Identification of Key Risk Factors in Mechanical Cross Passage Construction Based on the Decision-Making Test and Evaluation Test Method and the Interpretation Structure Model

Abstract

In order to solve the problem that the key risk factors in the construction of mechanical cross passages are relatively vague, the decision-making test and evaluation test method (DEMATEL) and the interpretation structure model (ISM) are combined to analyze safety risks in the construction of mechanical cross passages scientifically and reasonably. Based on the ‘personnel’, ‘material’, ‘machine’, ‘method’, and ‘environment’ of 4M1E comprehensive management, the construction safety risk index system of mechanical method cross passage, including 18 risk factors, is extracted. DEMATEL was used to divide the risk factors into four factor sets: strong cause, weak cause, strong result, and weak result. Furthermore, ISM is used to construct a hierarchical structure diagram of risk factors, and deep risk factors and risk factors with a high node degree are divided. The results show that the safety awareness of construction personnel, the safety technology level of construction personnel, the establishment and implementation of the safety construction system, the level of construction management, and the degree of geological complexity are strong cause-type risk factors with both high centrality and high causality. Additionally, the effect of reinforcement of the internal support system, the setting of the digging parameter, the effect of grouting, and the effect of assembling the pipe sheet comprise the strong result-type risk factors with high centrality. The risk factors with higher node degree in ISM are consistent with the risk factors with higher centrality in the DEMATEL model, which are key factors that play an important role in risk control.

1. Introduction

Taking the degree of mechanization of construction excavation as a criterion, the construction method of the cross passage can be divided into two categories: mechanical methods and non-mechanical methods. Non-mechanical construction technology is mainly used for the mine method of excavation after the ground is reinforced via the freezing method. However, this method has a long construction period, a long freeze–thaw settlement cycle, and quality hazards. Mechanical methods are a new technology in line with the development trend of engineering, and the use of mechanical methods to construct the cross passage can reduce the impact of soil freezing and thawing settlement on the surrounding buildings and greatly improve the construction efficiency [1,2,3]. Therefore, it is of great practical significance to study the safety risk management of mechanical cross passage construction for the safe and efficient completion of the construction of the cross passages.

In recent years, many scholars have carried out research on the construction technology of mechanical cross passages. Excavation of the cross passages will cause secondary disturbance to the stratum, which will affect the surface settlement and the safety of the main tunnel structure. Most of the existing research on cross passages focuses on the influence of construction on the main tunnel structure or the mechanical response of the structure under dynamic load [4,5,6,7,8]. However, there are few studies on the risk management of mechanical cross passages.

The basis of risk analysis is the identification and extraction of safety risk factors based on construction methods [9,10]. Traditional risk assessment methods, such as the fuzzy analytic hierarchy process, entropy weight method, fuzzy comprehensive evaluation method, cloud model, and Bayesian network have been widely used and verified in engineering practices [11,12,13,14,15]. However, the construction face of cross passages is narrow, and its safety risk is affected by many factors such as geological conditions, construction equipment, and construction methods. In the actual construction process, these influencing factors do not exist in isolation, and they affect each other. Because the risk is uncertain, it is necessary to analyze the risk of the project before construction [16]. The DEMATEL model can use matrix operations to calculate the influence degree and the affected degree of each risk factor on other risk factors in the system [17,18,19,20,21]. Interpretative structural modeling (ISM) is a method used to analyze the structural relationship of complex systems, which is often used to analyze the interactions between factors [22,23,24,25].

In the field of underground space tunnel engineering, most research is aimed at the risk management of shield main tunnel construction, and research on the risk management of mechanical cross passage construction is relatively sparse. In view of this, this paper constructs a construction risk analysis model of a mechanical cross passage based on DEMATEL-ISM. By identifying the safety risk factors of the construction of the mechanical cross passage, the DEMATEL analysis method is used to calculate and determine the influence degree of each risk factor, and the ISM model is used to divide the hierarchical structure between all risk factors. The combination of the two can accurately identify the key risk factors and formulate corresponding risk control measures. These research results can provide a reference for risk assessment of similar projects.

2. DEMATEL-ISM Analysis Method

The DEMATEL model can use matrix operations to calculate the influence degree and the affected degree of each risk factor on other risk factors in the system, then calculate the cause degree and center degree of each risk factor to determine the causal relationships between risk factors. However, this method cannot effectively identify the hierarchical structure of risk factors in the system. The ISM method gives the simplest hierarchical topology without losing system function through Boolean logic operation. It can clearly understand the ladder structure of each risk factor in the system. The combination of the two can not only identify the status of each risk factor in the system but also construct the hierarchical structure of the system risk factors. The specific calculation steps are as follows [26,27,28,29]:

(1) The risk factors that constitute the system are identified and marked as $x_1,x_2,\cdots ,x_3$.

