ABSTRACT. Timber plays a key role in reducing upfront carbon emissions in new buildings. A limitation of traditional structural fire engineering analyses is the lack of consideration for post-fire thermal wave propagation, which can continue to degrade strength and stiffness after fire extinguishment of a timber column. This residual degradation may lead to a sudden brittle collapse, posing risks to occupant egress, firefighter operations and adjacent structures. This paper presents results for the compressive parallel-to-grain strength and stiffness of New Zealand Radiata Pine structural grade SG8 and LVL8 under both heating and cooling, to assess the potential for recovery after cooling. Approximately 200 small-scale tests were conducted at the University of Canterbury, Christchurch, New Zealand. The results indicate that the reduction in strength and stiffness during heating is less severe than predicted by Eurocode 5, and the timber can recover its compressive mechanical properties on cooling, following exposure to temperatures below 180°C. The experimental results will be used to improve the design of timber columns exposed to realistic fires, using numerical methods.

ABSTRACT. This study presents a calibrated modelling framework for predicting ductile fracture in cold-formed steel (CFS) under elevated temperatures. Experimental tensile data from S280GD+Z steel coupons tested from 20 °C to 700 °C are used to establish accurate true stress–strain relationships, including the post-necking regime reconstructed using a weighted-average methodology. The Stress-Modified Critical Strain (SMCS) model and a ductile damage model with temperature-dependent evolution law are implemented in finite element simulations to capture fracture initiation and propagation. The calibrated approach is subsequently validated against additional CFS grades spanning yield strengths up to 1200 MPa, demonstrating strong agreement with experimental fracture behaviour. The results provide a reliable methodology for modelling fracture of thin-gauge CFS in fire scenarios, with potential for extension to hot-rolled steels.

ABSTRACT. This study develops probabilistic high-temperature material models for the elastic modulus and 0.2% proof stress of cold-formed steel (CFS) to address the limitations of current deterministic fire design provisions. A comprehensive database of nearly 400 experimental data points was compiled, covering a wide range of temperatures and steel grades. Ten candidate probability distributions were evaluated, and lognormal and continuous logistic formulations were identified as the most suitable models. Temperature-dependent distribution parameters were obtained through regression, enabling prediction of quantile curves that closely match experimental trends and encompass observed scatter. The proposed probabilistic models were incorporated into Monte Carlo analyses of eight CFS end-wall columns from a prototype metal building. The results revealed significant dispersion in failure temperatures, with 5%–95% ranges of approximately 100 °C. Despite their different mathematical forms, the lognormal and logistic models produced similar median failure temperatures and variability, demonstrating the robustness of the framework. The developed models provide a foundation for reliability-based performance-based fire design and support the derivation of fragility curves for fire risk assessment of CFS structures.

ABSTRACT. In recent years, the probability of post-earthquake fires in urban areas has risen with the widespread development of pipeline networks. For steel structures prevalent in seismic regions, the inherently poor fire resistance of steel, coupled with design codes that treat seismic and fire resistance separately, poses critical challenges to structural safety. Nonetheless, limited research has explored the fire performance of steel or novel materials under seismic damage conditions. To address this gap and provide a basis for material optimization, this study examines the fire resistance of cyclically-damaged materials, including iron-based shape memory alloy (Fe-SMA), conventional structural steel (Q235), and high-strength steel (Q550), through transient-state fire tests on 90 specimens. The findings demonstrate that Fe-SMA achieves recoverable strains of 0.11%-0.18% at 150-350℃, enhanced ductility, and significantly higher ultimate temperatures than structural steels. Conversely, cyclic damage deteriorates the fire resistance of Q235 and Q550, lowering their ultimate temperatures and fracture strains. The superior performance of Fe-SMA under combined high stress and temperature highlights its promise for post-earthquake fire-resistant design, advancing safety and resilience in seismic-prone urban regions.

ABSTRACT. This study introduces an AI enhanced beam column element for efficient and accurate structural fire analysis. The element embeds a trained neural network within a traditional beam column formulation, enabling it to capture nonlinear and post buckling behaviour of fire affected members without relying on shell or solid elements. A high resolution dataset of H section steel columns under varied fire and loading conditions was generated using shell element simulations and used to train the AI model. The AI enhanced element was implemented in OpenSees and validated against fire tests and system level building simulations. Results show that it reproduces the fidelity of shell elements while reducing computation time by about 75%, demonstrating strong potential for scalable performance based fire design and hybrid simulation applications.

ABSTRACT. Hybrid fire simulation is an innovative testing method that combines physical and numerical substructuring to better understand the global behaviour of structures subjected to complex fire loading. This approach addresses the limitations of traditional fire testing methods, which can be either prohibitively expensive when testing entire structures or overly simplistic when testing individual components by integrating physical substructures, tested in fire laboratories, with numerically simulated substructures. Despite its potential, the application of hybrid fire simulation has been limited, with only a few tests controlling multiple degrees of freedom at the physical and numerical substructure interface and fewer still conducted in full-scale fire testing furnaces. The major challenge is often due to difficulties maintaining stable control due to the complex interactions between multiple degrees of freedom and nonlinear behaviour both thermal and mechanical. Building on earlier developments of a model-based adaptive error compensation method for hybrid fire simulation, this work introduces a temperature-informed extension that explicitly incorporates thermal effects into the adaptive control model. In the original formulation, the adaptive system identified the relationship between actuator strokes and specimen deformations based solely on the evolving mechanical response. The enhanced algorithm integrates measured temperature fields directly into the recursive least-squares adaptation, enabling real-time capture of stiffness degradation, thermal expansion, altered boundary conditions, and evolving multi-degree-of-freedom coupling as the specimen heats. The improved methodology was implemented within the UT-SIM hybrid simulation framework, allowing simultaneous updating of both mechanical and thermo-mechanical components of the control model. Verification studies on heated steel and reinforced concrete substructures demonstrate that the temperature-informed model substantially improves actuator stability, reduces displacement tracking errors, and enhances prediction accuracy during periods of rapid thermal-mechanical stiffness loss. Relative to the previous generation of the adaptive algorithm, the new approach shows reduced overshoot, faster convergence, and improved robustness to non-uniform heating and thermal gradients. These findings underscore the importance of incorporating temperature effects directly into adaptive error compensation for hybrid fire simulation. The proposed method represents a significant advance toward more reliable, efficient, and performance-based evaluation of structures in fire, especially for complex assemblies where thermo-mechanical behaviour governs global response.

ABSTRACT. [Introduction] The quantification of axial force development at beam-end shear connections is a critical parameter to ensure safe structural fire design [1]. It determines the structural stability via considering the demand against capacity at the connections in the beam thermal expansion/contraction induced loading in fire. However, most research on steel- concrete composite floor structures in fire is limited to its scale (i.e., component-based standard fire tests), due to high cost and difficulties in implementing large-scale experiments [2,3]. The limited scale in beam length would induce underestimated axial forces during the thermal expansion in the heating phase. Therefore, apart from the Cardington Fire Test series in the 1990s [4], NIST carried out a two-story steel concrete composite-floor (CF) fire tests test in 2019-2021 [5]. The prototype structure had a 3 × 2 bay layout (18m × 11m in plan, 7.2m in height), with a large-compartment fire applied on the ground floor at the south-central bay (9.1m × 6.1m in plan, 3.8m in height), see Fig. 1(a). The structural steel was protected with a 2-hour fire resistance. A total mechanical load of 5.2 kPa was applied, including self-weight. The compartment was heated by natural gas burners following an ASTM E119 fire curve and with a natural cooling phase [6]. Further details about these NIST CF tests can be found in Ref. [7].

