Research on Coupling Effect Measurement and Coupling Risk Simulation of Green Building Construction Safety Risk Factors

Abstract

The construction of green buildings is an important direction for the transformation and development of the construction industry, but it is beset with problems such as a lack of construction experience, immature new technologies, and unstable material properties; these issues bring risks to the construction stage of green buildings, and the coupling of uncertain risk factors in the construction process of green buildings may lead to unfavorable results. The purpose of this study is to explore the coupling degree of green building construction safety risk factors and the changing trend in their coupling combinations at the system risk level. First, the risk factor index system was defined by reading the literature and gathering expert opinions, and the coupling degree between risk factors was measured using an improved coupling degree model. Then, a system dynamics model was established to simulate and analyze the coupling effects among the risk factors and determine the combinations with the greatest influence. The results show that the risk probability is proportional to the risk coupling value, the human–environment coupling value is the largest, and the material equipment–management coupling value is the smallest. The human–environment system simulation shows that reducing the coupling value of system factors will promote a decrease in the total level of system risk. According to the research conclusions, measures to prevent risk coupling are proposed, which offer theoretical references for green building practitioners carrying out risk management; these measures hold a certain guiding significance for the risk control and future development of green buildings.

1. Introduction

The development of the construction industry plays an important role worldwide. This industry consumes vast amounts of raw materials (nearly 3 billion tons) every year, including great quantities of wood and accounting for about half of the global steel production. It also involves massive energy and water consumption, accounting for about 30% and 16%, respectively [1]. Currently, during the whole life cycle, construction activities have certain negative impacts on the environment, economy, and society, which hinder the industry’s sustainable development [2]. These not only affect people’s lives, but also destroy the ecological environment. In order to overcome various problems, the construction industry has begun to transform from outputting general buildings to green buildings, with emphases on the environment and society [3]. A study shows that green buildings, as a kind of high-performance building, can not only reduce environmental pollution, but also improve people’s physical health and reduce resource consumption, such as raw materials, water, and energy [4]. In addition, another study shows that green building is a new phenomenon, which aims to reduce the adverse impact of buildings on the environment throughout their life cycle [5].

Accordingly, green buildings have attracted the attention of many countries. Since the 20th century, their construction has been increasing annually, and corresponding regulations and policies have been implemented and promulgated [6]. Achieving sustainability, good construction quality, a high degree of safety, and timely delivery are among the objectives of green building projects [7].

Nevertheless, although green buildings have advantages over general buildings, they also have risk characteristics throughout their whole life cycle (spanning decision-making, design, construction, trial operation, and operation and maintenance), where different stages will involve specific risks. The main obstacles to constructing green buildings in the decision-making stage are government bureaucracy and complicated approval procedures. In the design stage, the probability of risk occurrence is relatively small, but the degree of harm is large. The most important risk involves the low experience level of the designer. They are faced with stricter requirements for green buildings than those of general buildings, such as the pursuit of economic benefits over the whole life cycle of green buildings, the application of innovative construction technology, and the conservation of energy resources. In the construction stage, the risk occurrence is the most concentrated, with random and uncertain risk factors arising, and one of the most important problems to be solved is the lack of experienced green building practitioners. This is a concern since their construction is relatively complex compared to general buildings and involves new equipment, green materials, and new technologies, along with participation in a certification process and local policy incentives, which adversely affect the project if improperly handled [8]. Furthermore, the green building construction stage involves converting design drawings into entities, and those involved need to ensure the performance of green buildings in the trial operation stage; so, the construction stage plays a key role in linking the past and the future. Moreover, the external environment of the green building construction stage is complex and changeable. As such, during the green building construction stage, risks are most likely, and the environmental impact is most serious [9]. Then, in the operation and maintenance stage, the most important risk is an ineffective level of management from the property company [10].

If a risk factor arises during the green building stage, the project manager can take targeted measures to avoid it from harming the work; however, if multiple risks occur at the same time, the manager may not be able to achieve timely prevention and control, which may lead to the expansion of risks and adverse ecological environmental results. Accordingly, in this study, we comprehensively applied the risk coupling theory and improved the coupling degree model and system dynamics model to analyze the coupling effect of risk factors in the green building construction process, aiming to avoid coupling between key risk factors, and to control green building construction coupling risks effectively.

The contributions of this paper are as follows:

(1) This study presents the degree of coupling between risk factors in green building construction, which is calculated using an improved coupling degree model. The higher the coupling degree between the factors, the greater their adverse impact on the construction of green buildings.

(2) The dynamic coupling relationships between risk factors are charted using an improved coupling degree model. For the two first-level risk factors with the largest coupling degree, a causality diagram and corresponding function equation between the influencing factors are presented, to provide decision support for the risk management of green construction projects in China.

This paper is organized as follows: Section 2 offers a literature review, including the latest research on green building risks and risk coupling models; Section 3 introduces the risk coupling analysis and research methods; in Section 4, the simulation of a subsystem is discussed; Section 5 gives the results and analysis; and Section 6 summarizes the contributions of this study and the limitations of this paper.

2. Literature Review

2.1. Research on Green Building Risks

Four aspects of the environmental protection of green buildings have been studied by scholars, along with their greater risks than ordinary buildings. Taking Vietnam as an example, Nguyen and Macchion [11] conducted an exploratory factor analysis on the risk factors of green building projects and revealed the six most influential risk factors. Nguyen and Macchion [12] analyzed 64 studies on the risks of green buildings from 2006 to 2020, systematically reviewed the past studies, and proposed future research directions. Yang et al. [13] analyzed the risk network relationship in complex green building projects based on social network analysis (SNA) and studied the risk association relationships of project participants. Zhao et al. [14] established a risk assessment model using the fuzzy comprehensive evaluation method to assess the risks associated with green building projects in Singapore. Thanu et al. [15] developed the Building Performance Score (BPS), a tool used to evaluate the performance of green buildings. Koc et al. [16] adopted a combination of fuzzy AHP and fuzzy TOPSIS to establish a whole life cycle risk management framework for assessing the risk factors of green building projects, filling gaps in the existing literature. Recognizing that green building is a new industry, Nguyen et al. [17] discussed the impact of practitioner characteristics on risk assessment, based on the combination of ANOVA and the AHP, and revealed that green building project professionals with rich experience play an important role in risk management. From the perspective of the whole life cycle and multiple project risks, Guan et al. [18] adopted the integrated method of ISM-MICMAC to study the driving force and dependence of the interdependent factors in green building projects, and they ascertained the key restrictive risk factors. Huang et al. [19] discussed the relationship between the risk occurrence network and the risk hazard network of green building projects, and they showed that the age of their respondents was significantly related to the two networks. Wang et al. put forward a comprehensive review, pointing out that the current challenges faced by risk management research on green building projects include a lack of financial risks related to green building projects, a lack of specific risk assessment models, and insufficient research on risk mitigation strategies at different stages [20].

