Headnote
ABSTRACT
Objective: The objective of this study is to investigate the effectiveness of constructive barriers (compartmentalization) in containing the spread of fire between semi-detached single-story houses.
Theoretical Framework: The research focuses on the importance of fire safety in single-story semi-detached houses, highlighting the vulnerability of these buildings and their occupants to fires. The study covers concepts of computer simulation and fire behavior in buildings.
Method: The methodology adopted for this research involves computer simulations using the Fire Dynamics Simulator (FDS) software to analyze different fire scenarios. Data collection was performed using outputs generated by the FDS and Smokeview (SMV).
Results and Discussion: The results obtained revealed that: in the scenario where there is no compartmentalization between the semi-detached dwellings, the spread of fire occurs through the gap between the ceiling and the roof, revealing that this is the point of greatest vulnerability in the building; and, in the scenarios where there is some type of compartmentalization, the spread of fire is significantly delayed. Simulations with the FDS enabled predictions of fire behavior, temperatures, and Heat Release Rate (HRR) in the scenarios analyzed.
Research Implications: The practical and theoretical implications of this research are discussed, providing insights into how the results can be applied or influence practices in the field of design development based on the performance of buildings in fire situations. These implications may include the development of safer designs and the promotion of fire safety guidelines for semi-detached residential buildings in Brazil.
Originality/Value: This study contributes to the literature by advancing fire safety studies based on building performance. The relevance and value of this research are evidenced by the use of the FDS in conjunction with Brazilian fire safety standards.
Keywords: Fire Safety, Semi-Detached Housing, Compartmentalization, Fire Dynamics Simulator (FDS), Performance-Based Design.
RESUMO
Objetivo: O objetivo deste estudo é investigar a eficácia de barreiras construtivas (compartimentação) na contenção da propagação do incêndio entre habitações térreas geminadas.
Referencial Teórico: A pesquisa fundamenta-se na importância da segurança contra incêndios em habitações térreas geminadas, destacando a vulnerabilidade destas edificações e seus usuários a incêndios. São abordados conceitos de simulação computacional e comportamento do incêndio em edificações.
Método: A metodologia adotada headnote esta pesquisa compreende simulações computacionais utilizando o software Fire Dinamicys Simulator (FDS) headnote analisar diferentes cenários de incêndio. A coleta de dados foi realizada por meio dos outpus gerados pelo FDS e Smokeview (SMV).
Resultados e Discussão: Os resultados obtidos revelaram que: no cenário que não há nenhum tipo de compartimentação entre as habitações geminadas, a propagação do incêndio ocorre pelo vão entre o forro e a cobertura, revelando que este é o ponto de maior vulnerabilidade da edificação; e, nos cenários que há algum tipo de compartimentação, a propagação do incêndio é retardada significativamente. As simulações com o FDS permitiram prever o comportamento do fogo, temperaturas e taxa de liberação de calor (HRR) nos cenários analisados.
Implicações da Pesquisa: As implicações práticas e teóricas desta pesquisa são discutidas, fornecendo insights sobre como os resultados podem ser aplicados ou influenciar práticas no campo de desenvolvimento de projetos baseadas no desempenho de edificações em situações de incêndio. Essas implicações podem abranger o desenvolvimento de projetos mais seguros e promover diretrizes de segurança contra incêndio em edificações residenciais geminadas no Brasil.
Originalidade/Valor: Este estudo contribui headnote a literatura com o avanço de estudos de segurança contra incêndio baseado no desempenho de edificações. A relevância e o valor desta pesquisa são evidenciados pelo uso do FDS em conjunto com as normas de segurança contra incêndio do Brasil.
Palavras-chave: Segurança Contra Incêndios, Habitações Geminadas, Compartimentação, Fire Dynamics Simulator (FDS), Projeto Baseado no Desempenho.
RESUMEN
Objetivo: El objetivo de este estudio es investigar la efectividad de las barreras constructivas (compartimentación) en la contención de la propagación del fuego entre casas adosadas de una sola planta.
Marco Teórico: La investigación se centra en la importancia de la seguridad contra incendios en viviendas adosadas de una sola planta, destacando la vulnerabilidad de estos edificios y sus ocupantes a los incendios. Abarca conceptos de simulación por ordenador y el comportamiento del fuego en edificios.
