Exploring the Development Trajectory and Dynamic Frontiers of Numerical Simulation Method Applied in Thermal Protective Clothing Investigation

Abstract

With the continuous advancement of digital technology and intelligent algorithms, numerical simulation has become an important approach for investigating the thermal transport behavior and protective performance of thermal protective clothing (TPC). To provide a systematic understanding of the development and current research landscape of numerical simulation of thermal protective clothing (NSTPC), this study reviewed the development status of NSTPC through statistical analysis, bibliometric methods, and in-depth literature reading. The theoretical foundations, model construction, and development process of NSTPC were analyzed, with particular emphasis on heat transfer mechanisms, moisture transport, multilayer material structures, and fabric–air gap–skin interfaces. The results indicated that NSTPC has developed along multiple directions, with models ranging from one-dimensional to three-dimensional approaches, from material- and layer-level representations to garment-level simulations, and from single-physics heat transfer models to coupled multi-physics models. These approaches continue to coexist and are employed for different research purposes. Future research may further explore coupled multi-physics field simulation, dynamic simulations of the human body and heat exposure scenarios, the design and evaluation of novel high-performance thermal protective materials and coatings, and the integration of artificial intelligence and machine learning with numerical simulation. These findings provide a systematic knowledge base for understanding NSTPC development and support the design, performance evaluation, optimization, and safety assessment of TPC.

1. Introduction

Workers in high-temperature environments, such as emergency responders and industrial workers, are exposed to various heat sources, including flames, molten metal sputtering, and hot gases [1]. Thermal protective clothing (TPC), consisting of multilayer textile structures and insulating components, is widely applied to reduce heat penetration and protect underlying substrates from thermal damage. By utilizing flame-resistant fibers, insulating layers, and functional protective coatings, TPC systems can reduce or delay heat transfer to the wearer/skin. TPC can be classified according to materials, protective functions, protected body regions, and other characteristics. Materials include flame-resistant fibers and fabrics, aramid and other high-performance fibers, polymeric composites, and coated textile systems. Protective functions include protection against flame, thermal radiation, high-temperature liquids, and steam, among others, while TPC can be designed for whole-body protection or specific regions such as the hands and head. ISO 11612 specifies minimum performance requirements and test methods for protective clothing against heat and flame, covering limited flame spread, convective heat, radiant heat, contact heat, and molten metal splashes.

Numerous research centers and pioneering branches are dedicated to the study of TPC. Most of the research predominantly employs physical experiments. However, certain experimental methods may not adequately reproduce the dynamics of a real-world scenario. These methods tend to be costly, and their results can be affected by environmental conditions, clothing properties and structural characteristics, and experimental conditions [2]. With advancements in computer technology, numerical simulation methods have been increasingly integrated into the research of TPC. Numerical models can simulate specific fire/thermal exposure scenarios [3], predict the thermal protective performance (TPP) of TPC, and assess the severity of skin burns. Moreover, they can assist in identifying and characterizing critical parameters that influence the TPP of fabrics, thereby enhancing insight into complex heat and mass transfer phenomena [4]. Several scholars have summarized and discussed the application of numerical simulations in TPC. Advancements in simulating heat and mass transfer across porous textiles, heat transfer within the air gap (AG) between fabric and skin, and bio-heat transfer models were summarized by Udayraj et al. [4]. Deng et al. [5] examined the impact of AG characteristics within firefighters’ clothing on TPP. Weng et al. [6] synthesized knowledge on the body’s thermal regulation mechanisms, heat stress, skin burns, and methods for assessing inhalation injuries in fire environments. Shakeriaski et al. [7] investigated the heat transfer properties of fire-retardant fabrics. However, these studies tend to concentrate on specific topics, and the potential knowledge structures and developmental patterns across different research directions remain unexplored, resulting in an insufficient understanding of the overall development context and trends.

Bibliometrics investigates citation networks to identify and visualize research trends and shifts, offering quantitative insights into specific scientific domains [8]. Employing bibliometrics and knowledge mapping, Li et al. [9] examined the status of research, identified key areas of interest, and forecast potential future directions of functional clothing. Azam et al. [10] performed a systematic scientific review of wind energy literature, mapping the knowledge landscape and identifying emerging research frontiers. Thus, bibliometrics effectively aids researchers in understanding and exploring the knowledge structure and frontier dynamics of scientific research.

Through statistical analysis, in-depth literature reading, and bibliometric analysis of the literature in Web of Science (WoS) Core Collection regarding numerical simulation of thermal protective clothing (NSTPC), the development status, dynamic frontiers, and collaboration patterns within this field were comprehensively examined, while also exploring its potential knowledge structure. A systematic knowledge base and reference value for research in the area of NSTPC are provided in this work.

2. Methodology

2.1. Research Method

The bibliographic information retrieved from WoS can be statistically analyzed to provide an overview of research fields, helping researchers understand the distribution and impact of publications and identify research trends. Building upon this foundation, in-depth literature reading helps researchers develop a comprehensive understanding of the research field, analyze the research context, and accurately comprehend core concepts and theories. Bibliometrics enables research analysis by identifying emerging trends through burst detection, mapping knowledge structures via co-citation clusters, and revealing temporal evolution patterns of scientific fields. In this study, statistical analysis was used to characterize the distribution and development of scientific output, while bibliometric analysis was employed to further identify research hotspots, knowledge structures, and emerging trends. In-depth literature reading was conducted to examine the theoretical foundations, modeling approaches, and research development of NSTPC. By integrating these complementary analytical approaches, this study investigates the development trajectory, research hotspots, and potential future directions of NSTPC. The schematic diagram of the research framework is shown in Figure 1.

2.2. Data Collection and Screening

WoS Core Collection was used as the literature source, renowned as a prominent platform for citation analysis and bibliometric research in the scientific community [11]. WoS Core Collection is widely recognized for its rigorous indexing standards and provides good coverage of materials science, textile engineering, heat transfer, and protective technologies relevant to this study, together with standardized bibliographic and citation information. The literature search was conducted using terms related to “thermal protective clothing” and “numerical simulation,” with the complete search strategy provided in Appendix A. The time span was from 1 January 1998 to 31 July 2026. To ensure consistency in the types of publications included and to reduce potential noise from non-full research records, the search was restricted to articles and reviews published in English. A total of 2030 publications were identified through the initial screening, and 2030 records remained after removing duplicates from the database search results. The articles were screened based on their titles and abstracts, and any uncertainty regarding eligibility was discussed with another researcher to reach a consensus. The eligible articles were then assessed through full-text review, and studies that didn’t meet the predefined inclusion and exclusion criteria were excluded, as follows:

(a)

Numerical simulation studies focusing on thermal protective clothing or protective textile systems exposed to high temperature, radiant heat, flame, or steam were retained.