(2) The direct influence matrix is determined, which is A, where, $x_{ij}(i,j=1,2,\cdots ,n;i\ne j)$ represents the direct impact of risk $x_i$ on risk $x_j$. Because the risk factor has no effect on itself, when $i=j$, $x_{ij}=0$.

$$A=\left[\begin{array}{cccc}0& x_{12}& \cdots & x_{1n}\\ x_{21}& 0& \cdots & x_{2n}\\ ⋮& ⋮& 0& ⋮\\ x_{n1}& x_{n2}& \cdots & 0\end{array}\right]$$

(1)

(3) The direct influence matrix A is normalized using the row sum maximum method to obtain the normalized direct influence matrix B. Then, the comprehensive influence matrix T is obtained, which represents the comprehensive effects of direct and indirect influence between risk factors. In Formula (3), I represents the unit matrix.

$$b_{ij}=a_{ij}/\mathrm{m}\mathrm{a}\mathrm{x}({\sum }_{j=1}^n{a}_{ij})$$

(2)

$$T={\sum }_{K=1}^{\infty }{B}^k={B(I-B)}^{-1}$$

(3)

(4) The influence degree, centrality, and cause degree of each risk factor were calculated.

The influence degree $D_i$ refers to the sum of the values of each row in T, which indicates the comprehensive influence value of the risk factors corresponding to each row on all other risk factors.

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

(4)

The affected degree $C_i$ refers to the sum of the values of each column in T, which indicates the comprehensive influence value of the risk factors corresponding to each column on all other risk factors.

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

(5)

The centrality $M_i$ is the sum of the influence degree and the affected degree of the risk factor, which indicates the role of the risk factor in the system.

$$M_i=D_i+C_i$$

(6)

The cause degree $R_i$ is the influence degree of the risk factor minus the affected degree. If the cause degree is greater than 0, it means that the risk factor has a great influence on other risks, which is called the cause factor; otherwise, it is called the result factor.

$$R_i=D_i-C_i$$

(7)

(5) The overall influence matrix E is determined, which is the sum of the comprehensive influence matrix T and the unit matrix I.

$$E=T+I$$

(8)

(6) By introducing the threshold λ to eliminate the factors with less influence in matrix E, the reachable matrix F can be obtained. The value of λ is usually determined by experts or decision makers according to the actual situation.

$$f_{ij}=\left\{\begin{array}{c}1,e_{ij}\ge \lambda (i,j=1,2,\cdots ,n)\\ 0,e_{ij}<\lambda (i,j=1,2,\cdots ,n)\end{array}\right.$$

(9)

(7) The hierarchy is divided. The set of the risk factors corresponding to the columns with a value of 1 on a row of the reachable matrix is the reachable set; that is, $R\left(x_i\right)=\left\{x_i|f_{ij}=1\right\}$. The set of risk factors corresponding to the rows with a value of 1 on a column is the leading set; that is, $S\left(x_i\right)=\left\{x_i|f_{ij}=1\right\}$. When $R\left(x_i\right)\cap S\left(x_i\right)=R\left(x_i\right)$ is satisfied, the risk factor in $R\left(x_i\right)$ can find the antecedent in $S\left(x_i\right)$; that is, factor “i” is a high-level factor. Rows and columns corresponding to high-level risk factors are subtracted from F, and the calculation process is repeated until all rows and columns are eliminated. The directed hierarchical topology diagram between risk factors is drawn according to the order of elimination.

3. Construction Risk Factor Identification of Mechanical Cross Passage

3.1. Mechanical Cross Passage

Mechanical cross passage technology refers to the mechanical construction method of using a shield machine or pipe jacking machine to replace manual excavation. The principle is based on the theory of soil pressure balance. The full-section cutting cutterhead is used to cut the front soil into the storage sealed silo behind the cutterhead, and the appropriate pressure in the silo is balanced with the water and soil pressure on the excavation surface. The construction technology of the mechanical cross passage is shown in Figure 1, and the construction process is shown in Figure 2.

The construction of the cross passage needs to excavate the special segments reserved on the main tunnel, which greatly reduces the stability of the main tunnel and is prone to quicksand and water seepage in the tunnel. Although the amount of earthwork required for the construction of a single cross passage is small, it is based on two existing parallel tunnels. The construction conditions are more complex, and the long-term earth pressure and formation resistance will show different conditions from the construction of separate tunnels. Because the soil has been disturbed by the construction of two existing parallel main tunnels, the construction of the cross passage will cause the deformation of the soil again, aggravate the degree of disturbance of the soil, and make the construction more difficult. The construction of the cross passage will inevitably affect the two existing tunnels, causing additional deformation and stress.

3.2. Risk Factor Identification

First, we collected the design documents, construction schemes, and other information related to the mechanical cross passage to understand the basic situation of the project. Secondly, we determined the national and local norms and standards related to the construction of mechanical cross passages, such as ‘Code for Construction and Acceptance of Underground Railway Engineering’, ‘Code for Construction and Acceptance of Shield Tunnel’, ‘Code for Construction and Acceptance of Shield Mechanical Cross passage’, and so on. We carefully studied the technical requirements, quality control indicators, and safety regulations in the standard, and we sorted the terms and contents that may cause risks. Then, we consulted the relevant literature on the construction of the mechanical cross passages. Field investigations were conducted on multiple construction sites, and risk factors were initially identified in combination with construction characteristics. Then, we introduced the identified potential risk factors to senior experts in the industry and listened to the opinions and suggestions of the experts. Finally, the identified risk factors were merged in the same class to ensure that the correlation was not repeated. According to the five aspects of ‘people’, ‘materials’, ‘machinery’, ‘methods’, and ‘environment’ comprehensively managed by 4M1E, the construction risk factors of the mechanical cross passage are determined as shown in

Table 1. Construction risk factor system of mechanical cross passage.

