Risk Assessment of Prefabricated Building Projects Based on the G1-CRITIC Method and Cloud Model: A Case Study from China

Abstract

To enhance the ability to identify and analyze the construction safety risks of prefabricated building projects, this paper explores the risk factors affecting the construction safety of prefabricated buildings from the perspective of the construction stage. Based on the WSR theory, this paper identifies risk-influencing factors from five dimensions: personnel, materials, management, technology, and environment, and constructs a safety risk assessment index system. This paper establishes a risk assessment model based on the G1-CRITIC method and cloud model. Firstly, it quantitatively analyzes the weights of the risk indicators for prefabricated building construction, and then evaluates the specific degree of impact of each indicator on the construction risk of this type of project. The research results show that the project is at the low-risk level, but there are still some potential risks in terms of material and technical factors, which require close attention and targeted management. The evaluation results obtained by applying this model are consistent with the current actual situation of prefabricated building construction, further demonstrating the applicability of this model. The risk assessment model proposed in this paper, by focusing on a specific type of risk, comprehensively incorporates the fuzziness and randomness of risk factors, thereby more effectively capturing the dynamic characteristics of risk evolution. This model can effectively evaluate the level of safety risk management and plays a positive role in reducing the incidence of engineering accidents. Furthermore, it also provides practical experience that can be drawn upon by risk managers of similar projects which holds significant theoretical value and practical guiding significance.

1. Introduction

Compared with traditional buildings, prefabricated buildings have advantages such as high efficiency and fast construction, energy conservation and environmental protection, and green construction [1,2]. These advantages have drawn extensive attention from countries worldwide and have led to large-scale adoption [3]. For example, through assembly technology, the Kajstaden Tall Timber Building project in Sweden not only shortened the construction time by 50% but also achieved the environmental protection goal of reducing carbon emissions by 70%. Daiwa House Industry of Japan has adopted the technology of prefabricated concrete components, which has shortened the construction period of buildings by 40% and reduced on-site workers by 60%. These successful cases have driven the rapid development of the global prefabricated building market. At present, the prefabricated building industry in China is in a rapid development stage; however, its development is still restricted by issues such as incomplete policies and regulations, weak willingness of enterprises to promote it, and low public recognition [4]. To enhance support for prefabricated buildings, the relevant departments have promulgated a series of planning policies [5]. The “14th Five-Year Plan” explicitly proposed vigorously developing prefabricated buildings and set a goal that the proportion of prefabricated buildings in new buildings should reach more than 30% by 2025. For example, in Hainan Province, the proportion of the construction area of prefabricated buildings in the total construction area of newly built buildings exceeded 70% in 2023. Beijing has put forward the goal that by 2025, the proportion of prefabricated buildings in the total construction area of newly built buildings will reach 55%. According to the development plans for prefabricated buildings proposed by many provinces, the proportion of prefabricated buildings in new constructions has already reached 50% [6].

Prefabricated buildings encounter a multitude of risks during construction. These risks predominantly originate from their distinctive construction approaches. Due to the large volume and weight of prefabricated components, there are high safety risks in the transportation and hoisting process, which can easily lead to safety accidents. For example, during the hoisting operation of a prefabricated residential project in Changsha in 2021, the sudden breakage of the hoisting sling caused the prefabricated structure to fall from a height, resulting in a severe safety accident. In terms of node connections, a certain project in Hong Kong encountered problems with the quality of connection points, which led to overall structural safety hazards. In addition, the stability issue of the temporary support system and the risks brought by the cross-operation of multiple trades cannot be ignored either. Due to improper construction management in a hospital renovation project in China, an accident involving falling objects from heights occurred, resulting in injuries to workers. The above-mentioned accident cases fully demonstrate that there are numerous safety risk factors in the construction process of prefabricated buildings, which seriously restrict their further development. Due to the late start of prefabricated construction in our country, there are still many safety loopholes in the current management system and standards, leading to frequent construction safety accidents [7]. In conclusion, adopting scientific management methods is of great significance for reducing the construction risks of prefabricated buildings. To effectively reduce and avoid the losses caused by risks, it is necessary to conduct a systematic study on risk management during the construction stage of prefabricated buildings [8].

This paper is organized as follows: Section 2 presents the literature review on safety risk management of prefabricated building projects. Section 3 describes the research methodology of this paper, including the subjective and objective weight assignment methods as well as the relevant theoretical knowledge of the cloud model. Section 4 introduces a specific engineering case and utilizes the case to validate the proposed evaluation model. Section 5 discusses the model and puts forward relevant management suggestions. Section 6 summarizes the entire paper.