(b)

Numerical simulation studies addressing thermal protection performance or related thermal responses of clothing systems, including temperature distribution, heat flux, heat transfer, moisture transfer, or skin burn prediction, were retained.

(c)

Studies focusing solely on materials, human thermophysiology, or other topics unrelated to the numerical modeling of thermal protective clothing or relevant textile systems were excluded.

A total of 136 publications were included in the final study.

3. Statistical Analysis

3.1. Publication Analysis

The distribution of 136 publications and their citation frequencies over the years was analyzed using the citation report function of WoS (Figure 2). Between 1998 and 2026, both the overall number of publications and citations exhibited an upward trend. It was suggested that research on NSTPC has attracted increasing attention, as reflected by the increasing number of publications and citations. Based on this growth trend, the research can be categorized into three developmental stages: the initial stage, the development stage, and the prosperity stage. The period from 1998 to 2007 represented the initial stage, characterized by a low average annual number of publications and citations, with numerous instances of zero output. From 2008 to 2015, the field entered the development stage, during which the average annual number of publications saw a slight increase, while the average annual number of citations surged nearly 23-fold compared to the previous period. Overall, from 1998 to 2015, the number of publications fluctuated significantly and remained generally low. After 2016, publication output and citation activity increased substantially, with more than 9 publications and 200 citations recorded on average per year. Based on this marked increase in research output and citation activity, the period was identified as the prosperity stage in this article.

3.2. Research Domain and Journal Analysis

A total of 23 distinct research fields were identified using the analysis results function in WoS. Figure 3 illustrates the temporal development of the top four research fields based on the number of publications. Among these, engineering emerged as the predominant field, with 66 published papers, representing 48.5% of the total publications, followed by materials science, accounting for 45.6%. Additionally, significant research fields were observed in thermodynamics (39.7%) and mechanics (27.9%). These percentages are not mutually exclusive because individual publications may be assigned to multiple research fields. Research in materials science and engineering commenced earlier, with literature recorded dating back to before 2000, whereas the literature in thermodynamics and mechanics only began to appear in 2005. It was indicated that NSTPC spanned multiple disciplines. Materials science and engineering represented major research categories in the NSTPC literature, reflecting research attention to material properties, structural characteristics, and heat-transfer behavior. Due to the numerous limitations associated with physical experiments examining factors such as AG and moisture, the numerical simulation method has garnered increasing attention and application from scholars post-2005, leading to a gradual development in the fields of thermodynamics and mechanics.

To elucidate the distribution of journals within the research field and the citation relationships between the citing and cited fields, we employed the superposition diagram for further analysis (Figure 4). It was constructed using data extracted from over 10,000 journals in WoS [12]. The cluster of citing journals signifies the knowledge frontier on a specific topic, while the cluster of cited journals constitutes the knowledge base. This visual representation aids in clarifying the flow of knowledge between the citing and cited fields/journals.

The superposition diagram includes two base maps: the left and right maps depict citing and cited relationships, respectively [13]. Curves in this diagram illustrate citation connections, depicting citation flow between journals and supporting analysis of citation trajectories. The labels denote the relationship between journals and subject fields. The predominant citing subject field encompassed physics, materials, and chemistry, followed by mathematics, systems and mathematical. Notable representative journals within these fields included International Journal of Thermal Sciences, Journal of the Textile Institute, and Fire Technology, aligning with the publication distribution from WoS. The disciplines with the highest number of cited journal articles were chemistry, materials, and physics, followed by environmental, toxicology, nutrition and sports, rehabilitation, and sport. Key journals associated with the disciplines included Textile Research Journal, Fire Technology, and International Journal of Occupational Safety and Ergonomics. Research within NSTPC was characterized by its interdisciplinary nature, often closely linked to mathematics, materials science, environmental studies, human movement, and health. It is necessary that scholars employ multidisciplinary approaches to conduct systematic research in related areas.

4. Theoretical Foundations

Numerical simulation is a technical approach that employs mathematical models and computational algorithms to approximate the numerical solutions of complex physical problems through discretization and iterative calculations. This method utilizes computers to simulate physical processes, enabling the analysis of phenomena that are challenging to observe or measure directly through experiments. With rapid advancements in computer technology and artificial intelligence, numerical simulation has emerged as a vital method in the research of TPC [14]. It effectively simulates the heat and moisture transfer processes of TPC under complex environments, allowing for predictions regarding its thermal performance and comfort. The application of this technology in the study of TPC offers significant theoretical support for clothing design, performance optimization, and safety assessments [14,15].

4.1. Model Construction

The physical model comprises a thermal environment, multilayer clothing, AGs, and skin tissue (Figure 5). Given that the thickness of the fabric and skin is much smaller than their surface area, they can be approximated as infinitely large parallel planes, with heat and moisture transfer typically occurring in the direction perpendicular to the fabric surface. Consequently, some studies simplified the human–clothing–environment system into a one-dimensional (1D) model to mitigate complexity. However, to more accurately simulate complex phenomena such as airflow distribution and temperature field variations in fire scenes, researchers have developed a three-dimensional (3D) full-scale model. After establishing the material properties (e.g., thermal conductivity, specific heat capacity), initial conditions and boundary conditions (e.g., environmental temperature, wind speed), the model is solved numerically. The simulation results are then verified and analyzed against experimental data to ensure the model’s accuracy and reliability.

4.1.1. One-Dimensional Model

The 1D heat transfer model is constructed based on Fourier’s law and the law of energy conservation (Equation (1)). This model can perform single-field or multi-physics coupling simulations based on research requirements. If moisture transfer is considered, it is necessary to introduce the phase change effect of liquid water. Fick’s law is employed to describe the diffusion of water vapor, while Darcy’s law is used to simulate the seepage behavior of liquid water in porous media.

$$\rho c_p{\displaystyle \frac{\partial T}{\partial t}}={\displaystyle \frac{\partial }{\partial x}}\left(k{\displaystyle \frac{\partial T}{\partial x}}\right)+Q$$

(1)

where ρ is density, kg/m³; c_p is the specific heat, J/(kg·K); T is temperature, K; t is time, s; x is horizontal coordinate; k is thermal conductivity, W/(m·K); Q is internal heat source, W/m³.

Most of the incident radiation is absorbed within a path length equivalent to three fiber diameters. Only the radiation from the outer fabric layer needs to be considered, which is calculated using the two-flux radiative heat transfer model [16] (Equations (2) and (3)).

$${\displaystyle \frac{\partial F_R}{\partial x}}=-\beta (x)F_R+\beta (x)\sigma T^4(x,t)$$

(2)

$${\displaystyle \frac{\partial F_L}{\partial x}}=\beta (x)F_L-\beta (x)\sigma T^4(x,t)$$

(3)

where F_R and F_L represent the total thermal radiation incident at point x traveling to the right and left, respectively, W/m²; β is the radiative absorption constant of the fibers, 1/m; and σ is the Stefan–Boltzmann constant, 5.67 × 10⁻⁸ W/(m²·K⁴).