Factor Layer	Indicator Layer
Personnel and management factors	The safety awareness of construction personnel is insufficient $X_1$ The safety technology level of construction personnel is low $X_2$ The safety construction system is not perfect, and it is not fully constructed according to the established system $X_3$ Insufficient construction supervision $X_4$
Construction material factors	Poor quality of grouting material $X_5$ The segment quality is poor or the segment is damaged during transportation $X_6$
Construction machinery factors	Unreasonable selection of cutterhead system for excavation machine $X_7$ Unreasonable selection of supporting equipment for excavation machine $X_8$
Construction technical factors	The reinforcement effect of the starting and receiving ends does not meet expectations $X_9$ Special segment did not achieve the desired reinforcement effect $X_{10}$ The unreasonable loading of the internal support system leads to the deformation of the main tunnel segment $X_{11}$ Unreasonable setting of tunneling parameters $X_{12}$ Poor grouting effect $X_{13}$ The segments are not assembled in time, the segments are staggered, and the bolts are not tightened $X_{14}$ Low level of construction monitoring $X_{15}$
Construction environmental factors	Length of excavation $X_{16}$ The complexity of geology $X_{17}$ Surrounding pipelines and buildings $X_{18}$

.

4. Model Construction

4.1. DEMATEL Model

Since the calculation of DEMATEL requires non-technical data, it is the most reliable and convenient method to make full use of the knowledge and experience of experts. Therefore, through the investigation of practical engineering, 10 experts in the field of mechanical cross passages were invited to assign the degree of action between risk factors. These experts included university professors, research institute personnel, project managers, and construction personnel. They had rich theoretical knowledge and construction experience. Generally, the 0–4 integer scoring scale is used to score the degree of influence between two risk factors. Among them, ‘0’ means that there is no influence between the two factors; that is, one factor has no effect on the existence and change in the other factor. ‘1’ represents a small influence, which means that one factor has a certain influence on another factor, but this influence is relatively weak and not very obvious. ‘2’ indicates a general influence, indicating that the interaction between the two factors is at a medium level. ‘3’ represents a strong influence, indicating that one factor has a greater impact on another factor, which can significantly affect the state or change of another factor. ‘4’ indicates a strong influence, indicating that the relationship between the two factors is very close, and a change in one factor will greatly promote or restrict the other factor. Through the initial scoring, 10 direct influence matrices were obtained. In order to eliminate the individual differences of expert scoring, the integer of the average of 10 initial matrices was taken to construct the direct influence matrix of the construction risk factor of the mechanical method cross passage, as shown in

Table 2. The direct influence matrix.

A	${\mathit{X}}_1$	${\mathit{X}}_2$	${\mathit{X}}_3$	${\mathit{X}}_4$	${\mathit{X}}_5$	${\mathit{X}}_6$	${\mathit{X}}_7$	${\mathit{X}}_8$	${\mathit{X}}_9$	${\mathit{X}}_{10}$	${\mathit{X}}_{11}$	${\mathit{X}}_{12}$	${\mathit{X}}_{13}$	${\mathit{X}}_{14}$	${\mathit{X}}_{15}$	${\mathit{X}}_{16}$	${\mathit{X}}_{17}$	${\mathit{X}}_{18}$
${\mathit{X}}_{\mathbf{1}}$	0	2	4	3	2	2	3	3	2	2	2	2	2	2	1	0	0	0
${\mathit{X}}_{\mathbf{2}}$	3	0	2	3	0	0	3	3	2	2	3	3	3	3	2	0	0	0
${\mathit{X}}_{\mathbf{3}}$	2	1	0	3	2	2	2	2	2	2	2	2	2	3	3	0	0	0
${\mathit{X}}_{\mathbf{4}}$	2	1	3	0	2	2	2	2	2	2	2	2	2	3	2	0	0	0
${\mathit{X}}_{\mathbf{5}}$	0	0	0	0	0	0	0	0	4	0	0	2	3	1	0	0	3	0
${\mathit{X}}_{\mathbf{6}}$	0	0	0	0	0	0	0	0	0	3	2	0	0	3	0	0	0	0
${\mathit{X}}_{\mathbf{7}}$	0	0	0	0	0	0	0	2	0	2	0	2	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{8}}$	0	0	0	0	0	0	0	0	0	0	2	2	2	0	0	0	0	0
${\mathit{X}}_{\mathbf{9}}$	0	0	0	0	0	0	0	0	0	1	1	1	0	1	0	0	0	0
${\mathit{X}}_{\mathbf{10}}$	0	0	0	0	0	0	0	0	0	0	1	2	1	1	0	0	0	0
${\mathit{X}}_{\mathbf{11}}$	0	0	0	0	0	0	0	0	0	0	0	2	0	1	0	0	0	0
${\mathit{X}}_{\mathbf{12}}$	0	0	0	0	0	0	0	0	0	2	2	0	2	2	0	0	0	0
${\mathit{X}}_{\mathbf{13}}$	0	0	0	0	0	0	0	0	4	1	2	2	0	1	0	0	3	0
${\mathit{X}}_{\mathbf{14}}$	0	0	0	0	0	0	0	0	0	0	0	2	2	0	0	0	0	0
${\mathit{X}}_{\mathbf{15}}$	0	0	0	0	0	0	0	0	2	2	2	3	1	2	0	0	0	0
${\mathit{X}}_{\mathbf{16}}$	0	0	2	2	0	0	2	4	0	0	2	2	1	1	0	0	3	3
${\mathit{X}}_{\mathbf{17}}$	0	0	2	1	1	1	4	4	2	1	1	3	2	2	0	0	0	0
${\mathit{X}}_{\mathbf{18}}$	0	0	2	1	0	0	1	1	1	0	0	2	1	0	0	0	0	0

