A Study on Fire Prevention Strategies for Bamboo-Wood Frames and Natural Vegetation Roofs in Southwest China Based on FDS: A Case Study of Wengding Village, Yunnan

Abstract

In Southwest China, traditional wooden buildings in historic villages commonly feature natural vegetation roofing materials, such as thatch or bamboo shingles, which are highly susceptible to fire. Existing research has primarily focused on traditional timber-frame buildings with tiled roofs, while limited attention has been given to those with natural vegetation roofs. This study, taking Wengding village in Cangyuan Wa Autonomous County, Yunnan Province, as an exemplary case, conducts a fire risk assessment and explores fire prevention strategies for buildings with bamboo-wood frames and natural vegetation roofs on the basis of Fire Dynamics Simulator (FDS): the application of fire-retardant coatings, the use of synthetic thatched roofing materials, and a combination of both. The results indicate that the strategy employing synthetic thatched roofing materials offers the best fire resistance performance. By integrating traditional fire prevention knowledge with modern technologies, this study provides a scientifically grounded reference for mitigating fire risks in historic buildings with natural vegetation roofs in China’s ethnic minority regions, aiming to enhance fire safety while preserving architectural authenticity.

1. Introduction

Historical villages [1], witnesses to the evolution of human civilization, stand as fine testament to cultural diversity and richness [2]. Since the initiation of China’s Traditional Village Preservation Project in 2012, 8155 villages have been included in the national protection registry [3], the largest and most comprehensive agricultural heritage conservation network in the world. However, fire risk continues to pose a significant threat to the preservation of cultural heritage [4].

The fire disaster in Wengding village, Yunnan, in 2021, attracted widespread attention. In just seven hours, 104 traditional thatched houses were destroyed, exposing the vulnerability of vernacular historic architecture and revealing the deep-seated challenges in cultural heritage protection amid modernization [5]. In China’s traditional villages, wooden structures are the most classic architectural forms [6]. However, their poor fire-resistant performance, instead of preventing, often fuels rapid fire spread.

As such, strengthening fire safety measures in ancient villages is not only a challenge that must be addressed in architectural studies but also a cultural responsibility.

In fire safety research for ancient villages, the Fire Dynamics Simulator (FDS) is widely used in fire simulation research and is considered one of the most important tools for assessing fire risks [7]. Existing studies have primarily focused on two building systems, timber-frame buildings with tiled roofs and brick–wood frames with tiled roofs (Figure 1), providing important references for fire prevention in ancient villages.

However, research on bamboo-wood frames and natural vegetation roofs—widely found in ethnic minority regions of Southwest China—remains insufficient. The natural vegetation roof, a defining architectural feature of these traditional buildings, is a vulnerable link in fire protection.

Most current literature relies on empirical fire safety measures (e.g., fire hydrants, firebreaks, and timed sprinkler systems) and lacks quantitative analysis based on material combustion characteristics. The lack of research on the technical transformation of traditional fire prevention knowledge and a comprehensive evaluation system for fire protection retrofitting within the context of heritage conservation have resulted in incompatibility between modern fire safety measures and the principles of authenticity and integrity in heritage preservation. Consequently, these measures often fail to ensure both fire safety and heritage protection.

In light of this research gap, this study features the case of the traditional thatched-roof dwellings of Wengding village in Yunnan in particular. Based on FDS modeling, it explores optimal fire prevention strategies under the principles of authenticity in architectural heritage conservation.

The traditional dwellings in Wengding village feature a highly flammable combination of bamboo-wood frames and thatched roofs. The settlement itself is characterized by high-density clusters, with inter-building distances of only 1.2 to 1.5 m. In addition, important cultural customs involving open fires—such as the New Fire Festival—are still practiced. According to local residents [8], the water pressure in fire hydrants was extremely low, ineffective in fire control (Figure 2). These factors contributed to the rapid spread of fire in Wengding [9], at a speed much faster than in other historic buildings.

The fire at Wengding village revealed the extreme vulnerability of vernacular architecture characterized by combustible plant-based materials under fire dynamics. This study focuses on this critical issue by employing an integrated “field investigation—material testing—numerical simulation” methodology to address three key challenges:

(1)

Quantifying the fire spread behavior of bamboo-wood framed buildings with thatched roofs;

(2)

Evaluating the feasibility of combining traditional fire prevention methods with modern technologies;

(3)

Establishing an assessment model for retrofit effectiveness that ensures both heritage authenticity and safety.

The novel contributions of this research include the application of FDS simulation to fire studies of ethnic minority bamboo-wood frame and natural vegetation roof buildings; the development of a parameterized analytical framework for traditional fire prevention wisdom; and the proposal of a cost–benefit-based, phased retrofit strategy.

This study will systematically explore the current state and shortcomings of fire prevention research in traditional villages and propose fire protection improvement recommendations based on FDS simulations using the case of Wengding village.

This study employs FDS in fire simulations on traditional dwellings in Wengding village, with a core focus on fire spread suppression. Through simulations, it quantifies key fire dynamics indicators such as flue gas temperature and fire spread rate; it further explores the fire response performance of dwellings under different scenarios, while proposing targeted firefighting measures from an architectural perspective. The core objective of the study is to balance the needs of ancient village cultural heritage protection, fire safety protection, and sustainable development, ensuring that fire prevention measures do not damage the authenticity of the architectural heritage.

Academically, this study enriches the theoretical framework of fire prevention for vernacular architecture as it breaks away from the reliance on tiled roof structures in current research. Meanwhile, it provides technical support for the “preventive conservation” advocated in the Charter on Vernacular Architectural Heritage [10]. At the practical level, the fire prevention renovation strategies proposed in this study are applicable to similar villages in Southwest China and Southeast Asia and provide a paradigm reference for cultural heritage risk management in the context of rapid urbanization.