2. Literature Review

With the development of prefabricated buildings, many scholars have conducted extensive research on the construction risk management of prefabricated buildings and achieved certain theoretical research results. These studies have promoted the development of risk assessment methods for prefabricated building construction from different perspectives. In the application of fuzzy evaluation models, scholars have proposed various innovative methods to enhance the accuracy of risk assessment in the construction of prefabricated buildings. Ali Bavafa et al. [9] integrated the use of DEMATEL and fuzzy analysis techniques to systematically identify key safety factors in construction and develop a comprehensive risk prevention and control framework. Yang [10] innovatively constructed a fuzzy neural network evaluation model. The model integrates the advantages of qualitative knowledge and quantitative calculation, and its feasibility and applicability have been verified through extensive sample training. Amiri et al. [11] effectively enhanced the level of construction safety control by introducing the concept of fuzzy mathematics to construct an evaluation model. Shojaei [12] developed an integrated risk management model for the entire engineering process by incorporating fuzzy mathematics and the grey relational degree method. These risk assessment models have significantly enhanced the accuracy and applicability of construction risk assessment for prefabricated buildings by integrating multi-source uncertainty quantification methods. However, the majority of these risk assessment models depend on historical data, making them inherently limited in their ability to adapt to dynamic changes at construction sites. In terms of data research, Fard et al. [13] analyzed accident cases related to prefabricated buildings in the United States and identified structural instability as the primary cause of frequent accidents. Liao [14] applied interval approximation theory to assess risk levels, taking into account the interval fluctuation characteristics of risk evaluation indicators. Cao [15] constructed a structural equation model to deeply analyze the mutual influence relationships among various risk indicators under the EPC model, and verified the logical rationality of the model with the aid of scanning electron microscopy technology. Liu [16] established an Interpretative Structural Modeling (ISM) framework through a literature review and questionnaire surveys to identify the interrelationships among various influencing factors. The proposed method effectively enhances safety performance throughout the construction process and contributes to the reduction in safety incidents. Dang et al. [17] used literature analysis and expert interview methods to systematically identify 27 risk factors affecting the sustainability of prefabricated component construction, and prioritized these factors by combining entropy method and fuzzy analytic hierarchy process (FAHP). Based on the analysis results, the research focuses on the causes of the first five key risk factors, and then proposes targeted improvement measures to improve the sustainability level of prefabricated component construction. Wang [18] conducted a statistical analysis of the relevant literature and developed a questionnaire to assess construction risks in prefabricated buildings based on the Theory of Planned Behavior. The primary risk factors contributing to construction accidents were systematically summarized, and a comparative analysis was conducted to identify the differences in risk characteristics between prefabricated buildings and traditional cast-in-place buildings. Most of these studies adopt a static analytical framework, which makes it difficult to effectively capture the time-varying characteristics of risks. The research data sources are limited to specific cases or regions, and the ability to handle the coexistence of fuzziness and randomness in complex risks is insufficient.

Previous studies have primarily employed methods including case analysis, literature review, expert interviews, and experience-based summaries to identify construction safety risk factors [19,20]. Meanwhile, methods such as the fuzzy comprehensive evaluation approach, the fuzzy neural network evaluation model, and the integration of fuzzy mathematics with grey relational analysis have been applied to assess construction risks in prefabricated building projects [21,22,23]. Most studies focus on a single type of risk, with their evaluation models primarily relying on historical data. The static modeling approaches employed are insufficient for accurately capturing the dynamic evolution process of risks. To enhance the accuracy and efficiency of risk assessment, this study combines the G1-CRITIC method with the cloud model to construct a risk assessment model suitable for the construction of prefabricated buildings. The proposed evaluation model fully takes into account the fuzziness and randomness of risk factors, and more effectively reflects the objective characteristics of dynamic risk changes. This method provides new ideas for other similar studies and model evaluations. At the same time, it has accumulated valuable experience that can be drawn upon by risk managers of similar engineering projects and has significant theoretical value and practical guiding significance.

3. Research Method

This paper first identifies the safety risk factors of the project through a systematic literature review and expert investigation and ultimately constructs a safety risk assessment index system for prefabricated building projects. In this paper, the G1 method is chosen as the subjective evaluation method, and the CRITIC method is selected as the objective evaluation method. Meanwhile, this study incorporates relevant theoretical frameworks from game theory to systematically synthesize the evaluation outcomes of the two methods, thereby deriving the optimal comprehensive weight value. This method can effectively ensure the reliability of the assessment results. Finally, a safety risk assessment model for prefabricated building projects is developed through the integration of the cloud model. An overview of the research concept is shown in Figure 1.

3.1. Identification and Determination of Safety Risk Factors in Prefabricated Construction Projects

The construction risks of prefabricated building projects mainly include personnel operation risks, construction environment risks, and construction management risks. In this paper, the literature analysis method is adopted for the initial identification of influencing factors of project safety risks. Firstly, a literature search was carried out with the keywords “prefabricated buildings” and “risk assessment”. In this paper, 10 representative research papers were selected as the statistical subjects. The summarized influencing factors serve as the foundational basis for the construction of the evaluation index system. Then, based on the WSR methodology, the identified influencing factors are systematically classified and hierarchically ranked from the three dimensions of physical (W), system (S), and human (R) [24]. Finally, this paper identifies 38 influencing risk factors from five aspects: material factors, environmental factors, personnel factors, management factors, and technical factors. The content is shown in

Table 1. Screening of safety risk factors in prefabricated construction projects.

