首页|塔吊作业事故关联规则挖掘及贝叶斯建模分析

塔吊作业事故关联规则挖掘及贝叶斯建模分析

扫码查看
为探究塔吊作业安全风险,深入挖掘事故特征,明确塔吊作业事故成因机制,提出了一种基于Apriori算法关联规则挖掘的改进朴素贝叶斯网络结构,采用数据驱动的方式训练模型,提升塔吊作业事故推理的效率和精度.通过诊断推理得出,塔吊倒塌事故多由物因缺陷造成,高处坠落事故多由人因缺陷造成,吊物伤人事故多由物因缺陷和不良环境共同造成.通过因果推理对诊断结果进行了验证,表明该模型可在设定条件及场景下有效进行事故推理分析.
Tower crane operation accident association rule mining and Bayesian modeling analysis
To explore the safety risk of tower crane operation,deeply explore the characteristics of accidents,and clarify the causal mechanism of tower crane operation accidents,an improved naive Bayesian network structure based on Apriori algorithm association rule mining is proposed,the model is trained in a data-driven manner to improve the efficiency and accuracy of tower crane operation accidents reasoning.According to the diagnostic reasoning to explore the causal mechanism of the accident,combined with causal reasoning to analyze the development trend of the accident and predict the probability of the accident,the internal relationship between the risk factors of tower crane operation is deeply explored,providing a new idea for reducing the probability of tower crane operation accidents.(1)Tower crane operation accident is given priority to tower crane collapse(more than 50%),followed by hanging objects injury accident(31%),high falling accident probability is small.The relevant units should be in the existing safety management system,to strengthen the prevention and emergency management of tower crane collapse accidents.(2)The results of reverse reasoning show that the direct causes of tower crane collapse accidents are mostly the unsafe state of objects,the direct causes of high falling accidents are mostly the unsafe behavior of people,and the direct causes of hanging object injury accidents are mostly the unsafe state of objects and bad environment.(3)The results of forward causal reasoning show that under the conditions of illegal operator operation,tower crane functional component damage,failure,cross-operation,and inadequate on-site safety supervision,the most likely type of accident is hanging object injury.In the human-induced defect scenario,accidents involving high falling and hanging object injury are most likely to occur.In the material defect scenario,tower crane collapse accidents are most likely to occur,while the possibility of high falling accidents is very low.In the environmental defect scenario,the most likely accident is hanging object injury,and the probability of tower crane collapse and high falling accidents is very low.Finally,in the management defect scenario,tower crane collapse accidents are most likely to occur,while the possibility of hanging object injury to people is very low.

safety engineeringtower craneApriori algorithmdata-drivennaive Bayesian networksaccident reasoning

叶勇军、张笑语、张英朋

展开 >

南华大学资源环境与安全工程学院,湖南衡阳 421001

安全工程 塔吊 Apriori算法 数据驱动 朴素贝叶斯网络 事故推理

2024

安全与环境学报
北京理工大学 中国环境科学学会 中国职业安全健康协会

安全与环境学报

CSTPCD北大核心
影响因子:0.943
ISSN:1009-6094
年,卷(期):2024.24(2)
  • 14