2.2. Research on Risk Coupling Model

When studying risk coupling, scholars have cited a variety of models to solve coupling problems in different fields. Yan et al. [21] collected information on 108 social stability risk accidents of major projects at home and abroad from 2000 to 2020, and they evaluated social stability risk coupling based on the N-K model. Deng et al. [22] analyzed the coupling of four types of factors, including human–material and equipment–environment–management factors, calculated the coupling values of single factors, double factors, and multiple factors, and conducted a comparative analysis of the coupling characteristics of marine accidents. Guo et al. [23] applied the N-K model and the coupling degree model to analyze the relationship between the tunnel construction risk value and the components of each risk factor. Liu et al. [24] identified key coupling risk factors of subway operation accidents from macro and micro perspectives by integrating the DEMATEL and ISM-NK models. It was Professor Forrest of MIT who first proposed the system dynamics (SD) method. This method can produce a clear causal relationship diagram and stock flow diagram for various factors, and it can simulate and analyze the dynamic changes in coupling risks [25]. Zhang et al. [26] optimized the N-K model with system dynamics to study the coupling of tunnel construction safety risks. Xue et al. [27] studied the system dynamics model to identify risk factors with large coupling effects. Hai et al. [28] established a system dynamics simulation model for risk factors, discussed the evolutionary engineering of common tunnel risks, and identified some combinations of risk factors with great influence.

Meng et al. used the N-K model and Bayesian network (BN) method to analyze collision accidents, and the BN model further quantified the collision risk based on the calculation of the N-K model. The BN method can express the causal relationship and probabilistic dependence relationship between variables. The advantage of the BN method in dealing with uncertainty and logical relationships can reduce subjectivity [29]. However, it is necessary to consider the complex evolution of the system over time, and the BN model lacks the ability to model the flexible structure and the dependency between variables [30]. As a building information platform, the building information model (BIM) is used to analyze building information and enhance the communication process. The most important advantage of the BIM is that it can be updated simultaneously with design changes, which can minimize possible errors and changes in the construction process [31]. The application of the BIM in the design and construction stage can dynamically assess and reduce risks, thus improving the safety of the workplace [32]. Although the BIM can provide detailed spatial information, the overall dynamic ability for dealing with complex systems is limited. Few scholars use the BIM to solve problems when studying risk coupling problems.

Through this literature review, it was found that most of the current studies on the risk direction of green buildings focus on the risk assessment and risk identification of a single factor. Meanwhile, there is little research on the risk coupling between multiple factors, and despite the complexity of the green building construction system, the existing research rarely combines models to explore the coupling effect between risk factors and present a dynamic analysis of the coupling risk. On the one hand, this paper studies this field in the construction stage of green building projects from the perspective of risk coupling, and applies system dynamics flexibly, which is an innovation of this paper. At this stage, it will be affected by many uncertain risk factors. If these factors are not controlled in time, the coupling effect of multiple risk factors will make the system risk level change with the change in time, and the risk occurrence process is dynamic. On the other hand, this paper innovatively proposes a comprehensive research framework that combines the entropy weight–DEMATEL method, improved coupling degree model, and system dynamics model. In the parameter setting of the simulation model, the introduction of the entropy weight–DEMATEL method increases the rationality of the weight distribution, and the parameter of the coupling coefficient is determined according to the improved coupling degree model. Exploring the influence of the coupling relationships between homogeneous factors within a subsystem and heterogeneous factors between subsystems on the overall risk system. According to the simulation results, targeted risk management and control measures are proposed to provide a certain theoretical reference for risk management and prevention in green building construction, so as to realize the expected functions of green buildings.

3. Materials and Methods

In the existing literature, the commonly used models for risk coupling are the N-K model and the coupling degree model. Because the N-K model is based on a large quantity of objective historical data, and green building is an emerging field in which it is difficult to collect objective data, while the coupling degree model is suitable for a small-sample-size analysis, we adopted the latter. Since the traditional coupling degree model too often produces low or high coupling degree values, the coupling degree model was improved by referring to the paper of Yuan [33], whereby a coordination index was introduced to correct the coupling degree. Then, the improved coupling degree model was combined with the system dynamics model. Subsequently, the influence of the variation in the coupling coefficients of homogenous and heterogeneous factors in the subsystems on the overall risk level of the system was explored. Figure 1 outlines our overall research framework.

3.1. Coupling Analysis of Risk Factors

3.1.1. Identification of Risk Factors

We searched a large number of articles in green building construction risk research, using Web of Science, Elsevier, Emerald, CNKI, and other databases to collect articles, and searching with the keywords “green building project” and “risk management” or “green building project” and “risk factor”. The time range of the literature retrieval was from 2010 to 2023, and the quality of the articles, which was highly correlated with the search string, met the Peking University, SCI, CSSCI, CSCD, and EI standards, among others. In total, 23 Chinese and 49 English articles were preliminarily retrieved for screening. In this study, the main entry point was the construction stage, for which 26 risk factors were selected from the collected literature and combined with expert opinions, as shown in

Table 1. Risk factors affecting the green building construction stage.

