Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Der Artikel geht auf die strukturelle Verwendung tragender Holz-Glas-Verbundwände ein und adressiert die inhärenten Herausforderungen, die sich aus der Sprödigkeit und Temperatursensibilität von Glas ergeben. Es hebt das Potenzial der Kombination von Glas und Holz hervor, um die Festigkeit und Duktilität nach dem Versagen zu verbessern und damit die Robustheit der Konstruktion insgesamt zu verbessern. Die Studie präsentiert eine detaillierte experimentelle Analyse einer großflächigen Holz-Glas-Verbundwand unter Brandbedingungen und enthüllt die komplexen Wechselwirkungen zwischen Glas und Holzkomponenten. Numerische Simulationen ergänzen die experimentellen Ergebnisse und liefern ein umfassendes Verständnis der thermischen und mechanischen Leistungsfähigkeit dieser Verbundsysteme. Die Forschung unterstreicht die Bedeutung von Feuerwiderstandsmechanismen und Ausdauer und bietet praktische Empfehlungen für zukünftige Untersuchungen und Anwendungen. Der Artikel diskutiert auch das Verhalten von Glas und Holz bei erhöhten Temperaturen, die Effektivität verschiedener Verbundstrategien und die entscheidende Rolle der thermischen und mechanischen Leistung bei der Gewährleistung der strukturellen Integrität. Die Publikation widmet sich diesen Schlüsselthemen und zielt darauf ab, das Verständnis und die Anwendung von Holz-Glas-Verbundplatten im modernen Bauwesen zu fördern und so den Weg für widerstandsfähigere und nachhaltigere Baulösungen zu ebnen.
KI-Generiert
Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.
Abstract
Fire accidents are a critical design condition for load-bearing elements in general. Among others, ordinary glass and composite glass materials are even more susceptible to fire and require the use or definition of specific test protocols, simulation strategies, performance indicators and validation methods. In this paper, the structural performance of a full-scale composite timber-glass composite wall (consisting of a perimetral timber frame and a double thin insulating glass unit (IGU)) under the effects of sustained mechanical loads (25 kN/m, as in a typical two-story building) and fire exposure is investigated based on a standard test furnace. The mechanical concept uses a laminated system that can cover an area of up to 3.2 × 2.7 square meters, with a relatively low thickness (63.52 mm for the double insulating glass unit (IGU), including cavity). A great advantage to evaluate the potential and critical points of the composite timber-glass composite system comes from experimental and finite element (FE) thermomechanical investigations. A pilot test is being conducted on a prototype prefabricated timber-glass module, which is expected to function as an efficient load-bearing system in buildings, withstanding the typical mechanical loads from normal or extreme actions, but also providing adequate resistance to fire accidents. The laboratory investigation was carried out on the basis of conventional recommendations for the experimental assessment of building components in the event of fire, with the main focus on estimating fire resistance. It has been shown that the overall load-bearing capacity and the corresponding fire resistance are mainly determined by the intrinsic properties of the glass components, which may need to be protected or optimized to ensure adequate residual capacity.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
The inherent brittleness, its much lower tensile strength than compressive strength, the risk of fracture due to virtually unavoidable defects and its decreasing strength over time make glass a challenge for structural applications. To overcome these undesirable failure characteristics in glass structures, several promising solutions have been tested and developed in recent years. These include applications where glass is combined with other materials to increase post-failure strength and ductility, thus improving post-failure performance and overall design robustness [1‐5]. Recent trends in design also consider design aspects of sustainability and the circular economy, where timber-glass composites could offer some positive advantages over more traditional solutions. One such recently developed concept uses timber flanges connected to glass webs, creating timber-glass composite beams with effective load-bearing behaviour [6‐14] The same idea can be applied to walls where a pane of glass is connected to a timber frame [4, 15‐17].
A differently designed load-bearing system was presented in the studies by [18]. The load-bearing glass panel as a rigid diaphragm is not attached to the timber frame with rigid (glued) connections, but the connection is achieved by direct contact and friction [19]. On the other hand, the connections between the elements of the timber frame were realised by using glued-in rods. This type of system achieves a more ductile behaviour than the previously mentioned ones. If the load-bearing capacity is exceeded, cracks occur in the area of the corners of the timber frame, so that the glass remains intact, which contributes to safety [20]. In addition, the system can be easily retrofitted in the event of failure, as the individual elements can be replaced or repaired independently of each other. There are only a few practical applications of timber-glass composites in the literature and most designs are still at the concept stage.
Anzeige
The idea developed by Kreher was applied to the roof structure of the Hotel Palafitte in Monruz, Switzerland [21, 22]. The beams support a lightweight roof and transfer snow and wind loads to steel posts concealed in the outer walls. Each of the 6000 mm long and 580 mm deep I-shaped composite beams consists of a single glass plate and timber flanges bonded on either side of the glass walkway. The upper flange consisted of two solid timber blocks 100 × 160 mm2, while the lower flange consisted of two timber blocks 65 × 65 mm2.
Using a pedestrian bridge as an example [23] present the entire design process, from the early concept phase to numerical simulations, fabrication and final design. As far as the authors are aware, the only economically successful implementation of a timber-glass composite system in a façade system is reported in [24, 25]. This application includes glass panes that are glued to timber connecting strips with a silicone adhesive and mechanically attached to the main timber structure. Large glass panes are used in a recently developed concept called timber-glass buildings, where large windows together with the main timber structure enable high energy efficiency.
However, the use of timber-glass elements in practise raises an important issue related to the fire behaviour of the structure. The behaviour of timber in fire and of adhesives at high temperatures is a well-known phenomenon. To date, however, little is known about the effects of temperature variations on the mechanical performance of structural glazing components and assemblies [26, 27]. This applies in particular to the high temperatures that can be expected under real fire conditions [28].
In addition, glass elements absorb and transmit a considerable amount of solar radiation even under normal conditions [29]. The absorbed radiation significantly increases the temperature of the glass, which can reach 50–80 °C in summer (depending on whether and what type of coating is used). These temperatures also have a major influence on the mechanical properties of the interlayers and therefore on the bending properties of laminated glass elements out of plane.
Anzeige
This paper discusses some important issues that are relevant to the efficient application of timber-glass composite panels in practise, where fire safety plays an important role, and should be further elaborated. The timber-glass composite panels presented consist of true-to-scale glass panels that are held at the glass edges by wooden components. As shown, the thermal and mechanical performance and capacity of such a composite system depends not only on the behaviour of the individual materials and components, but also on their mutual interaction under both ordinary and extreme accidental loads, such as fire events. Accordingly, a sound scientific knowledge on fire resisting mechanisms and fire endurance is of utmost importance [30] but necessitates of specific attention and methodologies [28].
In this context, Sect. 2 first summarises the behaviour of glass and timber at elevated temperatures and under fire conditions. Some examples of strategies for timber-glass composites are briefly recalled. The methodology and corresponding results of an original test for a timber-glass composite wall under fire conditions are then presented in Sect. 3. The experimental findings are used to develop a full three-dimensional finite element (FE) numerical model and a detailed thermomechanical analysis of the same prototype, highlighting some important mechanisms and possible critical issues. Finally, general conclusions and practical recommendations are given for future investigations.
