An Investigation of the Process of Risk Coupling and the Main Elements of Coal-Mine Gas-Explosion Risk

Abstract

This study suggests a method for analyzing the risk of methane explosions using the N-K model and Social Network Analysis (SNA) to understand how different risk factors related to coal-mine methane explosions are connected and change over time, aiming to prevent these accidents effectively. We identified 41 secondary risk factors and four fundamental risk factors—human, equipment, environment, and management—based on the 4M accident causation theory. The SNA model was utilized to determine the main risk factors and their evolutionary routes, while the N-K model was utilized to quantify the degree of risk coupling. The findings show that the number of risk variables engaged in the methane-explosion risk system closely correlates with the number of accidents that occur and the maximum coupling level among the four elements. The primary control factors in the methane-explosion risk system are poor equipment management, broken safety monitoring and control systems, inadequate safety education and training, safety regulation violations, and poor safety production responsibility system implementation. We utilized the primary evolution paths and key elements to propose risk control approaches. A reference for ensuring safety in coal-mine operations can be found in the research findings.

1. Introduction

Gas disasters are still the main disasters of deep coal mining, and with increases in mining depth and mining intensity, the danger is becoming more and more serious. Influencing factors and coupling relationships of gas disasters are very complex, and disaster accidents happen occasionally [1,2]. The risk-triggering mechanism of coal-mine gas explosions is the result of the coupling of multiple factors, which is difficult to measure accurately; therefore, it is crucial to examine the risk coupling mechanism and calculate the coupling evolution analysis of the coupling effect for coal-mine gas-explosion risk management.

Recently, many scholars have carried out relevant research in the field of coal-mine gas-explosion risk. Li et al. [3] deducted the causal factors of coal-mine gas-explosion accidents from micro-, meso-, and macro-levels and found that violation of operating procedures, chaos in the ventilation system, and illegal organization of production are the factors that cause a high probability of coal-mine gas-explosion accidents. Guo et al. [4] investigated the interaction relationship and attribute characteristics among the causal factors of gas explosion accidents from a quantitative perspective. NIAN QF et al. [5] used the gray correlation method to analyze the main risk factors of coal-mine gas explosions and constructed a gas-explosion risk assessment system based on the four factors of “human-equipment-environment-manegement.” An approach for assessing the danger of gas explosions based on game theory and Bayesian networks was proposed by Lin et al. [6]. These studies have laid a solid foundation for the scientific management of coal-mine gas explosion risk; however, less systematic research has taken place on the coupling of risk factors and evolution mechanisms, etc.; Fox [7] and KHUFFMAN [8], who proposed the N-K model, aimed at starting from the objective data of accidents to quantify the measured values of the risk factors to support the evaluation results and to reduce subjective judgment in the evaluation process. Applications in the field of underground passage [9], well blowout accidents [10,11], ship accidents [12], subway construction risk [13,14], tunnel construction risks [15,16,17], etc., offer the theoretical underpinnings and methodological support for the coupled evolution of coal-mine gas-explosion risk, and show that the N-K coupling model is feasible for assessing safety state and analyzing risk factors, among other things.

Social network analysis uses images and mathematical methods to quantitatively analyze the roles and interrelationships of the factor nodes within the system, which can present the evolution paths between the nodes well and identify the key factors in the risk system [18]. Miao et al. [19] visualized the causes of coal-mine roof incidents using the SNA model. Using the SNA technique, Chen et al. [20] determined the primary risk factors in subterranean engineering projects and examined the risk dissemination effect. A complicated network model of the contributing factors of chemical accidents in China between 2015 and 2020 was built by Yang et al. [21]. To some extent, social network analysis ignores the effects of node combinations on the system in favor of concentrating on the relationships between nodes. Gas-explosion risk constitutes a complex risk system; there is a coupling effect between the risk factors, and only considering the interrelationship between the risk nodes may neglect the riskogenicity of the risk-factor coupling of the accident. There have been some studies integrating the N-K model and the SNA model in tailing-pond dam-failure risk [22], the explosion risk of chemical enterprises [23], construction risks of buildings [24], and fire risks of the buildings [25], and other areas with complex risk systems have seen some research carried out, providing research ideas for risk assessment and risk factors.