References
	Hu et al. [34]	Koc et al. [16]	Huang et al. [19]	Nguyen et al. [17]	Wuni et al. [35]	Guan et al. [18]	Shi et al. [36]	Wang et al. [37]	Bai et al. [38]	Jamalu-din et al. [39]
Construction personnel’s physical fitness status	*	—	—	—	—	—	—	—	*	—
Safety protection measures for construction personnel	—	—	—	—	—	*	—	*	—	—
Physiological condition of construction personnel	—	—	—	—	—	—	—	—	*	—
Green building construction experience	—	*	*	—	*	*	—	—	—	—
Green building construction consciousness level	*	—	—	—	—	—	*	—	—	—
New technology application proficiency	*	—	—	*	*	*	—	—	*	—
Completeness of construction site safety protection facilities	—	—	—	—	—	—	—	*	—	—
Operation capacity of green building construction equipment	*	—	*	—	—	—	*	—	—	—
Applicability of green building construction equipment	*	—	—	—	—	*	*	—	—	—
Equipment repair and maintenance measures	*	—	—	—	—	—	—	—	—	—
Aging and wear of green building construction equipment	—	—	—	—	—	*	—	—	—	—
Green building construction materials use technology	—	*	—	—	—	*	*	—	—	—
Green building construction material quality	*	—	—	*	*	—	*	—	—	*
Transportation of materials and components onsite	*	—	—	—	—	—	—	—	—	*
Site material and component storage yard layout	*	—	—	—	—	—	—	—	—	*
Construction noise and vibration pollution	—	—	—	—	—	—	—	*	*	—
Construction dust pollution	—	—	—	—	—	—	—	*	*	*
Construction waste	—	—	—	—	—	—	—	—	—	*
Unexpected bad weather	—	—	—	—	*	*	—	*	*	—
Unexpected site conditions	—	—	*	—	*	—	—	*	*	—
Personnel safety and health management training	*	—	—	—	—	*	—	—	—	—
Protection of the site construction environment	—	—	—	—	—	—	—	*	—	—
Green building construction safety management mechanism perfection	*	—	—	—	—	—	—	—	—	—
Rationality of the green building construction scheme	*	—	—	—	—	*	—	*	—	—
Emergency warning and emergency handling measures	*	—	—	—	—	—	—	—	—	—
Construction waste management	—	—	—	—	—	*	—	*	—	*

“*” indicates that the literature includes the factor and “—” indicates that the literature does not include the factor.

.

Applying system safety theory, the risk factors for green building construction were divided into four subsystems (Wu et al. [40]): human (A: physical quality of construction personnel, experience level of green building construction personnel, etc.), materials and equipment (B: operation capability of green building construction equipment, quality of green building construction materials, etc.), environment (C: unpredictable bad weather, site conditions, etc.), and management (D: personnel safety and health management training, safety management mechanism, etc.). The four subsystems were first-level risk factor indicators, each including second-level indicators, as shown in Figure 2.

3.1.2. Coupling Mechanism Analysis of Risk Factors

The term “coupling” is derived from physics and refers to a phenomenon of mutual influence and connection between two or more systems or modes of motion, which has been widely used in physics, the power industry, building science, chemistry, and other fields [41]. Risk coupling occurs between various factors within the system, affecting the system stability and risk value [42]. Green building construction safety risk factor coupling refers to the mutual influences and connections between risk factors in green building construction. The coupling phenomenon occurs, and the associated coupling risk arises, affecting the smooth implementation of green building projects.

In Figure 2, the four first-level risk factors of human, material equipment, environment, and management are recorded as A, B, C, and D, respectively. According to this classification, the types of green building construction risk couplings are divided into the following three categories:

(1) Homogeneous factor coupling refers to interactions between different secondary risk factors under the same first-level risk factor; it is the most common form of coupling, though its impact on the system is generally relatively slight, usually not breaking through the overall defenses of the system.

(2) Heterogeneous two-factor coupling refers to the mutual influences and interactions between secondary risk factors under different primary risk factors. Heterogeneous two-factor coupling includes human–material equipment coupling, human–environment coupling, human–management coupling, material equipment–environment coupling, and material equipment–management coupling.

(3) Heterogeneous multi-factor coupling refers to the interactions between secondary risk factors when three or more different primary risk factors are involved. Heterogeneous multi-factor coupling includes human–material equipment–environment coupling, human–material equipment–management coupling, human–environment–management coupling, material equipment–environment–management coupling, human–material equipment–environment–management coupling, and human–material equipment–environment–management coupling.

This paper analyzes the coupling mechanism of green building construction risk management from the perspective of “human–material equipment–environment–management” and is based on the complex system theory and the trigger principle, as shown in Figure 3.

Figure 3 shows that the risk system of green building construction is not only affected by the individual first-level human, material equipment, environment, and management risk factors, but also by coupling between first- and second-level risk factors. When a risk arises, the whole system has a certain self-healing ability, since when the risk threshold of a subsystem is not broken, measures can be taken to control and resolve the risk. However, when multiple risks occur, the internal risk factors of each subsystem interact with each other, causing the system risk level to rise and break the subsystem risk threshold. The coupled oscillator means that when some part of the system is affected by external or internal changes, this coupling can cause oscillations or fluctuations within the system, which in turn affects the stability of the overall system. After passing through the coupled oscillator of the system, under the disturbance of a sudden event, a risk factor may meet others, possibly producing heterogeneous two- or multi-factor coupling [43]. If two or more risk factors are coupled and the overall risk level is lower than the previous total risk, this is called the negative coupling phenomenon; if the total risk before and after coupling does not change, this is called the zero coupling phenomenon; and if coupling occurs between two or more risk factors and the overall risk level exceeds the previous total risk, this is called the positive coupling phenomenon. With positive coupling, the system risk will accumulate. If no protective measures are taken at this time, the coupling intensity will continue to increase, and the duration will be prolonged. The positive coupling effect on the system will also lead the risk accumulation to change from quantitative to qualitative, and the risk level of the system will continue to rise. Therefore, exploring the coupling mechanisms among risk factors, finding the key risk factors, and taking measures to reduce the positive coupling effect is key to reducing the total system risk.

3.2. Establish the Coupling Degree Model of Risk Factors

3.2.1. Entropy Weight–DEMATEL Method

The entropy weight method is based on a decision matrix for determining the weights of risk factors; it can overcome the shortcomings of fuzzy comprehensive evaluation, the analytic hierarchy process, and other subjective weighting methods, but this method cannot reveal the mutual influences between risk factors [44]. Green building is still under development in China, and it is not being sufficiently promoted, meaning little original data can be obtained. The DEMATEL method, also known as the decision experiment and evaluation laboratory method, is effective for assigning the causal relationships between various risk factors and the degrees of their influence by setting up an expert decision-making team [45]. Therefore, this paper combines the entropy weight method and DEMATEL method to calculate the combined weights, to not only quantify the relationships between risk factors but also make the established index system more scientific and reasonable.