2. Literature Review

Ancient Chinese architecture predominantly used wooden structures prone to combustion [11]; thus, fire prevention has always been a widely studied issue in the field of architecture. The evolution of China’s fire protection systems for wooden structures is essentially a history of human technological development in response to the combustibility of materials and the risks associated with spatial density [12]. To better investigate the harm caused by fires and explore effective fire prevention strategies, this study reviews traditional fire prevention practices in ancient China [13], outlines the development trajectory of modern fire research methods, and examines the latest applications of fire simulation software (PyroSim v.2023). Taking a historical perspective, this article systematically studies the evolution of fire protection technologies, with a focus on analyzing the dialectical relationship between traditional wisdom and modern scientific methods. The goal is to provide a solid theoretical foundation for research on fire prevention in ancient villages.

2.1. Fire Prevention Measures for Ancient Chinese Architecture

China’s ancient fire prevention measures, developed through a long history, clearly reflected the profound understanding of fire hazards and the gradual refinement of technical practices by our ancestors.

Ancient Chinese people developed an awareness of fire prevention as early as the 11th Century BCE, as recorded in historical texts. The Book of Changes records [14]: “Water rests above fire, already completed. The wise person foresees danger and prevents it.” This indicates that the ancient Chinese already recognized the importance of proactive prevention. Such understanding laid the theoretical foundation for the subsequent development of fire prevention technologies.

Early periods from 221 BCE to 475 BCE had witnessed the practice of specific, fire prevention measures. According to The Rites of Zhou [15], the Zhou Dynasty established official positions related to “fire administration” [16], responsible for enforcing fire bans and raising people’s fire prevention awareness. Additionally, Shen Zinan in Take an examination of art circles remit [17] mentions that during this period, the ancient Chinese used realgar to ward off fires. The Philosophical Works of Han Feizi [18] also records that people at the time employed methods such as sealing building gaps and applying fire-resistant coatings. These measures reflect the early technological exploration by ancient people.

From 221 BCE to 207 BCE, with the large-scale construction of cities and palaces, fire prevention technology saw significant advancements. Archaeological evidence shows that the urban planning of Luoyang City included a comprehensive drainage system [19]. By 907 CE, fire prevention measures had further improved, with palaces and temples commonly equipped with water tanks and wells. New firefighting tools were also developed, such as the Leather Bag-Bamboo Tube Extinguishing System (Figure 3), which became an effective method of firefighting [20], as described in the A Research on Tu-yu and His Work [21]. This period saw not only firefighting equipment innovations but also institutional improvement, laying a solid foundation for the development of fire prevention technologies in later years.

In the following periods, fire prevention technology in ancient China became more diversified and systematic. During the Song Dynasty, the government promoted brick and tile construction and removed illegal buildings that obstructed rescue channels [22]. In 1023 CE, after the Great Fire of Kaifeng, Emperor Renzong of Song established China’s first firefighting authority, the “Junxun Pu” [23], marking a further improvement of fire safety awareness and technology. In terms of architecture, the Yingzaofashi stipulated that “In all house construction [24], the foundation platform must be constructed with bricks.”, which effectively reduced fire occurrences. The Song Dynasty not only saw innovations in equipment and building materials but also the establishment of an organizational model for systematic fire management, providing valuable references for future generations.

Between 1368 and 1911, ancient Chinese fire prevention technology matured and reached a peak—an evolution clearly demonstrated by the Palace Museum in Beijing, which was completed in 1420. This complex was not only equipped with large copper water storage vats, but also a professional fire brigade with regular fire drills [25]. Additionally, during the Ming Dynasty, building upon the Song Dynasty’s squirt tube technology [26], the introduction of Western hydraulic pumps through the Illustrations and Descriptions of Extraordinary Machines from the Far West [27] led to the development of Jitong-type mechanical firefighting equipment. These innovations contributed to longer range and enhanced collaborative efficiency, further diversifying firefighting apparatuses. The Qing Dynasty inherited and refined the fire prevention technology of the Ming Dynasty. The ridge ornaments of the Palace Museum—known as “Chi-Wen” (Figure 4)—not only served as decorative elements but also lightning rod [28], effectively preventing fires caused by lightning. Moreover, the Qing Dynasty saw the establishment of “fire prevention squads [29]” dedicated to firefighting within the city, marking new advancements in organization and management.

Ancient fire prevention measures, developed and refined through history, are valuable assets still relevant in modern times. These ancient techniques reflect our ancestors’ deep understanding of and emphasis on fire safety, and have laid a solid foundation for the development of modern fire prevention technologies.

2.2. Review and Development of Modern Fire Research Methods

The development of fire research has undergone a significant transformation, evolving from intuitive observation and empirical conclusions to systematic scientific methods. Early studies primarily relied on qualitative analysis and case studies, lacking a comprehensive theoretical and technical framework. However, during this period, fire research was not yet a systematic academic discipline.

In the 1970s, with a change in the way of understanding, fire began to be viewed as a complex, systemic issue rather than an isolated event. “Fire science” gradually emerged as an independent research field. In 1985, the establishment of the International Association for Fire Safety Science (IAFSS) further promoted the international development of the field [30]. In China, after the 1987 Daxing’anling fire, the University of Science and Technology of China established the National Key Laboratory of Fire Science, marking the official beginning of fire science research in the country.

With the development of computer technology, fire research methods gradually became more systematic and rigorous. By the end of the 20th century, fire science entered the theoretical modeling phase, where combustion theory and scientific computing techniques provided new perspectives for research. Performance-based fire design (PBD) became a global trend, emphasizing the design of fire protection measures based on the specific conditions of a building to enhance safety and flexibility [31]. In 1996, Australia introduced the performance-based building code [32], marking the beginning of a new era for performance-based fire protection design.

Entering the 21st century, fire research methods were further developed with experimental simulations and computer simulations emerging as the primary tools [33]. The fire simulation software FDS (PyroSim v.2023), based on Large Eddy Simulation (LES) theory [34], can simulate temperature and heat release rates in fire scenarios, providing strong technical support for fire safety analysis [35]. Studies have shown that FDS offers high accuracy in predicting fire behavior, smoke movement, and heat radiation, among other aspects [36].

