Many factors cause the construction accidents of railway business lines and are inter-correlated. It is important to explore the correlation and causation between the causes of accidents for the prevention and control of accidents. The investigation report of a business line accident in a railway bureau from 2010 to 2022 was processed by text mining,and 51 accident cause factors were extracted. The cause factors were divided into four levels:human cause level,equipment level,environment level,and management level for hierarchical analysis based on the causal chain theory of accidents,and the cause system of railway business line construction accidents was constructed. The traditional FP-Growth algorithm was improved to mine the causation of railway business line construction accidents based on compression algorithm and differential coding,find high-support association rules and high-confidence association rules that satisfy the requirements of enhancement degree,and find the key causation association and accident causation law. Bayesian network for construction safety risk of business lines was established based on the Bayesian network theory,causative correlation relationship,and expert experiences. Combined with complex network theory,the characteristics of network node degree,clustering coefficient,and node internode number were analyzed to find the key causative factors. On this basis,the causal reasoning and fault diagnosis reasoning were used to further analyze the high-risk causes in the construction process of business lines,and the prevention and control measures were proposed from the three aspects of "civil defense,physical defense,and technical defense". The results of the case show that:improper operation of construction personnel,failure of overhead contact system caused by construction,encroachment of facilities and equipment caused by construction,damage to cable equipment caused by construction,track circuit failure caused by construction,and maintenance and cleaning equipment delay of train accidents are frequently related and are the cause of high probability of red light belt accidents. Multi-level prevention and control should be focused in the construction process. The research results provide a new risk analysis method for the construction safety management of railway business lines.
关键词
铁路营业线施工事故/改进FP-Growth算法/关联规则挖掘/贝叶斯网络推理/致因体系
Key words
railway business line construction accidents/improved FP-Growth algorithm/association rules mining/Bayesian network reasoning/causative system
维普
万方数据