scispace - formally typeset
Search or ask a question

Showing papers in "Fire Technology in 2021"


Journal ArticleDOI
TL;DR: This work addresses the possibility of AI-based detection and prediction of fire source and hazard, thus, providing scientifically based guidance for smart-firefighting technologies and paving the way for future emergency-response tactics in a smart city.
Abstract: The fire event in a tunnel creates a rapid spread of heat and smoke flows in a long and confined space, which not only endangers human life but also challenges the fire-evacuation and firefighting strategies. A quick and accurate identification for the location and size of the original fire source is of great scientific and practical value in guiding fire rescue and fighting the tunnel fire. Nevertheless, it is a big challenge to acquire fire-source information in an actual tunnel fire event. In this study, the framework of artificial intelligence (AI) and big data is applied to predict the fire source in a numerical model of the tunnel. A big tunnel fire database of numerical simulations, with varying fire locations, fire sizes, and ventilation conditions, is constructed. Temporally varied temperatures measured by multiple sensor devices are used to train a long-short term memory recurrent neural network. Results demonstrate that the location and size of the tunnel fire and the ventilation wind speed can be predicted by the trained model with an accuracy of 90%. Sensitivity analysis is also carried out to optimize the database configuration and spatial–temporal arrangement of sensors in order to achieve a fast and reliable fire prediction. This work addresses the possibility of AI-based detection and prediction of fire source and hazard, thus, providing scientifically based guidance for smart-firefighting technologies and paving the way for future emergency-response tactics in a smart city.

59 citations


Journal ArticleDOI
TL;DR: A comprehensive survey of the machine learning algorithms based forest fires prediction and detection systems is presented, highlighting the main issues and outcomes within each study.
Abstract: Forest fires are one of the major environmental concerns, each year millions of hectares are destroyed over the world, causing economic and ecological damage as well as human lives. Thus, predicting such an environmental issue becomes a critical concern to mitigate this threat. Several technologies and new methods have been proposed to predict and detect forest fires. The trend is toward the integration of artificial intelligence to automate the prediction and detection of fire occurrence. This paper presents a comprehensive survey of the machine learning algorithms based forest fires prediction and detection systems. First, a brief introduction to the forest fire concern is given. Then, various methods and systems in forest fires prediction and detection systems are reviewed. Besides works that reported fire prediction and detection systems, studies that assessed the factors influencing the fire occurrence and risk are discussed. The main issues and outcomes within each study are presented and discussed.

52 citations


Journal ArticleDOI
M.Z. Naser1
TL;DR: In this article, a review highlights the merit of adopting mechanistically-informed MI to answer some of the burning questions, multi-dimensional and ill-defined problems fire engineers and scientists are facing.
Abstract: Fire is a chaotic and extreme phenomenon. While the past few years have witnessed the success of integrating machine intelligence (MI) to tackle equally complex problems in parallel fields, we continue to shy away from leveraging MI to study fire behavior or to evaluate fire performance of materials and structures. In order to advocate for the use of MI, this review showcases the merit of adopting mechanistically-informed MI to answer some of the burning questions, multi-dimensional and ill-defined problems fire engineers and scientists are facing. This review also sympathizes with the fact that a traditional curriculum does not often cover principles of MI and hence it starts by introducing a number of machine learning (ML) and artificial intelligence (AI) techniques such as deep learning, metaheuristics, decision trees, random forest, support vector machines etc. Then, this review details recommended procedures associated with preparing databases and carrying out a proper MI-tailored fire analysis via examples; to enable researchers and practitioners from implementing MI with ease. Towards the end of this review, a number of concerns and challenges are identified to stimulate the curiosity of interested readers and accelerate future research works within fire engineering and sciences (FES).

40 citations


Journal ArticleDOI
TL;DR: CNNs are shown to have a great potential for smoke and fire detection and better development can help prepare a robust system that would greatly save human lives and monetary wealth from getting destroyed from fires.
Abstract: The risk of fires is ever increasing along with the boom of urban buildings. The current methods of detecting fire with the use of smoke sensors with large areas, however poses an issue. The introduction of video surveillance systems has given a great opportunity for identifying smoke and flame from faraway locations and tackles this risk. Processing this huge amount of data is a problem with using these video and image data. In recent times, a number of methods have been proposed to deal with this challenge and identify fire and smoke. Image processing algorithms for detecting flame and smoke, motion-based estimation of smoke, etc are some of the methods that are proposed earlier. Recently, there has been an array of methods proposed using Deep Learning, Convolutional Neural Networks (CNNs) to automatically detect and predict flame and smoke in videos and images. In this paper, we present a complete survey and analysis of these machine vision based fire/smoke detection methods and their performance. Firstly, we introduce the fundamentals of image processing methods, CNNs and their application prospect in video smoke and fire detection. Next, the existing datasets and summary of the recent methodologies used in this field are discussed. Finally, the challenges and suggested improvements to further the development of the application of CNNs in this field are discussed. CNNs are shown to have a great potential for smoke and fire detection and better development can help prepare a robust system that would greatly save human lives and monetary wealth from getting destroyed from fires. Finally, research guidelines are presented to fellow researchers regarding data augmentation, fire and smoke detection models which need to be investigated in the future to make progress in this crucial area of research.

