Analysis of characteristics and causes of gas explosion accidents: a historical review of coal mine accidents in China

Objectives This study aimed to provide greater insight into the characteristics of severe and extraordinarily severe gas explosion accidents (SESGEAs). Methods. The study analyzed the accident characteristics and causes of SESGEAs. As an example, we conducted a specialized case analysis using the 24Model (fourth edition) on the recent Baoma coal mine gas explosion. Results. SESGEA data are characterized by greater volatility, with significant differences in the geographical distribution, temporal distribution and attributed characteristics of the accidents. From the accident analysis: chaotic ventilation management was the most serious accident cause of SESGEAs; unsafe acts related to ventilation operations accounted for 18.51% of all unsafe acts; coal miners lack professional safety knowledge and have a serious fluke mentality in mining work; enterprises have insufficient enforcement of safety procedure documents, and lack of attention to the allocation of underground human resources and safety training systems; and the importance of safety, the role of the safety department and satisfaction with safety facilities have become the most serious missing items of safety culture. Conclusion. This study can provide important data support and management basis to assist mine operators in developing more targeted accident prevention strategies.


Introduction
As the most important industrial fuel and basic energy source in China, coal accounts for more than 60% of China's energy consumption [1].According to the current energy consumption structure in China, the dominant position of coal in the primary energy consumption structure will not change quickly [2].However, the death toll from coal mine accidents in China accounts for more than 70% of global coal mine deaths [3].Therefore, the safe production of coal enterprises has become a premise and required guarantee for the sustainable development of China's coal industry.In the past 10 years (from 2010 to 2020), the fatality rate per million tons in China's coal mines ranged from 0.059 to 0.749, and its rate of change was −0.063, showing an overall downward trend.Compared with the previous decade (from 2000 to 2009), the rate of change was −0.5353.Although the decline in the fatality rate per million tons gradually slowed and stabilized, it also greatly narrowed the gap with other major coal-producing countries worldwide.For example, in 2019 the fatality rate per million in China was 4.88 times that of the USA in the same period, whereas the figure in 2004 was 110.97 times.
However, despite the significant decline in the numbers of coal mine accidents and casualties, it remains unacceptable that the production accidents caused by mining, especially severe and extraordinarily severe accidents (SESAs) with massive social effects, have not been eliminated.The failure to effectively control the occurrence of accidents has also become a key factor restricting the development of coal enterprises [4].Of all coal mine accident types, gas explosion accidents are considered to be most fatal coal mine accidents [5,6].Figure 1 shows the statistics for 556 SESAs in China from 2000 to 2021; among these, 285 accidents were caused by gas explosions, accounting for 51.12% of the total.The death toll CONTACT Gui Fu fugui66@126.comwas 6378, accounting for 56.74%.For example, on December 3, 2016, 32 miners were killed in a gas explosion at Baoma Mining in Chifeng, Inner Mongolia.Therefore, behind these heartbreaking figures, the effective control of gas explosion accidents is crucial to the improvement of the safety regarding production in China's coal mines.While there have been numerous research efforts focused on the analysis and prevention of coal mine accidents in general [5,[7][8][9][10][11][12], relatively few have specifically examined the characteristics and causes of SESAs, especially severe and extraordinarily severe gas explosion accidents (SESGEAs) in coal mines.This is an important area for further study in order to improve mine safety.Therefore, it is necessary to systematically analyze the causes of gas explosion accidents and find the accident occurrence pattern to provide mine safety personnel with the necessary information for accident prevention.
From the current research status, the research on gas explosion accident prevention and control is still more concerned with the application of engineering technology tools, mainly in four aspects: • Explosion suppression techniques [13]: the use of inert gas, water mist, dry powder, halogen compounds and other explosive suppressant materials can help prevent and control gas explosions.These methods can effectively slow down and reduce the speed of flame propagation and explosion overpressure.• Explosion resistance techniques [14]: aim to halt the spread of flame in the event of an explosion.Porous materials are commonly used in the development of explosion-resistant materials for this purpose, as they can effectively prevent the backward propagation of the blast flame.• Explosion relief techniques [15]: an explosion relief door is the most common explosion relief device in daily production and life, in addition to the explosion relief film, explosion-proof cover, explosion-proof valve and other explosion relief devices.• Explosion isolation techniques [16]: isolation to stop the propagation of the explosion, and even the method of extinguishing the explosion flame.
The results of the aforementioned review also indicate that the preventive measures taken by the current research to prevent accidents are more concerned with changing the states of the machine or environment, such as by installing protective devices for mechanical equipment or adding complicated safety regulations.In addition, the unsafe acts of miners have received increasing attention in the analysis of accident causes [17].Human reliability has brought considerable uncertainty to the complex coal mine safety system [4].Among all accident causes, human factors are the direct cause of accidents [18], and 95.10% of accidents in Chinese coal mines are caused by human errors [12].Yin et al. [19] statistically analyzed the unsafe acts in gas explosion accidents in China during 2000-2014, and all identified unsafe acts could be analyzed in three dimensions: accident site; process and equipment; and installation.Patterson and Shappell [20] analyzed the causes of 508 coal mine accidents based on HFACS (Human Factors Analysis and Classification System) and found that skill-based errors were the most serious unsafe acts.Yuxin et al. [21] identified three critical unsafe acts and six critical behavior paths in gas explosion accidents by modeling and simulating the interrelationships of unsafe acts in gas explosion accidents.While past research on accident prevention has largely concentrated on certain causes of accidents, often referred to as 'unsafe conditions' and 'unsafe acts' in accident causation models, it has not adequately examined the organizational factors contributing to gas explosion accidents.This lack of a systematic framework for accident analysis can result in incomplete and insufficient findings.However, the failure of a coal mine defense system cannot only be attributed to the direct factors that induce accidents (unsafe acts and unsafe conditions).A positive organizational culture and a sound management system have been clearly demonstrated to be effective in intervening in miners' behavior [22].Accident causation models have been widely used in accident investigation and analysis due to their ability to analyze the causes of accidents in a comprehensive manner and the strong causal logic among the causes [23].In this approach, failure modes and potential states of hazards are often adopted to classify all causes of accidents [24], and the individual and organizational factors that contribute to the occurrence of accidents can be comprehensively considered.Therefore, it is a more scientific approach to apply the accident causation model to systematically analyze gas explosion accidents [25].
Accidents will occur repeatedly at the same location when the underlying factors that cause the accidents are not thoroughly solved.The aim of accident analysis is to identify the fatal factors and common characteristics that exist in many accidents and to prevent the recurrence of the similar accidents [26].SESGEAs in coal mines with serious social effects have similar causes and characteristics.This study statistically analyzes 588 cases of SESGEAs and studies the macroscopic patterns and characteristics of gas explosion accidents.In addition, the causes of gas explosion accidents are statistically analyzed using the 24Model (fourth edition), and the case study of the Baoma coal mine gas explosion accident is used as an example.This study hopes to help more people understand the characteristics of SESGEAs and the various causes of coal mine system failure, and the relevant research findings can provide important references for accident prevention and enhanced safety training.