Classification	Number	Influence Factors	Reference	Frequency
Personnel	1	Technical level of construction personnel	[25,26,27,28,29,30,31,32,33,34]	10
	2	Security personnel configuration	[25,26,27,28,29,32]	6
	3	Health status of construction personnel	[27,28,30,34,35]	5
	4	Safety awareness of construction personnel	[25,26,27,28,29,30,32,33,34,35]	10
	5	Degree of coordination between jobs	[26]	1
Management	6	Site safety management system	[26,27,29,30,33,34]	6
	7	Emergency management program	[25,26,29,30,32,33,34]	7
	8	Safety measures for component hoisting	[25,27,28,35]	4
	9	Implementation of security measures	[29,35]	2
	10	Storage of finished goods	[27,28,32,34]	4
	11	Safety supervision inspection	[25,30,33,35]	4
	12	Protective measures for working at heights	[32]	1
	13	Safety education and training	[29,30,33,34,35]	5
	14	Coordination management level	[34]	1
	15	Layout of temporary facilities	[32]	1
Technology	16	Deep design of embedded parts	[25,27,28]	3
	17	Component connection node design	[27,28,30,31,33,36]	6
	18	Component connection positioning technology	[25,27,28,30,34,36]	6
	19	Construction scheme	[25,26,29,30,33,34]	7
	20	Hoisting plan and supervision	[26,32,33]	3
	21	Hoisting technology	[29,31,34,35]	4
	22	Tower crane cross-operation	[35]	1
	23	Other technical recovery degree	[35]	1
	24	Exterior wall connection technology	[27]	1
Environment	25	Temporary hole protection	[28]	1
	26	State of the natural environment	[27,30,33,34]	4
	27	Construction site environmental status	[25,27,28,30,33,34,36]	6
	28	Surrounding environment of hoisting process	[26,29,31,32]	4
	29	Conditions of the surrounding transportation environment	[25,27,28,32]	4
	30	Safety standards policy environment	[25,27,28,29,34]	5
Materials	31	Reliability of mechanical equipment	[29,30]	2
	32	Hoisting machine selection	[25,26,27,28,32,33,34]	7
	33	Precast component quality	[25,26,28,29,30,32,33,34]	8
	34	Regular inspections of equipment	[25,27,28,33,34]	5
	35	Equipment maintenance	[26,29,30,36]	4
	36	Temporary support bearing strength	[25,26,27,28,29,30,34]	7
	37	Site stacking of components	[25,29,30,32,35]	5
	38	The carrying capacity of lifting equipment	[30]	1

.

In this paper, a statistical analysis was conducted on the influencing factors of safety risks in prefabricated building projects, and 29 factors with an occurrence frequency greater than one were identified. To further ensure representativeness and uniqueness, the method of eliminating indicators with high similarity through difference degree comparison was employed. The attribute levels of the indicators were classified into 1 to 5 levels, and 10 experts were invited to rate the levels of the initially screened 29 influencing factors. Taking the technical influencing factor module as an example, the attribute level ratings of the indicators are presented in

Table 2. Index attribute grade score.

Classification	Number	Influence Factors	Index Attribute Grade Score
			5	4	3	2	1
Technology	1	Deep design of embedded parts	6	3	1	0	0
	2	Component connection node design	4	2	2	2	0
	3	Component connection positioning technology	4	2	2	1	1
	4	Construction scheme	5	2	2	0	1
	5	Hoisting plan and supervision	3	2	2	3	0
	6	Hoisting technology	2	3	2	3	0

.

The normalization of the rating results for each level is conducted by employing the formula $x_i/m$ (where i represents the influencing factor and m denotes the number of experts), as presented in

Table 3. Normalization of rating results.

Classification	Number	Influence Factors	Index Attribute Grade Score
			5	4	3	2	1
Technology	1	Deep design of embedded parts	0.6	0.3	0.1	0.0	0.0
	2	Component connection node design	0.4	0.2	0.2	0.2	0.0
	3	Component connection positioning technology	0.4	0.2	0.2	0.1	0.1
	4	Construction scheme	0.5	0.2	0.2	0.0	0.1
	5	Hoisting plan and supervision	0.3	0.2	0.2	0.3	0.0
	6	Hoisting technology	0.2	0.3	0.2	0.3	0.0

.

Based on the normalized data processing, the degree of difference of technical influence factors is computed using Equation (1) [34].

$$\begin{array}{c}{r}_{ij}={\displaystyle \sum _{K=1}^n}\left|x_{ik}-x_{jk}\right|·S_K\end{array}$$

(1)

In Equation (1), $r_{ij}$ denotes the degree of mutual difference between influencing factor i and influencing factor j. $x_{ik}$ and $x_{jk}$ represent the evaluation results of influencing factor i and influencing factor j, respectively, at $S_K$. $S_K$ denotes the grade score corresponding to the Kth grade. The grade scores are numerical values within the range of 1 to 5. When $r_{ij}$ < 0.5, it suggests that there is a relatively high degree of correlation between influencing factor i and influencing factor j. In the process of screening indicator factors, only one of these influencing factors needs to be retained.