2 Glass and Timber in Fire
2.1 Glass
Glass is not flammable and does not cause dripping, which is of course an advantage. Nevertheless, the material also has serious disadvantages that should be optimally addressed, especially with regard to structural safety and efficiency. In a real fire scenario, glass heats up relatively quickly, resulting in large temperature gradients, both in thickness and surface area, and uneven temperature distributions [31]. In general, for components made of ordinary glass, as in the present study, it is assumed that the temperature gradient reached will most likely lead to breakage of the glass pane exposed to fire within a few minutes. This therefore inevitably requires the use of resistant sections consisting of several panes. The use of laminated glass elements is, as is well known, unavoidable in constructions for a number of reasons [32]. Laminated glass with multiple panes must be however examined with regard to the loss of the composite effect, due to the possible loss of stiffness of the interlayer (Fig. 1a). In the first (full) case, the laminated glass behaves as a complete, intact system, where both the glass panes and the interlayer contribute to the overall mechanical behaviour. The interlayer provides additional strength and energy absorption, allowing the glass to resist bending without significant failure.
Fig. 1
Example of a double laminated glass section [34], with b glass [34] and c PVB moduli as a function of temperature [35]
However, in the second (null) case the contribution of the interlayer is considered negligible or absent. This could represent a scenario where the interlayer is either defective or suffers for unfavourable external effects. The bending behaviour is dominated by the unbonded glass layers alone, resulting in a reduced overall strength and higher susceptibility to failure. Finally, in the third (partial) case, the interlayer somehow contributes to the composite mechanical behaviour of the laminated section. This corresponds to scenarios where the interlayer is only partially mechanically effective. The bending resistance is thus compromised, but not to the extent observed in the null case. Again, in fire conditions, the heat quickly heats and even melts all the panes and the interlayers [30]. The interlayers of polymeric materials soften very rapidly [26, 30] and this effect could result in major consequences especially for glass components in out-of-pane bending [33]. For shear walls, as for vertically oriented glass members, the overall load-bearing capacity and even buckling resistance decreases due to the lower bending stiffness of the glass sheets, which is further affected by non-uniform propagation of high temperature on the elevation [31].
Still little is known about the effects of temperature variations on the performance of structural glazing [26, 36‐38]. Glass-related phenomena such as thermal shock, frame constraint and shading effects are critical for glass when exposed to high temperatures and fire scenarios [29]. The density of glass as a function of temperature has been studied by [39, 40].
Regarding the material properties modification, a review of past literature efforts can be found in [26]. In general, the density of glass decreases with increasing temperature. Below the glass transition temperature (Tg), which is approximately in the range of 400–550 °C [34, 41], the decrease in density is primarily due to thermal expansion of the glass volume [42]. Above Tg, the decrease is faster due to the higher coefficient of expansion in the supercooled liquid state. In addition, near Tg there is a time dependence due to structural relaxation effects that depend on the thermal history of the glass. The density is reduced by about 5% at 900 °C compared to ambient temperatures.
The modulus of elasticity, tensile strength and coefficient of linear thermal expansion of glass can significantly affect fire exposure behaviour and subsequent cracking and fallout of broken glass [26]. The progressive increase of temperature corresponds to a major decrease of the modulus of elasticity [43, 44], due to accelerated corrosion [45], with a consequent loss of bending stiffness, migration of stress peaks in glass and increase of the measured deflection rate [30, 33].
Thermal expansion is an important property in predicting thermal stresses that develop in a component under fire conditions. In general, the linear coefficient of thermal expansion increases with increasing temperature, but the relationship is not linear. The temperature dependence of the expansion is closely related to the temperature dependence of the density [45].
Linear thermal expansion coefficients were established (using a polynomial curve fit to the results reported by [45] for the same temperature range as for the elastic modulus. Although tensile strength is a critical parameter for glass, it is the least known among the other parameters [46]. The strength parameters of glass at elevated temperature are very difficult to test and definitely require further studies [47, 48]. The literature results presented for ordinary glass under high temperature, in some cases, show considerable discrepancies and cannot be directly compared [26]. In general, the strength of glass decreases with temperature, but dropping dramatically especially at Tg and higher. The above review concludes that advanced numerical material models are needed to capture the material behaviour of glass, especially when the temperature is above Tg, and the corresponding thermo-mechanical load-bearing performance in fire [28, 30, 33]. More sophisticated material models have been discussed also by [49, 50], based predominantly on the past work of [51‐53].
2.2 Timber
Timber is understood as a material that burns easily, which may be a misinterpretation in the case of structural behaviour. In the case of structural timber, the size effect is an advantage, as the area exposed to fire determines how it burns. Timber burns from the outside towards the thickness at a fairly known rate, leaving the core material structurally intact in terms of stiffness and strength. The burnt timber part becomes a layer of charcoal that loses all strength but retains its role as an insulating layer that prevents excessive temperature rise in the core [54].
This is generally advantageous in timber structures compared to other building materials (steel loses stiffness when heated; concrete tends to flake from the outside, exposing the reinforcement to fire). The usual approach is therefore to oversize the timber cross-section to provide a sufficient core, or to protect part of the external surface from fire by means of suitable fire protection (e.g. plasterboard).
The unity of a timber-glass composite can also be studied from the point of view of the scenario in which the elements are used, as the requirements are different if the elements are used in a load-bearing case (REI classification according to [55] or “simply” as partitions (EI classification according to [55]. There are different solutions, depending on the degree and duration of fire protection through a heat-reflecting surface coating that minimises heat radiation or intermediate layers of water glass that swell and form a foam barrier that prevents the spread of fire. The solutions available on the market offer certifications not only for the glass part itself, but also for the respective frame details.
2.3 Composite Systems Made of Timber and Glass
Timber-glass (or glass-timber) composite systems with a bonded connection can be viewed from two different perspectives [56], also depending on the primary role of glass or timber respectively, and on the way they are expected to mechanically interact and ensure the required load-bearing capacity. In the first idea, the load-bearing glass pane is protected by a fire protection layer, which is also made of glass. The second approach relates to the loss of the adhesive bond between glass and timber, which is caused by the lack of resistance of the adhesive at elevated temperatures. This includes mechanical bonding systems that are activated after the failure of the adhesive bond.
In the event of glass breakage, the timber frame can still carry a significant part of the vertical load. If the timber frame is sufficiently resistant, on the other side, it is therefore necessary to allow and ensure a minimum resistance of the glass infill. The same can be achieved by using toughened glass for shorter periods and coatings or infills between the glass panes if a longer period of time is to be achieved during which the glass must not fail.