The primary constituents of air are nitrogen and oxygen, which are weak emitters of thermal radiation in the infrared region under typical fire conditions, due to their lack of an electric dipole moment and consequently limited infrared activity. Radiative heat transfer in air is modeled using Beer’s law (Equation (4)).

$$I(x)=I_0\exp (-\gamma x)$$

(4)

where I(x) is the transmitted radiation intensity, W/m²; I₀ is the incident radiation intensity, W/m²; and γ is the absorption coefficient, 1/m.

The skin tissue is composed of the epidermis, dermis, and subcutaneous tissue. The Pennes bioheat equation [17], which is founded on Fourier’s law of heat conduction, is widely utilized to model heat transfer within the skin (Equation (5)). The epidermis is avascular and thus does not necessitate the consideration of blood perfusion rates. Given that the dermis and subcutaneous tissue are vascularized and subject to blood perfusion, blood flow and metabolic heat generation should be considered in modeling heat transfer within these tissues [18].

$${(\rho {\mathrm{c}}_p)}_{skin}{\displaystyle \frac{\partial T}{\partial t}}={\displaystyle \frac{\partial }{\partial x}}\left(k_{skin}{\displaystyle \frac{\partial T}{\partial x}}\right)+{\omega }_b{(\rho c_p)}_b\left(T_{\mathrm{b}}-T\right)+G_m\sqrt{a^2+b^2}$$

(5)

where ρ_skin and ρ_b represent the density of skin tissue and blood, respectively, kg/m³; (c_p)_skin and (c_p)_b refer to the specific heat of skin tissue and blood, respectively, J/(kg·K); k_skin is the thermal conductivity of skin tissue, W/(m·K); ω_b donates the rate of blood perfusion, 0.00125 (m³/m³·s); T_b is the blood temperature, K; and G_m is the metabolic heat production, W/m³.

Biological systems exhibit heterogeneous structures with relatively long thermal relaxation times [19,20]. Under short-duration intense heating, when the exposure time approaches the thermal relaxation time, Pennes’ equation may overpredict the early temperature response, whereas the thermal wave model of bioheat transfer (TWMBT) accounts for the finite thermal wave velocity. Pennes’ and TWMBT predicted markedly different burn times of 0.097 and 2.04 s, respectively, at 83.2 kW/m² [21]. TWMBT (Equation (6)), based on the hyperbolic heat conduction equation (Equation (7)) proposed by Lui et al. [21], may help reduce potential deviations associated with the Pennes model.

$${(\rho c_p)}_{skin}\tau {\displaystyle \frac{{\partial }^{2}T}{\partial t^2}}+\left({\tau }_q{w}_b{\rho }_b{c}_b+{(\rho c_p)}_{skin}\right){\displaystyle \frac{\partial T}{\partial t}}+{\omega }_b{(\rho c_p)}_b\left(T-T_b\right)=k_{skin}{\displaystyle \frac{{\partial }^{2}T}{\partial x^2}}+Q_r$$

(6)

$$q+\tau {\displaystyle \frac{\partial q}{\partial t}}=-k\nabla T$$

(7)

where q is the heat flux density, W/m²; τ is the thermal relaxation time, s; and Q_r is the internal heat source in the space, W/m³.

The boundary conditions for the fabric layer, or skin layer, in contact with air typically require consideration of both convection and radiation. First- or second-type boundary conditions are generally applied under alternative conditions, indicating that regulating temperature or heat flux density is adequate.

The finite difference method is the primary approach for solving 1D models, incorporating discretization formats such as explicit, implicit, and Crank–Nicolson. The Crank–Nicolson difference method provides second-order temporal accuracy with relatively low computational cost. [16]. A commonly utilized approach for evaluating skin burn severity is the Henriques skin burn integral model (Equation (8)).

$$\Omega ={\displaystyle {\int }_0^{t}P}\exp \left(-{\displaystyle \frac{\Delta E}{RT}}\right)dt$$

(8)

where Ω is a quantitative measure of burn damage at the basal layer or at any depth in the dermis; P is the frequency factor; ΔE is the activation energy for skin, J/mol; R is the universal gas constant (8.314 J/mol·K); and T is the total time for which T is above 44 °C.

4.1.2. Two-Dimensional Model

The two-dimensional (2D) heat transfer model enhances the 1D model by incorporating lateral heat and moisture transfer. Both vertical and horizontal heat and moisture transfer within the fabric–AG–skin system are further accounted for. The non-uniformity of temperature distribution in the fabric is more accurately reflected in this approach, leading to improved calculations of heat transfer.

4.1.3. Three-Dimensional Model

The 3D heat transfer model is developed based on the theoretical principles of mass conservation, momentum conservation, and energy conservation. Both axial and radial heat transfer behaviors within the fabric–AG–skin system are simultaneously analyzed in this model [22]. By integrating natural convection and radiative heat transfer within the AG, it enables high-precision simulation of the full-scale heat transfer process.

(a)

The continuity equation is used to describe the mass conservation relationship in the flow field (Equation (9)).

$${\displaystyle \frac{\partial {\rho }_f}{\partial t}}+{\displaystyle \frac{\partial \left({\rho }_{f}u\right)}{\partial x}}+{\displaystyle \frac{\partial \left({\rho }_{f}v\right)}{\partial y}}+{\displaystyle \frac{\partial \left({\rho }_{f}w\right)}{\partial z}}=0$$

(9)

where ρ_f is the fluid density, kg/m³; u, v, w are fluid velocity vector in x, y, z direction, m/s.

(b)

The momentum conservation equations (Equations (10)–(12)), derived from the Navier–Stokes equations, reflect the changes in the velocity field and account for the effects of viscosity in the fluid.

$${\rho }_f\left({\displaystyle \frac{\partial u}{\partial t}}+u{\displaystyle \frac{\partial u}{\partial x}}+v{\displaystyle \frac{\partial u}{\partial y}}+w{\displaystyle \frac{\partial u}{\partial z}}\right)=\mathrm{div}(\mu \mathrm{grad}u)+S_u-{\displaystyle \frac{\partial p}{\partial x}}$$

(10)

$${\rho }_f\left({\displaystyle \frac{\partial v}{\partial t}}+u{\displaystyle \frac{\partial v}{\partial x}}+v{\displaystyle \frac{\partial v}{\partial y}}+w{\displaystyle \frac{\partial v}{\partial z}}\right)=\mathrm{div}(\mu \mathrm{grad}v)+S_v-{\displaystyle \frac{\partial p}{\partial y}}$$

(11)

$${\rho }_f\left({\displaystyle \frac{\partial w}{\partial t}}+u{\displaystyle \frac{\partial w}{\partial x}}+v{\displaystyle \frac{\partial w}{\partial y}}+w{\displaystyle \frac{\partial w}{\partial z}}\right)=\mathrm{div}(\mu \mathrm{grad}w)+S_w-{\displaystyle \frac{\partial p}{\partial z}}$$

(12)

where μ is dynamic viscosity, Pa·s; S_u, S_v, S_w are source terms, N/m³; and p is pressure, Pa.