.

Using Formula (2), the direct influence matrix A is normalized via the row sum maximum method, and the direct influence matrix B is obtained, in which the row sum maximum value of matrix A is 32. The comprehensive influence matrix T is obtained using Formula (3), as shown in

Table 3. The comprehensive influence matrix.

T	${\mathit{X}}_1$	${\mathit{X}}_2$	${\mathit{X}}_3$	${\mathit{X}}_4$	${\mathit{X}}_5$	${\mathit{X}}_6$	${\mathit{X}}_7$	${\mathit{X}}_8$	${\mathit{X}}_9$	${\mathit{X}}_{10}$	${\mathit{X}}_{11}$	${\mathit{X}}_{12}$	${\mathit{X}}_{13}$	${\mathit{X}}_{14}$	${\mathit{X}}_{15}$	${\mathit{X}}_{16}$	${\mathit{X}}_{17}$	${\mathit{X}}_{18}$
${\mathit{X}}_{\mathbf{1}}$	0.023	0.072	0.145	0.117	0.081	0.081	0.121	0.129	0.116	0.121	0.129	0.151	0.127	0.135	0.057	0.000	0.020	0.000
${\mathit{X}}_{\mathbf{2}}$	0.108	0.013	0.088	0.114	0.020	0.020	0.120	0.127	0.110	0.117	0.157	0.180	0.152	0.157	0.082	0.000	0.016	0.000
${\mathit{X}}_{\mathbf{3}}$	0.074	0.040	0.023	0.107	0.076	0.076	0.084	0.089	0.110	0.113	0.120	0.141	0.118	0.155	0.107	0.000	0.018	0.000
${\mathit{X}}_{\mathbf{4}}$	0.074	0.040	0.109	0.021	0.076	0.076	0.084	0.089	0.108	0.111	0.118	0.138	0.117	0.153	0.079	0.000	0.018	0.000
${\mathit{X}}_{\mathbf{5}}$	0.001	0.000	0.007	0.004	0.004	0.004	0.014	0.015	0.147	0.019	0.023	0.093	0.112	0.055	0.001	0.000	0.105	0.000
${\mathit{X}}_{\mathbf{6}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.095	0.067	0.017	0.010	0.100	0.000	0.000	0.001	0.000
${\mathit{X}}_{\mathbf{7}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.063	0.001	0.067	0.011	0.073	0.011	0.007	0.000	0.000	0.001	0.000
${\mathit{X}}_{\mathbf{8}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.001	0.009	0.007	0.072	0.073	0.068	0.010	0.000	0.000	0.006	0.000
${\mathit{X}}_{\mathbf{9}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.034	0.035	0.038	0.006	0.036	0.000	0.000	0.001	0.000
${\mathit{X}}_{\mathbf{10}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.005	0.006	0.039	0.071	0.039	0.039	0.000	0.000	0.004	0.000
${\mathit{X}}_{\mathbf{11}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.004	0.005	0.066	0.007	0.036	0.000	0.000	0.001	0.000
${\mathit{X}}_{\mathbf{12}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.001	0.009	0.066	0.071	0.019	0.071	0.071	0.000	0.000	0.007	0.000
${\mathit{X}}_{\mathbf{13}}$	0.001	0.000	0.006	0.004	0.004	0.004	0.013	0.014	0.134	0.047	0.080	0.091	0.018	0.053	0.001	0.000	0.096	0.000
${\mathit{X}}_{\mathbf{14}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.001	0.009	0.007	0.009	0.069	0.068	0.008	0.000	0.000	0.006	0.000
${\mathit{X}}_{\mathbf{15}}$	0.000	0.000	0.000	0.000	0.000	0.000	0.001	0.001	0.069	0.073	0.077	0.114	0.046	0.078	0.000	0.000	0.004	0.000
${\mathit{X}}_{\mathbf{16}}$	0.011	0.006	0.084	0.078	0.014	0.014	0.090	0.158	0.033	0.033	0.103	0.126	0.077	0.072	0.013	0.000	0.102	0.094
${\mathit{X}}_{\mathbf{17}}$	0.007	0.004	0.068	0.039	0.039	0.039	0.134	0.143	0.089	0.067	0.071	0.147	0.101	0.099	0.009	0.000	0.013	0.000
${\mathit{X}}_{\mathbf{18}}$	0.007	0.004	0.068	0.039	0.007	0.007	0.040	0.042	0.047	0.020	0.022	0.085	0.050	0.022	0.009	0.000	0.005	0.000

.