To achieve an R classification (or a combination with especially E and or I according to [55], additional measures must be taken. An earlier paper by [57] investigated the case of fire in timber-glass composites using load-bearing wall elements as an example. The case of composite wall elements was studied, with the main interest being the use in external walls, assuming that only one side is exposed to fire, i.e. the internal side, and excluding the case of fire from the outside. The study suggests separating the functions of fire protection and load-bearing capacity to ensure both aspects. A two-layer structure decouples the function so that there is an EI layer on the fire side and an R layer on the outside. An important aspect is to ensure the integrity of both layers, i.e. not to jeopardise the respective function, i.e. to ensure that the load-bearing layer does not become (too) hot and that the fire protection layer is not subjected to disproportionate loads.
Implementing the solution described for wall elements would result in a three-layer structure with a fire protection layer on both sides of the load-bearing layer. Although this solution is feasible, it may not be practical as the structure would become too thick, as the layers should not be too close together to ensure sufficient load-bearing capacity in the load-bearing layer. Another solution could be to combine only one layer of glass with the load-bearing and fire-protection layers. Here too, the inner glass panes are used for load-bearing capacity, while the outer panes are used for fire protection.
Depending on the security level, several sacrificial layers may have to be laminated together [56]. The biggest problem could be how to ensure the functionality of fire protection and load-bearing functions. As there is only one glass package, mechanical decoupling cannot be easily ensured as all layers function together. In the warm case, only the inner panes are taken into account for the load-bearing function, whereas in the cold case, all layers are taken into account. The warm case is also special because the outer layers are broken, while the composite effect of glass and timber frame elements is still intact.
3 Summary of Experimental Analysis
3.1 Full-Scale Specimen
The load-bearing prototype investigated consisted of a frame of timber components and an insulating glass unit (IGU) to support the composite system in Fig. 2 and cover an area of up to 3.2 m wide × 2.7 m high.
Fig. 2
Test specimen for the pilot experiment: axonometric view and detail drawings (measures in mm)
The transoms and mullions for the bracing frame were characterized by timber profiles with nominal dimensions of 160 × 100 mm. The angle connections were also made with wooden screws (Φ 10 × 280 mm, type Assy 3.0 SK). The IGU panels consisted of double-laminated glass elements (2 × 10 mm thick float layers, each with 1.52 mm PVB), with a 13 mm cavity between them. The IGU elements were held in place by timber frame battens that were attached along the glass edges, on both sides and around the entire perimeter of the frame. The battens were specially profiled to create grooves that prevented the glass panes from falling out, but at the same time prevented the possible ingress of air and water.
The sample was fixed to the foundation of the oven using 10-point fixings with L-shaped steel elements (285 mm apart), as shown in Fig. 3. The vertical edges of the sample were left free of mechanical constraints. At the same time, the space between the vertical edges of the sample and the furnace was filled with ceramic wool.
Fig. 3
Specimen preparation in the test furnace: a base connection detail and b front view
The furnace experiment on the full-scale wall prototype was carried out in the LTM—Laboratory for Thermal Measurements in Oroslavje (Croatia). More precisely, the sample was placed in the furnace as in Fig. 4a. Before the beginning of the fire experiment, the façade element was subjected to a constant, uniformly distributed vertical load of 25 kN / m (corresponding to a resultant vertical force of 80.55 kN), that was kept fix until failure (Fig. 4b).
Fig. 4
Experimental setup: a thermocouples in the oven, b thermocouples on panel; c A–A cross-section view
Fire exposure was carried out with six blowpipes on liquid fuel, according to [58], (see Fig. 4c). The air temperature in the test room was kept at 20 °C (± 5) for 24 h before the test.
The heating setup was then determined by using the standard ISO time–temperature curve [58]. In the meanwhile, the oven temperature was measured using 6 evenly distributed plate thermocouples (type K). The direct contact of these thermocouples with the open flame was properly avoided to protect instruments, and a distance of 100 mm was guaranteed from the wall surface exposed to fire. The static pre-pressure in the fire area was kept in the range of 15 Pa (± 2). The sensor was placed at a height of 2.3 m above the floor for pressure measurement.
3.3 Key Performance Parameters
In addition to the criteria that should be considered when analysing the fire resistance of load-bearing components and systems (see Sect. 3.4), it is important to remember that the thermomechanical performance of a full-scale prototype such as the system under study is strongly influenced and determined by the response of the glass material to high temperatures (Sect. 2).
As such, it is useful to report herein some of the conventional values of allowable thermal gradient, that standards and technical guidelines propose for a first simplified analysis of glass integrity (Table 1). In particular, Table 1 reports the relevant values for float glass only, which is the material in use for the present study.
Another important aspect to examine is the combined effect of simultaneous thermal and mechanical stresses [28]. Insofar as the prototype under investigation is subjected to a constant uniform vertical load of 25 kN/m (i.e. simulating the permanent load in a real building), the corresponding stresses and strains are superimposed on the effects of fire and thermal loading [30, 33]. As such, all these multiple aspects should be properly addressed for these types of load-bearing materials and assemblies, especially in consideration of failure mechanisms and fire endurance analyses [28].
3.4 Experimental Results
3.4.1 Temperature
Overall, the temperature experimental records are summarised in Fig. 5.
Fig. 5
Temperature evolution in time: a oven temperature (imposed and average measure), with b corresponding local measurements, and evolution of temperature trends on the unexposed side of the specimen, in terms of c maximum and d average measurements
In the analysis of experimental findings, the average temperature in the oven was found to agree rather well with the imposed standard ISO time–temperature curve. Figure 5a presents in fact the measured average temperature, according to the 6 evenly distributed plate thermocouples, and some deviations from the standard ISO curve can be observed for the first steps of test only.
Figure 5b, in this regard, proves the detailed temperature trend in the oven, and emphasises that during the test it was not completely evenly distributed. In any case, it is also worth noting that the partially lower temperature in Fig. 5b corresponds to control points P1 and P2, that were positioned in the bottom region of the oven and sample, while a more uniform temperature distribution can be noted for the other control points. Such evidence recalls the critical role of thermal loading distribution, which was also addressed in [31] for vertical glass components that usually suffer for additional thermal issues, due to the non-uniform spread of real fire events.
Regarding the thermal exposure of the wall specimen in fire, a set of 21 thermocouples in total was used to track its response. More specifically, as also schematically described in Fig. 4, the measurements for the average temperature rise on the unexposed side of the sample (i.e. the 140 K criterion [59]) were performed at 5 different control points on glass (T1–T5). One of them (T1) was specifically placed in the centre of the sample, while the others (T2–T5) were equally distributed on the wall surface, and placed in the centre of each quarter of the exposed glass surface.
In addition, the measurement of the maximum temperature rise on the unexposed side was also carried out (criterion 180 K [59]), using the following instruments:
T6–T10: glass (50 mm from slats).
T11–T13: slats (mid-span sections).
T14–T18: glass—next to the slats.
T19–T21: transoms (external edges).
