Fire risk assessment of high-rise buildings under construction based on unascertained measure theory.

To prevent fire accidents in high-rise buildings under construction, in this paper, the fire risk assessment of such buildings is studied. First, based on project investigation and a literature review, a fire risk assessment index system suitable for high-rise buildings under construction was established. Second, the unascertained measure theory was applied to establish a fire risk assessment model for high-rise buildings under construction. The index weight was determined by the entropy weight method. Finally, taking a high-rise building project in Xi'an, China, as an example, the feasibility and rationality of the fire risk assessment index system and assessment model were verified. This research provides a new method for objectively assessing the fire risk of high-rise buildings under construction and provides a certain reference for controlling the fire risk of high-rise buildings under construction.

Introduction

The continuous progress of society and the development of the economy have resulted in gradual urbanization and the rapid construction of various types of buildings [1, 2]. At present, the engineering projects under construction can be seen everywhere. Due to the large amount of combustible and flammable materials on the construction site, many fire and heat sources, poor fire fighting conditions, and high fire risk, and once a fire occurs, it is difficult to fight the fire. The frequent fires of engineering projects under construction have caused serious safety accidents, caused serious property losses and casualties, and brought extremely bad social impacts [3]. At the same time, it will also affect the sustainable development of society and the environment [4]. Therefore, the fire risk of engineering projects under construction has attracted the attention of all sectors of society.

The fire risks of high-rise building under construction are relatively large, with many influencing factors and thus high complexity and uncertainty, mainly in the following aspects. On the one hand, high-rise buildings under construction require a large amount of work, a long construction period, and a complex and variable construction environment, and these factors increase fire risk uncertainty. Especially during the peak period of construction, there are many mixed operations, open-fire operations, and fire hazards, which can easily lead to fire accidents. On the other hand, the high-rise buildings under construction are still in the construction state, most of the fire control systems of the buildings have not been installed in place, or are not in use even though they have been installed. Instead, there are only some temporary fire control facilities on the construction site, and the fire water source is limited. Therefore, in the event of a fire, the fire spreads quickly and is difficult to extinguish.

Therefore, it is important to understand the hidden fire hazards present in high-rise buildings under construction, to perform scientific risk assessments and to take effective preventive measures.

Literature review

The literature review and analysis are divided into two parts.

Building fire risk assessment

To reduce the probability of building fires and to prevent and control such fires, much research has been done on building fire risk. In the 1980s, with the continuous progress of science and technology, ultrahigh buildings and super-large buildings began to appear; however, the current design specifications could not meet the fire protection requirements of the new buildings. Therefore, performance-based fire protection design was proposed. In 1985, the UK promulgated the first performance-based fire protection code. Thereafter, some developed countries researched performance-based fire protection design and related fire safety engineering theories and technologies and explored performance-based fire protection design. From 1996 to 2002, four international symposia on performance-based design norms and design methods were held, showing that performance-based fire protection design had become an international trend. Since then, an increasing number of experts and scholars have researched building fire risk. Kang et al. calculated the fire safety assessment level of high-rise buildings by using fuzzy centralization theory, and the developed method was both applicable and practical [5]. Ding et al. developed a smart fire risk estimation model for high-rising buildings by using the back propagation (BP) neural network [6]. Ren designed a model for assessing the fire risks of logistics warehouses by using the analytic hierarchy process (AHP) method [7]. Xin et al. analyzed a large number of fire incidents in China, determined the characteristics and main factors of the fires, and evaluated the risk levels of residential buildings [8]. Wu et al. presented a diagnostic assessment of fire safety by using the extension engineering method, which can be applied for all kinds of buildings [9]. Gao et al. established a fuzzy analytic hierarchy process model that aimed to assess the tunnel fire risk of subways by combining the fuzzy consistent matrix with AHP [10]. Chen et al. established a building fire risk assessment system for factories, hotels, malls, schools and public buildings and applied AHP to evaluate building fire risk [11]. Sun et al. introduced fuzzy mathematics into the analytic hierarchy process; furthermore, the analytic hierarchy process-fuzzy comprehensive evaluation (AHP-FCE) method was used to evaluate risk, and the traditional quantitative evaluation method was integrated with the qualitative evaluation method [12]. Roshan et al. assessed fire risk and economic loss by using event tree analysis [13]. Lau et al. proposed a fire risk scoring system and applied it for determine the fire risk of residential buildings [14]. Wei et al. established a fire risk assessment model based on support vector machine (SVM) theory, and the method was precise even when given a small number of samples [15]. Wei et al. proposed a fast fire risk assessment method based on fuzzy mathematics and the SVM algorithm [16]. Liu et al. proposed a fire risk assessment system for large-scale commercial buildings by using the structure entropy weight method [17]. Li et al. established a mathematical model by using the gray risk degree method, the analytic hierarchy process and the fuzzy evaluation method [18]. Bart et al. developed a quantitative risk assessment method that can quantify the fire safety level through the failure probability, individual risk and social risk [19]. Sun et al. expounded the procedures and methods of fire risk assessment for super-high-rise buildings and quantified the possibility and consequences of a fire [20]. Qian et al. established an urban fire risk assessment index system according to the possibility and severity of the fire and constructed an evaluation model based on regression with latent variables [21]. Hassanain et al. developed a fire safety evaluation tool that can evaluate existing restaurant facilities to identify and eliminate fire hazards [22]. Omidvari et al. proposed a model based on the analytical hierarchy process and failure mode and effect analysis logic [23]. Li et al. built a fire risk assessment system for coal mines based on the TOPSIS method [24]. Zheng conducted a fire risk assessment of the stadium used for the National Games based on the AHP method to ensure safety [25].

