A BIM-Based Multi-Criteria Spatial Framework for Assessing Fire Risks in Indoor Environments

Abstract

Building fires are considered major disasters because of their significant effects on people, property, and the environment. This understanding has led to increased attention on developing preventive measures, particularly through the creation of effective methods for assessing fire risk. However, the effectiveness of these methods relies heavily on detailed physical and functional information of the building and data-driven decision-making. Building Information Modeling (BIM) has proven effective in representing structures, even in three dimensions. When integrated with Geographic Information Systems (GIS), it enhances spatial intelligence, leading to improved decision-making through robust multi-criteria approaches. Hence, this study develops a framework to assess fire risk in an indoor environment that deploys a BIM-based GIS and Multi-Criteria Decision-Making; this is specifically known as Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The developed framework consists of four steps: identifying fire risk parameters, calculating weights, conducting spatial fire risk assessments, and visualizing the results, where the developed concepts are tested and validated. According to the significant findings, the developed framework estimates that 18% of building rooms are at moderate risk, while the compared model identifies only 1%. This considerable difference could potentially arise from the detailed data structure of BIM and the spatial insights gained from GIS. By implementing the designed framework, key fire risk factors can be identified in three dimensions, accompanied by a comprehensive quantitative evaluation platform for fire risks within indoor environments.

1. Introduction

In complex urban environments, individuals often spend a considerable amount of time indoors, which heightens the risk of incidents like fires. These events can result in loss of life, damage to infrastructure, and substantial economic losses. Consequently, numerous studies are being undertaken to enhance incident management and mitigate the effects of such disasters. Fire risk assessment methods are categorized into three types: qualitative, quantitative, and semi-quantitative [1]. Qualitative assessments quickly evaluate fire hazards using descriptive measures and commonly employ methods such as checklists and event trees [2]. In contrast, quantitative assessments rely on numerical data, engineering principles, and software to analyze fire risks, making them more labor-intensive. Examples include FIRECAM, the FIERAsystem, and CRISP [3,4,5]. Furthermore, semi-quantitative methods assess the probability or consequences of fire hazards using techniques such as fuzzy comprehensive evaluation, Fire Risk Analysis Method for Engineering (FRAME) and the Analytical Hierarchy Process (AHP) [2,5]. In the realm of fire risk assessment, Two-Dimensional (2D) drawings, tabular data, building codes, and the advice of experts regarding buildings are commonly utilized [6,7]. However, the limitations of this type of information in terms of granularity can hinder a thorough evaluation of potential fire risks. As a result, it is essential to obtain comprehensive and detailed information regarding the building. This could be achieved by employing Building Information Modeling (BIM). BIM technology includes the physical and functional characteristics of buildings, such as walls, doors, and windows. It can generate three-dimensional (3D) digital models of buildings in virtual environments. This technology offers valuable insights for the analysis and management of fire safety performance in buildings [8]. Moreover, the implementation of BIM allows for the correction of discrepancies between 2D and 3D drawings. Additionally, BIM facilitates a more comprehensive visualization of building surroundings and facility locations, thereby enhancing the effectiveness of fire management strategies when compared to traditional 2D tools [9]. Therefore, BIM could be employed during the design and operation phases of buildings to manage fire risks [10]. In BIM, the Level of Detail (LoD) indicates the sophistication and accuracy of the information presented in the model. It defines the extent of detail included, from basic representations of elements to highly detailed and precise models [11]. BIM has recently been applied in the field of building fire research. Its applications include areas such as building evacuation [12,13], emergency management systems [14], fire prevention systems [15], and safety [8,16,17]. Despite its significant potential for fire risk assessment, there are few studies in the literature that specifically explore the use of BIM for this purpose. One of the most significant study conducted by Wang et al., analyzed the BIM model of a post office and employed the FRAME to evaluate fire risks within the building [18]. Additionally, Hosseini and Maghrebi explored the risks associated with fire emergency evacuation in complex construction sites via integrating 4D-BIM, fire quantitative risk assessment, and simulations using the Social Force Model [19]. Still, the spatial characteristics of fire and smoke could not be modeled using a spatial database equipped with advanced decision-making tools. However, utilizing Geographic Information Systems (GIS) to manage, analyze, and manipulate spatial data, along with BIM, could address these issues. GIS plays a vital role, offering a powerful tool to tackle various challenges associated with emergency management [20]. Fires possess spatial characteristics, enabling effective modeling, mapping, identification, management, and anticipation of fire incidents through GIS [21]. Therefore, integrating GIS with BIM could yield enhanced results in fire risk assessment thanks to GIS’s robust ability to analyze spatial relationships between different objects. Similar to most BIM studies, many of the research efforts involving GIS have primarily focused on fire safety and evacuation scenarios [22,23,24]. Multi-criteria decision-making (MCDM) is a vital aspect of GIS as it involves the evaluation of various alternative solutions based on multiple qualitative and quantitative criteria. In MCDM processes, decision-makers define and prioritize these criteria to assess the different options available [25]. The integration of GIS into MCDM enhances the decision-making process by providing spatial context to the criteria being evaluated. There is a considerable body of literature that examines the integration of GIS and MCDM methodologies to evaluate and enhance fire safety [26,27,28]. In this context, one of the prominent and used MCDM techniques is the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The TOPSIS method is usually utilized in the fire risk assessment field for calculating fire risks [29] and establishing evaluation models [30]. Moreover, the majority of the studies in the literature do not incorporate BIM with GIS [31]. While some studies utilize BIM, they often fail to integrate it with GIS [9,13].

Consequently, despite various studies conducted in the literature, there are still unresolved issues in the area of fire risk assessment for buildings: (a) The majority of existing research on fire risk assessment predominantly relies on 2D drawings and tabular data, and while BIM has been widely utilized across various domains of fire studies, its application in fire risk assessment remains limited. (b) Various parameters have been deployed in fire risk assessment; however, since their appropriateness heavily depend upon availability of spatial data, several granularities and limitations exists. (c) Research on the use of spatial analyses in assessing fire risk is currently limited. Nevertheless, the integration of GIS offers the potential to conduct comprehensive spatial analyses, thereby enhancing the effectiveness of BIM.

This study seeks to develop a novel fire risk assessment system for buildings by integrating the TOPSIS method with BIM and GIS. It focuses on evaluating fire risks within a complex public building that contains various functional spaces as part of a MCDM process. By combining GIS and BIM, this research aims to deliver a precise and comprehensive assessment of fire risks. BIM is used to support an in-depth evaluation of fire risks in the study area, while GIS facilitates the analysis, evaluation, and visualization of the findings. The framework is designed to achieve the following objectives: (a) the effectiveness of fire risk assessments is improved through the application of BIM, (b) the accurate identification of fire risk parameters using detailed BIM, and (c) the integration of BIM and GIS.