[Methodology and results] This paper aims to quantify the axial force development at beam-end shear connections of a steel-concrete composite floor structure during a large compartment fire. A finite element model (FEM) is developed using Vulcan [8], validated against the NIST CF tests [7], see Fig. 1(b), and followed by a parametric study. Apart from the validated deflections at critical locations of the composite slab (similar modelling efforts by Ma & Gernay using SAFIR [5]), see Fig. 1(c) and Fig. 2(a,b); this paper extends the validated parameters to the measured horizontal displacements of beam ends as illustrated in Fig. 1(d) and Fig. 2(c,d). Those extended validation parameters are critical to improve the model confidence in representing the expansion and contraction from the slab and steel beams during the heating phase and cooling phase, respectively, interacting with the restrained column stiffness. It enables the quantification of the estimated axial force at beam-end shear connections. As shown in Fig. 2, the deflection and horizontal displacements indicate overall close agreement, except for HD1 and HD7, located away from the fire bay, hence having limited influence on the subsequent analysis. In terms of the FEM setup, steel beams and columns are represented by line elements. Steel connections are simulated using zero-length spring elements as pinned, while the composite slab is modelled with layered shell elements based on Mindlin-Reissner plate theory [8] to capture through-thickness response and composite action. Beam section temperatures in the model are prescribed directly from experimental measurements. Eurocode 4 [9] lightweight concrete with a compressive strength of 67 MPa, fixed column bases, and a slab reinforcement ratio of 60 mm2/m are applied in the model. Figure 3 presents the estimated axial forces at the beam-end shear connections of NIST CF Test 1 through the FEM model. Compressive forces during the heating phase and subsequent tensile forces during the cooling phase are clearly captured in the model, due to the restrained beams in their thermal expansion and contraction, although the level of restraints is different depending upon the adjacent bay stiffness. For example, the north primary beam, tied into the non-fire adjacent bays, reaches around 650 kN in compression, whereas the less-restrained south primary beam peaks at around 300 kN since its position as an edge beam, as shown in Fig. 3. Furthermore, the north end of the east girder develops a maximum of 600 kN while the south end reaches around 200 kN and releases sooner to tensile force as presented in Fig. 3(c). This paper also presents a parametric study via the validated FEM model, including parameters: concrete type and strength, connection stiffness, column base rotational stiffness, and slab reinforcement ratios. This parametric study confirms that the axial forces are more sensitive than global deflections (e.g., VD5) to parameters. For example, for concrete strength, Fig. 4(a,b) indicate that reducing the concrete strength from 67 (measured after curing) to 28 MPa (design value) reduces the east-girder compressive force significantly, i.e., from 600 kN to 400 kN. This is consistent with early slab cracking and loss of in-plane stiffness in the lower-strength reinforcement concrete slab. Meanwhile, maximum mid-span deflections at VD5 for both strengths lie between 550 mm and 600 mm, compared with a measured peak of 580 mm, giving deflection deviations of about 3-5% from the test. It suggests that, in this configuration, concrete strength mainly influences axial restraint forces in the girder rather than the overall deflection. Moreover, per Fig. 4(c,d), a higher reinforcement ratio helps to reduce the load demand at the beam-end shear connections, via encouraging the development of tensile membrane action from the composite slab.

[Conclusions and future work] These results show that the axial force demands at connections are redistributed within the floor according to local stiffness, temperatures, and boundary restraint, and cannot be inferred directly from deflection alone (e.g., 650 kN at the north primary beam end and 300 kN at the south primary beam end). The parametric study demonstrates that the axial forces at the connections are affected (sensitivity from highest to lowest) by: connection stiffness, concrete strength, slab reinforcement ratio, column-base rotational stiffness, and concrete material type. In the full paper, comparisons with hand calculations are also included to further validate its applicability under alternative system configurations and to quantify the total axial forces (from slab and beams) that vary between parameter sets. For future work, the extension of the current established FEM model to the other two NIST CF tests, with different slab reinforcement ratios and fire protection layouts, will be discussed. The effect of connection ductility on axial force will also be investigated, and its impact on the redistribution of the internal forces during the heating-cooling cycle will be studied.

ABSTRACT. Composite steel–concrete floor systems are widely used in multi-story buildings but are vulnerable to fire because steel components degrade rapidly at elevated temperatures. Although prescriptive fire design remains the default approach, performance-based fire design can deliver higher performance and/or more efficient designs. However, broader adoption is limited by the cost and expertise required for nonlinear thermo-mechanical simulations. This study develops a fast, interpretable machine-learning surrogate to predict the fire response of composite floor systems and identify key design drivers. The surrogate is trained on experimentally validated thermo-mechanical simulations spanning a range of floor designs. A suite of regression models were evaluated to map geometric and design parameters to system response, and Shapley Additive Explanations (SHAP) analysis was performed to quantify feature importance and reveal governing trends. The best model accurately predicts both peak and residual displacements at the slab centre. Results indicate that fire load is the dominant driver of maximum and residual displacements, while slab geometry and ventilation primarily influence peak response.

ABSTRACT. This study presents a deep learning-based Heat Transfer Neural Operator for predicting temperature evolution in concrete structures under fire. Built on the Fourier Neural Operator framework, the model learns the nonlinear mapping between time-dependent Neumann boundary conditions and full 2D spatio-temporal temperature fields. The HTNO employs spectral convolution to capture global spatial–temporal dependencies and provides near-instantaneous inference once trained. Validation against Finite Difference Method (FDM) results shows excellent agreement, with a mean relative L2 error below 0.7 %. The trained model generates a complete 2D thermal field in less than 0.008 s, over 3800× faster than conventional numerical solvers. These findings highlight the potential of neural operators as efficient surrogate models for performance-based fire design, enabling rapid assessment of diverse fire scenarios in concrete structures.

ABSTRACT. This research work presents a framework for undertaking fire risk assessments of bridge structures (both existing and new) aimed to provide project stakeholders with detailed risk profiles appraising both the frequency and consequence of analysed fire hazards. It also helps with quantification of the likely risk reduction if different mitigation strategies are implemented. The output results serve to inform cost-benefit assessments, assessment as to whether risk to bridge users and bystanders is as low as reasonably practicable, and compliance with regulatory requirements. It can also support prioritisation of investments over time and across portfolios of multiple bridges. The intention is that this framework is used to inform international Clients considering major infrastructure projects in future. This framework is not limited to specific fire hazards, meaning it is equally applicable to assessing small scale fires hazards (e.g. passenger car fire), large scale fires with different escalation regimes (e.g. tanker road tanker fire), or complex fire scenarios involving multiple burning objects (e.g. natural wildfires). It is also analysis inclusive, meaning it can incorporate calculation procedures of varying complexity (e.g. hand calculations adopting technical guidance documents, or numerical intensive analysis utilising finite methods or computational fluid dynamics).

ABSTRACT. Introduction Open-sided car parks in the UK are predominantly constructed in concrete, or in steel-concrete composite floor structural systems. Their fire protection measures are typically assessed either through the performance-based design method using localised fire model in Eurocode 1 [1], or the prescriptive design method using Approved Document B (ADB) [2], which requires a 15 minutes fire resistance. It is based upon the assumption that the open- sided car parks are at the over-ventilated condition, hence limited energy can be accumulated to encourage the fire spread between vehicles. The evidence that underpins these design provisions assumes that the fire spread scenarios are limited to three vehicles [3], [4]. However, recent incidents challenge such assumptions, including the Kings Dock car park fire (concrete structure) in 2017 [5], and the London Luton Airport (LLA) car park fire (steel-concrete composite floor structure) in 2023 [6]. Both incidents involved more than 1000 vehicles in fire, demonstrated a clear scenario of travelling fires [7]. The later one ultimately resulted in the disproportionate collapse of the steel-concrete composite floor structure, see Figure 1 (a). More importantly, the travelling fire scenario was significantly affected by the wind in the LLA car park fire, emphasising the importance of investigating the wind-driven travelling fires and its impact on the disproportionate collapse with more than three vehicles in fire.

Research methodology This paper aims to investigate the impact of wind effect on the structural fire response of a large-scale steel- concrete composite floor structure in travelling fires via a real-life case study, i.e., the LLA car park fire. The computational fluid dynamics (CFD) tool, FDS 6.10.1 [8], was used to simulate travelling fire scenarios, which were subsequently applied as fire boundaries for analysing the structural utilisation of critical steel columns along the travelling fire trajectory. Note that this study is not a forensic reconstruction. Rather, it uses the incident as a real- scale benchmark scenario to develop, test, and generalise new numerical models for travelling fires and its effects on the disproportionate collapse of the structure. The geometry of the CFD model follows the dimensions of the LLA car park, with approx. 11000m2 floor area and 2.75m estimated storey height, see Figure 1 (b). The investigated steel columns have dimensions of universal column (UC) 305×305×198 and UC 305×305×158, represented in the model as “obstructions”. The analysis was restricted to the third floor, as the fire originated from a single vehicle on this level and subsequently led to disproportionate collapse in the floor above. A simple pyrolysis model using a fixed heat-flux ignition criterion (i.e., 20kW/m2), were applied in the model. The heat release rate (HRR) of each vehicle is prescribed following [9], with a peak of 6.5MW, as presented in Figure 2 (a). Given the high uncertainty in modelling the initial fire spread stage, it was assumed that seven vehicles were ignited at the “start” of the fire, i.e., 5 mins, see Figure 2 (b) and (c). Two travelling fire scenarios were considered: 1) no-wind, and 2) with wind at 5 m/s reported in the LLA car park fire. Due to the length limit, the CFD model and relevant sensitivity studies will be detailed in the full paper. Once the two travelling fire scenarios were developed, the Eurocode lumped-capacitance method [10] was used to estimate the heat transfer to the steel columns along the travelling fires trajectories. Note that the column compression resistances are 7610kN and 6035kN for the UC 305×305×198 and UC 305×305×158 respectively, and their utilisations are both around 10% at the ambient temperature. The utilisations of those steel columns were then transiently calculated with the estimated heat transfer results.