Accident case
The research data were acquired from the State Administration of Work Safety (SAWS) and State Administration of Coal Mine Safety (SACMS).Among them, the records of SES-GEAs in China from 1950 to 2020 were obtained from official accident case analysis reports [27][28][29].The SESGEAs mentioned here refer to severe accidents resulting in 10-29 fatalities and extraordinarily severe accidents resulting in more than 30 fatalities, reaching a total of 588.The data here also address the coal mine ownership, mine scale and gas grade, which are defined as coal mine attribute factors.Each accident case is analyzed, and accident causes are extracted from it.The mine ownership is divided into national strategic mines (NSM), government-supported local mines (GSLM) and township mines (TM); the mine scale is divided into small mines (SM), medium mines (MM) and large mines (LM); and the gas grade is divided into slightly gassy mines (SGM), highly gassy and outburst-prone mines (HGOPM) and unclassified mines (UCM).

Accident analysis model
The accident causation model is a fundamental theory in safety science that provides a framework for understanding and analyzing the causes of accidents, their progression and the factors that contribute to them [30].These models are used to trace the causes of accidents and develop strategies for preventing them.The 24Model is a comprehensive accident causation model developed by Fu in 2005 [31].It builds upon the principles of domino theory, the Swiss cheese model and the loss causation model, and has been continuously refined and improved over the past 17 years [31].The 24Model has reached its sixth version, which is more suitable for accident causation analysis and accident prevention in enterprises.In the 24Model, among the internal organizational factors that lead to accidents, all accident causes can be divided into two levels and four stages, of which the two levels refer to two categories of organizational factors and individual factors, and the four stages refer to the division of accident causes into four categories (individual act, individual capability, safety management system [SMS], safety culture) according to the causal logic and organizational relationships., as presented in Figures 2 and 3.
The 24Model offers several advantages for analyzing accidents: it clearly classifies and defines all types of causes; the causes of accidents are modular and have a clear logical structure; it allows for the creation of a complete accident path starting from the initial unsafe act; and it focuses on causes within the organization, providing a defined scope for accident prevention and management.Many scholars have used the 24Model to conduct valuable research on the causes of coal mine accidents in China and provide targeted prevention recommendations [25,[33][34][35][36][37][38].In this study, we used the 24Model (fourth edition) to perform a detailed analysis of the causes of gas explosion accidents.