This research seeks to develop a fire risk assessment system for buildings using the TOPSIS method. The system is first implemented in the Civil Engineering Faculty building at Istanbul Technical University (ITU). Afterward, the results obtained are compared with those from another fire risk assessment system found in the literature. The aim of the study is to provide a precise and comprehensive evaluation of fire risks by effectively integrating GIS and BIM within the developed system. This study addresses challenges such as insufficient data, the identification of key fire risk parameters, and the consideration of spatial relationships among various objects through the application of BIM and GIS environments. These environments are essential for conducting a thorough and effective fire risk assessment, and provide reliable information to stakeholders.

2. Data and Methodology

This study involves four steps, including (a) identification of fire risk parameters, (b) calculation of parameter weights, (c) GIS-based fire risk evaluation, and (d) visualization of risks in the BIM environment. Initially, the data and the study area are described. The Civil Engineering Faculty of ITU in Istanbul, Türkiye, was selected to test and verify the developed fire risk assessment concept. The building contains a total of 257 rooms, the majority of which are designated as offices, classrooms, and laboratories. The analysis concentrates on three distinct levels, which includes the first and second underground floors as well as the first floor in order to assess both high and low fire risk potentials. The second underground floor has 38 rooms, predominantly laboratories, including the soil laboratory, soil mechanics laboratory, and analysis room, all of which are equipped with materials that have the potential to cause a fire. The first underground floor consists of 32 rooms, primarily laboratories such as molecular biology laboratories, a static soil mechanics laboratory, and rooms for chemical processes, all of which could also pose a fire risk. On the first floor of the building, there are 45 rooms, the majority of which are designated for academicians, presenting a comparatively lower risk in the event of a fire within the building. All floors are equipped with two emergency exits, a venting system, and various fire protection systems including fire alarms, fire extinguishers, smoke detectors, and a sprinkler system.

To evaluate fire risks in the rooms of the study area, BIM is employed for its precision and comprehensive information. BIM objects vary in their levels of detail, offering different degrees of granularity in both geometric and non-geometric attributes [11]. The LoDs range from LoD100, which provides a conceptual visualization, to LoD500, reflecting the as-built geometry. Specifically, LoD200 includes approximate geometry, LoD300 delivers precise geometry, LoD350 features detailed connections, and LoD400 signifies fabrication-ready models [32]. In the study, the BIM object was generated using 2D floor plans and material specifications of independent sections, alongside electrical and mechanical plans. The model was further refined through the incorporation of laser scanning and photogrammetric measurements to enhance accuracy and detail. The generated BIM for the study area has a LoD of 300, reflecting a high degree of detail for the objects modeled. This level is adequate for performing a fire risk assessment. At LoD 300, building elements are accurately represented with precise geometry and specific information, facilitating a thorough analysis of spatial configurations and associated fire risks. This BIM model contains comprehensive information about the building and its components including columns, floors, stairs, different types of walls and plaster, various flooring and coatings, as well as mechanical, electrical, lighting, ventilation, and fire prevention installations. Regarding the fire risk assessment, this detailed information about the building has vital benefits. The BIM of the study area is illustrated in Figure 1.

2.1. System Architecture

The system architecture is presented in Figure 2. The system architecture includes four steps: identification of fire risk parameters, calculation of parameter weights, GIS-based fire risk determination, and visualization of results. Each of these steps is briefly explained in the following sections.

The initial stage entails identifying fire risk parameters in the study area. In this context, fire risk parameters were identified by assessing the available information in the BIM model, along with a review of applicable standards, codes, and related literature. In this step, a review was conducted of the following standards and codes: NFPA 101: Life Safety Code [33], NFPA 101A: Guide on Alternative Approaches to Life Safety [34], ISO 3941 [35], IFC 2024 [36], and BS 9999:2017 [37]. Fire risk parameters were selected based on the detailed data from the BIM model. Additional relevant parameters were identified through a review of applicable standards, codes, and literature. Then, the weights of the selected fire risk parameters were established through a two-step process. First, the values for each fire risk factor associated with the individual rooms in the study area were normalized using the minimum–maximum normalization method. The necessary data was gathered using BIM in conjunction with Dynamo and Python. Second, these normalized values were employed to implement the Entropy Weight Method (EWM), an established objective weighting technique. Later, the BIM model of the study area was integrated with GIS. This integration allowed for the calculation of fire risks for each room within the study area using the TOPSIS approach. Finally, fire risk maps were created to illustrate the fire risk situation of the study area. Additionally, a semi-quantitative fire risk assessment method known as FRAME is used to validate the results obtained.

2.2. Identification of Fire Risk Parameters

The identification of parameters in studies varies by their specific aims and focus areas, creating ambiguity in fire risk assessment. The evaluation of fire risk is significantly impacted by insufficient data and the presence of numerous unclassified parameters. Consequently, there is no established methodology for selecting the fire risk parameters that should be utilized in the assessment process [6,26,29,38,39,40,41]. Nonetheless, this issue can be resolved by using BIM objects and GIS, which will provide high-resolution data and enable the analysis of fire risks through spatial relationships. For this purpose, based on the characteristics of a public building, along with a review of the BIM model of the building, literature, and relevant building standards and codes, namely NFPA 101: Life Safety Code [33], NFPA 101A: Guide on Alternative Approaches to Life Safety [34], ISO 3941 [35], IFC2024 [36], and BS 9999:2017 [37], the fire risk factors were determined, and a two-level building fire risk assessment system was established. This developed fire risk assessment system offers significant advantages. It does not rely on tabular values, facilitates spatial analysis, and delivers a comprehensive perspective of the study area. Hence, the proposed fire risk assessment system incorporates the essential characteristics of the building related to fire event within the specified study area. Afterwards, the fire risk assessment system developed in this study was tested in the study area.

2.3. Calculation of Parameter Weights

In the existing literature, various methodologies are used to determine the weights of fire risk parameters [6,26,29]. However, the statistical weights of these parameters are subject to uncertainties in their accurate calculation when these methodologies are used. Two main types of weighting are identified: subjective weighting and objective weighting. Subjective weighting relies on the insights and expertise of professionals, whereas objective weighting is based on quantitative computations [42]. Numerous studies utilize either one of these approaches or a combination of both. For instance, the research conducted by Zhang et al. employs both the Analytical Hierarchy Process (AHP) and the EWM, effectively integrating both subjective and objective weighting techniques [30]. However, they assign equal weight to each weighting method. This equal weighting may overlook the varying levels of importance among the criteria, potentially leading to less accurate or meaningful assessment outcomes. In another study, four factors were identified as influencing fire damage: door fire resistance, window layout, structural aging, and building type. A weighting factor of 10% was applied [8]. However, this oversimplifies their impacts and limits broader applicability. Conversely, the FRAME method is unable to incorporate weighting, as it relies solely on tabular values for fire risk parameters. The EWM method is employed in this study due to its effectiveness in objectively assessing the weights of fire risk parameters using the BIM model of the study area, which provides in-depth information about the building. This approach aims to eliminate any subjective judgments related to these weights.