Preliminary results, conclusion, and next steps As presented in Figure 2, the case without wind has total of 14 vehicles in fire, leading to a maximum HRR of 50MW. In contrast, the case with wind has 69 additional vehicles in fire, leading to a maximum HRR of 145MW. The fire duration is also increased significantly by the scenario with wind, from around 45 mins to 120 mins. In addition, although both scenarios exhibited similar behaviour during the first 15mins (see Figure 3), the wind-driven case develops into a more extensive travelling fire, producing greater spatial and temporal non-uniform heating of the columns and higher internal temperatures, whereas fire spread in the no-wind case remained confined to a more limited area. As presented in Figure 4 (a), a limited number of columns were affected by the fire with no wind, i.e., only two columns increased their utilisation from 10% to 30% and 40%. In comparison, almost all columns were affected by the fire with wind, yielding a clear travelling fire scenario involving the increased utilisations from 10% to 25%, 50%, and 40% along the fire trajectory. This paper finds that once more than three vehicles are in fire, the subsequent fire development can be significantly encouraged by the wind effect, which is difficult to avoid for the open-sided car parks. The number of steel columns with higher utilisations (i.e., approx. 50%) under the wind-driven travelling fires also increased dramatically. In the full paper, we will broaden the travelling fire scenarios to include various wind conditions (e.g., different speed, directions, or transient in both), expand the investigation of structural columns and their utilisations, and include steel connections.

ABSTRACT. In order to investigate the potential of a steel-concrete composite (SC) modular solution for industrial applications compared to conventional reinforced concrete (RC) solutions, the RFSC research project SCIENCE (SC for industrial Energy and Nuclear Construction Efficiency) was launched in 2013. Within the scope of WP6 (Testing of SC elements in fire) of this project, several full-scale fire tests were carried out, including a full scale four-point bending test on a SC composite floor under ISO 834 fire condition. This paper focuses on presenting the results of a numerical simulation carried out on a 3D finite element model of the tested SC composite floor specimen using CAST3M software. These results are compared with those of the fire test. The numerical simulation is based on a hybrid structural model which takes account of structural steel, concrete slab and reinforcing steel mesh with different types of finite elements. The bond/friction between structural steel and concrete is assumed with a friction coefficient of 0.2. In both heat transfer and structural analysis, the thermal and mechanical laws of Eurocode 4 part 1.2 have been used for all SC composite floor materials. The behaviour of concrete material is simulated by using elastic model, the Drucker-Prager elastic-plastic model and the Mazars damage model. A good correlation is obtained between the Mazars damage model and the global fire behaviour of SC composite floor specimen, and therefore the assumptions adopted are suitable for reproducing the real behaviour of the test.

ABSTRACT. Structural fire design referred to the engineering process of ensuring that structural members maintained adequate fire resistance for a prescribed duration such that buildings neither collapsed nor experienced a loss of structural stability prior to fire suppression. This design philosophy secured safe evacuation routes for occupant, provided safe working environments for firefighters, and preserved the load-bearing functionality of structures under elevated temperatures. In general, structural fire design was categorized into prescriptive and performance-based approaches. Prescriptive design relied on uniform empirical specifications that overlooked the inherent differences among structural members, while performance-based design adopted a more flexible engineering framework encompassing fire loads, temperature-dependent member response, and structural failure criteria. Although prescriptive design enabled straightforward and convenient implementation, the limited degree of design freedom often resulted in uneconomical or overly conservative fire protection solutions. Conversely, performance-based design offered significant flexibility but required sophisticated thermos-mechanical analysis capabilities and considerable effort for verification and approval. Until 2019, structural fire design in Korea was implemented exclusively through prescriptive methods. Under the prescriptive code, the fire resistance of steel structures and composite structures with externally exposed steel elements, such as concrete-filled tubular members, was assumed to be zero, and compliance was achieved solely through sprays, boards, and intumescent paints. This framework frequently led to excessive construction costs, and the high market price of 3-hour fire-resistant intumescent paints became a major impediment to the construction industry. However, the code revision in 2019 established an institutional foundation for performance-based structural fire design, which became fully applicable in 2020. The first performance-based structural fire design in Korea was applied in 2020 to the exterior structural framing system of a 20-story office building, completed in 2022. The system was initially considered for performance-based design due to the inability to install or replace fireproofing materials under harsh external environments. Nevertheless, the project demonstrated notable benefits, including substantial reductions in construction cost and project duration, as well as the realization of architectural design intent. Although an increasing number of alternative performance-based proposals had since been attempted, only projects supported by academic expertise had been successfully executed due to the high technical barriers for practitioners. As of 2024, the second performance-based project, comprising concrete-filled tubular columns and composite steel beams in a 8-m-high data hall of a data center building, was under construction, and the third project, an unprotected 10-m-high steel truss system for a sports facility, was also underway. The recent advancement of structural fire design revised interest in fire-resistant steel, which had been developed in the early 2000s but not widely implemented in practice. Whereas conventional structural steel retained approximately two-thirds of its ambient temperature yield strength at 350°C, fire-resistant steel retained the same strength ratio at 600°C. Although fire-resistant steel did not exhibit superior thermal conductivity or ultra-high temperature (> 800°C) behavior, its enhanced strength within the mid-to-high temperature (400 ~ 800°C) range around 600°C provided clear advantages in performance-based design scenarios where the critical temperature remained in that range. This study aims to (i) provide a comprehensive overview of current structural fire design practices in Korea; (ii) discuss the limitations of prescriptive approaches; and (iii) describe the methodologies applicable to performance-based fire-resistance design. In Korea, performance-based design had been conducted using the critical temperature method, the simplified calculation method, and the advanced calculation method. Based on theses methodologies, this study presented three actual projects and investigated their economic outcomes in terms of cost savings and reduction in construction duration. Moreover, using experimentally measured thermal-mechanical properties of fire-resistant steel, the study explored future strategic directions for structural fire design. To further evaluate the potential of performance-based design with fire-resistant steel, additional case studies were conducted: a 15-story steel office building, a steel airport passenger terminal, a steel warehouse, and a sound-proof tunnel. For each case, structural fire design was performed using both conventional structural steel and fire-resistant steel, and the resulting benefits were quantified with respect to cost and construction time for both steel and composite structural systems. Finally, the expected economic savings and construction time reduction associated with the adoption of fire-resistant steel were estimated, and a set of practical strategies was proposed to enhance the effectiveness of future structural fire design.

ABSTRACT. In timber hybrid structures, which have been increasingly adopted in recent years, reinforced concrete (RC) slabs are often employed to prevent fire spread from floor to floor. When the RC slab and the glulam beam were integrated using shear connectors such as lag screws, the composite beam exhibited composite effect at elevated temperatures, resulting in a significant increase in fire resistance period compared with the glulam beam. The composite effect of shear connectors mainly depends on the embedment stiffness and strength parallel to the grain of a glulam beam. In case of a fire, the temperature rise of the embedment region caused a reduction in the embedment stiffness, which decreased the degree of composite effect and resulted in bending failure of the beam. This study developed a component-based numerical model of a partially composite beam to clarify its fire behaviour and stress distribution at ultimate state, taking into account the degradation in the embedment stiffness. The proposed analytical model was validated against the results of load-carrying fire tests. In addition, the influence of the temperature-induced degradation of the shear connector's horizontal spring on the deflection behaviour in fire was evaluated.

ABSTRACT. This work investigates the structural fire behaviour of slender concrete-filled cold-formed steel (CF-CFS) composite columns using coordinated experimental testing and finite element modelling. Twelve full-scale built-up column specimens, incorporating four different cross-sectional configurations constructed from Class 4 S280GD+Z steel profiles and filled with lightweight concrete, were tested under sustained axial loading in an ISO 834 fire environment. The results demonstrate the thermal stabilising role of the concrete core, the significance of geometry-dependent thermal gradients, and the dominant failure modes, including local, distortional and combined buckling interactions. A numerical modelling framework coupling heat transfer, linear buckling, and nonlinear structural analysis is developed and shown to accurately reproduce observed fire-induced degradation and load-bearing response. The study provides critical experimental validation for modelling of CF-CFS columns in fire and informs future refinement of design provisions for composite slender columns under elevated temperature.

ABSTRACT. Pipe racks are essential for the safe transport of hazardous materials, acting as the primary support system for pipelines in industrial facilities. However, these structures are highly vulnerable to wildfires, whose frequency and severity are increasing due to climate change, raising concerns over NATECH (natural hazard-triggered technological) events. Despite this growing risk, quantitative frameworks for assessing wildfire-induced damage to industrial steel structures remain limited. This study provides a methodology for developing fragility functions to evaluate the structural vulnerability of a steel pipe rack under wildfire exposure. Thermomechanical analyses were performed considering different wildfire scenarios, with fire severity quantified through the maximum radiative heat flux. Engineering demand parameters are used to define performance levels, with global collapse associated with interstorey drift limits. Cloud analysis was employed to establish intensity-demand relationships, and lognormal fragility functions were derived to estimate the probability of exceeding specified performance levels. Results indicate a strong correlation between increasing radiative heat flux and structural demand, with forest fire scenarios producing the most severe response. The methodologies and results provide a quantitative basis for wildfire risk assessment and support risk-informed mitigation and design of pipe racks located in Wildland–Industrial Interface areas.