Spatial and temporal distribution of accidents
Complex geological conditions have caused an uneven regional distribution of gas explosion accidents in China.In one recent study, it was found that the southern provinces, especially Guizhou, Sichuan, Chongqing, Hunan and Yunnan, contributed to 13.5% of the coal production in China, whereas they caused 55.3% of the gas explosion accidents [9].Seven provinces located in high-gas geological zones have become the hardest-hit areas with regards to gas 12 explosion accidents [39].As shown in Figure 4, provinces with more than 100 HGOPMs are mostly located south of the Yangtze River.Among them, the southwestern areas in China are dotted with more HGOPMs.These areas are mostly located in high-gas geological zones.These special geological conditions provide inducements for the occurrence of fatal gas explosion accidents in Chinese coal mines, leading to the high frequency of SES-GEAs in such areas.Shanxi Province has the highest number of SESGEAs, with 118 accidents and 3106 deaths.
As shown in Figure 5, most accidents occur in March, April, May and November, accounting for 45.06% of total accidents in a year; fewer accidents occur in February and June.SESGEAs in Chinese coal mines are more likely to occur 2 months before and after the Chinese New Year.As work resumption starts after the Chinese New Year, the lax mood after the holiday leads to a low awareness of worker safety, increasing the possibility of fatal accidents.Meanwhile, November is the month with the most frequent SESGEAs.As the winter arrives, northern China generally needs central heating to face the cold temperatures.During this period, coal is used as the main fuel supply, greatly increasing the pressure on the market supply and demand.Forced by the surge in market demand, coal mining enterprises need to increase the production intensity to meet the set target, and they encourage front-line workers to work excessively, e.g., by using a performance reward system.Worryingly, this incentive policy enables front-line workers to ignore hidden hazards in operation, increasing the possibility of fatal accidents.

Frequency distribution of accidents
The number of accidents, death toll and death rate per million tons (DRPMT) are crucial statistical indicators for the Chinese government to measure the status of safety production [40].As the randomness of the number of accidents varying with the time series is small, the number can represent a change in safety production in a certain area [41].In recent years, researchers [5,42] have discovered that the number of gas explosion accidents in Chinese coal mines also fluctuates over time.They mostly used 10 or 20 years as sample intervals to study the characteristics of Chinese coal mine gas explosion accidents.However, owing to the intensive time divisions, the sample size was too small and the results of the studies fluctuated greatly, which would be a disadvantage in conducting a macroscopic exploration of the changes in gas explosion accidents.Therefore, this study adopted the Mann -Kendall test to determine the time abrupt changes regarding the number of SESGEAs in Chinese coal mines from 1950 to 2020.
The trend in the number of accidents over the time variable is determined by comparing and analyzing the interrelationship of the statistical variables UF k and UB k .This is used to determine the point in time when accidents produce a sudden increase or a sudden decrease.UF denotes the number It can be seen from Figure 6 that the number of SESGEAs changed significantly in 1980, passing the significance test.The year 1980 is therefore considered to be the turning point in time when the number of SESGEAs changed significantly.The number of SESGEAs in coal mines has increased sharply after this time point.Considering that the market economy of China has been in a stage of accelerated development since the Chinese reform and opening-up in 1979, the vigorous development of the coal industry has led to the construction of a large number of TM accompanied by private enterprises [1].However, the low-entry barriers have also led to a remarkable increase in the number of fatal accidents.Notably, in this period, the SESGEAs were characterized by a marked improvement in production safety at NSM, whereas the TM became the hardest hit.
As shown in Figure 7, the temporal trend of the number of SESGEAs is divided into two stages, centering on the time node in 1980.The number of accidents in the two stages changed

Accident distribution of coal mine attribute factors
To explore whether the distributions of the number of accidents from the different coal mine attribute factors in the SES-GEAs are different, this study used a variance analysis to test the number of accidents in three groups of attribute variables (ownership, coal mine scale and gas grade), with three subvariables in each group of variables.The results are shown in Figure 8, which represents the mean accident counts for each type of coal mine.The image provides a central point of reference to compare the accident frequency among different mine types.The error bars in Figure 8 represent the standard deviation of the attribute variables of accidents in the distribution of the number of accidents, which is the measure of  the dispersion of each data point in the dataset relative to its mean.In other words, they depict the range within which the accident counts are distributed.
Different letters (a and b, c and d, e and f) were used to indicate differences in the distribution of the number of accidents at different mines.For better understanding, the coal mine scale group is denoted by letters a and b for quantitative differences, the gas grade group is denoted by letters c and d for quantitative differences and the ownership group is denoted by letters e and f for quantitative differences.This means that the comparisons here focus on within-group comparisons rather than between-group comparisons.The same letter in each group of variables indicates that there is no significant difference in the attribute distribution of the number of accidents, whereas different letters indicate the opposite.In the case of coal mine size, e.g., the mean value of the number of accidents in SM is much larger than the number of accidents in MM and the number of accidents in LM, as can be seen in Figure 8.Therefore, the number of accidents in SM is designated a, and the number of accidents in MM and the number of accidents in LM are assigned b.The letters a and b are used to reflect whether there is significant variability in the number of accidents in different types of coal mines.This result implies that the number of accidents in small coal mines accounts for the majority of the distribution of accidents in the coal mine scale, which provides a suggestion for the subsequent prevention strategy of the Chinese government's coal mine accidents.It is noteworthy to emphasize that the absence of a substantial disparity in accident frequencies among HGOPM, SCM and UCM within the gas grade group is denoted by the consistent employment of the letter 'c'.This standardized representation emphasizes the fact that there is no significant difference between the variables within the group in terms of accident occurrence.Consequently, the letter 'd' is notably omitted from this particular grouping.
There are significant differences in the number of SESGEAs in regards to mine scale and ownership.Among them, in the comparison of mine scales, it is found that there is no significant difference between MM and LM, and that the two values are close to the same.However, the number of accidents in SM is significantly different from, and much greater than, the other two.There was no significant difference with regards to gas grade in the accident number distribution (p = 0.196 > 0.05).
Meanwhile, in the comparison of ownership, the number of accidents in TM is much greater than in the other two.It was found that most of the SM belong to TM.The safety performance of township coal mines is far worse than that of NSM [43].The economic situation of coal mine owners and the remote location of township coal mines have resulted in poor safety infrastructure in small-scale mines and the fact of being outside the government's safety regulatory framework.In order to maximize short-term benefits, township coal mines engage in illegal production practices such as overcapacity and unlicensed mining.As can be seen from Figure 9, small and medium-sized mines, mainly township coal mines, are the focus of the impact on China's coal mine safety performance.