The EWM is a technique based on the concept of information entropy often used in multi-criteria decision analysis to objectively determine weights or importance factors [43]. The EWM assigns weights based on the information contained in each index, eliminating human subjectivity. This makes it an objective approach to assigning weights. When the index data is objective, the calculation of weights is free from subjective bias [44]. The major steps of EWM can be stated as follows [29,43]:

• Generation of decision matrix X that consists of m decisions and n indexes.

  $$X=\left[\begin{array}{cccc}{x}_{11}& x_{12}& \dots & x_{1n}\\ x_{21}& x_{22}& \dots & x_{2n}\\ \dots & \dots & \dots & \dots \\ x_{m1}& x_{m2}& \dots & x_{mn}\end{array}\right]$$

  (1)
  In Equation (1), $x_{ij}$ (i = 1, 2, …, m; j = 1, 2, …, n) represents the value of the ith decision in the jth index.
• Normalization of values. The indexes have different values because they are measured in different units. It is important to normalize these values using methods such as Min–Max normalization, Z-score standardization, regularization, and mean valuation. Min–Max normalization scales data to a specified range to accommodate risk indicator data. Higher positive indicators indicate increased fire risk, while greater negative indicators reflect lower fire risk [44]. The minimum–maximum normalization equations are given below.
  For positive indicators:

  $$x_{ij}={\displaystyle \frac{x-x_{min}}{x_{max}-x_{min}}}$$

  (2)
  For negative indicators:

  $$x_{ij}={\displaystyle \frac{x_{max}-x}{x_{max}-x_{min}}}$$

  (3)
  In Equations (2) and (3), $x_{ij}$ and x, respectively, represent the normalized value and the value of the relevant data in the specific index, while $x_{max}$ and $x_{min}$ denote the maximum and minimum values of the specific each index.
• Calculation of the information entropy of each index (E_j).

  $$E_j=-{\displaystyle \frac{1}{ln\left(n\right)}}\sum _{i=1}^n{p}_{ij}\times ln{(p}_{ij});i=1,2,\dots ,n$$

  (4)
  In Equation (4), n indicates the number of alternatives.
• Calculation of the difference coefficient of each index ($g_j$) regarding the $E_j$ value of each index.

  $$g_j=1-E_j$$

  (5)
• Weight calculation for each index.

  $$W_j={\displaystyle \frac{g_j}{{\sum }_{j=1}^m{g}_j}}$$

  (6)

2.4. GIS-Based Fire Risk Determination

In order to conduct a thorough and accurate assessment of fire risks under a multi-criteria decision process, the study combined BIM with GIS. This integration facilitated the collection of spatial data and enhanced the analysis of spatial relationships between building components and fire-related factors. It supported comprehensive evaluations of fire risk, encompassing aspects such as smoke, and allowed for the effective visualization of results through mapping techniques. Consequently, it greatly enhanced the precision and depth of the risk assessment. Therefore, as an initial phase of the fire risk assessment, the BIM of the study area was integrated with GIS in this study. Moreover, in the context of fire risk assessment, the utilization of MCDM methods, GIS, mathematical models, and their integration holds substantial significance [45]. Especially within the advancements in fields such as GIS and data analytics, fire risk assessments have significantly improved in accuracy and comprehensiveness [46]. Therefore, in this study, to calculate fire risks in the study area, a GIS-based TOPSIS methodology is implemented. This approach enables a structured evaluation of fire risks based on the principles of MCDM.

TOPSIS is based on the concept that the best solution is the one that is closest to the positive ideal solution, maximizing the benefit criteria and minimizing the cost criteria, and farthest from the negative ideal solution, which maximizes the cost criteria and minimizes the benefit criteria. At the end of the TOPSIS, the alternatives are ranked in accordance with their distances from the ideal solutions [25,47]. The calculation steps of TOPSIS are given below [25,47,48]:

• Calculation of the normalized decision matrix: the original values are represented as $x_{ij}$, while the normalized values are shown as $r_{ij}$.

  $$r_{ij}={\displaystyle \frac{x_{ij}}{\sqrt{{\sum }_{i=1}^m{x}_{ij}^2}}};i=1,2,\dots ,I;j=1,2,\dots ,J$$

  (7)
• Calculation of the weighted normalized decision matrix. The weighted normalized values are indicated as $v_{ij}$.

  $$v_{ij}=w_j\times r_{ij};i=1,2,\dots ,I;j=1,2,\dots ,J$$

  (8)
  In these equations $w_j$ denotes the weight of the jth attribute and ${\sum }_{j=1}^n{w}_j=1$.
• Determining the positive ideal and negative ideal solutions.

  $$A^{*}=[v_1^{*},v_2^{*},\dots ,v_1^{*}]$$

  (9)

  $$A^{-}=[v_1^{-},v_2^{-},\dots ,v_j^{-}]$$

  (10)

  $$\begin{array}{cc}\hfill v_j^{*}& =max{v}_{ij}\mathrm{if}j\mathrm{is}\mathrm{a}\mathrm{benefit}\mathrm{attribute},min{v}_{ij}\mathrm{if}j\mathrm{is}\mathrm{a}\mathrm{cost}\mathrm{attribute};\hfill \\ \hfill v_j^{-}& =min{v}_{ij}\mathrm{if}j\mathrm{is}\mathrm{a}\mathrm{benefit}\mathrm{attribute},max{v}_{ij}\mathrm{if}j\mathrm{is}\mathrm{a}\mathrm{cost}\mathrm{attribute}.\hfill \end{array}$$
  The $A^{*}$ and $A^{-}$ denote the positive and negative ideal solutions in Equations (9) and (10).
• Calculation of the separation values. The separation value is used to find distances from each alternative to both the positive and negative ideal solutions. To calculate this value, the Euclidean distance theory is utilized.

  $$\mathrm{For} \mathrm{the} \mathrm{positive} \mathrm{separation} \mathrm{value}:s_i^{*}=\sqrt{\sum _{j=1}^J{\left(v_{ij}-v_j^{*}\right)}^2},i=1,2,\dots ,I$$

  (11)

  $$\mathrm{For} \mathrm{the} \mathrm{negative} \mathrm{separation} \mathrm{value}:s_i^{-}=\sqrt{\sum _{j=1}^J{\left(v_{ij}-v_j^{-}\right)}^2},i=1,2,\dots ,I$$

  (12)
• Calculation the relative closeness to the ideal solution. The relative closeness of alternative $A_i$ respective to $A^{*}$ is calculated as follows:

  $$c_i^{*}={\displaystyle \frac{s_i^{-}}{\left(s_i^{+}+s_i^{-}\right)}};0<c_i<1,i=1,2,\dots ,I$$

  (13)
• In the final step, the alternatives are ranked from highest to lowest based on their $c_i^{*}$ value.