The main calculation steps of the entropy weight method are as follows:

(1) Construct the original information matrix $X$. According to the results of the questionnaire, the original information matrix of $m$ experts and $n$ evaluation indexes is constructed as follows:

$$X={\left(x_{ij}\right)}_{m\times n}$$

(1)

(2) Construct a standardized matrix $X^{\prime }$. The original information matrix is processed using range normalization to eliminate the influence of dimensionality, and the normalized matrix $X^{\prime }={\left({x_{ij}}^{\prime }\right)}_{m\times n}$ can be obtained:

$${x_{ij}}^{\prime }=\frac{\max \left\{x_j\right\}-x_{ij}}{\max \left\{x_j\right\}-\min \left\{x_j\right\}}$$

(2)

(3) The calculated entropy $E_j$ is as follows:

$$E_j=-\left({\displaystyle \frac{1}{lnm}}\right){{\displaystyle \sum }}_{i=1}^m\left({\displaystyle \frac{{x_{ij}}^{\prime }}{{{\displaystyle \sum }}_{i=1}^m{x_{ij}}^{\prime }}}\right)\ln \left({\displaystyle \frac{{x_{ij}}^{\prime }}{{{\displaystyle \sum }}_{i=1}^m{x_{ij}}^{\prime }}}\right)$$

(3)

(4) Calculate the entropy weight:

$$W_j={\displaystyle \frac{1-E_j}{n-{{\displaystyle \sum }}_{j=1}^n{E}_j}}$$

(4)

The main calculation steps of the DEMATEL method are as follows:

(1) Assume that $m$ is the number of experts and construction site personnel. According to the $k\left(k=1,2,\cdots ,m\right)$ years of experience of the experts and construction site personnel, the average value is processed to eliminate subjective error. The calculation formula is as follows:

$$F={\displaystyle \frac{1}{m}}{{\displaystyle \sum }}_{m=1}^m{F}^k$$

(5)

Thus, the direct influence relation matrix $F$ is established:

$$F={\left(F_{ij}\right)}_{n\times n}$$

(6)

(2) Normalize the direct influence matrix $F$ to obtain the normalized direct influence matrix $X$:

$$X={\displaystyle \frac{F}{\underset{1\le i\le n}{\underbrace{max}}{{\displaystyle \sum }}_{j=1}^n{F}_{ij}}}\left(1\le i\le n,1\le j\le n\right)$$

(7)

(3) The comprehensive influence matrix $T$ is obtained:

$$T=X{\left(E-X\right)}^{-1}$$

(8)

(4) Calculate the influence degree, the affected degree, and the centrality. The degree of influence $D_i$ represents the degree of influence of factor $i$ on other influencing factors, the degree of influence $C_i$ represents the degree of influence of factor $i$ on other factors, and the centrality $M_i$ is obtained by combining the two.

$$D_i={{\displaystyle \sum }}_{j=1}^n{t}_{ij},\left(i=1,2,3,\cdots ,n\right)$$

(9)

$$C_i={{\displaystyle \sum }}_{j=1}^n{t}_{ji},\left(i=1,2,3,\cdots ,n\right)$$

(10)

$$M_i=D_i+C_i,\left(i=1,2,3,\cdots ,n\right)$$

(11)

(5) Calculate the combined weight. According to the entropy weight obtained using the entropy weight method and the centrality obtained using the DEMATEL method, the combined weight $W_z$ is obtained from multiplicative combination:

$$W_z={\displaystyle \frac{W_j{M}_i}{{{\displaystyle \sum }}_{i,j=1}^n{w}_j{M}_i}}\left(i,j=1,2,3,\cdots ,n\right)$$

(12)

When calculating the coupling degree between green building construction risk factors, it is first necessary for experts to objectively evaluate the risk value of each risk factor index, and then convert the risk value into a digital feature that can directly calculate the coupling degree. Since the expert evaluation results are often subjective, applying the reverse cloud model can remove a certain degree of subjectivity. It can also test the degree of the dispersion of expert score data. If this is large, data can be deleted. In this paper, the inverse cloud model based on the expert survey method transforms the evaluated value of each risk factor index to one ready for processing. The specific calculation steps are as follows:

(1) Obtain the risk values of coupling factors. The expert scoring method is used to evaluate the risk values of risk factors. Assuming that $n$ represents the number of experts and $m$ represents the number of risk factors, the evaluation matrix is obtained:

$$X={\left(x_{ij}\right)}_{n\times m}$$

(13)

(2) Numerical characteristics are calculated from secondary risk factor indicators:

$$\overline{X}={\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n{x}_i}{n}}$$

(14)

$$S^2={\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n{\left(x_i-\overline{X}\right)}^2}{n-1}}$$

(15)

$$E_x=\overline{X}={\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n{x}_i}{n}}$$

(16)

$$E_n=\sqrt{{\displaystyle \frac{\pi }{2}}}{\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n\left|x_i-E_x\right|}{n}}$$

(17)

$$H_e=\sqrt{S^2-{E_n}^2}$$

(18)

In the above formula, $\overline{X}, S^2, E_x,E_n$, and $H_e$ represent the mean, variance, expected value, entropy value, and super entropy value of the secondary risk factor indicators.

(3) Calculate the digital characteristics of the first-level risk factor indicators:

$${E_x}^{\prime }={\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^m{E}_{x_i}{u}_i}{{{\displaystyle \sum }}_{i=1}^m{u}_i}}$$

(19)

$${E_n}^{\prime }={\displaystyle \frac{{u_1}^2{E}_{n1}+{u_2}^2{E}_{n2}+\cdots +{u_m}^2{E}_{nm}}{{{\displaystyle \sum }}_{i=1}^m{u_i}^2}}$$

(20)

$${H_e}^{\prime }={\displaystyle \frac{{u_1}^2{H}_{e1}+{u_2}^2{H}_{e2}+\cdots +{u_m}^2{H}_{em}}{{{\displaystyle \sum }}_{i=1}^m{u_i}^2}}$$

(21)

In the above formulas, ${E_x}^{\prime },{E_n}^{\prime },\mathrm{and}{H_e}^{\prime }$ represent the expected value, entropy value, and super entropy value of the first-level risk factor index.