Accordingly, this research is based on the investigation reports of 105 gas-explosion events that occurred between 2011 and 2013. To identify the main risk factors and crucial evolution paths of gas explosions, it merges the N-K and SNA models and examines the risk factors of gas-explosion events from four perspectives: human, equipment, environment, and management. The research findings offer a theoretical foundation for averting gas explosion mishaps, and they also suggest risk-prevention and -control strategies.

2. Materials and Methods

2.1. Identification of Risk Factors for Gas Explosions in Coal Mines

An analysis of 105 (non-complete statistics) coal-mine gas-explosion accident-investigation reports published by China’s Ministry of Emergency Management (MEM), the State Mine Safety Supervision Bureau (SMSSB), local mine safety supervision bureaus (MSBs), and the Safety Management Network (SMN) from 2011 to 2023 was undertaken because coal-mine accident-investigation reports are the most logical representation of the patterns and characteristics of coal-mine disaster accidents.

A grounded theoretical analysis of coal-mine gas-explosion risk factors is constructed by the article using NVivo11 Pro software to qualitatively analyze the obtained accident reports. This analysis provides data support for the construction of the index system of coal-mine gas-explosion risk factors. The safety system engineering theory [26,27,28,29,30,31] divides the risk factors for coal-mine gas explosions into four categories: management unsafe factors (M), environmental unsafe factors (E), equipment unsafe factors (P), and human unsafe factors (H). Figure 1 displays each risk factor’s distribution.

2.2. Coal-Mine Gas-Explosion Risk Elements and Their Coupling Mechanism

2.2.1. Coupling Mechanisms of Coal Mine Gas Explosion Risk Factors

The phenomenon of dependency, mutual effect, and interaction between the risk variables of different subsystems in the coal mine production process is known as coal-mine gas-explosion risk coupling. Coal-mine gas-power systems are examples of typical complex systems. In order to analyze the formation mechanism of coal-mine gas-explosion risk coupling, this paper uses the concept of trigger [32,33]. As illustrated in Figure 2, different risk factors combined with other risk-factor categories will destroy the system’s initial equilibrium state, increasing the coupled risk or creating new risks that will cause accidents.

2.2.2. Type of Coal-Mine Gas-Explosion Risk-Factor Pairing

The coal-mine gas-explosion risk-factor coupling is divided into three categories—single-factor coupling, two-factor coupling, and multi-factor coupling—based on the definition of the idea and the explanation of the risk coupling process.

(1)

Single-factor coupling: In other words, a risk subsystem’s mutual effect and interaction, including the risk coupling of similar factors like human, equipment, environment, and management, and the associated risk coupling value, are represented as T₁₁, T_12, T₁₃, and T₁₄.

(2)

Two-factor coupling: The reciprocal influence and interaction between two risk subsystems, which includes six different forms of coupling—human–equipment, human–environment, human–management, equipment–environment, equipment–equipment, and environment–management—is represented by the relevant risk coupling values, which are T₂₁, T₂₂, T_23, T_24, T₂₅, and T₂₆.

(3)

Multi-factor coupling: This study includes three-factor and four-factor risk coupling, which include five different types of coupling: human–equipment–environment, human–equipment–management, human–equipment–environment–management, equipment–environment–management, and human–equipment–environment-management. The corresponding risk coupling values are indicated as T₃₁, T₃₂, T₃₃, T₃₄, and T₄.

2.3. Method of Key Factor Analysis for Combining N-K and SNA Models

Although the N-K model somewhat decreases subjectivity by using objective data from accident cases to calculate risk-coupling effects, it still has trouble figuring out key risk factors. Nevertheless, the outcomes that result in dangerous events and the combinations of risk factors differ, suggesting that relying exclusively on network analysis could lead to bias. The SNA model can examine important factors based on the structure of the relationships between risk factors. Thus, the N-K and SNA models are combined in this study to create an analytical model of coal-mine gas-explosion risk variables. Figure 3 illustrates the analytical framework.