3. Results and Discussion

3.1. Identification of Fire Risk Parameters

The fire risk parameters were selected by reviewing the BIM of the study area, standards, codes, and the literature, resulting in the development of a two-level building fire risk assessment system. The Level-1 parameters consisted of fire prevention factors ($P_1$), hazard factors ($P_2$), building characteristics ($P_3$), and human factors ($P_4$) in the developed fire risk assessment system. The sub-parameters of these Level-1 parameters are derived from studies in the literature, analyses of the BIM model, and codes regarding NFPA 101, NFPA 101A, BS 9999:2017, and IFC 2024. $P_1$ consists of fire protection systems, including a sprinkler system ($P_{11}$), smoke and heat sensors ($P_{12}$), distance to fire alarms ($P_{13}$), distance to fire extinguishers ($P_{14}$), and distance to emergency exit doors ($P_{15}$). $P_2$ consists of fire classes ($P_{21}$), fire loads ($P_{22}$), explosive materials ($P_{23}$), flammable materials ($P_{24}$), and linkage between venting shafts and rooms ($P_{25}$). The P3 category pertains to room characteristics such as room area ($P_{31}$), door width ($P_{32}$), floor type ($P_{33}$), and height of the room based on the ground floor ($P_{34}$). The $P_4$ category relates to human density in rooms ($P_{41}$) and the number of disabled people in the rooms ($P_{42}$). The schema of the two-level building fire risk assessment system is presented in

Table 1. Two-level building fire risk assessment system and associated units.

Level-1	Level-2	Unit
Fire Prevention Factors $\left(P_1\right)$	Sprinkler system ($P_{11}$)	score
	Smoke and heat sensors ($P_{12}$)	count
	Distance to fire alarms ($P_{13}$)	meter
	Distance to fire extinguishers ($P_{14}$)	meter
	Distance to emergency exits ($P_{15}$)	meter
Hazard Factors $\left(P_2\right)$	Fire classes ($P_{21}$)	score
	Fire loads ($P_{22}$)	Kcal/${\mathrm{m}}^2$
	Explosive materials ($P_{23}$)	score
	Flammable materials ($P_{24}$)	score
	Linkage between venting shafts and rooms ($P_{25}$)	score
Building Characteristics $\left(P_3\right)$	Room area ($P_{31}$)	${\mathrm{m}}^2$
	Door width ($P_{32}$)	m
	Floor type ($P_{33}$)	score
	Height of the room based on the ground floor ($P_{34}$)	m
Human Factors $\left(P_4\right)$	Human density in rooms ($P_{41}$)	people/${\mathrm{m}}^2$
	Number of disabled people ($P_{42}$)	count

.

The importance of parameters within the BIM model is crucial for conducting an effective and reliable fire risk assessment. In this context, the BIM model is utilized at a LoD of 300, which typically offers sufficient information for various geometry-based parameters. These include details about sprinkler systems, smoke and heat sensors, fire alarms, extinguishers, and exits, as well as measurements related to room area, door width, floor type, and room height. However, it is important to note that certain hazard-related factors, such as fire classes, fire loads, explosive and flammable materials, human density, and the number of disabled occupants, are not directly available with LoD 300. To address this gap, additional data integration or a higher LoD might be necessary to achieve a complete assessment. A systematic evaluation of these parameters at LoD 300 was performed, and a comparison with LoD 350 was made to identify potential data gaps with using 2024 LOD Specification of BIMForum [49]. A summary of the availability of Level-2 parameters at both LoD levels is provided in

Table 2. Data sources of the fire risk parameters.

Level-2 Parameter	Data Source
Sprinkler system ($P_{11}$)	BIM (LoD 300-350)
Smoke and heat sensors ($P_{12}$)	BIM (LoD 300-350)
Distance to fire alarms ($P_{13}$)	Spatial Analysis
Distance to fire extinguishers ($P_{14}$)	Spatial Analysis
Distance to emergency exits ($P_{15}$)	Spatial Analysis
Fire classes ($P_{21}$)	Standards and Codes
Fire loads ($P_{22}$)	Standards and Codes
Explosive materials ($P_{23}$)	In-Situ Analysis
Flammable materials ($P_{24}$)	In-Situ Analysis
Linkage between venting shafts and rooms ($P_{25}$)	Spatial Analysis
Room area ($P_{31}$)	BIM (LoD 300-350)
Door width ($P_{32}$)	BIM (LoD 300-350)
Floor type ($P_{33}$)	BIM (LoD 300-350)
Height of the room based on the ground floor ($P_{34}$)	BIM (LoD 300-350)
Human density in rooms ($P_{41}$)	BIM (LoD 300-350) and Expert Assessments
Number of disabled peoples ($P_{42}$)	BIM (LoD 300-350) and Expert Assessments

, highlighting areas where external data sources are required and indicating that parameters not included in the BIM can be obtained from relevant standards, codes, and literature specific to each room.

The fire prevention factors ($P_1$) encompass parameters relating to active and passive fire protection systems that are essential in the containment and suppression of fires. The building within the study area is equipped with a comprehensive fire protection system, including sprinklers ($P_{11}$) and smoke–heat sensors ($P_{12}$). Each room is outfitted with a specific number of sprinklers tailored to its size, ensuring comprehensive coverage. Furthermore, smoke sensors are installed in nearly all rooms, while heat sensors are strategically positioned in high-risk areas, such as rooms housing technical and electrical equipment. Furthermore, the building is equipped with multiple fire alarms ($P_{13}$), fire extinguishers ($P_{14}$), and emergency exits ($P_{15}$). Each floor is outfitted with five fire extinguishers, three to five fire alarms, and three to four exits, with two designated as emergency exits and the others serving as standard exit points. The study utilizes distances from each room to fire alarms, fire extinguishers, and emergency exits. The hazard factors ($P_2$) involve parameters that represent potential fire-causing factors within the building. Among these parameters is the classification of fire types ($P_{21}$), which encompasses classes A, B, C, D, and K as specified in ISO 3941:2007 [35]. The other parameter is the fire load ($P_{22}$) of each room, indicating the potential severity of a hypothetical fire event. The explosive materials ($P_{23}$) and flammable materials ($P_{24}$) point to the presence status of these types of materials in the rooms. The parameter $P_{25}$, which assesses the linkage between venting shafts and rooms, serves to indicate the status of the connection with the venting shafts. This connection, if compromised, may pose a fire risk on other floors and in additional rooms. Furthermore, the building’s characteristics ($P_3$) significantly influence fire events. The rooms in the buildings vary from offices to laboratories, playing a crucial role in both fire and smoke propagation and people evacuation. Therefore, the size of the rooms ($P_{31}$) is a vital building characteristic to consider in fire risk assessment. Additionally, door widths ($P_{32}$), strongly correlated with evacuation speed, are pivotal for safe and efficient evacuation. The study area encompasses various floor types ($P_{33}$) including epoxy, linoleum, polyurethane, sheet-filled polyurethane, and concrete. Epoxy and linoleum are predominantly found in offices and laboratories, while fire-protected rooms, like technical areas, are constructed with concrete. Lastly, the height of rooms from the ground floor ($P_{34}$) significantly impacts evacuation speed. It is well-established that evacuation speed increases with the elevation of evacuated individuals. Human factors, referred to as $P_4$, involve the interactions between individuals and fires, playing a crucial role in evacuation procedures. This encompasses considerations such as human density within enclosed spaces ($P_{41}$) and the presence of individuals with disabilities ($P_{42}$), all of which are integral components of fire risk assessment.