(c)

The energy conservation equation is utilized to characterize the heat transfer and temperature variations within the fabric, air, and skin tissue (Equation (13)).

$${\rho }_f{\mathrm{c}}_p\left({\displaystyle \frac{\partial T}{\partial t}}+u{\displaystyle \frac{\partial T}{\partial x}}+v{\displaystyle \frac{\partial T}{\partial y}}+w{\displaystyle \frac{\partial T}{\partial z}}\right)=\nabla \cdot (k\nabla T)+S_T$$

(13)

where S_T is the source term, J.

The AG encompasses significant processes, including convective heat transfer, moisture diffusion, and phase change, all of which are influenced by complex geometries and boundary conditions. These processes are typically addressed using the finite volume method, which relies on the integral discretization of conservation equations over control volumes to ensure the local conservation of physical quantities, making it particularly suitable for fluid dynamics problems [23]. The fabric and skin layers exhibit anisotropic and thermal-moisture coupling characteristics [24]. The finite element method presents advantages when addressing non-uniform materials and complex structures. The weak form-based finite element method can adeptly manage multilayer structures and intricate boundaries, achieving high computational accuracy through the use of high-order shape functions or adaptive meshing.

4.1.4. Models Across Different Dimensions

The 1D model, which considers only heat and moisture transfer perpendicular to the fabric–AG–skin interface, has a small computational scale and demonstrates good convergence in solving. It effectively reflects the changing patterns of coupled heat and moisture processes, making it suitable for large-scale parameter studies and sensitivity analyses. However, it fails to capture phenomena such as the non-uniformity of fabric structure, AG distribution, and boundary effects. The 2D model can more accurately simulate abnormal spatial distributions of heat flow caused by local structural inhomogeneities, providing a more reliable computational basis for thermal safety assessments in cases of localized overheating or partial fabric damage. 3D models excel at simulating complex factors such as non-uniform heat flow, fabric pore distribution, fiber geometric morphology, and variations in AG thickness, thereby more closely approximating the actual working conditions of protective clothing. Their application is limited due to high computational costs, challenges in numerical stability, and difficulties in parameter acquisition. By coupling small-scale and full-scale models and utilizing methods such as machine learning, it is possible to enhance the operability and promotional value of 3D simulations while balancing computational efficiency and physical accuracy. A detailed comparison of models with different dimensions is presented in

Table 1. Comparison of 1D, 2D, and 3D numerical models for TPC.

Model Dimension	General Assumptions	Governing Physics	Input Parameters	Typical Applications	Common Validation Methods	Computational Cost	Common Software
1D	Planar; heat and mass transfer through the fabric thickness	One-dimensional heat and mass transfer equations; Fourier, Fick, Darcy, Pennes, etc.	Physical, geometric, and physiological parameters; initial and boundary conditions	Parameter studies; sensitivity analyses	Bench-scale tests; temperature and heat-flux measurements	Low	MATLAB
2D	Heat and mass transfer in the thickness and one lateral direction	Two-dimensional heat and mass transfer equations, accounting for variations in the thickness and lateral directions	Above parameters + lateral geometric and physical parameters	Local non-uniformities, localized damage, and lateral heat transfer	Bench-scale tests; temperature and heat-flux measurements	Medium	MATLAB, ANSYS Fluent, etc.
3D	Spatial variations in structure and boundary conditions are considered	Three-dimensional mass, momentum, and energy conservation equations; the Navier–Stokes equations can be used to describe airflow and its spatial distribution	Above parameters + 3D geometry, airflow and spatially varying exposure conditions	Complex garment structures, human-body curvature, and complex fire environments	Instrumented manikin tests; human subject tests	High	ANSYS Fluent, COMSOL, etc.

.

5. Bibliometric Analysis

5.1. Contributing Countries and Institutions

The main contributing regions and their cooperative relationships can be analyzed through a co-occurrence network. To manage node count and network size, both countries and institutions were selected as node types, resulting in the creation of a national and institutional cooperation network map (Figure 6). Utilizing the function of analyzing results from WoS, it was found that 22 countries contributed to a total of 136 articles. For the analysis of countries/regions, publications were attributed to all countries/institutions represented in their author affiliations, meaning that a publication involving authors from multiple countries/institutions could be counted for each corresponding country/institution. The top five countries in terms of published papers were China, the United States, India, Poland, and Canada, accounting for 44.9%, 11%, 10.3%, 9.6%, and 9.6% of the publications associated with each country, respectively. These percentages are not mutually exclusive because a publication may be associated with multiple countries through international collaboration. The output of China in this field was significantly higher than that of other countries, approaching half of the total publications. Focusing on China, Canada, and the United States as central nodes in the national cooperation network revealed strong academic exchanges with numerous collaborative countries, including Switzerland, India, Russia, and Japan. The color coding of nodes and lines indicated that the United States and Canada initiated relevant research earlier. In terms of institutional cooperation, Donghua University demonstrated the most significant influence, collaborating with institutions such as Zhongyuan University of Technology, Swiss Federal Institutes of Technology, Iowa State University, and Cornell University. The primary research topics included the impact of cylindrical geometric shapes on the thermal properties of flame-retardant fabrics [25,26], burn prediction [27,28], and heat stress [29]. Overall, while the research topics both domestically and internationally were diverse, there was a need for enhanced collaboration.

Detailed information regarding the first seven countries in terms of frequency, centrality, and bursts is presented in

Table 2. Top seven contributing countries of NSTPC.