2.3.1. Building the N-K Risk-Coupling Model

The N-K model’s two key parameters, N and K, indicate the total number of factors in a complex system and the number of mutual coupling links between those factors, respectively. There are n^N different types of coupling modes in the system if there are N influencing factors and each of those factors has n sub-factors. When K is greater than a particular value, the system’s coupling relationship can form a complex relational network where 0 ≤ K ≤ N − 1. The interaction information T in information theory is used to express the N classes in the complex system’s network state in order to more effectively assess the problem of the network state. The following formula is used to determine the interaction information between the group elements:

$$T\left(a,b,c,\dots ,n\right)={\displaystyle \sum _{I_1=1}^n{\displaystyle \sum _{I_2=1}^n{\displaystyle \sum _{I_3=1}^n\dots {\displaystyle \sum _{I_N=1}^n{P}_{I_1,I_2,\dots ,I_N}{\log }_2(\frac{P_{I_1,I_2,\dots ,I_N}}{P_{I_1}\cdot P_{I_2}\dots P_{I_N}})}}}}$$

(1)

$$T\left(H,P,E,M\right)={\displaystyle \sum _{l=1}^L{\displaystyle \sum _{m=1}^M{\displaystyle \sum _{n=1}^N{\displaystyle \sum _{o=1}^O{P}_{lmno}}{\log }_2(\frac{P_{lmno}}{P_l\cdot P_m\cdot P_n\cdot P_o})}}}$$

(2)

$$P_l={\displaystyle \sum _m^M{\displaystyle \sum _n^N{\displaystyle \sum _o^O{P}_{lmno}}}}\hspace{1em}\forall l\in \left\{1,\dots ,L\right\}$$

(3)

Among these, $a,b,c,\dots ,n$ represents the risk factors of category N, while $P_{I_1,I_2,\dots ,I_N}$ represents the likelihood of coupling under factors an in state I₁, b in state I₂, and n in state I_n. The higher the estimated T-value, the more the combination affects system safety. In the sequence of human, equipment, environment, and management, the letters H, P, E, and M stand for the four components of risk factors; the likelihood that behavioral factors in the ‘l’ state, equipment factors in the ‘m’ state, environmental elements in the ‘n’ state, and management factors in the ‘o’ state will couple is indicated by P_lmno.

Since the coupling between components within the four different types of risk factor systems is known as single-factor risk-factor coupling and cannot be measured using interactive information, this study only looks into two-factor and multi-factor coupling risk. Six different two-factor risk couplings and four different multi-factor risk couplings exist, such as “human-equipment”, “human-equipment-environment”, and “human-equipment-environment-management”. Here, (4) through (6) display the calculating formulas.

$$T_{21}\left(H\&P\right)={\displaystyle \sum _{l=1}^L{\displaystyle \sum _{m=1}^M{P}_{lm}{\log }_2(\frac{P_{lm}}{P_l\cdot P_m})}}$$

(4)

$$T_{31}\left(H\&P\&E\right)={\displaystyle \sum _{l=1}^L{\displaystyle \sum _{m=1}^M{\displaystyle \sum _{n=1}^n{P}_{l,m,n}}}{\log }_2(\frac{P_{lmn}}{P_l\cdot P_m\cdot P_n})}$$

(5)

$$T_{41}\left(H\&P\&E\&M\right)={\displaystyle \sum _{l=1}^L{\displaystyle \sum _{m=1}^M{\displaystyle \sum _{n=1}^N{\displaystyle \sum _{o=1}^O{P}_{l,m,n,o}}}}{\log }_2(\frac{P_{lmno}}{P_l\cdot P_m\cdot P_n\cdot P_o})}$$

(6)

2.3.2. Construction of a Model for SNA Risk Evolution Analysis

The linkages and interactions between influencing factors in complicated issues can be studied using the quantitative analysis method known as structural equation analysis (SNA). It constructs the network of complex problems by gathering and examining the interconnections and interactions between people or organizations. This allows one to determine the evolutionary relationships between elements and the influence of each component in the system. The investigation of coal-mine gas-explosion incidents is used as a case study in this research. The expert consultation approach is used to create an adjacency matrix of risk variables (Supplementary Materials). The 0–1 relationship adjacency matrix is then visualized using NetDraw, as seen in Figure 4. The induced consequences of risk factors are represented by arrows in this directed social network of risk factors.