3.2.2. Coupling Degree Model

$E\left(i=1,2,3,\cdots ,4;j=1,2,3,\cdots ,m\right)$ represents the expected values of the $j$-th and $i$-th risk factors in the green building construction risk system in the reverse cloud model. $A_{ij}$ and $B_{ij}$ represent the upper and lower limits of the risk factor index when the system reaches a stable state. Then, the efficacy coefficient of the secondary risk factor index to the entire green building construction risk system is expressed as follows:

$$U_{ij}={\displaystyle \frac{E_{x_{ij}}-B_{ij}}{A_{ij}-B_{ij}}}$$

(22)

Using the linear weighting method, the contribution of the first-level risk factor index to the entire green building construction risk system can be obtained, that is, the order parameter. The calculation expression is as follows:

$$U_i={{\displaystyle \sum }}_{j=1}^n{w}_{ij}{U}_{ij},{{\displaystyle \sum }}_{j=1}^n{w}_{ij}=1$$

(23)

Based on the above analysis, referring to the research of Yuan [33] on the coupling degree model, based on the coupling degree function measurement model and the concept of physics, assuming that there are $v$ subsystems in the green building construction risk system, the coupling degree calculation model is expressed as follows:

$$T_v={\left[{\displaystyle \frac{U_1{U}_2\cdots U_v}{\left(\frac{U_1+U_2+\cdots +U_v}{v}\right)}}\right]}^{{\displaystyle \frac{1}{v}}}$$

(24)

If the value of the efficiency coefficient is low or close, when calculating the coupling degree with this value, the result will be close to 1, which is inconsistent with the actual degree. After referring to the existing literature, we introduced the comprehensive coordination index $P$, which can correct the original coupling degree model. The calculation expression is as follows:

$$P=k_1{U}_1+k_2{U}_2+k_3{U}_3+\cdots +k_m{U}_m$$

(25)

$$T={\left(T_v\times P\right)}^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}$$

(26)

In the formula, $T$ is the modified coupling degree; $P$ is the comprehensive coordination index; and $k_1,k_2,k_3\cdots k_m$ are the undetermined coefficients. The combined weight of all risk factors must satisfy $k_1+k_2+k_3+\cdots +k_m=1$.

According to the classification of coupling states in physics, the value range of $T$ is $\left[0,1\right]$, and the coupling degree and its corresponding risk coupling state are shown in

Table 2. Coupling value interval and its corresponding coupling state.

$\mathit{T}$ Ratio	Coupling State
[0, 0.3]	Low coupling, less likely to induce risk
(0.3, 0.7]	Medium coupling, the possibility of inducing risk is greater
(0.7, 1]	Strong coupling, it is easy to induce risk and the greatest harm

.

3.2.3. System Dynamics Model

Since the coupling degree model cannot explore the causal relationship between secondary risk factors in subsystems, nor can it explore the impact of coupling changes in key factors on the system, a causal relationship diagram of secondary risk factors was constructed by referring to the research of Mo et al. [46] and using theoretical knowledge of system dynamics to reflect the interactions between various risk factors. In the following, the coupling of risk factors in the human and environment system (herein referred to as “human–environment”) is taken as an example. The coupling relationships between homogenous and heterogeneous factors in the two subsystems are shown in Figure 4.

In the human–environment coupling relationship, human factors can impact environmental factors through the transmission of risk factors, and in turn, the environment can also impact people. As such, the impact is mutual. The coupling between human and environmental factors is centered on the low proficiency of green building construction personnel in the application of new technology, which will lead to a waste of raw materials, resulting in high construction waste in green building construction, thus causing certain environmental risks. Moreover, unpredictable bad weather and other uncertain environmental factors, if coupled with green building construction personnel not performing the corresponding safety protection measures, will produce human risks. In addition, the improper selection of transportation modes for on-site materials and components in the environment, if coupled with the operators lacking green building construction experience, may lead to the unreasonable transportation of some untested innovative materials, which may cause harm to the operators, that is, a human risk.

4. Case Study

4.1. Project Overview

We take the Guiyu Tinglan project in Shijiazhuang City as an example, where the total construction area is 253,557.33 m² and the total investment is RMB 33,054,900, including residential buildings, kindergartens, basements, supporting facilities, and other forms of construction. In the green design, according to the ‘residential building energy-saving design standards’, ‘building energy-saving doors and windows engineering technical specifications’, ‘sound environment quality standards’, and other specifications, in the selection of materials, (ready-mixed) cast-in-place reinforced concrete, (ready-mixed) building mortar, and high-strength structural building materials are reasonably used. The design of foundational and structural components is optimized to achieve material savings. Furthermore, a water-saving system uses water metering devices set up according to use, and it is equipped to implement effective measures to avoid pipe network leakage. The project is in line with energy savings, water savings, land savings, material savings, and environmental protection. Therefore, it offers a representative example.

4.2. Application of Coupling Degree Model

4.2.1. Calculation of Portfolio Weights of Risk Factors and Indicators

Questionnaires were designed based on the risk list of green building construction projects, and a five-point Likert scale was used to investigate the impact degree of each risk factor through a combination of on-site inquiry and network questionnaires. Two types of questionnaires were distributed to experts and construction units, respectively. The first type of questionnaire datum was the entropy weight, obtained through sorting and calculation using the entropy weight method. The second type of questionnaire datum was the use of the DEMATEL method to determine the centrality, according to the results of questionnaire inspection and calculation.

Table 3. Expert background information.

Expert	Experience (Years)	Academic Degree	Position	Research Field
1	5–10 years	bachelor’s	field director	civil engineering
2	10–15 years	master’s	architect	building
3	5–10 years	bachelor’s	field director	construction management
4	15–20 years	master’s	project manager	construction management
5	5–10 years	bachelor’s	general contractor	civil engineering
6	5–10 years	bachelor’s	general contractor	civil engineering
7	15–20 years	master’s	field director	building
8	15–20 years	master’s	project manager	civil engineering
9	10–15 years	bachelor’s	field director	civil engineering
10	10–15 years	bachelor’s	field director	construction management
11	10–15 years	bachelor’s	field director	civil engineering

lists the background information of the experts involved in this study.