38 citations


Journal ArticleDOI
TL;DR: In this paper, the authors analyzed the research done in the last decades for improving the thermal behaviour and fire resistance of green materials as well as the efficient synergistic techniques and fire retardant treatments to counter this vulnerability.
Abstract: Stringent fire safety regulations have limited the use of green biocomposites in practical applications due to vulnerability of their constituents to heat and fire. To counter this weakness, several flame-retardant treatments and techniques have been introduced, such as halogenated and non-halogenated flame-retardants, nano fillers, layered silicates, copolymerization, grafting, and synergistic use of natural fibre and fire retardant. While the physical and chemical treatment of green biocomposites has improved their heat resistance to some extent, these materials still fail to comply with strict fire safety regulations such as the Federal Motor Vehicle Safety Standard No. 302 (FMVSS 302) and the code of Federal Aviation Regulation 25.853 applicable in the automotive and aerospace industry, respectively. Therefore, an in-depth study of thermal decomposition and fire behaviour of green biocomposites is inevitable to improve flame retardancy techniques, to discover flame-retardants that are more suitable and environment friendly, and to select appropriate natural fibres and biopolymers to develop fire safe biocomposite products. This article analyses the research done in the last decades for improving the thermal behaviour and fire resistance of green materials as well as the efficient synergistic techniques and fire retardant treatments to counter this vulnerability.

25 citations


Journal ArticleDOI
TL;DR: Comparison between the failure probabilities of steel and concrete columns subject to fire, considering the proposed consolidated model and two currently commonly used models, indicates that relative differences of the probability of failure can be in the order of 10%.
Abstract: Probabilistic analysis is receiving increased attention from fire engineers, assessment bodies and researchers. It is however often unclear which probabilistic models are appropriate for the analysis. For example, in probabilistic structural fire engineering, the models used to describe the permanent and live loads differ widely between studies. Through a literature review, it is observed that these diverging load models are based on surveys conducted between 1893 and 1976 and that widely adopted assumptions, such as the rule for combining permanent and live loads into the total load effect, are commonly adopted based on precedent. The diverging current models however relate to mostly the same underlying datasets and basic methodologies. Differences can be attributed (largely) to specific assumptions in different background papers, which have become consolidated through repeated use in research papers and adoption in background documents to codes. By reviewing the studies underlying currently applied probabilistic load models in structural fire engineering, a consolidated probabilistic load model is proposed in this paper. It is concluded that the total load effect is ideally described by KE·(G + Q), with KE the model uncertainty for the load effect, G the permanent load, and Q the imposed load. The model uncertainty KE can be described by a lognormal distribution with mean equal to unity and coefficient of variation (COV) of 0.10. The permanent load is preferably modelled by a normal distribution with mean equal to the nominal permanent load, and a COV which can either be assessed on a project specific basis, or can be set to 0.10 for a first assessment. For common occupancies (office, residential), the live load is preferably modelled by a Gamma distribution. The mean live load can be taken as 0.2 times the nominal, and the live load COV can be taken as 0.60 for large load areas (> 200 m2) and 0.95 for smaller load areas (< 100 m2). Comparison between the failure probabilities of steel and concrete columns subject to fire, considering the proposed consolidated model and two currently commonly used models, indicates that relative differences of the probability of failure can be in the order of 10%. Live load models for evacuation routes and warehouses require specific study and are outside the scope of the review.

22 citations


Journal ArticleDOI
TL;DR: In this paper, the authors show that the BS 8414 test is empirical and the criteria arbitrary: there is no scientific link between test performance and how a building will perform in the event of a fire; nor any detailed analysis of why fires spread through facade systems which have passed the test.
Abstract: The Grenfell Tower fire has brought the regulatory system that permitted combustible materials on high-rise buildings in England into question. At the heart of that system is the BS 8414 test, and the BR 135 criteria used to demonstrate compliance with the Building Regulations. The test is empirical and the criteria arbitrary: there is no scientific link between test performance and how a building will perform in the event of a fire; nor any detailed analysis of why fires spread through facade systems which have passed the test. Following the Grenfell tragedy, the UK government commissioned a series of tests on Grenfell Tower-type facades, using BS 8414. This paper critically analyses BS 8414, the BR 135 criteria and the government tests. It shows that important aspects of the standard are poorly defined: the heat flux imposed on the facade is not measured and the fire load can vary by at least a factor of 2; the ambient ventilation has a significant impact on the thermal attack but is not adequately controlled; judicious location of the cavity barriers can confer compliance or failure on a facade system. As the vehicle for allowing combustible products on tall buildings, the test does not specify the extent of cavity barrier deployment, while ignoring features present in real buildings, such as windows, vents or other openings, despite a test rig height of more than 8 m. There is no restriction on debris, or molten or burning droplets falling from the facade during the test. The BR 135 criteria only specify that the test must run for the full 60 min duration without flames reaching the top, and the temperature rise at thermocouples 5 m above the fire chamber must only remain below 600°C for the first 15 min. It is unclear how the fire safety of the occupants behind the facade system can be ensured, when the criteria specify such a high temperature for such a short period, so early in the test. There is no direct connection between the facade system in the test and the actual facade system the results deem compliant. Worse, “desktop studies”, using large-scale test data, have been allowed to confer compliance on systems which have never been subject to the test. The UK government tests used heavy-duty welded aluminium “window pods”, preventing flames from entering the cavity within the facade. They also used a disproportionately large number of vertical and horizontal cavity barriers of a higher specification than required by statutory guidance. These aids to meeting the criteria are not proscribed by BS 8414-1 but are not commonly found in actual rainscreen system designs.