In establishing the values of fire risk parameters, a variety of approaches were utilized. To begin with, the sprinkler systems, which are well-distributed throughout the building, were assessed using a scoring system (0-25-50-100) based on the number of sprinklers present in each room. The quantity of smoke and heat sensors was also measured. Distances to fire alarms, fire extinguishers, and emergency exits were determined by extracting the relative coordinates of these safety features through Dynamo, a tool specifically designed for analyzing BIM data. The information gathered was subsequently organized into a tabular format and imported into a Python environment, where any noisy or inconsistent data was filtered using the Pandas library. Using the refined coordinates, distances to the safety components were accurately calculated within Python. Fire classes were categorized according to their relative risk potential and quantified accordingly. Additionally, fire loads were calculated based on the mass of combustible materials present in each room, while the presence of explosive and flammable materials was evaluated using a scoring system based on their total mass. When linking venting shafts to rooms, a scoring method was applied: a score of 0 was assigned if no shaft was present, 50 for one shaft, and 100 for two or more shafts. Information regarding room areas and door widths was sourced from the BIM model, along with ceiling heights. Floor types were also categorized based on their potential contribution to fire, as some floor types can be combustible in the event of a fire. Furthermore, the human density within the rooms was assessed relative to typical daily conditions, and the number of disabled individuals was determined using a similar methodology.

All criteria were transformed to a common $[0,1]$ scale before applying the TOPSIS method, with directional logic based on fire risk. Distances to fire alarms, fire extinguishers, and emergency exits ($P_{13}$, $P_{14}$, $P_{15}$) were considered cost criteria (greater distance indicates higher risk) and then min–max normalized. The counts of devices—specifically sprinkler systems and smoke–heat sensors ($P_{11}$, $P_{12}$)—were treated as benefit criteria (more devices correspond to lower risk) and normalized accordingly. Venting shaft linkage ($P_{25}$) was considered a cost criterion because smoke from a fire can travel through venting shafts, affecting humans even if there is no fire on their floor. Building characteristics were encoded as benefits; larger room areas ($P_{31}$), wider door widths ($P_{32}$), and greater room heights relative to the ground floor ($P_{34}$) reduce risk after min–max scaling. The floor type ($P_{33}$) was mapped to an ordinal hazard scale, where more combustible or heat-contributing types received higher scores. Hazard factors, which increase risk, were handled as follows: fire classes ($P_{21}$) were ordered from lower to higher severity and mapped to $[0,1]$; fire load ($P_{22}$) was min–max scaled as a cost; and the presence or intensity of explosive and flammable materials ($P_{23}$, $P_{24}$) was converted to monotonically increasing ordinal scores. Human factors were modeled as costs, with higher human density ($P_{41}$) and a greater number of disabled individuals ($P_{42}$) producing higher normalized risk values. After this harmonization, all criteria are commensurate, directionally consistent (cost vs. benefit), and ready for weighting and aggregation within the TOPSIS framework.

3.2. Calculation of Parameter Weights

In the developed fire risk assessment system, EWM was employed to compute the weights for each factor. The weights for each parameter were calculated in a Python 3.10 environment, utilizing the fire risk parameter values assigned to each room. Initially, a decision matrix was established, incorporating the study area rooms as alternatives and Level-2 factors as criteria within the two-tier fire risk assessment system. Subsequently, both the quantitative and qualitative values of each factor were ascertained for every alternative. The qualitative values were then converted into quantitative values based on their respective impact in a fire scenario. Following this, normalized values were computed using Equations (2) and (3). Then, entropy values and difference coefficient values for each criterion were calculated. Ultimately, the normalized weights for the Level-2 fire factors were derived. The weights of fire factors are given in the

Table 3. Calculated weights of fire risk factors.

Level-1	Level-2	Weights
Fire Prevention Factors $\left(P_1\right)$	Sprinkler system ($P_{11}$)	0.12
	Smoke and heat sensors ($P_{12}$)	0.02
	Distance to fire alarms ($P_{13}$)	0.15
	Distance to fire extinguishers ($P_{14}$)	0.14
	Distance to emergency exits ($P_{15}$)	0.14
Hazard Factors $\left(P_2\right)$	Fire classes ($P_{21}$)	0.10
	Fire loads ($P_{22}$)	0.02
	Explosive materials ($P_{23}$)	0.07
	Flammable materials ($P_{24}$)	0.06
	Linkage between venting shafts and rooms ($P_{25}$)	0.03
Building Characteristics $\left(P_3\right)$	Room area ($P_{31}$)	0.02
	Door width ($P_{32}$)	0.02
	Floor type ($P_{33}$)	0.02
	Height of the room based on the ground floor ($P_{34}$)	0.01
Human Factors $\left(P_4\right)$	Human density in rooms ($P_{41}$)	0.07
	Number of disabled peoples ($P_{42}$)	0.01

. In the table, the most significant factors contributing to fire risks fall under the Fire Prevention Factors ($P_1$) category. Within this category, the most important Level-2 fire risk parameters are Distance to fire alarms ($P_{13}$), Distance to fire extinguishers ($P_{14}$), Distance to emergency exits ($P_{15}$), and Sprinkler system ($P_{11}$), with respective weights of 0.15, 0.14, 0.14, and 0.12. The second most important factors are found in the Hazard Factors ($P_2$) category, indicating potential factors that could cause or increase fire risks. Within this category, the most important fire risk factor is Fire classes ($P_{21}$), which has a weight of 0.10. In contrast, the Building Characteristics ($P_3$) and Human Factors ($P_4$) categories were found to be the least important fire risk factors in the study area. The observed outcomes can be attributed to two key factors. First, the majority of rooms within the study area exhibit similar geometric and thematic characteristics. Most rooms feature nearly identical floor areas, door widths, and ceiling heights. Additionally, flooring materials tend to be consistent, except in instances where a room serves a specific purpose, such as a laboratory or specialized facility. Second, the human density throughout the building remains relatively uniform. The study area is predominantly composed of classrooms and offices, where the number of occupants tends to be fairly stable. Furthermore, the low number of individuals with disabilities using the building contributes to a reduction in variability regarding occupancy.