Many researchers have introduced and systematically summarized models and methods for quantitatively assessing building fire risk [26]. In addition, there are many fire risk assessment methods [27, 28] (e.g., expert scoring, Delphi, analytic hierarchy process, fuzzy comprehensive evaluation, fault tree analysis, grey comprehensive evaluation, SVM, TOPSIS, artificial neural network, matter element extension). Researchers have applied these methods for fire risk assessment.

In summary, there are many studies on fire risk, and good applications have been developed. Relevant research is mainly focused on the fire risk of existing buildings. There is relatively little research on the fire risk of buildings that are under construction and even less research on the fire risks of high-rise buildings under construction. Therefore, in this paper, high-rise buildings under construction are considered as the research object, the characteristics of fire accidents in the high-rise building under construction are summarized, a fire risk assessment is conducted for high-rise buildings under construction.

However, the fire risks of high-rise buildings under construction are relatively large, with many influencing factors and thus high complexity and uncertainty. In addition, there are many uncertainties associated with the assessment index and the assessment process. Therefore, reasonable mathematical methods are needed for fire risk assessment of high-rise buildings under construction.

Unascertained measure theory

Unascertained information and its mathematical processing theory were first proposed by Wang Guangyuan in 1990 [29]. Unlike fuzzy information, random information and gray information, unascertained information indicates that people do not fully grasp the real quantitative relationships or states being considered, which causes subjective and cognitive uncertainty in the minds of decision makers and evaluators. It can be said that all systems with behavioral factors are unascertained. To develop a method for quantitatively describing the unascertained state or the unascertained size of something, Liu Kaidi et al. [30] established unascertained mathematical theory and proposed an evaluation model for unascertained measure theory to describe an unascertained state or an unascertained nature by using a real number in [0,1].

Thereafter, unascertained measure theory has been rapidly developed and widely applied in many fields, such as mining risk assessment [31, 32], geotechnical risk evaluation [33, 34], pipeline risk assessment [35], geological risk assessment [36], ecological risk assessment [37], chemical safety evaluation [38], and social evaluation [39]. According to the above description, unascertained measure theory can effectively and quantitatively analyze various uncertain factors. Furthermore, it can avoid the incompleteness of risk assessment indexes due to the uncertainty of the influencing factors, and it can avoid the shortcomings of the subjectivity of risk assessment results caused by expert scoring.

This paper applies unascertained measure theory to the fire risk assessment of high-rise buildings under construction to address many uncertainties of fire risk assessment. It is hoped that the model can provide a new idea for fire risk assessment of high-rise buildings under construction.

Model development

The purpose of this study is to conduct a fire risk assessment of high-rise buildings under construction. The algorithm used in this study is illustrated in Fig 1.

Fig 1

The algorithm for this study.