Unsafe acts
Coal miners, who always work in the front line of the mining workface, are at the 'sharp end of safety' and are potentially closest to the accident.When an accident occurs, injuries to workers are the most direct and fatal.Reducing the unsafe acts of front-line workers has become a key measure for preventing gas explosions in coal mines [44].Figure 10 shows the 10 unsafe acts with the highest frequency among the attribute factors.In view of the differences in the rankings of unsafe acts among the different attribute factors, there are 15 common unsafe acts.A1-A15 represent the 15 common unsafe acts with the highest frequencies among the attribute factors: A1 is 'chaotic ventilation management', A2 is 'illegal use or uninstalled ventilation', A3 is 'not checking gas concentration', A4 is 'inadequate safety education', A5 is 'mining under dangerous conditions', A6 is 'special operation personnel without a certificate', A7 is 'unattended electromechanical equipment', A8 is 'unlicensed or illegal mining', A9 is 'non-compliance with safety supervision orders', A10 is 'concealment and missed inspection', A11 is 'no hazards screening and identification', A12 is 'inadequate staffing', A13 is 'not checking the gas three times before explosion', A14 is 'mining design not in accordance with reality' and A15 is 'not to evacuate the workers in time in case of power and wind failure'.
It can be seen from Figure 10 that 'chaotic ventilation management' is one of the most serious unsafe acts, including acts such as wrong placement of ventilators, the turning on and off the ventilator at will and series ventilation.The unsafe acts related to ventilation operation accounted for 18.51% of all unsafe acts.Compared with NSM, which are more standardized in terms of professional quality and allocations of personnel, GSLM and TM are often criticized for their allocations of staff capacity and human resources.This defect is also reflected in two high-frequency unsafe actions: 'special operation personnel without a certificate' and 'inadequate staffing'.In addition, the inadequate allocation of human resources means that front-line workers usually perform multiple jobs under harsh working conditions.Meanwhile, 'unlicensed or illegal mining' and 'non-compliance with safety supervision orders' have become high-frequency unsafe acts in TM and small and medium-sized mines.
In the analysis of accident causes, it was found that the percentage of unsafe acts from front-line miners and management was 50.4 and 49.6%, respectively.Compared with the unsafe acts of front-line miners directly causing accidents, the unsafe acts of management, such as faulty mining methods or unlicensed or illegal mining, may be more insidious and destructive than those of front-line workers.However, finding their own mistakes is often a human weakness, and the management of coal mine enterprises usually attribute accidents to unprofessional operation behaviors and the low safety consciousness of front-line staff, while seldom investigating their own problems regarding safety supervision and decision-making.The long-term lack of effective safety management mechanisms in the coal industry and the lax safety supervision attitudes of management led to loopholes in the safety supervision of some coal mining enterprises.

Miners' safety capabilities
There is a significant correlation between people's cognitive levels and behavioral activities.In general, knowledge can change an individual's awareness toward things, upon which differences in behavior can be identified.A person's cognitive level, personality characteristics, mental state and degree of dedication to work will all affect their own behavior [45].The 24Model attributes the lack of safety knowledge, awareness, habit, physiological status and psychological status to the miners' lack of safety abilities, which is also considered to be an indirect cause of the accident (Figure 11).
For the purpose of maximizing the interests of coal mine employers, the managers involved in accidents are mostly temporary employees and relatives with no professional background.The serious lack of professional knowledge makes it possible to plant safety hazards at the beginning of the mining design.The difference in mastering safety knowledge is reflected in the lack of professional knowledge and excessive or insufficient related experience.Miners having accumulated a large amount of experience would be initially deemed as a positive factor in improving the accuracy of making judgments.However, out of laziness and blind selfconfidence, employees use their experience to perform activities such as illegal operation of a coal drill, random movement of cables, etc.In addition, not checking the gas three times before explosion and maintenance of electrical equipment under the energized conditions are further examples demonstrating the lack of professional safety knowledge of miners.Further, due to laziness, workers can often adopt the mentality of saving energy, and psychological shortcuts make miners rely on past experiences stored in their memories, simplifying the prescribed operating procedures.However, repeat occurrences of illegal activity have gone unpunished, which has allowed workers to continue conducting said illegal activity and turn them into habitual violations.Habitual violations by miners account for 90% of all human errors in accidents [46].