A total of 11 questionnaires were collected. However, two invalid questionnaires were removed as the respondents answered too quickly or continuously selected the same option. The data from the remaining questionnaires were analyzed, and they were standardized using the entropy weight method. According to Formulas (1)–(4), the entropy weight of each risk factor was obtained. Meanwhile, using the principles and methods of DEMATEL, six professionals from different work units (three supervision units, two construction units, and one scientific research institution) were invited to form an expert decision-making group. The “0–4 scale” was used to rate the degrees of influence of 26 factors through a questionnaire, and power processing was applied to the responses based on the working years of the experts. The arithmetic average value was taken as the direct influence matrix, and the comprehensive influence matrix was ascertained via Equation (8). The center degree was obtained with Equations (9)–(11), and according to Equation (12), the final combined weights were derived, as shown in

Table 4. Combined weights of risk factor indicators.

Primary Index	Secondary Index	Entropy Weight	Centrality	Combined Weight
Human A	Construction personnel’s physical fitness status A1	0.0212	2.6752	0.0166
	Safety protection measures for construction personnel A2	0.0640	3.3006	0.0618
	Physiological condition of construction personnel A3	0.0328	3.1417	0.0301
	Green building construction experience A4	0.0357	3.4987	0.0365
	Green building construction consciousness level A5	0.0691	3.1983	0.0647
	New technology application proficiency A6	0.0328	3.7064	0.0356
Material equipment B	Completeness of construction site safety protection facilities B1	0.0493	4.8184	0.0695
	Operation capacity of green building construction equipment B2	0.0521	3.9494	0.0602
	Applicability of green building construction equipment B3	0.0202	4.5208	0.0268
	Equipment repair and maintenance measures B4	0.0493	3.8664	0.0558
	Aging and wear of green building construction equipment B5	0.0469	3.4837	0.0478
	Green building construction materials use technology B6	0.0328	3.9861	0.0383
	Green building construction material quality B7	0.0202	3.7539	0.0222
Environment C	Transportation of materials and components onsite C1	0.0307	3.2043	0.0288
	Site material and component storage yard layout C2	0.0285	2.9086	0.0243
	Construction noise and vibration pollution C3	0.0681	3.6056	0.0718
	Construction dust pollution C4	0.0493	2.7183	0.0392
	Construction waste C5	0.0493	3.4178	0.0493
	Unexpected bad weather C6	0.0163	3.7871	0.0180
	Unexpected site conditions C7	0.0187	2.6363	0.0144
Manage D	Personnel safety and health management training D1	0.0168	2.0314	0.0100
	Protection of the site construction environment D2	0.0303	4.0314	0.0357
	Green building construction safety management mechanism perfection D3	0.0469	2.6065	0.0358
	Rationality of the green building construction scheme D4	0.0345	2.8227	0.0285
	Emergency warning and emergency handling measures D5	0.0328	3.7582	0.0361
	Construction waste management D6	0.0512	2.8239	0.0423

.

4.2.2. Calculation of Numerical Characteristics of Risk Factor Indicators

In this paper, according to the influence degree of risk factors, scores are normalized with respect to a [0–10] scale. The higher the value, the more important the influence of the risk factor. Six experts were invited to score all secondary risk factor indicators according to the actual project conditions of their green building construction. The scoring results are shown in

Table 5. Expert scoring results of secondary risk factor indicators and calculation results of the reverse cloud model.

	Expert 1	Expert 2	Expert 3	Expert 4	Expert 5	Expert 6	(Ex, En, He)
A1	3	6	5	6	7	8	(5.83, 1.532, 0.788)
A2	7	4	3	2	6	3	(4.17, 1.950, 0.185)
A3	6	4	5	6	4	8	(5.50, 1.462, 0.402)
A4	2	4	3	2	6	5	(3.67, 1.671, 0.355)
A5	3	5	3	1	3	2	(2.83, 1.114, 0.725)
A6	5	3	4	3	6	5	(4.33, 1.253, 0.323
B1	2	3	5	2	3	6	(3.50, 1.671, 0.304)
B2	3	1	3	4	4	2	(2.83, 1.114, 0.354)
B3	6	7	5	3	5	3	(4.83, 1.532, 0.469)
B4	2	3	3	4	6	5	(3.83, 1.462, 0.169)
B5	4	5	3	2	4	6	(4.00, 1.253, 0.655)
B6	6	4	5	5	4	3	(4.50, 1.044, 0.096)
B7	3	5	4	7	4	5	(4.67, 1.253, 0.544)
C1	4	3	5	6	5	7	(5.00, 1.253, 0.655)
C2	6	8	5	6	9	7	(6.83, 1.462, 0.169)
C3	3	4	3	4	5	3	(3.67, 0.836, 0.177)
C4	5	3	5	4	4	2	(3.83, 1.114, 0.354)
C5	2	3	6	5	4	4	(4.00, 1.253, 0.655)
C6	8	7	6	9	8	6	(7.33, 1.253, 0.323)
C7	6	5	4	6	7	3	(5.17, 1.462, 0.169)
D1	5	3	7	5	4	8	(5.33, 1.810, 0.435)
D2	4	7	4	6	3	5	(4.83, 1.462, 0.169)
D3	5	3	4	3	6	1	(3.67, 1.671, 0.524)
D4	3	5	6	4	5	2	(4.17, 1.462, 0.169)
D5	6	4	3	5	4	4	(4.33, 0.975, 0.341)
D6	4	2	4	5	3	2	(3.33, 1.253, 0.323)

.

According to Table 5 and Equations (19)–(21), the expected value, entropy, and super entropy of first-level risk factor indicators were obtained, as shown in

Table 6. The calculation results of the first-level risk factor index reverse cloud model.

First-Order Coupling Index	(Ex, En, He)
A	(4.0414, 1.499292189, 0.4416)
B	(3.8189, 1.253813694, 0.4181)
C	(4.5856, 1.28074314, 0.4350)
D	(4.1050, 0.456764997, 0.1320)

.

4.2.3. Calculation of the Efficacy Coefficient

According to the calculation results in Table 5 and Table 6, combined with Formula (22), the efficacy coefficients of primary and secondary risk factors were obtained, as shown in

Table 7. Calculation results of efficacy coefficients of risk factor indicators.