Fire risk assessment index system and index grading ranges

Establishment of the fire risk assessment index system

Establishing a scientific and comprehensive assessment index system is the key to risk assessment, as it affects the reliability and accuracy of the assessment results. For high-rise buildings, the chimney effect can occur, which greatly impacts fire risk. Furthermore, for high-rise buildings under construction, the site is complex and changing, and various fire influencing factors are uncertain. Therefore, the establishment of the assessment index system is very difficult.

Based on the above literature review in the “Building fire risk assessment” and relevant standards [40-43], visited and consulted the owner, supervisory unit, design unit, fire brigade and other related units, the index system and index grading standards were determined. A fire risk assessment index system that is suitable for high-rise buildings under construction is established as shown in Table 1.

Table 1

Fire risk assessment index system for high-rise buildings under construction.

First-level index	Second-level index	Third-level index
Passive fire prevention factor (V1)	Architectural design (V11)	Fireproof endurance rating of building material (V111)
		Fire and smoke prevention zoning (V112)
		Fire separation distance (V113)
		Building height (V114)
	Fire load (V12)	Fire load density (V121)
Active fire prevention factor (V2)	Building fire extinguishing facilities (V21)	Allocation situation of temporary fire fighting equipment (V211)
		Fire detection and automatic alarm (V212)
		Temporary fire water-supply system (V213)
	Fire-fighting capacity of fire brigade (V22)	Fire lane (V221)
		Firefighter business level (V222)
Safe evacuation (V3)	Safe evacuation channel (V31)	Evacuation channel condition (V311)
		Setup situation of safety export (V312)
		Safe evacuation route and distance (V313)
	Safe evacuation facilities (V32)	Safety evacuation indication sign (V321)
		Emergency lighting (V322)
	Site personnel situation (V33)	Pedestrian flow situation of construction site (V331)
		Fire awareness of on-site personnel (V332)
	Emergency plan (V34)	Emergency relief material (V341)
		Preparation of emergency rescue plan (V342)
		Drill of emergency rescue plan (V343)
Fire safety management (V4)	Establishment of management system (V41)	Establishment and implementation of fire control system (V411)
		Division of safety duty (V412)
	Daily fire management work (V42)	Maintenance of fire fighting facility (V421)
		Regular inspection of fire safety (V422)
		Education and training of fire safety (V423)
		Management of flammable and explosive material (V424)
	Managerial personnel factor (V43)	Mastery situation of fire control knowledge and skill (V431)
		Managerial personnel business level (V432)
Fire hazard source control (V5)	Human factor (V51)	Production hot work (V511)
		Careless use of electricity (V512)
		Careless smoking (V512)
	Material factor (V52)	Material stacking condition (V521)
		Electrical equipment condition (V522)

Determination of the index grading standards

This paper determines the index grading standards through theoretical research and expert experience and converts qualitative indexes into semiquantitative indexes. Based on the above literature review, relevant standards [40-43], project investigation and consulting expert opinions, the fire risk assessment is divided into five grades, i.e., Grade I, Grade II, Grade III, Grade IV and Grade V, which refer to very low, low, moderate, high, very high fire risks, respectively. The specific index grading standards are shown in Table 2.

Table 2

Index grading standards.

