optimization for the first frequency
3.6 Simplified FE model
3.6.1 Simplifying dowels by beam elements
3.6.2 Simplifying layer by shell elements
3.6.3 Simplifying shape of dowel
3.6.4 Computational cost effect
3.7 Academic contribution
3.8 Conclusion
Chapter 4: Variability with MSP
4.1 Introduction
4.1.1 Short overview on non-deterministic methods
4.1.2 Presentation of the Modal Stability Procedure (MSP)
4.2 General MSP formulation
4.3 Development of MSP formulation for 20-node hexahedral solid element
4.3.1 MSP mesh convergence
4.3.2 Comparison between MSP and FE model in the nominal case
4.4 Uncertain parameters and distribution law
4.5 Assessment of the MSP to calculate variability
4.5.1 Error indicator
4.5.2 Comparison between MSP and direct MCS for single layer beam and AFLB
4.5.3 Error indicator for AFCLT panels
4.6 Influence of uncertain properties of compressed wood dowels
4.7 Comparison between MSP and experimental results
4.7.1 Comparison for single layer beam
4.7.2 Comparison for AFLB
4.7.3 Comparison for AFCLT panel
4.8 Prediction variability of full-size AFCLT panel
4.9 Computational cost
4.10 Academic contribution
4.11 Conclusion
Chapter 5: General conclusion and perspectives
5.1 General conclusion
5.2 Perspectives
Bibliography