According to Formulas (4)~(7), the influence degree, affected degree, centrality, and cause degree of each risk factor are calculated, as shown in

Table 4. The calculation results of DEMATEL.

Risk Factors	Influence Degree	Affected Degree	Centrality Degree	Cause Degree
$X_1$	1.625	0.306	1.930	1.319
$X_2$	1.583	0.179	1.762	1.403
$X_3$	1.452	0.600	2.052	0.852
$X_4$	1.410	0.524	1.934	0.886
$X_5$	0.604	0.321	0.925	0.282
$X_6$	0.293	0.321	0.615	−0.028
$X_7$	0.236	0.703	0.938	−0.467
$X_8$	0.250	0.872	1.122	−0.622
$X_9$	0.151	0.999	1.149	−0.848
$X_{10}$	0.203	1.010	1.213	−0.807
$X_{11}$	0.119	1.210	1.329	−1.091
$X_{12}$	0.317	1.692	2.008	−1.375
$X_{13}$	0.565	1.199	1.764	−0.635
$X_{14}$	0.180	1.287	1.467	−1.107
$X_{15}$	0.463	0.360	0.823	0.103
$X_{16}$	1.108	0.000	1.108	1.108
$X_{17}$	1.069	0.424	1.493	0.645
$X_{18}$	0.473	0.094	0.567	0.380

. According to the calculation results in Table 4, a causal diagram of the construction risk factors of the mechanical cross passage is drawn, as shown in Figure 3.

According to the causal diagram (Figure 3) of the construction safety risk factors of the mechanical method cross passage calculated by DEMATEL, it can be seen that the construction safety risk factors of the mechanical method cross passage can be divided into four categories.

All factors in Region I are strong cause risk factors, which have a very significant impact on the formation of the construction risk of the mechanical cross passage and have a greater impact on other result-based risk factors. According to the degree of centrality, the risk factors in the first area are sorted from large to small, as follows: $x_3$ (the safety construction system is not perfect, and it is not fully constructed according to the established system), $x_4$ (insufficient construction supervision), $x_1$ (the safety awareness of the construction personnel is insufficient), $x_2$ (the safety technology level of the construction personnel is low) and $x_{17}$ (geological complexity).

All factors in Region II are weak outcome risk factors, which is the result of the combined effects of other causal risk factors. All factors in Region III are weak cause risk factors, which also have a certain impact on other outcome factors.

All factors in Region IV are strong outcome risk factors. This kind of risk factor is also the result of the combined action of other causal risk factors, which has a very important impact on the formation of the construction risk of the mechanical cross passage. According to the degree of centrality, the risk factors of the first area are sorted from large to small, as follows: $x_{12}$ (unreasonable setting of tunneling parameters), $x_{13}$ (poor grouting effect), $x_{14}$ (the segments are not assembled in time, the segments are staggered, and the bolts are not tightened) and $x_{11}$ (the unreasonable loading of the internal support system leads to the deformation of the main tunnel segment).

The degree of centrality indicates the role of risk factors in the construction process of the cross passage. Therefore, more attention should be paid to the strong cause risk factors in Zone I and the strong result risk factors in Zone IV.

4.2. ISM Model

According to Formula (8), the overall influence matrix E is obtained. In order to obtain the reachable matrix F, it is necessary to introduce a threshold λ to facilitate the division of the hierarchical structure. According to the overall influence matrix and expert opinions, λ = 0.13 is introduced. The reachable matrix is shown in

Table 5. The reachable matrix.

F	${\mathit{X}}_1$	${\mathit{X}}_2$	${\mathit{X}}_3$	${\mathit{X}}_4$	${\mathit{X}}_5$	${\mathit{X}}_6$	${\mathit{X}}_7$	${\mathit{X}}_8$	${\mathit{X}}_9$	${\mathit{X}}_{10}$	${\mathit{X}}_{11}$	${\mathit{X}}_{12}$	${\mathit{X}}_{13}$	${\mathit{X}}_{14}$	${\mathit{X}}_{15}$	${\mathit{X}}_{16}$	${\mathit{X}}_{17}$	${\mathit{X}}_{18}$
${\mathit{X}}_{\mathbf{1}}$	1	0	1	0	0	0	0	0	0	0	0	1	0	1	0	0	0	0
${\mathit{X}}_{\mathbf{2}}$	0	1	0	0	0	0	0	0	0	0	1	1	1	1	0	0	0	0
${\mathit{X}}_{\mathbf{3}}$	0	0	1	0	0	0	0	0	0	0	0	1	0	1	0	0	0	0
${\mathit{X}}_{\mathbf{4}}$	0	0	0	1	0	0	0	0	0	0	0	1	0	1	0	0	0	0
${\mathit{X}}_{\mathbf{5}}$	0	0	0	0	1	0	0	0	1	0	0	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{6}}$	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{7}}$	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{8}}$	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{9}}$	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{10}}$	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{11}}$	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{12}}$	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0
${\mathit{X}}_{\mathbf{13}}$	0	0	0	0	0	0	0	0	1	0	0	0	1	0	0	0	0	0
${\mathit{X}}_{\mathbf{14}}$	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0
${\mathit{X}}_{\mathbf{15}}$	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0
${\mathit{X}}_{\mathbf{16}}$	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0
${\mathit{X}}_{\mathbf{17}}$	0	0	0	0	0	0	1	1	0	0	0	1	0	0	0	0	1	0
${\mathit{X}}_{\mathbf{18}}$	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1

.