The results obtained by normalizing the data in Table 2 and conducting the difference degree calculation in accordance with Equation (1) are presented as follows:

$$\begin{gathered} r_{12}=2.1,r_{13}=2,r_{14}=1.3,r_{15}=2.8,r_{16}=2.9,r_{23}=0.3,r_{24}=1,r_{25}=0.7,r_{26} \\ =1.6,r_{34}=0.7,r_{35}=1,r_{36}=1.9,r_{45}=1.7,r_{46}=2.5,r_{56}=0.9 \end{gathered}$$

From the above results, it can be seen that $r_{23}<0.5$, indicating that factor 2 and factor 3 have a high degree of correlation. Therefore, the article retains the influence factor of the connection point design of the component. The calculation results indicate that five groups of influencing factors demonstrate a high level of correlation. The paper retains the influence factors of component lifting safety measures, connection point design, construction site environmental conditions, lifting scheme and supervision, and equipment regular inspection. The construction risk assessment index system of prefabricated building projects is shown in Figure 2.

3.2. The Evaluation System Based on the Combined Weighting of Game Theory and Cloud Model

3.2.1. Indicator Weighting Based on the G1-CRITIC Method

The G1 method, also known as the order relationship analysis method, is a subjective weight assignment approach that does not require a one-time verification compared to the Analytic Hierarchy Process (AHP) [37]. To accomplish the subjective weight assignment using the G1 method, the calculation steps are as follows:

The first step is to determine the ordering relationship. Rank the importance of $n$ indicators $({\mathrm{x}}_1\succ {\mathrm{x}}_2\succ {\mathrm{x}}_3\succ \cdots {\mathrm{x}}_{\mathrm{n}})$.

The second step is to determine the importance of adjacent indicators. The ratio of the importance of the adjacent index $x_k$ to $x_{k-1}.$

$$\begin{array}{c}{B}_k={\displaystyle \frac{x_{k-1}}{x_k}}\left(k=2,3\cdots ,n\right)\end{array}$$

(2)

In Equation (2), the weight of $x_{k-1}$ is $B_k$-fold that of $x_k$, and the value of $B_k$ is determined as shown in

Table 4. Reference table for assignment values.

${\mathbf{B}}_{\mathbf{k}}$	Value Description
1.0	$\mathrm{Indicator}{x}_{k-1}$  $\mathrm{is}\mathrm{of}\mathrm{equal}\mathrm{importance}\mathrm{to}\mathrm{Indicator}{x}_k$
1.2	$\mathrm{Indicator}{x}_{k-1}$  $\mathrm{is}\mathrm{marginally}\mathrm{more}\mathrm{important}\mathrm{than}\mathrm{indicator}{x}_k$
1.4	$\mathrm{Indicator}{x}_{k-1}$  $\mathrm{is}\mathrm{substantially}\mathrm{more}\mathrm{important}\mathrm{than}\mathrm{indicator}{x}_k$
1.6	$\mathrm{Indicator}{x}_{k-1}$  $\mathrm{is}\mathrm{much}\mathrm{more}\mathrm{important}\mathrm{than}\mathrm{indicator}{x}_k$
1.8	$\mathrm{Indicator}{x}_{k-1}$  $\mathrm{is}\mathrm{of}\mathrm{extreme}\mathrm{importance}\mathrm{compared}\mathrm{to}\mathrm{indicator}{x}_k$
1.1,1.3,1.5,1.7	The mid-value between the aforementioned two adjacent judgments

.

The third step is to determine the weight coefficient $W_n$. According to the value of $B_k$, the subjective weight of the evaluation index is determined by referring to Formulas (3) and (4).

$$\begin{array}{c}{W_n=\left(1+{\displaystyle \sum _{k=2}^n}{\displaystyle \prod _{i=k}^n}{B}_k\right)}^{-1}\end{array}$$

(3)

$$\begin{array}{c}{W}_{n-1}=B_k{W}_n\end{array}$$

(4)

The CRITIC method is an objective weight assignment approach that transforms the entropy weight method, and its fundamental concept proceeds from two aspects. One is the contrast intensity, and the other is the conflict among the evaluation indicators [38]. This method is applicable for judging the stability of the data and is suitable for analyzing data where there are certain correlations among the indicators or factors. To complete the objective weight assignment using the CRITIC method, the calculation steps are as follows:

The first step is data standardization.

If the larger the index value, the better, the positive indicator is used, as shown in Equation (5).

$$\begin{array}{c}\begin{array}{c}{x^{\prime }}_{ij}={\displaystyle \frac{x_{ij}-min\left\{x_j\right\}}{max\left\{x_j\right\}-min\left\{x_j\right\}}}\end{array}\end{array}$$

(5)

If the smaller the index value, the better, the reverse index is used, as shown in Equation (6).

$$\begin{array}{c}{x^{\prime }}_{ij}={\displaystyle \frac{max\left(x_j\right)-x_{ij}}{max\left(x_j\right)-min\left(x_j\right)}}\end{array}$$

(6)

The second step is to calculate the information load.