Frequency	Countries	Centrality	Countries	Bursts	Countries
61	China	0.33	China	3.94	USA (2003–2006)
15	USA	0.22	USA	3.82	Canada (2008–2013)
14	India	0.15	Switzerland	2	Poland (2017–2018)
13	Poland	0.09	Canada	1.49	Saudi Arabia (2020–2022)
13	Canada	0.09	France	1.3	Slovenia (2008–2009)
7	Portugal	/	/	1.06	Portugal (2020–2023)
5	Switzerland	/	/	0.98	Australia (2021–2022)

. The United States ranked highly in both centrality and bursts, with values of 0.22 and 3.94, respectively. This underscored the significant role of the United States in NSTPC, primarily focusing on TPP [30,31] and heat transfer simulation [32,33], while also engaging in the development and design of intelligent TPC [34,35]. China held the top position in centrality, with a centrality value of 0.33, and experienced rapid development with a total of 61 published papers. The most highly cited articles focused on simulating energy transfer and accumulation of fabrics under low radiation conditions [36], investigating the heat and moisture transfer mechanisms under varying moisture distributions [37], and examining the protective effects of embedding phase change materials (PCM) of different thicknesses and positions [38]. Canada ranked second in bursts, with a value of 3.82, and one of its primary research focuses was on heat transfer modeling of thin fiber fabrics under high heat flux [39].

An analysis of the top ten institutions, ranked according to frequency, centrality, and bursts, is provided in

Table 3. Top 10 contributing institutions of NSTPC.

Frequency	Institutions	Centrality	Institutions	Bursts	Institutions
33	Donghua University	0.02	Donghua University	3.69	University of Saskatchewan (2009–2013)
9	Indian Institute of Technology System	0.02	Soochow University	3.44	North Carolina State University (2004–2006)
7	University of Saskatchewan	0.02	Hong Kong Polytechnic University	2.59	Warsaw University of Technology (2017–2018)
7	Lodz University of Technology	0.01	Zhejiang Sci-Tech University	2.21	Donghua University (2015–2018)
6	North Carolina State University	0.01	Iowa State University	2.05	National Research Council Canada (1998–1999)
6	Warsaw University of Technology	0.01	University of Alberta	2.02	National Institute of Technology (2023–2024)
5	Swiss Federal Institutes of Technology	0.01	Qingdao University	2.02	Soochow University (2023–2024)
5	Soochow University	/	/	2.02	National Institute of Technology Silchar (2023–2024)
5	Zhejian Sci-Tech University	/	/	1.87	Zhongyuan University of Technology (2008–2009)
5	EMPA	/	/	1.6	Indian Institute of Technology System (2016–2017)

. Donghua University in China ranked first with 33 published papers. The remaining top six institutions, which were notable for their publication output, included the Indian Institute of Technology System (9), the University of Saskatchewan (7), the Lodz University of Technology (7), North Carolina State University (6), and Warsaw University of Technology (6), representing India, Canada, Poland, the United States, and Poland, respectively. This pattern matched the evaluation of the primary contributing countries. The University of Saskatchewan exhibited the highest bursts (3.69), with a burst period from 2009 to 2013, primarily focusing on AG [40,41,42]. Following closely was North Carolina State University (3.44), experiencing a burst period between 2004 and 2006.

5.2. Contributing Authors

The academic achievements of core authors are often associated with higher citation frequencies. We can identify high-impact authors by analyzing the co-citation author network. The minimum spanning tree algorithm was utilized to prune the network, and we ultimately generated a co-citation author network map comprising 350 nodes and 1910 connections (Figure 7). The three clusters containing the most cited authors were as follows: #0 simulations, #1 thermal response, and #2 firefighter garments. These clusters exhibited large silhouette values of 0.978, 0.925, and 0.971, indicating good cluster cohesion and separation. A partial enlarged view of the main cluster revealed that Torvi, Song, and Tian were among the most frequently co-cited authors in studies of simulations. Analysis of the CiteSpace summary report showed that several highly cited authors were distributed across different clusters. Chitrphiromsri and Su showed strong co-citation relationships in studies investigating the effects of heat and moisture transfer on TPP. Key research focuses included heat and moisture transfer, fabric property characterization, and skin burns. Li, Zhang, Talukdar, and Sawcyn were more frequently involved in studies under radiative heat exposure, which likely reflects the important role of radiative heat transfer in actual fire environments [43].

The cooperation network (Figure 8) reveals that Li Jun, Song Guowen, Tian Miao, and Su Yun serve as the central figures within this network, exhibiting the closest collaborative relationships. Among these scholars, Li Jun stood out with the highest publications (26). Following closely was Tian Miao, who had published 12 papers. Su Yun, Li Jun and Zhang Xianghui demonstrated relatively high centrality scores of 0.03 (

Table 4. Top 10 contributing authors of NSTPC.

Freq	Author	Centrality	Author	Bursts	Author
26	Li, Jun	0.03	Su, Yun	2.61	Chitrphiromsri, P
12	Tian, Miao	0.03	Li, Jun	2.53	Bergstrom, Donald J
11	Su, Yun	0.03	Zhang, Xianghui	2.37	Furmanski, Piotr
9	Ghazy Ahmed	0.02	He, Jiazhen	2.37	Lapka, Piotr
8	Song, GW	0.02	Rossi, Rene M	2.26	Udayraj
8	Das, Apurba	0.01	Song, Guowen	2.15	Acharya, Jnanaranjan
6	Furmanski, Piotr	0.01	Udayraj	2.05	Torvi, DA
6	Torvi, DA	0.01	Wang, Faming	2	Bhanja, Dipankar
6	He, Jiazhen	0.01	Xiang, Chunhui	1.95	Kuznetsov, AV
6	Lapka, Piotr	0.01	Wang, Zhaoli	1.73	Talukdar, Prabal

), signifying the critical importance of their research contributions. The research topics explored by Su Yun encompassed safety thermal exposure distance assessment, radiation heat transfer simulation, and heat and moisture transfer simulation. Li Jun, Tian Miao, and Su Yun are all affiliated with Donghua University, fostering close collaboration. A highly cited document was related to the coupled model [44]. He Jiazhen and Zhang Xianghui maintained close collaborative relationships with these three scholars. On the left side of the network, Song Guowen and other researchers constituted a significant collaborative network. Their research concentrated on the numerical simulation of heat and moisture transfer [18,45,46], TPP [47], and intelligent fire protective clothing [34]. As shown in Table 4, Udayraj also exhibited a relatively high burst strength. Udayraj, Das Apurba and Talukdar Prabal, from the Indian Institute of Technology, primarily investigated AG dynamics [48,49], coupled computational fluid dynamics (CFD)-radiation heat transfer models [50], fabric property analysis [51] and heat and moisture transfer modeling [52].

5.3. Intellectual Basis Based on the Cited References

A particular research domain may be understood as a temporal mapping connecting the research frontier to the knowledge basis [8]. By analyzing the references of publications obtained from the database, we can gain insights into the knowledge base related to NSTPC. Interpreting articles that exhibit significant bursts can enhance our comprehension of the research focus and emerging topics.