The safety management system
The SMS is the operating system for risk management and control, and is a collection of safety-related policies, practices, procedures, roles and functions [47].A sound SMS is considered to be an important constraint on miners' unsafe behavior, and the effectiveness of this constraint is mainly reflected in management procedures, organizational structure and safety policies [48].Deficiencies in safety management regulations, rules or operating procedures have been found to be the most common loopholes in management system in coal mine accidents.Compared with employees' unsafe acts that lead to the occurrence of accidents, this pattern of system failure tends to affect management decisions, organizational culture and construction design for a long time [37].The failure modes of SMSs in coal mine accidents were categorized and analyzed, the results of which are presented in Table 1.
From the presented statistical results, an observation can be made that the support and operation procedure (9008) is the most prominent problem in the SMS of gas explosion accidents.The items with a higher frequency in the support category further include the allocation of underground human resources, safety training and education of miners, dangerous mining and gas inspection management.As revealed in the accident analysis, human resource allocation and the safety education of miners are highly frequent, with management decisions rendering such problems particularly prominent.These problems have been found in a large number of coal mine accidents in China, such as illegal production, incorrectly opening closed areas, unauthorized resumption of work during shutdown, etc.
In the planning process, the inspection of gas and ventilation conditions is undoubtedly one of the most frequent items, consisting of items such as no gas inspection records, no ventilation equipment inspection procedures, lack of ventilation system air volume verification procedures, etc.As aforementioned, chaotic ventilation management is the biggest challenge faced by small coal mines and township coal mines in China, with a lack of appropriate program documents being the main reason for gas accumulation in coal mines.This can be illustrated by the severe gas explosion that occurred in the Dashan Coal Mine in Leping Township on May 10, 2013.Under the guise of technical reform, the mine crossed the mining boundary for illegal mining.At the same time, a chaotic ventilation system and serious shortages of air volume kept workers in a windless or low-wind working environment for a long period of time, resulting in gas accumulation at the Masi Lane operation site.In the end, 12 people were killed by a gas explosion caused by sparks from a coal drill.

Safety culture
Safety culture was derived from the accident investigation report of the International Atomic Energy Organization on the Chernobyl accident [49].Since then, safety culture has begun to be applied to investigate the underlying causes of major accidents [50].In the 24Model, safety culture, as the guiding ideology of organizational safety work, affects the behavior of organizational members by influencing the organizational SMS.Ultimately, safety culture affects the generation of actions, and has therefore also become the root cause of accidents.Fu et al. [51] quantified safety culture into 32 elements, and the missing cultural elements were judged by the unsafe acts in the accident.In the present study, the missing items of safety culture in accidents were analyzed, and the analysis results are shown in Figure 12.
Regarding the classification of safety culture, the frequencies of the items from high to low are safety commitment (30.71%), awareness of safety responsibility (19.67%), awareness of safety practice activities (15.27%), role of safety agencies (13.95%), quality of SMS (11.95%) and safety participation degree (8.44%).An observation can be made that items having a significant impact on the occurrence of accidents included the safety commitment of the management of coal mining enterprises, the safety responsibilities at all levels and the implementation of safety work.A positive safety  commitment is the key to building a corporate safety culture [52].Coal mine enterprises provide safety belief support to employees through the establishment of safety commitment by the management, which is followed by the establishment of an SMS that complies with laws and regulations.Finally, the miners perform safety work.In terms of the elements of safety culture, the importance of safety (1274), the role of the safety department (813) and the satisfaction with safety facilities (801) have become the most serious missing items of enterprise safety culture in SESGEAs.

Case study
In this section, we use the 24Model to thoroughly analyze one accident case, identifying all contributing factors and creating a clear causal chain for a more intuitive understanding of the accident's development.This helps researchers to better understand the underlying issues and prevent future accidents.

Accident process
On December 3, 2016, the Baoma Mining Co. in Chifeng City, Inner Mongolia suffered a catastrophic gas explosion which left 32 people dead and 20 injured, and resulted in direct economic damage of 43.99 million yuan. Figure 13 presents a timeline of the accident.The process of accident analysis using the 24Model is as follows.