Método: La metodología adoptada headnote esta investigación consiste en simulaciones por computadora con el software Fire Dynamics Simulator (FDS) headnote analizar diferentes escenarios de incendio. La recopilación de datos se realizó utilizando los resultados generados por el FDS y Smokeview (SMV).
Resultados y Discusión: Los resultados obtenidos revelaron que: en el escenario donde no existe compartimentación entre las viviendas adosadas, la propagación del fuego se produce a través del espacio entre el techo y la cubierta, lo que revela que este es el punto de mayor vulnerabilidad del edificio; y, en los escenarios donde existe algún tipo de compartimentación, la propagación del fuego se retrasa significativamente. Las simulaciones con el FDS permitieron predecir el comportamiento del fuego, las temperaturas y la tasa de liberación de calor (HRR) en los escenarios analizados.
Implicaciones de la investigación: Se discuten las implicaciones prácticas y teóricas de esta investigación, brindando perspectivas sobre cómo los resultados pueden aplicarse o influir en las prácticas en el desarrollo de diseños basados en el comportamiento de los edificios en situaciones de incendio. Estas implicaciones pueden incluir el desarrollo de diseños más seguros y la promoción de directrices de seguridad contra incendios headnote edificios residenciales adosados en Brasil.
Originalidad/Valor: Este estudio contribuye a la literatura al impulsar los estudios de seguridad contra incendios basados en el desempeño de los edificios. La relevancia y el valor de esta investigación se evidencian en el uso del FDS en conjunto con las normas brasileñas de seguridad contra incendios.
Palabras clave: Seguridad Contra Incendios, Viviendas Adosadas, Compartimentación, Fire Dynamics Simulator (FDS), Proyecto Basado en el Desempeño.
1 INTRODUCTION
Rapid urban expansion and changes in the use and occupation of buildings, especially in historic centres, have produced areas of high building and population density, in which fire risks are significantly increased (Palazzi et al., 2023). Tozo Neto & Ferreira (2020) highlight that high building density, proximity and shared walls between adjacent dwellings, and inadequate adaptation for new uses make such areas particularly vulnerable to fire.
In the context of housing types, semi-detached dwellings represent a particular case of vulnerability, especially in historic centres, as they are based on the juxtaposition of units that share walls and, in some cases, roofs, seeking to optimise the use of urban space and reduce implementation costs. However, this configuration favours the spread of fire and heat between adjacent dwellings, increasing the risk of property damage and loss of life in the event of a fire (Palazzi et al., 2023; Tozo Neto & Ferreira, 2020).
As for the location where fires start in buildings, the kitchen is consistently identified as the most frequent starting point, followed by bedrooms or living rooms, reflecting both national and international trends (Corrêa et al., 2017; Spearpoint & Hopkin, 2020). In Brazil, the most common causes of fires in homes include short circuits, high-voltage electrical discharges, overheating of pots, gas leaks, forgetting to turn off irons, and negligent use of electrical appliances, cigarettes, and lighters (Corrêa et al., 2017; Carnieletto et al., 2019; Falcão, 2024).
Fire safety in residential buildings presents a critical gap, especially with regard to Brazilian single-family homes (Corrêa et al., 2025). Fire and panic safety regulations are the responsibility of states and municipalities, with state military fire brigades primarily responsible for developing specific standards. However, unlike vertical buildings, which are covered by specific regulations, single-storey buildings, including semi-detached houses, are exempt by law from any active or passive fire prevention system (Corrêa et al., 2025). This regulatory gap means that fire protection measures, such as compartmentalisation, are not even addressed in their design and construction.
This lack of prevention and legal protection results in a high incidence of fires and casualties in single-family homes in Brazil. These dwellings are the type of building most affected by fires in the country and are the ones that cause the most deaths and injuries in the country (Falcão, 2024; Menezes & Corrêa, 2022; Carnieletto et al., 2019).
At the same time, there has been a significant increase in the adoption of performancebased design in fire safety engineering. Unlike traditional prescriptive methods, performancebased design allows for greater flexibility in the choice of technical solutions, provided that they are proven, through experimental testing or computer simulations, to comply with safety objectives. In this context, Fire Dynamics Simulator (FDS) software has established itself as a support tool for the application of performance-based design in fire safety, due to its ability to model combustion and heat transfer based on adapted Navier-Stokes equations (McGrattan et al., 2025). In conjunction with Smokeview (SMV), a visualisation programme used to display the outputs of FDS simulations, it is possible to analyse the evolution of flames, smoke formation and displacement, thermal stratification, heat release rate (Heat Release Rate - HRR), among others, providing a detailed view of fire behaviour in different construction scenarios.