21 citations


Journal ArticleDOI
TL;DR: In this paper, the authors used a multiscale charring model at the microscale (mg-samples), mesoscale (g-sample), and macroscale (kg-sample) for several wood species exposed to different heating regimes and boundary conditions.
Abstract: Engineered timber is an innovative and sustainable construction material, but its uptake has been hindered by concerns about its performance in fire Current building regulations measure the fire performance of timber using fire resistance tests In these tests, the charring rate is measured under a series of heat exposures (design fires) and from this the structural performance is deduced Charring rates are currently only properly understood for the heat exposure of a standard fire, not for other exposures, which restricts the use of performance-based design This paper studies the charring rates under a range of design fires We used a multiscale charring model at the microscale (mg-samples), mesoscale (g-samples), and macroscale (kg-samples) for several wood species exposed to different heating regimes and boundary conditions At the macroscale, the model blindly predicts in-depth temperatures and char depths during standard and parametric fires with an error between 5% and 22% Comparing simulations of charring under travelling fires, parametric fires, and the standard fire revealed two findings Firstly, their charring rates significantly differ, with maximum char depths of 42 mm (travelling), 46 mm (parametric), and 59 mm (standard fire), and one (standard fire) to four (travelling fire) charring stages (no charring, slow growth, fast growth, steady-state) Secondly, we observed zero-strength layers (depth between the 200 °C and 300 °C isotherm) of 7 to 12 mm from the exposed surface in travelling fires compared to 5 to 11 mm in parametric fires, and 7 mm in the standard fire Both traditional design fires and travelling fires, therefore, need to be considered in structural calculations These results help engineers to move towards performance-based design by allowing the calculation of charring rates for a wide range of design fires In turn, this will help engineers to build more sustainable and safe structures with timber

18 citations


Journal ArticleDOI
TL;DR: This work has proposed a novel fire detection method based on Red Green Blue and CIE $$L * a * b$$ color models, by combining motion detection with tracking fire objects, which operates on a reduced number of parameters compared to the existing methods.
Abstract: Emergency incidents and events of fires can be dangerous and required quick and accurate decision-making need quick and correct decision-making. The use of computer vision for fire detection can provide a efficient solution to deal with these situations. These systems handle the usual data, provide an automated solution, and discard non-relevant information without discarding relevant content. Researchers developed many techniques for fire detection in videos and still images by using different color-based models. However, for videos, these methods are unsuitable because of high false-positive results. These methods use few parameters with little physical meaning, which makes fire detection more difficult. To deal with this, we have proposed a novel fire detection method based on Red Green Blue and CIE $$L * a * b$$ color models, by combining motion detection with tracking fire objects. We have eliminated the moving region and calculate the growth rate of the fire to reduce false-alarm and calculate the risk. The proposed method operates on a reduced number of parameters compared to the existing methods. Experimental results demonstrate the effectiveness of our method of reducing false positives while keeping their precision compatible with the existing methods.

17 citations


Journal ArticleDOI
TL;DR: Results suggest that people typically choose the elevator for evacuation, even if their hotel room was located closer to the evacuation stair, and that a detector activated self-closing fire door without vision panels to the elevator lobby made it more difficult to find the evacuation elevators in an emergency.
Abstract: Past studies suggest that people are often reluctant to use occupant evacuation elevators in case of fire. However, existing research is scarce and current knowledge is based on questionnaire studies and laboratory experiments. An unannounced evacuation experiment was therefore performed on the 16th floor of a 35-floor high-rise hotel building. Sixty-seven participants took part and eye-tracking glasses were used to collect data on exit choice and eye fixations. Three different scenarios were studied, including two different hotel room locations on the floor and a variation of guidance system for one of these locations, i.e., flashing green lights next to the evacuation sign at the elevators. Results suggest that people typically choose the elevator for evacuation, even if their hotel room was located closer to the evacuation stair. Flashing green lights next to an evacuation sign made people look more at this sign. However, in spite of looking more at the sign, the flashing light was not shown to significantly improve compliance with the sign. Also, the results suggest that a detector activated self-closing fire door without vision panels to the elevator lobby made it more difficult to find the evacuation elevators in an emergency.

17 citations


Journal ArticleDOI
TL;DR: A feature-squeeze block is proposed that squeezes the feature maps spatially and channel-wise to effectively use the information from the multi-scale prediction of fire detection tasks.
Abstract: The automation of fire detection systems can reduce the loss of life and property by allowing a fast and accurate response to fire accidents. Although visual techniques have some advantages over sensor-based methods, conventional image processing-based methods frequently cause false alarms. Recent studies on convolutional neural networks have overcome these limitations and exhibited an outstanding performance in fire detection tasks. Nevertheless, previous studies have only used single-scale feature maps for fire image classification, which are insufficiently robust to fires of various sizes in the images. To address this issue, we propose a multi-scale prediction framework that exploits the feature maps of all the scales obtained by the deeply stacked convolutional layers. To utilize the feature maps of various scales in the final prediction, this paper proposes a feature-squeeze block. The feature-squeeze block squeezes the feature maps spatially and channel-wise to effectively use the information from the multi-scale prediction. Extensive evaluations demonstrate that the proposed method outperforms the state-of-the-art convolutional neural networks-based methods. As a result of the experiment, the proposed method shows 97.89% for F1-score and 0.0227 for false positive rate in the average of evaluations for multiple.