First-Grade Indexes	Efficiency Coefficient	Second Index	Efficiency Coefficient
Human A	0.40414	Construction personnel’s physical fitness status A1	0.58
		Safety protection measures for construction personnel A2	0.42
		Physiological condition of construction personnel A3	0.55
		Green building construction experience A4	0.37
		Green building construction consciousness level A5	0.28
		New technology application proficiency A6	0.43
Material equipment B	0.38189	Completeness of construction site safety protection facilities B1	0.35
		Operation capacity of green building construction equipment B2	0.28
		Applicability of green building construction equipment B3	0.48
		Equipment repair and maintenance measures B4	0.38
		Aging and wear of green building construction equipment B5	0.40
		Green building construction materials use technology B6	0.45
		Green building construction material quality B7	0.47
Environment C	0.45856	Transportation of materials and components onsite C1	0.50
		Site material and component storage yard layout C2	0.68
		Construction noise and vibration pollution C3	0.37
		Construction dust pollution C4	0.38
		Construction waste C5	0.40
		Unexpected bad weather C6	0.73
		Unexpected site conditions C7	0.52
Manage D	0.41050	Personnel safety and health management training D1	0.53
		Protection of the site construction environment D2	0.48
		Green building construction safety management mechanism perfection D3	0.37
		Rationality of the green building construction scheme D4	0.42
		Emergency warning and emergency handling measures D5	0.43
		Construction waste management D6	0.33

.

4.2.4. Calculation of Coupling Degree

According to the efficiency coefficients in Table 7, the coupling degrees of the four first-level indicators under different coupling combinations were calculated by applying Equations (23)–(26). For example, T (A, B) represents the coupling degree between A (human factor) and B (material and equipment factor). The specific calculation results are shown in

Table 8. Numerical results of coupling degree of first-order risk factors.

Heterogeneous Two-Factor Coupling		Heterogeneous Multi-Factor Coupling
Coupled Mode	Coupling Value	Coupled Mode	Coupling Value
T21(A, B)	0.6240	T31(A, B, C)	0.6405
T22(A, C)	0.6561	T32(A, B, D)	0.6019
T23(A, D)	0.6356	T33(A, C, D)	0.6495
T24(B, C)	0.5365	T34(B, C, D)	0.6380
T25(B, D)	0.6186	T4(A, B, C, D)	0.6374
T26(C, D)	0.6560	——	——

.

4.3. Human–Environment Subsystem Risk Factor Coupling Simulation

During the statistical analysis, it was found that human–environment system coupling was the largest, with the highest coupling degree, which will increase the system risk and even cause new risks. Therefore, we selected this coupling as the object of simulation using Vensim PLE x32 software version 7.3.5. The simulation included the complete subsystems of human and environmental factors in the green building construction risk system over 24 months, with a running step of 1 month. The system flow diagram is shown in Figure 5, where $L$ represents the total risk level of human–environment system factor coupling; $L\left(A\right)$ represents the risk level of human factors; $L\left(C\right)$ represents the level of environmental risk; $L\left(ij\right)$ represents the risk level of the $j$-th index of the $i$-type risk factor of the human–environment risk factor subsystem; $R\left(A\right)$ represents the change in the risk level of human factors; $R\left(C\right)$ represents the change in the environmental risk level; $R\left(ij\right)$ represents the change in the risk level of the $j$-th index of the $i$-th risk factor; and $T\left(ij,i^{\prime }{j}^{\prime }\right)$ represents the risk coupling value between the $j$-th index of the $i$-th risk factor and the $j^{\prime }$-th index of the $i^{\prime }$-th risk factor, where $i=A,C;j=1,2,\cdots ,n$.

Before the simulation, it was necessary to determine the weights of each secondary risk factor index, as described in detail in the previous section. Then, the coupling values for combinations of human–environment system factors were calculated successively according to Equations (22)–(26). The initial values of the state variables here were provided by five experts and then the arithmetic average was taken. The data results are shown in

Table 9. Initial values of risk factors.

Variable Code	Variable Name	Variable Value
L(A)	Human factor level	36.21
L(C)	Environmental risk level	35.99
L(A1)	Construction personnel’s physical fitness status	29.83
L(A2)	Safety protection measures for construction personnel	36.32
L(A3)	Physiological condition of construction personnel	31.92
L(A4)	Green building construction experience	41.62
L(A5)	Green building construction consciousness level	47.22
L(A6)	New technology application proficiency	33.19
L(C1)	Transportation of materials and components onsite	38.3
L(C2)	Site material and component storage yard layout	36.77
L(C3)	Construction noise and vibration pollution	29.27
L(C4)	Construction dust pollution	31.86
L(C5)	Construction waste	45.66
L(C6)	Unforeseen bad weather	32.46
L(C7)	Unforeseen site conditions	35.28

below. We inputted the equation required by the model according to the stock flow diagram in Figure 5, as shown in

Table 10. Modeling equations.

Variable	Functional Equation
State variable	L = INTEG(“L(A)” × 0.245 + “L(C)” × 0.246)
	L(A) = INTEG(“R(A)”, 36.21)
	L(C) = INTEG(“R(C)”, 35.99)
Rate variable	R(A) = INTEG(“L(A1)” × 0.0166 + “L(A2)” × 0.0618 + “L(A3)” × 0.0301 + “L(A4)” × 0.0365 + “L(A5)” × 0.0647 + “L(A6)” × 0.0356)
	R(C) = INTEG(“L(C1)” × 0.0288 + “L(C2)” × 0.0243 + “L(C3)” × 0.0718 + “L(C4)” × 0.0392 + “L(C5)” × 0.0493 + “L(C6)” × 0.018 + “L(C7)” × 0.0144)
	R(A1) = INTEG(“L(A3)” × “10T (A1, A3)” + “1T (A1, A2)” × “L(A2)”)
…	…

.

The data in Table 9 and the equations in Table 10 were substituted into the system dynamics model for simulation, and the change trend of the total value L of the human–environment system factor coupling risk level within 24 months was obtained, as shown in Figure 6. In this figure, under the coupling of risk factors, the growth rate and level of risk increase with time, and there is no big fluctuation in the first 12 months. Then, up to the 18th month, it starts to rise slowly, and from the 18th month, it starts to rise rapidly, that is, the risk level increases rapidly. In the 24th month, the total risk level of the system reaches 245.395.