We selected the top 50 cited documents from each time slice to construct a co-citation literature network by utilizing the extracted 2157 references (Figure 9), comprising 402 nodes and 1372 connections. The LLR algorithm was employed to extract keywords and generate cluster labels, thereby illustrating cluster topics. The modularity (0.8149) and the silhouette (0.9324) were both relatively high, indicating a clear cluster structure and significant homogeneity. The silhouette for each cluster exceeded 0.8, demonstrating that the cluster was reasonable.

Cluster firefighters clothing (#0) was the largest cluster, comprising 54 references with an average citation year of 2013. The highly cited documents within this cluster primarily investigated the AG model, fabric thermal shrinkage, and energy transfer models. Cluster flash fire exposure (#1) ranked as the second largest cluster, containing 47 references. Its high silhouette (0.98) indicated an exceptionally high level of homogeneity, with a focus on heat transfer during thermal exposure. Ghazy et al. [40] examined the coupled conduction-radiation heat transfer within each AG of the clothing system during flash fire, as well as the changes in thermal performance during exposure and subsequent cooling. Cluster #0 exhibited significant overlap with the cluster flash protective exposure (#1), cluster thermal protective clothing (#2), and cluster human body (#10). In numerical simulations, a manikin is commonly employed to assess the protective performance of TPC in practical scenarios. An unsteady CFD model based on a flame manikin was established by Li and Tian [53] to examine the impact of wind speed on flame shape and heat transfer processes. There was also considerable overlap between the cluster radiant source (#3) and cluster unexpected burn (#8), which primarily focused on the thermal environment, firefighter occupational safety, and burn injuries. Su et al. [15] formulated models for heat and moisture transfer in clothing, skin heat transfer, and skin burn prediction to assess firefighters’ minimum exposure time in thermal hazards, thereby enhancing firefighter safety.

5.4. Keyword Co-Occurrence Analysis for Research Hotspots and Frontiers

5.4.1. Analysis of Critical Keywords

The co-occurrence analysis of keywords provides a more intuitive representation of high-frequency keywords, facilitating researchers in analyzing the research hotspots and frontiers within this domain.

Table 5. Top 10 keywords of NSTPC.

Freq	Keywords	Centrality	Keywords	Bursts	Keywords
55	Heat transfer	0.29	Heat transfer	3.89	Flash fire exposure (2015–2018)
42	Numerical simulation	0.27	Model	3.21	Air gaps (2016–2018)
33	Performance	0.19	Protective clothing	2.95	Thermal protective performance (2023–2024)
31	Moisture transfer	0.16	Performance	2.76	Numerical simulation (2016–2017)
28	Firefighter protective clothing	0.15	Numerical simulation	2.25	Transport (2023–2024)
27	Model	0.15	Moisture transfer	2.13	Skin burn injury (2016–2017)
25	Air gaps	0.14	Fabrics	1.99	Moisture transfer (2015–2017)
25	Fabrics	0.14	Flash fire	1.96	System (2011–2013)
24	Flash fire	0.11	Flame resistant fabrics	1.94	Flash fire (2015–2016)
22	Protective clothing	0.09	Thermal protective performance	1.71	Natural convection (2017–2018)

presents the top 10 keywords in NSTPC, ranking according to frequency, centrality, and bursts. Heat transfer appeared with the highest frequency at 55 occurrences, followed by numerical simulation (42). Heat transfer represented the fundamental and primary research focus of NSTPC. Investigating the heat transfer characteristics of TPC can elucidate the heat transfer mechanisms, thereby offering a scientific basis for research and development of TPC, enhancing TPP, and minimizing human thermal load. NSTPC primarily focused on the fabric level. Simulating fabrics, as opposed to entire garments, reduces model complexity. The characteristics and structure of fabrics significantly influence the TPP of clothing. Consequently, the TPP of clothing can be approximately characterized through fabrics. Additionally, fabric characteristics considerably affect the accuracy of the model. Therefore, “performance” (frequency: 33) and “fabrics” (frequency: 25) were fundamental objectives and mediums in NSTPC.

The bursts of “flash fire exposure” were the highest, recorded at 3.89. Although the duration of flash fire exposure is very brief, lasting only a few seconds, its intensity is significant. During a flash fire, the flames spread rapidly, posing serious risks to firefighters. When the temperature at the basal layer (interface between the epidermis and the dermis layers) reaches 44 °C, skin burns could occur [40]. Fabrics can result in varying degrees of skin burns during thermal exposure and cooling. Consequently, researchers placed considerable emphasis on the study of burns. Burn grade and burn distribution are critical indicators for assessing the performance of TPC and hold considerable importance for the design of such clothing and the establishment of safety standards.

“Moisture transfer” (frequency: 31; bursts: 1.99) and “air gaps” (frequency: 25; bursts: 3.21) were significant research hotspots. The presence of sweat and environmental moisture substantially influences the TPP of fabrics. The presence and redistribution of moisture can alter the effective thermal properties and heat-transfer behavior of porous textile systems, which may affect the temperature distribution within the clothing system, potentially resulting in thermal damage to the human body. The research concentrated on heat and moisture transfer under various thermal exposure conditions, including flash fire, high-intensity radiation, low-intensity radiation, and high-temperature steam. AGs situated at the fabric-fabric and fabric–skin interfaces are crucial for protective performance. The heat transfer through AG determines the exposure level experienced by the skin. Current modeling research on AG primarily investigated its position, thickness, width, and dynamic variations.

5.4.2. Temporal Distribution and Evolution of Keywords

The time-zone perspective indicates temporal relationships between a research frontier and foundational knowledge. The evolution of keywords in NSTPC across distinct periods is illustrated in Figure 10, which is categorized into three stages as outlined in Chapter Three. Prior to 2000, research primarily concentrated on the heat transfer properties of TPC subjected to flash fire exposure. Following 2000, there was a heightened focus on firefighter safety, leading to the application of CFD technology to simulate the thermal environment. Concurrently, the concept of PCMs emerged, which were capable of absorbing or releasing heat at their phase change temperature. As a class of functional thermal-regulating materials, PCMs can be incorporated into protective coatings or surface-functionalized layers to regulate interfacial heat transfer. The integration of PCMs into TPC enhances the heat transfer performance and provides additional thermal protection for the wearer. Hence, a model combining heat and moisture transfer with fabrics and PCMs was developed [54]. By approximately 2006, research began to focus on predicting the extent and duration of skin burns [55] and the mechanisms of moisture transfer [45].