Accident cause analysis 3.3.2.1. Individual-level cause analysis.
In the 24Model, individual-level causes are divided into individual act causes (direct causes) and individual capability causes (indirect causes).Among them, individual acts can be divided into unsafe acts and unsafe physical conditions, and individual Note: SMS = safety management system; SESGEAs = severe and extraordinarily severe gas explosion accidents.
capability includes safety physiology, safety psychology, safety awareness, safety knowledge and safety habits.Due to the inability to directly assess the physiological state of the miners involved in the accident, factors related to physiological defects were not included in the analysis of the safety incidents.Table 2 presents an individual-level analysis of the causes of the gas explosion at the Baoma coal mine.
According to the accident investigation report, a total of 13 unsafe actions were identified as contributing to the accident.As presented in Table 2, illegal mining was found to be the most significant issue, encompassing behaviors such as crosslayer and transboundary mining, false closed doors, multiple working faces, forged mine production drawings and evasion of safety regulations.These unsafe acts reveal a disregard for legal regulations and a prioritization of economic efficiency over safety production by the Baoma coal industry.There are also long-term safety hazards regarding ventilation in this coal mine.Adequate airflow is crucial for providing underground miners with the necessary oxygen.The proper configuration of air paths is essential for maintaining a safe working environment and preventing the entry of dirty airflow into the work area.In this accident, several unsafe acts related to ventilation were identified, such as the use of a single ventilation fan to supply air to two different workfaces and the illegal release of high concentrations of gas into the work area.In particular, the decision to connect the air paths of two workfaces, which is prohibited by law, was a flaw in the initial project design and contributed to the accident.Additionally, as the 'trigger' in this accident, the miners' welding violation was more like an ordinary act of daily violation.Therefore, the accident was an inevitable result.According to Table 2, it can be seen that the miners in this accident had low safety awareness, which manifested in two ways: their inability to detect hazards in a timely manner and their failure to promptly address them.This was exemplified by their failure to check the gas concentration, unplanned power delivery and the production of fake mine drawings.Additionally, the miners demonstrated a lack of safety knowledge in areas such as the operation of special equipment, gas inspection, safety regulations, welding operation and ventilation design.In terms of safety psychology, poor attitudes toward safety were demonstrated through a tendency toward risky behavior and a reliance on luck.This was evident in the case of the miner who was welding the hydraulic support and disregarded the usual gas concentration checks due to a belief in luck.This also highlights that many of the unsafe actions in the accident were the result of habitual rule violations.

Organizational-level cause analysis.
The organizational causes of the 24Model accident can be divided into two categories: shortcomings in the management system, which is considered the radical cause of the accident; and inadequate development of safety culture, which has been identified as the root cause of the accident.Table 3 highlights the safety issues with the organizational management system revealed by this accident.
As presented in Table 3, Baoma Mining Group has disregarded national laws and industry regulations in its operations, as demonstrated by its cross-layer and transboundary mining and failure to seal blind alleys as required.The company has a lack of law consciousness and has not implemented a safety policy of 'safety first, prevention first, comprehensive management'.In addition, the coal mine's safety management procedures are not comprehensive and are not being properly implemented.This is evident in the lack of gas checking records, inadequate safety training systems, lack of airflow verification procedures and unplanned power delivery after power failure.While it is unclear whether violations such as the use of welding by miners are due to carelessness or the lack of strict implementation of management procedures, it is clear that Baoma Mining Group needs to improve skills training and safety education for miners.Note: HB1 = safety awareness; HB2 = safety habits; HB3 = safety knowledge; HB4 = safety psychology.
The safety culture of an organization is made up of its safety values and beliefs, and the level of understanding and adherence to these values by its employees plays a crucial role in preventing accidents.The 32 elements of safety culture shown in Figure 12 were used to assess the safety culture of Baoma Mining Group, with the results presented in Table 4.
Baoma Mining Group's deficiencies in safety culture were highlighted in the accident through 13 safety culture elements.These elements generally revealed weaknesses in six areas: the quality of the SMS, awareness of safety practice activities, awareness of safety responsibility, safety commitment, safety participation degree and role of safety agencies.The company's poor safety culture often leads to the following consequences: • Enterprise managers have a weak safety legal concept and safety awareness.• Enterprise benefits outweigh miner life safety.
• Inadequate and poorly implemented safety management procedure documents.• Miners lack adequate safety knowledge and emergency escape ability, and there are numerous violations in their operations.