Currently, fire research features multidisciplinary integration. For example, real fire environments are simulated through a bidirectional coupling of customized FDS and Abaqus structural analyzers, allowing for a comprehensive capture of the thermal-mechanical responses of structures [37]. Additionally, deep learning technology for fire image recognition has achieved new advancements, enabling greater recognition accuracy [38], which provides a new pathway for intelligent fire protection in ancient villages. Therefore, this study uses FDS for fire simulation experiments [39]. It should be specifically noted that traditional bamboo-wood frames exhibit high heterogeneity and complexity in terms of geometric forms (such as differences in cross-sectional dimensions of log components), material properties, and joint structures. Meanwhile, the primary objective of this study is clearly defined as proposing fire prevention strategies from an architectural perspective, with the core goal of slowing down fire spread to win critical time for personnel evacuation and emergency rescue—this practical demand is prioritized over building structure collapse assessment. Based on this, this study mainly focuses on the quantitative analysis of fire spread processes and the optimization of fire prevention strategies, without separate research on building structures.

2.3. Application of FDS in Fire Protection of Ancient Chinese Buildings

Current fire risk assessment [40] studies mainly focus on modern structures such as high-rise buildings [41], underground facilities, educational institutions, and commercial centers. Fire risk assessments for traditional rural residences and villages are relatively scarce. The lack of targeted research may lead to insufficient protection of these architectural groups, which hold significant cultural and historical value. Therefore, conducting fire risk assessment studies for ancient buildings is crucial for developing appropriate fire prevention strategies.

In the study of fire protection for ancient villages, timber-frame and brick–wood frame with tiled roofs buildings have already been under widespread attention. Several studies have explored the role and historical applications of these structures in fire protection. For example, Huai et al. [42] conducted research on a historic temple in Beijing (timber-frames with tiled roofs) using FDS software (PyroSim v.2023), and found that historical wooden structures with low density and pyrolysis parameters are more prone to flashover. Zhang et al. [43], using the Dong ethnic wind and rain bridge (timber-frames with tiled roofs) in a case study, performed fire simulation analysis with FDS and proposed and verified four fire prevention strategies. Wu et al. [44] studied three ancient villages in Huizhou (brick–wood frames with tiled roofs), creating fire simulation scenarios under 27 different wind speeds, building spacing, and ridge height conditions to analyze the performance of various fire-seal wall types. Liu et al. [45] conducted a comprehensive quantitative analysis of fire prevention strategies for traditional Dong villages in western Hunan (brick–wood frames with tiled roofs) and proposed reasonable fire protection recommendations.

In summary:

(1)

In the field of architecture, the application of fire simulation software has mainly focused on timber-frames with tiled roofs and brick–wood frames with tiled roofs buildings. These studies have provided the authors with insights into fire protection for ancient buildings. However, there has been little research on bamboo-wood frames with vegetated roofs. Due to the characteristics of the materials, these buildings are more susceptible to fire, which often spreads at an alarming speed once breaks out. Therefore, fire research on these types of buildings is crucial.

(2)

Wengding village, among architectural remains with bamboo-wood frames and natural vegetation roofs, which are small in number per se, serves as an ideal research subject. Fire protection research in Wengding village is relatively scarce, and studies using fire simulation software are particularly limited. This indicates that there is still a lot of room for development in fire protection research in this specific area. Therefore, this study fills this research gap by carrying out a study on fire protection for buildings with bamboo-wood frames and natural vegetation roofs, using Wengding village as an example.

(3)

Currently, research based on fire simulation software primarily focuses on fire risk assessment, with limited emphasis on the building itself as a key variable, whose effectiveness in fire protection shall be taken into account. Furthermore, due to the unique characteristics of the Wengding village case, studies on fire spread behavior in such types of buildings remain relatively scarce. This paper intends to dig deeper in this direction, aiming to enhance the understanding of buildings’ dynamic response mechanisms during fires and to provide references for optimizing fire protection strategies for ancient structures.

3. Materials and Methods

To more comprehensively and thoroughly investigate the fire risks of traditional buildings in Wengding village and mitigation strategies, this study deeply integrates field investigation with computer simulations, ensuring that the experimental results are both realistic and precise.

3.1. Field Investigation

This study selects Wengding village in Cangyuan Wa Autonomous County, Yunnan Province (23°12′ N, 99°03′ E) as the research subject. The settlement consists of six natural villages, with Wengding village being the core sample area. It currently houses 109 intact thatched stilt dwellings. The research team conducted field investigations during the dry season of 2024.

3.1.1. Location of the Ancient Village

Cangyuan, located at the southwestern border of China, covers a total area of 2446.43 square kilometers and experiences a low-latitude subtropical climate. The average annual temperature is 18 °C, with annual rainfall ranging from 1500 to 1800 mm. The region is characterized by abundant rainfall, a mild climate, indistinct seasons, distinct wet and dry seasons, and a significant vertical climate.

Wengding village sits in [46] the northwest of Mengjiao Dai, Yi, and Lahu Ethnic Township, on the southern extension of the Nu Mountains. The village is situated in a canyon surrounded by four mountains: Wokan Mountain to the east, Wenghei Mountain to the west, Gonghan Mountain to the south, and Genglao Mountain to the north (Figure 5).

3.1.2. Structure and Characteristics of the Dwellings of the Wa Ethnic Group

The Wa people believe that their villages are living entities, and the “heart” of the village holds sacred religious significance [47]. Under the influence of traditional architectural philosophy, Wa dwellings emphasize harmony with nature, adapting to the local environment and climate and using local materials. This philosophy has shaped the “stilted” architecture style [48], where the structures are elevated, reflecting both practicality and cultural values (Figure 6).

The Wa people’s stilted houses are supported by wooden columns, with a gable roof structure that forms a conical, umbrella-like shape, covered with thatch, and the eaves are relatively low. Without traditional windows, these houses rely on doors and small skylights on the roof, known as “tiger windows,” for natural light. The buildings are divided into two vertical parts (Figure 7): the upper part is used for living, while the lower one is for storing goods and raising livestock. The stilted architecture not only reflects the local cultural characteristics and practicality but also offers effective protection against floods, moisture damage, mosquitoes, and wild animals.