Index	Grade I (C1)	Grade II (C2)	Grade III (C3)	Grade IV (C4)	Grade V (C5)
	(90, 100]	(80, 90]	(70, 80]	(60, 70]	[0, 60]
V111	<1%	1%-5%	5%-10%	10%-15%	≥15%
V112	<1%	1%-5%	5%-10%	10%-15%	≥15%
V113	<1%	1%-5%	5%-10%	10%-15%	≥15%
V114	Below 3 floors	4 to 6 floors	7 to 9 floors	10 floors to the top	Capping and Decoration Stage
V121	≤200	200–400	400–600	600–800	>800
V211	<1%	1%-5%	5%-10%	10%-15%	≥15%
V212	Very good usability and very high safety	Good usability and high safety	General usability and general safety	Poor usability and low safety	Not usable and very low safety
V213	<1%	1%-5%	5%-10%	10%-15%	≥15%
V221	<1%	1%-5%	5%-10%	10%-15%	≥15%
V222	Very high level of fire fighting	High level of fire fighting	General level of fire fighting	Poor level of fire fighting	Very poor level of fire fighting
V311	<1%	1%-5%	5%-10%	10%-15%	≥15%
V312	<1%	1%-5%	5%-10%	10%-15%	≥15%
V313	<1%	1%-5%	5%-10%	10%-15%	≥15%
V321	<1%	1%-5%	5%-10%	10%-15%	≥15%
V322	<1%	1%-5%	5%-10%	10%-15%	≥15%
V331	<0.5	0.5–1	1–1.5	1.5–2	2–2.5
V332	Very high fire awareness	High fire awareness	General fire awareness	Low fire awareness	Very low fire awareness
V341	Very adequate	Adequate	General	Inadequate	Very inadequate
V342	Compliant	Basically compliant	Generally compliant	Less compliant	Incompliant
V343	Regular	General	Occasional	Few	Never
V411	Very good	Good	General	Bad	Very bad
V412	Very clear and reasonable	Basically clear and reasonable	Generally clear and reasonable	Less clear and reasonable	Not clear and reasonable
V421	Very good	Good	General	Bad	Very bad
V422	Regular	General	Occasional	Few	Never
V423	Regular	General	Occasional	Few	Never
V424	Very safe and reasonable	Basically safe and reasonable	Generally safe and reasonable	Less safe and reasonable	Not safe and reasonable
V431	Very skilled	Basically skilled	Generally skilled	Less skilled	Very unskilled
V432	Very strong	Strong	General	Weak	Very weak
V511	The operation is strictly in conformity with the specifications and very high safety	The operation is in conformity with the specifications and high safety	The operation is general in conformity with the specifications and general safety	The operation is not in conformity with the specifications and low safety	The operation is strictly not in conformity with the specifications and very low safety
V512	The operation is strictly in conformity with the specifications and very high safety	The operation is in conformity with the specifications and high safety	The operation is general in conformity with the specifications and general safety	The operation is not in conformity with the specifications and low safety	The operation is strictly not in conformity with the specifications and very low safety
V512	Compliant	Basically compliant	Generally compliant	Less compliant	Incompliant
V521	The stacking is strictly in conformity with the specifications and very high safety	The stacking is in conformity with the specifications and high safety	The stacking is general in conformity with the specifications and general safety	The stacking is not in conformity with the specifications and low safety	The stacking is strictly not in conformity with the specifications and very low safety
V522	The erection and utilization are strictly in conformity with the specifications and very high safety	The erection and utilization are in conformity with the specifications and high safety	The erection and utilization are general in conformity with the specifications and general safety	The erection and utilization are not in conformity with the specifications and low safety	The erection and utilization are strictly not in conformity with the specifications and very low safety

The index grading standards of V111, V112, V113, V211, V213, V311, V312, V313, V321, and V322 are expressed by the ratio of the number that does not meet the requirements of the specification to the total number, while the index grading standard V331 is expressed by the ratio of the number of people in the peak period to the total building area.

Suppose that the assessment object X = {X1,X2,⋯,X} and the assessment index set V = {V1,V2,⋯,V}. If x denotes the measured value of the i-th assessment object X with respect to the j-th assessment index V, then X can be expressed as an m-dimensional vector {x,x,⋯,x}. Suppose that the assessment grade space C = {C1,C2,⋯,C}, where C(k = 1,2,⋯,p) is the k-th assessment grade, and suppose that the k-th grade is higher than the k+1-th grade in the risk assessment process, i.e., C>C. If C1>C2>⋯>C or C1<C2<⋯C2,⋯,C} is an ordered segmentation class of assessment space C.

Single-index unascertained measure

If μ = μ(x∈C) denotes the degree to which the measured value x belongs to the k-th assessment grade C, then

Eq (1) is called nonnegative boundedness, Eq (2) is called normalization, and Eq (3) is called additivity. If μ satisfies Eqs (1)–(3), then μ is called the unascertained measure, which is abbreviated as measure.

For every assessment object X(i = 1,2,⋯,n), the matrix of (μ) is called the single-index unascertained measure matrix of X, as shown in Eq (4).