Given this scenario, the present study aims to evaluate, through computational simulations performed with FDS, the effectiveness of construction barriers in containing the spread of fire between semi-detached single-storey dwellings. Three fire scenarios are analysed, with and without the presence of compartmentalisation elements. The recommendations of Technical Instruction (IT) No. 07/2025 of the Fire Department of the Military Police of the State of São Paulo (CBPMESP) are tested, which, although not specifically addressing singlefamily buildings, provides recommendations on separation between buildings (risk isolation). The aim is to provide technical support for the enhancement of passive protection strategies based on the performance of single-storey semi-detached buildings, contributing to design guidelines aimed at promoting fire safety in high-density urban contexts.
2 THEORETICAL REFERENCE
FDS has been applied to simulate a wide range of fire scenarios, reflecting the diversity of safety engineering challenges. Applications include fire simulation in dormitories (Sá, 2018), two-storey concrete buildings (Byström et al., 2012), sprinkler performance in enclosed compartments (Hopkin et al., 2018), the influence of ventilation (Honma et al., 2013), smoke dynamics (Vigne et al., 2021), fire spread (Himoto et al., 2018), real fire investigation (Silva Filho et al., 2011; Chi, 2013; Husted et al., 2025), hospital fires (Huang, 2022), fires in furnished houses for firefighter training (Yuen et al., 2014), spread in semi-detached houses (Lin et al., 2012), adjacent wooden buildings (Lai et al., 2024), fires in educational environments (Guimarães et al., 2024; Cunha & Pinto, 2016). This diversity of applications reinforces that FDS has proven to be a robust tool for assessing fire behaviour in different building types and fire load conditions.
In simulations with FDS, it is necessary to include at least one combustible material. Wood, widely used as a simplified fuel, adequately represents fire behaviour, allowing essential parameters such as heat release rate (HRR), flame spread, and average temperature evolution to be reproduced with good accuracy (Sá, 2018; Himoto et al., 2018; Byström et al., 2012). This simplification is advantageous because wood has well-documented thermal and combustion properties, allowing models to be calibrated with relative accuracy. Despite limitations in representing complex phenomena such as multi-stage pyrolysis, slow combustion, or advanced carbonisation, wood remains an effective solution for fire simulations in FDS, especially when the goal is to analyse the overall behaviour of the fire rather than the chemical details of combustion (Cicione & Walls, 2020; Torero et al., 2014).
The rate of heat release per area (HRRPUA) is the property that determines the burning speed of combustible material in FDS simulations. This value is sensitive to changes in the model geometry and the mesh adopted (Cicione & Walls, 2020; Sá et al., 2019). In simulations where there are no experiments to calibrate the model, determining the HRRPUA value becomes a challenge. An alternative is to adjust the HRRPUA value to obtain HRR behaviour for a fully developed fire, similar to the classic model presented in standard EN 1991-1-2 (CEN, 2010). This alternative results in fire development behaviour consistent with experimental studies (Sá, 2018).
Processing time is one of the frequent concerns regarding simulations in FDS, as depending on the computer used, they can take hours, days, or even weeks to complete (Huang, 2022; Vigne et al., 2021; Cicione & Walls, 2020). Excessive processing time can make the practical application of simulations in projects unfeasible. In general, processing time depends on factors such as model complexity, size of the scenario analysed, desired simulation time, and mesh size. As for the mesh, the more refined it is, the more volume elements are inserted and, consequently, the longer the processing time will be. In general, meshes between 0.10 m and 0.40 m usually return satisfactory results (Sá et al., 2019; Vigne et al., 2021; Qin et al., 2024).
3 METHODOLOGY
The study was conducted by simulating three fire scenarios in a single-storey semidetached building containing three identical adjacent dwellings. Figure 1 shows the layout of the building. The three housing units are juxtaposed and share dividing walls and the roof. The floor area of each unit is 30.40 m2, totalling 91.20 m2. The ceiling height is 3.00 m and the total height (to the roof) is 5.00 m.