In Wengding village, there are 109 thatched houses located along the ridgeline, with roofs tightly packed together. However, this compact layout has made the village an easy victim of fire hazard—once a fire breaks out, the flames can spread rapidly [49].

3.1.3. Fire-Related Local Customs and Practices

The fire-related traditional customs of the Wa people in Wengding village are deeply embedded in their daily life and belief system. In the traditional wooden-structured villages, in addition to natural factors such as lightning strikes and spontaneous combustion, human activities also contribute significantly to fire hazards. Unlike other regions in China, Wengding village has unique fire-related customs and culture [50].

Apart from fire incidents caused by children’s playful activities, the villagers’ daily fire usage habits and fire-related rituals during festivals also pose significant fire risks, often leading to fire accidents. This study investigates the customs and practices of the Wa people and predicts common behaviors that may lead to fire hazards.

3.2. Software Simulations

Based on field investigations, the study systematically summarizes the current fire prevention methods employed in the area and scientifically verifies their effectiveness. The research utilizes Pyrosim (v.2023) fire simulation software to construct a fire scenario model of the village’s dwellings, providing an in-depth analysis of the fire prevention effectiveness of traditional methods in real-world applications. Based on the simulation results and field survey data, targeted optimization strategies are further proposed.

3.2.1. Simulation Model

This study uses Pyrosim software to simulate the fire hazards in ancient villages, with Wengding village’s dwelling as a case study. First, based on field measurement data, a 3D model of the Eight-Pillar Structural System in Wengding village was created using SketchUp (v.2022) software. The DXF format model files generated in SketchUp were then imported into Pyrosim as the input basis for fire simulation. In the PyroSim software, thermophysical parameters and combustion characteristic parameters of various materials were configured according to the experimental plan. Using a zoning approach, different materials were assigned to corresponding components of the model, and corresponding combustion modes were selected. A systematic simulation and analysis of the combustion behavior and fire spread process of the overall structure under fire conditions were then conducted.

Given the special features and complexity of the traditional buildings in Wengding village, this study set boundary conditions but with necessary simplifications during the FDS modeling process to ensure the model’s computability and the comparability of results. Considering the high heterogeneity and uncertainty in the combustible loads and layouts inside traditional Wa ethnic group dwellings (e.g., the first part used for livestock rearing and firewood storage and the upper part serving as residential space exhibit significant differences in combustible loads), this study adopted a unified simplification approach for the indoor layout across all simulation scenarios: no specific combustible materials such as furniture or debris were placed indoors, only the main structural components of the building were retained, and a basic state of “no additional combustible loads” was uniformly set. This measure aims to eliminate interference from non-core variables such as “differences in combustible types” and “randomness in layout,” and focus the research variables on the core objective—namely, investigating the impact of building structural materials and roof materials on fire spread, as well as the effectiveness of the fire prevention strategies themselves—to ensure the comparability of simulation results across different scenarios. Although these simplifications may lead to numerical deviations between the simulation results and the random process of a single real fire, they are a necessary prerequisite for ensuring scientific rigor and reproducibility in multi-scenario comparative analysis, fully demonstrating the rationality and feasibility of the simplified settings.

In numerical fire simulations, the grid is the basic computational unit of the FDS tool, and its size is crucial as it determines the reliability and accuracy of the simulation results. Generally speaking, the smaller the grid size, the more precise the simulation results will be, but the computational time and cost will increase significantly. Conversely, if the grid size is too large, it will lead to increased simulation errors. According to McGrattan’s study [51], accurate simulation results can be obtained when the ratio of grid size to the minimum grid size is between 4 and 16. The characteristic diameter of the fire source (D*) can be calculated using the following formula:

$${\mathrm{D}}^{\ast }={\left({\displaystyle \frac{\mathrm{Q}}{{\mathsf{\rho }}_0{\mathrm{c}}_{\mathrm{p}}{\mathrm{T}}_0\sqrt{\mathrm{g}}}}\right)}^{{\displaystyle \frac{2}{5}}}$$

(1)

The heat release rate (Q) of the fire source is 8000 kW (i.e., 8000 kJ/s, corresponding to the heat release amount per unit time), the air density (ρ₀) is 1.205 kg/m³, the specific heat capacity of air (c_p) is 1.0042 kJ/(kg·K), the initial environmental temperature (T₀) is 293 K (i.e., 20 °C), and the acceleration due to gravity (g) is 9.8 m/s². The calculation shows that the characteristic diameter (D*) is approximately 2.2 m. Therefore, a grid size in the range of 0.13 to 0.55 m would provide relatively accurate simulation results. Considering both accuracy and computational cost, this study selects a grid size of 0.4 m, with grid cells of 0.4 m × 0.4 m × 0.4 m, resulting in a total of 8100 grid cells. Based on the measured data of the dwelling in Wengding village, the simulation calculation region is determined to be 12 m × 6 m × 6 m, with a total volume of 432 m³.

3.2.2. Simulation Parameters: Fire Source Setting

In this fire simulation, the ignition source was set based on the actual cause of the 2021 Wengding fire, where a foreign fire accidentally ignited the thatched roof. Due to the narrow spacing between buildings in Wengding village, fires can spread easily and are difficult to control. Therefore, in this study, the fire source was placed at the lower part of the building roof (Figure 8) to simulate the scenario of fire spreading from an adjacent building to the area beneath the roof of the target structure. To evaluate the worst-case scenario, all doors and windows of the dwelling were kept open during the entire simulation process. This setup helps provide a comprehensive understanding of how the fire would spread when no protection measures are in place.

The setting of the fire source is a key factor influencing the accuracy of fire simulation results. Since the buildings in Wengding village are primarily constructed with wood, the t² fire growth model is used to describe the development of the fire. The heat release rate (Q) in this model is calculated using the formula:

$$\mathrm{Q}={\mathsf{\alpha }\mathrm{t}}^2$$

(2)

where Q is the heat release rate, t is the time of fire development, and α is the fire growth coefficient. According to the relevant provisions of the “Technical Standard for Smoke and Heat Ventilation Systems” (GB51251-2017) [52] and the findings from field investigations, the buildings in Wengding village conform to the characteristics of fast-burning materials (

Table 1. The four levels of the t² fire growth model.