Journal ArticleDOI
TL;DR: The prototyping steps required to develop a non-immersive VR serious game (SG) to train the staff of Vincent Van Gogh (VVG) hospital in Belgium are described and the paper finally validates the VR SG comparing its effectiveness against slide-based lecture training.
Abstract: In a healthcare context, the success of a fire safety procedure in a real-life emergency mainly depends on staff decisions and actions. One of the factors influencing staff decision-making is their training. In most healthcare facilities, safety educators use slide-based lectures as a training tool. Virtual Reality (VR) is gaining fire safety community attention for being an interesting training tool. However, few studies have assessed the effectiveness of VR-based fire safety training simulators compared with a slide-based lecture. The present research proposes a novel non-immersive VR-based training for healthcare fire safety education. This paper describes the prototyping steps required to develop a non-immersive VR serious game (SG) to train the staff of Vincent Van Gogh (VVG) hospital in Belgium. The paper finally validates the VR SG comparing its effectiveness against slide-based lecture training. 78 staff from VVG hospital in Belgium participated in this study. They were divided into two groups: Group A was trained using a slide-based lecture, and Group B was trained using the VR SG. The results indicated that the VR SG was more effective than the slide-based lecture in terms of knowledge acquisition and retention and in terms of self-efficacy increment in short and long terms than the slide-based lecture.

Journal ArticleDOI
TL;DR: In this article, the authors provide additional context to the historical narrative of the development of the standard temperature-time heating curve used for the determination of the fire resistance of structural elements.
Abstract: This review aims to provide additional context to the historical narrative of the development of the standard temperature–time heating curve used for the determination of the fire resistance of structural elements. While historical narratives of the development of the standard temperature–time heating curve exist, there are portions of the timeline with missing contributions and contributions deserving of additional examination. Herein, additional newly available contributions (owing to recent digitization efforts) from the original standard development cycle not distinctly covered by existing historical narratives are introduced and reviewed. Though some engineers have long been recognized for their contributions to the curve’s development, lesser-recognized influences are re-examined. These include contributions to fire resistance testing from Sylvanus Reed, that are acknowledged for the first time in a contemporary light. Practitioners will find discussion from the temperature–time heating curve’s development period that is useful for current philosophical discussions pertaining to the curve’s use for combustible material testing. This study identifies that no currently available historical literature can support the definition of the temperature points which describe the standard temperature–time heating curve. This reinforces contemporary discussion that the heating curve lacks scientific basis in its representation of a real fire.


Journal ArticleDOI
TL;DR: In this article, a series of large-scale experiments on cross-laminated timber (CLT) slabs that were exposed to fire from below, using four different heating scenarios, with a sustained mechanical loading of 6.3 kNm per metre width of slab.
Abstract: This paper describes selected observations, measurements, and analysis from a series of large-scale experiments on cross-laminated timber (CLT) slabs that were exposed to fire from below, using four different heating scenarios, with a sustained mechanical loading of 6.3 kN m per metre width of slab. The deflection response and in-depth timber temperatures are used to compare the experimental response against a relatively simple structural fire model to assess the load bearing capacity of CLT elements in fire, including during the decay phase of natural fires. It is demonstrated that the ventilation conditions in experiments with a fixed fuel load are important in achieving burnout of the contents before structural collapse occurs. A mechanics-based structural fire model is shown to provide reasonably accurate predictions of structural failure (or lack thereof) for the experiments presented herein. The results confirm the importance of the ventilation conditions on the fire dynamics, burning duration, and the achievement of functional fire safety objectives (i.e. maintaining stability and compartmentation), in compartments with exposed CLT.

Journal ArticleDOI
TL;DR: In this paper, a firebrands were represented by a dry wood ball with a diameter of 20mm and a weight of 2.9 grams, which carried a flame with the heat release rate of 250 W. The firebrand was held by a pendulum system to adjust the velocity.
Abstract: Firebrands are a widely observed phenomenon in wildland fires, which can transport for a long distance, cause spot ignition in the wildland–urban interface (WUI) and increase the rate of wildfire spread. The flame attached to a moving firebrand behaves as a potential pilot source for ignition, so extinguishing such a flame in the process of moving can effectively minimize its fire hazard. In this work, firebrands were represented by a dry wood ball with a diameter of 20 mm and a weight of 2.9 g, which carried a flame with the heat release rate of 250 W. The firebrand was held by a pendulum system to adjust the velocity. Results showed that there is a minimum sound pressure to extinguish the firebrand flame, which increases slightly with the sound frequency. As the firebrand velocity increases from 0 m/s to 4.2 m/s, the minimum sound pressure for extinction decreases significantly from 114 dB to 90 dB. The cumulative effect of firebrand motion and acoustic oscillation was found to facilitate flame extinction. A characteristic Damkohler number (~ 1), with the ratio of the fuel residence time to the flame chemical time, is used to quantify the extinction limit of the flaming firebrand. This work provides a potential technical solution to mitigate the hazard of firebrand flame and spotting ignition in WUI and helps understand the influence of acoustic waves on the flame stability on the solid fuel.