Accident path analysis
Based on the aforementioned analysis, the interrelationship of the various causes of this accident is constructed and the path of occurrence is formed according to the logical framework of the 24Model, as shown in Figure 14.
To aid in accident prevention and clearly illustrate the path leading to an accident, this article breaks down the accident path into three levels: triggering electrical sparks; causing gas accumulation; and management violations.Only easily identifiable safety knowledge and habitual violations are discussed with regards to safety capability, and safety culture is not addressed in this section.
Path A is the causal path of the electrical worker's welding violation that directly triggered the accident, as shown in Figure 15.After in-depth analysis, it was found that there were no special operating procedures to guide workers in safe operation before this welding process.Such a phenomenon revealed that Baoma Mining Group did not strictly implement the management system for special operators and the management system for the operation of mining equipment and devices.These procedure documents should have been developed and supervised by the mining captain, deputy production manager and safety mine manager.In addition, the lack of Note: SMS = safety management system; SMS1 = safety policy; SMS2 = structural establishment; SMS3 = personnel allotment; SMS4 = duty assignment; SMS5 = planning; SMS6 = support and operation; SMS7 = improvement; SMS8 = performance evaluation.
miners' knowledge of safety operations highlights the inadequacy of the Baoma Mining Group's safety education and skills training system for miners.
Path B is the gas accumulation that led to the accident, as shown in Figure 16, that can be traced back to the design of the tandem air path, which connected the air paths of multiple working faces.When ventilation was restored, the strong air flow pushed the high concentration of accumulated gas into the 6040 working face.Additionally, the illegal discharge of gas Note: SC1 = quality of safety management system; SC2 = awareness of safety practice activities; SC3 = awareness of safety practice activities; SC4 = safety commitment; SC5 = safety participation degree; SC6 = role of safety agencies.
and not carrying a gas detector also contributed to the buildup of gas concentration.Even without a subsequent explosion due to welding violations, the combination of these factors led to an accumulation of gas at the explosion limit, creating the risk of asphyxiation accidents.Path C is the direct causal path of management violations that led to the accident, as shown in Figure 17.The illegal mining practices of Baoma Mining Group created a significant safety hazard that ultimately caused the accident.Specifically, the group engaged in cross-layer and transboundary mining of adjacent mines with clearly defined mineral rights boundaries, and fabricated mining drawings to avoid supervision.Despite receiving an order to halt production from the supervision department, the company continued to secretly mine using the tunnel mining method, exposing miners to dangerous conditions for an extended period of time.The mine production technician, deputy production manager and safety manager were responsible for developing a comprehensive mining plan and design, but in this case there was a lack of geological drilling information and power supply system drawings    in the cross-border area, significantly increasing the risk of coal mining.
This section explored the specific causes of accidents and accident paths, using the Baoma coal mine gas explosion accident as an example, to provide a more comprehensive understanding of the process of accident development.The three accident cause paths presented are also control paths for accident prevention, and cutting off any one of them can serve the purpose of preventing accidents from occurring.

Discussion
Through the presented study, it is found that unsafe acts are the most direct cause of accidents, while unsafe acts by management are the deeper hidden and more destructive behaviors.In addition, the lack of adequate safety training for miners is also an influential cause of accidents.Both of these problems are prevalent in most gas explosion accidents and are urgent issues for coal mine accident prevention in China.This study discusses these two problems and proposes targeted safety recommendations in the hope that more researchers will focus on coal mine accident prevention.

How to reduce human error of management
Mining accidents are often caused by the unsafe actions of miners, and addressing these unsafe acts is critical in preventing accidents [21].However, recent data have shown that the root causes of accidents are increasingly shifting from the unsafe acts of front-line workers to those of management.Management has significant influence on the selection of production processes, safety investments and design solutions [25], and their behavior can contribute to accidents through violations of laws and regulations, inadequate implementation of safety measures and risky decision-making.Additionally, the manager, as mediator between management and constraint of workers' behavior, plays a key role in shaping the safety behaviors and attitudes of workers [20].
When management ignores safety regulations, this can demotivate front-line workers from practicing safe behaviors and make the reward and punishment system ineffective [53].
To correct this, it is necessary, first, to increase the emphasis on safety within the company, so that overall safety levels can be improved -when miners feel that their leaders prioritize their safety and are committed to ensuring it, they are more likely to adopt a positive attitude toward safety [44].Second, it is necessary to raise awareness of legal issues -illegal practices such as unlicensed mining and cross-border mining reveal a lack of legal understanding within management.Only when employers have a clear legal mindset will they consciously take responsibility for safety and proactively increase their safety investments [1].

How to strengthen the safety training of miners
For safe production, it is not sufficient to simply have good safety awareness.Without the proper safety knowledge and operational skills, accidents are likely to occur.Safety education is an essential tool to raise miners' safety awareness and help them acquire safety knowledge and skills.Safety training should be enhanced from the following aspects.First, it is necessary to introduce more specific case training programs to improve miners' understanding of the accident process in conjunction with the accident causation model.Second, since miners' ability to identify hazards and perceive safety risks is largely determined by safety training [54], coal mining companies should focus on strengthening safety education and skills training in specialized areas such as legal regulation, hazard identification and emergency response.Third, it is necessary to develop annual training plans and increase investment in safety education, construct specific personnel training files and conduct regular assessments.Fourth, while increasing the bias for safety education for young, inexperienced miners, there should also be a sustained increase in safety education for experienced miners.Experienced miners rely more heavily on work experience than younger miners and are far less likely to comply with rules and regulations than younger miners [44].Fifth, managers also need to enhance safety training to increase their awareness of safety boundaries and their sense of safety responsibility.

Coal mine gas explosion accident prevention suggestions
According to the results of this study, the government should pay more attention to the safety supervision of SM and TM.
Similarly, SGM also need to establish and improve gas monitoring and control systems.For coal mining enterprises, the unsafe actions of management are more worthy of attention, such as 'wrong mining methods' and 'unlicensed or illegal mining'.More importantly, it is important to establish a scientific enterprise SMS, and to form a good safety culture atmosphere for coal mining enterprises.When formulating measures to reduce the unsafe actions of front-line workers, enterprises should propose targeted measures according to the type of work.For example, in severe ventilation work types, it is necessary to reasonably install local ventilators, provide timely inspection of ventilation equipment and prohibit turning on and off the ventilator at will.For township coal mines with a shortage of personnel, the auxiliary office staff should be reduced, and human resources should be allocated to key jobs to meet the minimum requirements of staffing.