Growth Type	A (kW/s2)	Typical Combustible Materials
Superhigh speed	0.18760	Oil pool fire, flammable decorative home
High speed	0.04689	Wooden shelf pallets, foam
Medium speed	0.01172	Cotton and polyester items, wooden offices
Low speed	0.00293	Heavy wood products

). Therefore, the fire growth coefficient α is set to 0.04689 kW/s².

In this simulation, the maximum heat release rate used is 8 MW (

Table 2. Maximum heat release rate values for typical fire locations [44].

Typical Fire Locations	Maximum Heat Release Rate /MW (MJ/s)
Shopping malls with sprinklers	5
Offices and guest rooms with sprinklers	1.5
Public places with sprinklers	2.5
Supermarkets and warehouses with sprinklers	4
Offices and rooms without sprinklers	6
Public places without sprinklers	8
Supermarkets and warehouses without sprinklers	20

). Based on the parameters mentioned above, the time required to reach the maximum heat release rate is calculated to be 413 s, at which point the fire reaches a stable burning phase. To comprehensively cover all stages of fire development, the total simulation time is set to 800 s. Additionally, the fire source area is set to 1 m × 1 m to simulate the initial scale of the fire. This setup allows for a more accurate reflection of the combustion characteristics of the Wengding village dwellings and fire spread behavior.

3.2.3. Slice and Measurement Points Parameters

In this fire simulation, the Wengding village dwelling model’s simulation slices are set as shown in Figure 9. The slices are placed along the main axes (x-axis and y-axis) to observe the temperature changes in the upper space of the building. A horizontal slice is set along the z-axis at a height of 2.95 m from the ground, corresponding to the elevated layer of the Wa people’s stilted building (with a base height of 1.35 m), with the addition of the average human eye level (1.6 m) to form the characteristic observation plane. This slice is used to capture the effects of heat and smoke stratification on the evacuation of residents. Longitudinal slices are set along the x-axis at 3.65 m (entrance area), 6.15 m (sacred hearth area), and 8.65 m (bedroom boundary area), covering the full 12-m depth of the space sequence. This setup allows for better observation of temperature variations at different locations during the fire. In particular, the slice at X = 6.15 m is located around the sacred hearth, which is the main activity area for residents, and measuring points here can better visualize the dynamic temperature changes of gases.

The placement of measurement points is strictly aligned with the building structure and fire behavior characteristics. In the sacred hearth core area (X = 6.15 m, Z = 3.15 m) along the axis, four layers of measurement points are set along the vertical direction:

(1)

At the ground level (0 m), the temperature at the base of the fire spread is captured.

(2)

At a height of 1.4 m, corresponding to the combustion plane of the sacred hearth located in the upper part of the dwelling.

(3)

At 2.95 m height, representing the breathing zone of an adult in an upright posture (with the combination of the 1.35 m elevated stilted house layer and average human height).

(4)

At 4.5 m height, the point touches the lower bamboo purlin structure of the sloped roof, used to analyze the heat damage caused by the roof jet flow on building components.

(5)

The locations of these measurement points are shown in Figure 9. Through these points, temperature changes of gases during the fire process can be carefully recorded, providing scientific data for formulating fire prevention and control measures. The setup of these slices and measurement points helps to comprehensively understand the fire spread characteristics and temperature variation patterns inside the dwelling, offering critical data support for future fire prevention and control research.

3.3. Strategy Research

Wengding village dwellings are at high risk of fire, given its highly inflammable construction materials—wood for the frame and thatch for roofs. Currently, residents at Wengding village take rooftop water-sprinkling as the main preventative measure against fire. To validate the effectiveness of this method, this study set up two preliminary experimental groups (

Table 3. Experimental setup to validate traditional fire prevention measures in Wengding village.

Experimental Projects	Variant	Practical Method
Control group	Normal state	Maintain the roof thatch at ≤8 per cent moisture content
Experimental group	Sprinkler intervention	Simulation of daily firefighting operations to increase the water content of the roof thatch to 35 per cent

). In the FDS simulations, all boundary conditions and control variables remained consistent. The burning behavior of roofs under different moisture content conditions was simulated by adjusting the ignition temperature and heat release rate of the thatch material.

A control group (dry roof) and an experimental group (water-sprayed roof). By comparing the results of the two experimental groups, we can assess the fire prevention effectiveness of the water-spraying measure in real-life applications.

If the current sprinkling measure proves to be ineffective, this study further proposes three improvement plans from the perspective of architecture to address the fire risks and combustion conditions of the Wengding village dwellings. The specific experimental design is shown below (

Table 4. Experimental setup of fire prevention strategies proposed for Wengding village.

Experimental Projects	Variant	Practical Method
Experimental group A	Timber frame	Painting of timber frames of dwellings with fireproofing material
Experimental group B	Roofing material	Replacement of roof thatching material with synthetic Class A fireproof thatching material (Figure 10)
Experimental group C	Timber framing and roofing materials	Painting of timber frames of dwellings with fire-resistant materials, while replacing roof thatch materials with synthetic Class A fire-resistant thatch materials

).

In the FDS simulations, different fire prevention strategies were implemented by adjusting the parameters of the building materials and fire mitigation materials (

Table 5. Material parameters.

Material	Density (kg/m3)	Specific Heat (kJ (kg·K))	Conductivity (W/(m·K))	Emissivity
Schima wallichii [53]	742	1.76	0.18	0.8
Moso bamboo	780	1.26	0.2	0.9
Thatch	300	1.3	0.05	0.9
Artificial thatch (HDPE)	950	2.3	0.4	0.9

). The above experimental design creates three scenarios where different measures are assessed. Group A applies fireproof materials to the wooden frame, Group B replaces the thatched roof with fire-resistant materials, and Group C is a combination of both. Comparison of the original control group with these three experimental groups clearly shows the impact of different fire prevention strategies and therefore provides scientific basis for fire prevention design in Wengding village dwellings.

Figure 10. Artificial Class A fireproof thatch material (left) and the use of natural thatch material in Wengding dwellings (right).