ABSTRACT. Shear failure in fibre-reinforced concrete (FRC) beams under fire conditions is challenging to predict due to its complex nature. As FRC is increasingly common to be used as building materials, there is an urgent need for a prediction tool for engineers to predict shear behaviour of FRC beams under fire. This work attempts to fill that demand by formulating a semi-analytical model based on the strut-and-tie model (STM) to predict the fire endurance of shear-critical FRC beams. The effect of continuity and the moment redistribution due to the stiffness loss of the heated span could also be considered. This novel model is shown to predict the failure time of six shear-critical fibre-reinforced lightweight aggregate concrete beams with good accuracy, highlighting its potential to be deployed as a straightforward and user-friendly prediction tool for engineers.

ABSTRACT. This paper presents original research that is important for evaluating and enhancing the post-fire shear strength of reinforced concrete structures. In such structures, shear may have to be transferred across interfaces between two members or two parts of a member that can slip relative to one another. These interfaces include planes of existing cracks, such as the interface between concretes cast at different times, construction joints between footings and shear walls, and the junction between flanges and webs in T-beams. Shear transfer can also happen across planes that are initially uncracked such as in the case of simply supported girders or the vertical interface between corbels and their supporting columns [1]. Steel fibres are added to concrete to enhance its engineering properties, such as strength flexural and shear strengths, fracture toughness, fatigue resistance, and impact resistance. The inclusion of steel fibres has also been proven to provide significant improvement in shear transfer strength of concrete. Exposure of concrete to elevated temperatures due to fire causes a significant deterioration in its mechanical properties resulting in reduced shear transfer capacity of reinforced concrete elements. Although shear is often the cause of several structural failures during real fire events [2], most fire tests have focused on the flexural behaviour of beams and slabs and the axial behaviour of columns. While a few investigations have revealed the effectiveness of steel fibres in the enhancement of residual mechanical response after exposure to elevated temperatures, only a little attention [3] has been directed to study the effect of elevated temperatures on the shear transfer behaviour of steel fibre reinforced concrete (SFRC). This paper presents the results of an extensive series of tests on the shear transfer performance of SFRC after exposure to elevated temperatures. Twelve initially un-cracked push-off specimens were cast and tested under direct shear. Two mixes were employed in this phase, one mix being made with conventional concrete to serve as the reference mix; the other was made with SFRC at 0.67% volume of steel fibres. The steel fibres were hooked-end with 0.54 mm diameter and 35mm length (aspect ratio L/d= 65) with an ultimate strength of 1100 MPa. The push-off specimens chosen for the residual tests were heated to 500oC with a heating rate of 5oC/minute and a soaking period of 1 hour, after which the specimens were allowed to cool naturally to room temperature as shown in Figure 1a. Measurements were taken for residual shear transfer strength, crack widths and shear slip (see Figure 1b). Digital Image Correlation (DIC) was conducted on some specimens of each of the conventional concrete and SFRC. The test results showed that conventional concrete push-off specimens lost, on average, two-thirds of their shear transfer capacity after exposure to 500oC. This reduction in shear transfer strength was associated with larger crack widths and shear slip compared with those of the unheated specimens. Steel fibres considerably enhanced the residual shear transfer strength after exposure to elevated temperature, as shown in Figure 1c. The test results showed that the addition of fibres also led to a ductile failure and improvements in the energy dissipation capacity of SFRC in the post-cracking regime at ambient or after exposure to high temperatures, as shown in Figure 1d. DIC analysis showed that lower shear strains were developed in SFRC than conventional concrete (see Figure 1e). This reduction in shear strains developed in SFRC specimens can be attributed to the effectiveness of steel fibres in bridging cracks through the pullout and dowel actions. The reduction in concrete shear transfer in push-off specimens due to elevated temperatures reveals the need to conduct more research to study the effect of fire on the shear transfer strength in concrete structures.

ABSTRACT. This study investigates the stability of reinforced concrete (RC) columns during natural fire tests, focusing on the cooling phase and residual capacity. Three full-scale R60-designed columns, stored outdoors with high moisture content (3.7-5.9 %), were tested under realistic fire loads. Despite experiencing spalling, two of the three columns withstood the entire fire exposure, including the cooling phase. Residual load-bearing capacities ranging from 1200 to 1540 kN were recorded 70-80 hours after the fire. Numerical simulations using SAFIR qualitatively matched experimental results but overestimated temperatures and prematurely predicted failure in spalled models. The data will aid in calibrating numerical models for improved fire safety predictions.

ABSTRACT. In this paper, the influence of fire loading on the shear capacity of reinforced concrete beams is inspected both experimentally and numerically.

ABSTRACT. Concrete spalling under high-temperature conditions, such as those encountered in fires, presents a critical challenge to the structural integrity and safety of modern infrastructure. The coupled thermo-hygro-mechanical (THM) process are widely recognized as the fundamental mechanism of driving spalling behavior. However, the dominant governing parameters remain contentious, leading to two competing hypotheses: the pore pressure mechanism and the thermally induced stress mechanism. Through an integrated program of modeling and controlled experimentation, this work aims to pinpoint the dominant mechanism of concrete spalling. Specifically, we develop a fully coupled thermo-hygro-mechanical peridynamic (PD) framework to simulate spalling progression and analyze stress to track the fracture path. In parallel, we design a series of spalling experiments to evaluate the influence of mechanical restraint condition and fibre content. The systematic comparison of results from controlled numerical and experimental case studies will quantify the contribution of each mechanism.

ABSTRACT. Columns often experience fire loading after sustaining structural damage due to extreme mechanical actions such as earthquakes or blasts. This work examines the impact of cyclic damage on the fire resistance of CFST columns with fibre reinforced concrete through a comprehensive experimental investigation. Table 1 presents a summary of the experiments conducted, along with key findings and observations. This study investigated six square CFT columns with outer dimensions of 220 mm × 220 mm. All specimens were fabricated using commercially available cold-formed square steel tubes. The tests were conducted in two stages: (1) quasi-static cyclic testing to induce earthquake-type damage, followed by (2) fire testing. The square CFST columns were subjected to cyclic loading in lateral direction while a constant axial compressive force (load ratio of 0.3) was maintained throughout the quasi-static tests until the desired level of seismic damage was achieved. The FEMA 461 loading protocol was used to apply the cyclic loading. Subsequently, the damaged columns were exposed to fire loading in accordance with the ISO 834 standard fire curve, while the axial load was kept constant. Displacement transducers, strain gauges, and a 3D digital image correlation system were used to measure strain fields, deformations, and buckling or bulging of the tube. The test setups for the fire and cyclic tests are shown in Figures 1 and 2. The hysteresis curve for specimen S08-F-DR6 is presented in Figure 3(a), while Figure 3(b) shows a representative envelope curve. Figures 3(c) and 3(d) compare axial deformation over time for the CFST columns, and Figures 3(e) and 3(f) illustrate the temperature profiles within the specimens. The failure times of the undamaged columns indicate that the specimen with the 6-mm-thick tube (S06-F-DR0) exhibited a higher level of fire resistance than the specimen with the 8-mm-thick tube (S08-F-DR0). This counterintuitive result is attributed to the larger steel contribution in the 8-mm specimen, whose load-carrying capacity is almost entirely diminished at elevated temperatures. Additionally, both specimens showed comparable axial-shortening trends. As shown in Table 1, the opposite trend was observed when the 6-mm specimens were subjected to cyclic damage: they exhibited lower fire resistance than the 8-mm specimens. The reduction in fire resistance is attributed to local buckling of the tube, damage to concrete caused by cyclic loading, and reduced confinement resulting from the local buckling/bulging of the tube. The addition of fibres significantly improved the earthquake performance of the 6-mm-tube specimens, whereas its effect on the 8-mm-tube specimens was less pronounced. Increasing the level of cyclic damage correspondingly reduced the fire resistance, demonstrating the detrimental effect of prior cyclic damage on structural performance under fire conditions. The inclusion of fibres enhanced fire performance across all specimens, contributing to improved resistance and structural stability at elevated temperatures.

ABSTRACT. Fire design of steel structures can rely on the simplified critical temperature method of EN 1993-1-2, which evaluates member resistance assuming uniform temperature and no interaction with the surrounding structure. In practice, however, beams are connected to adjacent elements that restrain thermal expansion, which generates axial forces and moments during fire. Prior studies have shown that restraint can influence internal forces, deflection, and the development of catenary action, while recent experiments suggest that the influence on the temperature at which bending failure occurs may be limited. However, there remains a lack of clarity on how axial and rotational restraint influence the bending failure temperature of members, and whether the Eurocode simplified method needs adjustment to account for these effects. This study addresses this gap through a systematic numerical investigation of restrained steel beams under fire, quantifying the effects of axial and rotational restraint on the critical temperature. The analysis covered more than 5000 combinations of load ratio, section size, beam length, restraint stiffness (axial and rotational), and restraint position. The results show that axial restraint has little influence on the critical temperature for bending failure, supporting the applicability of the simplified EN 1993-1-2 critical temperature method to restrained beams under uniform heating, for the resistance criteria. Axial restraint, however, significantly increases deformations of the beams, which may create potential integrity concerns.