The walls of the building are made of 9 cm thick ceramic bricks, covered with a 3 cm layer of cement mortar on both sides, totalling 15 cm. The floor has a 5 cm layer of concrete, with 5 mm thick ceramic tiles. The ceiling and roof structure are made of wood, and the roof is covered with ceramic tiles. The doors are made of wood, 3 cm thick and measuring 0.90 m x 2.10 m (except for the bathroom door, which measures 0.70 m x 2.10 m). The windows are made of glass and measure 1.00 m x 1.00 m with a 1.00 m sill.
The fire load of the dwellings was calculated based on the methodology suggested by IT No. 14/2025 (CBPMESP, 2025b), surveying the mass of combustible material in the rooms and assigning to the objects the calorific value of the predominant material in their composition.
As this is a hypothetical building, the masses of the furniture were estimated based on manufacturers' catalogues. The calorific value of the materials was adopted as indicated in IT No. 14/2025 (CBPMESP, 2025b). The fire load density per dwelling was 479.70 MJ/m2.
3.1 SCENARIOS ANALYSED
Three scenarios were modelled for the study. In all cases, the internal walls of the dwellings are 3.00 m high (only up to the ceiling) and the walls shared between dwellings have been modified to provide different levels of compartmentalisation:
Scenario 1 (S1): The walls shared between dwellings are 3.00 m high (only up to the ceiling). This is the most common scenario among single-storey semi-detached buildings. Figure 2 shows this scenario modelled in FDS;
Scenario 2 (S2): The walls shared between dwellings are over 3.00 m high, extending to below the roof. Figure 3 shows this scenario modelled in FDS;
Scenario 3 (C3): The walls shared between dwellings are over 3.00 m high, extending above the roof and continuing vertically for 0.90 m above it. In addition, the shared walls extend 0.90 m horizontally beyond the front and rear façades. This configuration adopts the solution recommended by CBMSP, as provided for in IT 07/2025. Figure 4 shows this scenario modelled in FDS;
In all scenarios, internal doors and windows were kept open, and external doors were kept closed during the simulations.
3.2 COMPUTATIONAL DOMAIN, MESH AND SIMULATION TIME
The scenarios were modelled using a rectangular domain measuring 21.60 m x 11.10 m x 7.80 m. This made it possible to model the building and leave the domain boundaries a few metres away from the walls and ceiling, as indicated in the FDS user manual (McGrattan et al., 2025). The domain was divided into a cubic mesh with dimensions of 0.30 m x 0.30 m x 0.30 m, totalling 69,264 elements.
The upper and lateral boundaries of the domain were considered open, so that there could be exchanges between the exterior and interior of the domain, as indicated in the FDS user manual (McGrattan et al., 2025).
The simulations were performed on a standard desktop computer, and the total simulation time was 120 min.
3.3 NON-COMBUSTIBLE OBSTRUCTIONS
To match the dimensions of the adopted grid, the building elements were modelled using obstructions with minimum edge dimensions of 0.30 m. The divergence between the actual and modelled dimensions was corrected by assigning surfaces (SURF) with information on the actual thicknesses of the obstructions. The properties of non-combustible materials are presented in Table 1.
3.4 FIRE LOAD
The fire load was distributed in the form of equivalent wood volume. To this end, the fire load for each combustible object present in the building was converted into wood volume and arranged in the same position as the objects in the building. The properties of density of 400.00 kg/m3, emissivity of 0.90, thermal conductivity of 0.12 W/m.K and specific heat of 1.34 kJ/kg.K were considered, in accordance with NBR 15220 (ABNT NBR 15220, 2005). The calorific value was 19,000.00 kJ/m3, according to IT No. 14/2025 (CBPMESP, 2025b). The ignition temperature was 210°C (Figueroa & Moraes, 2009).
The HRRPUA was 230 kW/m2. This value was established through 60-minute test simulations, modelling only the dormitory. All parameters defined for the scenario simulations were maintained, adjusting only the HRRPUA values of the wood to obtain an HRR graph similar to the classic theoretical model presented in standard EN 1991-1-2 (2010) for fully developed fires. The following were considered for the theoretical graph: average fire growth rate; maximum heat release rate per unit area (RHRf) of 250 kW/m2; time required to reach a heat release rate of 1 MW, tα of 300 s.
3.5 IGNITION AND FIRE
According to McGrattan et al. (2025), for the gaseous fuel that will act in the fire, at least the following parameters must be specified: calorific value, chemical formula, carbon monoxide and soot yields. For the simulations, values referring to wood were adopted: calorific value of 19,000 kJ/kg (CBPMESP, 2025b); chemical formula equal to CH1.7O0.74N0.002, CO yield of 0.004 kg/kg, and soot yield of 0.015 kg/kg (McGrattan et al., 2025).