The parameter weights obtained in this study could be compared with those from other research. For example, this study reports the fire load density as 0.02. This value is relatively low compared to most other parameters, primarily because the study area lacks high levels of fire loads, aside from a few laboratories that contain chemical materials and equipment. In contrast, Zhang et al. [30] measured a fire load density of 0.0164, while Guang-Wang and Hua-Li [50] recorded it at 0.0260. According to the study, the weight of the sprinkler system is measured at 0.12, making it one of the highest parameter weight values observed. Although multiple sprinklers are installed across various locations, some areas are entirely without them, resulting in inconsistencies in the number of sprinkler systems at different sites. Zhang et al. [30] and Zhang et al. [42] recorded this value as 0.0427 and 0.0460, respectively, while Guang-Wang and Hua-Li [50] measured this value at 0.0740. Lastly, the measured weight of human density in the rooms is 0.07 in this study. This figure is notably among the highest when compared to other parameters, largely due to the fact that the study area is a faculty building that typically hosts over a thousand individuals. In a study conducted by Zhang et al. [30], this value was noted as 0.0406. In conclusion, it is evident that the common fire risk parameters exhibit almost identical weight values. This consistency is largely due to the variations in parameters used across different studies, which depend on their specific objectives and study areas. Additionally, it is worth noting that the EWM effectively assigns weights that are aligned with the characteristics of the study area in this research.

3.3. GIS-Based Fire Risk Determination

This study evaluated the fire risks of different rooms using the TOPSIS method. The fire risk calculations for the TOPSIS method were performed in a Python environment, utilizing the assigned weights and Level-2 parameter values for each room. In contrast, calculations were manually conducted using tables in the FRAME method to validate and compare results with TOPSIS. This approach was found to be more straightforward than implementing the method in Python. To compare FRAME with TOPSIS on a common basis, FRAME outputs were linearly normalized to $r\in [0,1]$. Given that FRAME does not set a specific maximum limit, the normalized axis was discretized using empirically derived cutpoints while preserving ordinal severity: $r\le 0.342$ (Very Low), $0.342<r\le 0.445$ (Low), $0.445<r\le 0.513$ (Moderate), $0.513<r\le 0.582$ (High), $0.582<r\le 0.650$ (Very High), and $0.650<r\le 1.000$ (Critical). These bands enable a direct 1:1 comparison between FRAME categories and TOPSIS scores on the same $[0,1]$ scale. The categories and their ranges are summarized in

Table 4. FRAME classes with original R ranges and normalized r bands for comparison with TOPSIS.

Risk Category	FRAME R Range	Normalized r Band $[0,1]$
Very Low	$0.0\le R\le 1.0$	$0.000\le r\le 0.342$
Low	$1.0<R\le 1.3$	$0.342<r\le 0.445$
Moderate	$1.3<R\le 1.5$	$0.445<r\le 0.513$
High	$1.5<R\le 1.6$	$0.513<r\le 0.582$
Very High	$1.6<R\le 1.9$	$0.582<r\le 0.650$
Critical	$R>1.9$	$0.650<r\le 1.000$

.

The weighted parameters for fire risk assessment were employed to evaluate the fire risk associated with each room within the study area. Initially, the BIM for the study area was integrated with GIS. Following this integration, the TOPSIS method was implemented for the assessment of fire risk. A visual comparison of the results is presented in Figure 3. Figure 3a–c show the fire risk assessment results for the first floor, first underground floor, and second underground floor of the building, respectively. The first floor of the building demonstrates a relatively high level of fire protection, with the majority of rooms classified as having very-low or low fire risks. Nonetheless, there are three offices on this floor that are assessed to have moderate fire risks. This elevation in risk is attributed to their proximity to critical safety features, such as emergency exits, fire alarms, and fire extinguishers. Conversely, the majority of rooms located on the first underground floor, which predominantly comprises laboratories, present moderate fire risks. Specifically, the static laboratory (B-B1-04), the physical processes laboratory (A-B1-15), and the ecotoxicology laboratory (A-B1-02) are identified to have moderate fire risks. These risks are primarily due to the presence of flammable and explosive materials within these spaces. Lastly, the evaluation of fire risk levels on the second underground floor of the building reveals significant variation among the rooms. Notably, the stock room (A-B2-21) and the solid laboratory (A-B2-13) are classified as having very-high fire risks. This assessment indicates that these rooms are situated at considerable distances from emergency exits, fire alarms, and fire extinguishers. Furthermore, these areas are linked to venting shafts and contain hazardous materials at dangerous levels. In addition, several other spaces on this floor, including the wastewater laboratories (A-B2-14, A-B2-15), the instrumental analysis laboratory (A-B2-16), and another solid laboratory (A-B2-19), exhibit high fire risks. Similar to the previously mentioned rooms with very-high fire risks, these areas face increased vulnerability due to their remoteness from critical safety features. A further concern arises from the elevated human density and the presence of individuals with disabilities in these rooms. Conversely, some rooms, such as the soil laboratory (A-B2-23) and the sample laboratory (B-B2-05), are assessed as having moderate fire risks. These laboratories benefit from proximity to safety features. The remaining rooms on the second underground floor are categorized as low or very-low fire risks due to their nearness to safety equipment and the limited presence of hazardous materials.

The fire risk assessment conducted utilizing the TOPSIS method has yielded the following results: 18 rooms, representing 14% of the total, were classified as having very-low fire risk; 76 rooms, accounting for 60%, were categorized as low risk; 23 rooms, or 18%, were deemed to have a moderate risk; 7 rooms, which constitute 6%, were identified as high risk; and 2 rooms, equivalent to 2%, were classified as having very-high risk. Rooms classified with low or very-low fire risks are predominantly situated on the first floor and the first underground floor. Conversely, rooms designated as having high or very-high fire risks are located on the second underground floor. The TOPSIS results indicate that 14% of the rooms possess a very-low fire risk, 60% exhibit a low fire risk, 18% demonstrate a moderate fire risk, 6% are assessed as high risk, and 2% are classified as having a very-high fire risk. The TOPSIS results presented in the Figure 4.

