Research On Network Model Method Of Tower Crane Accident Cause

With the increasing volume of construction projects and the more complex processes,tower cranes are widely applied in construction,and the operation frequency of them continues to increase.Due to complex structure,high-altitude operation,lifting objects with large volume and irregular shapes,coupled with the current safety risk caused by technical and management problems,tower cranes accidents prone which make great threat to project quality,personnel safety,economic property,etc.However,due to the diversity and nonlinear relationship of tower crane safety factors,its risk evolution law and accident formation mechanism are more complicated.Therefore,it is of great significance to focus on the complex system characteristics of tower crane accidents,and study the coupling and correlation problems of accident causative factors and the scientific model of the internal mechanism,which can help to propose targeted management and control measures to improve the safe operation of tower cranes.The paper firstly uses the tower crane accident investigation report as the corpus to carry out text mining.the accident causative factors are extracted by identifying and screening key textual feature,and then the text information is transformed into a structured accident causative factor set based on the spatial vector model.,to provide data support for mining factor correlation.Secondly,using the data set of causative factors of tower crane accidents as mining samples,the causal relationship and interaction between causative elements are studied by using association rule.Based on high support and high confidence,the correlation and coupling relationship between different causal factors is analyzed from three types of rules,namely binomial,three-item and four-item set rules,and the risk transfer form between factors in the process of accident is revealed,which become the data basis for subsequent accident causation network research.Thirdly,based on the association rules and lifting degree of accident causative factors,combined with complex network theory,the tower crane accident causal network model with 72 nodes and 296 edges is constructed.The accident network structure and topology are analyzed from four dimensions including key nodes,critical paths,community detection and aggregation characteristics,indicating the scale-free characteristics of the accident network and the small-world network effect,and through identifying the key causative factors at different levels,the inherent causal law and important evolution path of the accident is revealed,which provides a new perspective for the formulation of tower crane accident prevention and response measures.Finally,based on the full use of objective historical data,the results of accident correlation mining and network analysis are used to summarize the characteristics of two kind of associations between causative factors and between causative factors and accident consequences.And the core factors are taken as accident prevention entry point to propose precise security management and control strategies for each key node.Based on a large number of tower crane accident reports,combined with its complex characteristics,this paper conducts data mining and network analysis.The research method can be applied to risk identification and accurate response in tower crane safety management,which is conducive to promote the continuous optimization and improvement of tower crane safety management level.