Journal ArticleDOI
TL;DR: In this paper, the residual stress-strain behavior of lightweight concrete (LWC) containing pumice coarse aggregate and rock wool waste (consisting of mineral fibers) were explored prior to and following thermal loading.
Abstract: In the present study, the mechanical properties and the residual stress–strain behavior of lightweight concrete (LWC) containing pumice coarse aggregate and rock wool waste (consisting of mineral fibers) were explored prior to and following thermal loading. Key variables included the volume percentage of rock wool waste (0%, 2.5%, 5%, 7.5%, and 10%) and exposure temperature (20°C, 200°C, 400°C, and 600°C). Here, parameters playing a role in the compressive performance of LWC containing rock wool waste were examined. These parameters included the elastic modulus, compressive strength, strain at peak stress, ultimate strain, toughness index, stress–strain relationship, and failure mode. Then, several empirical relationships were proposed to predict different mechanical characteristics in terms of temperature and volume percentage of rock wool. Furthermore, the compressive strength, elastic modulus, and strain at peak stress values were compared to the prediction results of the ACI 216, EN 1994-1-2, ASCE, and CEB-FIP 1990 codes. The results demonstrated that the mechanical properties of the LWC specimens were degraded with temperature. The highest degradation in the temperature range under study occurred at 600°C, with around 50% and 80% drop in the compressive strength and elastic modulus, respectively. Furthermore, after exposure to 600°C, an increase of 2 to 2.8 folds occurred in the strain at peak stress and an increase of 2.6 to 3.4 folds occurred in the ultimate strain of the specimens relative to those at the ambient temperature. In the end, two stress–strain models were presented to capture the compressive performance of LWC including rock wool waste after elevated temperature exposure based on the empirical equations obtained for the mechanical characteristics. These models showed a relatively good correlation with the experimental data.

Journal ArticleDOI
TL;DR: An improved long short-term memory neural network model with multiply input layers and attention mechanism to increase its performance, named as Multi-AM-LSTM is proposed, which is capable to predict the wildland fire burned areas at a high accuracy in a reasonable computational complexity.
Abstract: Wildland fire is a major natural disaster that causes the environmental hazards and severe negative impacts on the lives. Therefore, early prediction is the key to control such phenomenon and avoid the economic and ecological damage. This paper proposes an improved long short-term memory (LSTM) neural network model with multiply input layers and attention mechanism to increase its performance, which is named as Multi-AM-LSTM. The proposed model is used to predict the wildland fire burned areas, as well as to carry out an analysis of the multilateral interactive relationship among the related variables. The Montesinho Natural Park dataset is introduced to provide the wildland fire and meteorological data between Jan. 2000 and Dec. 2003. And then, the correlation analysis performs to figure out the related variables with the wildland fire and the features of the variables can be extracted by using the convolutional neural network. Next, the improved LSTM model is established to predict the wildland fire and further the attention mechanism is injected into the LSTM model, which can help reduce the loss of historical information and strengthen the impact of important information. Quantitative results show that the designed model outperforms the recent outstanding methods with the highest accuracy over 96%. The results indicate that the model is capable to predict the wildland fire burned areas at a high accuracy in a reasonable computational complexity.

Journal ArticleDOI
TL;DR: The paper presents the challenges and prerequisites to build such an expert system and acts as a proof of concept for the use of machine learning for the assessment of the insulation performance of shallow floor system under the guidance of BS EN 1363-1:2012.
Abstract: This paper is proposing an machine learning based expert system for preliminary prediction of the insulation fire resisting performance of shallow floor systems when subject to exposure to the ISO 834 Standard Fire Curve. The proposed system is a digital tool which incorporates a machine learning algorithm trained on the outcomes of pre-run two-dimensional finite element heat transfer analyses of shallow floor system details. The algorithm predicts the insulation performance of similar details with a measurable accuracy of 96% (i.e the insulation rating band was predicted correctly in 96% of the cases) without the requirement for an explicit deterministic analysis. The paper presents the challenges and prerequisites to build such an expert system and acts as a proof of concept for the use of machine learning for the assessment of the insulation performance of shallow floor system under the guidance of BS EN 1363-1:2012. A Support Vector Machine machine learning algorithm is adopted in this work. The required processes that were needed for the development of the expert system include the stages of data acquisition, exploratory data analysis, choice of machine learning algorithm, model training, tuning, and validation. This expert system is useful for practitioners to rapidly assess the feasibility of different construction details at early stages of the design process.