Figure 5c, d present the corresponding evolution in time of the maximum and average temperatures respectively. For clarity of presentation, Fig. 5c groups the experimental results for instruments T1–T10 and T11–T21. In particular, Fig. 5c shows a maximum temperature peak in the centre of the specimen (T1), which mostly coincides with the top right corner (T3), and raises to ≈ 82 °C and ≈ 81 °C respectively in about ≈ 11 min of test. The other control points on glass revealed maximum temperature peaks in the order of ≈ 68 °C at the top left corner (T2), ≈ 42 °C at the bottom right corner (T5) and ≈ 30 °C at the bottom left one (T4).
The analysis of average temperature as in Fig. 5d shows partially reduced peaks at the end of the experiment. The highest temperature peak was in fact reached at the right upper-middle edge (≈ 51 °C in T9 and ≈ 49 °C in T10), followed by ≈ 28 °C in T7 and T8 (upper middle and upper left) and ≈ 17 °C in the left middle (T6). Globally, an average temperature up to 60 °C was derived, at the end of the test, from the elaboration of T1-to-T5 measurements, as it can be seen in Fig. 5d.
3.4.2 Vertical Deflection
During the test, the vertical deflection was measured at two points next to the free edges. In other words, the vertical deflection of the horizontal beam was measured. The layout of measurement points is shown in Fig. 6.
The experimental measurements are shown in Fig. 7 for deflection parameters. It is worth noting that the failure time of the specimen is associated with values for deflection and deflection rate that are significantly lower compared to the standard reference values. The absolute vertical deflection at the end of the test was estimated to be less than ≈8 mm (corresponding to less than ≈ 1/338 of the wall height), while the deflection rate at the end of the test was calculated to be approximately ≈ 4.5 mm/min.
Fig. 7
Measured a deflection and b deflection rate of the specimen
3.5 Criteria, Observed Events and Failure Mechanism
The resistance to fire is the capacity of a structure or an individual compartment (as in this case—for a wall) to resist a determined amount of time, in terms of stability (R), integrity (E) and isolation (I) capability. Experimental tests and observations are therefore of utmost importance for any element that fulfil the criteria. Classification is in fact carried out by checking the time value determined for the estimation of mechanical resistance under fire conditions with the reference nominal one.
According to Fig. 8, it can be thus provisionally concluded that all criteria are met in at least 10 min for the investigated full-size composite wall.
In addition, however, the observation and analysis of experimental results should also draw attention to some key events or critical issues. To this end, Table 2 summarises some important observations that emerged during the test execution. As it can be seen, the occurrence of first cracks in the glass components is quite premature and starts after > 2 min of fire exposure on the outer glass layer of the exposed panel, where the imposed temperatures are highest. To note that at this stage the timber-glass wall still provides some residual load-bearing capacity and integrity before failre, which manifests in a larger fall-out of the exposed LG panel (after about ≈ 7 min of fire exposure) and in a progressive development of the failure mechanisms until the end of the experiment (after > 11 min).
Table 2
Experimental observations
Time (min:s)
Specimen side
Observation
0:00
–
Start
2:21
E
Cracks in the outer layer of glass
6:53
E
Fall-out of the exposed laminated glass panel
7:49
E
Cracks in the second/ inner glass layer
9:40
U
Cracks in the exposed layer of glass
10:55
U
Fracture of laminated glass – unexposed side + Fire break at the top centre of the wall
11:05
–
End of test
Key: E exposed side, U unexposed side
Figure 9a shows the specimen when the second/inner glass layer of the exposed LG panel cracks, while Fig. 9b was taken at the end of the experiment.
Fig. 9
Experimental observations: a failure of the first glass panel (8th minute) and, b view of the specimen at the end of experiment (11th minute)
As in Fig. 9, it is interesting to note that no damage was observed on the exterior side of the of the timber joints. On the other hand, Fig. 10 shows charring effects for the timber frame on both the exposed (Fig. 10a) and unexposed (Fig. 10b) sides of the wall. In this case, the reason for charring of timber on the non-exposed side was found in the failure of the glass panel, which occurred in the final phase of the text only.
Fig. 10
Timber frame at the end of experiment: a exposed (bottom corner) and b unexposed (top corner) sides
The numerical analysis was performed in ABAQUS v.6.14 to further investigate the experimental evidence. Overall, the construction and characterization of the full 3D FE model followed and extended the assumptions from the literature, which have been experimentally validated for various monolithic and laminated members in fire conditions, see for example [28, 30].
In the present study, the thermal simulation consisted in detail of a single uncoupled, transient “heat transfer” step of ABAQUS/Standard, in which the nominal geometry of the timber-glass wall specimen was subjected to the average experimental temperature–time curve measured in Fig. 5a. In a preliminary stage, it was assumed that this temperature–time curve is uniformly distributed on the exposed glass and timber surfaces.
To this end, the glass panes, the intervening PVB interlayer, the air infill and the timber frame components were initially modelled in ABAQUS/Standard with their nominal geometry by using DC3D8 brick elements (Fig. 11). For simplicity, the glass layers are labelled E1, E2, U1 and U2 in the following, as indicated in Fig. 11a. The steel parts at the top and bottom edges of the frame, which are used to attach the prototype to the furnace structure, have not been included in the model. In addition, the ceramic wool and other secondary details of the joint were not considered in the modelling, in place of ideal boundary conditions. This means that the nominal geometry of the frame members, glass and PVB was considered. To minimize the computational cost of the numerical studies, finally, 1/4 of the nominal specimen was described in ABAQUS to investigate the heat distribution (Fig. 11b). Preliminary trials and sensitivity studies were focused especially on the mesh pattern and size, considering their effects on local and global estimates. The selected solution resulted in 8-node mesh scheme for all the model components, with an average edge size of 2 cm (Fig. 11c). Three mesh elements were used in the thickness of each glass layer and PVB bond, and four elements were used in the thickness of air cavity.
Fig. 11
Numerical analysis: a schematic drawing of model assembly (cross-section) and b general view of 1/4th of specimen, with c mesh details (ABAQ US)
A regular 8-node mesh pattern was used for the individual components of the 3D assembly by utilising the local subdivisions in the solid members. The final result consisted of ≈ 17,400 bricks (≈ 23,000 DOFs).
It should be noted that a rigid, temperature-independent connection was used at the interface between the glass layers and the PVB film in between. Similarly, a rigid contact was considered at the interface of all FE components, as between glass and the adjacent timber elements. Symmetry considerations were taken into account for the thermal boundaries of the 1/4th model. Finally, an ideally insulated condition was assumed at the bottom and side surfaces.
Finally, the experimental setup for thermal exposure was described numerically in the form of a series of radiation and convection interactions (“surface radiation” and “surface film condition” from the ABAQUS library), which were assigned to all model surfaces exposed or not exposed to fire, see for example [30].
4.2 Materials and Air Cavity
Glass, PVB and timber materials were numerically described by thermo-physical properties that are strongly dependent on temperature, given that this sensitivity represents a crucial aspect for fire resistance considerations. Many literature studies (see for example [26, 28, 30, 31, 33]) proved that the temperature variation of material properties is of utmost importance for the numerical investigation of similar components and systems. However, in most cases, the major challenge for numerical investigations is represented by the lack of detailed material properties that specifically correspond to the basic components of the examined full-scale experimental prototypes [26]. As such, literature applications like [28, 30, 31, 33], where the same numerical strategy has been implemented for different geometrical and loading conditions, represent a useful support for new investigations. For this reason, the present numerical analyses were carried out based on earlier extended investigations for laminated glass in fire, considering a similar modelling strategy and in particular the same temperature-dependent properties for basic materials.