Between 2010 and 2016, the keywords “moisture transfer”, “CFD simulation”, “flame resistant fabrics”, “exposure”, and “protective performance” emerged as high-frequency terms. These keywords centered on TPP optimization, encompassing core research areas from the mechanisms of heat and moisture transfer to material design and the effects of environmental exposure. Researchers have focused on selecting various flame-retardant and fireproof fabrics or materials to develop heat and moisture transfer models in multiple thermal environments, enabling predictions of temperature and humidity distributions, heat distribution, skin burns, and vapor density distributions. During this period, modeling techniques including the finite element method and the finite volume method gained prominence, while research on CFD simulations and PCMs continued. The flame manikin found extensive application as a primary method for model establishment and validation.

After 2016, “air gap” and “heat and moisture transfer” became the research themes, focusing on the static and dynamic changes of AG, the impact on the total energy stored within the fabric system, and both closed and open AGs [56]. Studies also explored the effects of moisture on the TPP of firefighter clothing containing PCMs. A model considering all major heat and moisture transfer mechanisms was established during this period [16].

5.4.3. Knowledge Structuring of NSTPC Research

To achieve a more systematic understanding of the application of numerical simulation methods of TPC, a cluster analysis of keywords was conducted. The analysis revealed ten distinct clusters, which include the following: #0 dry heat loss, #1 moisture transfer model, #2 extended model, #3 skin model, #4 fabrics, #5 protective suit, #6 heat transfer model, #7 multilayer protective clothing, #8 firefighter, and #9 exposure environment. Based on the co-occurrence and cluster results of these keywords, as well as the overall knowledge structure of this research area, the ten clusters were categorized into four groups: TPC, wearer, thermal exposure environment, and model characteristics (Figure 11).

The first category was related to model characteristics, including clusters #1, #2, #3, and #6. Initially, the focus was on a 1D heat transfer model for single-layer fabrics, which simulated the mechanisms of heat conduction and radiation under high-temperature conditions. The thermal physical properties, such as the fabrics’ heat conduction coefficient and specific heat capacity, were determined experimentally to analyze the thermal exchange boundary conditions [55]. By simulating parameters including the epidermis, dermis, subcutaneous tissue, and blood flow of human skin, a skin heat transfer model was established to predict burn distribution and duration. Air, being an excellent thermal insulator, effectively isolates external heat transfer to the human body, thereby enhancing the thermal insulation performance of clothing. A heat transfer model encompassing clothing, AG and skin was developed to examine the effects of parameters encompassing thickness, location, heterogeneity, orientation, and dynamics of AG. The findings indicated that AG increased the total energy stored in the fabric system and the percentage of energy released to the skin tissue [36]. During the cooling stage, the accumulated heat is released, resulting in an increase in skin temperature that exacerbates skin burns. Hence, a quantitative analysis of the relationship between heat transfer and heat accumulation, along with the proportion of accumulated heat transferred to human tissues and the effects of various parameters, is significantly important for reducing skin burns and optimizing the design of protective clothing. Moisture alters the thermal characteristics of fabrics. For instance, during the heat exposure stage, the evaporation and vaporization of moisture dissipate some of the heat, thereby positively impacting thermal protection. However, during the cooling stage, the presence of moisture may impede heat release, resulting in higher skin temperature. Investigating the influence of humidity on TPP through pre-soaking technology, as well as developing moisture transfer models and coupled heat and moisture models, were also key research areas in this field.

The second category pertained to TPC, encompassing clusters #4, #5, and #7, which emphasized the investigation of the TPP of fabrics and clothing. Initially, the focus was primarily on the basic modeling and performance assessment of flame-retardant fabrics and firefighting clothing materials. To enhance model accuracy, research on fabric performance parameters became more comprehensive, considering various influencing factors such as fiber moisture regain, fabric thickness, specific heat capacity, extinction coefficient, surface emissivity, fabric deformation, thermal resistance, and moisture resistance [57]. A cylindrical test device and a skin simulation sensor were developed to account for the influence of fabric shrinkage during heat transfer measurements [26]. Porous materials exhibit very low thermal conductivity, and their unique continuous network of interconnected pores effectively inhibits heat conduction, thereby enhancing the thermal insulation properties of clothing. This material constitutes a multiphase mixture of fibers, water in three phases, and air. The thermal physical properties of the fabric may be influenced by the moisture present in the porous material [58]. Thus, the study of heat and moisture transfer mechanisms in porous materials via numerical simulation is of significant importance.

PCMs, which absorb heat during phase transition, effectively extend the duration of heat transfer to the skin, thereby enhancing TPP [59]. The thermal protective effect was strongly influenced by the textile latent heat and PCM melting/solidification temperature [60]. Research demonstrated that the integration of microencapsulated phase change materials into multilayer clothing systems can significantly alter temperature distributions and heat flux transfer within the fabric assembly [61]. Su et al. [62] developed flame-retardant PCM-coated fabrics using a dry-coating process and found that the thermal protective performance of the clothing system was influenced by the PCM type, phase change temperature, and the position of the PCM-coated fabric within the protective clothing system. These investigations highlighted the potential of PCM-based functional treatments for regulating coupled heat and moisture transfer and mitigating thermal hazards in protective clothing, thereby promoting the development of advanced materials for TPC. From a numerical modeling perspective, coating thickness, thermal conductivity, emissivity, and interfacial thermal resistance can be incorporated as layer-specific thermophysical and boundary parameters to quantitatively characterize their effects on heat transfer. For PCM-containing coatings, phase change temperature and latent heat can be introduced through temperature-dependent enthalpy or specific heat formulations, providing a basis for predicting their transient thermal protection behavior [60]. The validation of simulations regarding the TPP of fabrics and garments was primarily conducted through experimental testing, which included empirical analyses such as bench tests and mannequin tests. Clusters #4, #5, and #7 offered crucial reference values for the development and optimization of TPC.

The third category pertained to the environment, focusing on the thermal exposure conditions, including cluster #9. Under varying thermal exposure conditions, the predominant methods of heat transfer differ, leading to significant variations in heat and moisture transfer mechanisms, temperature distribution, and skin burn severity. Precisely assessing the safe thermal exposure distance is essential for protecting fire rescue personnel. Su et al. [63] determined the safe thermal exposure distance in thermal radiation environments by simulating the heat transfer process through fire scenes, clothing, AG, and human skin, providing theoretical guidance for fire safety operations. CFD was extensively utilized in fire disaster simulations, capable of modeling phenomena such as airflow, temperature distribution, and smoke propagation resulting from fires. Hence, this capability allowed for the prediction of fire development trends and impacts, providing a critical foundation for fire safety assessment and prevention.

The fourth category was related to the human body and encompasses two clusters: #0 and #8. Occupational health and safety concerns for workers engaged in high-temperature operations have garnered significant attention. The focus of research in this area included dry heat loss, human physiological thermoregulation, and skin burns. Dry heat loss primarily occurs through three mechanisms: conduction, radiation, and convection, which are also central to numerical simulation studies. Human thermal models and skin models are employed to calculate heat exchange within the human body and between the body and the thermal environment, facilitating the evaluation of physiological responses and heat stress of firefighters in such environments [29]. During firefighting work, physical activity caused clothing to shift periodically relative to the body, thereby impacting the TPP of garments [49,64,65].