The proposed framework has not yet been implemented in a full-scale engineering case; rather, its validation was conducted through expert consultation. Feedback was gathered from professionals in the fields of fire safety, BIM, and GIS to evaluate the relevance of the selected criteria, the logical integrity of the framework, and its practical applicability. To further enhance the efficiency and reliability of the proposed TOPSIS-based fire risk assessment system, a focused meeting was convened with fire safety experts and occupational safety specialists. They emphasized that specific areas, such as laboratories, represent significant fire hazards and highlighted the importance of factors such as the presence of shafts, flooring types, and potential explosion risks in fire risk evaluations. A comparison of their expert opinions with the findings of the study revealed a high degree of consistency, confirming that the proposed system effectively reflects the actual fire risk conditions present within the building.

This study also employed the FRAME method, which assesses fire risks in each room [5] to validate and compare the results obtained from the TOPSIS methodology. Fire risk calculation is based on a three-level index system in the FRAME method. In the first level, there are the main indicators, namely potential risks (P), acceptable risks (A), and protection level (D). The second level contains the factors of the first level indicators such as fire load factor, area factor, evacuation time factor, and environment factor. The final level also includes the factors of the second level indicators such as total area, length, and number of persons [51]. The general framework of FRAME method is given in Figure 5.

The FRAME method provides fire risk calculations for property, occupants, and activities in the building [52]. The main equation for calculating fire risk is provided in Equation (14).

$$R={\displaystyle \frac{P}{A\times D}}$$

(14)

When the FRAME method is employed, only the risks for property were calculated because of the aim of the study. The achieved fire risks from FRAME method are illustrated in Figure 6. Figure 6a–c present the fire risk assessment results for the first floor, the first underground floor, and the second underground floor of the building, respectively. As illustrated in Figure 6a, all rooms on the first floor of the building exhibit a very-low level of fire risk. Similarly, the first underground floor primarily comprises rooms with very-low fire risks. However, there is one exception: the static laboratory (B-B1-04), which is associated with a critical level of fire risk. Additionally, the biotechnology laboratory (A-B1-10) is identified as having a low fire risk. On the second underground floor, most rooms also demonstrate very-low fire risks. The notable exception in this area is the sample laboratory (B-B2-05), which has been classified with a moderate fire risk. Furthermore, the solid laboratory (A-B2-19) presents a low fire risk.

The fire risk assessment results obtained through the FRAME method indicate that 122 rooms (97%) are classified as very-low fire risk, 2 rooms (2%) as low fire risk, 1 room (1%) as moderate fire risk, and 1 room (1%) as critical fire risk. Within the first floor, all rooms have been classified as very-low fire risk. In the first underground floor, the analysis reveals that 33 rooms (94%) are classified as very-low fire risk, 1 room (3%) as low fire risk, and 1 room (3%) as critical fire risk. Finally, on the second underground floor, 40 rooms (95%) are classified as very-low fire risk, 1 room (2.5%) as low fire risk, and 1 room (2.5%) as moderate fire risk. The detailed results of FRAME are given in the Figure 7.

The findings from the fire risk assessment methodologies, as illustrated in Figure 3 and Figure 6, demonstrate notable variations. On the first floor of the building, predominantly occupied by offices, the FRAME method identified most rooms as exhibiting very-low fire risks. In contrast, the TOPSIS method classified the fire risks across these rooms into categories of very low, low, and moderate. It is observed that rooms with moderate fire risks are located away from the exits compared to other rooms. On the first underground floor, where the majority of the space is allocated to laboratories, the TOPSIS method assessed many of the rooms as having moderate fire risks. This assessment is largely due to the presence of explosive and flammable materials in these areas. Furthermore, some rooms are connected to shafts, which pose a considerable fire risk by allowing smoke to travel through these pathways. However, the FRAME method indicated that most rooms posed very-low fire risks, with the exception of the static laboratory (B-B1-04), which was classified as having a moderate fire risk according to the TOPSIS assessment. This issue mainly stems from two factors in the FRAME method. First, the large area of the room, coupled with materials that can ignite at very-low temperatures, significantly heightens the potential fire risk. Second, the acceptable risk level is lowered due to the presence of industrial materials with high fire loads. Moreover, the number of personnel in the area is also considerable. In contrast, the TOPSIS-based method for fire risk assessment considers the characteristics of buildings by evaluating multiple factors and utilizing a weighted calculation to produce more accurate results. Similar discrepancies in fire risk evaluations were observed on the second underground floor. Although both methods agreed that none of the rooms represented very-high or critical fire risks, the TOPSIS analysis revealed that several laboratory rooms were assessed as having high or very-high fire risk levels, while FRAME method shows that almost all of the rooms have very-low and low fire risks. Also, the results could be quantitatively represented. According to the TOPSIS analysis in Figure 4, 74% of the rooms were classified as having very-low or low fire risks. In contrast, 18% of the rooms were categorized as exhibiting moderate fire risks, while 8% were classified as having high or even higher fire risk levels. Conversely, the results of the FRAME method indicate that nearly all the rooms in the study area have a very-low fire risk, at 97%. Therefore, it can be concluded that the TOPSIS method provides better results by considering various aspects of fire risks more comprehensively than the FRAME method.

The results of the TOPSIS and FRAME fire risk assessment methodologies vary considerably, especially on the second underground floor, as depicted in Figure 3c and Figure 6c. These discrepancies are presented in Figure 8 and are particularly evident on the left side of the building. The most notable changes can be observed in the solid laboratory (A13) and the stock room (A21). According to the TOPSIS results, room A13 is among the most distant from fire alarms, fire extinguishers, and emergency exits. This room contains explosive and flammable materials, and its flooring is also classified as flammable. Consequently, the fire risk classification for room A13 is deemed high, with an average number of sprinklers installed. In contrast, the FRAME analysis indicates that this room has a significant fire load potential as well as a low destruction temperature. Moreover, it benefits from effective protective measures, including sprinkler systems and fire detectors, which substantially mitigate fire risks. Therefore, while the TOPSIS method suggests that the room is at a very-high fire risk, the FRAME results present a different perspective, indicating that the room has a very-low fire risk. Similarly, the TOPSIS results indicate that A21 has a very-high fire risk due to several factors. It is situated far from fire alarms, extinguishers, and emergency exits. Furthermore, the flooring is made of flammable materials, and the area lacks smoke and heat detectors. Also, this room consists of flammable and explosive materials. Despite being recognized as having a high fire load, this room was classified as very-low risk based on the FRAME method. This discrepancy arises primarily from the protection level parameter utilized in the FRAME method. Notably, the room lacks both heat and smoke detectors, which would typically suggest a heightened fire risk. However, the FRAME method assigns approximately similar values to the protection level factors, despite differences in their significance. For example, one of the sub-parameters that contribute to the calculation of the protection level is the configuration of automatic fire detectors. This sub-parameter includes details about the fire detectors themselves. The "Small fire zone with individual identification" sub-parameter is assigned a value of 2, the same value given to smoke detectors. Moreover, other protection level sub-parameters, such as those pertaining to fire extinguishers, only provide qualitative information. Consequently, this can result in similar risk assessments for rooms that possess distinctly different characteristics and relationships with their surroundings. The differences are also apparent in other rooms, which are classified as having a high, moderate or low fire risks according to the TOPSIS-based fire risk assessment method. In contrast, the FRAME method categorizes these rooms as having very-low fire risk. The primary reason for this discrepancy lies in the parameters utilized in the TOPSIS method. For example, the rooms marked as high risk exhibit considerable distances from fire alarms, extinguishers, or emergency exits. Furthermore, the fire loads in these rooms were assessed based on actual conditions, whereas the FRAME method adopts a theoretical approach.