For annealed glass, the material characterization was in fact carried out in accordance with past successful experiences of thermo-mechanical numerical analyses, see [28] and Fig. 12. Glass emissivity was set to 0.95 [30]. At the same time, the PVB layers were described in the form of a thermal-sensitive elastic material, with temperature-dependent input features according to [30]. For structural mechanics considerations, this means that a rather small modulus of elasticity was taken into account at room temperature (≈ 0.00012 times the modulus of glass), and further reduced as a function of thermal exposure, for temperatures higher than ≈ 50 °C [30].
Fig. 12
Selected mechanical and thermo-physical material properties for ordinary float glass, with evidence of their sensitivity to temperature variations (literature experimental data in use for the FE numerical simulations, based on [28]: a modulus of elasticity; b density; c specific heat; d thermal conductivity; e linear thermal expansion coefficient
The thermo-physical and mechanical characterization of timber was carried out according to Fig. 13, where the modification of most important features for mechanical analysis in fire conditions, compared to the ambient temperature of 20 °C, is proposed in terms of modulus of elasticity, density, specific heat and conductivity. Due to lack of more specific material properties, the evolution with temperature was defined according to the relationships suggested by EN 1995-1-2 [62], and widely used in practice for numerical modelling purposes.
Fig. 13
Selected mechanical and thermo-physical material properties for timber, with evidence of their sensitivity to temperature variations: a modulus of elasticity and density; b thermal conductivity and specific heat
In general terms, it is to note that all the fire exposed surfaces, the convective heat transfer coefficient was conventionally set to 25 W/m2K [30, 61] and assumed in 8 W/m2K for the unexposed surfaces [30, 61].
Due to lack of more specific material properties, as suggested by EN 1991-1-5 [63], emissivity for timber was set to 0.8, while conductivity and specific heat at ambient temperature were set equal to 0.12 W/mK and 1.53 kJ/kgK respectively [63].
Specific attention was given also to the IGU cavity description, even under basic assumptions. For the air infill interposed to the laminated glass panels, more in detail, the numerical description of input properties was based on the specification of thermal conductivity and specific heat, which have in general a key role for transient thermal analyses. Besides, it is also important to remind that the cavity itself represents in general a challenging model component, and for the present numerical study it was accounted in a simplified approach only. Assumed that an equivalent material was used to describe the air infill, the input features from [60] were in fact taken into account for its characterisation, and their variation with temperature is proposed in Fig. 14a, b. For such a fictious air material, the density was set in 1.20 kg/m3 at room temperature (1 atm pressure), and modified according to Fig. 14c.
Fig. 14
Selected thermo-physical material properties for the simplified numerical characterisation of air in the IGU cavity, with evidence of their sensitivity to temperature variations: a thermal conductivity, b specific heat and c density
In parallel to such a kind of material calibration, the corresponding mesh elements were characterised accordingly, assuming an ideal mechanical bond with the surrounding glass and timber elements of the model. To note that this last aspect represents a further simplification of the modelling strategy, compared to reality. It is in fact worth to remind that the presently examined composite timber-glass wall prototype was not assembled with a perfectly sealed air cavity, and thus the assumption of an ideal insulation for the air infill is.
For the same reason, radiation was considered towards ambient, for the exposed and unexposed surfaces only. In this regard, a robust model approach for future applications should possibly account for gaps in the cavity seal and possible thermal interactions of the air infill itself with ambient. At the same time, however, experimental registrations for the cavity would be required in support of model detailing and refined calibration.
Finally, the last simplified assumption for numerical modelling purposes regarded the transmission effect of glass. This effect is known to represent a significant factor in some contexts, particularly for thinner panes. In the present study, however, it was chosen to neglect the transmission effect for several reasons. First of all, the glass panes used for the full-scale wall specimen were set in 10 mm of thickness. For such thickness value, the percentage of transmission is in general relatively low, compared to thinner glass, thus its influence on the overall system behaviour was deemed minor. Furthermore, most importantly, the assembled numerical model aimed to prioritize the primary factors that can significantly influence the performance parameters under consideration, such as structural integrity, thermal response, and load-bearing capacities. Including the transmission effect would certainly increase model complexity, with rather limited additional insight for the specific objectives of the work reported herein. By neglecting the transmission effects, it is also generally recognized that there may be a minor underestimation of heat or light transfer, which could influence the numerical results in scenarios where such effects are non-negligible. However, given the relatively low transmission for 10 mm thick glass panes and the focus of this study, such a simplification was considered acceptable and well-balanced to the scope of the investigation.
4.3 Numerical Results
The analysis of numerical results was first focused on the temperature assessment in the glass layers, given that they represent the most critical component in similar composite systems, but with a specific attention also for the timber frame.
4.3.1 Timber Frame and Charring Rate
The thermal analysis of timber frame components was carried out in terms of temperature evolution and corresponding charring rate, which can be efficiently monitored towards the reference limit value of 300 °C.
To this aim, Fig. 15 reports the temperature evolution in a timber cross-section, where the legend values are fixed to 300 °C to emphasise the charred region over time. In this sense, the grey model region is associated to temperatures that exceed 300 °C and denote full charred timber.
Fig. 15
Numerical results (ABAQUS): local detail of temperature evolution in the exposed timber components, as a function of time (vertical cross-section of the transom, with IGU and mesh hidden). Legend values in °C
Basically, and in accordance with the experimental evidence, the resisting timber section in use proved to offer a relatively stable performance in time, as it can be seen from the temperature peaks and their distribution in time of Fig. 15. The numerically recorded temperatures in timber do not exceed the limit of 300 °C for the first 3 min of fire exposure. Successively, the resisting cross-section of the frame is progressively affected by fire exposure, and especially after 7–8 min, there is a clear propagation of the char front and charred timber region in the exposed side. Worth to note, in any case, is that fire exposure still affects a limited portion of the resisting cross-section, and this finding is again in accordance with the experimental outcomes (see for example Fig. 10). The blue elements in the contour plots of Fig. 15 are in fact associated to timber that is still at ambient temperature.
4.3.2 Temperature in the Exposed Glass Panel
A more detailed analysis is required for the thermal analysis of the IGU components, and first of all the exposed laminated glass panel (E1 and E2 glass layers, according to Fig. 11a).
Fig. 16
Numerical results (ABAQUS): local detail of temperature evolution in the exposed laminated glass panel (E1 and E2 glass layers), as a function of time (vertical cross-section, with timber frame, PVB foil and mesh hidden). Legend values in °C
Due to the relatively small thickness, and relevant exposed surface, the glass components are in fact subjected to major temperature gradients that could have major effects for integrity and load-bearing performance considerations.