According to Table 5, the reachable set $R\left(x_i\right)$ and the leading set $S\left(x_i\right)$ of the construction risk factors of the mechanical cross passage can be obtained. The calculation shows that when $i=6,7,8,9,10,11,12,14,15,18$, $R\left(x_i\right)\cap S\left(x_i\right)=R\left(x_i\right)$ is satisfied. That is, $x_6,x_7,x_8,x_9,x_{10},x_{11},x_{12},x_{14},x_{15}$ and $x_{18}$ are the first-level risk factors. The first-level risk factors are eliminated in the rows and columns of matrix F to obtain a new matrix. By repeating the above steps in the new matrix, $x_3,x_4,x_5,x_{13},x_{16},$ and $x_{17}$ are the second-level risk factors, and $x_1$ and $x_2$ are the third-level risk factors. Therefore, the construction risk factors of the mechanical cross passage are divided into three levels, as shown in Figure 4.

According to the calculation results of the ISM model (Figure 4), the construction safety risk factors of the mechanical cross passage can be divided into three levels. Among them, $x_1$ (the safety awareness of the construction personnel is insufficient) and $x_2$ (the safety technology level of the construction personnel is low) are deep risk factors. $x_3$ (the safety construction system is not perfect, and it is not fully constructed according to the established system), $x_4$ (insufficient construction supervision), $x_5$ (poor quality of grouting material), $x_{13}$ (poor grouting effect), $x_{16}$ (length of excavation), and $x_{17}$ (geological complexity) are intermediate risk factors. The rest are direct risk factors.

In the ISM model, $x_1$ (the safety awareness of the construction personnel is insufficient), $x_2$ (the safety technology level of the construction personnel is low), $x_{12}$ (unreasonable setting of tunneling parameters), $x_{13}$ (poor grouting effect), $x_{14}$ (the segments are not assembled in time, the segments are staggered, and the bolts are not tightened), and $x_{17}$ (geological complexity) are the risk factors with more nodes. These factors are consistent with the risk factors with larger centrality values calculated by the DEMATEL model.

5. Risk Response Measures

In the calculation results of DEMATEL, the strong cause risk factor in Zone I and the strong result risk factor in Zone IV are the key risk factors. The cause and centrality of the strong cause risk factors are both large, which include $x_1$ (the safety awareness of the construction personnel is insufficient), $x_2$ (the safety technology level of the construction personnel is low), $x_3$ (the safety construction system is not perfect, and it is not fully constructed according to the established system), $x_4$ (insufficient construction supervision), and $x_{17}$ (geological complexity). The centrality of the strong outcome risk factor is large, which includes $x_{11}$ (the unreasonable loading of the internal support system leads to the deformation of the main tunnel segment), $x_{12}$ (unreasonable setting of tunneling parameters), $x_{13}$ (poor grouting effect), and $x_{14}$ (the segments are not assembled in time, the segments are staggered, and the bolts are not tightened). Based on the key risk factors, appropriate risk control measures are developed.

5.1. Measures to Deal with Risk Factors $x_1$, $x_2$, $x_3,$ and $x_4$

The four risk factors of personnel and management factors are the key risk factors. It can be seen that people occupy a dominant position in the construction process of the mechanical cross passage. All construction relies on people: safe construction environments need to be built by humans, advanced technical methods need to be operated by humans, and comprehensive management systems need to be implemented by humans.

Therefore, a safe construction system should be formulated before construction, and regular training should be organized for operators and managers to enhance the level of construction management. Personnel should carry out pre-class education before taking office and avoid fluke psychology. Shield machine and pipe jacking machine are special equipment, so the operator must have a special equipment operating license and standardize the operation of the equipment. Machinery is controlled by human beings, who have subjective consciousness, so it is necessary to maintain a high degree of vigilance and concentration throughout the construction work cycle.

5.2. Measures to Deal with Risk Factor $x_{17}$ (Geological Complexity)

A comprehensive investigation of the geological conditions is the basis for the construction of the cross passage. Samples should be obtained by drilling cores to understand the type of formation. According to different geological conditions, the corresponding tunneling equipment is selected. For typical soft soil layers such as silt and silty clay, an earth pressure balance shield machine is the first choice. It can keep balance with the water and soil pressure of the excavation face by controlling the soil pressure in the soil bin, and it can effectively prevent the collapse of the excavation face. For the water-rich sand layer, a slurry shield machine has unique advantages. It uses slurry pressure to balance the water and soil pressure of the excavation face, and it discharges the cut sand through the slurry circulation system, which can better control the stability of the excavation face and prevent water and sand gushing. TBM is an ideal choice for rock strata with good integrity and high strength. It has a strong rock breaking ability, which can squeeze and break the rock through the disc cutter on the cutterhead to achieve rapid excavation. In short, according to different geological conditions, the selection of corresponding roadheader equipment plays an important role in the safe construction of the mechanical cross passage.