Volatility, where $\overline{x_j}$ is the mean value of each indicator column, as shown in Equation (7).

$$\begin{array}{c}{S}_j=\sqrt{{\displaystyle \frac{{\sum }_{i=1}^m{\left(x_{ij}-\overline{x_j}\right)}^2}{n-1}}}\end{array}$$

(7)

Conflict, where $r_{ij}$ represents the correlation coefficient between index $i$ and index $j$, as shown in Equation (8).

$$\begin{array}{c}{A}_j={\displaystyle \sum _{i=1}^n}\left(1-r_{ij}\right)\end{array}$$

(8)

The amount of information is shown in Equation (9).

$$\begin{array}{c}{C}_j=S_j\times A_j\end{array}$$

(9)

The third step is to calculate the weight, as shown in Equation (10).

$$\begin{array}{c}{W}_j={\displaystyle \frac{C_j}{{\sum }_{j=1}^n{C}_j}}\end{array}$$

(10)

3.2.2. The Game Combination Weighting Method Determines the Comprehensive Weight

The game theory-based combination weighting method systematically investigates the coordination and consistency between subjective and objective weighting approaches under conflict conditions, with the aim of achieving an optimized balance between them. It optimizes the weights calculated by different methods to find the optimal weight value [39,40]. The calculation steps are as follows:

In the first step, the weight obtained by the G1 method and the CRITIC method is combined and optimized, and the deviation of weight is minimized. See Equation (11) for a linear equation equivalent to the optimal first derivative condition based on the differential properties of the matrix.

$$\left[\begin{array}{cc}{w}_1{w}_1^T& w_1{w}_2^T\\ w_2{w}_1^T& w_2{w}_2^T\end{array}\right]\left[\begin{array}{c}\alpha 1\\ \alpha 2\end{array}\right]=\left[\begin{array}{c}{w}_1{w}_1^T\\ w_2{w}_2^T\end{array}\right]$$

(11)

$w1$ is the set of subjective weight vectors; $w2$ is the set of objective weight vectors;$\alpha 1$ is the coefficient of the G1 method;$\alpha 2$ is the coefficient of the GRITIC method.

In the second step, the optimized combination coefficients are calculated and normalized, as shown in Equations (12) and (13).

$${\alpha }_1^{*}={\displaystyle \frac{\alpha 1}{\alpha 1+\alpha 2}}$$

(12)

$${\alpha }_2^{*}={\displaystyle \frac{\alpha 2}{\alpha 1+\alpha 2}}$$

(13)

The third step is to calculate the comprehensive weight of indicators, as shown in Equation (14).

$$\begin{array}{c}W={\alpha }_1^{*}{w}_1^T+{\alpha }_2^{*}{w}_2^T\end{array}$$

(14)

3.2.3. The Basic Theory of Cloud Models

The cloud model is a mathematical model put forward by Li Deyi [41], an academic of the Chinese Academy of Engineering. It is capable of converting the uncertain language in the evaluation into quantitative analysis to reduce the randomness and fuzziness of the evaluation indicators [42]. The cloud model is composed of numerous cloud droplets. A cloud typically encompasses three elements: Ex (expected value), En (entropy), and He (hyper entropy). These three numerical characteristics are used to reflect the safety risk information of prefabricated building projects. Among them, the expected value Ex indicates the point value that most representatively characterizes the risk classification level of prefabricated building projects; entropy En and hyper entropy He represent the interval and thickness, respectively, of the cloud droplets, reflecting the degree of dispersion [43,44].

(1)

Standard cloud

The risk is divided into different interval levels, and each interval level corresponds to its own sub-interval. The calculation formula of the three characteristic numbers of the standard cloud corresponding to the sub-interval is shown in Equation (15).

$$\left\{\begin{array}{l}Ex={\displaystyle \frac{\left(T_{min}+T_{max}\right)}{2}}\\ En={\displaystyle \frac{\left(T_{max}-T_{min}\right)}{6}}\\ He=k\end{array}\right.$$

(15)

(2)

Index comment cloud

Invite n experts to rate the safety risk evaluation indicators of prefabricated building projects and calculate the digital characteristic values of the evaluation cloud. The specific calculation formula is presented as Equation (16).

$$\begin{array}{c}\left\{\begin{array}{l}{Ex}_j={\displaystyle \frac{\sum _{i=1}^n{x}_{ij}}{n}}\\ {En}_j=\sqrt{{\displaystyle \frac{\pi }{2}}}·{\displaystyle \sum _{i=1}^n}{\displaystyle \frac{\left|x_{ij}-{Ex}_j\right|}{n}}\\ {He}_j=\sqrt{\left|s_j^2-{En}_j^2\right|}\\ s_j^2={\displaystyle \sum _{i=1}^n}{\displaystyle \frac{{\left(x_{ij}-{Ex}_j\right)}^2}{\left(n-1\right)}}\end{array}\right.\end{array}$$

(16)

In Equation (16), n is the number of experts; m is the number of risk indicators; $s_j^2$ is the sample variance.