ABSTRACT. Iron-based shape memory alloy (Fe-SMA) has overcome the limitations of traditional materials due to its superelasticity and shape memory effects. However, the mechanical performance of bolted connections between Fe-SMA and steel plates (FSBCs) under fire conditions remains unexplored, which significantly limits the practical application of Fe-SMA. To address this research gap, 14 steady-state high-temperature tests were conducted on FSBCs, revealing the failure mechanism of FSBCs under fire exposure. Furthermore, a validated numerical model was developed in ABAQUS, and the effects of geometric parameters of the Fe-SMA plate and bolt on the mechanical behavior of FSBCs under fire conditions were systematically investigated. The results demonstrated that the optimal geometric parameters of the Fe-SMA plate and bolt contribute to superior mechanical performance and economic efficiency of FSBCs. Additionally, the strength reduction method should be appropriately modified to ensure high accuracy in assessing the behavior of FSBCs under fire conditions.

ABSTRACT. Iron-based shape memory alloy (Fe-SMA), as an emerging advanced material with distinctive shape memory effect and superelasticity, has attracted extensive attention in engineering fields. This study experimentally investigates the transient-state mechanical behaviour of Fe-SMA and the fire resistance of T-stub joints incorporating Fe-SMA bolts. Fe-SMA outperforms high-strength steels and bolt materials in terms of reduction factors in mechanical properties at elevated temperatures, making it a promising alternative metal for structural fire-resistant applications of steel structures. Besides, predictive models for reduction factors of transient-state mechanical properties of Fe-SMA are established as a basis for engineering applications at elevated temperatures. Furthermore, fire testing of T-stub joints employing Fe-SMA bolts further confirms their enhanced fire resistance compared to those with high-strength bolts. The parametric investigations covering bolt material, bolt size, flange thickness, and bolt positioning offer a novel perspective for the development of active fire-resistant strategies in beam-to-column connections.

ABSTRACT. Introduction Fire safety design of steel structures requires more advanced assessment tools than traditional pre-scriptive checks, as steel undergoes significant loss of strength and stiffness with increasing tem-perature, potentially compromising the global stability of the structure. Although modern fire engi-neering provides models capable of describing the thermo-mechanical behaviour of structural sys-tems, their integration into an optimized decision-making process is often hindered by the difficulty of linking structural performance to a reliable quantification of life-cycle economic benefits. In this context, this work proposes the development and validation of an integrated methodology that al-lows steel structures with passive fire protection to be evaluated not only in terms of technical effec-tiveness, but also through a cost–benefit analysis founded on structural vulnerability derived from fragility curves under natural fire conditions. Methodology Rather than limiting the assessment to fire resistance characterization, the methodology aims to con-struct a quantitative process in which structural behaviour is linked to the probability of exceeding predefined performance levels and to the expected economic impact over the structure’s life cycle. The methodological framework includes several consecutive phases (see Fig. 1), such as the identifi-cation of representative structural typologies, the definition of realistic localized fire scenarios, ad-vanced thermo-mechanical modelling at elevated temperatures, and the selection of performance indicators, such as maximum column-top displacements or mid-span deflections of beams. These results feed into the development of fragility curves, derived through modified Cloud Analysis [1] and constructed using different fire intensity measures, such as fire load energy and peak heat re-lease rate, in order to statistically represent the structural vulnerability associated with the proposed performance levels.

Figure 1: Flowchart of the proposed methodology Once the fragility curves are obtained, they are integrated into a Life Cycle Cost-Benefit Analysis model [2] that includes initial costs, maintenance expenses [3], and the economic consequences associated with direct [4] and indirect [5] damage linked to the probability of exceeding the perfor-mance thresholds. This approach allows passive protection to be evaluated as an optimizable design choice rather than a mandatory requirement, making it possible to identify the most advantageous configuration in relation to expected fire severity and the desired performance level. Proof-of-Concept application The methodology is applied to a steel building used as a vehicle storage facility, characterized by a structural system composed of steel trusses with hollow sections and HEB columns. The structure is analysed in both unprotected and protected configurations, considering different layer of fire-resistant sprayed plaster. A total of 72 fire scenarios are defined, derived from combinations of ve-hicles whose heat release rate curves are represented through simplified modelling approaches. Thermal and thermo-mechanical analyses, conducted using SAFIR®, show that passive protection significantly reduces steel temperature and limits deformation development, thereby improving the structure’s capacity to satisfy higher performance levels. The fragility curves confirm that the unpro-tected configuration exhibits substantially higher vulnerability, whereas passive protection drastically reduces the probability of performance exceedance. Increasing the thickness of the protection further enhances this improvement, particularly in high-severity fire scenarios. The life-cycle economic analysis demonstrates that protected configurations are cost-effective across a wide range of fire intensities. The initial cost of passive protection is largely offset by the reduction in expected direct and indirect losses, with even greater economic benefits observed for larger protection thicknesses due to the reduced structural fragility (see Fig. 2). The proposed ap-proach shows that combining fragility-based vulnerability assessment with cost-benefit analysis pro-vides a robust framework to support more informed design decisions, balancing safety, perfor-mance, and economic sustainability.

Figure 2: Comparison between PVLCC of unprotected and protected configurations, by assuming two different IMs: fire load (a) and HRR peak (b). References [1] Jalayer, F., Ebrahimian, H., Miano, A., Manfredi, G., and Sezen, H. (2017), Analytical fragility as-sessment using unscaled ground motion records. Earthquake Engineering and Structural Dynam-ics. https://doi.org/10.1002/eqe.2922 [2] Ma, C., Van Coile, R., and Gernay, T. (2024), Fire protection costs in composite buildings for cost-benefit analysis of fire designs. Journal of Constructional Steel Research. https:// doi.org/10.1016/j.jcsr.2024.108517 [3] Miano, A., Sezen, H., Jalayer, F., and Prota, A. (2019), Performance-based assessment method-ology for retrofit of buildings. Journal of Structural Engineering. https://doi.org/ 10.1061/(ASCE)ST.1943-541X.0002419 [4] National Institute of Building Science (NIBS). (2003), HAZUS-MH MR1 Technical Manual, devel-oped by the Federal Emergency Management Agency, Washington, D.C. [5] Hicks, H. L., and Liebermann, R. R. (1979), Study of Indirect Fire Losses in Non-Residential Prop-erties. FoU-brand, 8-14, January 1979.

ABSTRACT. This work reports the results of a numerical investigation on the post-buckling behaviour, ultimate strength and Direct Strength Method (DSM) design of cold-formed steel (CFS) simply supported lipped channel beams (LCB) subjected to uniform major-axis bending and exhibiting distortional-global (D-G) buckling interaction at elevated temperatures. The study builds on previous work by the authors concerning members failing in pure distortional or pure global (lateral-torsional  LT) modes. The beams considered (i) exhibit different levels of D-G interaction, (ii) are exposed to temperatures up to 800°C, (iii) cover various cross-section geometries, lengths and yield stresses, and (iv) have two support conditions differing only in end cross-section warping and local displacement/rotation restraints (partially restrained or prevented). The results presented and discussed include post-buckling equilibrium paths, failure moments and associated collapse modes (distortional, global or interactive), obtained from shell finite element geometrically and materially non-linear analyses with imperfections (GMNIA) performed in ABAQUS. The temperature-dependent material behaviour follows the EN 1993-1-2 model. The assembled D-G interactive failure moment data are used to develop and assess the merits of the DSM-based design approaches accounting for interaction and elevated temperature effects  they are shown to handle adequately interactive failures in LCB, thus providing a good starting point for a rational and efficient design approach for arbitrary CFS beams under fire conditions

ABSTRACT. Localised fires commonly occur in large single-storey industrial halls where combustible materials are concentrated in limited areas. Nominal fire curves used in prescriptive design cannot represent the non-uniform thermal environment produced by such fires, particularly when flashover does not develop. This study applies performance-based fire design by integrating simplified localised fire models with two-zone modelling to more realistically estimate thermal actions on structural components. A 3D thermal-mechanical model of a steel industrial hall was developed in SAFIR, using thermal inputs from OZone, Hasemi and LOCAFI for a 50 MW, 8 m-diameter localised fire. Two strategies for applying the Hasemi model were examined: using the Eurocode H/10 height criterion and using the zone-interface height predicted by the two-zone model. Preliminary results show that the choice of height threshold influences column deformation and structural fire resistance, and that the H/10 limit may not always provide a sufficiently conservative boundary when hot-layer temperatures become dominant. A parametric study on opening factor and fire severity further highlights conditions under which two-zone modelling should supplement localised-fire analysis. The findings aim to support improved guidance on localised fires in future Eurocode developments.