5.3. Measures to Deal with Risk Factor $x_{11}$ (the Unreasonable Loading of the Internal Support System Leads to the Deformation of the Main Tunnel Segment)

In the initial excavation of the cross passage, the horizontal thrust and reaction force along the excavation direction act on the main tunnel segment, so the deformation of the main tunnel segment will be increased, which may cause the segment to break and endanger the safety of the construction of the cross passage. Therefore, the integrated internal support system is set up on the launching trolley and the receiving trolley in the position of the hole door. It plays the role of temporary support in the process of stress redistribution after cutting the special segment of the main tunnel. The schematic diagram of the support system is shown in Figure 5.

From the calculation results of ISM (Figure 4), it can be seen that the deep risk factor $x_2$ has a very significant impact on the direct risk factor $x_{11}$. Therefore, corresponding technical measures should be taken to reduce the risk. First, one should check whether the range of the jack meets the loading requirements and calibrate the accuracy to ensure that the loading force can be accurately controlled. Secondly, monitoring instruments should be installed, including strain gauges and pressure sensors. The strain gauges should be pasted on the key parts of the support structure, such as the mid-span and nodes of the support, to accurately measure the strain change of the support system during the loading process. The pressure sensor is used to monitor the magnitude of the loading force. In addition, the internal support system needs to be loaded in stages, and the loading speed should be moderate. In general, the loading speed should be determined according to factors such as geological conditions. For example, in the cross passage in the soft soil layer, a loading speed that is too fast may cause the soil around the passage to fail to adapt to the change in support force, resulting in excessive soil deformation.

During the whole process of the construction of the cross passage, the deformation and stress of the supporting system are continuously monitored. During the loading process, the strain, stress, and loading force data of the support are obtained in real time through monitoring equipment such as strain gauges and pressure sensors. The loading scheme is adjusted in time according to the monitoring data. If the monitoring data are abnormal—for example, if the deformation of the support exceeds the warning value allowed by the design—then the loading should be stopped immediately, the reasons should be analyzed, and corresponding measures should be taken, such as strengthening the support, adjusting the loading sequence, etc. Then, loading can continue.

5.4. Measures to Deal with Risk Factor $x_{12}$ (Unreasonable Setting of Tunneling Parameters)

From the ISM calculation results (Figure 4), it can be seen that the deep risk factors $x_1$ (the safety awareness of the construction personnel is insufficient) and $x_2$ (the safety technology level of the construction personnel is low) and the intermediate risk factors $x_3$ (the safety construction system is not perfect, and it is not fully constructed according to the established system), $x_4$ (insufficient construction supervision), and $x_{17}$ (geological complexity) have an important influence on the direct risk factor $x_{12}$ (unreasonable setting of tunneling parameters).

The comprehensive investigation of the geological conditions of the location of the cross passage is the basis. According to the geological survey results, the problems that may be encountered in the tunneling process can be preliminarily estimated to provide a basis for determining the correct tunneling parameters. After the equipment is installed, it should be fully debugged. One should check whether the adjustment function of the cutterhead speed, propulsion speed, torque, and other parameters of the excavation machine is normal to prepare for adjusting the tunneling parameters according to the actual situation in the construction process, considering the adjustment strategy of tunneling parameters under different geological conditions. For example, when the stratum changes, one should consider how to adjust the cutter head speed, propulsion, grouting pressure, grouting volume, and so on according to the hardness change of the stratum.

Various monitoring equipment should be installed to obtain tunneling parameters in real time. A high-precision sensor is equipped on the shield machine to monitor parameters such as cutterhead speed, propulsion speed, torque, and soil chamber pressure. At the same time, the deformation of the stratum during the excavation process is monitored. Through the settlement monitoring points and horizontal displacement monitoring points arranged around the cross passage, the deformation of the stratum is measured regularly using instruments such as level and total station. According to the real-time monitoring data, the tunneling parameters are adjusted in time. When it is found that the ground settlement exceeds the warning value, it is necessary to reduce the pressure of the soil bin, appropriately reduce the propulsion speed, and increase the amount of synchronous grouting to fill the gap left by the shield machine after propulsion and reduce the ground settlement.

Regarding professional training for construction operators, operators should be familiar with the working principle and operation process of tunneling equipment, and they should be able to correctly understand and use various tunneling parameters. Strict operating norms should also be established. It is stipulated that during the excavation process, the operator must operate in accordance with the construction plan and operating procedures.

5.5. Measures to Deal with Risk Factor $x_{13}$ (Poor Grouting Effect)

During the tunneling process of the cross passage, the concrete segments need to be polished at both the starting and receiving ends. In order to improve the compactness of the soil layer behind the main tunnel segment at the location of the cross passage and improve the soil at the starting and receiving portals, it is necessary to grout and reinforce the soil at the starting and receiving ends before the construction of the cross passage to prevent the excessive settlement of the stratum and the instability of the main tunnel structure during the excavation process. The grouting range is shown in Figure 6.