Key influencing factors and important transmission pathways in the methane explosion risk system are identified in this study using centrality analysis. Closeness, which gauges the effectiveness of information transfer between nodes, shows how close a node is to other nodes. It is the clearest indicator of how significant risk variables are. A factor’s influence in the system increases with its Closeness. The equation is as follows:

$$C_c{\left(e\right)}^{-1}={\displaystyle \sum _{f=1}^m{d}_{ef}}\hspace{1em}e\ne f$$

(7)

Betweenness, which is determined by the following formula, is a measure of a factor’s positional relevance in a risk system in relation to other factors. It is the frequency of a node’s occurrence in all of the network’s shortest paths.

$$C_B\left(e\right)={\displaystyle \sum _{f=1}^m{\displaystyle \sum _{l=1}^m\frac{g_{fl}\left(e\right)}{g_{fl}}}}\hspace{1em}e\ne f\ne l,f<l$$

(8)

3. Result

3.1. N-K Model Risk-Coupling Analysis

The frequency of occurrence of risk coupling between variables is determined based on the findings of the prior identification of coal-mine gas-explosion risk factors; a value of “0” denotes the absence of a risk factor, while a value of “1” denotes its occurrence [34]. For instance, “0110” denotes the presence of environmental and equipment hazards but not of behavior or management.

Table 1. Combining the likelihood and frequency of explosive risk factors for gas explosions.

Single-Factor Coupling	Number	Frequency	Two-Factor Coupling	Number	Frequency	Multi-Factor Coupling	Number	Frequency
0000	2	0.019	1100	0	0.000	1110	11	0.105
1000	0	0.00	1010	1	0.009	1101	7	0.067
0100	0	0.00	1001	5	0.048	1011	12	0.114
0010	3	0.029	0110	1	0.009	0111	4	0.038
0001	2	0.019	0101	1	0.009	1111	55	0.524
			0011	1	0.009

displays statistics on 105 coal-mine gas-explosion incidents, the frequency of risk factors occurring, and the frequency and number of instances of risk coupling.

Equations (2)–(6) are employed to determine, using Table 1′s risk-coupling frequency and frequency, the coupling level of multi-factor risk coupling and two-factor risk coupling of gas-explosion events.

Table 2. Values of the risk coupling under various coupling forms.

T2	T21	T22	T23	T24	T25	T26
	0.0394	0.0426	0.0249	0.0174	0.0153	0.0198
T3	T31	T32	T33	T34	T4	T4
	0.0634	0.1089	0.0812	0.0985		0.2035

presents the findings, showing the coupling level from high to low, as follows: the mean value of risk coupling for each heterogeneous factor is determined as follows: T₄ > T₃₂ > T₃₄ > T₃₃ > T₃₁ > T₂₂ > T₂₁ > T₂₃ > T₂₆ > T₂₄ > T₂₅, $\overline{{\mathrm{T}}_4}$ = 0.2035, $\overline{{\mathrm{T}}_3}$ = 0.088, $\overline{{\mathrm{T}}_2}$ = 0.0267. This can be inferred from the T-value results.

Visualize the content of Table 2 to obtain the result as shown in Figure 5.

Increasing the number of coupled factors has a positive correlation with the T-value of the interaction information of risk-factor coupling. The highest level of four-factor risk coupling has the highest T-value (0.2035), followed by three-factor risk coupling and two-factor risk coupling. As the number of couplings increases, so does the mean coupling value; T₄ is about ten times higher than ${\overline{\mathrm{T}}}_2$ and more than twice as high as ${\overline{\mathrm{T}}}_3$. As a result, employees should actively look for risk factors that influence the likelihood of accidents in their regular job, minimize the impact of risk, and find fewer risk factor couplings.