The findings are presented in Figure 9. The figure illustrates that 104 rooms categorized as having a very-low fire risk by the FRAME method do not show similar fire risk levels when assessed using the TOPSIS method. This indicates that the FRAME method identifies a greater number of rooms with very-low fire risks than the TOPSIS method. In addition, noteworthy variations were observed in the low fire risk category. The results indicate that 74 rooms, which were not classified as low fire risk by the FRAME method, were identified as low fire risk by the TOPSIS method. Changes were also noted in the classification of rooms categorized as moderate fire risks, with 22 rooms identified as moderate fire risks based on the results. The most significant variations were found in the categories of high, very-high, and critical fire risks. According to the figure, seven rooms were classified as high fire risks and two rooms as very-high fire risks—categories that were not recognized by the FRAME method but were noted by the TOPSIS method. Lastly, the room determined to be at critical fire risk by the FRAME method was not classified as critical by the TOPSIS method.

The findings demonstrate that the BIM-GIS-based TOPSIS fire risk assessment system developed in this study is more effective than the FRAME method. This improvement is primarily attributed to several key factors, as the TOPSIS method considers various aspects of fire risks more comprehensively and integrates multiple criteria in a systematic way: (a) The FRAME method does not utilize BIM and therefore lacks comprehensive building data. In contrast, the TOPSIS fire risk assessment system effectively leverages BIM to extract valuable information about the building and integrates it with GIS. Given that fire risk is influenced by multiple factors, including building materials and geometric configuration, it is essential to establish a robust and adaptable fire risk assessment methodology. This integration enables rapid fire risk assessments and enhances the visualization of results through mapping, capitalizing on the capabilities of BIM. (b) The FRAME method is fundamentally based on strict tabular values. In this approach, sub-factors are derived from designated tables. For instance, the fire load is calculated using a formula that incorporates two distinct parameters: fixed and mobile fire load densities. These densities are sourced from tables; for example, the mobile fire load density for schools is 200 Kcal/${\mathrm{m}}^2$ [18]. However, schools and other types of buildings do not always possess the same characteristics. As a result, a customized fire risk assessment system is essential in these situations. The developed system is tailored for the specific study area and can be adapted for similar environments. (c) The FRAME method is designed to include all potential parameters essential for a thorough fire risk assessment. However, it is possible for buildings not to meet every one of these criteria. Consequently, the proposed TOPSIS fire risk assessment system concentrates on the most vital features that are crucial for accurately assessing fire risk. (d) The FRAME method encompasses a range of qualitative parameters, including “enough fire extinguisher”, “insufficient fire extinguisher”, and “there is no staircase for the exit” [18]. These qualitative parameters are converted into quantitative values. However, the inclusion of spatial analyses can lead to more accurate assessments. As such, the developed fire risk assessment system leverages spatial correlations between various elements, such as the distances to emergency exits, fire alarms, and fire extinguishers from the central nodes of the rooms. Furthermore, certain qualitative parameters, such as “floor type” and “fire classes,” are transformed into quantitative values based on their significance or the number of items involved. Consequently, it should be noted that when the fire risk assessment is performed using the BIM-GIS integrated model, the developed TOPSIS fire risk assessment methodology yields better results than FRAME in this study.

4. Conclusions

The present study aimed to develop a fire risk assessment system using the TOPSIS, leveraging the integration of BIM and GIS. This system is designed to assess fire risks within a selected building, the Civil Engineering Faculty of ITU, which comprises diverse room types, including laboratories and offices, each representing varying levels of fire risk. Dynamo was employed to obtain the relative coordinates of safety components, enabling the calculation of distances to fire alarms, extinguishers, and emergency exits. Additionally, parameters such as sprinkler systems, smoke and heat sensors, fire classes, fire loads, venting shafts, room areas, door widths, ceiling heights, floor types, human density, and the number of disabled individuals were quantified through scoring systems and categorization methods. Weights for each parameter were determined in Python, and fire risk assessments were conducted using the TOPSIS method. Meanwhile, FRAME method calculations were performed manually with predefined tables for simplicity. In this framework, a BIM model with LoD 300 was utilized. The proposed framework consists of four distinct stages: identification of fire risk parameters, calculation of parameter weights, GIS-based fire risk determination, and visualization of results. The fire risk parameters were identified through a comprehensive review of various sources, including BIM of the study area, relevant literature, standards, and established codes. To assign objective weights to these parameters, the EWM was utilized. Furthermore, the FRAME method was employed to compare and validate the results obtained from the developed TOPSIS-based fire risk assessment approach. The results indicate that the developed TOPSIS fire risk assessment system yielded improved outcomes in relation to FRAME in this study. This is achieved by employing a BIM model of the study area, which provides comprehensive information about the building. It effectively measures spatial relationships between objects by incorporating dynamic values that comprehensively cover the study area using GIS, rather than relying on the rigid tabular values commonly found in FRAME. Also, the parameters of this model thoroughly encompass the study area. Moreover, the developed fire risk assessment system effectively converts qualitative information into quantitative data by considering the significance of the information or the number of relevant items involved. Consequently, this study emphasizes the crucial role BIM and GIS for fire risk assessment. These tools offer detailed data about buildings, clarify spatial relationships between objects, and present results in a user-friendly manner. Additionally, by employing more detailed BIM of the buildings in subsequent research, the limitations associated with detail level can be addressed, potentially resulting in enhanced outcomes. In summary, the proposed BIM-GIS-integrated TOPSIS framework presents a scalable and adaptable methodology suitable for a diverse range of facilities, from individual buildings to intricate environments such as hospitals, airports, and shopping centers. By offering quick, practical, and reliable insights into fire safety, this framework acts as a decision support system that aids decision-makers in identifying critical risks and effectively prioritizing mitigation strategies. Future efforts will focus on testing the framework in larger-scale scenarios to further validate its generalizability, robustness, and practical applicability. Additionally, the accessibility analysis of the building will be carried out using network-based approaches. It is important to acknowledge that its scalability is ultimately dependent on the availability of detailed data and the specific characteristics of the study area, as fire risk assessment involves inherently dynamic factors.