Before establishing the single-index unascertained measure matrix, it is necessary to establish a single-index unascertained measure function. At present, the construction methods of a single-index unascertained measure function mainly include linear, exponential, parabolic and sinusoidal methods [44]. Regardless of the type of simulation function used, it must satisfy the limiting conditions of Eqs (1)–(3). To operate simply and easily, this paper adopted the linear unascertained measure function, and the calculation expression is as follows [45]:

Multi-index comprehensive unascertained measure

Given that μ = μ(X∈C) denotes the degree to which the assessment object X belongs to the k-th assessment grade C, as shown in Eq (6), where 0≤μ≤1 and are satisfied, the vector {μ,μ,⋯,μ} is called the multi-index comprehensive unascertained measure vector of X. where w is the index weight. The specific index weight calculation process is detailed in the following section.

Credible degree recognition

To get the final assessment result, credible degree recognition criteria are introduced. Supposed that λ (λ≥0.5; usually, λ = 0.6 or 0.7) is the credible degree, if C1>C2>⋯>C is satisfied and p0 is satisfied by Eq (7), then the assessment object X belongs to the assessment grade .

Determination of the index weight

The entropy weight method [46, 47] is used to determine the weight of each index, and this method can make full use of the values of the single-index unascertained measure matrix.

Suppose w denotes the relative degree of importance of an index compared with other indexes. If w satisfies 0≤w≤1 and , then w is called the index weight of V, and w = (w1,w2,⋯,w) is called the vector of the index weight. According to the matrix (μ), the index weight w can be obtained from Eqs (8) and (9). where H>0; ; t is a coefficient and t = 1/ln p; and when μ = 0, μ ln μ = 0 (i = 1,2,⋯,n).

Case study

To verify the effectiveness of the model proposed herein, a high-rise building under construction is taken as an example. The building, which is an inpatient building in a hospital in Xi’ an, China, has a shear wall structure with a length of approximately 116 m, a width of approximately 53.7 m and a height of approximately 81.8 m. The building has 21 floors, 2 of which are underground and 19 of which are above ground. The total floor area is 77,260 m2, the underground floor area is 10,600 m2, and the ground floor area is 66,660 m2.

Data collection

This paper used the expert scoring method to determine the actual assessment value of each index. Ten experts in related fields were invited to inspect the construction site, the basic information of these ten experts is shown in Table 3.

Table 3

The basic information of the experts.

Expert code	Professional title	Academic qualification	Working years
E1	Senior title	Specialty	25
E2	Intermediate title	Undergraduate	15
E3	Senior title	Undergraduate	18
E4	Intermediate title	Master	8
E5	Intermediate title	Undergraduate	12
E6	Senior title	Master	21
E7	Intermediate title	Undergraduate	14
E8	Intermediate title	Master	9
E9	Senior title	Master	17
E10	Intermediate title	Undergraduate	15

And the third-level indexes were scored by these ten experts according to the index grading standards in Table 2. The specific scores are shown in Table 4. To ensure the objectivity and authenticity of the scores, the highest and lowest scores of the indexes are first eliminated, and then the average value of each index was obtained. The fire risk assessment index value of the project is shown in Table 4.

Table 4

The fire risk assessment index value of the project.

Index	Scoring by ten experts										Final indexvalue
	E1	E2	E3	E4	E5	E6	E7	E8	E9	E10
V111	85	83	87	84	87	83	84	82	84	83	84.13
V112	86	88	83	84	85	86	84	84	85	87	85.13
V113	87	89	92	90	88	90	88	89	87	89	88.75
V114	65	63	64	68	63	64	65	64	66	63	64.25
V121	63	65	67	66	68	65	67	66	63	66	65.63
V211	92	85	87	90	88	86	91	87	89	86	88.00
V212	56	60	58	61	63	58	59	58	61	56	58.88
V213	88	83	86	85	87	88	85	87	85	84	85.88
V221	89	91	88	85	86	87	90	86	90	87	87.88
V222	88	90	93	87	92	87	91	92	91	88	89.88
V311	84	87	86	83	82	83	85	84	85	82	84.00
V312	85	88	87	84	83	89	83	87	86	83	85.38
V313	88	83	85	84	89	83	88	85	88	84	85.63
V321	90	88	85	91	89	89	90	89	87	85	88.38
V322	87	83	84	86	87	85	84	86	83	87	85.25
V331	83	85	79	80	83	82	84	81	85	84	82.75
V332	88	90	86	87	83	87	88	85	89	86	87.00
V341	89	92	94	90	91	91	90	90	93	91	91.00
V342	83	85	86	83	82	86	86	86	82	86	84.63
V343	87	89	84	86	87	88	87	86	84	87	86.50
V411	84	80	83	81	85	82	84	82	83	81	82.50
V412	81	81	79	82	80	81	79	82	82	81	80.88
V421	83	81	84	81	82	81	80	84	80	81	81.63
V422	83	81	85	81	84	83	81	84	83	83	82.75
V423	85	83	87	83	85	85	87	86	84	86	85.13
V424	85	86	84	88	84	87	85	84	87	85	85.38
V431	87	85	89	85	86	85	85	85	88	85	85.75
V432	89	87	88	84	82	86	87	86	85	85	86.00
V511	84	88	87	83	82	87	85	82	86	84	84.75
V512	85	82	81	83	80	82	81	80	82	84	81.88
V512	85	87	84	83	88	83	86	83	84	85	84.63
V521	86	87	85	88	83	84	87	87	87	85	86.00
V522	80	81	80	84	85	82	80	80	84	80	81.38