The source of ignition for the fire was defined as a burning surface located on the stove in the kitchen of dwelling 1 (see Figure 1 and Figure 5). This surface had an area of 0.09 m2 (30 cm x 30 cm) and burned for 150 s with an HRRPUA of 2400 kW/m2.
3.6 PARAMETERS ANALYSED
The results were analyzed using images generated by the SMV, tabulated data automatically recorded by the FDS (such as HRR), and data recorded by meters positioned on the models (thermocouples). Images and videos generated by the SMV were used to analyze the overall behavior of the fire. In addition, a temperature measurement slice was positioned longitudinally in the middle of the building. To measure temperatures, thermocouples were positioned in the center of all compartments at heights of 0.30 m, 0.60 m, 0.90 m, 1.20 m, 1.50 m, 1.80 m, 2.10 m, 2.40 m, and 2.70 m from the floor. Figure 5 shows these meters in scenario C1, and they were considered in the other scenarios.
4 RESULTS AND DISCUSSIONS
The simulation processing time was 13 hours for scenarios C1 and C2, and 20 hours for scenario C3.
Figure 6 shows the HRR graph for the analyzed scenarios.
Figure 6 shows that in all scenarios, there was a peak energy release at approximately 20 min into the simulation. In the SMV image analysis (Figure 7), it can be seen that at this moment, dwelling 1 is completely consumed by the fire and flames are coming out of the dwelling's external doors and windows. At this point, the external doors were already completely consumed. Thus, it was possible to verify that, in all scenarios, the flashover in dwelling 1 occurred approximately 20 minutes into the simulation.
Figure 8 shows that, for all scenarios, at 13 min of simulation, the ceiling of unit 1 had not yet been perforated, so the space between the ceiling and the roof above the dwellings remained at room temperature. However, at 16 min, the ceiling was perforated, indicating that it had begun to be consumed by the fire. At this point, in scenario C1, high temperatures (approximately 800°C) spread through the space above all dwellings. In scenarios C2 and C3, this spread did not occur, as the walls shared between dwellings extended above the ceiling.
In all scenarios, after the peak energy release at 20 min, there was a decay in this release until 44 min (Figure 6). During this time interval, the combustible material was being consumed by the fire and the heat energy contained in dwelling 1 was being depleted, indicating that the fire was in the decay phase.
In scenario C1, at 47 min, the HRR began to increase again (Figure 6), reaching a plateau at 50 min. This indicates that the fire resumed its growth as it spread to dwellings 2 and 3. As shown in Figure 8.b, this spread occurred through the gap between the ceiling and the roof.
In scenario C2, the HRR only began to increase again at 69 min and reached a plateau at 72 min, indicating the spread of the fire from dwelling 1 to dwelling 2. In this scenario, there was no spread to dwelling 3. As shown in Figure 9, in this scenario, the spread occurred through the external ceiling, on the eaves of the front facade between dwellings 1 and 2.
In scenario C3, the HRR did not increase again, indicating that the fire did not spread to dwellings 2 and 3. In this case, the energy release ended at 52 min (Figure 6), indicating the total consumption of combustible materials in unit 1 and the extinction of the fire.
Table 2 presents a summary of the main events observed in the simulations. In all scenarios, the maximum fire temperature occurred in the kitchen of dwelling 1 (the room where the fire originated), measured by the thermocouple 2.70 m above the floor, with the absolute maximum temperature occurring in scenario C1, at 1175 °C at 20 min.
5 CONCLUSION
In this study, three fire scenarios were analyzed in a single-story semi-detached building containing three dwellings. The efficiency of compartmentalization was analyzed, concluding that:
In all scenarios analyzed, the flashover in dwelling 1 (where the fire started) occurred approximately 20 minutes into the simulation;
In scenario C1 (without any type of compartmentalization), the fire spread through the space between the ceiling and the roof shared between the dwellings, highlighting that this is the building's greatest point of vulnerability;
In scenario C2 (with partition walls extending below the roof), the fire spread through the front facade of the building due to the burning of the soffit shared between dwellings 1 and 2, showing that shared facades are also a vulnerability point of the building;
In scenario C3 (with partition walls extending above the roof and beyond the facade), the fire did not spread to adjacent dwellings, showing that the compartmentalization of the roof and facades was effective in preventing the spread of fire in the building;
Although scenario C2 is not able to completely compartmentalize the fire, this solution was able to cause a delay of about 22 min in the spread of the fire from dwelling 1 to dwelling 2. This can ensure a longer time interval for the fire department to be called and arrive at the scene of the fire;
The behavior of temperature evolution, consumption of combustible materials, and the maximum temperature reached in dwelling 1 (the dwelling where the fire originated) were similar in all scenarios analyzed. This shows that compartmentalization between dwellings has little influence on the evolution of the fire in the rooms of the dwelling where the fire originated.