Figure 16, in this regard, shows the rapid temperature increase and distribution in glass, as a function of time, for the exposed laminated glass panel (layers E1 and E2). For graphical convenience, and to better emphasize the non-uniform temperature evolution in the region of timber frame and in the thickness of glass layers, the maximum legend value is fixed in 120 °C.
It can be noted that the region of glass which is contact with the timber frame, can take advantage of a minimum protection, and hence the temperature at the edges of glass is relatively low, compared to the exposed surface. Besides, in less than ≈ 3 min of exposure, the thermal gradient in the E1 and E2 glass layers is relatively high and involves temperature peaks above ≈ 80 °C also for the second E2 glass layer.
The local front view of Fig. 17, for the E1 glass layer that is directly exposed to fire, is even more strategic for the analysis of resisting and failure mechanisms of the examined composite system.
The localized analysis of thermal peaks, but most importantly the analysis of maximum thermal gradients that affect the glass component in fire conditions, is in fact a first important verification step.
In this regard, considering that the corner and the centre of E1 glass are the coolest and hottest regions respectively for the timber-glass wall prototype in fire conditions, a more quantitative analysis of the imposed fire scenario was carried out towards the earlier reported experimental evidences.
To note that the attention was primarily focused on the 3 initial minutes of exposure, given that first cracks in glass were experimentally observed in the E1 layer after 2:21 min (Table 2).
To this aim, Fig. 18 reveals some temperature trends that can justify the experimental observations. The collected charts show in fact the local temperature estimation in the centre and in the corner respectively of the reference FE models. The corresponding thermal gradient ΔT, which should not exceed some critical conventional values (Table 2), is also highlighted in Fig. 18. It is thus interesting to correlate the numerical evidence with the experimental events summarised in Table 2.
Fig. 17
Numerical results (ABAQUS): local detail of temperature evolution in the exposed laminated glass panel (E1 layers), as a function of time (front view, with timber frame and mesh hidden). Legend values in °C
As it can be seen from Fig. 18a, the temperature rise at both the centre and corner control points is rather fast but at least for 6–8 min of exposure does not exceeds the reference transition temperature for glass (i.e., 400–500 °C, according to Sect. 2). Otherwise, in the same time interval, there is evidence of experimental events that denote the damage propagation in the exposed glass panel.
Fig. 18
Numerical results (ABAQUS): temperature evolution in the exposed laminated glass panel (E1 layers), as a function of time: a corner and centre temperature; b maximum thermal gradient
As such, it can be useful to examine more in detail the measured thermal gradient of Fig. 18b, which interestingly reveals a maximum gradient in the order of ≈ 40 °C after ≈ 2 min of fire exposure, and is in close correlation with the first observed cracks.
While a similar thermal analysis can offer some useful feedback in support of the interpretation of experimental outcomes, the same strategy has some major intrinsic limits in the analysis of similar composite systems, due to the fact that the progressive failure of the exposed glass panel cannot be directly taken into account. Any further consideration related to the thermal performance or even thermo-mechanical and load-bearing performance of the examined system, necessarily requires a different approach.
4.4 Uncoupled Mechanical Numerical Analysis in Fire
A mechanical analysis was also carried out on the examined system, in order to verify and assess the expected load-bearing performance and residual capacity [28, 30]. In this case, a general static step was used by assuming the temperature distribution and evolution in time, as obtained from the previous “heat transfer” simulation, as a key input. Accordingly, the element type for brick elements was changed to C3D8R in ABAQUS library.
The experimental boundary and loading conditions were thus numerically reproduced. As such, both gravity loads and the superimposed distributed load of 25 kN/m were kept fix during the whole analysis, while the input temperature was imposed at each node, for the 12 min of simulated time of fire exposure.
To note, compared to the transient thermal step, that the prototype was investigated in its full elevation, still under single symmetry conditions, in order to capture the in-plane and out-of-plane mechanical response. As a major effect of the presently adopted modelling strategy, the input temperature in the time of analysis was equally applied to the “top” and “bottom” quarter of the mechanical model, and possible non-uniform thermal distribution in the elevation of the prototype was disregarded.
Another implicit limit is associated to the type of failure mechanisms captured by the numerical model, considering that the characterization of materials was based on temperature-sensitive properties. This means that possible relaxation of the modulus of elasticity was properly taken into account through the mechanical analysis in fire conditions. Besides, the possible detachment of glass fragments (i.e., to account for fallout of the exposed glass layer) was not included.
Figure 19a compares the vertical deflection of the system and the past experimental measurements. Given the complexity of the examined configuration, as well as the combination of multiple phenomena, it can be seen that the FE model is still able to capture the global response of the composite wall.
Fig. 19
Numerical results (ABAQUS): a vertical deflection of the composite wall in fire conditions, as a function of time and b corresponding out-of-plane deflection
Accordingly, Fig. 19b presents the evolution of maximum out-of-plane displacement of glass, as a function of time. The progressive bending is a key parameter that depends on the progressive decrease of Young’s modulus due to high temperature, and manifests in a progressively increase of deflections. It is thus worth to note that after ≈ 4 min of fire exposure, the maximum bending of the glass wall is still relatively small, and equal to about ≈ 4.5 mm, corresponding to ≈ 1/600 the wall height. The progressive increase of out-of-plane displacements (up to ≈ 11 mm ≈ 1/245 the wall height) are estimated in around ≈ 7–8 min of fire exposure, that approximately coincide with the experimental fall-out of the exposed glass panel.
Also in this case, it can be thus observed a certain correlation of experimental and numerical evidences.
A very important performance indicator for vertical load-bearing components and systems is in fact represented by the out-of-plane deflection in fire. While the EN-1363-1 reminds the limits for structural systems in fire “mainly loaded in-plane”, as it is for a vertical composite wall, the out-of-plane deflections and deflection rates become relevant due to non-linear effects that take place in the overall resisting mechanisms of the timber-glass wall in fire.
As such, the EN-1363-1 limit values for structural systems in fire and “mainly loaded in bending” are:
For the presently examined timber-glass wall, it can be thus noted from Fig. 19 and other experimental evidences that the wall sustained the superimposed mechanical loads for about 11 min (with major damage in its component). Besides, as also emphasised by the numerical analysis, the predicted in-plane and out-of-plane deflections—while still relatively important in amplitude—did not match with the current referent values from standards [i.e., Eqs. (1)–(4)], based on whom the examined wall should not suffer for any failure mechanism in fire.
In conclusion, it is clear that the explored composite system can offer major structural benefits and capacities in buildings, even in fire conditions. At the same time, however, it is still uncertain and sometimes difficult to assess their fire resistance and endurance capacities, based on conventional parameters from standards that should be properly adapted and extended to structural configurations and material properties that fit with structural glass applications. In this sense, future research will be carried out to further optimise and support the promotion of the composite timber-glass wall concept.