Calculation process

(1) Constructing the single-index unascertained measure function

The unascertained measure function studied in this paper is linear. The single-index unascertained measure function is constructed as follows:

In this paper, the single-index unascertained measure function is represented by a graph, as shown in Fig 2.

Fig 2

The single-index unascertained measure function.

(2) Calculating the single-index unascertained measure matrix

According to the index value in Table 4 and based on the single-index unascertained measure function established by Eq (10)–Eq (14), the three-level index unascertained measure matrix can be obtained.

(3) Determining the index weight

According to the previous content, the index weight is calculated, as shown in Table 5.

Table 5

Index weights.

First-level index	First-level index weight	Second-level index	Second-level index weight	Third-level index	Third-level index weight
V1	0.178	V11	0.255	V111	0.261
				V112	0.296
				V113	0.208
				V114	0.236
		V12	0.745	V121	1.000
V2	0.158	V21	0.327	V211	0.254
				V212	0.436
				V213	0.310
		V22	0.673	V221	0.383
				V222	0.617
V3	0.206	V31	0.317	V311	0.333
				V312	0.348
				V313	0.319
		V32	0.257	V321	0.410
				V322	0.590
		V33	0.204	V331	0.535
				V332	0.465
		V34	0.222	V341	0.397
				V342	0.357
				V343	0.246
V4	0.218	V41	0.301	V411	0.529
				V412	0.471
		V42	0.346	V421	0.199
				V422	0.220
				V423	0.306
				V424	0.275
		V43	0.353	V431	0.517
				V432	0.483
V5	0.240	V51	0.605	V511	0.380
				V512	0.251
				V512	0.369
		V52	0.395	V521	0.537
				V522	0.463

(4) Calculating the comprehensive unascertained measure matrix

1) Calculating the second-level index comprehensive measure matrix

According to the third-level index unascertained measure matrix and the third-level index weight, the second-level index comprehensive measure matrix can be obtained.

2) Calculating the first-level index comprehensive measure matrix

According to the second-level index unascertained measure matrix and the second-level index weight, the first-level index comprehensive measure matrix can be obtained.

3) Calculating the total target comprehensive measure matrix

According to the first-level index unascertained measure matrix and the first-level index weight, the total target comprehensive measure matrix can be obtained

(5) Credible degree recognition

The total target comprehensive measure matrix of fire risk of high-rise buildings under construction is μ = [0.181 0.584 0.078 0.133 0.024].

According to the obtained measure matrix, the credible degree recognition criterion is used for fire risk assessment. λ, which is generally 0.6 or 0.7, is set to 0.6 in this paper. As 0.181+0.584 = 0.765>0.7, p0 = 2; that is, the fire risk assessment grade of the high-rise building under construction is Grade II and the risk is low, and this assessment result is consistent with the result determined by the fire department. Similarly, the fire risk assessment grades of all indexes can be obtained are shown in Table 6.

Table 6

Fire risk assessment grade of each index.