T₃₂ and T₃₄ rank highest in the three-factor risk coupling, suggesting that when the personnel, equipment, and management factors are working in tandem and share common characteristics, it is simple to overcome risk thresholds and barriers against risk. The similar features of the equipment and management elements show that the equipment was in a hazardous state and that the key variables causing the gas explosion were not properly managed.

In the two-factor risk coupling, T₂₂, T₂₁, T₂₃, and T₂₆ exhibit larger coupling-value risks; it is evident that the environmental and personnel factors are coupled to the second highest degree, the personnel and equipment factors are coupled to the second highest degree, and the personnel and environmental factors are coupled to the highest risk. The system is more vulnerable when personnel, environmental, and equipment risks are combined, as is evident in the second-highest risk. During the production process, attention should be paid to worker education and training in addition to the dependability of the gas monitoring equipment and the prompt detection of gas accumulation to avoid operational errors.

3.2. Examination of SNA Model Output

3.2.1. Centrality Analysis

The Closeness and Betweenness of each node are determined using Ucinet6.0 software using Equations (7) and (8), using the adjacency matrix of risk factors displayed in Figure 3 as the data-base.(

Table 3. Values of the risk coupling under various coupling forms.

Factors	Closeness		Betweenness	Factors	Closeness		Betweenness
	inCloseness	outCloseness			inCloseness	outCloseness
H1	11.396	70.175	33.484	E6	11.494	62.500	14.541
H2	11.869	37.736	44.822	E7	13.423	2.439	0
H3	11.765	62.500	65.906	E8	12.739	2.439	0
H4	11.236	93.023	19.086	M1	11.364	61.538	40.484
H5	11.765	42.105	4.667	M2	11.730	55.556	35.371
H6	11.429	88.889	50.689	M3	11.331	34.783	0.25
H7	11.834	67.797	107.839	M4	11.268	95.238	9.732
H8	11.527	57.143	16.341	M5	11.799	63.492	36.543
H9	11.461	48.193	0.702	M6	11.332	83.333	19.408
H10	11.730	64.516	75.833	M7	11.494	75.472	33.571
P1	11.628	36.364	11.819	M8	11.527	39.604	5.127
P2	11.662	46.512	43.757	M9	11.662	55.556	11.516
P3	11.461	47.059	14.794	M10	13.245	2.439	0
P4	11.111	39.604	0.633	M11	11.331	93.023	25.075
P5	11.696	54.795	74.172	M12	11.494	88.889	28.431
P6	11.561	97.561	46.947	M13	11.299	80.000	18.586
E1	11.594	95.238	45.789	M14	11.594	76.923	65.288
E2	13.201	2.439	0	M15	11.594	42.105	32.514
E3	11.696	28.169	8.979	M16	13.158	2.439	0
E4	13.333	2.439	0	M17	11.834	52.632	17.305
E5	13.468	2.4398	0

) Since the gas-explosion risk-system network in this paper is a directed network, Closeness includes inCloseness and outCloseness. As can be observed, inCloseness ranked the top 5 risk nodes in terms of degree of influence: electric sparks (E5), friction collision (E7), smoking open flame (E4), failure to take safety supervision seriously (M10), and large-scale roof plate collapse (E2). As can be observed, the top-five ranked risk factors in Closeness are primarily unsafe factors in the coal mine’s own environment, which are the direct causes of accidents; the top-five ranked risk nodes in OutCloseness are complex geological structure (E1), equipment management not being in place (M4), unauthorized command (H4), abnormal operation of the safety monitoring and control system (P6), and the lack of safety training and education (M11), of which complex geological structure (E1) and equipment management not being in place (M4) are of note. Unauthorized command (H4) and inadequate safety training and education (M11) have the same value as inadequate management of geological structure (E1) and equipment (M4). These risk factors originate from the four dimensions, suggesting that risk control is a result of all aspects and that accident prevention and control are inextricably linked to each link of control. The top five risk nodes, according to Betweenness, are illegal production organization (H3), failure to implement adequate safety measures (H7), and illegal production organization (H3). The majority of these factors—daily false gas reports (H10), confusion in the ventilation system (P5), and inadequate emergency response (M14)—are caused by personnel factors, which are thought to be the most crucial link in determining how the risk network spreads. Equipment and management factors are also significant mediators that can strengthen the control of this risk in order to prevent accidents.