5 Conclusions
The paper investigated the behaviour of a shear wall in form of the hybrid system composed of load-bearing glass panels infilled into a timber frame. In addition to the fire load, the panel was subjected to the simultaneous action of a vertical distributed mechanical load, in order to possibly simulate a real loading condition in buildings. The experimental analysis reported a fire performance characterized by premature and progressive damage propagation in the glass components (starting after about 2 min of exposure), but an overall capacity which resulted in the conclusion of the experiment after 10 min of exposure. Moreover, a pilot numerical study was also carried out in support of the experimental observations, to investigate further some local and global phenomena. The presented findings, as shown, proved the great potential of the examined timber-glass composite panels, as well as fact that this concept can represent an efficient alternative to other solutions of load-bearing structural components in buildings.
The research investigation, in particular, has shown that the timber frame has sufficient bearing capacity for vertical load, e and due to the failure of the glass panel at high temperatures. However, the bearing capacity for lateral loads is significantly lower because the form of stabilization is lost when the glass fails. Due to the significant effect of wood in insulating against high temperatures, the detail of the connection between the column and the beam of the timber frame is safe until the moment of failure of the glass itself, so it does not represent a problem in determining the load capacity of the panel. In conclusion, the main problem is to achieve better behaviour of the glass panel in fire.
Although a proof of concept for the application of this type of shear wall has been demonstrated, however, its widespread use is not yet feasible. Serviceability Limit State (SLS) and Ultimate Limit State (ULS) are fundamental design concepts for structural engineering. SLS ensures that the structure remains functional and comfortable for users under normal conditions, while ULS ensures the safety and integrity of the structure under maximum unfavourable loads. Since there is a lack of a clear definition of SLS and ULS in fire conditions, and even more for post-fire situations, it is difficult to define possible performance indicators that could be used to assess the actual capacity in fire conditions of the investigated timber-glass structural panel concept. The use of glass as a structural, load-bearing component for full-scale walls is currently very limited, and future studies are required, both in experimental and numerical terms. Investing financial resources in research, to solve these problems and go beyond the present knowledge would open possibilities for even enhanced constructive and architectural concepts of timber-glass solutions.
Among the possible alternatives that should be taken into account to maximize the fire response of the explored prototype, for example, it has to be mentioned the possible use of tempered glass, as well as the application of additional safety layers, i.e. fireproof coatings, to further reduce the impact of high temperatures on glass components, as much as possible, and preserve further its load-bearing capacity.
Acknowledgements
This research was funded by the Croatian Science Foundation (Project no. IP-2016-06-3811 VE-TROLIGNUM - “Prototype of multipurpose composite timber-load bearing glass panel”, coordinator Prof. Vlatka Rajcic, University of Zagreb, Croatia). Open Access publication fees are covered the University of Trieste (CRUI-CARE transformative agreement).
Declarations
Conflicts of Interest
The authors declare no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Kozlowski M (2012) Hybrid glass beams. Review of research projects and applications, architecture. Arch Civ Eng Environ 3:53–62
2.
Rajčić V, Žarnić R (2012) Racking performance of wood-framed glass panels. In: Quenneville P (ed) WCTE 2012—world conference on timber engineering, Auckland, New Zealand, conference proceedings
3.
Rajčić V, Žarnić R (2012) Seismic response of timber frames with laminated glass infill. In: Proceedings of CIB-W18 meeting 45, CIB-W18/45-15-4, Växjo, Sweden, pp 11–19
4.
Ber B, Premrov M, Strukelj A, Kuhta M (2014) Experimental investigations of timber-glass composite wall panels. Constr Build Mater 66:235–246CrossRef
5.
Martens K, Caspeele R, Belis J (2015) Development of composite glass beams—a review. Eng Struct 101:1–15CrossRef
6.
Hamm J (2001) Development of timber-glass prefabricated structural elements. In: Lahti, Finland. IABSE Symposium Report 85(10): 41–46. https://doi.org/10.2749/222137801796348340
7.
Hamm, J (1999) Tragverhalten von holz und holzwerkstoffen im statischen verbund mit glas, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland. Doctoral thesis, 161 pages, https://doi.org/10.5075/epfl-thesis-2065
Blyberg L, Serrano E (2011) Timber/glass adhesively bonded I-beams. In: Glass performance days conference, conference proceedings, Tampere, Finland
10.
Blyberg L, Lang M, Lundstedt K, Schander M, Serrano E, Silfverhielm M, Stalhandske C (2014) Glass, timber and adhesive joints—innovative load bearing building components. Constr Build Mater 55:470–478CrossRef
11.
Kozlowski M, Serrano E, Enquist B (2014) Experimental investigation on timber-glass composite I-beams. In: Louter C, Bos F, Belis J, Lebet J-P (eds) Challenging glass 4 & COST action TU0905 final conference, Lausanne, Switzerland, conference proceedings
12.
Kozlowski M, Kadela M, Hulimka J (2016) Numerical investigation of structural behaviour of timber-glass composite beams. Procedia Eng 161:990–1000CrossRef
13.
Premrov M, Zlatinek M, Strukelj A (2014) Experimental analysis of load-bearing timber-glass I-beam. Constr Unique Build Struct 4:11–20
14.
Furtak K, Rodacki K (2018) Experimental investigations of load-bearing capacity of composite timber-glass I-beams. Arch Civ Mech Eng 18:956–964CrossRef
15.
Dorn M, Kozlowski M, Serrano E, Enquist B (2014) Development of large-scale load-bearing timber-glass structural elements. In: Salenikovich A (ed) WCTE 2014—world conference on timber engineering, Quebec City, Canada, Proceedings
16.
Kozlowski M, Dorn M, Serrano E (2015) Experimental testing of load-bearing timber-glass composite shear walls and beams. Wood Mater Sci Eng 10:276–286CrossRef
17.
Štrukelj A, Ber B, Premrov M (2015) Racking resistance of timber-glass wall elements using different types of adhesives. Constr Build Mater 93:130–143CrossRef
18.
Rajcic V, Perkovic N, Bedon C, Barbalic J, Žarnić R (2020) Thermal and energy-efficiency assessment of hybrid CLT–glass façade elements. Appl Sci 10:1–22CrossRef
19.
Rajcic V, Perkovic N, Damjanovic D, Barbalic J (2022) Influence of friction on the behavior and performance of prefabricated timber–bearing glass composite systems. Sustainability 14:1–23CrossRef
20.
Rajcic V, Žarnić R, Perković N, Barbalić J (2019) Optimisation test study of connections and joints between elements of the hybrid timber—structural glass panel. In: Miko (ed) World construction forum 2019: buildings and infrastructure resilience—book of abstracts with programme, Ljubljana, Slovenia
21.
Natterer J, Kreher K, Natterer J (2002)New joining techniques for modern architecture. Rosenheimer, Fenstertage
22.
Kreher K (2004) Tragverhalten und bemessung von holz-glas-verbundträgern unter berücksichtigung der eigenspannungen im glas, ecole polytechnique federale de Lausanne, Lausanne, Switzerland
23.
Vallee T, Cordula G, Milchert L, Fecht S (2016) Design and dimensioning of a complex timber-glass hybrid structure—the IFAM pedestrian bridge. Glass Struct Eng 1:3–18CrossRef
24.