Conclusion
SESGEAs have serious social effects in China, and the large number of casualties and large economic and property losses have made them a key target of the Chinese government.As there is a scarcity of relevant research, analyzing the SES-GEAs in Chinese coal mines and exploring the characteristics and accident causes have become increasingly necessary.The specific conclusions of the present study are summarized as follows:.
• The data of SESGEAs are characterized by greater volatility, with significant differences in the geographical distribution, temporal distribution and attribute characteristics of the accidents.The year 1980 was a key time point for abrupt changes in the number of SESGEAs, and Shanxi and Heilongjiang are the most seriously affected provinces in SESGEAs.There is a high probability of SESGEAs in the two months before and after the Chinese New Year.The safety supervision department of the Chinese government should implement a strict safety inspection system in the aforementioned provinces and time.• Chaotic ventilation management has become the most serious unsafe act among various attribute factors, including wrong placement of ventilators, the turning on and off the ventilator at will and series ventilation.The unsafe acts related to ventilation operation accounted for 18.51% of all unsafe acts, and the percentage of unsafe acts from front-line miners and management was 50.4 and 49.6%, respectively.• Coal miners lack safety professional knowledge and have serious fluke mentality in mining work.The mining risks of miners are elevated by the lack of safety awareness and the adventurous psychology of the management in decisionmaking.• In the coal mine SMS, enterprises have insufficient enforcement of safety procedure documents, such as gas inspection management and dangerous mining, and lack of attention to the allocation of underground human resources and safety training systems.• The safety culture of coal mining enterprises is weak, and the safety commitment of management (30.71%),safety responsibilities at all levels (19.67%) and the implementation of safety work (15.27%) have a significant impact on the occurrence of accidents.In terms of the elements of safety culture, the importance of safety (1274), the role of the safety department (813) and the satisfaction with safety facilities (801) have become the most serious missing items of enterprise safety culture in SESGEAs.

Figure 1 .
Figure 1.Statistics for severe and extraordinarily severe accidents in Chinese coal mines from 2000 to 2021.Note: The full color version of this figure is available online.

Figure 3 .
Figure 3.The 24Model application process in accident analysis.

Figure 4 .
Figure 4. Number of severe and extraordinarily severe gas explosion accidents (SESGEAs) and fatalities in each province from 1950 to 2021.

Figure 5 .
Figure 5.Time distribution characteristics of SESGEAs.Note: SESGEAs = severe and extraordinarily severe gas explosion accidents.

Figure 6 .
Figure 6.Time test of the number of the SESGEAs.Note: SESGEAs = severe and extraordinarily severe gas explosion accidents; UF, UB = number of non-synonymous mutations in the foreground and background, respectively (for detailed explanations, refer to Section 3.1.2).

Figure 7 .
Figure 7. Variation trend of the number of SESGEAs.Note: Vertical dashed line indicates the point in time at which the accident surge occurred, identified by Figure 6.SESGEAs = severe and extraordinarily severe gas explosion accidents.

Figure 8 .
Figure 8. Analysis of variance for different attribute factors.Note: The coal mine scale group is denoted by letters a and b for quantitative differences, the gas grade group is denoted by letters c and d for quantitative differences and the ownership group is denoted by letters e and f for quantitative differences.The same letter in each group of variables indicates that there is no significant difference in the attribute distribution of the number of accidents, whereas different letters indicate the opposite.HGOPM = highly gassy and outburst-prone mines; LM = large mines; MM = medium mines; NSM = national strategic mines; SGM = slightly gassy mines; SLM = government-supported local mines; SM = small mines; TM = township mines; UCM = unclassified mines.

Figure 9 .
Figure 9. Characteristics of different types of coal mine accidents.Note: GSLM = government-supported local mines; NSM = national strategic mines; TM = township mines.

Figure 11 .
Figure 11.Classification of miners' capabilities.Note:(a)The safety awareness factor in accidents; (b)The safety psychological status factor in accidents; (c)The safety knowledge factor in accidents; (d)The safety habits factor in accidents

Figure 12 .
Figure 12.Lack of safety culture in SESGEAs.Note: SESGEAs = severe and extraordinarily severe gas explosion accidents.

Figure 15 .
Figure 15.Path A: accident path caused by electrical sparks.

Figure 16 .
Figure 16.Path B: accident path caused by gas accumulation.

Figure 17 .
Figure 17.Path C: accident path caused by management violations.
The current coal mine safety training system has several shortcomings: training content and training methods are too simple and lack a specific safety knowledge training framework; the training method is based on the enterprise's own training, and the training effect and training time cannot be guaranteed; and training content is limited to case training and skills training, and operators generally lack understanding of laws and regulations and knowledge of emergency rescue.

Table 1 .
Results of insufficient SMS in SESGEAs.

Table 2 .
Individual-level causes of gas explosion accidents in Baoma coal mine.

Table 3 .
Problems of the organizational management system of Baoma Mining Group.

Table 4 .
Problems of organizational safety culture in Baoma Mining Group.