5. Results and Analysis

This study discusses the impact of multi-risk-factor coupling on the green building construction risk management system and identifies key coupling risk factors that have the greatest impacts on its total risk level. Although previous scholars have studied the coupling effects of multiple risk factors on public tunnels, integrated pipeline corridors, water transportation, and other fields [27,28,47], few studies have analyzed risk coupling in the field of green building construction; so, this paper constitutes original work.

The coupling coefficients of homogeneous and heterogeneous factors were adjusted. There were eight types of homogeneous factors—namely, T (A1, A2), T (A4, A5), T (A5, A6), T (A1, A3), T (C6, C7), T (C4, C6), T (C3, C6), and T (C1, C7)—and three types of heterogeneous factors—T (A4, C1), T (A6, C5), and T (A2, C6). To clarify the impact of a coupling value change on the overall risk level, the coupling values of 11 combinations were reduced by 50%, respectively, and a total risk numerical simulation was performed over 24 months to observe the change in the total risk. The greater the value, the greater the impact of the coupling effect between the corresponding factors on the overall risk.

(1) Adjust the coupling coefficients of homogeneous factors

The coupling coefficient values of eight combinations were reduced by 50% each to observe the horizontal change in the human–environment coupling risk, as shown in

Table 11. Adjusted results for the coupling coefficient of internal factors in the subsystem.

Scheme Code	Coupling Coefficient	Initial Value	Adjustment Proportion	Value after Adjustment
current 1	T (A1, A2)	0.6348	−50%	0.3174
current 2	T (A4, A5)	0.5564	−50%	0.2782
current 3	T (A5, A6)	0.5790	−50%	0.2895
current 4	T (A1, A3)	0.7364	−50%	0.3682
current 5	T (C6, C7)	0.7821	−50%	0.3910
current 6	T (C4, C6)	0.7018	−50%	0.3509
current 7	T (C3, C6)	0.6445	−50%	0.3222
current 8	T (C1, C7)	0.6922	−50%	0.3461

.

As shown in Figure 7 and Figure 8, the impact trend of a 50% reduction in the coupling coefficient of the internal homogeneous factors involved a decline in the total risk level of the “human–environment” system. In particular, the coupling combinations of A1 and A3, A1 and A2 in the human factor subsystem greatly impacted the system risk level, along with the coupling combination of C1 and C7, C6 and C7 in the environmental factor subsystem.

(2) Adjusting the heterogeneous coupling coefficients

The coupling coefficient values of three heterogeneous combinations were reduced by 50% each to observe the horizontal change in the human–environment coupling risk, as shown in

Table 12. Adjustment results of the coupling coefficient between subsystems.

Scheme Code	Coupling Coefficient	Initial Value	Adjustment Proportion	Value after Adjustment
current 1	T (A4, C1)	0.6520	−50%	0.326
current 2	T (A6, C5)	0.6410	−50%	0.3205
current 3	T (A2, C6)	0.6800	−50%	0.34

.

Figure 9 shows that when the coupling coefficient of heterogeneous factors between subsystems was reduced by 50%, this produced a downward trend in the overall risk level of the “human–environment” system and reduced the growth rate of the coupling risk level. Among the heterogeneous factors, the coupling combination of A2 and C6 had the greatest impact on the change in the overall trend of the system, followed by the coupling combination of A4 and C1.

6. Conclusions

This paper proposes a risk coupling analysis method based on an improved coupling degree model and system dynamics, which can reveal the coupling relationships between various factors and accurately quantify the degree of risk of a coupling effect between risk factors. The method may be applied to practical cases of green building construction projects. Our specific conclusions are as follows:

(1) We used the improved coupling degree model to calculate the degrees of coupling among the first-level risk factor indicators, and we found that for heterogeneous two-, three-, and four-factor combinations, these were all between (0.3, 0.7], constituting medium coupling that may induce great risks. The coupling degree of the four factors was 0.6374, which indicated that although the probability of the four risk factors influencing green building construction safety simultaneously is small, their combined influence may be large. In addition, the three-factor risk coupling value was generally greater than that for two factors, though there were special cases in which the human–environment coupling value was larger. The coupling degree was 0.6561, which indicated that the coupling of factors within the green building construction risk system had a great influence and that the coupling relationships in the system were complex.

(2) The SD model was used to simulate the coupling risk of human–environment system factors, and the results showed that the degree of coupling trended upward with time, indicating a dynamic change in green building construction risk. When we simulated adjusting the coupling coefficient between risk factors, it was found that the level of risk could be effectively reduced.

In summary, according to our simulation of the risk factors in green building construction, the coupling combinations of homogeneous factors with the greatest impacts on the risk are the physical fitness and physiological condition, and the physical fitness and safety protection measures, while those for heterogeneous factors are safety protection measures and unexpected bad weather, and new technology application proficiency and construction waste. Therefore, we advise always considering the forecasted changes in the weather in the next few weeks, to minimize the adverse consequences of bad weather on the construction process, while construction personnel should also consider their health and daily safety protection. Moreover, in the construction process, considerately using innovative construction technology may reduce the adverse consequences on the environment. The findings of this study provide decision-making support for green building risk management, so that decision-makers can prevent risks from coupling and plan appropriate risk response measures.

Although, in this study, we systematically analyzed the risk coupling relationships in green building construction, there were still some limitations to our work. First, we mainly studied the coupling factors of personnel, materials and equipment, environment, and management within the risk system, while we did not fully consider the coupling factors outside the system. To remedy this, future research should perform a combined coupling effect analysis of internal and external system factors. Second, we mainly studied risk factors in the construction stage, without considering uncertainties in other stages. The third limitation of this paper is that the risk factors of uncertainty in the environmental data of the construction material should be considered when studying the risk factors. The environmental data of materials used in the construction process are usually affected by various factors, which may lead to the inaccuracy of the data. Therefore, more uncertainty factors should be considered in the future, and appropriate models should be selected to further analyze this problem. Finally, energy consumption and efficiency are important aspects of green buildings, and the relationship between energy consumption and efficiency and green buildings can be deeply explored in future studies. In the future, risk coupling may be studied with a view to the whole life cycle, for a more comprehensive analysis of the risks in green building work.