Fadai A, Winter W (2014) Application of timber-glass composite (TGC) structures for building construction. In: Louter C, Bos F, Belis J, Lebet J-P (eds) Challenging glass 4 & COST action TU0905 final conference, Lausanne, Switzerland, conference proceedings
25.
Jiang Y, Schaffrath J, Knorz M, Winter S (2014) Bonding of various wood species—studies about their applicability in glued laminated timber. In: Aicher S, Reinhardt H-W, Garrecht H (eds) RILEM conference—materials and joints in timber structures - recent advancement of technology. Stuttgart, Germany, Conference Proceedings
26.
Bedon C (2017) Structural glass systems under fire: overview of design issues—experimental research and developments. Adv Civ Eng 2120570:1–18
27.
Bedon C, Zhang X, Santos F, Honfi FK, Arrigoni M, Figuli M, L., Lange, D. (2018) Performance of structural glass facades under extreme loads—design methods, existing research, current issues and trends. Constr Build Mater 163:921–937CrossRef
28.
Bedon C, Louter C (2023) Thermo-mechanical numerical analyses in support of fire endurance assessment of ordinary soda-lime structural glass elements. J Struct Fire Eng 14:522–546CrossRef
29.
Galuppi L, Franco A, Bedon C (2023) architectural glass under climatic actions and fire: review of state of the art—open problems and future perspectives. Buildings 13:939CrossRef
30.
Louter C, Bedon C, Kozlowski M, Nussbaumer A (2021) Structural response of fire-exposed laminated glass beams under sustained loads—exploratory experiments and FE-simulations. Fire Saf J 123:1–18CrossRef
31.
Vedrtnam A, Bedon C, Youssef MA, Chaturvedi S (2021) Effect of non-uniform temperature exposure on the out-of-plane bending performance of ordinary laminated glass panels. Compos Struct 275:114517CrossRef
32.
Oikonomopoulou F, Bristogianni T, Barou L, Veer FA, Nijsse R (2018) The potential of cast glass in structural applications. Lessons learned from large-scale castings and state-of-the art load-bearing cast glass in architecture. J Build Eng 20:213–234CrossRef
33.
Kozlowski M, Bedon C (2021) Sensitivity to input parameters of failure detection methods for out-of-plane loaded glass panels in fire. Fire 4:5CrossRef
34.
Rouxel T (2007) Elastic properties and short-to medium-range order in glasses. J Am Ceram Soc 90:3019–3039CrossRef
35.
Pelayo F, Lamela-Rey MJ, Muniz-Calvente M, López-Aenlle M, Álvarez-Vázquez A, Fernández-Canteli A (2017) Study of the time-temperature-dependent behaviour of PVB: application to laminated glass elements. Thin-Walled Structures 119:324–331CrossRef
36.
Vandebroek M (2014) Thermal fracture of glass. Ghent University, Ghent
37.
Sabsabi A, Youssef MA, El-Fitiany SF, Vedrtnam A (2024) Simplified structural analysis of laminated glass panels during fire exposure. Fire Saf J 146:1–7CrossRef
38.
Corrado A, Magisano D, Leonetti LGG (2024) Sensitivity to intensity and distribution of the temperature field in the nonlinear thermo-mechanical analysis of laminated glass plates. Int J Non Linear Mech 165:1–15CrossRef
39.
de Sa JC (1986) Numerical modelling of glass forming processes. Eng Comput (Swansea) 3:266–275CrossRef
40.
Fluegel A, Earl DA, Varshneya AK, Seward TP (2008) Density and thermal expansion calculation of silicate glass melts from 1000 to 1400°C. Phys Chem Glasses Eur J Glass Sci Technol B 49:245–257
41.
Le Bourhis E, Rouxel T (2003) Indentation response of glass with temperature. J Non Cryst Solids 316:153–159CrossRef
42.
Zarzycki J (1982) Glasses and the vitreous state. Cambridge University Press, Cambridge
43.
Badger AE, Silverman WB (1935) Modulus of elasticity of glass in relation to temperature. J Am Ceram Soc 18:276–280CrossRef
44.
Shen J, Green DJ, Tressler RE, Shelleman DL (2003) Stress relaxation of a soda lime silicate glass below the glass transition temperature. J Non Cryst Solids 324:277–288CrossRef
45.
Scholze H (1991) Glass. Nature, structure, and properties. Springer, New York
46.
Wang Y, Wang Q, Shao G, Chen H, Sun J, He L, Liew K (2014) Experimental study on critical breaking stress of float glass under elevated temperature. Mater Des 60:41–49CrossRef
47.
Rouxel T, Sanglebuf J-C (2000) The brittle to ductile transition in a sodalimesilica glass. J Non Cryst Solids 271:224–235CrossRef
48.
Xie Q, Zhang H, Si D (2011) experimental study on critical breakage stress of float glass with different thicknesses under conditions with temperatures of 25 and 200 °C. Fire Mater 35:275–283CrossRef
49.
Aronen A (2012) Modelling of deformations and stresses in glass tempering. Tempere University of Technology, Tampere
50.
Nielsen JH (2009) Tempered glass—bolted connections and related problems. Technical University of Denmark, Lyngby
51.
Narayanaswamy OS (1978) Stress and structural relaxation in tempering glass. J Am Ceram Soc 61:146–152CrossRef
52.
Carré H, Daudeville L (1999) Load-bearing capacity of tempered structural glass. J Eng Mech 125:914–921
53.
Daudeville L, Bernard F, Gy R (2002) Residual stresses near holes in tempered glass plates. Mater Sci Forum 404–407:43–48CrossRef
54.
Buchanan AH (2001) Structural design for fire safety. Wiley, Chichester
55.
CEN (2016) EN 13501-2:2016—fire classification of construction products and building elements—part 2: classification using data from fire resistance tests, excluding ventilation services. European Commitiee for Standardization, Brussels
56.
Kozłowski M, Lenk P, Dorn M, Honfi D, Sjöström J (2018) Structural considerations on timber-glass composites at fire scenarios. Chall Glass Confer Proc 6:229–240
57.
Karlsson V, Wärnelöv M, Dorn M, Östman B (2017) Brandpåverkan På lastbärande trä-glasväggar. Bygg & Teknik 109(109):44–47
58.
CEN (2020) EN 1363-1:2020—fire resistance tests—part 1: general requirements. European Committee for Standardization, Brussels
Bergman TL, Incropera FP (2011) Fundamentals of heat and mass transfer (Appendix A). Wiley, Hoboken
61.
ISO 10077-2:2012 (2017) Thermal performance of windows, doors and shutters—calculation of thermal transmittance—part 2: numerical method for frames. International Organization for Standardization, Switzerland
62.
CEN (2005) EN 1995-1-2. Eurocode 1—design of timber structures—part 1–2: general—structural fire design. European Committee for Standardization, Brussels
63.
CEN (2004) EN 1991-1. Eurocode 1—actions on structures—part 1–5: general actions—thermal actions. European Committee for Standardization, Brussels