Journal ArticleDOI
TL;DR: In this article, the authors focus on quantifying uncertainties in the temperature-time evolution of railway tunnel fires considering fire spread between train cars, and propose a framework to establish guidelines for temperature demands in the design of concrete tunnel linings within risk-based frameworks.
Abstract: Extreme fire events in tunnels can have catastrophic consequences, including loss of lives, structural damage, and major socioeconomic impacts. The fire scenario itself is one of the primary parameters that would influence the level of damage in a tunnel. Standard hydrocarbon fire temperature–time curves exist but they are idealized and do not consider the actual fire duration and potential for fire spread within the tunnel. Furthermore, risk-based decision-making frameworks and performance-based design of tunnel linings require realistic sets of fire scenarios to quantify damage. This paper focuses on quantifying uncertainties in the temperature–time evolution of railway tunnel fires considering fire spread between train cars. In this study, 540 numerical simulations are conducted in fire dynamics simulator by varying ventilation velocity, amount of fuel, tunnel slope, ignition point, and criteria for fire spread between railcars. Temporal and spatial distribution of fire temperature in the tunnel is studied. The resulting 540 temperature–time curves at sections with the highest temperature are analyzed and statistics of the maximum fire temperature and duration, heating rate, and decay rate are provided. Fires with heat release rates larger than 40 MW are categorized as high-intensity with mean maximum temperature of 1007°C. Fires with heat release rates smaller than 30 MW are categorized as low-intensity with a mean maximum temperature of 245°C. The proposed framework can be expanded in future to establish guidelines for temperature demands in the design of concrete tunnel linings within risk-based frameworks to achieve required performance levels in railway tunnel fire events.

Journal ArticleDOI
TL;DR: In this article, the effects of removing fluorocarbon surfactants from AFFF leads to an apparent decrease in film-forming ability, foam stability, foam spreading property, and corresponding fire extinguishing performance.
Abstract: The application of conventional aqueous film-forming foam (AFFF) has been restricted due to the environmental hazards caused by long-chain fluorocarbon surfactants. Environmentally friendly firefighting foams are urgently needed. In this study, AFFFs based on a long-chain fluorocarbon surfactant and a short-chain fluorocarbon surfactant, and fluorine-free foams based on a silicone surfactant and a mixture of foam stabilizers are prepared. The critical properties, including film-forming ability, foam stability, and foam spreading property, of foams and a commercial AFFF are investigated systematically. The fire extinguishing and burn-back performances are evaluated by a small-scale test method. Results indicate that the removal of fluorocarbon surfactants from AFFF leads to an apparent decrease in film-forming ability, foam stability, foam spreading property, and corresponding fire extinguishing performance. AFFF based on short-chain fluorocarbon surfactant shows excellent extinguishing and burn-back performances even if its film-forming ability, foam stability, and foam spreading are slightly worse than those of conventional AFFF. Although FfreeF cannot form an aqueous film and exhibits poor foam spreading, it demonstrates satisfactory fire extinguishing performance and optimal burn-back performance depending on its superhigh stability. This study can provide guidance in the development of environmentally friendly firefighting foams.

Journal ArticleDOI
TL;DR: In this paper, a Gaussian process framework is used to develop a generative model that produces random viable HRR ramps for a transient heat release rate (HRR) input.
Abstract: This work describes a deep learning methodology for “emulating” temperature outputs produced by the Fire Dynamics Simulator (FDS), a CFD software. An array of artificial neural networks (ANNs) is trained to predict transient temperatures at specified locations for a transient heat release rate (HRR) input. These locations correspond to the locations of thermocouples used in an experimental burn structure. In order to build the training set, A Gaussian process (GP) framework is used to develop a generative model that produces random viable HRR ramps. Although this procedure may require thousands of FDS runs to build a sufficient training set, the application of transfer learning can reduce the required number of runs by nearly an order of magnitude. This refers to the process of initially training an ANN to predict the output of the Consolidated Model of Fire and Smoke Transport (CFAST) and then transferring its knowledge to an ANN that learns to predict FDS outputs. CFAST is a much faster model than FDS, so a large training set can be generated quickly. The final state of the ANN trained to emulate CFAST is used as the initial state of an ANN that learns to emulate FDS. The result is a model that produces FDS temperature predictions with a mean absolute error (MAE) of less than 2°C and runs over five orders of magnitude faster than FDS. The emulators are also capable of learning inverse mappings; i.e. for a given temperature output, they can predict the HRR ramp that would cause FDS to produce the temperature response. This ability to invert for the HRR profile is exercised on data collected from eight fire experiments with peak HRRs up to 200 kW, including four propane burner fires, two methanol pool fires, and two n-Hexane pool fires. The model inverts for the experimental HRR with a MAE of 5.8 kW-15.4 kW (11.3%–16.7%) for the burner tests and 5.0 kW–25.5 kW (12.1%–28.6%) for the pool fire tests, with a tendency to underestimate the HRR of the pool fires. Finally, the computational speed of the emulators allows for the incorporation of CFD physics in Bayesian parameter inversion. As an example, this is demonstrated to infer the radiative fraction from experimental and synthetic data in conjunction with reported uncertainties from the FDS Validation Guide.