Figure 10. Artificial Class A fireproof thatch material (left) and the use of natural thatch material in Wengding dwellings (right).

4. Results and Discussion

This section systematically summarizes the core findings from the field investigation and simulation analysis, and based on these, conducts an in-depth discussion of targeted fire prevention and mitigation strategies.

4.1. Survey Result

The layout of Wengding village, the architectural features of its dwellings, and the ethnic customs hold significant guiding value for fire risk assessment and mitigation strategies. This study collected authentic and reliable data based on field investigations.

4.1.1. Design Features of Wengding Village

The location of Wengding village is cleverly chosen, leveraging the natural mountain terrain. It is situated on a gently sloping area halfway up the sunlit side of the mountain. The village dwellings, distributed in accordance with the terrain, are in a centripetal form where residences encircle religious buildings and sacrificial sites for gods’ protection. Such a design makes it easier for villagers to take care of each other and defend against external threats together.

The main roads within the village form the shape of an octagonal trigram (Figure 11) with narrow lanes stretching out like leaf veins, all blending perfectly with the natural environment. To ensure safety, the village has specially designated fire exits for quick evacuation in case of emergencies. Other pathways within the village range from 0.6 to 1.2 m in width. The core area of Wengding village is the residential zone, while the outer perimeter consists of sacred forests, ponds, barns, head piles, and cemeteries. These parts together make up a complete settlement.

Located within the valley of a subtropical low-latitude mountainous area, Wengding village is often affected by intense monsoon winds [54]. This unique natural condition has shaped the living environment of the village as well as its people’s lifestyle and architectural composition.

4.1.2. Structure and Characteristic of Wa Dwellings

The architectural design and material usage of Wa people’s dwellings reflect distinctive ethnic and regional features. Apart from the wooden framework, beams, columns, and stairs, other parts of the dwellings primarily use bamboo, including bamboo fences, bamboo flooring, and bamboo roof beams. The construction techniques often combine mortise and tenon joints with binding, without the need for special column foundations, simple yet practical.

The Wa people’s dwellings (Figure 12) feature a unique roof design, with bamboo and wood at both ends of the roof intersecting at the ridge in a dovetail shape. This design is not only aesthetically pleasing but also contributes to the structural stability of the building [55]. The overall structure of the house is made of wood, with Schima wallichii for beams and load-bearing columns. The walls are woven with thin wood strips or split bamboo, creating an indoor space. According to existing survey data, the Wa people’s stilted houses are based on an integrated frame system [56], a structural form that is well adapted to the local natural environment and way of life.

As time progresses, the dwellings of the Wa people have also evolved. Currently, in Wengding village, the dwellings are primarily divided into three categories [57], with “the Eight-Pillar Structural System” being the most classic architectural form, widely used in the construction of houses in the village. This study focuses on the Eight-Pillar Structural System and the building materials and parameters for fire-resistant materials are shown in Table 5. The interior layout of the dwellings in Wengding village is simple and clear, with all spaces organized around the hearth (Figure 13).

In terms of material performance and maintenance costs, artificial thatch roofs stand out. Made of high-density polyethylene (HDPE) [58], they exhibit excellent durability with a designed service life of 15 to 30 years—far exceeding the three- to five-year lifespan of natural thatch. They are also resistant to rot, insect infestation, and UV, requiring almost no special maintenance beyond simple daily cleaning, thus highly cost-effective in whole-life maintenance.

Regarding cultural appropriateness and style harmony, high-quality artificial thatch products can be customized to accurately mimic the color, texture, and laying thickness of natural thatch. This allows them to visually replicate the original appearance of traditional roofs and meet landscape harmony requirements. However, from the perspective of cultural heritage authenticity, synthetic materials essentially alter the intrinsic nature of roof materials. Therefore, this replacement strategy is recommended to be treated as a reversible intervention measure. Prior to implementation, full communication with local communities and stakeholders is essential, and this strategy is recommended to be initially applied to non-core protection areas or reconstructed buildings to minimize impacts on the core value of heritage.

4.1.3. Fire-Involved Cultural Practices of the Wa People

• Sacred hearth

In the traditional dwellings of the Wa people, the hearth is the core, located at the center of the upper part of the house. It serves multiple functions, such as cooking and lighting, holds great symbolic significance for the Wa people, and its fire burns continuously throughout the year. However, the dwellings are divided into upper and lower parts by bamboo floorboards; the floorboards around the hearth on the upper part have gaps, large enough for embers to fall to the lower part (where firewood and sundries are stored), which creates a significant fire hazard.

• The New Fire Festival

The New Fire Festival is an important traditional festival of the Wa people, symbolizing “out with the old and in with the new.” In Wengding village, the festival includes such key rituals [59] as “sending off the old fire” and “fetching the new fire.” However, significant safety hazards exist during the festival celebrations: unextinguished embers from the old fire, or improper handling in the new fire-fetching process, can easily ignite fires; additionally, paper scraps, wood chips, and other debris left after the festival may also cause fires.

In addition to the hazards posed by traditional fire-related practices such as the sacred hearth culture and the New Fire Festival, the development of Wengding village as a tourist scenic area has also brought a variety of foreign fire risks [60]. The first one is improper tourist behaviors, which have become a major source of hidden dangers. For instance, tourists randomly discarding cigarette butts, particularly common during peak tourist seasons, directly increases the possibility of fire outbreaks. Second, the local custom of children setting off fireworks and firecrackers during the New Year is highly likely to trigger widespread fires in such an environment with dense wooden-structured buildings. Furthermore, foreign fire sources such as open fire in scenic areas and sparks from equipment during maintenance work also pose serious threats. On top of traditional fire-related practices, foreign fire sources have now become the most prominent cause of fire threat to Wengding village.

4.2. Simulation Results

4.2.1. Burning Behavior

The fire spread in the Wengding village dwellings before and after water spraying at different time points is shown in Figure 14. After the fire ignites, thick smoke rapidly disperses. In the early stages of the fire (at 100 s), the fire spread rate of both experimental groups is essentially the same, both demonstrating a rapid spread of smoke in all directions. However, by 200 s, the fire has rapidly engulfed the entire dwelling, indicating that the fire intensifies quickly within a short period. At this point, the fire spread rate in the untreated roof group is slightly higher than that in the water-sprayed roof group, although the difference is not statistically significant.

