数据为证:各类煤矿事故致因差异化分析与危险源管控研究

Evidence based on data:Risk analysis of accident-causing factors and hazards precautions in coal mines

为分析各类煤矿事故致因要素间的差异,本文构建了基于危险源和隐患的事故致因模型,并以事故文本数据为证据来源展开研究。首先,提取煤矿安全生产过程中人员、设备、环境和管理4类危险源的外在表现形式;其次,将TAN贝叶斯网络模型和Apriori关联算法作为证据生产方法,利用前者分析煤矿事故类型与事故致因的关系,借助后者从管理的角度找出各类煤矿事故中亟须重点防控的有关员工、设备和环境的不安全要素。研究发现:外部监管和组织氛围对瓦斯事故、水灾事故和火灾事故的影响较大,员工的决策失误和违章作业与爆破事故关联较大,安全监督人员的不作为是顶板事故和其他事故发生的主要原因,而员工自身较差的精神和生理状态可分别导致运输事故和机电事故。最后,基于上述证据,通过知识转化,本文提出了针对各类煤矿事故的具体危险源管控方案。

Coal mine safety has always been an important problem that unsettles China′s economic development.Constrained by many factors,such as traditional industrial production modes and increasing market demand,the intensity and difficulty of coal mining have increased in recent years,along with the drop in coal resource stock,resulting in increasingly complex mining conditions.These objective realities have made the coal industry one of the most serious casualties in China′s industrial production.Although the death rate per million tons in China′s coal industry is decreasing year by year,there is still a large gap between China and major coal-producing countries in the world,and major accidents still occur from time to time.The main disaster factors for the coal mine industry include gas,roof,water disaster,fire,transportation,electromechanical and blasting.Such factors are characterized by sudden,sharp and violent effects that not only have a great degree of harm and wide influence but also easily induce other serious accidents.Traditional risk management and control focuses on the investigation and treatment of hidden dangers,which aim at eliminating such dangers before accidents happen.However,according to the accident prevention effect of such methods and the occurrence mechanism of hazard sources,hidden dangers and accidents,the threshold of accident prevention needs to further be moved forward,with the prevention and control of hazard sources as the core and effective internal and external management means to eliminate risks before hidden dangers happen.In addition,considering the different relationship patterns among different types of hazard sources in coal mine accidents of different nature,the emphasis of hazard source control should be different to make preventive measures more targeted.The existing research has made a comprehensive analysis of the risk management of coal mine accident cause factors,but it ignores the integrity of four types of hazard sources,such as personnel,equipment,environment and management,and fails to consider the differences among different coal mine accident cause factors from the perspective of accident types,which leads the accident prevention measures to lack pertinence to some extent.In addition,the pre-control research of coal mine accidents is mostly based on past accident experience,and the policy suggestions put forward by researchers often regard unsafe factors as independent factors,ignoring the coupling effect between factors.At the same time,due to the lack of objectivity or extensive generalization,the quality of this research cannot be considered,so it is difficult to translate into effective decision-making information in the follow-up knowledge application process.Based on this,to provide evidence for the safety management of the coal mine industry,this paper follows the concept and method of evidence-based science,takes objective data,namely,883 investigation reports of various types of accidents from 2011 to 2020,as the evidence source,and uses the data-driven evidence production method based on reorganizing the specific causes of coal mine accidents.The Bayes risk evaluation model of the causative factors of various types of coal mine accidents is constructed based on the revised HFACS-CM framework to clarify the importance of various causative factors in different coal mine accidents.Furthermore,the hidden correlation network and occurrence regularity of coal mine accidents are obtained by using the coupling relationship between unsafe factors,which directly forms the solution to the problems while generating evidence,and provides a feasible decision-making basis for coal mine enterprise policy-makers.The results indicate that:1)the key causes of different types of coal mine accidents are different:in gas accidents,flood accidents and fire accidents,external environmental factors and the organizational atmosphere play a decisive role;decision-making mistakes and illegal operations of employees are more likely to lead to blasting accidents;and the unsafe supervision behavior of safety inspectors is the main cause of roof accidents and other accidents.Poor mental state and physiological state are the crux of frequent transportation accidents and electromechanical accidents respectively.2)The defects of external supervision and internal management are the fundamental reasons for the unsafe behavior of coal mine employees in addition to the unsafe state of machines and the environment in terms of changing from the expected state to an objective existence.Based on the key causes of all kinds of coal mine accidents,we continue to determine the internal correlation mode between the management system and other types of hazard sources,take it as evidence,and finally form specific hazard source control schemes for different coal mine accidents through knowledge transformation.