Journal ArticleDOI
TL;DR: In this paper, the effect of separation distance, dwelling height and dwelling length on the times-to-ignition of informal settlement fire spread is investigated, where it is shown that the heat flux received by an adjacent dwelling decreases approximately exponentially as the distance between dwellings increases.
Abstract: Globally, the number of informal settlement dwellings are increasing rapidly; these areas are often associated with numerous large fires. Unfortunately, until recently, very little research has been focused on informal settlement fire issues leaving any attempts to improve their fire safety lacking the evidence base to support effective-decision making. However, over the past 4 years, a limited number of researchers have looked at better understanding these fires through full-scale experimentation and numerical modelling; starting to provide the necessary evidence base and future research directions. It is with this background in mind that this paper seeks to provide a more fundamental understanding of the effect of dwelling separation distance on informal settlement fire spread based on full-scale experiments and analytical equations. In this paper two full-scale experiments were conducted. Both experiments consisted of multiple dwellings, with the main difference between the experiments being the separation distance. Fire spread times, heat release rates, door and window flow velocities, ceiling temperatures and incident heat fluxes were recorded and are reported for both experiments. Theoretical neutral planes are derived and compared to the experimental neutral planes, which show relatively good correlation. The paper continues by calculating the expected incident radiation and time-to-ignition, using the flux-time product method, of the two fire scenarios (i.e., the two experiments) through means of analytical equations, and these findings are compared to the experimental results. Through configuration factors, the paper shows the effect of separation distance, dwelling height and dwelling length on the times-to-ignition, where it is clear that the heat flux received by an adjacent dwelling decrease approximately exponentially as the distance between dwellings increases, and consequently, the time-to-ignition increases exponentially as the separation distance between dwellings increases.

Journal ArticleDOI
TL;DR: In this paper, the outer shell and thermal liner tear strength of turnout gear was measured after 40 laundering cycles, and the results indicated that some important protective properties of bunker gear can be decreased after repeated exposure/cleaning cycles relative to their levels when tested in a new condition.
Abstract: The US fire service has become acutely aware of the need to clean PPE after fires. However, there is concern that damage from repeated cleaning may impact critical protection from fireground risk. Using a protocol that included repeated simulated fireground exposures (between 0 cycles and 40 cycles) and/or repeated cleaning with techniques common in the fire service, we found that several important protective properties of NFPA 1971 compliant turnout gear are significantly changed. Outer shell and thermal liner tear strength showed a statistically significant reduction when laundered as compared to wet or dry decontamination. Larger changes in outer shell tear strength resulted when the coat closure incorporated hook & dee clasps as compared with garments using zippered closures. Total Heat Loss was reduced for all samples that underwent any form of cleaning while Thermal Protective Performance was only increased in the gear that was laundered. These results suggest that some important protective properties of bunker gear can be decreased after repeated exposure/cleaning cycles relative to their levels when tested in a new condition. For the specific materials tested, outer shell trap tear strength in the fill direction and seam strength dropped below NFPA 1971 requirements after 40 laundering cycles. The findings for this study may have utility for setting preconditions for the measurement of certain performance properties in future editions of NFPA 1971.

Journal ArticleDOI
TL;DR: The main research gaps identified include the lack of studies concerning the impact of cognitive limitations on egress, and the need to add the temporal dimension to the methods adopted in accessibility research to allow for their use in the egress field.
Abstract: This scoping review addresses the role of functional limitations on evacuation performance of adults in public buildings. Although this topic has been addressed in evacuation research, no linkage is currently available between functional limitations, the predominant activities affected by them and evacuation performance. This review strives to open a debate on the need to classify the impact of disability in terms of functional limitations on evacuation performance according to methods adopted in health science. This paper reviews literature concerning evacuation from public buildings with adults aged ≥ 60 years and/or adults aged ≥ 18 years with functional limitations. The International Classification of Functioning, Disability and Health has been used to identify predominant activities during an evacuation and to perform a structured classification at different levels of resolution to address self-evacuation possibilities. Results of the review are presented in a tabular form linking predominant activities in terms of the International Classification of Functioning, Disability and Health and six categories of functional limitations with the engineering evacuation time-line. The suggested classification can facilitate the assessment of the evacuation-related issues in buildings in relation to the population under consideration. The main research gaps identified include the lack of studies concerning the impact of cognitive limitations on egress, and the need to add the temporal dimension to the methods adopted in accessibility research to allow for their use in the egress field.

Journal ArticleDOI
TL;DR: Condensed aerosol based fire extinguishing systems (CAFES) have found versatile fire safety applications worldwide in many areas such as power generation, transportation, storage rooms, public buildings, heavy industries, battery storage systems, defence etc as discussed by the authors.
Abstract: Implementation of Montreal Protocols-1987 enforced phase wise ban on production and application of ozone depleting chemicals (Halons). Since then, condensed aerosol-based fire extinguishing technology as an alternative to Halons has been a subject for numerous investigations for its research and applications worldwide. It has come up as the most efficient halon alternative technology in comparison to other alternatives such as water mist, dry powders, inert gases, hydro fluorocarbons, and carbon dioxide. Even it is three times more efficient than that of Halon 1301 on weight to volume basis. Over other Halon alternatives, it has many advantages, e.g., zero ozone depletion potential -atmospheric lifetime -global warming potential, modular structure, low space requirements, easy & cost-effective installation & maintenance, no requirement of piping & pressurized cylinders, no oxygen depression etc. Condensed aerosol based fire extinguishing systems (CAFES) have found versatile fire safety applications worldwide in many areas such as power generation, transportation, storage rooms, public buildings, heavy industries, battery storage systems, defence etc. This review mainly focuses on applications of CAFES for extinguishment of: class A fires in libraries, archives, storage rooms etc.; class B fires in machinery spaces, gas turbine enclosures, combat vehicles, chemical storage rooms; electrical fires occurring in control rooms, UPS rooms, electrical/power substations & panels. Various reports and case studies demonstrating its testing and implementation methodologies have also been covered. The future prospects of aerosol technology and potential research/applications areas are also discussed.