(3)

Comprehensive comment cloud

According to the comprehensive comment cloud algorithm, the determined index weight is combined with the index cloud parameters of each index to calculate the comprehensive comment cloud ${\mathrm{C}}^{\mathrm{*}}$. The specific calculation formula is shown in Equation (17).

$$\begin{array}{c}\left\{\begin{array}{c}Ex={\displaystyle \sum _{j=1}^m}E{x}_j\cdot w_i\\ En=\sqrt{{\displaystyle \sum _{j=1}^m}{En}_j^2\cdot w_i}\\ He={\displaystyle \sum _{j=1}^m}H{e}_j\cdot w_i\end{array}\right.\end{array}$$

(17)

(4)

Determine the level of safety risk assessment

Based on the expert questionnaire, the standard cloud and the comprehensive comment cloud are solved, the contrast cloud map is drawn, and the risk level is determined according to the relative positions of the two in the same coordinate. The nearest standard cloud rating is the final security evaluation rating [45].

4. Case Analysis

4.1. Project Overview

The W prefabricated building project adopts the construction master contract model, and the high-rise residences involved include G8#, G9#, and G12#, which are the core of the building project. Due to the height of the building, the complex structure, and the difficulty of construction, a number of technical innovation measures have been taken to ensure project quality. The plan actively introduces the latest technology, and the use of fully shear concrete as exterior wall coating during construction can effectively improve the thermal insulation and waterproof performance of the exterior wall. The project achieved an assembly rate of about 30% to meet the requirements of the government land transfer conditions.

4.2. Weights Calculation

In accordance with the determined safety risk evaluation index system, ten related experts and construction engineers were invited to rank and score the 24 secondary indicators based on their importance. Firstly, the G1 method and the CRITIC method were employed, and Equations (2) to (4) and (5) to (10) were utilized to determine the subjective and objective weights of the indicators. The results are shown in

Table 5. Subjective and objective weight values of secondary indicators.

First-Level Indexes	Second-Level Indexes	G1 Method Subjective Weight	CRITIC Method Objective Weight
A1	A11	0.0838	0.0681
	A12	0.0430	0.0466
	A13	0.0762	0.0666
	A14	0.0247	0.0373
A2	A21	0.0296	0.0382
	A22	0.0326	0.0493
	A23	0.0520	0.0423
	A24	0.0326	0.0381
	A25	0.0572	0.0456
	A26	0.0206	0.0377
A3	A31	0.0629	0.0495
	A32	0.0326	0.0377
	A33	0.0247	0.0501
	A34	0.0572	0.0484
A4	A41	0.0358	0.0398
	A42	0.0430	0.0344
	A43	0.0473	0.0377
	A44	0.0692	0.0431
	A45	0.0247	0.0422
	A46	0.0296	0.0414
A5	A51	0.0296	0.0437
	A52	0.0187	0.0348
	A53	0.0520	0.0368
	A54	0.0206	0.0405

.

The commonly employed comprehensive weighting method determines the comprehensive weight through a weighted combination. The coefficients of the subjective weight and the objective weight are each set to 0.5, so that the subjective weight and the objective weight each account for 50%. The game-theory-based combination weighting approach is capable of identifying the equilibrium outcome between subjective and objective weights. The comprehensive weight is calculated based on Equations (11) to (14). The calculation results of the aforementioned two types of comprehensive weights are presented in

Table 6. Comparison of comprehensive weight values.

First-Level Indexes	Second-Level Indexes	Comprehensive Weights (Game Theory)		Comprehensive Weights (Average Value Method)
		Weight Value	Ranking	Weight Value	Ranking
A1	A11	0.0806	1	0.0760	1
	A12	0.0450	10	0.0448	8
	A13	0.0748	2	0.0714	2
	A14	0.0291	21	0.0310	21
A2	A21	0.0329	19	0.0339	19
	A22	0.0385	12	0.0410	11
	A23	0.0501	7	0.0472	7
	A24	0.0350	14	0.0354	17
	A25	0.0548	6	0.0514	6
	A26	0.0263	23	0.0292	23
A3	A31	0.0601	4	0.0562	3
	A32	0.0349	15	0.0352	18
	A33	0.0331	18	0.0374	14
	A34	0.0557	5	0.0528	5
A4	A41	0.0378	13	0.0378	13
	A42	0.0412	11	0.0387	12
	A43	0.0453	9	0.0425	10
	A44	0.0625	3	0.0562	4
	A45	0.0307	20	0.0335	20
	A46	0.0339	17	0.0355	16
A5	A51	0.0346	16	0.0367	15
	A52	0.0241	24	0.0268	24
	A53	0.0484	8	0.0444	9
	A54	0.0272	22	0.0306	22

.

As can be inferred from the calculation results of the two types of comprehensive weights presented in Table 6, for the secondary indicators A12, A22, A24, A31, A32, A33, A42, A43, A44, A46, A51, and A53, there are certain changes in the ranking of their weight values. This study adopts the game theory combined weighting method to optimize the allocation of subjective and objective weights to ensure the rationality of weight allocation. This method effectively enhances the scientificity and reliability of weight distribution by seeking the optimal equilibrium solution of subjective and objective weights.

4.3. Build the Project Standard Cloud

The standard cloud serves as the benchmark reference diagram for the safety risk evaluation of the project. The safety risk evaluation grades are divided into five intervals: extremely low risk [90,100], relatively low risk [75,90), medium risk [50,75), relatively high risk [25,50), and extremely high risk [0,25). By using MATLAB (v23.2) software and conducting 2000 calculations, the standard cloud diagram of the evaluation index with the optimal cloud entropy is generated. The digital characteristics of the standard cloud in each rating interval are presented in

Table 7. Standard cloud model digital features.