As the experiment continued, at 300 s, smoke began to spread outside. It is worth noting that before this (i.e., before 300 s), there was almost no significant difference in the combustion situation between the two experimental groups, indicating that the water-spraying measure had limited control over the fire in the early stages. However, at 400 s, the fire in the untreated roof group was significantly more intense than that in the water-sprayed roof group. By 800 s, the fire in the untreated roof group had almost spread throughout the entire house, while the fire in the water-sprayed roof group, although somewhat smaller, still did not completely prevent the spread of the flames.

4.2.2. Parameter Changes in Slices and Measurement Points

In this study, temperature monitoring (Figure 15) was conducted by setting two groups of measurement points at the same locations (THCP-01, THCP-02, THCP-03, THCP-04), recording the temperature variations of both the untreated (Control Group, abbreviated as CG) and water-sprayed roofs (Experimental Group, abbreviated as EG). The experimental results showed that the temperature at THCP-01 exhibited an overall upward trend over time, and the temperature of the experimental group was slightly lower than that of the control group on the whole, indicating that water spraying had a certain cooling effect at this measurement point. However, both groups followed the pattern of “gradual temperature rise + slight fluctuation in the later stage”. The temperature at THCP-02 rose rapidly to the peak at around 200 s and then remained relatively stable. The temperature curves of the control group and the experimental group almost coincided, yet the temperature of the experimental group became higher instead after 600 s. The temperatures at THCP-03 and THCP-04 rose the fastest, and both groups of experiments reached the peak temperature at 200 s. After reaching the peak, the temperatures of both roofs dropped rapidly to 250 °C and then entered a slow and gradual recovery stage. Additionally, the temperature data at points THCP-02 and THCP-03, corresponding to typical human activity zones, showed minimal differences between the two groups, suggesting that the water-spraying method, as a traditional fire prevention measure, did not significantly reduce the temperature in practical application.

Further analysis of the temperature variation at the slice settings (Z = 2.95 m, x = 3.65 m, 6.15 m, 8.65 m) is presented in Figure 16. Taking Z = 2.95 m as an example, the temperature in both groups of slices rapidly increased to 300 °C within the first 200 s and stabilized at around 250 °C after 500 s. At 800 s, the highest temperature for the untreated roof group reached 470 °C, while the highest temperature for the water-sprayed roof group reached 420 °C. From the overall trend, it can be observed that the high-temperature zones were mainly concentrated in the fire source area, ventilation area, and bedroom area within the dwelling.

Based on the above experimental results, it can be concluded that the water-spraying method has limited effectiveness in fire prevention. Although the peak temperature in the water-sprayed roof group was lower than that of the untreated roof group, the cooling effect was not significant, and it failed to effectively control the distribution of high-temperature zones. Therefore, based on existing research, this study proposes three new fire prevention strategies and evaluates their effectiveness through fire simulations to determine the optimal solution.

4.3. Mitigation Strategies

4.3.1. Comparison of Burning Behavior Under Different Fire Prevention Strategies

This study explores the limitations of traditional fire prevention measures in Wengding village and proposes three new fire prevention strategies to evaluate their effectiveness. For the experiment, the untreated group (with dry roofs) was used as the control group (Figure 17a).

The study shows that the frame structures of Wengding dwellings mostly use local Schima wallichii wood, which is an extremely flammable material. To improve fire resistance performance, the research team applied fire-retardant coatings on the surface of the wooden structures, setting this as Experimental Group A. However, the experimental results (Figure 17b) indicate that the burning behavior of the control group and Experimental Group A was almost identical, with both experiencing intense fire spread. This suggests that simply applying fire-retardant coatings on the wooden frames has little effect.

A prominent feature of Wengding dwellings is their thick thatch roofs, and thatch is actually more flammable than wood. Therefore, the researchers attempted to replace natural thatch roofs with synthetic thatch materials, setting this as Experimental Group B. The experiment found (Figure 17c) that within 300 s after ignition, both the control group and Experimental Group B produced a large amount of thick smoke, filling the entire building. However, at 400 s, the fire in the control group began to significantly surpass that in Experimental Group B; by 600 s, the flames in the control group almost completely engulfed the entire house, while the fire spread in Experimental Group B was relatively slow; at 800 s, the control group was fully surrounded by intense fire, whereas Experimental Group B still maintained a relatively small combustion range. This indicates that synthetic thatch materials have a significant impact on the overall fire prevention effect.

To further explore whether the superposition of variables would bring additional effects, the study set up Experimental Group C, which adopted both fire-retardant coatings and synthetic thatch roofs. The results showed that (Figure 17d) the combustion processes of Experimental Group B and Experimental Group C were basically consistent, indicating that, on the premise of using flame-retardant roofs, the additional application of fire-retardant coatings on wooden structures did not bring significant improvement. From the perspective of building fire prevention, the key to improving the fire resistance performance of Wengding dwellings lies in the optimization of the flame-retardant performance of roof materials, rather than the surface treatment of the main structure.

4.3.2. Comparison of Temperature Transfer Under Different Fire Prevention Strategies

The experiment set up measurement points at the same locations for temperature monitoring and recorded the temperature changes for experimental groups A, B, and C. The measurement points include THCP-01, THCP-02, THCP-03, and THCP-04. The temperature variations at the measurement points over time are shown in Figure 18. The slice settings include Z = 2.95 m, X = 3.65 m, 6.15 m, and 8.65 m. The temperature changes of the slices over time are shown in Figure 19.

• Group A

The temperature changes between experimental group A and the control group at the four measurement points (THCP-01, THCP-02, THCP-03, THCP-04) are nearly identical. Specifically, at the Z = 2.95 m slice, the maximum temperature for both groups is 470 °C. High-temperature areas are mainly concentrated in the fire source, ventilation, and bedroom zones of the dwelling. This indicates that experimental group A did not demonstrate significant advantages in fire prevention performance.