3.2.2. Critical Risk Path Accessibility Analysis

Using Ucinet 6.0, the mediated centrality of edges is calculated on the evolution path of the network for gas-explosion risk. The top-ranked paths are displayed in

Table 4. Risk system critical path for gas explosion.

Pathway	Betweenness	Pathway	Betweenness	Pathway	Betweenness
P5–H2	71.326	P2–M3	21.730	M5–P5	16.191
H7–P2	65.809	M7–M2	21.639	M6–H7	15.606
H2–E3	41.805	M14–M17	20.990	M11–M9	15.372
M1–P4	37.226	P6–H3	19.274	M15–P1	14.819
H10–M15	34.428	P2–P1	19.257	M13–H8	14.563
H1–P5	28.122	E1–H3	18.684	M15–M3	14.186
H3–M15	28.208	P6–H10	17.467	M14–H8	13.567
M2–M8	26.983	E1–H10	16.877	H6–H7	13.441

, and the top five evolution paths are as follows: In social network analysis, the mediated centrality of edges is used to indicate the criticality of the propagation path of risk factors in the network, which can also reveal its vulnerability. The ventilation system is unclear. The ranking is as follows: violation of operating procedure—H2 (71.326) and P5 (ineffective safety measures); H7— poor equipment management and P2 (65.809)—operating procedure infraction; H2—gas outflow and E3 (41.805)—indicates a lack of management and examination regarding concealed threats; M1—fake daily gas reports and electrical equipment failure—P4 (37.226); H10—poor organizational design and M15 (34.428). In conjunction with the aforementioned study, it is possible to stop the risk’s evolution by managing the key risk factors because they have a positive correlation with the key risk-evolution path.

4. Model Modification and Discussion

4.1. Combining the N-K and SNA Models to Rectify the Results

The risk-coupling value T (Table 3) is used as the correction coefficient of the risk nodes to correct the risk-node proximity centrality out-degree because the SNA model is more subjective in quantifying the interrelationships of risk factors, whereas the N-K model is the risk coupling value calculated from the data of the accidents that have already occurred, which has stronger objectivity. The results are displayed in Figure 6, Figure 7 and Figure 8.

The top five risk indicators did not change in order when comparing the node degree centrality before and after rectification. Following correction, the top five risk factors were illegal command (H4), complex geological structure (E1), insufficient safety training and education (M11), smoking open flame (M4), and abnormal operation of the safety monitoring and surveillance system (P6). Prior to the adjustment, environmental and equipment factors were better at linking risks. Following the correction, management factors gained the greatest traction, demonstrating their impact on the whole coal-mine gas-explosion risk network.

It is evident from the centrality degree following coupling correction that management risk factors result in a larger risk coupling; nonetheless, the top five risk factors still include inadequate equipment management. This result is essentially the same as the findings prior to correction, demonstrating that the use of objective data after correction aligns closely with the results derived from the SNA model. M4 (safety monitoring and control system operation is not normal), P6 (safety training and education is not in place), and M11 were identified. Furthermore, the system for production safety responsibilities is not in place, and the unauthorized command H4 is central. After the correction, M12 is ranked in the top five as well, indicating that supervisory departments and relevant enterprises should concentrate on raising the overall level of personnel safety production and strengthening the system of responsibility for production safety, safety skills, and safety training and education.