Risk Level	Score Interval	Standard Features
Very low risk	[90,100]	(95.0,1.667,0.5)
Low risk	[75,90)	(82.5,2.500,0.5)
Medium risk	[50,75)	(62.5,4.167,0.5)
High risk	[25,50)	(37.5,4.167,0.5)
Very high risk	[0,25)	(12.5,4.167,0.5)

. The standard cloud diagram is shown in Figure 3.

4.4. Build the Project Comment Cloud

According to the survey results of the experts, the data were quantitatively processed, and the comment cloud of all levels of indicators was solved by Equation (16), and the results are shown in

Table 8. All levels of indicators comment cloud value.

First-Level Indexes	Comment Cloud	Second-Level Indexes	Comment Cloud
A1	(82.3,3.2531,0.6769)	A11	(77.0,3.7599,0.5995)
		A12	(87.0,3.5093,0.3579)
		A13	(80.3,2.4565,0.6839)
		A14	(72.2,2.6069,1.1755)
A2	(84.6,2.7117,0.7932)	A21	(89.6,2.6069,0.5997)
		A22	(87.0,2.7573,0.2170)
		A23	(71.2,2.0554,0.5189)
		A24	(87.3,3.1333,1.6912)
		A25	(77.5,2.8826,1.0690)
		A26	(84.2,2.4565,0.3787)
A3	(88.1,3.3724,0.5597)	A31	(84.3,3.8853,0.4420)
		A32	(87.2,3.0080,0.1374)
		A33	(80.5,3.5093,0.5202)
		A34	(86.9,2.6320,0.9095)
A4	(82.9,2.8063,0.6379)	A41	(80.9,1.8800,0.3177)
		A42	(80.6,1.7546,0.7220)
		A43	(80.4,1.9552,0.6663)
		A44	(81.6,3.0080,0.8838)
		A45	(73.4,4.0106,0.2013)
		A46	(80.6,3.5093,0.6293)
A5	(90.2,4.1774,0.9961)	A51	(87.1,5.1637,0.8863)
		A52	(87.9,3.4090,0.7953)
		A53	(90.0,2.5066,1.1280)
		A54	(80.9,5.2639,0.9055)

. Then, according to Equation (17), the determined index weights and index cloud parameters are combined to calculate the comprehensive comment cloud of the project as ${\mathrm{C}}^{\mathrm{*}}$= (85.0, 3.1962, 0.7147).

4.5. Determine the Project Construction Safety Risk Level

Based on the calculation results of the first-level evaluation cloud and the comprehensive evaluation cloud, cloud diagrams were generated using MATLAB (v23.2) software. The cloud diagram representing the first-level indicators is shown in Figure 4, while the comprehensive evaluation cloud diagram is shown in Figure 5.

It can be seen from Figure 5 that the red color represents the construction safety risk level of this project, which is low risk. From the perspective of the first-level indicators, as shown in Figure 4, technical factors (A1) and material factors (A4) have a significant impact on the safety risk management of this project. Therefore, it is necessary to take effective preventive measures at the initial stage of the project. Although personnel factors (A3), management factors (A2), and environmental factors (A5) are categorized as very low risk elements with relatively minor impacts on the project construction risk assessment, they still warrant adequate attention due to their potential influence on the overall project risk profile. From the perspective of secondary indicators, construction site environmental status (A53) is at a very low risk level. Deep design of embedded parts (A11), component connection node design (A12), construction scheme (A13), site safety management system (A21), emergency management program (A22), storage of finished goods (A24), safety supervision inspection (A25), safety education and training (A26), technical level of construction personnel (A31), security personnel configuration (A32), health status of construction personnel (A33), safety awareness of construction personnel (A34), hoisting machine selection (A41), precast component quality (A42), temporary support bearing strength (A43), reliability of mechanical equipment (A44), regular inspections of equipment (A46), safety standards policy environment (A51), state of natural environment (A52), and conditions of the surrounding transportation environment (A54) are at a low risk level. Hoisting plan and supervision (A14), safety measures for component hoisting (A23), and site stacking of components (A45) are at a medium risk level.

5. Discussion

5.1. Model Validation

In order to validate the effectiveness of the risk assessment model for prefabricated building construction, which is constructed based on the G1-CRITIC method and the cloud model. This paper employs the fuzzy comprehensive evaluation method to quantitatively assess the membership degrees of key indicators within the risk assessment framework for prefabricated building construction. The evaluation set is V = {very high risk, high risk, medium risk, low risk, very low risk}. The calculated results of the comprehensive membership degree levels are (0.049, 0.078, 0.190, 0.352, 0.330). According to the principle of maximum membership degree, the construction risk level of the project is determined to be low risk. The calculation results obtained in this paper demonstrate consistency with the traditional risk assessment model, thereby further confirming the effectiveness of the G1-CRITIC method and the cloud model in the context of risk assessment. In addition, the evaluation results of the model are consistent with the actual construction conditions of the project, further proving the objectivity and accuracy of the model.