From the ISM calculation results (Figure 4), it can be seen that deep risk factors $x_1$ and $x_2$ have a very significant impact on the formation of intermediate risk factor $x_{13}$.

Detailed geological exploration is carried out on the soil at the starting end and the receiving end. According to the geological exploration results, the appropriate grouting scheme is determined, including the grouting method, grouting pressure, grouting amount, grouting hole layout, and other parameters.

The selection of grouting materials has a very important influence on the grouting effect. For sandy soil, a cement–water glass double slurry can be used, which has a short setting time and high early strength. It can quickly fill the pores of sand and improve the strength and impermeability of soil. For cohesive soil, ordinary cement slurry or cement slurry with an appropriate amount of admixture can be used to improve the performance of the cement slurry and improve its diffusion effect in cohesive soil. After grouting reinforcement is completed, drill core sampling inspection is carried out on the reinforcement. After the reinforcement strength meets the design requirements, the hole will be sealed.

The grouting pressure and grouting amount should be strictly controlled. If the grouting pressure is too small, the slurry may not be able to effectively fill the pores of the soil. If the grouting pressure is too large, it may cause problems such as soil splitting or ground uplift. During the grouting process and after the grouting is completed, the surrounding environment is monitored, including ground subsidence, building deformation, and groundwater level changes. According to the monitoring data, the grouting parameters are adjusted in time to ensure good grouting effect.

5.6. Measures to Deal with Risk Factor $x_{14}$ (the Segments Are Not Assembled in Time, the Segments Are Staggered, and the Bolts Are Not Tightened)

From the ISM calculation results (Figure 4), it can be seen that the deep risk factor $x_2$ (the safety technology level of construction personnel is low) and the intermediate risk factors $x_3$ (the safety construction system is not perfect, and it is not fully constructed according to the established system) and $x_4$ (insufficient construction supervision) have an important influence on the direct risk factor $x_{14}$ (the segments are not assembled in time, the segments are staggered, and the bolts are not tightened).

After the pipe section arrives at the construction site, the quality inspector and the supervisor must carry out strict quality inspection of the segment, including dimensional accuracy, appearance quality, concrete strength, impermeability, and other performance indicators. The segments should be stacked according to the requirements of the specifications, and the segments of each layer should be separated by padding wood to prevent the deformation of the segments.

Following the correct segment assembly sequence, the bottom segment is generally installed first, and then the two sides and the top segment are installed in turn. Appropriate assembly methods are adopted, and uniform extrusion and symmetrical assembly are adopted to avoid the problems of dislocation and damage caused by uneven stress on segments. The connecting bolts between the segments should be installed and tightened according to the design requirements. In the assembly process, the bolts should be tightened first, and after the assembly of the whole ring segment is completed, the bolts should be tightened again to make the bolts reach the specified torque value and ensure the reliability of the segment connection.

High-precision measuring instruments are used to accurately measure the axis and elevation of the connecting channel to provide an accurate reference for segment assembly. In the assembly process, the position and attitude of the segment should be monitored in real time and adjusted in time to ensure the accuracy of the segment assembly.

6. Conclusions

In this paper, a risk analysis model for the construction of a mechanical cross passage based on DEMATEL-ISM is constructed. Through the identification of the construction safety risk factors of the mechanical cross passage, the DEMATEL analysis method is used to calculate the centrality and cause degree of each risk factor, then the influence degree of each risk factor is determined. The deep risk factors are calculated by the ISM model, and the hierarchical structure between all risk factors is divided to determine the risk factors with higher node degrees.

(1) In the DEMATEL model, the factors in Zone I are strong causal risk factors, including the establishment and implementation of a safety construction system ${(x}_3)$, construction management level ${(x}_4)$, safety awareness of construction personnel ${(x}_1)$, safety technology level of construction personnel ${(x}_2)$, and geological complexity ${(x}_{17})$. Such risk factors have a very significant impact on the formation of the construction risk of the mechanical cross passage and have a greater impact on other result-based risk factors. The factors in Zone IV are strong outcome risk factors, including tunneling parameter setting ${(x}_{12})$, grouting effect ${(x}_{13})$, segment assembly effect ${(x}_{14}),$ and internal support system reinforcement effect ${(x}_{11})$. This kind of risk factor is the result of the comprehensive effect of other causal risk factors, which also has a very important influence on the formation of the construction risk of the mechanical cross passage.

(2) In the ISM model, the safety awareness of the construction personnel ($x_1$), the safety technology level of the construction personnel ($x_2$), the setting of tunneling parameters ($x_{12}$), the grouting effect ($x_{13}$), the effect of segment assembly ($x_{14}$), and the geological complexity ($x_{17}$) are the risk factors with more nodes.

(3) The risk factors with a higher node degree in ISM are consistent with the risk factors with higher centrality in the DEMATEL model, which are key factors and play an important role in risk control. Based on the identified key risk factors, appropriate risk control measures can be developed in a timely and effective manner.