First-level index	Risk grade	Second-level index	Risk grade	Third-level index	Risk grade
V1	IV	V11	II	V111	II
				V112	II
				V113	I
				V114	IV
		V12	IV	V121	IV
V2	II	V21	V	V211	II
				V212	V
				V213	II
		V22	I	V221	II
				V222	I
V3	II	V31	II	V311	II
				V312	II
				V313	II
		V32	II	V321	II
				V322	II
		V33	II	V331	II
				V332	II
		V34	II	V341	I
				V342	II
				V343	II
V4	II	V41	III	V411	II
				V412	III
		V42	II	V421	III
				V422	II
				V423	II
				V424	II
		V43	II	V431	II
				V432	II
V5	II	V51	II	V511	II
				V512	III
				V512	II
		V52	II	V521	II
				V522	III

Results and discussion

Through the analysis of the above assessment results, it is concluded that the actual fire risk level of the case is Grade II, which represents a low fire risk and an acceptable state. From Table 6, we can see the assessment results of each index level. Among them, the risk level of the “passive fire prevention factor” is Grade IV, while the risk levels of the “active fire prevention factor, the safe evacuation factor, fire safety management and fire hazard control” are Grade II.

After obtaining the fire risk grade of the project, the project leader, management personnel, technical personnel and relevant experts conducted a field investigation, discussed and analyzed the actual situation of the project. For the “passive fire prevention factor”, the project is in the sealing stage of the main body structure, and the building is very tall. Once a fire occurs, the chimney effect and fire load will be very large. In addition, the construction site of the project occupies a relatively small area and piles a lot of materials. There are a large number of people on the site, food and accommodation are all on the construction site, and most of the building materials and workers' daily necessities are flammable, which greatly increases the fire load on the construction site. For the “active fire prevention factor”, the fire protection of the project is relatively in place, but the fire detection and automatic alarm equipment is just like being idle and cannot be used. For the “safe evacuation factor”, the overall situation is good. For the “fire safety management”, the on-site fire fighting facilities are well equipped, a sufficient number of fire extinguishers are installed, and the fire fighting facilities are reasonably distributed, but most of these facilities are idle and unusable. For the “fire hazard control”, there are illegal operations of electrical equipment and unlawful pulling of wires on site.

Although the overall fire risk level of this project is Grade II, as the height of the building increases, it is very important to pay attention to fire prevention in the subsequent construction process. For those indexes with risk level of Grade V and Grade IV, measures must be taken immediately to deal with them, and they must be carefully supervised and inspected.; for those indexes with risk level of Grade III and Grade II, measures should be taken to deal with them according to the specific circumstances; for those indexes with risk level of Grade I, it is not necessary to deal with them.

By analyzing the basic situation of the project and comparing the assessment results of the model with the on-site situation, it is confirmed that the assessment results are scientific, reasonable, and consistent with the on-site situation. Based on the results, the project managers should strengthen the configuration and supervision of the fire extinguishing facilities of the building itself, as well as the fire control management, and adopt reward and punishment measures. In addition, the project managers should strengthen the fire safety investigation at the construction site, monitor the potential fire hazards at all times, and rectify them immediately to minimize the possibility of a fire.

According to the above analysis, the assessment model established in this paper is feasible. It shows that in view of the various uncertainties in the fire risk assessment of high-rise buildings under construction, this method can obtain reasonable assessment results. In addition, the calculation process of the model is simple. In summary, the assessment model can handle the uncertainty of fire risk assessment, and it is very suitable for the fire risk assessment of high-rise buildings under construction.

Limitations

The unascertained measure theory can be used for the fire risk assessment of high-rise buildings under construction, but the study has several limitations. Due to the project is under construction and the fire hazards are in a state of change in each stage. The index system established in this paper is a static index system. More research is needed to establish a dynamic assessment index system and to determine the fire hazards in each stage from time to time.

Conclusions

To discover hidden fire hazards and to reduce the occurrence of fire accidents in high-rise building under construction, this paper studied the fire risk assessment of high-rise buildings under construction. First, a fire risk assessment index system for high-rise buildings under construction was established, including 5 first-level indexes, 13 second-level indexes and 33 third-level indexes. Second, according to the uncertainty of the fire influence factors of high-rise buildings under construction and the uncertainty of the assessment process, a fire risk assessment model for high-rise buildings under construction based on unascertained measure theory was proposed. Finally, the feasibility and rationality of the proposed fire risk assessment index system and assessment model were verified by taking an inpatient building project of a hospital in Xi’an as an example. This study can solve the problem of fire risk assessment and provide new ideas and methods for the fire risk assessment and control of high-rise buildings under construction in the future.