ABSTRACT. The post-fire mechanical properties of bridge steels, which were rarely investigated, are the foundation for assessing the load-bearing capacity of bridges after fire events. This study investigates the post-fire mechanical properties of bridge steel Q370qD, considering two different cooling methods, namely cooled in air and cooled in water. Results indicate that after the ultimate experienced temperature Tu surpasses 700 ℃, the post-fire mechanical properties of Q370qD steel are significantly influenced by Tu. Compared to the similar grade structural steel Q355, the post-fire reduction factors of mechanical property parameters of both steels are similar when Tu is below 700 ℃. However, when Tu exceeds 700 ℃, the post-fire reduction factors of mechanical property parameters of Q370qD and Q355 are significantly different, especially for water cooling.

ABSTRACT. Concrete-filled steel tubular (CFST) columns with high-strength embedded steel cores provide high axial capacity, slender geometries, and robust fire performance. However, in composite structural systems, overall reliability is strongly governed by connection behavior and the efficiency of force transfer to the core while preserving floor and frame integrity. This study proposes and numerically investigates two connection concepts developed specifically for CFST columns with bar-bundle cores: a prefabricated beam-to-column connection head and a slab-to-column punching shear head.

The beam connection replaces conventional bolted tube attachments and thick welded diaphragms with a compact head and stub arrangement that locates most steel components within the concrete slab, thereby protecting critical load-transfer elements at elevated temperatures. The slab connection features a recessed punching shear head, anchored to the protruding bar bundle and combined with continuous slab reinforcement, to improve punching resistance and promote membrane action in fire.

The thermo-mechanical response is analyzed in Abaqus using GMNIA under ISO-834 fire exposure, with concrete modeled via damage plasticity, steel assigned temperature-dependent high-strength properties, and bond deterioration included to capture slip and partial decoupling. The results indicate a hierarchical, ductile behavior: tube resistance degrades first while the cooler core and head maintain axial capacity; load paths progressively shift into the bar bundle; punching failure is delayed; and calibrated semi-rigid rotational restraint enhances global stability. This work provides a foundation for subsequent experimental validation and parametric studies aimed at developing simplified temperature-dependent design rules and equivalent rotational spring stiffnesses for composite systems employing CFST columns with embedded steel cores.

ABSTRACT. Extended Abstract: Background: Although the research on the fire resistance of ordinary steel-concrete composite beams has been increasing in the past few decades, there is still a lack of in-depth research on the fire resistance of high-strength steel-concrete composite beams. Therefore, the significance of this study is to improve the fire resistance design theory of high-performance steel-concrete composite beams with end restraints, provide technical support for exploring the failure mechanism of high-performance steel-concrete composite beams with end restraints under fire, and accurately carry out the fire resistance design and checking calculation of high-strength steel-concrete composite beams, and provide a new perspective and direction for the research of new steel-concrete composite structures. To fill the current research gap, our researches carry out the fire resistance test of high-strength steel-concrete composite beams, and obtain the temperature distribution, flexural deformation, fire resistance limit, critical temperature and different end constraint conditions, so as to comprehensively investigate the influence of steel strength and end constraint forms on the fire resistance and failure mode of composite beams, and provide test data for finite element simulation and theoretical verification. Methods: All high-strength steel-concrete composite beams were designed with sufficient dimensions. The net span of the composite beams is 5700 mm, and the beam height is 370 mm. The steel beams were manufactured with hot-rolled HN section steel, with dimensions of the Chinese national standard HN250 mm × 125 mm × 6 mm × 9 mm. Steel beam strength was grade Q460. The concrete strength grade was C30. The concrete slab was 120mm thick. Inside the concrete slab, there were double layered and bi-directional steel bars HRB400 with a diameter of 8mm, with a spacing of 150mm between the horizontal and vertical directions. The beam column nodes of the constrained composite beam were connected by end plate bolt welding, which was a rigid connection. The loading mode of the composite beam was two-point concentrated loading, and the loading point was the third point of the net span of the composite beam. Before starting the ignition, a load should be applied in advance, and after being loaded in stages and proportions to the design load value, the ignition temperature should be increased. During the test, the load value should be kept constant according to the changes in the deflection of the component. The temperature rise was carried out according to the ISO-834 standard temperature rise curve. The steel column at the end of the composite beam was connected to the horizontal reaction frame through tension and compression sensors to provide end constraints for the specimen. Two tension and compression sensors were placed on each side of the composite beam specimen, namely the upper and lower parts of the steel column, to measure the internal force changes of the end constrained composite beam during deformation under fire. The experiment tests included a simple supported composite beam fire test and three end restrained beam fire tests with different load ratios, complete shear resistance, and partial shear resistance. Results: The failure mode of simply supported composite beam under fire was bending failure, with few surface cracks and mainly longitudinal cracks. There were water marks on the concrete slab surface at the end of the beam, and only a small amount of water vapor is emitted. Under fire, a significant plastic hinge was formed at the negative bending moment of the restrained composite beam at the beam end, and the steel beam end undergone significant buckling, resulting in severe cracking of the concrete at the beam end. During the experiment, a small number of cracks appeared in the concrete at the end of the beam, followed by an increase in crack width. There were more water marks on the surface of the mid span position, with a small amount of water accumulation and a large amount of evaporation. Finally, a large number of transverse cracks appeared at the end of the concrete slab surface beam, and the crack width significantly increased, with longitudinal cracks in the mid span. For load ratios of 0.3 and 0.5, the fire resistance limit of beam end restrained composite beams under high temperature decreased with the increase of load ratio, and is significantly higher than that of simply supported composite beams. The fire resistance limit of fully shear restrained composite beams under a load ratio of 0.5 was roughly the same as that of partially shear restrained composite beams, with a fire resistance time of about 40 minutes and basically the same failure mode. Under the action of fire, the steel columns of the restrained composite beam are subjected to tensile and compressive forces respectively. As the temperature increased, the axial force at the end rapidly increased to its peak and then gradually decreased. Conclusions: The failure mode of high-strength steel composite beams under fire is bending failure, and local buckling occurs at the negative bending moment at the beam end of the restrained composite beam under fire. The fire resistance limit of beam end restrained composite beams decreases with the increase of load ratio and is significantly higher than that of simply supported composite beams. The fire resistance limit of fully shear restrained composite beams is roughly the same as that of partially shear restrained composite beams. During the stage of large deflection deformation, the load-bearing mode of composite beams changes from bending to a combination of bending and catenary effects. The catenary effect is beneficial for improving the fire resistance performance of composite beams. The composite beam with beam end constraints has significantly higher fire resistance performance than simply supported beams, providing guidance for fire design in practical construction projects.

ABSTRACT. In this study, the bond stress–slip relationship between reinforcing bars and concrete at an elevated temperature was investigated through a series of pull-out tests. In the experiments, the load normally acting on reinforced concrete members was simulated by applying a predetermined pull-out load to the reinforcing bar before and during heating. This approach enabled a systematic evaluation of the influence of initial loading levels on subsequent bond–slip behavior under fire exposure. The experimental results indicated that the pull-out load applied to the reinforcing bar before and during heating had a negligible effect on bond strength. However, a significant thermal slip of approximately 0.8 mm, corresponding to approximately 1% of the bonded length was observed. These findings suggest that although ultimate bond capacity is not significantly affected by initial stress states, the heating process could induce significant interfacial deformations, potentially leading to stress relaxation between reinforcing bars and concrete. Therefore, such heating-induced slip should be considered when assessing the structural safety of fire-exposed reinforced concrete.