5.2. Management Suggestions

By applying the risk assessment model proposed in this paper to prefabricated building projects, the research results show that the construction risk assessment level of this project is low risk. But it is still necessary to take countermeasures to reduce the construction risks.

In terms of personnel management, enhancing the skill levels of personnel and strengthening their safety awareness are the keys to ensuring the smooth progress of the project. In the construction preparation phase, it is necessary to provide employees with comprehensive operation manuals and technical guidance documents to ensure that they have a thorough understanding of the relevant procedures and specifications. By means of systematic training and management measures, the safety awareness of those engaged in the construction industry can be effectively heightened, thus safeguarding safety throughout the construction process.

In terms of technical management, measures including enhancing the quality of drawings, improving design efficiency, optimizing the processing technology of embedded parts, and intensifying quality management can effectively elevate the overall standard of the detailed design of embedded parts. Before developing the hoisting plan, a comprehensive and systematic evaluation, together with scientifically based planning of the overall scheme, should be conducted. Hoisting operation procedures must be clearly defined, and detailed operational guidelines and safety protection measures should be established for each stage of the operational process.

In terms of material management, it is necessary to enhance the quality inspection of factory-produced components, strictly implement the incoming inspection and acceptance procedures, and effectively prevent non-compliant products from entering the construction site. It is necessary to scientifically plan and rationally arrange the stacking layout of components based on the specific conditions of the construction site. Furthermore, considering the technical specifications of hoisting operations, the selection of hoisting equipment should be conducted in a systematic and evidence-based manner. Moreover, a structured mechanism for periodic updating and maintenance of the equipment must be implemented to ensure its continuous optimal performance.

In terms of environmental management, concerning the safety standards policy environment, it is imperative to rigorously comply with national and industry-level regulations in order to ensure full conformity with the prescribed review criteria for prefabrication rate and assembly rate. Considering the specific conditions of the surrounding transportation environment, route reconnaissance should be conducted in advance, and transportation vehicles complying with applicable regulatory standards should be selected.

In terms of the safety management system, to enhance the safety management standards in prefabricated building construction, it is essential to establish a specialized safety management system tailored specifically for prefabricated construction projects. For high-risk operational scenarios such as component overturning, specialized emergency response plans should be developed, and practical emergency drills should be systematically organized. Specialized emergency rescue equipment should be strategically deployed at construction sites to ensure an efficient and timely response to construction-related risk incidents.

6. Conclusions

The development of prefabricated construction in China is becoming increasingly rapid, but it is also accompanied by frequent construction safety accidents. Therefore, from both social and economic perspectives, the analysis of safety risks in prefabricated building construction plays a critical role in enhancing safety management throughout the entire construction process. In this paper, the G1-CRITIC method is integrated with the cloud model and systematically applied to risk assessment in prefabricated building construction, leading to the development of a novel evaluation framework. This model demonstrates strong predictive capability in identifying potential risks throughout the construction process, thereby facilitating the mitigation of construction-related risks and safety hazards. This paper reaches the following conclusions:

(1)

Compared with previous construction risk assessment models, the construction risk assessment model proposed in this paper is based on the integration of the G1-CRITIC method and the cloud model. By focusing on a specific category of risk, the model systematically incorporates the ambiguity and stochastic characteristics of risk factors, thereby enhancing the objectivity and accuracy of dynamic risk characterization. Meanwhile, this evaluation model provides new ideas for other similar studies and offers experience to risk managers of similar projects, which is of significant guiding value.

(2)

Based on the research results of this article, from the perspective of the first-level indicators, technical factors and material factors have a greater impact on the safety risk management of this project, while personnel factors, management factors and environmental factors have a relatively smaller influence on the construction risks. From the perspective of the secondary indicators, the key points of risk management include the proper stacking and protection of materials at the construction site, the optimization of embedded part design, and the upgrading of hoisting plans. By implementing effective management measures, the incidence of engineering safety accidents has decreased by approximately 30%, and the efficiency of hazard identification has also significantly improved. Strengthened management of the above key factors supports risk control and enhances construction safety.

(3)

To rigorously validate the effectiveness of the established evaluation model, this paper applies the conventional fuzzy comprehensive evaluation methodology to systematically analyze and quantitatively assess the membership degrees of key risk indicators in prefabricated building construction. The final determination is that the construction risk level of this project is low risk, further confirming the effectiveness of the G1-CRITIC method and cloud model in risk assessment. Moreover, the model’s evaluation results align with the actual construction conditions of the project, further validating the objectivity and accuracy of the model.

(4)

Based on the existing research, future studies will subsequently consider the dynamic changes of risks. We intend to incorporate Internet of Things technologies and introduce a real-time data acquisition system, such as real-time monitoring of environmental temperature and humidity, wind speed during high-altitude operations, and using sensors to monitor the lifting load capacity of components. In the course of prefabricated building construction, the probability of construction risks can be quantified by means of machine learning algorithms. Moreover, the incidence of construction risks can be mitigated through proactive intervention strategies. To address the dynamic nature of risks related to environmental conditions, material properties, management practices, and personnel performance during the construction process, a comprehensive strategy incorporating monitoring, prediction, and intervention mechanisms will be implemented.