6. Development Trajectory of NSTPC

The developmental trajectory of NSTPC was delineated through in-depth literature reading and bibliometric analysis (Figure 12). The technical characteristics, research hotspots, and frontiers of NSTPC across various stages were clearly defined, thereby elucidating the evolutionary trajectory of NSTPC. The initial stage represented the foundational phase, during which the 1D heat transfer model of the fabric–AG–skin system, accounting for flash fire conditions, was predominantly utilized, while 2D models remained relatively scarce. Preliminary investigations into the thermal physical properties of fabrics [47,55], AG [47], and moisture transfer [66] were conducted during this phase; however, these models generally overlooked critical factors such as thermal decomposition, mass transfer, and optical characteristics of the fabrics. The subsequent stage was characterized by model development. Variations in the human body’s clothing form led to non-uniform temperature distribution on the fabric’s inner surface and AG characteristics. Three-dimensional full-scale simulations based on CFD were increasingly explored to account for the complexity of clothed human systems. During this period, the AG model underwent extensive investigation, examining the effects of changes in thermal performance of AG [67], coupled conduction-radiation heat transfer [41], and dynamic variations [64] on clothing performance. The third phase, corresponding to the prosperity stage identified from the publication and citation trends, involved a variety of modeling approaches adopted for different research purposes. Research primarily focused on multilayer fabrics, with increasing studies on fabric thermophysical properties [68,69], exposure conditions [70,71,72,73,74], AGs [75,76,77], moisture [78,79], and other parameters. Research on moisture transfer was mainly based on 1D models, while 3D models were primarily used to study heat transfer. Notably, 3D full-scale models demonstrated a developmental trend encompassing fire scene environment simulation [80], heat transfer simulation of naked manikins [81], heat transfer simulation of a flame manikin clad in close-fitting clothing [82], and parametric analysis of various factors using clothed manikins [22,53,83,84,85].

7. Future Development Directions

Despite the significant progress achieved in the development of NSTPC, several critical issues remain that require urgent resolution. With the rapid advancements in artificial intelligence, high-performance computing, and new material technologies, the future of NSTPC is expected to integrate these emerging technologies. The research focus will be directed towards the following aspects:

(a)

Integrated models consider a variety of factors. During the practical application of TPC, it needs to account for heat transfer under high-temperature exposure, moisture evaporation and absorption, mechanical interactions between the clothing and the body, and chemical erosion in certain specialized environments. In the future, greater emphasis will be placed on the coupled simulation of multiple physical fields, including heat, humidity, force, and chemical interactions. To gain a deeper comprehension of clothing performance under complex conditions, simulations can be used to investigate the effects of relevant physical and chemical factors and material properties on the thermal protective performance of clothing under high-temperature exposure. Additionally, the accuracy of models that account for the performance parameters of clothing materials and complex AGs requires enhancement, and the development of full-scale models that encompass these intricate factors is essential.

(b)

Realistic simulations of thermal exposure scenes and human movement. Thermal environments frequently undergo dynamic changes, such as the spread of fire at a fire scene and fluctuations in temperature. Real-time or transient simulations may provide a more realistic representation of the evolution of thermal exposure and clothing performance at different stages. Additionally, factors such as postural changes and muscle activities during movement significantly influence clothing performance. The consideration of thermal exposure scenarios alongside the dynamic changes of the human body may provide useful information for optimizing clothing structural design and evaluating clothing performance and comfort.

(c)

The optimization of TPC performance and the simulation of functional material integration. The continuous development of advanced thermal protective materials, including nanomaterials, PCMs, and smart temperature-regulating materials, requires enhanced numerical simulations to investigate their thermal transport characteristics and protective mechanisms. The integration of functional coatings and surface-modified layers into multilayer thermal protective systems provides new opportunities for improving thermal barrier performance and regulating interfacial heat transfer. By simulating thermal conduction, thermal radiation, moisture transport, and other relevant properties under high-temperature environments, researchers can establish a theoretical foundation for the application and optimization of these novel materials and coating systems in TPC. It is essential to explore innovative protective structures, such as multilayer composite architectures and modular designs, with particular attention to material interfaces and thermal resistance.

(d)

The integration of artificial intelligence and machine learning with numerical simulation. Artificial intelligence and machine learning can provide new approaches for improving the computational efficiency and optimization capability of NSTPC models. Recent studies have demonstrated that surrogate models can be constructed from numerical simulation data to rapidly predict the thermal protective performance of multilayer fabric systems and subsequently support multi-objective optimization of fabric parameters [69]. Physics-informed neural networks (PINNs) provide another potential approach by incorporating governing heat-transfer equations and boundary conditions into the learning process, which can improve physical consistency and reduce reliance on extensive training data [86].

8. Conclusions

A comprehensive statistical analysis, in-depth literature reading, and bibliometric evaluation of 136 relevant documents retrieved from WoS Core Collection spanning from 1998 to 2026 were presented in this study. The current development status, theoretical foundations, collaborative modes, research hotspots and frontiers, as well as the developmental trajectory in the field of NSTPC were thoroughly examined.

From 1998 to 2026, NSTPC exhibited a rapid growth trend, encompassing various fields such as engineering, materials science, and thermodynamics. The predominant citing subject fields encompassed physics, materials, and chemistry, followed by mathematics, systems, and mathematical. Notable representative journals within these fields included International Journal of Thermal Sciences, Journal of the Textile Institute, and Fire Technology. A thorough review of the literature revealed that NSTPC is becoming increasingly comprehensive. The analysis revealed a trend toward the development of models with increasing dimensionality, system complexity, and multi-physics coupling, while different modeling approaches continue to coexist and serve different research purposes. Keyword analysis indicated that topics such as heat transfer, moisture transfer, flash fire exposure, AGs, and skin models garnered significant attention within the NSTPC field. The application of advanced new materials emerged as a research hotspot in this domain, aimed at the design and optimization of TPC. Through cluster analysis, a knowledge structure was developed, encompassing model characteristics of NSTPC, the human body, development and design, performance evaluation, and thermal exposure scenarios. By integrating interdisciplinary approaches, future research could explore incorporating complex material properties and surface/interface effects into numerical models and place greater emphasis on performance evaluations in complex environments, thereby advancing the development of TPC towards greater efficiency, intelligence, and comfort. The results of this paper establish a theoretical basis for NSTPC, as well as theoretical support for the design, development, and safety assessment of such garments.