ABSTRACT. Concrete-filled steel tube (CFST) columns are known to possess superior structural performance and enhanced fire resistance rating (FRR) due to the thermal mass effect of the concrete core, which significantly reduces the heating rate of the steel shell. While unprotected CFST columns often achieve a one-hour FRR, achieving longer durations, such as the three-hour requirement common in international building codes (e.g., IBC), typically requires either internal reinforcement or the application of passive fire protection (PFP) systems. Despite extensive research on the fire performance of unprotected CFST members, studies on their behavior when combined with external PFP are limited. Critically, the quantitative role of the concrete core in reducing the required PFP thickness compared to empty hollow structural sections (HSS) has not been thoroughly investigated or quantified. This is an essential distinction, as the external insulation significantly alters the temperature gradient within the CFST section, necessitating a separate and specialized study. Furthermore, reliable numerical simulation of CFST columns remains challenging due to the difficulty in modelling the air gap that forms between the steel and concrete at elevated temperatures. To address this critical gap, an experimental study was conducted on twelve square HSS columns, ten of which were concrete-filled (CFST). These columns were subjected to fire resistance testing in accordance with BS EN 13381-4 and BS EN 13381-6 (Figure 1). Ten specimens were protected using sprayed applied fire resistive material (SFRM), including both gypsum-based and Portland cement-based coatings, a cost-effective and common PFP solution. To ensure accurate temperature field mapping and inform future numerical models, detailed measurements were taken using eight thermocouples on the steel wall and ten thermocouples embedded within the concrete core of each specimen (Figure 2). The tests confirmed that the CFST columns required significantly less PFP thickness than similar protected HSS columns to achieve the same FRR. Using the evaluation procedures of BS EN 13381-4/6, the experimental data was converted into design tables quantifying the PFP thickness reduction across different FRRs (1 to 4 hours). For a two-hour FRR, the presence of concrete infill resulted in a thickness reduction between 35% and 68% across both coating types and section factors (from 54 m-1 to 103 m-1). For a three-hour FRR, the reduction ranged from 47% to 63% (Figure 3). A key finding is that the effect of the concrete core becomes more pronounced as the steel wall section factor increases (i.e., thinner steel walls). These quantitative results provide essential, practical data for civil engineering projects, enabling the optimized and economical design of CFST columns using external PFP. Furthermore, the detailed temperature history recorded in the protected CFST sections supports the assessment of column reusability after a fire, a significant advantage, as the PFP limits temperature rise, potentially preserving the mechanical properties of the steel and concrete.

ABSTRACT. This study investigates the residual mechanical properties of corroded Q235 structural steel after elevated-temperature exposure. Specimens were subjected to 0, 42, and 72 days of salt-spray corrosion, then exposed to temperatures ranging from 400 °C to 1000 °C and cooled either in air or water. Tensile tests were conducted at room temperature to evaluate post-fire strength and ductility. Results show that water-cooled specimens exhibited significant strength increases up to 800 °C, followed by strength loss at 1000 °C. In contrast, corrosion reduced both yield strength and ductility, with more pronounced degradation after high-temperature exposure. These findings highlight the combined influence of corrosion level, peak temperature, and cooling method on the post-fire mechanical behavior of Q235 steel, providing insight into the performance of aging steel structures exposed to fire.

ABSTRACT. The present study investigates the behaviour of cold-formed steel (CFS) beam-to-column bolted shear connections using clip-angle connectors under fire conditions. The experimental program focused on examining the influence of connector width and depth on the overall connection performance. Both ambient temperature and elevated temperature tests were conducted to evaluate the behaviour of the connections. A customised furnace was specifically developed to test beam-to-column assemblies, equipped with provisions for monitoring strain and deformation characteristics of the connection components using the 3-D DIC technique. Ambient temperature tests were first conducted to establish the baseline connection behaviour. In the subsequent fire tests, a constant load corresponding to 50% of the ambient load-carrying capacity was applied to represent realistic serviceability conditions. The specimens were then exposed to the ISO 834 standard fire curve until failure. Failure modes were identified and classified for both ambient and elevated temperature conditions, providing valuable insights into the deformation mechanisms, strength degradation, and rotational behaviour. The critical temperature corresponding to connection failure was determined. The results contribute to improving the understanding of connection behaviour and form a basis for developing enhanced design approaches for cold-formed steel beam-to-column connections subjected to fire.

ABSTRACT. Hempcrete is a composite material mainly composed of hemp shiv, a lime-based binder, and water. It is recognized for its ecological properties, thermal insulation capacity, and ability to store carbon. Despite its hygrothermal and environmental advantages, fire resistance remains a major challenge for its use in construction. Existing studies on the fire behavior of hempcrete are limited, especially for formulations where lime is partially replaced by natural pozzolan.

This study aims to experimentally characterize the fire behavior of hempcrete with a partial substitution of lime by natural pozzolan. The materials used (hemp shiv, hydrated lime, natural pozzolan) were sourced from Quebec. Fire resistance tests were conducted using a cone calorimeter, in accordance with ISO 5660 and CAN/ULC S135 standards. Twelve samples were tested under an irradiance of 50 kW/m², measuring mass loss, heat release rate, critical heat flux, flammability, smoke production, and structural stability. Gas emissions, notably CO and CO₂, were also analyzed.

The integration of pozzolan into hempcrete could improve its thermal stability and fire resistance. The results obtained will serve as a basis for developing guidelines for the use of this material in construction, aiming to meet modern safety requirements.

ABSTRACT. Post-fire assessment of concrete structures requires reliable methods to evaluate residual load-bearing capacity and inform decisions on reuse, repair, or demolition. This study examines the feasibility and performance of common destructive and non-destructive testing techniques through controlled heating experiments on concrete cube specimens. Eight 10 cm cubes were subjected to one-dimensional heat transfer using radiant panel heating, producing maximum surface temperatures of 600-700 °C, with well-characterised internal temperature profiles verified through numerical modelling. Ultrasonic pulse velocity (UPV), rebound hammer, thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and compressive and tensile strength testing were systematically applied to assess thermal damage. Results indicate that non-destructive methods (UPV and rebound hammer) exhibit clear sensitivity to damage only when temperatures exceed approximately 400-500 °C. Compressive strength testing showed reasonable correlation with code-based strength retention models, though loading direction relative to thermal gradients significantly affects results. Material characterisation techniques (TGA and DSC) confirmed expected decomposition patterns but proved challenging for quantitative assessment. An integrated framework combining rapid non-destructive screening with selective destructive verification is recommended for practical post-fire structural evaluation.

ABSTRACT. During the 2020-2025 National Model Code development cycle, the Canadian Wood Council submitted a proposed Code change request (CCR) to the 2020 National Fire Code of Canada (NFCC) to permit mass timber ceilings to remain exposed during the construction of tall wood buildings, which was subsequently approved by the Canadian Board for Harmonized Construction Codes for publication in the 2025 edition of the NFCC. While the submission of this CCR was driven by difficulties encountered when reconciling the prescriptive protective encapsulation (gypsum board) requirements of the 2020 NFCC during construction with measures to control moisture and mold affecting the encapsulated mass timber elements in ongoing projects across Canada, its technical rationale was largely based on the results from a series of full-scale fire tests titled ‘Mass Timber Demonstration Fire Test Program’ (MTDFTP) conducted in Ottawa in 2022. This paper summarizes key findings from the two construction-site-fire-scenario tests, Tests 3 and 4, conducted as part of the wider MTDFTP study, focusing on (a) time-histories of fire development within the compartments, (b) fire impact on exposed mass timber elements, (c) fire exposure outside the compartments, and (d) post-fire suppression operations observed across the two tests.

ABSTRACT. This paper presents a numerical validation study of enclosure fires featuring exposed Cross-Laminated Timber (CLT) ceilings using the Fire Dynamics Simulator (FDS). To address the complexities of modelling timber pyrolysis, the study utilises the SPyro (Scaling Pyrolysis) model, which adopts a data-driven approach by scaling bench-scale cone calorimeter data to predict full-scale burning behaviour. The model is validated against a large-scale compartment fire experiment featuring a 2.5 MW propane fire source and a CLT ceiling bonded with heat-resistant adhesive. Results indicate that the SPyro model effectively captures the insulating effects of the char layer and predicts key fire dynamics indicators, including gas-phase temperatures and radiative heat flux, with reasonable accuracy. The simulation successfully reproduced the self-extinction behaviour observed in the experiment. The study concludes that FDS with SPyro is a robust engineering tool, provided that mechanical failure modes such as delamination are prevented.

ABSTRACT. Introduction: While heritage structures have immense value due to their historical and cultural significance, the long use and exposure to the natural elements makes them extremely susceptible to fires. The International Association of Fire and Rescue Services, CTIF, estimates that one historic building is affected by fire every day [1]. Examples of heritage building partially or destroyed by fire include The Mackintosh Building in Glasgow, Scotland in 2014 and 2018, Notre Dame cathedral in Paris, France and Shuri Castle in Okinawa, Japan in 2019. The centuries long use of timber as a construction material makes it present in many historic buildings and thus, understanding its fire performance is crucial for the preservation of heritage buildings.

The long-term use of timber is associated with the development of defects such as cracks as well as changes in the appearance and the chemical composition and microstructure. All the above result in decreased mechanical properties [2]. Yet the exact effects of the aging of wood and exposure to the elements on the fire performance of heritage timber are not fully understood. Therefore, this research applies a combination of fire testing methods with revolutionary image analysis techniques for the monitoring of fire performance of heritage timber. Video recordings of the fire testing of heritage timber are obtained by exposing an Eastern Hemlock mortise and tenon connection to an open pool fire practical test combined with narrow-spectrum illumination technique [3]. The open pool fire allows for the monitoring of localised fire damage on the connection while narrow-spectrum illumination reduces the effects of thermal radiation on the collected image data, providing clear images for analysis. The collected data is then analysed by applying digital image correlation, DIC, techniques for observing the formation of charring on the connection.