• Group B

Compared to the control group, experimental group B shows a different temperature change trend at the THCP-01 point. After 400 s, its temperature gradually stabilizes, while the control group continues to rise. At the THCP-02 point, the peak temperature in experimental group B occurs after 300 s and then quickly drops to 180 °C, while the control group stabilizes at 200 °C. At the THCP-03 point, the peak temperature in experimental group B occurs after 200 s and drops rapidly to 200 °C, while the control group stabilizes at 250 °C. Similarly, at the THCP-04 point, the peak temperature in experimental group B also occurs after 200 s and rapidly drops to 250 °C, while the control group stabilizes at 300 °C.

Additionally, at 100 s, there is no significant difference in temperature changes between the two groups, with temperatures in all regions not exceeding 100 °C. However, at 200 s, the control group temperature quickly rises to 300 °C, while experimental group B rises to 270 °C. At 400 s, the control group temperature gradually stabilizes at 270 °C, while experimental group B stabilizes at 210 °C. The maximum temperature for the control group reaches 470 °C, while for experimental group B, it reaches 420 °C.

• Group C

The temperature changes in experimental group C are almost identical to those in experimental group B, although its average temperature is slightly lower than that of group B.

By comparing the temperature changes at different measurement points and slices for experimental groups A, B, C, and the control group, the following conclusions can be drawn: Experimental groups B and C show more significant fire prevention effects, with only minor differences between them. Since experimental group C includes the additional variable of applying fire-resistant coatings, but its performance is comparable to that of experimental group B, it can be concluded that applying fire retardant coating to the frame of the Wengding traditional dwellings does not provide significant improvements in fire resistance.

Considering factors such as economic cost, this study concludes that experimental group B is the optimal solution. Experimental group B not only demonstrates superior fire protection performance but is also cost-effective. Therefore, it is recommended that the solution of experimental group B shall be the first choice in practice to achieve the best results both in fire prevention effect and cost control.

5. Conclusions

In this study, we conducted a fire disaster simulation analysis of the traditional dwellings of Wengding village in Cangyuan County based on Computational Fluid Dynamics (CFD). After first identifying potential fire hazards through in-depth investigation of factors such as the structural form, building materials, and fire-related customs of local residents, we undertook both visual and quantitative fire simulation analyses, using FDS and SketchUp modeling software to evaluate fire risks in a comprehensive manner and propose corresponding fire prevention strategies. The findings indicated that the traditional dwellings in Wengding are highly susceptible to fire hazards, and the use of fire-resistant roofing materials can significantly improve their fire prevention performance. The details of the findings are as follows:

(1)

The investigation shows that the traditional dwellings in Wengding village are at high risk of fire. Wood and thatch, the primary building materials, are fire-prone when exposed to the natural environment. Additionally, the Wa people’s fire-related customs, such as the New Fire Festival and sacrificial activities, lead to increasing fire risks. Once a fire breaks out, the surrounding forests and dwellings are likely to be engulfed, causing severe damage.

(2)

Fire simulation indicates that the traditional dwellings in Wengding village are highly susceptible to damage in the event of a fire. Through experimentation, we found that the current roof sprinkling system is not effective enough to control the rapid spread of fire once it breaks out.

(3)

To address these issues and in consideration of the features of traditional dwellings in Wengding village, this study proposes three fire prevention strategies (

Table 6. Three fire prevention strategies and their corresponding effects.

Fire Prevention Strategies	Effect
Fire-resistant coatings	Poor
The use of synthetic thatch roofing materials	Good
Combination of both	Good

): ① the application of fire-resistant coatings; ② the use of synthetic thatch roofing materials; and ③ a combination of both.

Simulation evaluated the effectiveness of these mitigation strategies. The results show that the combination of fire-resistant coatings and the use of synthetic thatch roofing materials yields the best performance, followed by artificial roofing alone, and with fire-resistant coatings being the least effective. However, the combination of fire-resistant coatings and artificial roofing, though the most effective, is neither cost-efficient nor easy-to-apply in practice. The study therefore recommends artificial roofing to be the first choice, as it is feasible in enhancing fire resistance performance while preserving traditional architectural features. The use of fire-resistant coatings alone is not advisable, as its effect is limited.

Despite its pursuit of rigor to the utmost, the study still has certain limitations, which are elaborated as follows: First, in the construction of the simulation model, practical factors such as residents’ daily activities and dynamic combustible loads were not incorporated. It should be clarified that this simplification is not an oversight in the research design, but a deliberate choice to focus on the core objective of “the effectiveness of building structures and fire prevention strategies.” By controlling variables within a reasonable scope, the key mechanisms of the core research objects can be more prominently revealed. Second, due to constraints of computer hardware performance, this study had to make necessary simplifications to the model of fire simulation scenarios. This may have a certain impact on the precision of the simulation results; however, the overall experimental design still maintains scientific rigor and integrity. Additionally, factors affecting fire risks in traditional villages are diverse and complex. This study mainly conducts experimental analysis from the perspective of architecture and has not fully covered multidimensional factors such as social and environmental aspects, thus resulting in certain limitations in terms of applicability.

It is also worth noting that traditional dwellings in other regions of Yunnan Province differ from those in Wengding, Cangyuan, in aspects such as structure, form, and cultural practices. Therefore, the fire simulation conclusions of this study may not be fully applicable to scenarios with significant differences in fire types. Nevertheless, the fire risk assessment methods and fire prevention strategies proposed in this study can still provide valuable references for disaster prevention practices in similar traditional settlements.

In conclusion, this study provides valuable references for optimizing and renovating buildings with bamboo-wood frames and natural vegetation roofs. Future research can focus on refining fire dynamics models, with an emphasis on conducting dynamic coupling simulations of occupant evacuation and fire spread. Meanwhile, it is also crucial to explore the introduction of stochastic fire sources and advanced combustion models; furthermore, multiple practical environmental factors can be incorporated, as well as the integrated analysis of optimized firefighting equipment. These multi-dimensional research directions will provide more systematic and accurate solutions for the fire protection design and safety assessment of traditional vernacular dwellings.