ACKNOWLEDGMENTS
The authors would like to thank the Brazilian Federal Government's Coordination for the Improvement of Higher Education Personnel (CAPES) for its financial support.
References
REFERENCES
ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS - ABNT NBR 15220-2: Desempenho térmico de edificações - Parte 2: Métodos de cálculo da transmitância térmica, da capacidade térmica, do atraso térmico e do fator solar de elementos e componentes de edificações. Rio de Janeiro, 2005.
Byström, A.; Cheng, X.; Wickström, U. & Veljkovic, M. (2012). Full-scale experimental and numerical studies on compartment fire under low ambient temperature. Building and Environment, 51, 255-262.
Carnieletto, E C., Castro, M. G. & Araújo, R. L. (2019). Analise da importância de medidas de Segurança Contra Incêndio em Residências Unifamiliares. Revista FLAMMAE, 5(12), 77- 103.
Chi, J.-H. (2013). Reconstruction of an Inn Fire Scene Using the Fire Dynamics Simulator (FDS) Program·. Journal of Forensic Sciences, 58(S1).
Cicione, A. & Walls, R. S. (2020). Towards a simplified fire dynamic simulator model to analyse fire spread between multiple informal settlement dwellings based on full-scale experiments. Fire and Materials, 45(6), 720-736.
CORPO DE BOMBEIROS DA POLÍCIA MILITAR DO ESTADO DE SÃO PAULO - CBPMSP. (2025a). Instrução Técnica n 07/2025 - Separação entre edificações (isolamento de risco). São Paulo: CBPMESP.
______. (2025b). Instrução Técnica n 14/2025 - Carga de incêndio nas edificações e áreas de risco. São Paulo: CBPMESP.
Corrêa, C.; Silva, J. J. R.; Pires, T. A.; Vieira De Melo, I. A. & Braga, G. C. (2017). Edifício Modal: um modelo para o estudo de Incêndios. Interações (Campo Grande), 18(4).
Corrêa, C.; Vasconcelos, R. S. S.; Wavrik, J. R. A. & Duarte, D. C. L. (2025). A urgência de normas específicas para proteção contra incêndio em residências unifamiliares no brasil. ARACÊ, 7(5), 26838-26848.
Cunha, L. J. B. F. & Pinto, E. M. (2016). Compartimentação horizontal seletiva na promoção da segurança contra incêndio em edificações acadêmicas, por meio da simulação computacional. Revista FLAMMAE, 2(4), 01-28.
European Committee for Standardization. (2010). EN 1991-1-2: Eurocode 1 - Actions on structures - Part 1-2: General actions - Actions on structures exposed to fire. Brussels: CEN.
Falcão, G. M. (2024). Perfil de Ocorrências de Incêndio em Edificações atendidas pelo CBMMG de 2018 a 2023. Revista FLAMMAE, 10(30), 01-43.
Figueroa, M. J. M. & Moraes, P. D. (2009). Comportamento da madeira a temperaturas elevadas. Ambiente Construído, Porto Alegre, 9(4), 157-174.
Guimarães, F. R. J.; Araújo. F. G. S. & Gouveia, A. M. C. (2024). O emprego do modelamento de incêndio pelo FDS (Fire Dynamics System - NIST) na determinação da arquitetura de espaços educacionais para crianças e jovens. Rev. Gest. Soc. Ambient., Miami, 18(8), 1- 16.
Himoto, K.; Shinohara, M.; Sekizawa, A.; Takanashi, K. & Saiki, H. (2018). A field experiment on fire spread within a group of model houses. Fire Safety Journal, 96, 105-114.
Honma, M.; Otake, T.; Miyamoto, H.; Okamoto, K. & Watanabe, N. (2013). Simulation of fire development and spread in japanese-style rooms under different ventilation conditions. In: 13th INTERFLAM, International Fire Science & Engineering Conference, Londres, Reino Unido, p. 1371-1382.