Colour Thresholding: The monitoring of the formation of charring on the connection is performed by determining a colour threshold to distinguish between the colour of charred and uncharred timber, a process referred to as colour segmentation. When choosing the colour model applied for the extraction of char formation, the change in the light and colour due to the fire has to be considered. While a RGB (reg, green, blue) colour model is often applied for colour segmentation, the accuracy of the results is affected by changes in the light [4]. A novel approach, applying and HSV (hue, saturation, value) colour model, which separates the characteristics of the colour from the brightness [5], is applied to minimise the effects of the changes in the light caused by the fire on the accuracy of the observed charring.

Monitoring the formation of charring: Once the colour threshold for the charred and uncharred timber is determined, the fire behaviour of the connection is monitored using Python code for choosing points of interest (POIs) and edge detection. Similarly to Gatien et al. [6], the POIs were determined by applying a mesh of equally spaced points within the boundaries of the connection but avoiding the area of the defects. In addition to the meshing, edge recognition is applied to determine the boundaries of the defects. To ensure that full observations of the effect of defects on the charring behaviour are made, the full defects edges are monitored instead of a limited number of POIs on them.

Conclusion: In order to determine the fire behaviour of heritage timber, an Eastern Hemlock connection is subjected to an open pool fire test. To clearly monitor the development of charring, narrow-spectrum illumination technique is used to minimise the effects of fire on the recorded videos. The recordings are analysed using HSV colour model for colour segmentation as well as a combination of meshing and edge detection techniques for monitoring the formation of charring during the test. The novel combination of standard fire testing with DIC techniques for colour thresholding and monitoring of POIs provides crucial knowledge about the charring behaviour of heritage timber and how it is affected by defects in the material.

ABSTRACT. This pioneering experimental study investigates the feasibility of using a Fabric-Reinforced Cementitious Matrix (FRCM) system as a repair technique for fire-damaged glued-laminated timber (glulam) beams. The broader adoption of mass timber is, to some extent, limited by uncertainties about fire safety and post-fire repairability. Existing approaches for post-fire repair focus on wood prosthesis or epoxy-bonded reinforcement, both of which face practical and thermal limitations. To address this gap, two full-scale glulam beams that had previously undergone a standard fire at the Lakehead University Fire Testing and Research Laboratory (LUFTRL) were selected. Their post-fire residual sections were assessed through char-depth surveys and cross-section loss analysis. A repair system was then designed, consisting of slag-modified cementitious mortar shell reinforced with carbon fibre mesh. The experimental program of this study evaluates whether the retrofitting technique using the FRCM system can (i) provide the minimum acceptable fire encapsulation rating as per the National Building Code of Canada, and (ii) restore the flexural strength of the glulam beams after the fire damage. The findings aim to demonstrate a viable repair technique for fire-damaged mass timber elements, promoting more sustainable and resilient timber construction practices in Canada.

ABSTRACT. Due to the lack of unified design guidelines for mass timber connections with glued-in rods (GIR), their broader application remains limited, particularly regarding their performance as moment-resisting joints under fire conditions. In this new study, the structural performance of two full-size mass timber GIR beam-to-column connections has been evaluated through laboratory testing. Each connection configuration utilized six 400-mm-long rods, made of either carbon steel or glass fibre-reinforced polymer (GFRP). Both test assemblies were subjected to 30% of the connection's maximum ambient design load during standard fire exposure, aiming to achieve a minimum of 90 minutes of fire resistance at this loading level. The fire tests involved applying a vertical load at the end of a glued-laminated timber (glulam) cantilever beam to develop shear force (50 kN) and bending moment (75 kN · m) at the beam-to-column connection. Thermal and mechanical data were collected for each test assembly to determine their charring and rotational behaviour during the fire tests. Test results show that the connection with steel rods and that with GFRP rods began to lose load-carrying capacity at approximately 57 and 67 minutes, respectively, under standard fire exposure. The results also show that the adhesive's glass transition temperature significantly influenced the failure time and behaviour of the connections. Since both connection configurations failed before the 90-minute fire-resistance time, further design improvements are needed to achieve the targeted fire-resistance time. The outcomes of this study provide a scientific basis for advancing the fire design and testing of full-size, loaded mass timber GIR connections to ensure their reliability and integrity under prolonged exposure to fire conditions.

ABSTRACT. As the world moves toward more sustainable design, mass timber offers a promising alternative to conventional construction materials. With recent advancements in engineered wood manufacturing and the development of modern heavy-duty connection systems, the application of mass timber in mid- and high-rise buildings has expanded significantly in Canada and worldwide. However, fire safety remains a critical challenge, particularly for beam-to-column connections, which are often the most vulnerable components in a structural system in fire conditions. To fully realize the potential of mass timber construction, a transition away from conventional connections, which are vulnerable in fire conditions, to more advanced and practical connecting mechanisms is essential. Interlocking connections for mass timber structures have the potential to transform construction practices; however, their fire performance remains inadequately understood and is a key concern. In this study, the fire performance of modern concealed, interlocking mass-timber beam-to-column connections was evaluated through full-scale testing under standard fire exposure conditions, accounting for practical beam-column gap sizes (~3 mm) and load-ratio variations. The experimental results demonstrate that this connection type provides superior fire resistance compared to conventional connection systems, achieving 125 and 105 minutes of fire resistance at load ratios of 50% and 70%, respectively, without fastener or connector failure and without the application of any fire protection or sealant.

ABSTRACT. This study investigates the collapse behavior of timber structures under fire and its impact on fire dynamics using controlled experiments representative of external-fire ignition scenarios in wildland-urban interface areas. Key variables include external fire location, column base design, and opening geometry. Fires impinging on large openings directly ignited interior fuel loads, while fires on small openings primarily ignited wall assemblies before transitioning to interior ignition. Increased sill height reduced flame penetration into the interior. The column bases in this study did not significantly affect the overall results due to their limited height. Two collapse patterns were identified: footprint-confined and footprint-expanding, distinguished by whether structural debris remained within or extended beyond the original plan area. Within an individual structure, the collapse pattern was primarily governed by the temporal development of fire exposure associated with fire spread and growth, and the resulting nonuniform thermal degradation, which controlled residual strength imbalance and failure direction. The footprint-expanding collapse can deposit structural debris into separation spaces between structures, potentially accelerating fire spread. In all but one non-collapsing test, failure occurred after peak fire intensity, indicating that sustained internal heating progressively reduced load-bearing capacity until collapse. Structural collapse, in turn, influenced fire development. Collapse and full enclosure involvement were closely coupled as the fire transitioned from a compartment-controlled to an external phase, and in multi-compartment or closely spaced structures, such a loss of compartmentation would likely increase fire spread. In footprint-expanding collapse tests, a secondary HRR peak emerged shortly after collapse, as redistributed burning members increased oxygen availability and briefly intensified combustion. Moreover, mechanical fracture and fragmentation of burning wood during structural collapse increased ember production, while collapse-induced aerodynamic and thermal disturbances further enhanced ember generation and lofting.

ABSTRACT. Timber construction is increasingly adopted for low-carbon and green development, yet current structural fire design often relies on charring rates derived from standard fire exposure, which can be overly conservative for large-span timber structures. This study developed a performance-based structural fire design concept tailored to large-space fire conditions, reflecting their fundamentally lower fire severity and distinct thermal characteristics. Extensive experiments on glued laminated timber (GLT) under both ISO 834 and large-space fires were conducted to quantify ignition behavior, temperature evolution, and charring rates. Results showed that GLT exhibits significantly reduced charring under large-space fire scenarios. A simplified method for estimating charring rates in such environments was proposed. The design concept was further implemented in a 102 m-span timber-steel hybrid structure, where performance-based fire assessment achieved the required fire resistance without additional protective char layers. This approach provides a rational pathway to ensure fire safety while reducing unnecessary protection, lowering construction costs, and enhancing material efficiency.

ABSTRACT. This contribution presents an extended theoretical model describing the coupled mechanisms of wood charring and crack formation within the charred layer during fire exposure. Building on previous experimental and analytical studies, which demonstrated that surface cracks significantly accelerate heat transfer, alter pyrolysis progression, and reduce the residual load-bearing capacity of timber members, the new model unifies Baroudi’s mechanical crack-spacing concept with Hietaniemi’s probabilistic charring-rate formulation. The resulting analytical expression accounts for nonlinear charring behavior, temperature-dependent stiffness reduction, moisture content, density, and oxygen availability. It links crack initiation and spacing to the evolving hot-layer thickness, heat flux, mechanical half-wave formation, and the anisotropic elastic properties of wood. Validation against medium-scale furnace tests shows that the coupled analytical–probabilistic model captures the measured evolution of crack spacing more accurately than the original Baroudi model. The work also outlines future integration with AI-based infrared monitoring, which could enable real-time parameter calibration and deeper insight into transient crack processes such as coalescence or branching during fire exposure.