Journal ArticleDOI
TL;DR: This paper presents a novel fire detection algorithm based on motion analysis using fractal and spatio-temporal features that outperforms the relevant state-of-the-art algorithms.
Abstract: Fire detection is one of the most important needs of surveillance and security systems in industrial applications. In this paper, a novel fire detection algorithm based on motion analysis using fractal and spatio-temporal features is presented. Initially, in each frame, dynamic textures are detected through three different fractal analysis methods and thresholding techniques. In the first method, Kernel Principal Component Analysis technique is used with fractal analysis and in the next a temporal blanket method is proposed. Finally, the third method is introduced based on temporal local fractal analysis and Laplace method. An RGB probability model is provided to separate the moving regions that have similar colors to the fire regions in each frame. Then, several spatio-temporal features such as correlation coefficient and mutual information are extracted from the candidate regions. Lastly, a two-class SVM classifier is used to classify these candidate regions. Various experimental results show that our proposed algorithm outperforms the relevant state-of-the-art algorithms.

Journal ArticleDOI
TL;DR: Comparative analysis shows that the proposed model outperforms previous fire segmentation deep learning models and image processing algorithms.
Abstract: In this paper, we proposed a semantic fire image segmentation method using a convolutional neural network. The simple but powerful method proposed is middle skip connection achieved through the residual network, which is widely used in image-based deep learning. To enhance the middle skip connection, we constructed a pair of convolution layers, hereafter referred to as input convolution and output convolution, to be inserted in front and behind of the entire architecture. Consequently, the middle skip connection yields a stronger feedback effect compared to when only the short skip connection of the residual block and the long skip connection are used. The validity of the proposed method has been confirmed by using the FiSmo dataset and the Corsican Fire Database based on various evaluation metrics. Comparative analysis shows that the proposed model outperforms previous fire segmentation deep learning models and image processing algorithms.

Journal ArticleDOI
TL;DR: In this article, a traveling fire concept and travelling fire models were proposed to provide an engineering description to this type of natural fire behaviour. But, the adequacy of homogenous temperature distribution in fully developed fires was questioned by researchers after reviewing the existing fire test data, which suggested a localised burning nature of the fires in relatively large compartments.
Abstract: Understanding the fire behaviour in buildings is fundamental and crucial to the practice of structural fire safety design. Traditionally, time-temperature curves associated with a burning rate developed from the “compartment fire framework” are most widely used by structural engineers and applied as a load to the structure. However, the adequacy of homogenous temperature distribution in fully developed fires was questioned by researchers after reviewing the existing fire test data, which suggested a localised burning nature of the fires in relatively large compartments. A groundbreaking travelling fire concept and travelling fire models were then proposed intending to provide an engineering description to this type of natural fire behaviour. The work in this paper was driven by such a trend and first summarises the modelling infrastructure in OpenSees to estimate the thermal response of structural members subjected to various scenario fires, followed by providing a smart application interface to capture the appropriate form of natural fire model through Python-OpenSees framework. The developed modelling infrastructure is validated against uniform and localised fire tests, which are also discussed regarding the smoke effect afterwards. Using the Python-OpenSees infrastructure, a real-scale localised fire test and the Malveira travelling fire test are modelled to demonstrate the modelling strategy. The work as preliminary attempts has shown the necessity of introducing additional variables when describing the natural fire impact, and this framework can be further improved in future by including more fire dynamics research and full-scale fire test input.

Journal ArticleDOI
TL;DR: A new approach for fire detection based on machine learning and optimization techniques, to monitor various types of fire by utilizing information obtained via multichannel fire sensor signals is proposed and achieves superior performance in terms of both fire detection time and false alarm rate.
Abstract: In recent years, the development of fire detectors has attracted the attention of researchers for the purpose of protecting human lives and properties from catastrophic fire disasters. However, monitoring fires is challenging due to several unique characteristics of fire sensor signals, such as the existence of temporal dependency and diverse signal patterns for different fire types, including flaming, heating, and smoldering fires. In this study, we propose a new approach for fire detection based on machine learning and optimization techniques, to monitor various types of fire by utilizing information obtained via multichannel fire sensor signals. The contribution of this study is to improve an existing fire detector by developing a new fire monitoring framework to identify fire based on support vector machine with dynamic time warping kernel function (SVM-DTWK), which considers the temporal dynamics existing in the sensor signals of different fire types. In addition, multichannel sensor signals are further considered by the SVM-DTWK with a multi-modeling framework that constructs multiple classifiers for each sensor type and effectively utilizes sensor information that is critical for the detection of fires without prior knowledge of the fire type. Using real-life fire data, the proposed approach is compared with existing fire monitoring methods and achieves superior performance in terms of both fire detection time and false alarm rate.