The intermediate level of gas-explosion risk factors has been modified, but the top five risk factors in that order remain the same: ineffective safety measures (H7), fabricated daily gas reports (H10), a disorganized ventilation system (P5), unlawful production organization (H3), and insufficient emergency response (M14). This suggests that across the entire risk system, human considerations have always been the most important component. Human behavior can change the risk status of the entire risk network, whether it is due to management mistakes, harmful ambient conditions, or unsafe equipment conditions. Therefore, the main goals for risk management in coal mines are to reinforce managers’ supervision duties, increase their safety awareness, and regularly provide safety training to employees.

4.2. Strategies and Ideas for Risk Prevention

As indicated in

Table 5. Risk prevention and control countermeasures.

Grouping	Factors	Countermeasures
critical risk factors	M4	Enhance the equipment archive management system and fortify the construction of equipment management systems.
	P6	Bolster monitoring and control system administration and maintenance, and enhance monitoring and control system design.
	M11	To improve staff safety knowledge and skills, and increase safety education and training.
	H4	Boost safety education and training, and when working, adhere to all laws and guidelines.
	M12	Boost the organization’s safety production responsibility framework and make clear what each level of staff is responsible for in terms of safety production management.
critical risk paths	P5-H2	Optimize ventilation shaft arrangement and tunnel design.
	H7-P2	Boost worker safety awareness and strengthen the coal mine safety risk assessment system.
	H2-E3	Boost safety education and training, and when working, adhere to all laws and guidelines.
	M1-P4	Identify and resolve any possible risks by conducting routine safety inspections of the mine.
	H10-M15	Create a structure of accountability and encourage the use of sophisticated monitoring tools.

, risk prevention and control methods are developed with the goal of removing important risk variables and obstructing important risk linkages. Parameters like risk factor and risk-node degree centrality, intermediate centrality, and proximity centrality are used to assess how well risk prevention and control are working [35].

4.3. Future Research Directions

This paper’s data sources span the years 2011 through 2023. The accuracy of findings should be improved by future research by increasing the sample size and data scale. Developing methods from a single-entity perspective has also been the main emphasis of the research that has been done on accident prevention and control. For the risk management of coal-mine gas explosions, more research into the collaborative governance approach is also required.

5. Conclusions

By merging the N-K model and SNA, a collection of accident case sentences, and the coupling of risk variables for gas explosion accidents, this study builds the coupling mechanism and evolution mechanism of gas-explosion risk in coal mines within social networks. The study yields the following results.

(1)

The ranking of factor coupling degrees, which is T4 > T32 > T34 > T3(3) > T31 > T22 > T21 > T2(3) > T26 > T24 > T25, was obtained by applying the N-K coupling effect metric model to the study of coal-mine gas-explosion risk evolution. This suggests that as coupling factors rise, so do their hazards. Preventing multi-risk factor coupling is an essential way to reduce accidents. The parameters “human-equipment-environment-management”, “human-equipment-management”, and “human-environment” are more closely associated with the risk of gas explosions among the various heterogeneous multi-risk couplings. The incidence of accidents is directly linked to the danger to personnel, and prevention and control should be the main priorities.

(2)

The SNA model’s computation results indicate that the mediator centrality of the entire risk network ranked high for the following: ineffective safety measures (H7), unlawful production organization (H3), falsification of daily gas reports (H10), ventilation system confusion (P5), and insufficient emergency response (M14). For the critical path reachability study, critical risk factors have a positive correlation with the critical risk evolution path. This evidence suggests that managing critical risk factors can successfully prevent accidents by blocking the evolution of the risk propagation path.

(3)

The results of the fusion of the N-K and SNA models on the rectification of risk proximity centrality indicate that there is no equipment management in place. M4: Safety monitoring and control systems typically do not operate as expected. P6: No safety instruction or training is available. There is no system for production safety responsibilities in place—M11; there is unauthorized command—H4. M12 and the other five factors are the main components of the risk system for gas explosion. The management factor is the risk system’s weak link, which makes it easy to create a multi-factor risk coupling. By enhancing employee safety education and training, raising employee active safety awareness, bolstering the stability of the monitoring and control system, and combining the safety production responsibility system, the system risk can be effectively avoided and the coal-mine gas-power system safely improved.