Hopkin, C.; Spearpoint, M. & Bittern, A. (2018). Using experimental sprinkler actuation times to assess the performance of Fire Dynamics Simulator. Journal of Fire Sciences, 36(4), 342-361.
Huang, Y.-H. (2022). Using Fire Dynamics Simulator to reconstruct a fire scene in a hospitalbased long-term care facility. Journal of Loss Prevention in the Process Industries, 80, 104863.
Husted, B. P.; Livkiss, K. & Sauca, A. (2025). Guldborgsund Arson House Fire Experiment and Numerical Investigation. Fire Technology, 61, 247-274.
Lai, C.-C.; Cao, C.-R. & Shu C.-M. (2024). Fire accident report verification using fire dynamics simulator: a case study from Chiayi City, Taiwan. Journal of Thermal Analysis and Calorimetry, 149, 11057-11070.
Lin, C.-S.; Chen, T.-C. & Wang, S.-C. (2012). Computer Modeling and Simulation of A Townhouse Fire Dynamic Field Characteristics. In: IEEE Symposium on Electrical & Electronics Engineering (EEESYM).
McGrattan, K.; Hostikka, S.; Floyd, J.; McDermott, R.; Vanella, M.; Mueller, E. & Paul, C. (2025). Fire Dynamics Simulator: User's Guide. Sexta Edição. Gaithersburg, Maryland, USA: National Institute of Standards and Technology, 504 p. NIST Special Publication 1019.
Menezes, R.R.F & Corrêa, C. (2022). 'Entre mortos e feridos': mapeamento, caracterização e análise dos incêndios com vítimas na região metropolitana do Recife. OPEN SCIENCE RESEARCH VI, 6(1), 1493-1508.
Palazzi, N. C., Baquedano Juliá, P., Ferreira, T. M., Rosas, J., Monsalve, M., & de la Llera, J. C. (2023). Fire risk assessment of historic urban Aggregates: an application to the Yungay neighborhood in Santiago, Chile. International Journal of Disaster Risk Reduction, 86, 103550.
Qin, R.; Dai, X.; Ding, C.; Zhang, Z.; Jiao, Y. & Ren, X. (2024). A Study on the Impact of Window Partition Walls on the Spread of Fire on Building Facades. Fire, 7(7), 253.
Sá, R. T. Simulação numérica de incêndios em edificações. 2018. 177 f. Dissertação (Mestrado em Engenharia Civil) - Universidade Federal de Pernambuco, Recife, 2018.
Sá, R. T.; Plens, D. P.; Oliveira, E. F.; Pires, T. A. C. & Silva, J. J. R. (2019), Considerações sobre o tamanho da malha em simulações com o Fire Dynamics Simulator. In: 5 CILASCI - Congresso Ibero-Latino-Americano em Segurança contra Incêndio, Porto, Portugal.
Silva Filho, L. C. P.; Ruschel, F.; Dolvitsch, J. N.; Lima, R. C. A. & Braga, G. C. B. (2011). Utilização de métodos computacionais nareconstituição de incêndios: aplicação ao caso do shopping total. Revista Sul-Americana de Engenharia Estrutural, Passo Fundo, 8(3).
Spearpoint, M. & Hopkin, C. (2020). A Study of the Time of Day and Room of Fire Origin for Dwelling Fires. Fire Technology, v.56, 1465-1485.
Tozo Neto, J. & Ferreira, T. M. (2020). Assessing and mitigating vulnerability and fire risk in historic centres: cost-benefit analysis. Journal of Cultural Heritage, 45, 279-290.
Torero, J.; Majdalani, A.; Abecassis-Empis, C. & Cowlard, A. (2014). Revisiting the compartment fire. Fire Safety Science, 11, 28-45.
Vigne, G.; Węgrzyński, W.; Cantizano, A.; Ayala, P.; Rein, G.; & Gutiérrez-Montes, C. (2021). Experimental and computational study of smoke dynamics from multiple fire sources inside a large-volume building. Building Simulation, 14(4), 1147-1161.
Yuen, A. C. Y.; Yeoh, G. H.; Alexander, R., & Cook, M. (2014). Fire scene reconstruction of a furnished compartment room in a house fire. Case Studies in Fire Safety, 1, 29-35.
